Background and Objective: Foodborne pathogens and cross-contamination of food products pose a serious risk to the food industry as many outbreaks are associated with biofilm formation, which increases post-processing contaminations and risks to public health. This review aimed to study the biofilm formation of spoilage and pathogenic bacteria in foods and on food contact surfaces, which subsequently represent serious challenges to the food industry and may decrease shelf life and increase transmission of diseases.

Results and Conclusion: Chemical and physical methods (e.g. sanitizing with chemicals and heat treatment) are not sufficiently applicable for biofilm removal in food sectors due to the increase of bacterial resistances, ingredient damages and possible residues in food matrix. During meat processing, the environment is filled with complex multispecies communities of microorganisms, majorly connected to the surface forming biofilms that are difficult to treat. Furthermore, bacterial cell relationships between various genera and species play a key role in the attachment process and formation of strong biofilms, as well as in the resistance of the biofilm community members against antimicrobial treatments. Thus, control of these biofilms are difficult in food industries since the biofilm cells secrete exopolymeteric substances that include preventing barrier or lessening contact with environmental stresses such as antimicrobial agents as well as the host immune system. Biofilms are highly resistant to conventional antimicrobial therapies and lead to persistent infections. Hence, there is a high need for novel strategies other than conventional antibiotic therapies to control biofilm-based infections. Bacterial biofilm formation and its problems in the food industry were discussed in this study in addition to various safety strategies aiming to provide novel insights into biofilm control in the food industry for improving food quality and safety.

Conflict of interest: The authors declare no conflict of interest.

1. Introduction

A biofilm is a complex community of microorganisms that adhere to a surface, forming multiple layers that protect their growth, proliferation and survival [1]. It can lead to antibiotic resistances, nosocomial infections and food-borne illnesses. However, biofilms benefit the microbes by helping them in adhesion, metabolite exchange, quorum sensing and drug resistance [2]. Microbial biofilms are composed of diverse bacteria surrounded by their exopolysaccharides and typically attached to biotic and abiotic surfaces, resulting in food poisoning with diarrhea, vomiting, enteritis, stomach discomfort and headaches in humans [3, 4]. Presence of biofilms in food-processing environments, food contact surfaces, processing equipment such as stainless steel, rubber, plastic and Teflon and completed products increases danger of spoiling, diminishes shelf life and increases possibilities of infectious disease outbreaks associated with foods [5]. Search for efficient ways to control microbes and their biofilms still needs further efforts [6]. This review covered topics associated with particular microorganisms that create biofilms in the food sector, illustrating the biofilm formation process, stages of development, interactions between microorganisms and various novel methods and strategies of biocontrol. 

1. Results and Discussion
2. Biofilm Formation

Several factors affect biofilm formation, including metabolism, signaling molecules, culture media, matrix and variations in cellular and genetic makeup [7]. Generally, biofilm formation consists of four common steps (Figure 1), initially produced on biotic or abiotic surfaces through reversible and irreversible adhesion, using adhesive proteins, lipopolysaccharides, flagella and pili [8]. Furthermore, biofilm maturation occurs in two stages of cell-to-cell communication and production of auto-inducer molecules. These molecules primarily consist of proteins, exopolysaccharides, DNA, RNA, enzymes, microbial cells and water with water being the major component responsible for nutrient movement within the biofilm matrix. Exopolysaccharides serve as a protective shield, enhancing microbial adhesion within biofilms, ensuring their structural integrity and facilitating nutrient acquisition [9,10,11].

1.1 Formation of Biofilm in Food Industry

In the food industry, surfaces and equipment that come into contact with foods are often occupied by microorganisms that can form biofilms [12,13]. Bacterial biofilms in foods pose severe hazards to human health, leading to systemic diseases, food intoxication and gastroenteritis and presence of bacterial biofilms on tables, staff gloves, animal carcasses, water, milk and other liquid pipelines has been documented [14].

1.2 Biofilm Resistance to Antimicrobial Agents 

Bacteria living in biofilms show 10 to 1000-fold increases in drug resistance, compared to their planktonic stages. Various multidrug-resistant (MDR) bacteria such as Salmonella spp., methicillin-resistant Staphylococcus aureus (MRSA), Listeria monocytogenes, Campylobacter jejuni, Escherichia coli O157:H7 and vancomycin-resistant enterococci (VRE) have been linked to foodborne outbreaks, presenting a significant public health threat [15]. Based on the estimates, biofilm matrix and EPS prevent bacteria from antibiotic exposure, providing them an adaptive advantage by preventing chemical stressors from penetrating deeper biofilm regions [16]. In biofilms, quorum sensing and horizontal gene transfer are the most commonly observed mechanisms [17]. Biofilm awareness in fight against MDR bacteria needs further discussion and persistence of biofilms in foods creates an ideal environment for resistance mechanism exchange; hence, greater awareness of these dangers is necessary.

1.3 Quorum Sensing System

Quorum sensing system is a communication system between the cells that allows them to send chemical signals, enabling cooperative gene expression, which leads to increased population density, enhanced biofilm formation and increased production of extracellular polymeric substances [18]. Recent studies have shown that Gram-negative bacteria produce acylated homoserin lactones (AHLs) as autoinducers, while Gram-positive bacteria use peptides (AIPs) [19].

1.4 Horizontal Gene Transfer

Horizontal gene transfer is a widely recognized mechanism; through which, bacteria adapt and spread resistance to antimicrobial agents using mobile genetic elements (MGEs), which pose a significant threat to global public health [20]. Release and transfer of bacterial DNA play a role in biofilm synthesis and contribute to the spread of antibiotic resistance. Resistance plasmids can spread through the conjugation process, promoting development of resistant biofilms within the densely populated structure of biofilms. Over time, drug resistance leads to the preferential expression of certain genes, resulting in increased productions of proteins associated with virulence and antibiotic resistance, which can alter characteristics of biofilm resistance.

 1.5 Common Foodborne Pathogens Forming Biofilms

Bacterial pathogens can contaminate meats because meats are rich in vitamins, minerals and proteins and include high water contents (75%) and acceptable pH ranges [21, 22]. The greatest dangers to food safety worldwide are environmentally hazardous microorganisms that can infect cattle during various processing procedures and foodborne illnesses associated with uncooked meats. These pollutants are challenging to clean off and disinfect, putting customers' health at major risks. A variety of bacterial pathogens can cause meat-borne diseases, by infecting animals or contaminating meat during meat processing such as Salmonella spp., E. coli, Campylobacter spp., L. monocytogenes, Yersinia enterocolitica, Brucella spp., Mycobacterium bovis, Bacillus anthracis and toxin-producing Staphylococcus aureus, Clostridium spp. and B. cereus [23]. Meat-borne diseases can be categorized into infections, intoxications, allergies, metabolic food disorders and idiosyncratic illnesses [24]. Harmful bacteria can build up on various equipment and biotic and abiotic surfaces and eventually create biofilms whereas over 90% of bacteria live. They create biofilms on gloves and surfaces of silicon, rubber plastic, glass and stainless steel [25]. Significance and effects of biofilms on the food industry have been demonstrated in several studies, where a variety of pathogens such as L. monocytogenes, Y. enterocolitica, C. jejuni, B. cereus and E. coli O157:H7 frequently cross-contaminate these food products [26]. Available studies have shown that the coexistence of multiple bacterial species could increase biofilm development and enhance pathogen persistence by promoting EPS production. Examples of these relevant biofilm-forming pathogens in the food industry are briefly described.

1.5.1 Gram-negative Bacteria

Approximately 80% of the available foods in Saudi Arabian markets are imported with 15.71% of these imports are meat-based. Escherichia coli, Salmonella spp. and Pseudomonas aeruginosa include several important virulence factors, form biofilms and easily contaminate meats. However, handling and consuming animal-derived products contaminated with E. coli biofilms can pose health risks while Shiga toxin-producing E. coli and Enterohaemorrhagic E. coli are important enteric pathogens linked to outbreaks and severe gastroenteritis. Verotoxigenic E. coli produce verotoxins while E. coli O157:H7 is a human pathogen responsible for outbreaks of bloody diarrhea and hemolytic uremic syndrome (HUS) and can be transmitted through raw milks, drinking waters, fresh meats and vegetables. A major element affecting production of E. coli biofilms is temperature. For example, after 7 d of incubation at 15 °C, quantity of adhering and planktonic cells increased on beef surfaces, which included a serious issue for meat processing plants [27]. Isolates with a higher capacity for mature biofilms showed resistance to sanitization [28].  

Salmonella spp. propagate at 35–37 °C and includes two species of S. bongori and S. enterica, the most prevalent pathogens in the food industry and the causative agent in several foodborne outbreaks [29, 30]. Salmonella enterica is commonly associated with refrigerated poultry products stored on shelves during food processing or in supermarkets. Three various types of Salmonella are important for human health, including non-typhoid Salmonella, S. typhi and S. paratyphi. Fresh poultry and meat are highly prone due to their nutrient-rich content, high water activity and near-neutral pH (5.5–6.5), creating optimal environmental conditions for Salmonella spp., which are not spore-formers and can easily be destroyed by heat at 60 °C for 15–20 min. Furthermore, growth of most isolates was inhibited below 7 °C and pH 4.5, while nontyphoidal salmonellosis in the US is nearly 1.35 million illnesses per year [31]. Because many people reside in Saudi Arabia during haj    and umrah seasons, a significant prevalence of Salmonella infection occurs [32].  Pseudomonas aeruginosa is wide spread on meat surfaces and in low-acid dairy products and affecting more than 2 million individuals and killing roughly 90,000 of them annually [33]. Because of its adaptability, P. aeruginosa may grow at temperatures lower than 7°C and contaminate fresh meat sold in stores, causing its spoilage via lipolytic, saccharolytic and proteolytic processes [34]. In addition, the microorganism secretes extracellular enzymes that cause breakdown of foods and includes a high degree of medication resistance, which can result in serious acute and chronic infections in immunocompromised people. Human infections usually affect the respiratory tract (RT), soft tissues, blood vessels, urinary tract (UT) and wounds [35]. Carbapenem-resistant strains of P. aeruginosa pose a hazard to public health [36]. Due to the abundance of EPS, cells can adhere to stainless-steel surfaces and create biofilms alone or with other pathogens and produce multispecies biofilms, increasing their stability and resistance [37].

1.5.2 Gram-positive bacteria

Listeria monocytogenes is a rod-shaped, non-spore-forming, facultative-anaerobic Gram-positive bacterium. It causes human infections of listeriosis, a serious illness that includes septicemia and meningitis, particularly in immunocompromised individuals and is capable of growing at temperatures ranging 3–45 °C with the optimal temperature of 30–37 °C [38]. Listeria monocytogenes is a harmful foodborne microorganism that is killed by pasteurization. Consumption of dairy products, meats, fishes, fruits, soft cheeses, ice creams and poultries has been linked to listeriosis epidemics [39]. It can form biofilms on surfaces commonly detected in the food industry and is resistant to treatments with heat up to 60 °C [40]. It can thrive in a broad range of conditions, including high salinities (10%), cold temperatures (4 °C), low water activities (< 0.9) and wide pH ranges (4.1–9.6) [41]. Post-processing contamination with Listeria spp. may be resulted from inadequate cleaning and poor separation techniques between ready-to-eat and raw foods [42]. In addition, L. monocytogenes is one of the most significant pathogenic microorganisms due to its high mortality rates (15.6%) and one of the major causes of hospitalizations and deaths in the US [43].

Staphylococcus aureus is responsible for staphylococcal food poisoning (SFP) and produces enterotoxins within the temperature range of 10–46 °C. Staphylococcus genus includes more than 50 recognized species; of which, S. aureus is commonly detected in food products and reaches foods through raw materials and grows best on meat, poultry and egg products. In food production chain, it may develop biofilms on living and non-living surfaces, resist desiccation and thrive on a variety of surfaces. Furthermore, strains of S. aureus that produce enterotoxins have been identified in a variety of food samples. [44]. Other Gram-positive bacteria such as Brochothrix thermosphacta and Carnobacterium spp. can form biofilms in the meat-processing environment [45].

1.6 Strategies for Controlling Biofilm Formation in the Food Industry

Pathogenic bacteria that form biofilms create strong defenses against antibiotics and are difficult to treat. Removing these biofilms is a critical challenge due to the severe effects on public health [46, 47]. Chemical and physical methods have been used to inhibit bacterial biofilms in the food industry. Chemical treatments can help; however, mechanical treatments such as clean-in-place are not effective. The most reliable way to prevent bacterial biofilm growth is through aseptic processing, routine disinfection and equipment sterilization. Various disinfectants and novel biofilm elimination methods of the food industry are briefly summarized (Figure 2).

1.6.1 Chemical and Physical Treatments

Biofilms can be treated with concentration and time-dependent chemical sanitizers. Decreasing bacterial populations to human-safe levels is the goal of sanitation. Sanitizing food-processing equipment is necessary to avoid cross-contamination between batches of foods. Stages of general cleaning methods for places that handle and manufacture foods include physical pre-cleaning, detergent washing, rinsing, sanitation, final rinsing and drying. Spraying detergents in form of foam or aerosol spray is possible as long as the right doses and time are used for surface contact. Alkaline and acidic chemicals are widely used as detergents in the food industry. A majority of disinfectants are safe to use on non-food-contact surfaces; nevertheless, food-contact and occasionally non-contact surfaces should be rinsed with high-quality water. The most popular sanitizer in the food industry is aqueous ClO₂, which acts well against B. cereus endospores in biofilms on steel surfaces [48]. In the food industry, chlorine-based sanitizers are most frequently used; nevertheless, several microorganisms have developed resistance to chlorine treatments. Food factories frequently use sodium hypochlorite or NaOCl [49]. Moreover, hydrogen peroxide (H₂O₂) and NaClO were successful in removing biofilms of S. aureus and P. aeruginosa; however, aqueous ClO₂ was more effective than NaOCl in eliminating E. coli O₁₅₇:H₇ biofilms [50]. In the food industry, H₂O₂ is a powerful oxidizing disinfectant that is often used. When it is exposed to biofilms, H₂O₂ produces free radicals that kill the bacteria at concentrations of 0.08–5%, without harmful side effects. Quaternary ammonium compounds are frequently used as sanitizers, removing biofilms and leading to bacterial lysis [51]. Steam heat treatment is a method used to decrease number of harmful bacteria and biofilm populations in production areas [52]. Non-thermal plasma is a partially ionized gas with low temperature and promising antibacterial characteristics. It can destroy bacterial biofilms of Pseudomonas spp., S. enterica and Bacillus spp. Ozone breaks down the cellular envelopes of a variety of microorganisms, including viruses, bacterial biofilms and protozoans.

1.6.2 Elimination of Biofilms Using Biological Strategies

In recent years, a more efficient and ecologically friendly control method for the elimination of or managing growth of dangerous biofilms is use of enzymes, bacteriophages, bacteriocin and plant extracts, which have been discussed based on safe and green approaches to control pathogen biofilm formation.

1.6.2.1 Enzyme against bacterial Biofilm

Enzymes or proteins are biologically active macromolecules against biofilm formation since proteases or other degrading enzymes have shown the ability to inhibit biofilm formation [53].   Enzymes are detected to include therapeutic functions in removal of pathogenic biofilms and can widely be used in detergents of food industries. In recent times, a variety of enzymes enriched products have been commercialized that include tablets, rinsing solutions, chewing gums for dental treatments and denitrifies containing enzymes such as lysins, dextranase, mutants that can serve to play an effective role in disintegration of the biofilm matrix [54]. The most often used enzyme types vary depending on the makeup of the biofilm as proteases, cellulases, polysaccharide depolymerases, alginate lyases and dispersin B [55]. Proteinase K and lysozyme have been verified to include promising antibiofilm activities [56]. The α-amylase enzyme includes potential to operate as an antibiofilm agent against bacterial species that produce biofilms, including S. aureus and P. aeruginosa [57]. Protease formulations were effective in eliminating S. aureus biofilms from polystyrene surfaces; however, combinations of protease, amylase and cellulase were needed to eradicate biofilms of P. aeruginosa. Cellulase effectively inhibited biomass and microcolony formation by P. aeruginosa on glass surfaces in partial [58].

1.6.2.2. Bacteriophages against Biofilm

Bacteriophages (phages) are bacterial viruses, acknowledged as the most diverse and abundant entities. Bacteriophages are mostly used in primary production to ensure food safety, biosanitization and biopreservation [59]. Phages can break down biofilms spread through developed biofilms and then show their antimicrobial characteristics inside them. Phage treatments are injected directly into food products during the biopreservation processes to extend the food shelf life and used in biosanitization to avoid biofilms on equipment surfaces [60]. Bacteriophages can create enzymes that break down the biofilm structure and presence of phage receptor sites such as endolysins and depolymerases. Use of phages as biocontrol agents in foods is affected by various factors, including the food matrix, surface area and structure, bacterial species, inhibitory compound and phage dose [61]. A commercial product, LISTEXTM, has been developed from the bacteriophage P100, which uses an enzymatic process to cause cell lysis and EPS breakdown. The US Department of Agriculture (USDA) has approved use of this natural, non-toxic phage product. It is effective against L. monocytogenes. Additionally, it seems that L. monocytogenes biofilms are susceptible to phage biocontrol. Phage Guard Listex, which uses phage P100, effectively removes biofilms from stainless steel surfaces. A user of Listeria phage P100 (under the commercial name of Listex P100) is a biological agent, formed to remove the biofilms in processed meat products [62]. In addition, a phage cocktail was used for 1 h to destroy and decrease pathogen populations of E. coli O₁₅₇:H₇ on stainless steels, ceramic tiles and high-density polyethylene coupons [63]. Endolysin is the second kind of enzyme produced by phages that include potential uses for sanitization. During the final stage of their lytic cycle, they release progeny of virions through the breakdown of the cell wall, which were active against Gram-positive bacteria [64]. Depolymerases are types of enzyme that may prevent production of biofilms and break down capsular polysaccharides in Gram-negative bacteria [65].

1.6.2.3 Bacteriocins against biofilms

Lactic acid bacteria (LAB) are used to produce fermented foods and the most important genera in controlling spoilage and pathogenic microbes are Lactobacillus and Bifidobacterium due to the production of bacteriocins and acids. Bacteriocins from LAB are used as alternatives to chemical food preservatives. They can spread through cell membranes and release internal components such as K⁺ and inorganic phosphate or they can prevent production of proteins, RNA and DNA [66]. Due to its safety in the gastrointestinal system, bacteriocin has extensively been used as a food preservative in the food sector for several years. Use of bacteriocin in biopreservation systems can meet consumers' demands and numerous compounds, including nisin, natamycin, subtilin, pediocin, tylosin and carnocyclin A, are used as food preservatives [67].

1.6.2.4. Plant Extracts against Bacterial Biofilm

Numerous substances from plants such as complex mixtures of monoterpenoids, plant-based essential oils, sesquiterpenoids and flavonoids have shown anti-biofilm characteristics [68, 69]. Previous materials can be used as an alternative to synthetic preservatives. Specifically, several flavonoids inhibited generation of bacterial toxins in various food products. In addition, bacterial cell adherence to stainless steel was strongly inhibited by essential oils and other plant components that are abundant in almost all plants such as phenolic chemicals, tannins, terpenoids, glucosinolates derivatives, alkaloids and thiols. Although the food industry uses a variety of plant-based extracts and essential oils in meat preservation, pomegranate and cranberry extracts are particularly popular due to their antibacterial and antifungal characteristics [70, 71, 72]. 

1. Conclusion

Controlling biofilm formation in the food industry is essential for food safety, quality and hygiene. Traditional methods such as cleaning and sanitizing with chemicals and heat treatment are critical; however, effectiveness was limited due to the increase of antimicrobial resistance and vitamin damage by heat. Moreover, biofilm cells secrete extracellular polymeric substances that include a barrier preventing or lessening contact with environmental stresses and decreasing effects of antimicrobial agents and the host immune system. Various methods have been investigated to prevent and remove biofilms, each offering unique advantages. However effectiveness of traditional methods was limited, biological control methods such as bacteriophages and probiotics offer promising sustainable solutions and may be further effective in specific settings due to targeting harmful pathogens without affecting the environment. Recent advances in combination of physical, chemical and biological methods in food processing environments may be effective keys to biocontrol biofilms, ensuring safety and quality of food products.

1. Conflict of Interest

The author reports no conflict of interest.

1. Authors Contributions

H.A, NZ and MMA conceptualized the idea and prepared the manuscript.

Background and Objective: Microalgae are rich sources of bioactive metabolites and one of the major focuses of the pharmaceutical industry is the use of secondary metabolites from plant sources. Chlorella vulgaris, a microalga with high economic values, includes a high protein content and significant bioactive compounds and polysaccharides. Therefore, this microalga can be used as a dietary supplement and medicinal product. In this study, inhibition of the growth of colon cancer cells was investigated.

Material and Methods: Proteins of Chlorella vulgaris were extracted using enzymatic hydrolysis using proteolytic enzymes of pepsin and Promod (Bacillus subtilis protease). Separation of the peptides was carried out using ultrafiltration techniques. Cytotoxic effects of the extracted peptides were assessed using MTT assay on mouse colon tumor cell lines (CT-26).

Results and Conclusion: Results indicated that the pepsin protein hydrolysates (Pep1, Pep2 and Pep3) at a concentration of 1000 mg.ml^-1 decreased the viability of the CT-26 colon cancer cell line by 24.34%, 36.00% and 40.08%, respectively, while the Promod protein hydrolysates (Pro1, Pro2 and Pro3) decreased the viability by 26.26, 35.91 and 37.13%, respectively. The Pep1 and Pro1 showed the highest cytotoxicity effects (P < 0.05). Findings of this study suggest that the bioactive peptides present in C. vulgaris may include beneficial functional compounds for cancer prevention.

Background and Objective: Chemical treatments in chitin extraction from shrimp shell wastes have affected the environment. Shrimp shell primarily bonds chitin with inorganic salts, lipids, proteins and pigments. Extraction of chitin from shrimp shells involves protein separation processes. Deproteinization process of chitin from dried shrimp (Litopenaeus vannamei) shells with papain enzyme was optimized and nanochitin as a derivative product of chitin was characterized.

Material and Methods: Effect of hydrolysis time, temperature and enzyme concentration were optimized using RSM Box-Behnken method to maximize chitin yields. Nanochitin was prepared using dialysis and ultrasonic methods and characterized for physical characteristics using scanning electron microscope, particle size analysis and Fourier transforms infrared spectroscopy.

Results and Conclusion: Optimum conditions using enzymatic hydrolysis at 6 h, 50 ^oC and 1.25% papain decreased the protein content from 33.66 to 2.31% and produced a high chitin yield (46.03%). Deproteinization using enzymatic hydrolysis method was more efficient than that using fermentation. Data of scanning electron microscope, particle size analysis and Fourier transforms infrared spectroscopy showed that the characteristics of chitin and nanochitin products were similar to those of chemical treatments for chitin products.

Conflict of interest: The authors declare no conflict of interest.

1. Introduction

Litopenaeus vannamei is one of the shrimp species that includes high commercial values and produces abundant shrimp shell wastes. Production of shell wastes from crustaceans was predicted to be 3.14 million metric tons per year worldwide [1]. Hundreds of shellfish wastes are generated from seafood manufacturing and daily Asian consumption [2]. Shell wastes from crustaceans contain a high quantity of chitin, a polysaccharide material that is important in biological functions and is biodegradable and compatible. Chitin and its derivatives are used in various fields such as pharmaceutical, food, textile and waste water-treatment industries [3]. Chitin in the shrimp shells is bonded with majorly inorganic salts, calcium carbonate, proteins, lipids and pigments. Therefore, isolation of chitin from shrimp shells involves protein separation processes and mineral separation [4]. Structure of chitin is arranged with N-acetylated glucosamine and glucosamine units, linked by β(1,4) covalent bonds. Corresponding to this structure, chitin is stable to chemical and biological actions and the linkage of chitin is similar to the linkage of cellulose [5]. Generally, chitin is extracted through demin-eralization using acid treatment and deproteinization using alkali treatment. These treatments affect the environment and finding an alternative process that is more friendly to the environment is still necessary. Deproteinization process for chitin extraction from shrimp shells can be carried out via chemical, enzymatic and microbial processes [6, 2]. The chemical treatment involves mineral acid at high temperatures, resulting in high volumes of polluted waste containing mineral acids in the washing process. These treatments are harmful to the environment due to high concentrations of mineral acids [7]. Deproteinization with enzymes is a zero waste system resulting in high yields of chitin products. Protease hydrolyzes proteins in the matrix efficiently [8]. Commercially purified enzymes such as alcalase, papain, pepsin and trypsin have been used in chitin extraction studies to remove protein from crustacean shells [9].

 Chitin is a biopolymer containing microfibrillar and semicrystalline structures. Based on data of the infrared (IR) spectra and X-ray crystallography (XRD), chitin is naturally in the forms of α-chitin, β-chitin and ɣ-chitin. Characteristics of chitin such as solubility, porosity and surface area restrict its uses. To solve this problem, various derivatives such as chitosan, chitin nanofibers and chitin nanowhiskers are produced [10]. Chitin nanofibers have been prepared via several methods such as ultrasonication, mechanical treatment, gelation and electrospinning [11]. Chitin nanofibers from species such as crabs, prawns and mushrooms have been prepared using mechanical and chemical treatments. The acidic medium was verified in the decrease of chitin nanofibers extracted from crab shells [12]. Under certain extraction conditions, chitin microfibrils are isolated in the form of nanocrystals and nanofibers. Their unique characteristics have been studied and used in food, cosmetics and medical industries [13]. Characteristics of chitin depend on the organisms and chitins may lay in α and β allomorphs shapes. These forms were assessed by the orientation of microfibrils that could be characterized using infrared, nuclear magnetic resonance (NMR) spectroscopy and XRD analysis [14]. Probiotic microorganisms have been studied for the demineralization treatment of crustacean shells. Chitin extraction using microorganisms was carried out simultaneously. Shrimp shells (Penaeus monodon) were fermented with lactic acid bacteria (LAB) and chitin was separated by adding carbohydrates [10]. Based on XRD and NMR data, chitin extraction via enzymatic process is an alternative method to preserve its native structure [1]. Shelma et al. reported the chitin nanofiber preparation via acid hydrolysis of the chitin powder followed by dialysis and ultrasonication [15]. Chitin from P. vannamae byproducts was prepared by associating enzymatic acid-alkaline strategies to achieve further sustainable processes [16]. Moreover, chitosan was produced through papain extract to help deproteinization process. Papain is achieved from the papaya plant with the endopeptidase, dipeptidase and exopeptidase activities. The optimum condition of this process was at 7 h of enzymatic hydrolysis and 25% of papain [8]. The current study was aimed to optimize deproteinization process of chitin from dried shrimp (L. vannamei) shells using low concentration papain (0.75–1.25%) and to achieve nanochitin, which was prepared via dialysis and ultrasonic methods. Furthermore, nanochitin products were characterized through physicochemical characteristics to verify their quality.

1. Materials and Methods

2.1. Materials

              Dried white-shrimp (L. vannamei) shells were provided as byproducts of a shrimp processing industry at Muara Gading City Bekasi, West Java, Indonesia. Commercial papain (CAS no. 2323.627-2) (Xian Arisun ChemParm, Shaanxi, China) was purchased in powder form. All chemicals used included laboratory grades.

2.2. Chitin extraction from the shrimp shells

Chitin from the sample was extracted using method of Hongkulsup et al. [1] with some modification. The extraction was carried out at two steps, including demineralization and deproteinization. In demineralization process, shrimp shells were ground to achieve a size of 100 mesh. Shrimp shell powder was extracted using 1.5 M HCl (ratio 1:10, w/v) at 25 ^oC for 6 h and at 150 rpm. Mixture was filtered using vacuum filter and the residue was mixed with distilled water (DW) to achieve neutral pH. Then, residue was dried at 50 ^oC for 6 h. Dried residue was mixed with 0.75–1.25% w/v papain in a phosphate buffer pH 7 and heated at 40–50 ^oC for 3–6 h. Hydrolysis was stopped at 90 ^oC and set for 20 min. Mixture was filtered and the residue was mixed with DW until neutral pH was achieved. Then, residue was dried at 50 ^oC for 6 h. Total residue was assessed gravimetrically and the soluble proein content in the residue was analyzed using modified Lowry method. Briefly, 1 g of residue was diluted with DW up to 1 ml and filtered using Whatman filter papers. Then, 0.5 ml filtrate was mixed with 5.5 ml of alkaline CuSO₄ reagent and incubated at room temperature (RT) for 10 min. Solution was mixed with 0.5 ml of folin phenol reagent. Then, sample solution was mixed with 3.5 ml of DW and the absorbance was measured at 650 nm. The protein soluble content was assessed by plotting bovine serum albumin (BSA) standard curve [17].

2.3. Optimization of deproteinization of shrimp shells using Box-Behnken method

               Optimum condition of the deproteinization process was predicted using response surface methodology (RSM)- Box-Behnken method. Optimization of deproteinization was carried out using three factors of effects of hydrolysis time, temperature and enzyme concentration (Table 1). Proportions of total residue, chitin and protein concentration were used as the responses data. Fifteen trials were carried out indiscriminately. The center value

was chosen based on the references, which were 1% papain, 45 ^oC and 6 h [18, 19]. Design Expert 13.0 software was used in this study.

2.4. Assessment of chitin

              Chitin content was assessed using adaptation of the Morrow method [20] with some modification. One gram of the sample was mixed with 40 ml of 1 M HCI and mixed at RT for 2 h. Chitin residue was separated using vacuum filter with a porous sintered glass disc and washed several times with water to reach a neutral pH. The residue was washed off and transferred into a beaker containing 40 ml of 5% NaOH and stirred at 100 °C for 2 h. Chitin product was separated using filter paper (Whatman no. 41, USA) and then rinsed with water until a neutral pH was achieved. Content of the chitin (%) was assessed gravimetrically.

2.5. Nanochitin preparation

              The selected chitin sample, which was prepared at optimized conditions, was soaked in 3 M HCl for 90 min at

90  ^oC. Suspension was precipitated by centrifugation at 6000 rpm for 10 min. Nanochitin from the precipitated fraction was prepared for dialysis and ultrasonic treatments using Mincea method [11] with modifications. Suspension of chitin was transferred to a dialysis bag (cellulose membrane with cut-off proteins mol. wt ≥ 12,000) and dialyzed in DW by changing the water every 2 h for three times. Dialysis was carried out until pH 6 was reached. Ultrasonic treatment of the chitin sample was carried out at pulse of 1/1 and amplitude of 60% (750 W, 20 kHz) for 6 h to 0.1% (w/v) of the suspension. Based on the modification of Wu and Meredith method [21], these samples were freeze-dried at -60 ^oC for 10 h.

2.6. Microstructure identification

              Microstructure of the freeze-dried samples was assessed using scanning electron microscope (SEM) (JSM-IT30, Jeol., Akhishima, Tokyo, Japan). These samples were put in a sample holder and layered with a thin layer of gold (±10 nm). Observation was carried out by accelerating voltage at 20 kV based on a previous method.

2.7. Particle size distribution

              Particle size distribution of the samples was analyzed using particle size analyzer (Zetasizer Nano ZS Malvern,UK)  based on Shelma et al. method [15] with modifications. Sample was dispersed in Tween 80 (0,4%; w/v) with a ratio of 1:4.

2.8. Fourier transforms infrared spectroscopy (FTIR)

              Spectra of the samples were analyzed using Fourier transforms infrared spectroscopy (FTIR 1000, Perkin-Elmer, USA) at mild conditions and method of KBr pellet scanning. Based on previous studies, KBr (100 mg) and the sample (1 mg) were mixed entirely until KBr pellet was formed. Then, samples were scanned at spectral ranges of 400, 4200 and 4200 cm-¹.

1. Results and Discussion

3.1. Optimization of deproteinization of the shrimp shells

              The optimum conditions of the enzymatic hydrolysis in the deproteinization process of white shrimp shell powder were predicted using RSM. Fifteen trials were carried out based on the RSM-Box Behnken design. The Box–Behnken design (BBD) is a widely used RSM design that is useful for ascertaining cause-and-effect correlations between factors and responses in experiments. The BBD needs three levels and can be used for factors of 3–21 [22]. Hydrolysis factors and their responses are provided in Table 1. Data showed that the total residue of the products ranged 74.14–80.76%, chitin content ranged 41.52–49.06% and protein content ranged 2.31–6.82%. Analysis of variances (ANOVA) was calculated and p-values of the total residue, soluble protein and chitin content are present in Table 2. Papain concentration (C) and its interaction with temperature (AC) and hydrolysis time (BC) significantly (p < 0.05) affected the total residue of shrimp shell powder. The hydrolysis time (A) and its interaction with the papain concentration (AC) significantly (p < 0.05) affected the chitin content. However, p-values of the soluble protein contents showed that treatments were not significant (p > 0.05). The equation for estimating the optimal condition for all responses (Y1, Y2 and Y3) from the shrimp shells is present in Table 3. Total residue included the yield of the dried product after the deproteinization process with the papain enzyme. Chitin extraction via enzymatic hydrolysis needs removing proteins from the crustacean shells, minimizing the deacetylation and depolymerization processes. This process may be carried out before or after the demineralization step of solid materials for accessibility of the reactants. Efficiency of the enzymatic treatments is inferior to chemical methods ranging 5–10% of the residual protein attached to chitin [9]. Commercial enzymes such as alcalase, econase, pancreatin and other proteases were used in the chitin extraction of shrimp and crustacean shells. The objective of these treatments was to eliminate the protein contained in the waste of shells. Proportion of the chitin ranged 16.5–22% [7]. Combination of the chemical agents and enzymes has been studied to increase yields of the chitin products. Use of sodium sulfite and alcalase was the best treatment for protein recovery. Characteristics of the chitin sample were similar to those of the commercial food-grade products [6].

Three-dimensional (3D) response surfaces of the response; of which, one of the factors is fixed at the central point and the other is varied, are present in Figure 1. The highest predicted chitin content is indicated by the surface confined in the smallest ellipse in two-dimensional (2D) contour plots. This indication was correlated with the interaction between hydrolysis time and papain concentration significantly. This was similar to the results of ANOVA analysis (Table 2). The 2D contour plots showed effects of hydrolysis time (A) in the chitin content prediction (Fig. 1c). However, stagnation was observed in the chitin content with increasing temperature (Fig. 1b). To achieve the optimum condition of the deproteinization process, an optimization process was analyzed using Design Expert 13.0 RSM optimizer software. The three factors (time, temperature and papain concentration) were adjusted in the importance level 3 (+++) and responses (total residue, soluble protein and chitin yield) were adjusted in the importance level 5 (+++++). The optimum condition with desirability of 0.619 was observed for chitin extraction at 6 h, 50 ^oC and 1.25% papain. Further, all the responses of the products were validated through laboratory experiments. The experimental and predicted values are present in Table 4. Data showed that the experimental and predicted values were in the range (95% prediction interval); thus, reliability of the optimized condition was verified. The RSM-Box Behnken design was successfully used to assess effects of hydrolysis time, temperature and papain concentration on deproteinization process to produce higher chitin contents. Chitin from the molted shrimp shells was extracted using a chemical method. The optimum condition of deproteinization was achieved in 3% NaOH at 50 ^oC for 6 h with a residual protein content ≤ 1% [23]. Yulirohyami et al. (2024) reported that chitosan was prepared through processes, including depigmentation, demineralization, deproteiniza-tion and deacetylation. The optimum condition of depro-teinization process was reached at 25% of papain for 7 h of hydrolysis. This study showed that the hydrolysis time of chitin deproteinization affected deacetylation degrees of chitosan [8].

              Proteases have been used for chitin extraction from shrimp byproducts. Residual proteins in shrimp wastes included 1.3 and 2.8% after treatment with chymotrypsin and papain enzymes. Combination of papain with other proteases that was used for deproteinization of shrimp wastes showed that the protein removal rates were low [24]. As an alternative to chemicals and decreasing shrimp wastes, 0.2% alcalase was shown to include activities in decreasing protein contents in shrimp wastes from 49.43 to 4.12% [25]. The enzymatic deproteinization of shrimp processing wastes has limited chitin yields nealy  4.4 to 7.9% of the total weight. This might be due to the residual of short peptides appropriately bonded to the compound of chitin. Use of combination agents with protease significantly decreased the protein fraction. Through this combination, protein fraction significantly decreased, assuming that protease degraded disulfide bonds of the shrimp head waste proteins that facilitated entry of sulfite ions [6]. Use of exoenzymes and proteolytic bacteria in deproteinization of demineralized shells produced liquid protein and solid chitin fractions [2]. Papain is a commercial enzyme, which includes endopeptidase, dipeptidase and exopeptidase activities. Binding affinity and catalytic efficiency of papain are affected by the substrate, temperature and incubation time [8].

3.2 Physicochemical characterization of Chitin

Characteristics of chitin, including degree of deacetylation, morphology and molecular mass, vary depending on the extraction method and origin of chitin [14]. For example, chitin achieved by the chemical extraction showed a tightly packed morphology, while a slightly microfibrillar structure was shown by chitins extracted via enzyme treatment [1]. Use of chitin increased significantly due to the prominent characteristics of its derivatives and nanostructure configuration, which are met for industrial processing. Techniques have been developed to produce chitin derivatives. For example, dialysis and ultrasonic methods  to produce nanochitin from shrimp shells; similar to those of the present study. Surface morphologies of the prepared chitin and nanochitin are present in Figure 2. Accordingly, porous-like honeycomb structure with no nanofibers on the surfaces was observed in the chitins (Figure 2a) and nanochitin achieved via dialysis process (Figure 2b). The only difference between these two products was in the pore size as the pore width of chitin (3 μm ±5) was smaller than that of nanochitin (5–15 μm). This result indicated that during the acid hydrolysis process of nanochitin preparation, the amorphous part of chitin was removed, leaving the crystalline side and leading to increases in pore size. The nanochitin generated from the ultrasonic technique (Figure 2c) showed a nanofibrillar structure with a diameter of nearly 160 nm. This reveals that hydrolysis with a strong acid followed by the ultrasonication treatment further facilitated dissolution process of the amorphous chitin [16]. Ultrasonication is a method to change the natural cithin into chitin nanofibers. Fibrillating chitin at 900–1000 W and 20 kHz in water (pH ±7) created nanofiber widths of 25–120 nm. High frequency of ultrasonication induced startling waves on the chitin surface that promoted their factorization with the axial way [26].

 Chitin is naturally detected in crystalline microfibrils as a structural component, serving as a functional material that is needed by many organisms [14]. The pH of a solution in chitin treatment affects the surface morphology of chitin nanostructures, as a previous study demonstrated that the nanofiber structures of chitin were destroyed to small irregular shapes under high alkaline environments [23]. Furthermore, chitin nanofibrous structure formed due to chitin nanofibers are not soluble and result in versatile porous structures of the products by adjusting the freezing temperature. Freeze-drying technique includes the potential for the assembly of the nanofibrous structure of water-dispersible materials [21].

Particle size distribution is an important characteristic that affects functionality of the chitin products. The chitin sample included two peaks in the spectra, which were dissolved in 0.4% of Tween 80 solution (Fig. 3a). The Z-average of chitin from the shrimp shells was 511.7 nm and the highest intensity was 21.8%. Moreover, nanochitin samples showed three peaks with Z-averages of 101.7 (Fig. 3b) and 345.4 nm (Fig. 3c), respectively. Nanochitin produced via dialysis method showed a Z-average of the particles smaller than that produced by the ultrasonic method. However, the intensity of nanochitin products was still lower than that of untreated chitin samples. Particle size distribution of the chitin nanofibers demonstrated a bimodal curve with majority sizes of 20–300 nm [15]. In this study, additional peaks in nanochitin products were assumed as degraded chitin products. Temperature of the experiments affected number of the peaks in spectra. For higher temperatures, large particles were observed, which might be caused by degradation of the chitin particles. The lower temperature of ionic liquids was further favorable, resulting in a narrow particle range of particle size distribution spectra [27]. Ionic liquids could change the chitin structure, able to modify the particle size [28].

The FTIR spectrum of chitin is present in Figure 4. Chitin sample showed similar spectra with nanochitin, which was prepared via dialysis and ultrasonic methods. Spectra at 3258 and 2924 cm-¹ were recognized as N-H and C-H stretching vibrations. The amide I band was distributed into two peaks of 1652 and 1621 cm-¹. Absorption at 1557cm-¹ was assigned to N-H bend and C-N to 1310 cm-¹. Peaks at 1069 and 1010 cm-¹ were recognized as C-O stretching. Chitin from the shrimp shells has been extracted via two-step extraction using citric acids and deep eutectic solvents (DESs). The study showed that spectra of DESs-extracted chitin included N-H stretching, which was limited by the bonds of intermolecular hydrogen  and the bonding of NH groups. Band of amide I was generated by bonding between the intra-chain hydrogen with the NH groups and the bonding between inter-chain hydrogen with the primary OH [29]. Moreover, two absorption bands at 1018 and 1172 cm-¹ were identified for C-O stretching vibration of chitin from snail shells [30]. Two-step fermentation method was used for demineralization and deproteinization of chitin extraction from shrimp shell powder. The FTIR spectra of the samples showed characteristic peaks corresponding to the amide I (1652 and 1620 cm-¹) and amide II (1554 cm-¹) regions. Peaks at 1375 and 950–1200 cm^-1 were C-H, C-O-C and C-O bonding [31]. Nanochitin was produced using microwave method to observe the difference of α and β structures in amide I. Two bands at 1654 and 1621 cm superscript were assigned as single H-bonded and double H-bonded α structures. A unique single band at 1631 cm^-1 was assigned as β structure of nanochitin [13]. Structure of α chitin is known stable than chitin due to strong hydrogen bonding in inter and intra-sheets [32].

Based on the characteristic data of nanocithin from white shrimp shells (L. vannamei), the product is potential for creating nanochitin-based materials. Morphological and chemical characteristics, including helical and its structure, encourage material developments. Nanochitin is a promising product as the support material at various dimensional aspects [26]. Chitin nanofiber includes potential uses such as in biomedical and biodegradable materials and waste treatments. Limitations of nanochitin production include high-energy demands, high catalyst costs and unstable yields [33]. Thus, further studies should focus on improving yields of chitin and assessing nanochitin uses in biomedical functions.

1. Conclusion

This study assessed effects of the deproteinization process of chitin from dried shrimp shells using enzymatic hydrolysis as an alternative method of fermentation and chemical processes. In this study, deproteinization process was optimized using RSM-Box Behnken design to maximize the chitin yield. Results of this study represented the optimum condition of chitin deproteinization process from dried shrimp shells using papain enzyme, with a chitin content ranging 41.52–49.06%. Interactions between the hydrolysis time and papain concentration included the most significant effect on the chitin content. Removing protein content in chitin extraction through protease

enzymes was verified as a further efficient process, compared to the fermentation process. Use of chitin in industries needs specific characteristics to meet the industries' needs. The SEM analysis showed that acid hydrolysis affected surface morphology of the chitin nanostructure and ultrasonication treatment demonstrated nanofibrillar structure of the chitin. Nanochitin products included similar spectra with chitin samples, which included typical groups of the chitin structure. The present study indicates that alternative method of chitin production may decrease effects of chemical residues on the environment.

1. Acknowledgements

This study was supported technically and financially by the Research Center for Applied Microbiology National Research and Innovation Agency, Republic of Indonesia.

1. Conflict of Interest

The authors report no conflicts of interest.

1. Authors Contributions

Conceptualization and Methodology, SP and DZA; Investigation: DAW, WK and DR; writing—original draft preparation: SP, IP; review and editing: CR, IP.

1. Using Artificial Intelligent chatbot

The authors declare no artificial intelligent chatbot use.

References

1. Hongkulsup C, Khutoryanskiy V, Niranjan K. Enzyme assisted extraction of chitin from shrimp shells (Litopen-aeus vannamei). J Chem Technol Biotechnol. 2016; 91: 1250-1256. https://doi.org/10.1002/jctb.4714
2. Arbia W, Arbia L, Adour L, Amrane A. Chitin extraction from crustacean shells using biological methods – A Review. Food Technol Biotechnol. 2013; 51: 12-25.
3. Kumari S, Rath P, Sri Hari Kumar A. Chitosan from shrimp shell (Crangon crangon) and fish scales (Labeorohita): Extraction and characterization. African J Biotechnol. 2016; 15: 1258-1268. https://doi.org/10.5897/AJB2015.15138
4. Kurniasih M, Dwiasi DW. Preparation and characterization of chitin from white shrimp shells (Litophenaeus vannamei) Molekul. 2007; 2: 79-87. https://doi.org/10.3390/md15050141
5. Berezina N. Production and application of chitin. Phys Sci Rev. 2016; 1-8. https://doi.org/10.1515/psr-2016-0048
6. Mizani AM, Aminlari BMA. New process for deproteinization of chitin from shrimp head waste. Proc Eur Congr Chem Eng. 2007; 1-8.
7. Gortari MC, Hours RA. Biotechnological processes for chitin recovery out of crustacean waste : A biotechnol-ogical processes for chitin recovery out of crustacean waste: A mini-review. Electron J Biotechnol. 2013; 16: 1-14. https://doi.org/10.2225/vol16-issue3-fulltext-10
8. Yulirohyami, Banowati R, Puspasari I, Risky A, Hidayat H. Papain assisted preparation of chitosan : Feasibility study for candidate in-house reference material and antibacterial agent. Food Humanit. 2024; 2: 100313. https://doi:10.1016/j.foohum.2024.100313
9. Younes I, Rinaudo M. Chitin and chitosan preparation from marine sources. Structure, characteristics and applications. Mar Drugs. 2015; 13: 1133-1174. http://dx.doi.org/10.3390/md13031133

10 Rameshthangam P, Solairaj D, Arunachalam PR. Chitin and chitinases: Biomedical and environmental applications of chitin and its derivatives. Enzym. 2018; 1: 24. https://doi.org/10.14302/issn.2690-4829.jen-18-2043

1. Mincea M, Negrulescu A Ostafe V. Preparation, modification and applications of chitin nanowhiskers : a Review. Rev Adv Mater Sci. 2012; 30: 225-242.
2. Ifuku S, Shervani Z, Saimoto H. Chitin nanofibers, preparations and applications. Adv Nanofibers 2013; 3: 85-100. http://dx.doi.org/10.1039/c2nr30383c
3. Marin RF, Ramos FH, Salberria AM andres MA, Labidi J, F. S. Eco-friendly isolation and characterization of nanochitin from different origins by microwave irradiation: Optimization using response surface methodology. Int J Biol Macromol. 2021; 186: 218-226. http://dx.doi.org/10.1016/j.ijbiomac.2021.07.048
4. Elieh D, Komi A. Chitinhgh and Chitosan : Production and application of versatile biomedical nanomaterials. Int J Adv Res. 2016; 4: 411-427.
5. Shelma R, Paul W, Shalma CP. Chitin nanofibre reinforced thin chitosan films for wound healing application. Trends Biomater. Artif Organs. 2008; 22(2): 111-115.
6. Vazquez JA, Ramos P, Mirón J, Valcarcel J, Carmen GS, Martín RIP. Production of chitin from Penaeus vannamei by-products to pilot plant scale using a combination of enzymatic and chemical processes and subsequent optimization of the chemical production of chitosan by response surface methodology. Mar Drugs. 2017; 15: 2-14. http://doi.org/10.3390/md15060180
7. AOAC: Official Methods of Analysis. Association of Official Analytical Chemists, Washington. 1990.
8. Annisa S, Sastro Y, Amalia U. The effect of various fish species on fish protein hydrolysate with the addition of papain enzyme. Indonesia. J Fish Sci Tech. 2017; 13: 24-30.
9. Priatni S, Harimadi K, Buana E, Kosasih W, Rohmatus-solihat. Production and characterization of spray-dried swamp eel (Monopterus albus) protein hydrolysate prepared by papain. Sains Malaysiana. 2020; 49: 545-552. http://dx.doi.org/10.17576/jsm-2020-4903-09
10. Morrow JC. Utilisation of nephrops shell waste. (Glasgow University, 2002).
11. Wu J, Meredith JC. Assembly of chitin nanofibers into porous biomimetic structures via freeze drying. ACS Macro Lett. 2014; 3: 185-190. http://dx.doi.org/10.1021/mz400543f
12. Al-tayawi AN, Horváth ZH, László Z. Methods for experimental design , central composite design and the Box -Behnken design to optimize operational parameters : A review. Acta Aliment. 2023; 52: 521-537. http://dx.doi.org/10.1556/066.2023.00235
13. Phuong PTD, Trung TS, Stevens WF, Minh NC, Bao HND, Hoa NV. Valorization of heavy waste of modern intensive shrimp farming as a potential source for chitin and chitosan production. Waste Biomass Valor. 2022; 13: 823-830. http://dx.doi.org/10.1007/s12649-021-01557-0
14. Kim Y, Park RD. Progress in bioextraction processes of chitin from crustacean biowastes. J Korean Soc Appl Biol Chem. 2015; 58: 545-554. http://dx.doi.org/10.1007/s13765-015-0080-4
15. Minh TLT, Truc TT, Osako K. The effect of deprotein-ization methods on the characteristics of glucosamine hydrochloride from shells of white leg shrimp (Litopenaeus vannamei) and black tiger shrimp (Penaeus monodon). Ciência Rural St Maria. 2022; 52: 1–6. http://doi.org/10.1590/0103-8478cr20200723
16. Bai L, Liu L, Esquivel M, Blaise LT, Huan S, Niu X, Liu S, Yang G, Fan Y, Rojas RJ. Nanochitin : Chemistry, structure, assembly and applications. Chem Rev. 2022; 122: 11604-11674. http://dx.doi:10.1021/acs.chemrev.2c00125
17. Jaworska MM, Gorak A. New ionic liquids for modification of chitin particles. Res Chem Intermed. 2018; 44: 4841–4854. https://doi.org/10.1007/s11164-018-3337-9
18. Jaworska MM, Górak A, Zdunek J. Modification of chitin particles with ionic liquids containing ethyl substituent in a cation. Hindawi, Adv Mater Sci Eng. 2017. https://doi.org/10.1155/2017/3961318
19. Zhao D, Huang W, Guo N, Zhang S, Xue C. Two-step separation of chitin from shrimp shells using citric acid and deep eutectic solvents with the assistance of microwave. Polymers (Basel). 2019; 11: 409. https://doi.org/10.3390/polym11030409
20. Oyekunle DT, Omoleye JA. Effect of particle sizes on the kinetics of demineralization of snail shell for chitin synthesis using acetic acid. Heliyon. 2019; 5. https://doi.org/10.1016/j.heliyon.2019.e02828
21. Liu Y, Xing R, Yang H, Liu S, Qin Y, Li K, Yu H, Li P. Chitin extraction from shrim (Litopenaeus vannamae) shells by successive two-step fermentation with Lactobac-illus rhamnoides and Bacillus amyloliquefaciens. Int J Biol Macromol. 2020; 148: 424-433. http://dx.doi.org/10.1016/j.ijbiomac.2020.01.124
22. Yihun FA. Nanochitin preparation and its application in polymer nanocomposites: a review. Emergent Mater. 2022; 5: 2031-2060. http://dx.doi:10.1007/s42247-022-00427-2
23. Du X, Wen J. Preparation and application prospects of nano chitin. E3S Web Conf. 2024; 02014: 1-4. https://doi.org/10.1051/e3sconf/202455302014

Background and Objective: The aim of this study was to increase the viable cell and biomass production of a potential probiotic strain, Lactiplantibacillus plantarum DLBSK207, by optimizing the ideal concentrations of key nutrients and fermentation conditions parameters using statistical method. such as (RSM) with Box-Behnken Design (BBD).

Material and Methods: The experiments investigated two key variables for medium composition and fermentation conditions. Based on the OFAT result, six factors were selected for the Plackett-Burman Design to evaluate whether the variables had significant effects to the response. The medium contains carbon (glucose) and nitrogen sources (yeast extract and peptone), while the fermentation conditions include initial pH and temperature. The basal medium, consisting of sodium acetate, MgSO₄ 7H₂O, K₂HPO₄, MnSO₄ H₂O, and Tween 80, was kept constant. Using RSM, the concentrations of glucose, yeast extract, and peptone, as well as the initial pH and temperature, were optimized to maximize viable cell counts and biomass.

Results and Conclusion: The optimum medium concentrations determined by RSM were 33.76 g l^˗1 glucose, 32.59 g l^˗1 yeast extract, and 28.38 g l^˗1peptone at an initial pH of 6.0 and a temperature of 35 °C. Under these optimized conditions, this study achieved a viable cell counts of 9.30 log CFU.ml^˗1 and a dry cell weight of 4.319 g l^˗1, representing a 1.82-fold increase compared to standard MRS broth. The experimental results were in closely matched the predicted values of 9.30 log CFU.ml^˗1 and 4.280 g l^˗1. Scaling up the process in a 10-l bioreactor controlled at pH 6.0 resulted in even higher biomass production, reaching a maximum viable cell counts of 9.88 log CFU.ml^˗1 and a dry cell weight of 5.819 g l^˗1after 20 h of incubation.  

Conflict of interest: The authors declare no conflict of interest.

Abstract

Background and Objective: Although biochemically similar, two laccase enzymes isolated from basidiomycetes (Trametes sp.) showed differences in their affinity to two types of phenolic compounds, interacting stronger with diphenols (catechol and caffeic acid), compared to interactions with benzenetriols (pyrogallol and gallic acid). Catalytic efficiency of Trametes pubescens laccase was detected 4-5 times higher than determined for commercial laccase (Trametes versicolor). In this study, the interactions of the two immobilized enzymes with di and triphenols were examined by various electrochemical techniques.

Material and Methods: Following electrochemical techniques: cyclic voltammetry, chronoamperometry and differential pulse voltammetry were used in this study. Experiments were carried out in varying substrate concentrations. Activity and sensitivity of the two alternative laccase – based biosensors were compared using DPV and chronoamperometry.

Results and Conclusion: Constant potential amperometric measurements indicated that the biosensor produced with Trametes pubescens laccase was much more active than biosensor based on laccase from Trametes versicolor when interacting with caffeic and gallic acids. The phenolic content of three different herbal extracts was evaluated with the developed laccase biosensors and results were found to be similar to those from chromatographic analysis used as a reference method. Therefore, biosensors can be used for rapid testing of phenolic content in real samples.

1. Introduction

Daily intake of antioxidant-rich foods and drinks is considered an important factor of a healthy regimen [1]. Antioxidants, the physiological role of which is to scavenge reactive oxygen species thus preventing oxidative damage of living cells [2] are micro-ingredients of plant cells such as those of fruits, vegetables, cereals and herbs. From natural antioxidants, phenolic acids are especially important not only because their antioxidant activity is within the highest ones, but also because they can rapidly be digested and adsorbed by gastrointestinal tract [3]. Due to their exceptional pharmacological, nutritional and wellbeing effects on humans, a wide spectrum of analytical methods for the assay of phenolic acids has been developed. Chromatographic analytical techniques receive extensive application for the quantification of these antioxidants in a variety of food samples [4-6]; however the required time-consuming sample preparation procedures stimulated the advance of various optical methods such as visible spectroscopy and fluorometry [7], or even paper-based colorimetric sensors [8]. Because of the susceptibility of phenolic acids to participate in redox processes, a range of electroanalytical methods such as voltammetry [9,10], pulse voltammetry [11-15] and amperometric detection [9,10,16,17] have been developed.

Modern electrochemical approaches for the quantify-cation of phenolic acids include phenolic acid detection with electrodes modified with advanced materials such as polymers [4], nitrogen doped carbon [12], carbon nanotubes [18,19] and gold nanostructures [10]. The use of nanostruc-tured materials or composites for electrode modification offers enhanced selectivity of the determination due to either pronounced electrocatalytic effect or greatly enhanced electrode surface area [10].

Assessment of antioxidant capacity through biosensing is a novel trend in contemporary studies [9,11,20-22]. Numerous authors report an improved selectivity of the analysis when using biosensing systems for the assessment of antioxidant content [23]. Copper –containing enzymes laccase or tyrosinase were the primary choice for developing biosensing systems for the analysis of phenols [24-26]; however, whole-cells [27] and DNA-based [28] elements for molecular recognition have also been used for this purpose. Due to the formation of colored products of biocatalyzed transformations of phenolic compounds, most of the highlighted biosensing platforms rely on optical detection principle.

Biosensors with electrochemical detection for phenolic antioxidants analysis, reported in current literature [24], require nanostructured electrode surfaces and sophisticated bioreceptor immobilization protocol. Unlike these, a simple enzyme attachment to an unmodified glassy carbon electrode is discussed, which ensures electrochemical response sensitive enough to guarantee phenolic acid assay at micromolar concentrations. Therefore, the focus of the present study was to develop and optimize an electrochemical method based on two identically prepared laccase biosensors for the determination of di- and trihydroxy aromatic compounds – one bearing commercial laccase from Trametes versicolor, and another laccase – isolated from T. pubescens, which is used seldom in biosensor development. Laccases are widely used in a variety of industrial cycles such as pulp and paper production, wastewater treatment (e.g. from olive-oil mills and textile industry), brewing and food industries, pharmacy or in the construction of fuel cells [25] and biosensors for the quantification of di-substituted aromatic compounds [29].

Laccases are complex enzymes with more than one active site, which embeds three copper clusters (type T1, T2 and T3) differing in both function, and spectroscopic character-istics [25,26]. The reaction mechanism of laccases involves uptake of one electron from the substrate- a hydrogen donor, which is oxidizing to form radical with concomitant 4-electron reduction of molecular oxygen to form two water molecules [26]. The T1 copper site is responsible for binding the aromatic substrate to be oxidized via 1-electron pathway, while the T2-T3 copper cluster binds molecular oxygen and catalyzes its 4-electron reduction to water [26]. Electrons are transferred from T1 site to T2-T3 trinuclear cluster through internal molecular electron transfer. The difference between laccases from T. versicolor and T pubescens is linked to their amino acid sequences. Despite these belong to the family of fungal laccases, variations are seen in their primary structures that may affect their catalytic efficiency, substrate binding specificity, and thermal stability. Thus, amino acids surrounding the less conservative T1 copper-binding site and the overall folding of the enzymes might differ partially, which may result in differences in catalytic efficiency of the laccase manifested as substantially different catalytic constants.

Here reported biosensors function on the following principle: laccases electrocatalytically reduce the dissolved in the working medium oxygen, thus generating reductive current, which is enhanced in the presence of phenol derivatives that act as electron shuttles (mediators) between the enzyme active site and electrode surface. Gallic and caffeic acids are two representatives of phenolic acids, the positive effects of which are commented not only in terms of their antioxidant abilities, but also with respect to their potential anti-inflammatory [30] and anticancer [31] pharmacological activity that was the major reason for their use as laccase substrates in this study. As a demonstration of the applicability of biosensing method, a series of three different types rich in phenolic compounds herbal extracts were analyzed for their phenolic content and results were compared with those from high-performance liquid chromatography (HPLC) analysis.

1. Materials and Methods

2.1. Reagents

Laccase (Е.С. 1.10.3.2, polyphenol oxidoreductase) enzymes from T. versicolor (Fluka, USA) and T. pubescens (a generous gift from Prof. Roland Ludwig, Department of Food Science and Technology, BOKU University of Natural Resources and Life Science, Vienna, Austria) were with homogeneous specific enzyme activities of 21 and 46 U mg^-1, respectively. One unit is the amount of enzyme necessary for the oxidation of 1.0 μmol of ABTS (2,2-azino-bis(3-ethylbenzothiazoline-6-sulfonic acid)) per min at pH 4 and 30 °C. Laccases were dissolved in 0.05 M sodium-citrate buffer, pH 4, in such quantity to form enzyme solutions with concentrations of 950 U.ml^-1. The two enzymes were used without further purification.

Catechol, resorcinol, pyrogallol, caffeic acid, gallic acid, ABTS and reagents for the preparation of buffer solutions (sodium citrate, citric acid monohydrate and NaClO₄) were of analytical grade (Acros, Belgium) and used without further purification. All stock solutions of the enzyme substrates were prepared with a concentration of 10 mM.

2.2. Enzyme immobilization

Enzyme electrodes were prepared based on commercial glassy carbon electrodes (2-mm diameter; Metrohm, Utrecht, The Netherlands). Prior to modification, electrodes were polished with 0.05 µm alumina slurry on a polishing cloth (Kulzer, Hanau, Germany), water-rinsed and cleaned using ultrasonication in ultrapure water for 1 min for at least two consecutive times.

Enzyme immobilization was carried out as follows [29]: 2 ml of enzyme solution were drop-cast on the electrode surface. Then, a 4-µl drop of the binder (Nafion 117 diluted with ultrapure water to 0.2%) was applied. Surface was dried at room temperature (RT). The two types of laccases were immobilized on the electrode surfaces identically and the amount of the immobilized enzyme in terms of enzyme units was equal.

After electrochemical measurements, enzyme electrodes were rinsed with ultrapure water and refrigerated at 4°C, when not in use. Regeneration of the working enzyme electrodes could be carried out after the mechanical removal of the enzyme-polymer layer via polishing procedure and following the above steps.

2.3. Electrochemical measurements

All electrochemical experiments were performed in a conventional single compartment three-electrode cell with working volume of 10 ml, connected to a computer-controlled electrochemical workstation Autolab PGSTAT 302 N (Metrohm-Autolab, Utrecht, The Netherlands) controlled by NOVA 2.1.6 software. Either a modified with enzyme glassy carbon electrode, or an enzyme– free electrode (for control experiments) was used as working electrode. A Ag|AgCl, sat. KCl (Metrohm, Utrecht, The Netherlands) was the reference and a platinum foil was the auxiliary electrode. If not otherwise specified, all reported potentials were stated against this reference electrode (Ag|AgCl, sat. KCl electrode) [29]. Cyclic voltammetry was run at scan rates of 5–20 mV.s^-1. Volt-ampere curves (voltammograms) were obtained in both background electrolyte - 10 ml of citrate buffer (pH 4, containing 0.1 M NaClO₄) and in the presence of enzyme substrates with a stock concentration of 10 mM until a 30 mM concentration was achieved in the cell.

Amperometric detection has been carried out at a constant potential of -0.2 V through successive additions of aliquots of 10 mM substrate stock solutions (typically from 20 to 500 ml) to 10 ml of the electrolyte in the cell. Chronoamperometric detection was carried out under constant stirring at 500 rpm.

Differential pulse voltammograms were recorded both in the absence and presence of studied compounds over the potential range from +0.6 to –0.6 V at a scan rate of 10 mV s^-1, pulse duration of 50 ms and an amplitude of 0.025 V, as optimized in the authors’ previous studies [29,34].

Data analysis was implemented with Origin Pro 8.5 software. Non-linear regressions and statistics were carried out using embedded software module for enzyme kinetics.

2.4. Herbal extracts preparation

Herbal extracts were prepared as follows: 20 g of dry herbal mixtures were added to 1 l of pure water and boiled for 20 min at atmospheric pressure (~101 kPa). Then, the herbal extract was set to cool down to RT, filtered through nylon cloth 6.6, packed in 50 ml sealed containers and refrigerated at 4˚C until analyses. Three types of herbal extracts were subjected to electrochemical and HPLC analyses, further referred as PM1, PM3 and PM7. The three types of herbal extracts consisted of following herbs:

PM1: Geranium sanguineous, Arctostaphylos uva-ursi, Betula alba, Polygonum hydropiper, Achillea millefolium;

PM3: Crataegus monogyna, Equisetum arvense, Geranium sanguineum, Urtica dioica; and

PM7: Fragaria vesca, Hypericum perforatum, Calendula arvensis, Frangula alnus, Polygonum hydropiper.

These herbal combinations were selected due to their use as healing teas (pharmaceutical products) with high antioxidant content.

2.5. Chromatographic analysis

The HPLC analysis has been carried out as a referent method for phenolic acid quantification, as follows: The phenolic acid composition of the extracts was assessed using chromatographic system (Shimadzu, Japan), which consist of auto sampler (Nexera X2, SIL- 30AC); CTO-20AC column and SPD-20A UV detector (Shimadzu, Japan). Analysis was carried out using column Metitaranea Sea RP-18e (150 mm × 4.6 mm × 2 μm) (Teknokrom, Spain), mobile phase of 4% acetic acid and 100% AcCN (80:20), flow rate of 0.65 ml.min^-1, λ = 280 nm and temperature of 35 ˚C. Results were analyzed using Lab-Solution Nexera-XR-RF software and the standard phenolic acids (gallic and caffeic acids) [32,33].

The content of the phenolic compounds was calculated against a standard line constructed with its solutions at concentrations ranging from 500 µg.ml^-1 to 2.5 µg.ml^-1 of the corresponding phenolic acid with a correlation coefficient of R² > 0.9991. Total phenolic content was calculated by summing the content of each determined phenolic compound and re-calculating the total phenolic content in equivalents of gallic acid.

1. Results and Discussion

3.1. Studies of two different laccases in di – and trihydroxy aromatic compounds present by cyclic voltammetry (CV)

Electrochemical behavior of the two types of laccase-based bioelectrodes was probed using cyclic voltammetry (CV) in the absence and presence of both laccase substrates – oxygen and phenolic compounds. Voltammetric studies have shown that for the two types of laccase biosensors, reductive wave starts at potentials more negative than -0.25 V in aerated buffer solutions (i.e. in the presence of oxygen), which was not seen in deaerated solutions. It is well known that laccase is a metalloprotein capable of exchanging electrons with underlying electrode surfaces directly [34] without the need for additional electron shuttles (mediators). The efficiency of the electrical communication between the electrode and laccase depends on the enzyme orientation and distance between its active site and electrode surface [34]. Most probably, the negatively charged Nafion membrane electrostatically repulsed the negatively charged laccase active site, this way orienting the enzyme to electrode surface. The latter conformation is favorable for the electron exchange with the underlying electrode, which was manifested by a reductive wave appearing on the CV in the presence of molecular oxygen. Therefore, voltammetric studies verified the ability of immobilized laccase to carry out bioelectrocatalytic O₂ reduction to water molecules, thus proving that enzymes were electrochemically active.

Comparison of the CVs of the enzyme electrode recorded in aerated solutions in the absence and presence of pyrogallol and catechol as substrates (Fig. 1) revealed laccase-catalyzed oxidation of the two phenols to semi-quinones, followed by electrochemical regeneration of the oxidized products. When resorcinol was tested as laccase substrate, the resulting voltammograms did not show interactions between either of the immobilized enzymes, as no reduction of the product of its enzyme-catalyzed oxidation was noticed (Fig. S1, Supplementary information). These phenolic compound-depending differences in the performance of the two types of laccase biosensors were due to the difference between catechol and resorcinol in their spatial structure. The first benzenediol is with two vicinal hydroxy-groups, while resorcinol is its meta-isomer.

Caffeic and gallic acids could be considered as phenolic compounds derived from catechol and pyrogallol, respect-ively. Their structural similarities with dihydroxyl and trihydroxyl aromatic compounds, as well as the fact that they are the usual constituents of polyphenolic complex in various natural products, motivated further interest in probing the voltammetric behavior of the produced biosensors in the presence of the two phenolic acids. On the CVs recorded in the absence of either phenolic compound (Fig. 2, black lines), a reductive wave was recorded with potentials more negative than -0.3 V that resulted from the electrochemical reduction of the dissolved molecular oxygen catalyzed by the immobilized enzyme, undoubtedly verifying that the two laccases were not only electrochemically, but also catalytically active.

The CVs recorded in the presence of either gallic or caffeic acid (Fig. 2, red lines), showed a clearly expressed reductive wave starting much earlier below +0.1 V, which was due to the fact that the two phenolic acids mediated the electrochemical reduction of dissolved oxygen and therefore significantly decreased the overpotential of the oxygen reduction on laccase-bearing electrodes. As seen from the presented plots, the interactions of the two laccases with the two phenolic acids resembled the shapes of the voltammograms recorded in the presence of catechol and pyrogallol. It is noteworthy that the efficiency of the enzyme interaction with the two phenolic acids is different being much higher in the presence of caffeic acid as it could be deduced from the pronounced reductive waves (Fig. 2C, D). Reaction with lower intensity between either of the two laccases with gallic acid was possibly resulting from electrostatic repulsion of its anionic form generated at the operating pH 4.0 [35] and the electrode surface, which also bore negative charges due to the coverage with a Nafion film.

3.2. Differential pulse voltamperometric response of laccase-based bioelectrodes to gallic and caffeic acids using laccases from Trametes versicolor and Trametes pubescens

Differential pulse voltammograms (DPV) of same laccase electrodes are depicted in Figure 3 (Fig. 3A, B; dashed curves). No peaks were identified on the DPVs of laccase-bearing electrodes in background electrolyte and the sharp current decay at potentials more negative than -0.4 V confirmed that oxygen reduction reaction occurred on the bioelectrode’s surface. To investigate further the voltammetric behavior of immobilized laccase in the presence of the two phenolic acids – gallic and caffeic acids, the differential pulse voltammograms were recorded at varied substrates’ concentrations. The addition of gallic acid aliquots to the buffer followed by the record of resulting DPV (Fig. 3A) caused a significant increase in the current with a peak at -0.2 V on the voltammograms, the height of which increased with increasing gallic acid concentration.

No shift of the peak position was reported upon raising its concentration. Similarly, in the presence of caffeic acid (Fig. 3B), a clearly expressed peak at a more positive potential than the one for gallic acid was recorded, the height of which increased proportionally to substrate concentration. The difference in the behavior of the laccase biosensor in the presence of caffeic acid as an enzyme substrate was that the reductive peak occurred at a potential of +0.2 V and the peak potential slightly shifted positively with increasing substrate concentration. It is plausible that the penetration of gallic acid was hampered by the deprotonation of its carboxylic group at the working pH due to the electrostatic repulsion between the negatively charged Nafion membrane and the gallic acid anionic form, resulting in a significant shift of the reduction potential to more negative values than those of caffeic acid. The latter was not deprotonated at the operating pH of the media [36] and hence its molecules penetrated the membrane easier. A similar behavior of the second laccase from T. pubescens was seen under equivalent experimental conditions.

A Michaelis type dependence between the DPV peak height and caffeic acid concentration was observed over a range from 0.01 up to 1 mM (Fig. 4A). Enzyme inhibition by the substrate of T. pubescens laccase became obvious at concentrations exceeding 0.5 mM, while enzyme isolated from T. versicolor seemed unaffected by substrate inhibition even at 1 mM concentration. Under equivalent experimental conditions, the dependence of the DPV peak current on gallic acid concentration (Fig. 4B) was based on Michaelis type kinetics only for T. versicolor laccase, while differential response of the biosensor based on the enzyme from T. pubescens decayed sharply in the presence of trihydroxy aromatic compound (Fig. 4B, red).

Despite the substantial differences between the DP voltammograms and the large peak separations recorded in the presence of either gallic or caffeic acid, DPV studies performed in combinations of the two phenolic acids did not allow discrimination between di- and triphenols. On DPVs recorded in 1:1 mixture of gallic and caffeic acids (Fig. S2, Supplementary information), the peak at -0.2 V appearing in the presence of gallic acid alone merged with the one typical for caffeic acid and the peak potential shifted negatively. With increasing the gallic acid quota up to 10 times, the two peaks broadened and turned into humps, the position and height of which varied irregularly with increasing the concentration of the mixture.

 Information from DPV studies clearly demonstrated that this electrochemical technique could hardly be used for analysis of mixtures of the studied di- and triphenolic compounds, which motivated further investigating an alternative electrochemical approach that could potentially be further useful for the analysis of complex mixtures as amperometric detection.

3.3. Amperometric detection of gallic and caffeic acids with laccase-based biosensors using laccases from Trametes versicolor and Trametes pubescens

In Fig. 5 are depicted the dependencies of the electrode response on the concentration of gallic and caffeic acids assessed with the two types of laccase-based electrodes under the working conditions selected as optimal: working potential of -0.2 V at pH 4.0 [29]. Similarities in the shapes of the curves were substantial. The two types of electrodes showed hyperbolic trends of electrode response as a function of substrate concentration. However, the apparent kinetic constants for the two immobilized laccases, determined from non-linear regression analysis of experimental data, showed significant differences. The apparent Michaelis constants for immobilized T. versicolor laccase (Fig. 5, A) with respect to caffeic and gallic acids have been found to be very similar (Table 1), while the corresponding apparent maximum rates of the enzyme catalyzed reaction differed significantly with the  for the caffeic acid being almost 5 times higher than that for gallic acid.

Unlike these, for the immobilized laccase from T. pubescens (Fig. 5B) with respect to caffeic and gallic acids differed more than 3 times (0.19 mM ±0.03 against 0.06 mM ±0.04, respectively, Table 1) whereas the apparent maximum reaction rates diverged more than those estimated for T. versicolor laccase. These differences in the apparent kinetic constants of the two identically immobilized enzymes suggested that despite biochemical similarities of the two laccases, significant differences were reported in enzyme affinity to dihydroxy and trihydroxy aromatic compounds. Since the reaction rate was equal to the electrode response (current, in amperes [A]), significant differences in  for the two laccases (Table 1) indicated that T. pubescens laccase respectively oxidized caffeic and gallic acids 5 and 4 times as fast as that from T. versicolor. Moreover, these differences in the heterogeneous enzymatic activities were observed regardless of equivalent units of the immobilized laccases.

In all cases except one, significant inhibition effects by the substrate were seen at concentrations exceeding 0.1 mM with inhibition constants calculated by non-linear regression (Table 1). Only T. versicolor laccase showed no inhibition upon addition of gallic acid at concentrations exceeding 0.15 mM.

From the non-linear regression analysis of the kinetic curves, it was found that the electrode response obeyed the Eq.1:

                                                                                                                      Eq.1

Where,  is the apparent maximum rate of the enzyme-catalyzed reaction, A;

 are the apparent Michaelis constant and inhibition constant, respectively, M;

C is the concentration of the respective phenolic acid, M;

I is the electrode response directly proportional to the rate of enzyme-catalyzed reaction, A.

All electrochemical measurements were done in triplicate with a RSD not exceeding 3.5 %.

3.4. Cyclic voltammetry analysis of phenolic acids in herbal extracts using laccases from Trametes versicolor and Trametes pubescens

The available three types of herbal extracts were tested for their redox behavior by performing cyclic voltammetry with the produced two types of laccase-based enzyme electrodes in the operating buffer; to which, equal volumes of each extract were added. Shapes of the resulting CVs (Figs. S3 and S4, Supplementary information) suggested that in the extract of type PM1, triphenols predominated while the other two types of herbal extracts contained a combination of di- and trihydroxy aromatic compounds. It is noticeable that the biosensor based on laccase from T. pubescens was responding ca. three times more intensely than the analogous one based on laccase from T. versicolor.

Quantitative analysis of herbal extracts has been carried out by means of constant potential amperometry (amperometric detection) using standard addition method. Method of standard additions (MSA) is typically carried out by adding small volumes of concentrated solution of the analyte to the sample [37]. The major advantage of the standard addition method is the opportunity to practically eliminate effects of the complex matrix in the real samples. As an external standard, 10 mM solutions of gallic acid were chosen and the MSA was implemented with the two discussed types of laccase biosensors. The one produced based on the laccase from T. pubescens (Fig. 6) guaranteed at least ten times more intense biosensor response to the same amount of PM1 herbal extract, compared to the signal of the T. versicolor laccase-based biosensor (Fig. 7). The latter finding led to greater RSD and significantly overestimated levels of the analyzed compounds; from which, it could be concluded that the second type of biosensor was not appropriate for further analyses. Based on the latter finding, the content of phenolic acids of the other two types of herbal extracts was analyzed using T. pubescens laccase-based biosensor, method of standard addition and gallic acid as an external standard.

The outcomes from biosensor analysis were compared with those from HPLC analyses of the three types of extracts (Table 2). Quantities of the phenolic compounds, estimated from the MSA, were further multiplied by dilution factor and the resulted values were recalculated into gallic acid equivalents per grams of dry weight herbs (DW). Based on HPLC analysis of the phenolic acids, the studied herbal extracts included gallic, caffeic, chlorogenic, p-coumaric and trans-ferulic acids.

As concluded from the present results, the recovery percentage was good only for the type PM1 herbal extract while for the other two extracts biosensor method showed significant deviations from the satisfactory recovery percentages (95-105%). This might be due to the presence of significant quantities of m-benzenediols (e.g. resorcinol and its derivatives) and/or other types of polyphenols, which did not react with laccase.

The two electroanalytical techniques, DPV and constant potential amperometry, were used to find the most convenient approach for the assessment of phenolic content in model solutions to adopt it for the analysis of real samples comprising numerous interferents. However, the hypothesis that the two-model phenolic compounds, benzenediol caffeic acid and benzenetriol gallic acid, could be discriminated in their mixtures based on their DPV peaks that were not well-separated, was not verified due to the strong interferences between them when combined. The constant potential amperometry was shown as a better technique for the quantification of benzenediols and benzenetriols. Large differences in the rates of their electrochemical conversion did not provide means for their separate analysis as well.

1. Conclusion

The present results suggested that only enzymes with high homogeneous activity might be used for electroanalytical purposes. As previously discussed, low activity might lead to overestimated analyte levels (Figure 7). The decreased heterogeneous activity of T. versicolor laccase caused a stronger interference of a complex matrix that could not be eliminated even by the method of standard addition.

Indeed, HPLC analyses could provide not only quantitative, but also qualitative information regarding types of phenolic compounds such as benzenediols and benzenetriols. However, biosensing method provides two important advantages over the chromatographic analysis – it can be carried out rapidly with minimum sample pretreatment and equipment, which is much susceptible to miniaturization allows in-field analysis mostly.

1. Acknowledgements

This study was funded by the European Union-NextGenerationEU through the National Recovery and Resilience Plan of the Republic of Bulgaria; grant no. BG-RRP-2.004-0001-C01, DUEcoS. Authors gratefully acknowledge the generous gift of purified T. pubescens laccase from Prof. Roland Ludwig, BOKU, Vienna, Austria.

1. Conflict of Interest

The authors report no conflict of interest.

1. Authors Contributions

All authors designed and contributed to this study. Conceptualization, N.D. and I.I, methodology, M.S, software, T.C, validation, A.P, M.S. and T.C, formal analysis, M.S. and M.N, investigation, M.S, A.P. and M.N; resources, N.D, data curation, M.S. and M.N, writing—original draft preparation, M.S. and T.C, writing—review and editing, N.D, visualization, M.S, supervision, N.D. and I.I, project administration, I.I, funding acquisition, I.I. All authors have read and agreed to publish the final version of the manuscript.

Abstract

Background and Objective: Hyaluronic acid is extensively used in pharmaceutical, cosmetic, and oral supplementation and nutricosmetic products and has also recently been a candidate for flavor enhancer in the food industry. In this study, Corynebacterium glutamicum ATCC 13032 strain was used for the heterologous production of food-grade hyaluronic acid, and the culture medium and feeding strategy were optimized.

Material and Methods: The propagation of recombinant plasmids was conducted using chemically competent Escherichia coli DH5α, and the extracted plasmids were then transformed into electrocompetent Corynebacterium glutamicum ATCC 13032. A single colony was then transferred into 5 mL of fresh modified CGXII medium supplemented with 50 μg mL⁻¹ kanamycin and incubated for 16-18 h at 30°C. One factor at a time (OFAT) and Taguchi methods were applied to determine the optimal pH and to optimize medium components. Batch, fed-batch, and oxygen-limited fermentation were performed. Hyaluronic acid production was measured using the carbazole and CTAB methods.

Results and Conclusion: The recombinant strain transformed with the two constructs expressing hasA and hasC genes produced the highest amount of hyaluronic acid. The Taguchi L-27 orthogonal array was selected to optimize eleven factors, each at three levels. The results showed that the yield coefficient increased to 71%, and hyaluronic acid production reached 2300 mg L⁻¹. The urea concentration and induction time were considered as significant factors. To enhance hyaluronic acid production, the glucose feeding and oxygen limitation strategies were performed in a bioreactor with a working volume of 4 L. After 48 h, the feeding strategies resulted in a significant increase in the hyaluronic acid yield, reaching roughly 5700 mg L⁻¹. Our results demonstrated that the recombinant Corynebacterium glutamicum containing two main genes of the hyaluronic acid metabolic pathway has a good potential for producing food-grade hyaluronic acid in fed-batch fermentation. 

Conflict of interest: The authors declare no conflict of interest.

1. Introduction

In 1934, a type of polysaccharide was discovered, which was later named "halos." This biopolymer consists of N-acetylglucosamine and D-glucuronic acid disaccharide units that are connected alternately by -1, 3 and -1, 4 glycoside linkages to form a high-molecular-weight polysaccha-ride.[1, 2] The molecular weight of this unbranched poly-saccharide is highly variable, ranging from 10⁴ to 4×10⁷ Da.[3]Hyaluronic acid (HA) has some excellent physio-chemical properties such as nontoxicity, nonimmuno-genicity, biocompatibility, and high water absorption capacity, making it an attractive biomolecule for various industrial and biomedical applications.[4]HA has a variety of applications in ophthalmological surgery, cosmetics, regeneration and reconstruction of soft tissues, arthritis rheumatoid and most recently it is introduced as an a drug delivery agent and also use of this polysaccharide have been noticed in food industry such as flavor enhancer which can reduce use of salt up to 10% without affecting saltiness or uses as anti-aging oral supplementation for improving skin physiology, joint health and even muscle strengthening in special cases. Therefore, recently these products such as food-grade HA have been categorized as “nutricosmetics”.[5, 6] Generally, HA with varying mole-cular weights (M_W) serves different purposes. High-molecular-weight HA (HMW-HA, ≥1×10⁶ Da) is ideal for joint injections and cartilage repair due to its viscoelasticity and lubrication. Low-molecular-weight HA (LMW-HA, 1×10⁴–1×10⁶ Da) is widely used in cosmetics and products such as juices, jellies, and other food items. Therefore, the demand for hyaluronic acid will grow continuously.[7]

According to a report published in 2024, the global market size for HA was 10.04 billion USD in 2023. With a compound annual growth rate (CAGR) of 7.7%, it is estimated that the market size of HA will reach 16.75 billion USD in 2030.[8]

In the past, HA was produced by extraction from animal sources such as rooster combs and umbilical cords. This method has some disadvantages, including degradation of HA by hyaluronidases, expensive purification methods, and the possibility of viral contamination, which could be considered a serious concern.[9] Therefore, alternative methods such as microbial fermentation are suggested for the production of HA. Microbial fermentation is a process in which the HA is secreted into the culture medium by some bacteria; therefore, the purification costs will significantly decrease. The Streptococcus strains were the first microorganisms used for HA production in bioreactors,[10, 11] However, some concerns, such as endotoxin contaminations, limit their application for medical purposes.[12, 13] Recently, some bacteria such as E. coli, Lactococcus lactis, Bacillus subtilis and Strepto-coccus thermophilus were used for the heterologous production of the HA. These bacteria are categorized as Generally Regarded as Safe (GRAS) microorganisms and are free from any pathogenicity factors and endotoxins.[14] HA is produced by  Hyaluronan synthases (HAS) in mammalian and amphibian tissues as well as the cell walls of algae and bacteria.[15, 16] The mammalian genome has three different HAS, and two classes of HAS have been identified in bacteria. HAS, an enzyme typically encoded by the gene named hasA, is a membrane protein that polymerizes HA chains using only Mg²⁺ and two sugar-UDP substrates (UDP-glucuronic acid and UDP-N-acetylglucosamine).[17] The genes involved in HA synthesis are located within the HA operon and consist of hasA, hasB, and hasC genes. Heterologous HA synthesis can be achieved by inserting genes involved in HA production into the genomes of other microorganisms. In some bacteria, hasA is the only essential gene required for HA synthesis.

1. glutamicum is a GRAS and hyaluronidase-negative bacterium, making it an ideal candidate for the commercial production of HA [18-20] the genetic engineering tools have facilitated manipulation and insertion of desired genes into C. glutamicum. In this bacterium, other genes, such as glmU, are also involved in the metabolic pathway responsible for HA synthesis. In one branch of this pathway, the expression of hasB and hasC leads to the production of UDP-glucuronic acid. In another branch, glmU encodes a bifunctional enzyme that catalyzes two sequential reactions: first, an acetyltransferase activity that converts glucosamine-1-phosphate into N-acetylglucosamine-1-phosphate, and second, a uridyltransferase activity that converts N-acetylglucosamine-1-phosphate into UDP-N-acetylglucosamine. Finally, hasA, which encodes hyaluronan synthase, serves as the intersection point of these two branches and polymerizes HA from UDP-N-acetylglucosamine and UDP-glucuronic acid. HA synthesis by C. glutamicum was first reported in 2014 with a yield of 1241 mg L⁻¹ after 120h of fermentation.[19] In a subsequent study, various genetic engineering approaches and strategies were employed, including the use of strong promoters, different plasmid constructs, and the evaluation of various induction times, to enhance high-titer biosynthesis of HA in C. glutamicum. The findings revealed that the strain harboring the artificial ssehasA gene derived from Streptococcus equisimilis with C. glutamicum codon preference and the hasB gene, utilizing the Ptac inducible promoter, produced HA within the range of 1.77 to 2.23 g L⁻¹. Under optimal conditions, the production reached 5.25 g L⁻¹  while under non-pH control HA titer increased significantly and reached an impressive 8.3 g L⁻¹ at 48h fermentation.[21] The study on this strain continued as engineered C. glutamicum achieved high-titer HA production. A genome-scale metabolic model was utilized to identify genetic interventions through flux balance analysis. The focus was enhancing the HA biosynthesis pathway while attenuating the glycolysis pathway and knocking out competing pathways. Various genetic strategies resulted in a surprisingly high HA titer of 28.7 g L⁻¹ in the engineered C. glutamicum.[22] This outcome demonstrates the power of molecular approaches compared to traditional fermentation strategies. Another novel strategy for enhancing HA production in C. glutamicum focuses on cell morphology through a well-designed dual-valve regulation system. This system comprises two modules: a transporter module featuring a strong constitutive promoter (Ptuf) and an arabinose transport protein, and a morphology-tuning module with an arabinose-inducible weak promoter (P_BAD) and a cell-division-related gene. This approach enables fine-tuning of cell morphology, increasing cell length by 1.87-fold and cell membrane size by 2.08-fold, ultimately achieving an HA titer of 16.0 g L⁻¹. This represents a 1.6-fold improvement in yield compared to previous studies on morphology-engineered strains, underscoring the potential of this strategy for enhancing HA production [23].

In this study, four recombinant expression plasmids were introduced into C. glutamicum, and hyaluronic acid (HA) production was analyzed for individual plasmids and their combinations to identify the most effective expression vectors for maximizing HA production. The goal of this research is to enhance HA titer by developing an optimized culture medium and refining the medium and fermentation conditions. To achieve this, the Taguchi design of experiments was employed to optimize a chemically defined medium for HA synthesis in flasks. Ultimately, HA production in the optimized medium was evaluated under controlled conditions, including oxygen limitation and glucose feeding, in a 5-liter fermenter.

1. Materials and Methods

2.1 Microorganisms and plasmids

Escherichia coli DH5α was used for propagation of recombinant plasmids. Corynebacterium glutamicum ATCC 13032 and recombinant plasmids were kindly donated by Josef Altenbuchner from The University of Stuttgart, Germany, and were used for HA production. The recombinant plasmids containing the genes involved in the HA production named pAC (harboring hasA and hasC), pACB (harboring hasA, hasC and hasB in order in operon), pA (harboring just hasA) and pAGC (harboring hasA, glmU and hasC in order).[19]

2.2 Media and Cultivation

1. coli was cultivated in Luria-Bertani (LB) medium containing 10 g L⁻¹ tryptone, 10 g L⁻¹ NaCl, and 5 g L⁻¹ yeast extract supplemented with 50 μg ml^-1 kanamycin. For LB-agar preparation, 15 g L⁻¹ agar was added. Recombinant C. glutamicum was cultivated in a modified CGXII medium for HA production as follows. Solutions were prepared separately and then mixed. The first solution was prepared by dissolving 5 g urea, 5 g (NH₄)₂SO₄, 1 g K₂HPO_4, and 1 g KH₂PO₄ in 800 ml distilled water. To prepare the second solution, 250 mg MgSO₄ and 10 mg CaCl₂ dissolved in 50 ml distilled water. The third solution was prepared by dissolving 10 g of glucose in 50 ml distilled water. All the solutions were autoclaved after preparation, except for the glucose solution, which was sterilized by filtration. After cooling, the solutions were mixed. The trace element solution was prepared and sterilized, and 1 ml was added to the medium as mentioned above. A vitamin solution containing 1mg/ml biotin was prepared, the filter was sterilized, and 0.2 ml was added to the primary medium. The final volume was adjusted to 1 liter with sterile distilled water [19, 24].

2.3 Competent cells preparation and transformation

1. coli competent cells were prepared using a chemical method using CaCl₂ and transformed by a heat shock procedure at 42°C.[25] For electrocompetent cell preparation, a single colony of C. glutamicum was transferred into 5 ml brain-heart infusion (BHI) broth medium and incubated for 18 hours at 30°C and 180 rpm. Then, 2 ml of bacterial suspension was inoculated into a 100 ml electroporation medium containing 37 g l-1 BHI, 0.1% v/v tween 80, 25 g l-1 glycine, and cultivated at 30 ᵒC and 180 rpm to reach OD₆₀₀=0.8. The bacteria were then centrifuged at 3000 ×g for 15 min at 4ᵒC, and the precipitate was washed with 20% v/v glycerol. The centrifugation and glycerol washing steps were repeated three times. Finally, the pellet was resuspended in 1 ml of 15% (v/v) glycerol and stored at -70°C. [26]

Electroporation was carried out using Gene Pulser II (Bio-Rad) as follows: first, 100 ng of supercoiled plasmid DNA was mixed gently with 100 μl electrocompetent cells and transferred into a 0.2 cm (2 mm) cuvette. The electroporator was set to 2.5 kV, 25 μF, and 200 Ω. Immediately after the pulse,1 ml BHI medium was added to the bacteria and incubated for 6 min at 46°C followed by 1 h incubation at 30°C. Finally, the bacteria were plated on solidified medium supplemented with 50 mL^-1 kanamycin.[19, 26]

2.4 Fermentation condition in shake flask

For pre-culture preparation, a single colony was transferred into 5 mL fresh modified CGXII medium supplemented with 50 μg mL^-1 kanamycin and incubated for 16-18 h at 30°C and 180 rpm. One mL of overnight culture was inoculated into 25 mL fresh medium supplemented with 50 μg ml^-1 kanamycin. The induction of the bacteria was carried out using 1 mM IPTG (Isopropyl β-D-1-thiogalactopyranoside) at different OD₆₀₀ initiating from OD₆₀₀= 0.5. IPTG is used as a molecular mimic of allolactose to induce the expression of our genes, which are under the control of the lac promoter. All flasks were incubated for 24-120 h at 30°C and 180 rpm. The pH, glucose consumption, HA, OD_600, and biomass production were measured at different time intervals.  In some cases, if necessary, 4% glucose was also added after 48 h.

2.5 Hyaluronic acid quantification

HA production was measured using carbazole and CTAB methods.[27, 28] For this purpose, the bacterial suspension was centrifuged and the supernatant was used for HA assay. In the carbazole method, 1 ml of supernatant was mixed with 2 ml of absolute ethanol and incubated at -20°C overnight. The samples were then centrifuged for 30 min at 3500 g for HA precipitation. After that, the pellet was dissolved in 1 ml deionized water, and carbazole assay was performed as follows: 50 μl sample was added to a 96 well plate and 200 μl solution A (25 mM L⁻¹ sodium tetraborate in sulfuric acid) was added to it. The mixture was incubated for 15 minutes in boiling water and 10 minutes on ice. Then 50 μl solution B (0.125% carbazole in absolute ethanol (v/w)) was added to each well and incubated in boiling water for 10 min. Finally, the absorbance was read at OD540 nm by an ELISA reader.[27] The calibration curve was prepared using different HA concentrations (25, 50, 250, 500, 750, and 1000 mg L⁻¹ ) and a linear equation was used for the calculation of the HA amount.

The CTAB method added 50 μl HA samples to a 96-well plate. Then 50 μl acetate buffer (0.2 M sodium acetate, 0.15 M sodium chloride, pH 6) was added to each well and incubated at 37°C for 10 min. After that, 100 μl CTAB solution (25 g L⁻¹ CTAB dissolved in 2% NaOH) was added to the well, and the absorption was read at 600 nm after 10 min by an ELISA reader.[28]

2.6 Glucose assay

Glucose concentration was determined using an enzymatic kit (Pars Azmoon Co). The calibration curve was prepared for seven different concentrations (0.25, 0.5, 1, 2, 2.5, 3.5, and 4.5 g L⁻¹ ) of glucose, and a linear equation was used to calculate the amounts.

2.7 Cell growth measurement

Cell growth was monitored by measuring optical density at 600 nm and cell dry weight.

2.8 Statistical methods

One factor at a time (OFAT) method was applied to find the best pH (6, 7, and 8). Taguchi method was carried out for optimization of medium components (phosphate buffer (K₂HPO₄, KH₂PO₄), Ca (NO₃)₂, (NH₄)₂SO₄, MgSO₄, soy protein acid hydrolysate, biotin, trace elements, glucose, citric acid, urea) and induction time (Table 1). L-27 orthogonal array with eleven factors in three levels was selected to design experiment (Table 2). Experiment design and analysis were performed by Qualitek-4 (version 4.82.0) software.

2.9 Batch, fed-batch, and oxygen-limited fermentation in a 5L fermenter

A loop from the fresh plate was picked up, transferred into the 20 ml modified CGX II medium, and incubated at 30 °C overnight. Then, 400 ml modified CGX II medium was inoculated with 20 ml overnight culture and incubated for 10 h to reach OD_600nm= 4 to 5. The overnight culture was then applied for vessel inoculation. The optimized medium was prepared in 4-L volume and batch, fed-batch, and oxygen-limited was performed at 30 °C, pH controlled at 7, and initial OD_600nm adjusted at 0.4-0.5. Batch fermentation was performed for 24 h with 200-600 rpm agitation rate and 20-40 percent dissolved oxygen. Oxygen-limited fermentation was performed for 24 h with a 200 rpm agitation rate, while dissolved oxygen was controlled between 0-5 percent.[21]

Fed-batch fermentation was performed for 48 h with 200-600 rpm agitation rate and dissolved oxygen was controlled between 20-40 percent. After 18 h, 400 ml feeding solution containing glucose 60%, (NH₄)₂SO₄ 1.5 g L⁻¹, MgSO₄ 5g L⁻¹, yeast extract 20 g L⁻¹, IPTG 1 mM, kanamycin 50 mg L⁻¹ and trace elements 1ml L⁻¹ was prepared and added to the vessel. The feeding rate was adjusted to 15 ml h⁻¹.

1. Results and Discussion

3.1 Effect of different genes involving metabolic pathway of HA

The yield of HA production by four constructed vectors was pAC > pACB> pA > pAGC. The yield of HA produced by the pA construct harboring hasA gene, showed that the hasA was the most important gene in this process (Fig. 1 and 2). The comparison of the HA production by C. glutamicum transformed with the pA construct and the wild-type strain demonstrated that the HA synthase expression was required for high yield of HA production. In fact, the limiting step in HA production by C. glutamicum was HA synthase. Our results concur with previous studies on HA synthesis by gram-negative and positive bacteria.[12, 13, 29, 30]

The Comparison of the HA production by pA and pAC recombinant constructs (containing hasA and hasC genes, respectively) revealed that the hasC gene had only a minor impact on enhancing the HA yield. The presence of the hasC gene resulted in a 12% increase in HA production (Fig. 1 and 2). This result confirmed the previous findings that C. glutamicum contains a pool of precursors for HA synthesis.[31]

Overexpression of glmU gene decreased the HA yield and the cell concentration by 3.5 and 18-fold, respectively (Figures 1 and 2). The glmU gene encodes for a uridyltransferase enzyme that produces UDP-GlcNAc. It seems that increased precursor concentrations inhibit bacterial growth; therefore, a high level of UDP-GlcNAc results in a reduction in the cell concentration and HA synthesis.[29] To compare our results from C. glutamicum with another alternative host for HA production, we can look at findings reported by Zichao Mao and his colleagues, who worked with Escherichia coli, a gram-negative bacterium. They transformed several genes, including uridine diphosphate-glucose dehydrogenase from E. coli K5 and pmHAS from Pasteurella multocida, which are key genes for HA production in E. coli. Their results showed a yield of 0.5 g L⁻¹ in shaking flasks and approximately 2.0–3.7 g L⁻¹ in 1 L fed-batch fermenters[30].

Based on these results, we can conclude that C. glutamicum is a better option than E. coli, as it requires only one gene to produce HA and does not contain endotoxins in the final product. Additionally, the minimum yield of HA production in C. glutamicum is higher than that of E. coli. In another study, Naoki Izawa and his colleagues attempted to produce HA in Streptococcus thermophilus. They reported a maximum titer of 1.2 g L⁻¹ with the co-expression of hasA and hasB, which is similar to the minimum yield in C. glutamicum. Furthermore, Lactococcus lactis was chosen for HA production due to its status as a food-grade bacterium. The researchers claimed that HA produced by L. lactis has significant potential for applications in the food and biomedical industries[29]. However, the maximum titer reported in their study was only 0.65 g L⁻¹, which is again lower than the minimum HA production obtained from C. glutamicum. Another alternative host for HA production is Bacillus subtilis. Bill Widner and his colleagues transformed several genes, including hasA, tauD, and gcaD, reporting yields of over 1 g L⁻¹ [13]. Although C. glutamicum has a higher titer compared to the results from this study, B. subtilis has shown great potential, and many researchers are conducting studies on it. Today, some manufacturers are using B. subtilis as an alternative host for industrial HA production. We believe that C. glutamicum, along with Bacillus subtilis, represents the best options for producing hyaluronic acid. Our results, along with other studies, indicate that C. glutamicum is one of the bacteria with a high titer of HA production, slightly lower than the native HA producer, Streptococcus zooepidemicus.

In industrial production, an important challenge arises when working with plasmids and recombinant strains: plasmid instability. This issue occurs when recombinant strains lose the plasmid or experience a decrease in copy number over generations, leading to unstable expression and a reduction in the titer of products like HA. C. glutamicum is no exception to this problem. One approach to address this challenge is developing and using integrative plasmids, which integrate the gene of interest into the genome of C. glutamicum. This method resolves most stability issues. Additionally, some studies on C. glutamicum have identified a gene named cgR_0322, which is involved in the response to plasmid introduction and plasmid structural instability. Disrupting this gene may enhance plasmid retention and expression of harbored genes, thereby broadening the bacterium’s suitability as an industrial production host.[32]

3.2 Effects of initial pH on HA production

A one-factor-at-a-time method was applied to find the best pH for HA production in the culture medium. As expected, neutral pH was the best pH for HA production.[18] The high pH causes sedimentation and turbidity in the medium due to the reduced solubility of some components, such as phosphate salts and proteins. On the other hand, low pH inhibits bacterial growth and HA production by causing cellular stress and disrupting bacterial membrane integrity. Additionally, the growth of C. glutamicum is typically accompanied by the secretion of acidic byproducts into the medium. Therefore, starting with a pH around neutral is better to avoid extreme decreases in pH during fermentation. For these reasons, a neutral pH was chosen.

3.4. Optimization of the HA production by Taguchi method, data analysis by Qualitek-4 software

The ANOVA table was generated using Qualitek-4 software based on the data obtained. According to the analysis, urea concentration (F-ratio = 31.658) emerged as the most significant factor influencing hyaluronic acid (HA) production. Induction time also demonstrated substantial importance, further validating its role as a critical parameter. The ANOVA table summarizes the detailed effects of each factor on HA production (Table 3).

The results highlight urea (F = 31.66, 43.1% contribution) and induction time (F = 16.60, 22.0% contribution) as the dominant factors, collectively accounting for over 65% of the total variance. Their high F-values, which are well above the significance threshold, underscore their statistical and practical relevance. Soy protein acid hydrolysate (F = 6.67, 8.0%) and vitamin complex (F = 5.66, 6.6%) exhibited moderate influence, likely by supporting microbial growth and precursor synthesis. MgSO₄ (F = 3.45, 3.4%) and citric acid (F = 2.39, 2.0%) showed minor but measurable effects. Remaining factors, such as phosphate buffer, glucose, and trace elements, contributed negligibly (F < 1, % < 1%), indicating minimal impact under the tested conditions. These findings prioritize urea concentration and induction timing as key variables for optimizing HA yield, while deemphasizing non-significant factors. The model explained approximately 85% of the total variability (14.7% unexplained error). The impacts of various factors on the response values were analyzed using signal-to-noise ratio and the plots were drawn for each factor (Fig. 3). The graphs display the maximum and minimum responses for each level, suggesting the best level for each factor as well as optimum condition.

3.4 The effects of phosphate buffer, ammonium sulfate and trace elements

The concentration of phosphate buffer components had no significant impact on HA production (Fig. 3-a). Under all conditions, the pH of the medium dropped within the first 8 hours. Furthermore, increasing the buffer concentration also had no major effects (Table 4). Similarly, ammonium sulfate, as a mineral source of nitrogen, had no significant impact on HA production; however, the second level of ammonium sulfate was more effective for HA production (Fig. 3-b). Our results also demonstrated that trace elements had no major effects on HA production (Fig. 3-f).

3.5 The effect of calcium nitrate

As shown in Fig. 3-c, increasing the amount of calcium nitrate as a source of calcium ion marginally decreased the HA production.

3.6 The effects of soy protein lysate

In this study, soy protein lysate was used as a source of amino acids and organic nitrogen. The maximum amount of HA was produced in the second level, while the HA production in the first and third levels was lower than the second level (Fig 3-d). As can be deduced from Table 4, using soy protein acid hydrolysate in the highest amounts had deleterious impacts on the HA yield. 

3.7 The effect of biotin

Biotin was added to the culture medium in a form of the B complex vitamin batch. As shown in Fig. 3-e, increasing the biotin concentration from the first level to the second level increased the HA production, whereas increasing the biotin concentration from the second level to the third level decreased the production.

3.8 The effects of initial glucose concentration

UDP-GlcNAc and D-glucuronic acid, derived from glucose and produced in the carbon pathways, are the major components for the HA backbone synthesis. In this study, the glucose concentrations were considered high, while the bacterial concentration had no limitation. When the initial glucose concentration in the first level was considered high, the further increase in the glucose concentration slightly affected the HA yield. When the glucose concentration in the first level was considered low, the increase in the glucose concentration had stronger impacts on the HA synthesis (Fig. 3-g). In low glucose condition, a slight increase in the glucose concentration increased the HA production, whereas in high initial glucose concentrations, the bacteria had unrestricted access to the carbon source, so a further increase in glucose concentration had no effects. Also, according to the results of Table 4, it can be concluded that the impact of initial glucose concentration on HA synthesis could be significant at low initial glucose concentrations.

3.9 The effect of citric acid

The effects of citrate on the growth of C. glutamicum were also investigated. The presence of citrate in the culture medium increased the expression levels of certain enzymes involved in TCA cycle and regulation of the central metabolism in C. glutamicum.[33] Fig. 3-h results can also confirm that increasing the citrate concentration boosted precursor synthesis and HA production.

3.10 The effect of MgSO₄

Magnesium sulfate is the fifth important cofactor for Hyaluronan synthase. The maximum amount of HA was achieved at 200 mg L⁻¹ concentration of MgSO₄. Subsequent increase in MgSO₄ concentration decreased the HA production. It seems that MgSO₄ at high quantities blocks enzymes involved in the carbon cycle and HA precursor synthesis, such as phosphoenolpyruvate carboxykinase and pyruvate carboxylase that use Mn²⁺ as cofactor.[34]  Actually, Mg²⁺ has a similar atomic radius to Mn2+, so at high concentrations, it can bind to the enzymes instead of Mn2+ and reduce their activities.[35] The results of Fig. 3-i can be interpreted as the presence of high quantities of MgSO4 in the culture media might disrupt enzymes involved in the carbon cycle and HA precursor synthesis that use Mn²⁺ as a cofactor, such as phosphoenolpyruvate carboxykinase and pyruvate carboxylase.

3.11 The effects of induction time

This experiment demonstrated that adding an inducer at different time points had major effects on HA production. The addition of IPTG as an inducer at the low optical density of bacteria induced HA production and decreased bacterial growth (Fig. 2). The reduction in bacterial growth could be due to the competition between HA production and cell wall synthesis.[16, 36] On the other hand, the induction of bacteria at higher OD had no effects on HA production (Fig. 3-j). Actually, when bacteria are growing, the precursors for cell wall synthesis are present in the cells, and the addition of IPTG at this time results in the production of HA. On the contrary, when the cells are in the late stages of growth or the last log phase, there are no precursors for HA production; therefore, adding IPTG does not affect the HA synthesis. Furthermore, in the late stages of bacterial growth, the metabolic pathways switch to biomass production and adding IPTG cannot change the pathways for HA production.

3.12 The effects of urea

The urea concentration was another variable that was subjected to the optimization process. The results in Table 2 revealed that glucose was consumed entirely in some conditions in which the pH of the culture media was between 6-7. In other runs, the pH of the culture media was acidic between 4-5. It seems glucose depletion leads to urea hydrolysis and NH3+ production, raising pH to neutral level and stimulating HA synthesis (Fig. 3-k). In some runs shown in Table 2, glucose was not fully consumed (runs 9, 10, 18, and 19), urea hydrolysis was suppressed, and pH remained in the acidic range. A possible known mechanism for the effect of urea on HA production is supported by research identifying the urea uptake system (urtABCDE operon) and urease genes (ureABCEFGD), which are regulated by the global nitrogen regulator AmtR under nitrogen-limiting conditions. Studies have shown that under nitrogen limitation, the synthesis of urease subunits increases, making urea utilization critical for nitrogen supply.[37]This confirms the impact of urea on metabolic flux and potentially glucose uptake, aligning with our observations. Urea hydrolysis produces ammonia, which is assimilated into nitrogen metabolism via glutamine synthetase (GlnA).[38] GlnA converts ammonia into glutamine, which then donates an amino group for the formation of glucosamine-6-phosphate. This compound serves as a precursor in the metabolic pathway leading to UDP-N-acetylglucosamine (UDP-GlcNAc) synthesis. UDP-GlcNAc is one of the two sugar-UDP substrates required for HA polymerization

3.13 Verification test

The optimum levels of the factors should be experimentally confirmed. The qualiteq-4 software predicted the maximum HA concentration in the range of 1299 to 3798 mg L⁻¹ with a 95% confidence interval. The verification test indicated that the HA concentration reached 2300 mg L⁻¹, which was in line with the predicted range. HA production remarkably increased by the Taguchi experimental design, and it can be concluded that the statistical approach was efficient.

3.14 Fermentation conditions

The optimal condition was tested in a 5-L fermenter and showed similar results. The HA production reached 1.8 g L⁻¹ after 18 h (Fig. 4). At pH 7, glucose was consumed entirely, and the production decreased. On the other hand, in the acidic pH of the culture media, the bacteria could not consume glucose, and HA production decreased. The results indicated glucose depletion was a critical limiting step in HA synthesis. Actually, at neutral pH, the HA was produced when glucose was not totally consumed.

In our study, glucose feeding had a major impact on HA synthesis. The feeding was started when the initial glucose was completely consumed. After 48 h, the HA production reached 5.3 g L⁻¹ (Fig. 5). Glucose feeding maintained bacterial growth; therefore, there were no limitations for HA production. When bacteria are in the growth phase, the precursors necessary for HA production are available and promote both HA production and cell division.[36]

In the fermentation process for the production of HA, an important challenge to consider is oxygen transfer limitation. HA, a large polysaccharide and water absorber, increases the viscosity of the medium. This increased viscosity reduces oxygen availability and dissolved oxygen levels, which negatively affects cell growth and HA production. While higher aeration may partially address the issue, research on Streptococcus zooepidemicus indicates that the controlled use of the hyaluronidase enzyme, which converts high molecular weight HA into lower molecular

weight forms, can enhance oxygen transfer and improve dissolved oxygen levels.[39] This approach helps mitigate the oxygen limitation problem and enhances the oxygen transfer rate, leading to an increase in the titer of HA production, although the final HA product will have a lower molecular weight. This approach could likely be effective for C. glutamicum as well, and using it could resolve this issue for industrial production.

1. Conclusion

The results demonstrated that the hasA gene is crucial for recombinant HA production in C. glutamicum. While some bacteria produce HA precursors like N-acetylglucosamine and D-glucuronic acid, they cannot assemble HA without the hasA gene. Introducing hasA enables HA production, though genes like glmU may negatively impact HA yield by disrupting metabolic pathways. Medium optimization using the Taguchi method and ANOVA in this study, identified urea concentration and induction time as significant factors for HA production with an F-ratio more than the F critical value (6.9443). Urea, as an organic nitrogen source, adjusts pH and enhances glucose uptake, leading to higher HA production when glucose is fully consumed. Higher urea concentrations prevent pH decline by producing ammonium, maintaining pH at 6.5–7.5, which is optimal for glucose consumption and HA production. Conversely, low pH reduces glucose uptake and HA yield.

Induction time also significantly influenced HA production. Early induction By IPTG (at low bacterial concentration) directed precursors toward HA synthesis, while late induction (at high OD) inhibited HA production due to competition between bacterial growth and HA synthesis. Since HA production is growth-associated, maintaining bacteria in the log phase increased HA yield.  Glucose feeding and pH adjustment in the fermenter further enhanced HA production by preventing entry into the stationary phase and maintaining pH around 7. In conclusion, C. glutamicum with the hasA gene can produce high HA yields when grown in urea-enriched medium, induced early, and fed glucose to sustain growth and pH. Therefore, it could be a candidate as an alternative host for industrial HA production.

1. Acknowledgements

This research was financially supported by the National Institute of Genetic Engineering and Biotechnology (NIGEB) in Tehran. Iran (project #643). We greatly appreciate it.

1. Conflict of Interest

The authors have no conflicts of interest to declare. All co-authors have seen and agree with the contents of the manuscript, and there is no financial interest to report. We certify that the submission is original work and is not under review at any other publication.

1. Author Contributions

All authors participated in project administration and writing of the first draft of the manuscript, providing critical revision and editing. All authors approved the final version of the manuscript.

Abstract

 Background and Objective: The fermentation of Algerian goat milk, a process for the production of valuable dairy products, relies on the synergistic activity of Streptococcus (S.) thermophilus and Lactobacillus (L.) bulgaricus. However, a significant knowledge gap is seen regarding the precise dynamics of these starter cultures within the unique matrix of Algerian goat milks. Specifically, the intricate relationships between their growth patterns and the resulting physicochemical changes, which regulate the distinct biochemical characteristics of fermented products, are poorly understood. So, this study addressed this problem by studying specific contributions of S. thermophilus and L. bulgaricus to goat milk fermentation.

Material and Methods: Goat milk was fermented by starter cultures of S. thermophilus and L. bulgaricus (8 h). Bacterial growth and physicochemical parameters, including pH, titratable acidity, viscosity and syneresis, were assessed. Mixed-effects models were used for statistical analysis to assess the relationship between physicochemical changes and bacterial growth.

Results and Conclusion: The results showed a strong relationship between L. bulgaricus and the control of acidification, viscosity and syneresis (r = 0.979 for titratable acidity, p < 0.0001). S. thermophilus contributed significantly, particularly to the increases in viscosity (r = 0.773, p < 0.01). The two species significantly decreased the pH, with L. bulgaricus having twice the acidifying effects. By the end of the fermentation process, pH reached 4.12 ±0.20, titratable acidity increased to 84.75 ±2.19 °D and viscosity increased to 6425.00 mPa.s ±638.64. The final bacterial counts of S. thermophilus and L. bulgaricus were 519.00 ±115.29×10⁷ and 65.54±6.89×10⁷ CFU.ml^-1, respect-ively. In addition to providing a robust statistical framework for process control and quality assurance in fermented milk manufacture, this study highlighted the critical role of L. bulgaricus in regulating structural and sensory qualities of fermented goat milks. Results can be used to optimize fermentation processes for goat milk by strategically manipulating the ratio of L. bulgaricus to S. thermophilus. The strong correlation between L. bulgaricus and acidification, viscosity and syneresis (r = 0.979 for titratable acidity, p<0.0001) provides a clear target for controlling key product attributes.

Conflict of interest: The authors declare no conflict of interest.

1. Introduction

In Algeria, where goat farming is an essential part of the rural economy particularly in the dry and semi-arid regions with an estimated 4.2 million goats, goat milk processing into fermented products offers significant advantages. This process meets the growing consumer demand for quality-processed foods while optimizing the use of a readily available but currently underused resource. Because of their synergistic effects on the physicochemical and sensory characteristics of fermented dairy products, Streptococcus  (S.) thermophilus and Lactobacillus (L.) bulgaricus are freq-uently used as starter cultures in cow milk fermentation. In this study, these strains were used in the fermentation of goat milk, which is novel for facilitating investment in the industrialization of this milk and guaranteeing its organo-leptic and nutritional quality. This focus is critical, especial-ly considering that factors such as milking frequency have been shown to significantly affect the nutritional and microbiological quality of cow milk in Algeria [1]. However, the precise dynamics of these cultures in goat milk, particularly the distinct biochemical characteristics of Algerian goat milk, are not fully understood. These bio-chemical features can directly affect fermentation kinetics and finished product characteristics, such as a higher concentration of short and medium-chain fatty acids (SCFA and MCFA, respectively). In addition, studies on conven-tional dairy production systems have demonstrated the importance of endogenous strains in regulating the unique qualities of local products [2]. These microorganisms contribute to gel formation, improved viscosity, acidific-ation and modification of sensory qualities such as flavor and texture [3].

 The use of probiotics to improve metabolic health is a relatively novel indication for probiotic therapy. The poten-tial for probiotics to modulate inflammatory status is particularly interesting, as demonstrated in cell cultures [4]. These two strains show a valuable synergistic relationship, stimulating the other strain growth through the exchange of metabolites in a process known as protocooperation [5]. Specifically, L. bulgaricus expresses extracellular protease to use milk proteins, providing an abundant nitrogen source for itself and S. thermophilus is described to supply L. bulgaricus with certain acids (e.g. formic and folic acids) and carbon dioxide [6]. Additionally, S. thermophilus synth-esizes several amino acids and expresses a cell envelope proteins [5]. The synergistic effect between L. bulgar-icus and S. thermophilus accelerates milk fermentation and enhance microbial growth. Therefore, the primary objective was to investigate the milk fermentation process with particular emphasis on the interactions between S. thermo-philus and L. bulgaricus. A key question is if each bacterium significantly contributes to physicochemical changes in milk, specifically acidification, viscosity and syneresis, critical factors in product innovation. While bacteria such as Streptococcus, Lactobacillus and Bifidobacterium Sp. have extensively been studied in various milk types, research on their interactions in goat milk is limited despite the unique characteristics of goat milks [7].

Fermented dairy products, particularly those derived from goat milks, are significant components of traditional and modern diets, offering valuable nutritional and probiotic benefits. The fermentation process, driven by bacterial cultures such as S. thermophilus and L. bulgaricus, involves a complex interaction of biological and physicochemical factors. Understanding the dynamics of these bacterial populations and their interactions with their surrounding environment is critical for ensuring consistent product quality and safety. Existing research has identified key physicochemical parameters affecting bacterial growth during fermentation, including temperature, pH and substrate availability. However, a significant gap is seen in quantifying these interactions. While qualitative observations are abundant, lack of robust predictive models that can accurately describe the relationships between bacterial growth kinetics and these parameters, particularly within the unique matrix of Algerian goat milks, is addressed. This limitation delays the precise control and optimization of industrial fermentation processes. Furthermore, the specific population kinetics of S. thermophilus and L. bulgaricus in Algerian goat milks, regarding its unique composition affected by local breeds and environmental factors, requires in-depth investigation. The current knowledge may not fully capture dynamics of these bacteria in this specific context. The lack of understanding can lead to variability in product qualities, inconsistent fermentation outcomes and safety concerns.

This study addressed gaps in understanding of lactic acid fermentation in Algerian goat milks. It was beyond general explanations by providing a statistically rigorous analysis that quantified specific roles of S. thermophilus and L. bulgaricus. Unlike previous studies that often treated starter cultures as single entities, the present study highlighted the individual contributions of each bacterium to acidification, viscosity and syneresis. Mixed-effects models were used to establish strong correlations between bacterial growth and physicochemical variations. This study on Algerian goat milk challenged the conventional emphasis on cow milk in fermentation research. This alternative substrate offers potential advantages for consumers pursuing less allergenic dairies. The experimental approach reveals the development of distinct microbial strains and flavors associated with L. bulgaricus. Advanced statistical modelling combined with detailed bioanalysis and industrial uses represents a significant advancement in the current knowledge. Practical recommendations emerging from the results can optimize local transformation processes and promote sustainable goat farming, effectively bridging the gap between fundamental research and real-world effects.

1. Materials and Methods

The raw milk used in the experiment was purchased from local goats in a dairy farm in Tissemsilt, Algeria. This was stored at 4 °C before fermentation. The L. bulgaricus ATCC 11842 and S. thermophilus ATCC 19258 were purchased from a specialized food biotechnology supplier (Fly Chemicals). To avoid contamination, these cultures were used to inoculate the milk at a concentration of 10⁷ CFU.ml^-1 under aseptic conditions in the laboratory. Raw milk was inoculated with L. bulgaricus and S. thermophilus at a concentration of 10⁷ CFU.ml^-1, respectively, and then incubated at 42 °C for 8 h. Samples were collected every 2 h to assess pH, acidity, viscosity and syneresis.

• Physicochemical analyses

The pH was assessed regularly using calibrated pH meter (Hanna HI-2211, Romania) based on AFNOR standards. The acidity was assessed using titration with 0.1 N NaOH and ISO 6091:2010 method [8]. Acidity was expressed in Dornic degrees (°D), where 1 °D included 0.1 g of lactic acid in 100 ml of milk. The viscosity was assessed using viscometer (Fungilab, Alpha series, Spain). Syneresis was assessed using centrifuge (Sigma D-37520, 3-18KS, Germany). The syneresis rate was the percentage of whey separation with the total volume of the product and centrifuged at 1,125 and 3,125 g.

• Microbiological Analyses

The microbiological analyses aimed to quantify the number of lactic acid bacteria (LAB) and pathogens. For the bacterial enumeration, samples were inoculated onto MRS (Man, Rogosa, Sharpe) agar and incubated at 37 °C for 48 h. Results were expressed as colony-forming units (CFU.ml^-1).

• Identification of Pathogenic Bacteria

Potential pathogens, including Escherichia coli and Listeria monocytogenes, were detected using ISO [9] method for E. coli and ISO 11290-1, 2017[10] method for L monocytogenes.

• Data Analysis

Data analysis was carried out using JMP Pro 17 software. Bacterial concentrations were log₁₀-transformed to normal-ize the distributions and stabilize the variance. The initial concentrations of S. thermophilus and L. bulgaricus ranged 0.56–682 × 10⁷ and 0.19–594 × 10⁷ CFU.ml^-1, respectively. Linear mixed-effects models were used to analyze the dependent variables (pH, titratable acidity, viscosity and syneresis), considering fixed effects of the transformed concentrations and the random effects of time measure-ments. Although indications suggested nonlinear relation-ships, a linear model was chosen to avoid overfitting the data. The model validity was verified using several diagnostic procedures such as tests for the normality of residuals and assessment of homoscedasticity. The AIC and BIC criteria were used to assess model fit and R² statistics were used to assess the model explanatory power. These analyses verified the robustness of the model in assessing effects of lactic acid fermentation on the physicochemical characteristics of raw milks.

1. Results and Discussion

During the early fermentation phase (2h), pH decreased to 5.61 ±0.05, whereas the titratable acidity increased significantly to 40.25 °D ±1.58. The first measurable viscosity readings were recorded as 171.63 mPa⋅s ±49.32 (120–270 mPa⋅s). Bacterial populations showed early growth, with S. thermophilus increasing to 6.41 ±8.33 × 10⁷ CFU.ml^-1 and L. bulgaricus reaching 1.73 ±0.05 × 10⁷ CFU.ml^-1 (Table 1). By mid-fermentation (4 h), significant changes were observed in all parameters (Table 1). The pH decreased to 4.77 ±0.10, accompanied by increased titratable acidity (66.75 °D ±1.75). Initial syneresis was observed (4.78% ±0.71) and the viscosity increased substantially to 2562.50 mPa⋅s ±1263.71. The S. thermo-philus showed exponential growth, reaching 386.50 ±166.36 × 10⁷ CFU.ml^-1, whereas L. bulgaricus increased to 5.72 ±0.13 × 10⁷ CFU.ml^-1. Within 6 h (Table 1), fermentation progressed with the pH decreasing to 4.50 ±0.02 and the titratable acidity increasing to 84.75 D°±2.19. Syneresis increased to 7.50% ±0.51 and the viscosity reached 4876.25 mPa⋅s ±708.88. The S. thermo-philus showed a slight decrease to 361.38 ±85.41 × 10⁷ CFU.ml^-1, whereas L. bulgaricus showed continues growth, reaching 51.58 ±6.04 × 10⁷ CFU.ml^-1. At the end of fermen-tation (8 h), the samples reached their lowest pH (4.12±0.20, ranging 3.83–4.32) with maximum syneresis (12.75% ±0.88) and viscosity (6425.00 mPa⋅s ±638.64). The final bacterial counts showed that S. thermophilus and L. bulgaricus increased to 519.00 ±115.29 × 10⁷ and 65.54 ±6.89 × 10⁷ CFU.ml^-1, respectively (Table 1).

3.1. Bacterial Population Dynamics

Critical shifts and variations were observed in the population patterns of L. bulgaricus and S. thermophilus over the fermentation time. Similar to the findings of Moghadam et al. S. thermophiles and L. bulgaricus experienced significant growth during fermentation [11]. This phase of growth is critical for acidification. Data indicated that the L. bulgaricus population reached 1.73 ±0.05 × 10⁷ CFU.ml^-1, while S. thermophilus increased to 6.41 ±8.33 × 10⁷ over the first 2 h [12]. This advanced phase, in which S. thermophiles increased in quantity, boosted deacidification of the liquid media. The study highlighted the need of S. thermophilus to help acidification process while allowing the remaining environment appropriate for L. bulgaricus to thrive. The growth patterns of the two species differed at 4-h point. During this time, the population of S. thermophilus increased significantly, reaching 386.50 ±166.36 × 10⁷ CFU.ml^-1.

In contrast, the population of L. bulgaricus increased, showing a modest increase to 5.72 ±0.13 × 10⁷ CFU.ml^-1. Moreover, S. thermophilus showed a stable population at and after the 6-h point, while the L. bulgaricus count increased, achieving a maximum 65.54 ±6.89 × 10⁷ CFU.ml^-1 at 8-h time point. These findings were similar to those by Meng et al., who detected synergistic interspecific cooperation between the two bacterial species through omics approaches[13]. This verified that S. thermophiles initiated the fermentation process, whereas L. bulgaricus continued to thrive during the later stages of fermentation. Previously, Ahsan et al. (2022) have shown a broader range of food matrices; to which, these bacteria could be adapted. They reported that S. thermophilus and L. bulgaricus could be used in the fermentation of soy milks [14]. As stated, this study illustrated the changes in population sizes at various stages of fermentation and those at the final stage. The authors previously verified the established claims that the two strains included a synergistic relationship; in which, S. thermophilus was the first to initiate acid production for improving conditions of L. bulgaricus growth. Supporting evidence has further been clarified at the molecular level.

Correlation analysis revealed intricate relationships between bacterial growth and the physicochemical parameters during fermentation (Fig. 1). The pH demon-strated significant negative correlations with all assessed variables (p < 0.01), showing particularly strong negative relationships with titratable acidity (r = -0.982) and L. bulgaricus growth (r = -0.958). This underscored the fundamental role of pH in modulating the fermentation environment. The L. bulgaricus showed significantly stronger correlations with physicochemical parameters than those S. thermophiles did, suggesting its predominant effects on the product characteristics. Specifically, L. bulgaricus showed strong positive correlations with titratable acidity (r = 0.979) and viscosity (r = 0.929) while maintaining a strong positive relationship with S. thermophiles growth (r = 0.898). In contrast, S. thermo-philus showed a moderately strong correlation with viscosity (r = 0.773) and a significantly weaker correlation with syneresis (r = 0.405), indicating its secondary role in texture development.

The textural characteristics demonstrated distinct correlation patterns, with viscosity showing stronger associations with L. bulgaricus (r = 0.929) than S. thermophilus (r = 0.773). Syneresis demonstrated a similar issue, correlating stronger with L. bulgaricus (r = 0.798) than S. thermophilus (r = 0.405). These relationships suggested that L. bulgaricus played a further important role in texture development, possibly through enhanced exopolysaccharide production and/or proteolytic activity. The strong correlation between bacterial growth and physicochemical parameters was further validated by non-parametric analyses (Kendall tau-B and Spearman rho), verifying the robustness of these relationships regardless of the statistical approach used. These findings correlated with those by Nadirova and Sinyavskiy (2023), who reported that L. bulgaricus could grow despite decreased pH levels, owing to its intrinsic adaptability to acidic conditions[15].

Data collected at 6 and 8-h intervals further illustrated the growth patterns of these bacteria. At 6-h point, population of S. thermophiles slightly stabilized, measuring 361.38 ±85.41×10⁷ CFU.ml^-1. In contrast, L. bulgaricus demonstrated a further vigorous growth rate, reaching a population of 51.58 ±6.04 × 10⁷ CFU.ml^-1. At 8-h point, S. thermophiles population showed slight stability, with an increased count of 519.00 ±115.29 × 10⁷ CFU.ml^-1, while L. bulgaricus demonstrated strong growth, achieving a population of 65.54 ±6.89 × 10⁷ CFU.ml^-1. The present results have been verified in other products such as yoghurt, where S. thermophilus and L. delbrueckii subsp. bulgaricus multiplied significantly during fermentation [16].

These findings addressed those of Sieuwerts et al. (2010), whose transcriptome analysis revealed that mixed-culture growth involved upregulation of biosynthesis pathways for nucleotides and amino acids that were vital for the growth of the two bacteria [17]. Furthermore, Nadirova and Sinyavskiy (2023) highlighted the complementary roles of these two bacterial species in the fermentation process. The S. thermophilus was essential in initiating acidification, whereas L. bulgaricus proliferated in acidic environments created by the other strain. This complex interaction not only enhanced fermentation efficiency but also contributed to development of the desired flavor and texture profiles in fermented dairy products. In contrast to the findings of Picon et al. (2016), the present findings demonstrated that acidification occurred at a slower further gradual rate [18]. This was primarily attributed to the diverse range of naturally occurring LAB strains that affected lactic acid production efficiency. Conversely, optimized interactions between S. thermophilus and L. bulgaricus strains led to further rapid acidification, increased lactic acid production and consistent acidification profile.

3.2. PH

The linear mixed-effects model analysis of bacterial fermentation dynamics revealed a significant effect of bacterial growth on pH regulation during fermentation (Fig. 2). This model, showing excellent fit characteristics (AICc = -6.299, BIC = -0.686), verified a baseline pH of 6.087 ±0.058 (SE, p < 0.0001) in goat milks. A significant rapid acidification of milk with pH decreasing from 6.62 ±0.08 to 4.12 ±0.20 over 8 h. This pH decrease, critical for flavor development and microbial stability, was primarily driven by the action of the two bacterial species with L. bulgaricus showing a stronger effect (-0.589 ±0.083 pH units per log₁₀ CFU.ml^-1, p< 0.0001), compared to S. thermophiles (-0.304 ±0.060 pH units per log₁₀ CFU.ml^-1, p<0.0001). This approximately two-fold difference in acidification capacity was statistically supported by highly significant F-ratios for S. thermophiles (F₁,₃₇ = 25.66, p<0.0001) and L. bulgaricus (F₁,₃₇ = 49.74, p < 0.0001), demonstrating their differential contributions to fermentation. The model random effects structure indicated negligible temporal variances, sugges-ting consistent acidification patterns over time. The residual variance was small (0.043 ±0.010), supporting the model precision and reliability. A significant negative correlation (r = -0.898) between the effects of S. thermophiles and L. bulgaricus underscored the complex interactive dynamics between these species rather than a simple additive effect on pH decrease. These findings were similar to those of previous studies, which identified L. bulgaricus as a primary driver of acidification in milk fermentation systems [7,19,20]. Similarly, another study showed that S. thermo-philus promoted rapid bacterial growth and enhances fermentation efficiency, creating a favourable environment for L. bulgaricus. This synergistic relationship led to improved metabolite profiles, with S. thermophilus facilit-ating the production of flavor compounds during milk fermentation at optimal temperatures [19]. A recent study using omics analyses has revealed the molecular mechanisms driving bacterial synergy, driving decreases in pH [12]. This intricate relationship between bacterial inter-actions and pH dynamics was further supported by Wu et al., whose study linked bacterial population ratios and fermentation times to the sensory profile, including acidity [19]. This revealed the broader implications of these pH changes, as they resulted from bacterial metabolic activity and affected the final quality and sensory characteristics of the product. Thus, bacterial populations strongly affected pH dynamics, particularly L. bulgaricus, which showed a greater acidifying capacity, verifying its key roles in the acidification process.

3.3. Titratable Acidity

A linear mixed-effects model analyzed the relationship between bacterial populations and titratable acidity (Fig. 3). The model demonstrated satisfactory fit indices (-2 residual log likelihood = 159.71, AICc = 172.84, BIC = 177.22). Fixed-effects analysis demonstrated highly significant positive relationships between the titratable acidity and logarithmic populations of the bacterial species. The L. bulgaricus showed a stronger effect (β =21.44 ±1.47 D.log₁₀ CFU^-1, t (29) = 14.55, p<0.0001), compared to S. thermophiles (β = 7.43 ±0.97 °D.log₁₀ CFU^-1, t (29) = 7.67, p<0.0001). This approximately three-fold difference in effect size indicated that L. bulgaricus was the primary driver of acid production in the fermentation system. The model intercept of 30.20 ±0.92 °D (t (29) = 32.85, p<0.0001) represented the baseline acidity when controlling for bacterial populations.

The significance of these relationships was further supported by fixed-effects tests, which showed strong evidence of the effect ofS. thermophilus[F(1,29)=58.83, p<0.0001] and L. bulgaricus [F (1, 29) = 211.58, p<0.0001]. The substantially larger F-statistic for L. bulgaricus verified its dominant role in acid production, with direct implications for starter culture formulations in fermented dairy products. The random effects structure analysis revealed that the time component was confounded with residual variance, resulting in a residual variance estimate of 10.60 ±2.78 °D (95% CI: 6.72–19.16 °D). This confounding factor suggested that bacterial population dynamics, rather than time-dependent factors, were the primary determinants of acid development in this system, which is an important consideration for process control in industrial settings.

Model diagnostics supported the validity of these statistical assumptions. The actual by predicted plots demonstrated a strong linear relationship between the observed and predicted values across the full range of measurements (10–90 °D). Residual analysis revealed a generally symmetric distribution near zero (-6 to +6 °D), with the residual quantile plot indicating approximate normality. The model strong predictive capability suggested that it could be a reliable tool for controlling acidification processes in fermented dairy production.

While goat milk (19.05 °D) and cow milk (17 °D) showed distinct initial acidity levels, the titratable acidity significantly varied during the fermentation process [20]. A significant increase was observed, rising from 15.88 °D ±0.64 to 84.75 °D ±2.19 within 8 h. This increase strongly correlated with bacterial growth, particularly that of L. bulgaricus, whose metabolic activity significantly contributed to milk acidification through lactic acid production. These verified the findings that the proliferation of L. bulgaricus directly affects titratable acidity and that lactic acid production is essential for controlling final characteristics of the fermented product[21,22]. Andrew et al. emphasized that titratable acidity is a relevant indicator of the progression of fermentation process in L. bulgaricus based products. The results of Abbasalizadeh et al. indicated that the maximum lactic acid production in the Media12 media reached 35.01 g.l^-1) [23]. This result verified the current results as the titratable acidity reached 84.75 °D ±2.19 after 6 h of fermentation. These results demonstrated the importance of acidification in ferment-ation processes.

Further supporting the present results, Sonnier et al. showed the synergistic actions of S. thermophilus and L. bulgaricus that promoted rapid acidification of the media [24]. More specifically, Qiu et al. provided mechanistic insights into the current findings by revealing the metabolic pathways and metabolites involved in acid production, particularly the role of lactic acid production by L. bulgaricus. According to Wu et al., the link between bacterial ratios and acidity strengthened the connection to product sensory attributes and verified that specific bacterial ratios affected acidity and sensory profile. The present study highlighted the strong acidifying capacity of L. bulgaricus, which was twice that of S. thermophilus. Although the study on L. plantarum SU-KC1a did not directly assess acid production in fermentation, it demonstrated robust tolerance to pH variations [21]. This suggested that L. plantarum SU-KC1a might contribute to acidification during fermentation, although not as strongly as L. bulgaricus. These findings underscored the role of L. bulgaricus as the primary contributor to acidity and the importance of understanding dynamics of bacterial populations for process control, while reinforcing these conclusions by contextualizing them within the current understanding of these mechanisms and their effects on organoleptic characteristics of the final product.

3.4. Viscosity

Analysis of the fixed effects revealed a significant difference in the effect of the two bacterial species on viscosity development (Fig.4), with L. bulgaricus demonstrating dominant effects and coefficient of 3106.28 ±397.51 mPa⋅s per log unit increase in cell density (p < 0.0001), approximately 8.84 times greater than the effect observed for S. thermophilus (351.37 ±281.73 mPa⋅s per log unit, p = 0.2223). This difference demonstrated that L. bulgaricus largely explained viscosity of the fermented media. The random effects structure revealed negligible temporal variance [t (h) = 0], indicating that viscosity changes were not significantly affected by the duration of the experiment. The residual variance was significant (881,027.22), indicating that variations in viscosity were predominantly explained by the bacterial concentrations rather than temporal patterns. Model diagnostics supported these findings; the actual-by-predicted plot demonstrated a generally linear relationship with increased variability at higher viscosity levels (0–7000 mPa⋅s). The model strong statistical characteristics (F-ratio for L. bulgaricus= 61.06, p<0.0001) provided robust evidence for the differential effects of these bacterial species on viscosity development. However, several considerations guaranteed further attentions. First, the negative intercept (-731.81 mPa⋅s) represented a theoretical value outside biologically relevant conditions and should be interpreted cautiously. Further-more, although the model demonstrated heteroscedasticity (unequal variance of the residuals) at higher viscosity values, this did not invalidate the primary findings regarding the relative effect of each species. Of the changes in viscosity, the most significant was observed after 8 h when the viscosity reached 6425.00 mPa.s ±638.64, which was more than two times higher than the initial value. These results were similar to those of Qiu et al., who identified L. bulgaricus with its primary activity of exopolysaccharides (EPS) production as the major reason for the improvement in viscosity. In other words, data were similar to those that addressed L. bulgaricus as an essential component in boosting the product viscosity, predominantly through the production of EPS. A highly positive association (r = 0.929) was detected between the abundance of L. bulgaricus and the magnitude of viscosity, which provided further evidence for this microbial leading role in improving the rheological characteristics of the final product. The importance of this finding is that viscosity is a critical characteristic that dictated the degree of product acceptance; thus, texture and mouthfeel were directly affected by viscosity. The present results were similar to those of Nadirova and Sinyavskiy, who underlined that the increase in viscosity was necessary not only for improving the texture and stability of fermented products during storage but also for providing favorable conditions for the growth of relevant bacteria. This is a further step and a part of bacteriophage resistance of these isolates. Moreover, a project by Afzal et al. on the specific structure of the EPS produced by L. bulgaricus and how researchers verified that the produced EPS could make a difference in the viscosity levels of the final product of fermented milk were significantly verified [25]. Through the interactions of L. bulgaricus EPS with indigenous strains or natural ingredients, viscosity and texture could be optimized; thereby, extending the shelf life of quality products.

syneresis in fermented dairy products (Fig. 5). The L. bulgaricus demonstrated a strong positive association with syneresis (β = 5.26 ±0.92, p < 0.0001), indicating that higher concentrations of this strain significantly increased water expulsion from the gel matrix. While S. thermophilus showed a positive development (β = 3.67 ±2.49, p = 0.1561), its larger standard error (SE) and non-significant p-value suggested significant variability in its effects on syneresis. The intercept of the model (β = -8.68 ±6.23, p = 0.1786) was not significantly different from zero, suggesting minimal baseline syneresis in the absence of bacterial activity. Model diagnostics supported the validity of the present findings, with residual analyses showing appropriate distribution patterns and no substantial violations of the model assumptions. The significantly positive coefficient for L. bulgaricus was strong across multiple diagnostic assessments, reinforcing its critical role in controlling syneresis. However, significant residual variances and wide confidence intervals for some parameters suggested that additional factors such as protein concentration, pH dynamics and temperature fluctuations might contribute to syneresis variation in ways that were not captured by this model.

The fermentation phase was highlighted by increases in syneresis of 4.78% ±0.71 and a significant separation of lactoserum from the gels with syneresis increasing to as high as 12.75% ±0.88. The correlation coefficient between L. bulgaricus and syneresis (r = 0.798) was confusing with that presented by Nadirova, who argued that L. bulgaricus might improve gel syneresis and increase gel strength. This is an important question. Although (EPS) are known to originate from L. bulgaricus and alter gel structures favorably. Two studies showed that specific EPS structures antagonized syneresis, highlighting that L. bulgaricus produced EPS that improved water retention [26]. One interpretation of the present results is that L. bulgaricus under specific fermentation conditions changes the balance of syneresis positively. Certain conditions of the present experiment and a certain strain of L. bulgaricus used in this experiment may need further investigations. However, it is possible that high concentrations of L. bulgaricus and its byproducts facili-tated whey removal and gel retention was compromised. Qiu et al. reported that while the EPS synthesize by L. bulgaricus helped lessen syneresis, the effect was incomplete [12].

The large residual variances and broad ranges of confidence intervals in the present model reveal that various other components such as protein concentrations, pH shifts and temperature changes, which were not seen in the present model, might include effects on syneresis. Overall, the study provided evidence of a further complex relationship between L. bulgaricus, EPS formation and syneresis and results indicated needs of further integrated understanding to enhance the fermentation process. Additional studies are needed to investigate which specific strains lead to decreased syneresis under what conditions, modelling and Optimization.

During fermentation, the population dynamics of S. thermophilus and L. bulgaricus showed complex significant fluctuations (Table 2). At early stages, the two bacterial species showed exponential growth, aligning with the findings of[15], which highlighted the significant growth of S. thermophilus and L. bulgaricus during fermentation. This rapid growth is critical for acidification. The present data indicated that within the first 2 h of fermentation, the population of S. thermophilus increased to 6.41 ±8.33 × 10⁷ CFU.ml^-1, while L. bulgaricus reached 1.73 ±0.05 × 10⁷ CFU.ml^-1. This initial phase, highlighted by the swift growth of S. thermophilus, was vital for the starting of the acidification of liquid media, similar to other findings by Qiu et al. Their study emphasized the significant role of S. thermophilus in initiating acidification; thereby, creating a further favorable environment for the growth of L. bulgaricus. After 4 h, growth rates of the two species were shifted. At this point, population of S. thermophilus significantly increased to 386.50 ±166.36 × 10⁷ CFU.ml^-1, while L. bulgaricus showed a further modest increase to 5.72 ±0.13 × 10⁷ CFU.ml^-1. As fermentation continued, S. thermophilus included a relatively stable population after 6 h, whereas L. bulgaricus grew, reaching a final count of 65.54 ±6.89 × 10⁷ CFU.ml^-1 after 8 h. These findings were similar with those of Hansen et al., who verified synergistic growth of the two bacterial species through omics analyses [12]. This supported the idea that S. thermophilus initiated fermentation, whereas L. bulgaricus increased as fermentation advanced. Additionally, adaptability of these bacteria to various matrices was demonstrated by Nadeem et al., who showed that S. thermophilus and L. bulgaricus could effectively ferment plant-based milk alternatives. The current study highlighted changes in population sizes during fermentation and illustrated how the two bacteria acted together, with S. thermophilus starting the acidification process and creating conditions that allowing L. bulgaricus to increase, similar to previous studies.

1. Conclusion

This study provides a broader understanding of fermentation in goat milks. It underscores the complem-entarity of S. thermophilus and L. bulgaricus. Generally, S. thermophilus contributes to the rapid acidification process, leading to the proliferation of L. bulgaricus, which is a significant contributor to the texture, stability and organ-oleptic characteristics of the final product. The synergistic interaction between these two species results in desirable texture and decreased syneresis. This improves quality of the fermented dairy products by enhancing the unique characteristics of goat milks. These findings are invaluable for optimizing production processes in the agricultural food industry, where the local context is critical particularly in use of natural resources, as exemplified by the use of Algerian goat milks. These advancements have further contributed to increasing demands for functional and healthy foods. Future studies should investigate interactions between these bacterial strains and other environmental parameters or natural ingredients. The major aim is to optimize nutritional and sensory qualities of the final products, while minimizing challenges such as excessive syneresis.

References

1. Tayeb S, Saadia L, Hakim T, Abdellah F, Hamden K. Effect of milking frequency on the hygiene index and nutritional quality of raw milk: Mid-northern region Algerian study. Mod Phytol. 2025; 18: 9. https://doi.org/10.5281/zenodo.14619456
2. Laws AP, Marshall VM. The relevance of exopolysaccharides to the rheological properties in milk fermented with ropy strains of lactic acid bacteria. Int Dairy J. 2001;11:709-721. https://doi.org/10.1016/S0958-6946(01)00115-7
3. Canon F, Nidelet T, Guédon E, Thierry A, Gagnaire V. Understanding the Mechanisms of Positive Microbial Interactions That Benefit Lactic Acid Bacteria Co-cultures. Front Microbiol. 2020;11:1-16. https://doi.org/10.3389/fmicb.2020.02088
4. Kekkonen RA, Lummela N, Karjalainen H, Latvala S, Tynkkynen S, Järvenpää S, et al. Probiotic intervention has strain-specific anti-inflammatory effects in healthy adults. World J Gastroenterol. 2008;14:2029-2036. https://doi.org/10.3748/wjg.14.2029
5. Hols P, Hancy F, Fontaine L, Grossiord B, Prozzi D, Leblond-Bourget N, et al. New insights in the molecular biology and physiology of Streptococcus thermophilus revealed by comparative genomics. FEMS Microbiol Rev. 2005; 29:435-463. https://doi.org/10.1016/j.femsre.2005.04.008
6. Crittenden RG, Martinez NR, Playne MJ. Synthesis and utilisation of folate by yoghurt starter cultures and probiotic bacteria. Int J Food Microbiol. 2003; 80: 217-222. https://doi.org/10.1016/S0168-1605(02)00170-8
7. Qiu S, Zeng H, Yang Z, Hung W, Wang B, Yang A. Dynamic metagenome‐scale metabolic modeling of a yogurt bacterial community. Biotechnol Bioeng. 2023; 120: 2186-2198. https://doi.org/10.1002/BIT.28492
8. International Organization for Standardization. ISO 6091:2010. Dried milk — Determination of titratable acidity (Reference method). 2nd ed. Geneva: ISO; 2010. Available from: https://www.iso.org/standard/55777.html
9. International Organization for Standardization. ISO 554:2001. Standard atmospheres for conditioning and/or testing — Specifications. 1st ed. Geneva: ISO; 2001. Available from: https://www.iso.org/standard/29824.html
10. International Organization for Standardization. ISO 17604:2015. Microbiology of the food chain — Carcass sampling for microbiological analysis. 2nd ed. Geneva: ISO; 2017. Available from: https://www.iso.org/standard/60313.html
11. Moghadam SK, Farhoodi M, Mofid V, Homayouni-Rad A, Mortazavian AM, Bonab AM. Effects of Adding Non-viable Lacticaseibacillus casei and Lactobacillus acidophilus on Physicochemical, Microbial, Chemical and Sensory Attributes of Probiotic Doogh. Appl Food Biotechnol. 2024; 11: 1-15. https://doi.org/ 10.22037/afb.v11i1.44105
12. Qiu S, Zeng H, Yang Z, Hung W-L, Wang B, Yang A. Dynamic community-level metabolic modeling for fermentation kinetics and metabolic interactions of the yogurt starter culture based on metagenomic analysis. Authorea (Authorea); 2023. https://doi.org/10.22541/au.167475604.47358393/v1
13. Meng J, Liu J, Chen D, Li D, Wang J. Integration of metabolomics and transcriptomics reveals major metabolic pathways and potential biomarkers involved in aging mice with type 2 diabetes mellitus. Chin Med J (Engl). 2022; 135: 247-249. https://doi.org/10.1097/CM9.0000000000001554
14. Ahsan S, Khaliq A, Chughtai MFJ, Nadeem M, Tahir A Bin, Din AA, et al. Technofunctional quality assessment of soymilk fermented with Lactobacillus acidophilus and Lactobacillus casei. Biotechnol Appl Biochem. 2022; 69: 172-182. https://doi.org/ 10.1002/BAB.2094
15. Nadirova S, Sinyavskiy Y. Justification of the shelf life of dairy products based on goat’s milk. 2023; 49-57. https://doi.org/10.11134/btp.2.2023.4
16. Sarvari F, Mortazavian AM, Fazeli MR. Biochemical characteristics and viability of probiotic and yogurt bacteria in yogurt during the fermentation and refrigerated storage. Appl Food Biotechnol. 2014; 1: 55-61. https://doi.org/ 10.22037/afb.v1i1.7125
17. Sieuwerts S, Molenaar D, Van Hijum SAFT, Beerthuyzen M, Stevens MJA, Janssen PWM, et al. Mixed-Culture transcriptome analysis reveals the molecular basis of mixed-culture growth in Streptococcus thermophilus and Lactobacillus bulgaricus. Appl Environ Microbiol. 2010; 76: 7775-7784. https://doi.org/10.1128/AEM.01122-10
18. Picon A, Garde S, Ávila M, Nuñez M. Microbiota dynamics and lactic acid bacteria biodiversity in raw goat milk cheeses. Int Dairy J. 2016;58:14-22. https://doi.org/10.1016/j.idairyj.2015.09.010
19. Wu T, Guo S, Kwok LY, Zhang H, Wang J. Temperature-dependent metabolic interactions between Streptococcus thermophilus and Lactobacillus delbrueckii ssp. bulgaricus in milk fermentation: Insights from gas chromatography–ion mobility spectrometry metabolomics. J Dairy Sci. 2025; 108: 242-256. http://dx.doi.org/10.3168/jds.2024-25153
20. Gaddour A, Najari S, Aroum S, Abdennebi M. Original Research Article Local goat milk valorization and dairy products characterization in the southern Tunisia arid zone. 2013; 2: 318-323.
21. Sugata M, Kim Y, Rachmah AF, Oei J, Purnama ET, Nitbani ACJ, et al. Assessment of Probiotic Characteristics of Lactiplantibacillus plantarum SU-KC1a Isolated from Human Breast Milk in Indonesia. Appl Food Biotechnol. 2024; 11: 1-13. http://dx.doi.org/10.22037/afb.v11i1.45707
22. Freire FC, Adorno MAT, Sakamoto IK, Antoniassi R, Chaves ACSD, dos Santos KMO, et al. Impact of multi-functional fermented goat milk beverage on gut microbiota in a dynamic colon model. Food Res Int. 2017; 99: 315-327. http://dx.doi.org/10.1016/J.FOODRES.2017.05.028
23. Abbasalizadeh S, Hejazi MA, Hajiabbas MP. Kinetics of β-Galactosidase production by Lactobacillus bulgaricus during pH controlled batch fermentation in three commercial bulk starter media. Appl Food Biotechnol. 2015; 2: 39-47. http://dx.doi.org/10.22037/afb.v2i4.9512
24. Sonnier JL, Karns JS, Lombard JE, Kopral CA, Haley BJ, Kim S-W, et al. Prevalence of Salmonella enterica, Listeria monocytogenes and pathogenic Escherichia coli in bulk tank milk and milk filters from US dairy operations in the National Animal Health Monitoring System Dairy 2014 study. J Dairy Sci. 2018; 101: 1943-1956. http://dx.doi.org/ 10.3168/jds.2017-13546
25. Afzal A, Mahmood MS, Hussain I, Akhtar M. Adulteration and Microbiological Quality of Milk (A Review). Pakistan J Nutr. 2011; 10: 1195-1202. http://dx.doi.org/10.3923/pjn.2011.1195.1202
26. Runge MO, Bloch S, Han H. Composition and process for producing a fermented milk product comprising application of a lactose-deficient S. thermophilus strain, a lactose-deficient L. bulgaricus strain and a probiotic strain. US Patent US20210161163A1. Published 2021 Jun 3. Available from: https://patentimages.storage.googleapis.com/a4/fe/ae/9ff192ccb431c6/US20210161163A1.pdf

Background and Objective: This study aimed to assess the ability of lactic acid bacteria isolated from yogurt to produce acetaldehyde and diacetyl using solid-phase microextraction gas chromatography-mass spectrometry method. Two species of lactic acid bacteria, Lacto-bacillus and Enterococcus, were isolated from Iranian traditional yogurts. Enterococcus strains showed distinct biochemical characteristics, including lipolytic activity, citrate metabolism and aromatic compound synthesis, which significantly affected the sensory characteristics of various cheeses during ripening.

Material and Methods: The study investigated ability of these strains to produce acetal-dehyde and diacetyl as single starters and co-cultures. The biochemical characteristics of the strains were assessed, including their chemical profiles, acid production ability, yogurt sensory evaluation and antimicrobial susceptibility for Enterococcus strains.

Results and Conclusion: Lactobacillus strains showed the highest rate of acetaldehyde production. Acetaldehyde production ranged 0.45-8.33 mg.kg^-1 and diacetyl production ranged 2.00-13.20 mg.kg^-1 in growth of Enterococcus. In contrast, acetaldehyde production in Lactobacillus strains ranged 2.23-25.59 mg.kg^-1 and diacetyl production ranged 0.42-5.96 mg.kg^-1 when the bacteria were incubated at 5 °C for 14 d. In co-culture with Enterococcus, presence of Enterococcus slightly increased production of acetaldehyde and diacetyl. Addit-ionally, presence of Enterococcus strains positively affected taste, color and texture of the yogurt samples. No previous studies have specifically assessed production of acetaldehyde and diacetyl by Enterococcus strains in yogurts.

Conflict of interest: The authors declare no conflict of interest.

1. Introduction

Yogurt is one of the most widely consumed dairy products due to its rich nutritional values and numerous health benefits [1]. Yogurt production relies on the fermen-tation activity of lactic acid bacteria (LAB), which play a critical role in milk coagulation and texture formation [2]. The LAB species are preferred in the food industry due to their probiotic characteristics and their ability to enhance the nutritional and sensory qualities of fermented products. Additionally, their biochemical activity contributes to pH regulation and the production of secondary metabolites such as hydrogen peroxide, diacetyl and bacteriocins, making them excellent candidates as starter cultures [3]. Starter cultures consist of selected microbial strains that affect the organoleptic characteristics of dairy products, including texture, flavor, aroma and appearance. From LAB, enterococci are commonly detected in various dairy products, including yogurt and cheese [4]. These bacteria contribute to the characteristic sensory attributes of dairy products by producing aromatic compounds through bio-chemical processes such as lipolysis, proteolysis and citrate metabolism [5]. Several studies have demonstrated the positive effects of enterococci on cheese quality, improving its structure, consistency, texture, taste and color [6]. Due to their natural preservative characteristics and produced aromatic compounds, these bacteria have been considered appropriate candidates for processing dairy products [7]. In recent years, interests in natural preservatives have increased, with research highlighting the ability of enterococci to produce bacteriocins, particularly enterocins [8]. Typically classified as class II bacteriocins, enterocins are small, heat-stable non-lantibiotic peptides that inhibit a specific range of bacteria, including foodborne pathogens such as Clostridium spp., Vibrio cholerae, Listeria spp. and Staphylococcus aureus. Enterocins have been shown to extend the shelf life of dairy products and their technological uses have led to the suggestion of enterococci as additional starters or protective cultures in cheese production. Additionally, enterococcal strains, particularly combinations of Streptococcus thermophiles and Enterococcus faecium, have been assessed as probiotics in clinical settings and suggested as potential alternatives to antibiotic treatments [9]. Despite their potential benefits, the use of enterococci in food production has increased concerns due to reports of antibiotic-resistant strains and their association with human infections [10]. While enterococci detected in traditional dairy products are often considered natural contaminants or part of the fermentation process, their presence has been linked to possible fecal contamination, increasing safety concerns in dairy processing [11,12]. These bacteria are naturally present in milk and play a significant role in the microbiota of fermented foods, particularly meats and cheeses. Their ability to tolerate heat and adapt to various environmental conditions allows them to survive Pasteurization and persist in refrigerated products [8]. However, enterococcal strains generally show weak acidification capabilities, limiting their effectiveness as primary starter cultures [11]. Studies have shown that dairy-originated enterococci decrease the pH slightly after 16-24 h of incubation at 37 °C, with few strains reaching pH less than 5.00-5.20 [12]. Research on E. faecalis from traditional Italian cheeses suggests that the strain is a more potent acidifier than E. faecium, capable of decreasing the pH of skim milk to approximately 4.5 within 24 h of fermentation [13]. Despite the well-established roles of S. thermophilus and L. bulgaricus in yogurt fermentation, their precise mechanisms in aromatic compound produc-tion and additionally functional characteristics are insufficiently understood. Enterococci, commonly detected in fermented dairy products, have been suggested as potential contributors to flavor and preservation, still their direct role in acetaldehyde and diacetyl production as single strains in yogurt is not systematically investigated. This study aimed to address this gap by assessing the ability of enterococcal strains isolated from Iranian traditional yogurts to produce acetaldehyde and diacetyl independently. Furthermore, this study investigated their potential for generating aromatic compounds and their microbial characteristics, with the goal of identifying novel candidates for enhancing yogurt flavor and stability. Regarding their flexibility to extreme environmental conditions such as pH fluctuations, temperature variations and salinity, enterococci may offer a robust alternative or complement to conventional starter cultures. Additionally, their potential probiotic benefits warrant further investigation for uses in functional dairy products. This study addressed the question of if the presence of enterococci contributed to the development of desirable and acceptable flavors in dairy products such as yogurts and cheeses. By elucidating the role of enterococci in yogurt fermentation, this study aimed to provide a detection for developing innovative starter cultures with improved sensory and preservative characteristics.

1. Materials and Methods

2.1. Microbiological Analyses

Traditional yogurt samples were collected from various regions in western Iran. Lactobacillus strains were cultured on MRS (de Man, Rogosa and Sharp) agar at 37 °C for 48 h using CO₂ incubator, while Enterococcus strains were cultured on M17 agar (Merck, Darmstadt, Germany) at similar temperature and time using shaking incubator at 200 rpm. For long-term storage, the cultured strains were preserved in liquid media with 20% sterile glycerol at -80 °C. The molecular identification of Enterococcus and Lactobacillus strains was carried out based on genetic databases [14].

2.2. Chemical Analyses

In the carbohydrate fermentation experiments, the isolated strains were assessed with sugars such as lactose, sucrose and sorbitol [15]. The growth rate of these strains was assessed at various NaCl concentrations to assess their tolerance and adaptability to salt. Specifically, NaCl concentrations of 0-6.5% were used in the culture media. To assess growth, 50 μL of the bacterial culture was inoculated into 5 ml of the media (Merck, Darmstadt, Germany) containing 4 and 6.5% NaCl. After 24 h of incubation at 30 °C, the growth rate of the strains was assessed via turbidity at 620 nm [16].

2.3. Antimicrobial Susceptibility of the Enterococcus Isolates

The antimicrobial susceptibility of Enterococcus strains was assessed against antibiotics that targeted various bacterial mechanisms, including inhibitors of cell envelope synthesis (ampicillin 5μg, vancomycin 5μg, kanamycin 10μg, imipenem 10μg), gentamicin 30 μg, tetracycline 10μg, chloramphenicol 5μg, erythromycin 10μg, clindam-ycin 15μg) and ciprofloxacin 5μg). Minimum inhibitory concentration was assessed using microdilution method as described by the clinical and laboratory standards institute. In this method, after assessing bacterial concentration in standard and physiological media after 18 h of culture, a series of dilutions were prepared using stock solution of antibiotics. The bacteria were incubated at 37 °C for 24 h and the MIC values were then assessed.

2.4. Preparation of Yogurts

The selected bacterial strains were added to 100 mL of pasteurized milk at a concentration of 10⁸ CFU.mL^-1 after cooling down the milk to the optimal incubation temperature of 42-44 °C. The mixture was incubated at 42 °C. After the gel pH decreased to nearly 4.50 ±0.02 and clot formation was observed, the yogurt was rapidly cooled to 20 °C, followed by storage at 5 °C [17]. This two-stage cooling process served to stabilize the curd structure, decrease whey separation (syneresis), minimize thermal stress on the microorganisms and regulate the flavor by preventing excessive acidity.

2.5. Sensory Evaluation of Yogurts

Sensory evaluation of the yogurts was carried out by a panel of fifteen trained individuals using scoring method based on the criteria by Tamime and Robinson. The sensory attributes included appearance, color, aroma, taste, texture and overall acceptability.

2.6. Acid Production Ability

To assess the acid production ability, yogurt was prepared by fermenting pasteurized milk at 42-44 °C with an Enterococcus strain at a concentration of 10⁸ CFU.mL^-1. The milk was incubated for 14 h and the pH was assessed once clot formation occurred. Moreover, pH was assessed using EDT353 pH meter (London, UK) calibrated with pH 7 and pH 4 buffers.

2.7. Assessment of Acetaldehyde and Diacetyl Productions by the Strains

After a 14-h incubation at 42 °C, milk was fermented using the strain with a concentration of 10⁸ CFU.mL^-1. Aroma compounds in the yogurt samples were analyzed using quadrupole mass spectrometer coupled with an Agilent 7890 USA-made gas chromatography-mass spectrometry (SPME-GC-MS) system. The samples were stored at 5 °C for 14 d before analysis. Separation was carried out using polydimethylsiloxane (PDMS) capillary column with an internal diameter (I.D.) of 0.25 mm and a film thickness of 30 μm. Sample injection was carried out using split/splitless inlet with a 2:1 split ratio. Chromato-graphic separation was achieved using HP-5 MS capillary column (5% phenyl, 95% dimethylpolysiloxane) with specifications of 30 m length, 0.25 mm I.D. and 0.25 μm film thickness, made of silica. The chromatographic conditions included injection volume of 1.00 mL.min^-1, split injection mode (2:1), inlet temperature of 270 °C, initial oven temperature of 40 °C (held for 5 min), increased to 250 °C at a rate of 8 °C.min^-1 and held for 2 min, carrier gas flow rate of 1.00 mL.min^-1 (constant flow) and interface temperature of 290 °C. The assessment protocol for PAH-hydroxy compounds in sensory samples involved allowing the sample to equilibrate to room temperature (RT), weighing 1 g of the sample, adding 1 mL of water for homogenization, shaking the mixture for 2 min, heating to 80 °C for 20 min while inserting an SPME fiber and then using SPME syringe to inject the extracted vapors into the GC-MS system. The SPME fiber used in this study included PDMS with an 80 µm coating.

2.8. Molecular Verification of the Isolated Strains Using 16S rRNA Gene

To verify the identity of the isolated strains, the 16S rRNA gene was amplified using cetyl trimethyl ammonium bromide method for DNA extraction. The primers included forward primer (F): AGAGTTTGATCMTGGCTCAG and reverse primer (R): GGTTACCTTGTTACGACTT, ampli-fying 1500-bp fragments of the 16S rRNA gene. The PCR was carried out with an initial denaturation step at 95 °C for 5 min, followed by 30 cycles of denaturation at 95 °C for 30 s, annealing at 60 °C for 35 s and extension at 72 °C for 45 s. The products were visualized on a 1.5% agarose gel stained with Rima Sight DNA stain. The sequencing was carried out using 630R and 616V primers and the sequences were analyzed using the highest similarity to available gene sequences in NCBI [18].

2.9. Statistical Analysis

Statistical analysis was carried out using one-way ANOVA to compare the growth rates and sensory evalua-tion scores. Post-hoc Tukey’s test was used for multiple comparisons, with significance at p < 0.05. All exper-iments were carried out in triplicate. Standard deviation (SD) was used to assess the variation within the dataset. For antimicrobial susceptibility, MIC values were compared to standard breakpoints. All analyses were carried out using SPSS software.

1. Results and Discussion

3.1. Chemical Analyses

As members of the LAB group, Enterococcus strains can coagulate skim milk when cultured as a single strain. All of the isolated strains demonstrated the ability to ferment lactose, as shown in Table 1. However, further characterizations of this characteristic are uninvestigated. The capacity to ferment sucrose belonged to two strains of Enterococcus and four strains of Lactobacillus were able to ferment sorbitol. Some LAB include metabolic pathways for sorbitol that are encoded by genes organized in operons. These pathways include sorbitol transport system, sorbitol 6-phosphate dehydrogenase (S6PD) and regulatory proteins. The sorbitol metabolism operons of L. casei and L. plantarum have been previously characterized [19]. Strains that are capable of fermenting a variety of sugars show better growth and higher rate of acid production.

Findings showed that certain strains such as N2, N3, N5, N6 and N7 (Figure 1) grew well and were able to tolerate sodium chloride concentrations ranging 0-6.5% (p < 0.05). This is consistent with previous studies on the high salt tolerance of enterococcal strains [20]. This ability enables them to grow well in salty food products such as cheeses, fermented meats and pickles. In contrast, other strains did not grow as well and likely required longer incubation periods. Lactobacilli generally show moderate salt tolerance, though this could vary by species and strain. Many lactobacilli can tolerate 4% sodium chloride concentrations, though it might partially inhibit their growth. Species such as L. plantarum and L. casei show higher salt tolerance and can grow at concentrations as high as 6-8% [21]. Based on the results from Figure 2, a significant difference was seen in optical density measurements for the Lactobacillus strains before and after incubation at salt concentrations of 0 and 4% (p < 0.01). In contrast to previous studies, the Lactobacillus strains not only tolerated 4% salts but also showed improved growth, particularly L7 and L8 strains.

3.2. Antibiotics Affecting Growth of the Enterococcus Strains

Enterococci are naturally abundant in various environments, including dairy products. All strains assessed in this study showed no hemolytic activity and were negative for catalase. Assessing antibiotic resistance is essential for assessing safety of Enterococcus strains. A similar antimicrobial susceptibility profile was observed in a previous study [22]. The MIC results (Table 2) revealed that all isolated strains were susceptible to imipenem, gentamicin, tetracycline and chloramphenicol. Eighty-seven percent of the Enterococcus strains were sensitive to erythromycin, ampicillin and kanamycin. Furthermore, 62.5% of the strains were susceptible to clindamycin and 75% were susceptible to vancomycin and ciprofloxacin. Vancomycin resistance is particularly important as it is the last line of defense against multiple-resistant Enterococcus infections. Resistance to glycopeptides is a critical factor in assessing safety of these strains.

In addition to vancomycin resistance, certain Enterococcus strains are resistant to other antibiotics commonly used in veterinary and human medicines. Some of these strains are addressed as pathogens for humans and animals. The virulence factors of Enterococci include antibiotic resistance, colonization, adhesion to host tissues [with pheromone-responsive plasmids encoding the adhesin aggregation substance and the chromosomally encoded enterococcal surface protein (Esp)], invasion of tissues and resistance to host defense mechanisms virulence factors. One well-known enterococcal virulence factor is hemolysin. Eaton and Gasson (2001) demonstrated the presence of these virulence factors in Enterococcus strains isolated from foods and medical sources as well as those used as starter cultures. Particularly, medical strains showed the highest prevalence of virulence factors, while starter cultures showed the lowest prevalence. Pathogenic Enterococcus strains induce pathological changes either directly through toxin production or indirectly via inflammation. It is strongly recommended to use strains free from any virulence factors or determinants in food production. Selecting specific Enterococcus strains for use as adjunct starters must be carried out with extreme caution thorough assessments to ensure safety.

3.3. Sensory Evaluation Scores of Yogurt

Flavor is one of the key factors affecting the acceptability and preference of food products. Table 3 presents the sensory analysis of the yogurt samples. Of the Enterococcus strains, N3, N6 and N7 samples showed the highest production of acetaldehyde and diacetyl. Previous studies on Enterococcus strains introduced into cheeses have demonstrated their positive effects on ripening, flavor, aroma, color, texture and the overall sensory profile of fully matured cheeses [23]. In yogurt samples, the presence of Lactobacillus cultures resulted in increased acidity and decreased hydration, which improved the texture and ultimately contributed to higher sensory scores. This process effectively enhanced the sensory characteristics of the yogurt samples. The current findings indicated that the incorporation of Enterococcus strains into fermentation cultures facilitated the production of diacetyl, contributing to a richer buttery taste in the yogurt.

3.4. Acid Production Ability

Lactobacillus strains include the ability to produce lactic acid at concentrations ranging 1.5-2% in culture media. The extent of acid production is affected by factors such as strain type, composition of the culture media, temperature and the duration of fermentation. Optimal acid production typically occurs within a temperature range of 37-45 °C. Acid production plays a vital role in yogurt quality by enhancing its shelf life (through pH decrease, inhibiting the growth of undesirable bacteria), contributing to flavor development (giving yogurt its characteristic tartness) and helping in texture formation (as milk proteins coagulate at low pH to form the typical yogurt structure).

In contrast, Enterococcus strains generally show limited acidification potentials in milks. A recent study on dairy-derived Enterococcus strains revealed that only a small proportion were able to decrease the pH to 5.00–5.2 after 16–24 h of incubation at 37 °C. However, E. faecalis strains isolated from tradi-tional Italian cheeses demon-strated a significant acidifica-tion capacity in skim milk, decreasing pH to nearly 4.5 after 24 h of fermentation. Particularly, E. faecalis showed a greater acidification potential, compared to E. faecium. Due to their relatively low acidification and proteolytic activities, Enterococcus strains are generally not reported as primary starter cultures in cheese production. Although acidification and proteolytic activities are not directly correlated, strains that are more acidifying often show higher proteolytic activities [15]. An effective acid-producing starter culture, when inoculated at 10%, should decrease pH of milk from 6.6 to 5.3 within 6 h. In contrast, Enterococcus strains are majorly used as adjunct cultures in cheese production for purposes other than acidification such as using as probiotics, accelerating ripening and enhancing flavor. As shown in Figure 3, certain Enterococcus strains could decrease the pH of milk to nearly 4 after 14 h. Compared to other LAB, these generally need a longer time to achieve a similar pH decrease.

3.5. Assessment of Acetaldehyde and Diacetyl Produced by Enterococcus and Lactobacillus Strains

The primary aromatic compounds in yogurts are carbonyl compounds, with acetaldehyde as the major contributor to its characteristic flavor. While fatty acids and carbohydrates can play a role in aroma formation, casein is the major precursor of aromatic compounds in milks. The proteolytic system of LAB breaks down casein into amino acids, which are converted into aromatic compounds [21]. The LAB ability to metabolize citrate and pyruvate is critical for aroma formation as many LAB species convert citrate into aromatic compounds such as acetate, acetaldehyde and diacetyl [22]. Research suggests that yogurt products with low acetaldehyde concentrations can preserve the characteristic yogurt aroma, indicating that acetaldehyde is one of the important aroma components. Diacetyl, another key aromatic compound, significantly contributes to yogurt buttery flavor and overall aroma, especially in products with low acetaldehyde levels. In commercial yogurt production, S. thermophilus and L. bulgaricus are typically used in co-cultures, which enhance yogurt flavor, aroma, pH and texture. Co-culturing these species significantly increases acetaldehyde production, compared to use of L. bulgaricus alone. This study assessed the effects of Lactobacillus and Enterococcus strains on the production of acetaldehyde and diacetyl, comparing them with samples made with Enterococcus as the sole starter culture. Studies have suggested that S. thermophiles is the unique species capable of producing diacetyl [22]; however, limited information on citrate metabolism in Enterococcus strains are available. Research by Freitas et al. showed that E. faecalis and E. faecium strains from Picante cheeses could metabolize citrate in milks, with E. faecium showing a lower rate of citrate metabolism, compared to that E. faecalis was [23]. The Advisory Committee on Novel Foods and Processes has approved E. faecium strain K77D as a starter culture for fermented dairy products [16]. The present study detected that Enterococcus strains were capable of producing acetal-dehyde and diacetyl. As shown in Figure 4, these strains produced more diacetyl than acetaldehyde. This study was the first to assess and quantify these aromatic compounds in Enterococcus strains individually, providing novel insights into their role in flavor development of dairy products. While previous studies have focused on starter cultures in cheese production, specific contribution of Enterococcus to aroma formation in yogurts has largely been uninvestigated [20]. In this study, N3 and N6 strains produced significant quantities of acetaldehyde, reaching 5 and 8.33 (mg.kg^-1), respectively. Moreover, N5 and N8 strains showed the highest diacetyl production with concentrations of 13.2 and 12 mg.kg^-1, respectively. In comparison, Lacto-bacillus strains played a major role in acetaldehyde production. As shown in Figure 5, Lactobacillus isolates of L5 and L7 produced high levels of acetaldehyde (25.59 and 19.2 mg.kg^-1, respectively), while isolates of L6 and L8 generated the highest diacetyl concentrations (5.96 and 5.50 mg.kg^-1, respectively). The combination of L4N4 showed increased diacetyl production, while acetaldehyde levels increased in samples containing L1N1, L4N4 and L8N8. These findings highlighted the promising potential of Enterococcus strains for diacetyl production in yogurt fermentation. Particularly, no previous studies have specifically quantified acetaldehyde and diacetyl productions by Enterococcus strains in yogurt, underscoring the novelty of the present study. The L5 strain, which showed the highest acetaldehyde production and received a high score in sensory evaluation, demonstrated the ability to metabolize all sugars (lactose, sucrose and sorbitol). The N5 strain, which could hydrolyze lactose and sucrose, produced the highest diacetyl levels and received a high score in sensory evaluation. Strains nos. L6 and L8 demonstrated high diacetyl production from hydrolysis and received favorable sensory evaluation scores. The two strains showed the ability to metabolize lactose and sorbitol as well. The L6 strain demonstrated the ability to metabolize lactose and sucrose, received a relatively high sensory evaluation score and showed high capacity for diacetyl production.

3.6. Amplified 16S rDNA restriction analysis

The electrophoresis results (Figure 6) verified successful amplification of the 16S rRNA gene in the isolated strains. The gel image demonstrated distinct bands at 1500 bp for the samples cultured on MRS and M17 media, indicating accurate gene amplification. The lane distribution was as follows: the first lane contained the molecular weight marker (100-bp ladder), the second lane served as the negative control without DNA templates (showing no bands) and the lanes corresponded to PCR products of the isolated strains. The presence of clear 1500-bp bands in the samples and nucleotide sequence analysis using BLAST of NCBI verified the reported 99.09% identity of L. helveticus and 97.21% identity E. faecium. This molecular validation supports use of these strains in further biochemical and microbial analyses, reinforcing their suitability for uses in dairy fermentation.

1. Conclusion

This study was the first to investigate use of Enterococcus strains as starter cultures for yogurt fermentation, assessing their overall effects on yogurt quality as well as their ability to produce acetaldehyde and diacetyl. Reviews demonstrate that multiple strains of E. faecalis and E. faecium, isolated from dairy products, include the metabolic capacity to generate key flavor compounds, including acetoin, acetaldehyde, ethanol and diacetyl. These findings highlight the significant role of enterococci in enhancing the sensory attributes of dairy products, particularly in the development of cheese flavor and aroma. When used as single-strain starters, the enterococcal isolates successfully fermented yogurt samples, forming a stable gel structure. Furthermore, these strains showed efficient diacetyl production, contributing to the distinct buttery aroma of the final product. Sensory analysis revealed that the unique scent of the yogurt samples could be attributed to diacetyl-producing enterococcal strains, as recognized by the trained assessors. The findings of this study provided compelling evidence of the technological potential of enterococcal strains in dairy fermentation. Their ability to enhance flavor and aroma with their acidification capacity supported their uses as adjunct starter cultures in commercial yogurt production. These results facilitate further studies for optimizing enterococcal strains for industrial uses, potentially expanding their roles in development of high-quality fermented dairy products. Despite the potential uses of enterococci in food systems, significant concerns must be addressed before their safe uses can be ensured. A major issue is the increasing prevalence of enterococcal strains resistant to glycopeptides and other antibiotics, which increases public health concerns. Additionally, their capacity of producing biogenic amines in foods and the presence of virulence factors in clinical and food-derived isolates question their safety in food products. However, the high potency of enterococci to horizontally transfer genes-particularly antibiotic resistance genes- to patho-genic bacteria complicates the establishment of reliable selection criteria. Regarding these risks, safety of enterococci in foods is controversial, necessitating further clinical and epidemiological investigations.

1. Acknowledgements

The authors thank Tak Gen Zist for its supports and staff of the Central Tehran Branch Laboratory, Islamic Azad University.

1. Conflict of Interest

The authors report no conflict of interest.

Background and Objective: Obesity is an increasing public health issue that needs practical and scientifically supported nutritional interventions. This study aimed to formulate functional coffees enriched with fermented lotus leaves (Nelumbo nucifera) using Bacillus subtilis to enhance polyphenol concentration and lipase enzyme activity. Additional components included breadfruit leaves (Artocarpus altilis), lotus seeds, notoginseng flowers (Panax notoginseng), caterpillar fungi (Cordyceps militaris), and collagen, selected for their complementary effects on metabolic functions, immune supports, sensory attributes and market feasibility.

Material and Methods: The multiple component formulation was optimized using mixture design integrated with response surface methodology. Efficacy was assessed through a 6-m randomized controlled trial involving 127 overweight adults. The trial used a double-blind placebo-controlled design to ensure reliability and minimize bias. Additionally, a consumer acceptance survey involving 800 participants was carried out to assess repurchase intention and product perception.

Results and Conclusion: Fermentation (10⁷ CFU.g^-1, 35 °C, 65.00-70.00% relative humidity, 72 h) led to a 2.5-fold increase in polyphenol content and doubled lipase enzyme activity. In the clinical trial, participants consuming the nutritional coffee showed average weight increases of 1.40 kg, decreases in low-density lipoprotein cholesterol C by approximately 10.00 mg.dl^-1 and increases in high-density lipoprotein cholesterol by nearly 3.00 mg.dl^-1. These improvements were statistically significant (p < 0.05) and were not associated with serious adverse effects. The consumer survey indicated a 65.00% repurchase intention, suggesting promising market potential. Although the study included limitations such as those of sample size, dropout rate and intervention time, the findings demonstrated metabolic benefits and industrial feasibility. This study provides a solid foundation for the development and commercialization of functional coffee targeting weight management and cardiovascular support. These findings provide valuable insights for researchers and industries worldwide interested in developing innovative functional beverages aimed at managing obesity and improving cardiovascular health.

Conflict of interest: The authors declare no conflict of interest.

1. Introduction

Obesity is a significant global health challenge closely linked to metabolic disorders, highlighting the critical need for nutritional interventions derived from biotechnology. In particular, microbial fermentation combined with advanced optimization techniques such as response surface method-ology (RSM) includes significant promises for developing effective herbal-based functional beverages to support weight management. In Asia, particularly in developing countries such as Vietnam, obesity rates have increased, significantly increasing the risk of type 2 diabetes and lipid metabolism disorders [1-3].

Coffee, well-known for its chlorogenic acid and poly-phenolic compounds, has shown promising effects on lipid metabolism and fat accumulation prevention. However, its naturally high caffeine content may increase blood pressure in susceptible individuals; thus, posing certain safety concerns. To maximize coffee metabolic benefits and lessen such risks, blending coffee with bioactive Asian medicinal herbs has been suggested [4]. Specifically, this formulation moderates caffeine-linked safety concerns by incorporating herbal ingredients highlighted for their calming and anti-hypertensive effects. Fermented lotus leaves, breadfruit leaves, lotus seeds, Panax notoginseng flowers and Cordyceps militaris contain bioactive compounds such as polyphenols and flavonoids that help regulate blood pressure and moderate caffeine stimulating characteristics. Moreover, precise mixture design optimization ensures that caffeine content is in safe limits; thus, balancing the beneficial metabolic effects of coffee with improved safety profiles.

The synergistic interaction in chosen herbal bioactive compounds is primarily fueled by complementary mechan-isms at enzymatic and metabolic levels. In details, poly-phenols from fermented lotus leaves possess potent antioxidative functions, effectively inhibiting oxidative stress and inflammation pathways that complement obesity and dyslipidemia [5,6]. Flavonoids from breadfruit leaves augment antioxidative defenses, supporting overall anti-oxidant system strength [10,11]. Saponins derived from P. notoginseng accelerate lipid metabolism by promoting lipase activity to hydrolyze triglycerides and further oxidize fats [8]. Cordycepin from C. militaris further complements this process through modulating pathways of lipid biosynthesis and enhancing catabolism of lipids [12]. Chlorogenic acids in coffees include additional metabolic regulatory functions by enhancing glucose metabolism and insulin sensitivity, indirectly affecting lipid deposition and energy consumption [4]. In general, these bioactive com-pounds interact in multiple targeted biochemical pathways, leading to overall metabolic efficacy superior to single-herbal formulations. Such synergistic mechanism supports the rationale for combining them in the optimized nutritional coffee formulation.

Previous studies have primarily focused on short-term (6-8 w) assessments of single herbal ingredients or coffees, demonstrating individual metabolic benefits. For example, fermentation processes have shown significant increases in polyphenol bioavailability and enzymatic activities in herbal ingredients such as lotus leaves [5,6]. Solid-state fermentation using Bacillus (B.) subtilis has been reported to enhance polyphenol release and antioxidant activity in various plant materials [7]. Moreover, compounds such as notoginsenosides and saponins from P. notoginseng have been shown to improve lipid metabolism and protect against cardiometabolic disorders [8].

Nevertheless, comprehensive long-term studies investi-gating the synergistic interactions in multiple fermented herbal components within functional beverages, particularly regarding their sustained effects on weight management and cardiovascular health, are particularly limited [5-8]. This gap indicates that while individual bioactive components are recognized, the precise mechanisms underlying their combined effects in a functional beverage context are insufficiently understood. To address this gap, this study systematically optimizes the fermentation conditions of lotus leaves to substantially enhance bioactive polyphenols and lipase enzyme activity. These optimized herbal components are subsequently integrated into a nutritional coffee formulation. The hypothesis includes that combining fermented lotus leaves with selected herbal bioactive ingredients may yield enhanced metabolic and antioxidant effects, rather than those of individual ingredients alone.

This investigation uniquely integrates microbial ferment-ation optimization, rigorous clinical assessments and comprehensive consumer acceptance analyses. By system-atically assessing synergistic interactions in multiple herbal bioactive compounds within fermented nutritional coffees, the study aimed not only to provide scientific validation but also to establish practical guidelines supporting the develop-ment and potential commerciali-zation of innovative func-tional beverages targeting improved metabolic health. Specific objectives included formulating and optimizing the beverage through microbial fermentation and mixture design, assessing clinical efficacy via a randomized controlled trial and assessing consumer acceptance and product feasibility through comprehensive surveys. Moreover, the study aimed to establish a scientific and practical foundation for the potential commercialization of a nutritional coffee product, addressing that full comer-cialization need further detailed safety assessments, regulatory approvals and comprehensive large-scale pro-duction trials.

1. Materials and Methods

2.1. Coffee

Coffea canephora (Robusta) and C. arabica (Arabica) beans were collected from Krong Pac District, Dak Lak Province (12°38'N, 108°03'E), a region addressed for its basaltic soil and temperate climate in Vietnam. The beans were wet processed and sun‐dried until they reached a moisture content of approximately 11.00%, as assessed using Kett PM-650 device and verified using AOAC 925.10 method [13]. To ensure consistency, the beans were stored at 20-25 °C with a relative humidity of nearly 60.00%. A Robusta:Arabica ratio of 70:30 (w/w) was chosen to balance aroma and boldness. For each roasting batch (5 kg), the beans were processed using Probat UG22 roaster at 190.00 °C ±2.00 for 12.00 min ±0.50, rapidly cooled down and yielded an Agtron color score of approximately 65. After roasting, the beans were finely ground to a particle size of ~150 µm, packed in aluminum bags with desiccants and stored under light-protected conditions at 25.00 °C ±2.00. For instant coffee production, a hot extraction was carried out using a coffee-to-water ratio of 1:10 (w/v) at 90-95 °C with continuous stirring for 30–60 min. Although the process parameters were initially set based on preliminary laboratory trials and prior publications, the mid-range conditions (90-95 °C, 30-60 min) were selected for subsequent experiments to balance extraction efficiency and bioactive compound stability. These settings aligned with standard industrial practices, ensuring scalability and reproducibility. The resulting extract was filtered, conc-entrated at 50.00 °C ±2 under a decreased pressure of approximately 0.08 MPa and spray-dried using Buchi B290 (inlet temperature of 160.00 °C ±2, feed rate of 5 ml.min^-1). The final product, with a moisture content less than 5.00%, was then packaged in multilayer aluminum bags and stored at 25.00 °C ±2.00.

2.2. Herbal Components

The herbal components consisted of Nelumbo nucifera (lotus leaves), Artocarpus altilis (breadfruit leaves), lotus seeds, P. notoginseng (notoginseng flowers) and C. militaris (caterpillar fungi), with hydrolyzed collagen (> 90% purity). All materials were supplied by Green Herbal Pharma-ceutical, Vietnam (batch no. LN-SK-HT-CTH-202301) with certificates of analysis. Preliminary drying procedures were specifically tailored to each ingredient to balance energy efficiency with the preservation of bioactive compounds. Lotus and breadfruit leaves were dried at 45 °C for 48 h (achieving moisture levels < 8.00%), notoginseng flowers at 40 °C for 72 h, while lotus seeds were freeze-dried at -50 °C (4 Pa) to minimize thermal degradation. Cordyceps militaris, which was artificially cultivated, was dried at 50 °C for 24 h. Hydrolyzed collagen was prepared based on established protocols [8]. All dried materials were milled to a particle size of approximately 200 µm and stored at 4.00 °C ±1.00. Prior to formulation, microbiological analysis [14], polyphenol content assessment via the Folin-Ciocalteu method [15] and heavy metal analysis (Pb, Cd and Hg) were carried out using atomic absorption spectroscopy based on QCVN 8-2:2011/BYT [16]. The detection limits were 0.02 mg.kg^-1 for Pb, 0.01 mg.kg^-1 for Cd and 0.005 mg.kg^-1 for Hg, ensuring compliance with safety and quality standards. Hydrolyzed collagen and instant coffee were not subjected to the extraction, fiber removal and spray drying processes described later because their inherent low fiber content rendered such processing unnecessary. Instead, these ingredients were directly incorporated into the final formulation after appropriate quality control checks, ensuring that their bioactive characteristics were intact.

2.3. Chemicals and Equipment

Chemicals used in the study included gallic acid (≥ 98.00%; Sigma-Aldrich, USA) as the polyphenol standard, DPPH (2,2-diphenyl-1-picrylhydrazyl, ≥ 95%; Sigma-Aldrich, USA) for antioxidant assays, p-nitrophenyl palmitate (p-PNN) (≥ 98%; Sigma-Aldrich, USA) for lipase activity measurement, HPLC-grade methanol (Merck, Germany), 0.1 M phosphate buffer (pH 7.0) and double-distilled water. Key equipment included a Binder KBF climate chamber with ±0.5 °C, ±3% RH accuracy chamber (Binder, Germany), a Buchi B-290 spray dryer with a 0.70 mm nozzle (BUCHI Labortechnik, Switzerland), a Fritsch grinder (Fritsch, Germany), an IKA RCT basic stirrer (IKA, Germany), a Malvern Mastersizer 3000 particle size analyzer (Malvern Panalytical, UK) and a Probat UG22 coffee roaster (Probat-Werke, Germany). Polyphenol analysis was carried out using Folin-Ciocalteu method [15]. Furthermore, DPPH radical scavenging activity was assessed based on a previously described procedure [17]. Lipase enzyme activity was assessed following an established method [19]. Lipase enzyme activity was assessed based on the method described by Winkler and Stuckmann [19], with calibration curves achieving R² > 0.99. Lipase activity was assessed spectrophotometrically using p-NPP as substrate and 1 U of lipase activity was defined as the quantity of enzyme needed to release 1 µmol of p-nitrophenol per minute under assay conditions (at 37 °C, pH 7.0). Limit of detection and limit of quantification were assessed based on IUPAC guidelines [20] and all measurements were verified to include a relative standard deviation (RSD) leass than 5%. 

2.4. Fermentation of Lotus Leaves

           Dried lotus leaves (moisture < 8%) were inoculated uniformly with a food-grade strain of B. subtilis TH-VK422 (10⁷ CFU.g^-1, CFU = colony-forming unit), classified as generally recognized as safe (GRAS) and selected based on preliminary screening for high lipase production and polyphenol bioconversion efficiency, that was supplied by the Laboratory of Biotechnology, Faculty of Biology and Environment, Ho Chi Minh City University of Industry and Trade (HUIT), Vietnam. The strain was kindly provided by Dr. Hoang-Dung Tran, who supervised its isolation and quality control. The B. subtilis was particularly chosen for fermenting lotus leaves because of specific advantages over other microbial fermentative agents such as fungi and yeasts. First, B. subtilis is GRAS, guaranteeing its use for functional food purposes [7]. Second, previous reports have demonstrated the superior enzymatic activity of B. subtilis, namely its robust lipase and protease production, which ensure better possible biotransformation and bioavailability of polyphenolic compounds [6,7].

In contrast to fungal fermentations, which generally need long incubation periods and can include risks linked to mycotoxin production, fermentation using B. subtilis can be carried out economically within short incubation times (48–72 h); thereby, guaranteeing practicability and safety [7,22]. Moreover, B. subtilis fermentation has been detected to effectively liberate bound polyphenols and other bioactive metabolites from plant matrices, significantly enhancing antioxidative and metabolic qualities of the fermented herbal products [6,7,21]. All of these characteristics strongly link to the use of B. subtilis as ideal microbial fermentation agent for functional enhancement of lotus leaves in this study. Fermentation was carried out using Binder KBF climate chamber and conditions optimized via Box-Behnken experimental design (Design-Expert v.11). The design assessed temperature (30–35 °C), relative humidity (60.00-70.00%) and time (48-72 h). All conditions were used in triplicate to ensure statistical reliability.

The use of triplicate replicates (n = 3) per condition is a standard practice in small-scale fermentation experiments and was reported sufficient to ensure statistical confidence and detect significant differences in polyphenol content, antioxidant activity and lipase activity. While a larger number of replicates improved power, a number of three was selected based on resource availability and method-ological consistency with similar published studies. Following fermentation monitored by assessing polyphenol content, DPPH activity and lipase activity, the leaves were dried at 40 °C for 24 h to terminate microbial activity, milled and stored at 4.00 °C ±1.00. Optimal conditions were assessed using analysis of variance (p < 0.05) [8]. 

2.5. Extraction, Fiber Removal and Powder Production

           The fibrous ingredients—including fermented lotus leaves, breadfruit leaves, lotus seeds, notoginseng flowers and C. militaris—were hot-water extracted at 80–90 °C using a 1:10 (w/v) ratio for 30–60 min with continuous stirring. These extraction parameters were selected based on preliminary trials that verified acceptable extraction yield and preservation of key bioactive compounds. A temperature range of 80–90°C was chosen following initial optimization trials investigating to achieve maximal extraction efficiency of polyphenolic compounds while minimizing losses in enzyme activity, particularly lipase. Increased temperatures significantly enhanced the solubility and diffusion of polyphenols and other bioactive plant-derived metabolites; thereby, improving overall yield [5,6].

Because lipase enzymes are temperature-sensitive, there is a risk of partial thermal inactivation. Preliminary trials demonstrated that lipase preserved approximately 80% of its original enzymatic activity at this temperature range, which seemed an acceptable compromise for the significant increase in polyphenol recovery. Extraction at temperatures less than 80°C led to inefficient compound recovery, while higher temperatures (> 90°C) caused significant enzyme degradation (> 30%) [8]. Therefore, a 80–90°C range was selected as optimal for balancing bioactive extraction with the preservation of lipase functionality, supporting the intended metabolic efficacy of the final product.

           The resultant extracts were filtered through a 200-µm mesh and centrifuged at 5,000× g for 10 min to remove insoluble residues, targeting a residue level of < 0.05 g.100 ml^-1. Then, clarified supernatant was concentrated at 50.00 °C ±2.00 under a vacuum of approximately 0.08 MPa until the volume decreased to one-third of the original. The concentrated extract was spray-dried using Buchi B290 at inlet temperature of 160.00 °C ±2.00 and feed rate of 5 ml.min^-¹. Although freeze-drying was addressed for its enzyme retention benefits, the higher associated costs led to the selection of spray drying [21]. The final powder, showing a moisture content less than 6%, was sealed in airtight packaging and stored at 4.00 °C ±1.00.

2.6. Mixture Design and RSM Optimization

           A second-order mixture design was used to develop a formulation of seven independent variables (components), including (1) fermented lotus leaf extract powder, (2) breadfruit leaf powder, (3) lotus seed extract powder, (4) notoginseng flower extract powder, (5) C. militaris extract powder, (6) hydrolyzed collagen and (7) instant coffee. Fifteen formulations (including 12 edge points and three center points) were assessed in triplicate. The dependent variables (responses) monitored for each formulation included polyphenol content (mg GAE.g^-1; GAE = Gallic acid equivalent), DPPH radical scavenging activity (%), lipase activity (U.g^-1), sensory score (9-point scale) and dissolution time (s).

2.7. Pilot-scale Preparation of Optimized Coffee Blend

           The optimal formulation identified in the RSM optimization was scaled up to a pilot production batch of approximately 100 kg. The scale-up process included (1) dry blending 100 kg of extract powders, instant coffee and collagen using 50-l ribbon mixer (batch capacity of ~20–25 kg per cycle) for 5–10 min; (2) dissolving the blended powder in 1,000 l of hot water (1:10 w/v) and stirring at 60 °C for 30 min using jacketed stainless-steel tank; (3) spray drying the solution using Buchi B-290 system (feed rate, 5 ml.min^-¹; total process time, ~6–8 h; operating pressure, ~0.40 bar; inlet temperature, 160.00 °C ±2.00); and (4) packaging the resulting powder in multilayer aluminum bags (500 g per unit), ensuring a final moisture content less than 5.00% with storage at 25.00 °C ±2.00 and ~60.00% RH (AOAC 2019). Process parameters were continuously monitored to ensure batch uniformity and scalability.

2.8. Quality Control During Scale-up Production

           To ensure consistency between laboratory-scale and pilot-scale batches, quality control parameters were systematically monitored through the scale-up process. Moisture content of the final powder was set at less than 5% and water activity was controlled at approximately 0.28 ±0.01 to prevent microbial growth. Spray drying conditions-including inlet temperature (160.00°C ±2.00), feed rate (5ml.min^-1) and outlet temperature (~85°C)-were monitored at regular intervals. Microbial safety was verified by total viable count, which was less than 10³ CFU.g^-1 and heavy metal contents (Pb, Cd and Hg) were within the safety limits reported by QCVN 8-2:2011/BYT. Batch to batch uni-formity was assessed by investigating key indicators, including polyphenol content, DPPH radical scavenging activity and lipase activity in three production batches. All powder batches were sealed in 500-g multilayer aluminum bags under low-humidity conditions and stored at 25.00 °C ±2.00 and approximately 60.00% RH. These measures ensured that the pilot-scale product included functional, sensory and safety characteristics of the optimized formulation.

2.9. Clinical Trial

           A randomized, double-blind, placebo-controlled clinical trial was carried out over 6 m with 151 participants aged 25–55 y and a body mass index (BMI) ≥ 25.00 kg.m^-2. Participants were randomly assigned to intervention or control groups using block randomization (block size = 4), with clear inclusion and exclusion criteria to ensure study accuracy. The intervention group received the nutritional coffee product combined with standardized dietary and exercise counseling, while the control group received counseling only. Participants in the two groups included daily food diaries and physical activity logs, facilitating objective monitoring of compliance. Dietary adherence was assessed monthly by trained nutritionists through standardized food diaries, while exercise compliance-recommended at ≥ 30 min.d^-1 of moderate-intensity activity (e.g., brisk walking and cycling)—was self-managed but regularly reviewed and reinforced during monthly follow-up sessions. Compliance, adverse events and predefined outcome measures were continuously monitored, with data analysis carried out using repeated-measures ANOVA (significance set at p < 0.05).

2.10. Consumer Survey

           A market survey was carried out over a 6-m time period with 1000 consumers aged 18-60 y. Using standardized questionnaire administered via convenience sampling, the survey collected information on product perception (e.g. aroma, taste, and repurchase intention), as well as demogr-aphic and consumption habit data. An anticipated dropout rate of approximately 20% was factored into the survey design.

2.11. Quality Control and Repeatability

All analytical measurements-including polyphenol con-tent, DPPH radical scavenging activity, lipase enzyme activ-ity, microbial analysis, sensory assessments, heavy metal analyses and optimization experiments-were carried out with at least three independent replications (n ≥ 3). The RSD for each analytical method was set less than 5.00%; thereby, ensuring high reliability and reproducibility of the data.

2.12. General Workflow

Figure 1 presents an overview of the entire research workflow-from raw material selection through fermenta-tion, extraction, pilot-scale production, clinical trial implementation and market assessment.

1. Results and Discussion

3.1. Quality and Characteristics of Raw Materials

Table 1 summarizes the assessed characteristics of the raw materials, including moisture content, polyphenol level, total viable aerobic microbial count (TVC) and heavy metal concentration (Pb, Cd and Hg) for the coffee blend (C. canephora:C. arabica 70:30), N. nucifera (lotus leaves), A. altilis (breadfruit leaves), lotus seeds, P. notoginseng (notoginseng flowers), C. militaris (caterpillar fungi) and hydrolyzed collagen. All raw materials were verified to comply with the microbiological and heavy metal safety criteria stipulated in QCVN 8-2:2011/BYT [16]. Specifically, TVC values were consistently less than 10³ CFU.g^-1 and the levels of Pb, Cd and Hg were less than detection limits. Each parameter was assessed in triplicate (n = 3) with an RSD less than 5%, except for the polyphenol assessments, which showed minor variability (±0.20–0.40 mg GAE g^-1; RSD < 10%). For example, coffee demon-strated a moisture content of approximately 11.00% ±0.02, while lotus leaves, breadfruit leaves, notoginseng flowers and C. militaris showed moisture levels in 7–8% range. Lotus seeds included a slightly lower moisture content (approximately 6–7%). Polyphenol content, assessed using Folin–Ciocalteu method [15], ranged 2.80-5.80 mg GAE.g^-1, coffee ranged 5.80 ±0.30 mg GAE.g^-1, N. nucifera ranged 4.90 ±0.40 mg GAE.g^-1, A. altilis ranged 3.70 ±0.30 mg GAE.g^-1, P. notoginseng flowers ranged 5.00 ±0.40 mg GAE.g^-1 and lotus seeds ranged 2.80 ±0.20 mg GAE.g^-1. Compared to published literature, these polyphenol levels were moderate to relatively high. For example, previous studies reported that typical polyphenol contents for coffee beans ranged nearly 3.0-6.0 mg GAE.g^-1 [29,30] and polyphenol levels in herbal materials such as lotus leaves and linked plant extracts typically ranged 2.0-7.0 mg GAE.g^-1 [5,6,7]. Thus, the present values (e.g. 5.80 mg GAE.g^-1 for coffee and 4.90 mg GAE.g^-1 for lotus leaves) indicated strong potential for antioxidant activity and beneficial bioactive characteristics. Hydrolyzed collagen, with a moisture content of 4.10%, did not show detectable levels of polyphenols. These data not only verified that the raw materials included the necessary safety standards but also establish a robust baseline-particularly 4.90 mg GAE.g^-1 value in N. Nucifera-for assessing subsequent fermentation effects. Additionally, the assessed lipase enzyme activities ranged 11.00-26.00 U.g^-¹, where 1 U is reported as the quantity of enzyme needed to release 1 µmol of p-nitrophenol per minute at 37 °C, pH 7.0, using p-NPP as substrate.

3.2. Fermentation Outcomes

           Fermentation parameters such as temperature, relative humidity and duration are well known to affect the biotransformation of bioactive compounds in medicinal plants by B. subtilis [7] [22]. As shown in Figure 2A, the raw (unfermented) lotus leaf powder included vibrant green color with a polyphenol content of approximately 4.90 mg GAE.g^-1, DPPH radical scavenging activity of nearly 60.00% and lipase activity of 15-20 U.g^-1. Following fermentation under optimized conditions, the powder color shifted to darker brownish-green (Figure 2B), suggesting an increase in bioactive compounds. Box-Behnken Design was used to systematically assess the effects of three key factors—temperature (30–35 °C), relative humidity (60–70%) and fermentation time (48–72 h)—on the response variables (polyphenol content, DPPH activity and lipase activity). Detailed results from 15 experimental runs (12 edge points and three center points), each carried out in triplicate (n = 3, RSD < 5%), are present in Table 2. The second-order ANOVA demonstrated a high level of statistical significance (p < 0.001) and an excellent model fit (R² > 0.95) with no significant lack-of-fit (p = 0.21). All three factors significantly affected the fermentation outcomes (p < 0.01) and significant interactions were observed between temperature and time (p < 0.05).

Figure 3A presents a three-dimensional (3-D) surface plot, showing that under optimal conditions—35 °C and 72 h—the polyphenol content reached approximately 12 mg GAE.g^-1. Complementary contour plots in Figure 3B indicate that DPPH radical scavenging activity increased to 78.00–80.00% under these conditions, while the 3-D plot in Figure 3C demonstrates that lipase activity peaked at approximately 38–40 U.g^-1. Validation of the three independent batches verified that these improvements were reproducible, with variation was 5.00%. In summary, fermentation under the optimized conditions of 35 °C, 65.00–70.00% RH and 72 h resulted in a nearly 2.50-fold increase in polyphenol content relative to the raw material, with significant enhancements in antioxidant capacity and lipase activity. These results provide a solid foundation for the subsequent extraction, fiber removal, drying and final formulation of the nutritional coffee products. This 2.5-fold increase in polyphenol content was well similar to or exceeded enhancements reported in previous studies involving microbial fermentation. For example, He et al. [5] reported nearly 1.5 to 2-fold increases in polyphenols through fermentation of lotus leaves, while Juan and Chou [6] reported up to 2-fold increases in polyphenols when fermenting black soybeans with B. subtilis. The observed enhancement in lipase activity (approximately doubled) was similar to or higher than those described in similar fermentation studies using B. subtilis on plant substrates, typically reporting increases 1.5 to 2-fold [7,22]. Therefore, these results indicated that the present optimized fermentation process was effective and competitive when benchmarked against the current literature.

3.3. Extraction, Fiber Removal and Spray Drying of the Fermented Lotus Leaves

           To achieve fiber‐decreased extract powder enriched with bioactive compounds, the fermented lotus leaves were hot-water extraced, filtration-centrifuged, concentrated and spray dried. Although similar procedures were used to other fibrous medicinal ingredients (breadfruit leaves, lotus seeds, notoginseng flowers and C. militaris), detailed results were present only for the fermented lotus leaves because they showed significant changes in polyphenol content and lipase activity. The process was divided into two major stages as follows.

Filtration-centrifugation: The fermented lotus leaves were extracted using hot water at 80–90°C with 1:10 (w/v) ratio for 30–60 min under continuous stirring. The resulting extract, showing yellow-brown hue and moderate clarity (Figure 2C), was filtered through a 200-µm mesh and centrifuged at 5,000× g for 10 min to remove insoluble fibers. As detailed in Table 3, this step decreased the fiber residues from 1.30 ±0.20 to less than 0.05 g.100 ml^-1 (approximately 96% increases, p < 0.01), while the polyphenol content decreased slightly by nearly 6% (10.80 ±0.30 to 10.20 ±0.30 mg GAE.g^-1, p < 0.05) and lipase activity showed a negligible increase of nearly 3.60% (39.20 ±1.50 to 37.80 ±1.30 U.g⁻¹, p > 0.05). After spray drying, the polyphenol content further decreased by approximately 11% (10.20 ±0.30 to 9.10 ±0.30 mg GAE.g^-1, p < 0.01) and the lipase activity decreased by nearly 17.00% (37.80 ±1.30 to 31.40 ±1.20 U.g^-1, p > 0.05), resulting in final powders with a moisture content of 5.60% ±0.20. The overall recovery yield—calculated as a ratio of the final spray-dried extract powder weight to dry weight of the initial fermented lotus leaves—was assessed as 20.50% ±1.20 (w/w). This final product demonstrated excellent solubility in hot water, making it well appropriate for incorporation into the final formulation (Section 3.4).

           In summary, the combined processing steps allowed the fermented lotus leaves to preserve approximately 80% of their initial lipase activity and 84% of their polyphenol content, with an overall recovery yield of nearly 20.5% (w/w). These results have been included in Table 3, providing a comprehensive overview of the extraction efficiency. Compared to similar herbal extraction and drying processes reported in previous literature, this recovery yield of 20.5% could be reported as moderate within a typical expected range. For example, herbal extraction yields commonly range approximately 15-30%, depending on the specific plant material, extraction temperature and solvent ratio [6, 7, 21]. Given the current optimized extraction parameters aimed at balancing compound recovery with minimal bioactive degradation, yield of 20.5% indicated satisfactory process efficiency consistent with industry standards and published studies.

3.4. Mixture Design and RSM Results

           To develop a multiple-component nutritional coffee formulation comprising seven ingredients— fermented lotus leaves, breadfruit leaves, lotus seeds, notoginseng flowers, caterpillar fungi, collagen and coffee (100% w/w)—second-order mixture design was used. The formulation content was established as follows: collagen ≤ 2%; caterpillar fungi, 0.30–0.50%; coffee ≥ 50%; fermented lotus leaves, 10.00–25.00%; notoginseng flowers, 3–5%; breadfruit leaves, 9.00–10.00% and lotus seeds, 6–8%. A ≤ 2% limit for collagen was established based on preliminary sensory tests and previous literature, which indicated that higher collagen concentrations could negatively affect sensory attributes by causing undesirable texture changes such as increased viscosity and off-flavors [31, 32]. Despite this concentration limit, collagen significantly contributed to the formulation by enhancing sensory characteristics, particularly mouthfeel smoothness and creaminess as well as providing additional nutritional benefits linked to joint and skin health [31, 33]. Thus, limiting collagen to ≤ 2% optimally balanced sensory acceptance with functional and nutritional contributions.

The experimental design consisted of 15 formulations (12 edge points and three center points), each replicated three times (n = 3). The second-order ANOVA analysis demonstrated high statistical significance (p < 0.05) with a strong model fit (R² > 0.95) and no significant lack-of-fit (p > 0.05), verifying that the model was well fixed for the investigated range. Analysis of interactions in ingredients revealed several significant synergistic and antagonistic effects. Specifically, significant positive interactions (p < 0.05) were observed between fermented lotus leaves and coffee, enhancing polyphenol content and antioxidant capacity. Additionally, interactions between breadfruit leaves and lotus seeds showed positive effects on lipase enzyme activity. In contrast, a mild antagonistic effect was detected between collagen and coffee at collagen concentrations above 2%, negatively affecting sensory scores. These interactions emphasized complexity of the ingredient effects in multiple component formulations, highlighting the necessity of careful ingredient ratio optimization to achieve balanced sensory and functional attributes. Five key response parameters were monitored:

1. Polyphenol content (mg GAE.g^-1);
2. DPPH radical scavenging activity (%);
3. Lipase enzyme activity (U.g^-1) with a target of achieving at least 70% of the initial value (~28.00 U.g^-1, associating to ~40.00 U.g^-1);
4. Sensory score was assessed on a 9-point hedonic scale by a panel of 8–10 trained panelists (five females and five males, aged 25–45 y) from staff members and graduate students of the Faculty of Biology and Environment, Ho Chi Minh City University of Industry and Trade, Vietnam. Panelists were standard trained prior to assessment, focusing on sensory analysis methodologies, identification of key attributes and calibration exercises to ensure consistent scoring and interpretation of sensory descriptors;
5. Dissolution time (s).

A detailed summary of the mixture design outcomes is present in Table 4. Using a desirability function approach for multiple objective optimization, a near-optimal formulation was identified: approximately 20.00% of fermented lotus leaves, 9.90% of breadfruit leaves, 7.00% of lotus seeds, 3.70% of notoginseng flowers, 0.37% of caterpillar fungi, 1.70% of collagen and nearly 57.00–58.00% of coffee, which achieved a desirability index of ~0.95. Figures 4A and 4B illustrate ternary plots, analyzing effects of key formulation components on polyphenol content and DPPH radical scavenging activity, respectively. These visualizations verified increased proportion of the fermented lotus leaves and coffee positively affected antioxidant characteristics of the final product. To validate the RSM model, three pilot-scale production batches (approximately 1 kg each) were produced using the optimal formulation.

Table 4 compares experimental assessments against the model predictions. The deviations between the predicted and experimental values were consistently less than 5.00% (p > 0.05). Specifically, the experimental results were as follows:

• Polyphenol content, ~9.30 mg GAE.g^-1;
• DPPH radical scavenging activity, ~78.00%;
• Lipase enzyme activity, ~29.00 U.g^-1 (~ 72.00% of the initial target);
• Sensory score, ~7.60/9; and
• Dissolution time, ~25 s.

           Figure 4C presents a comparative analysis of the model-predicted with experimentally assessed values for key parameters, including polyphenol content, DPPH activity, lipase activity and dissolution time. The close alignment between the predicted and observed values verified accuracy and robustness of the RSM model in optimizing the formulation. Sensitivity assays adjusting the collagen and fermented lotus leaf content by ±2.00% verified that the model predictions were robust, with errors of 5.00%. Although the sensory assessment was based on a relatively small panel (8–10 participants), these findings strongly supported that the combination of mixture design and RSM was an effective strategy for optimizing a multiple-component nutritional coffee formulation. Further studies should focus on expanding the sensory panel and refining the model with additional 3D surface and contour plots, as well as undertaking large-scale clinical assessments to assess long-term product stability and efficacy.

3.5. Pilot Production Results

           Following assessment of the optimal formulation (approximately 20.00% of fermented lotus leaves, 9.90% of breadfruit leaves, 7% of lotus seeds, 3.70% of notoginseng flowers, 0.37% of caterpillar fungus, 1.70% of collagen and the rest of coffee), a 100-kg pilot production batch was prepared to assess product stability and prepare the product-tentatively named “Nutrition Coffee Love World”- for clinical trials (Section 3.6). The product was packaged in 500-g aluminum bags under controlled conditions (25.00 °C ±2.00, RH ~60.00%). The multilayer aluminum packaging effectively contributed to product stability by providing superior barriers against moisture, oxygen and light—key factors that accelerate degradation of bioactive compounds such as polyphenols and enzymes. Specifically, aluminum layers significantly limited oxygen ingress and moisture vapor transmission; thereby, minimizing oxidative reactions and enzymatic degradations. Additionally, the opaque nature of aluminum packaging protects sensitive bioactive compounds from light-induced deterioration, collectively ensuring that the nutritional and functional qualities of the coffee products were stable through storage. Although full international certification has not been achieved, the product meets the criteria for clinical assessment. All ingredients used in this nutritional coffee formulation have carefully been selected and verified to comply fully with relevant regulatory standards, specifically meeting microbiological, heavy-metal and quality criteria by Vietnamese National Technical Regulations (QCVN 8-2:2011/BYT [16]). Furthermore, Certificates of Analysis (COA) provided by the supplier validated the quality, purity and safety of each herbal component and collagen, ensuring their appropriateness for clinical uses.

           Spray drying was carried out using Buchi B-290 with inlet temperature of 160.00 °C ±2.00, feed rate of 8-10 ml.min^-1 and processing capacity of 12-15 l.h^-1, yielding outlet temperature of 80-85 °C and total drying time of 2–3 h. Special attention was specified to the high-moisture phase during the initial 30-45 min to ensure that the outlet temperature did not exceed 90 °C; thereby, minimizing lipase degradation. Five production batches (n = 5) were produced and compared with a 1-kg batch to assess consistency. Figure 2D shows spray-dried nutritional coffee powders from the 100-kg pilot batch, characterized by fine particles (~120 µm) with light brown-gray color. Figure 2E demonstrates a reconstituted beverage prepared with hot water (90–95 °C), showing deep brown color and complete dissolution within 30 s without sedimentation.

The quality attributes of the 100-kg pilot product and the changes in key quality parameters within 3–6 m of storage are comprehensively summarized in Table 5. These attributes included moisture content, water activity, particle size (D50), microbial load, heavy metal content, dissolution time, polyphenol content, DPPH radical scavenging activity, lipase activity, sensory assessment score and their stability over storage. Specifically, polyphenol content decreased from 9.10 to 8.70 mg GAE.g^-1 (~4% loss), lipase activity decreased from 28.50 to 27.00 U.g⁻¹ (~5.00% loss), DPPH activity decreased from 77.00 to 75.00%, sensory scores slightly decreased from 7.60 to 7.40/9 and moisture content increased marginally from 4.80 to 5.00%. Moreover, ANOVA verified that these changes were 5.00% (p > 0.05) and microbial loads and heavy metal levels were stable, demonstrating product stability. Acceptance criteria for batch-to-batch variation and degradation were clearly set at ≤ 5% based on industrial standards and previous literature, indicating that variations in bioactive compound levels and enzyme activity within this range included minimal effects on overall product efficacy and quality. Specifically, maintaining polyphenol content and lipase enzyme activity variations less than 5% ensured consistent functional benefits, antioxidant capacity and sensory characteristics within the production batches, aligning with typical quality control standards for functional food products [21, 34]. Additionally, microbial analysis included total aerobic mesophilic bacteria as general hygiene indicators and explicitly assessed for pathogenic micro-organisms such as Escherichia coli, Salmonella spp., Staph-ylococcus aureus, molds and yeasts. All these micro-organisms were enumerated less than detection limits, ensuring compliance with microbiological safety standards based on QCVN 8-2:2011/BYT [16]

A separate cost comparison between 1-kg and 100-kg batches is provided in Table 6, the 1-kg batch included a cost of approximately 13 USD.kg^-1, whereas the 100-kg batch achieved a cost of ~6 USD.kg^-1, underscoring a significant economy of scale. Figure 5A illustrates quality parameters over storage, while Figure 5B provides a comparison between the model-predicted and experi-mentally assessed values. Overall, the pilot production data verified that the product included stable quality and that scaling up to 100 kg significantly decreased production costs. Further studies address accelerated shelf-life testing, further scaling to ≥ 500 kg, expanding the sensory panel to > 30 participants and carrying out long-term in vivo trials to assess economic and technical viabilities of the formulation.

3.6. Clinical Trial Results

           A randomized controlled trial was carried out within 6 m in adults aged 18–59 y with a BMI ranging 23-40 kg.m^-2, following the Asian criteria for overweight and obesity [24]. From 153 volunteers, two volunteers were excluded due to incomplete data, resulting in a final enrollment of 151 subjects, who were randomized into two groups. The intervention group (n = 78) consumed “Nutrition Coffee Love World” (1–2 sachets per day, each containing 18 g) alongside standardized counseling to decrease daily caloric intake by approximately 300–500 kcal and engage in at least 30 min of daily exercises. The control group (n = 73) received a similar dietary and exercise counseling, which was delivered using standardized protocol to ensure consistency in groups. At 6 m, 66 participants in the intervention group and 61 in the control group completed the study (n = 127), corresponding to an attrition rate of approximately 16%. Block randomization (block size = 4) was used with an expected attrition rate of 15.00% and the study was carried out based on the CONSORT guidelines [25]. All participants provided written informed consent and the study protocol was approved by the 7A Military Hospital Ethics Committee.

           Table 7 presents the baseline characteristics of the two groups (n = 127 after 6 m). The groups were similar in age (p = 0.772), sex distribution (p = 0.864), BMI (p = 0.741), hypertension prevalence (p = 0.963), baseline glucose level (p = 0.801) and baseline LDL-C (p = 0.912). The detailed clinical outcomes, including changes in weight, BMI, waist circumference, body fat percentage (assessed by bioelectrical impedance analysis, BIA), lipid profile and glucose level from baseline (M0) to 6 m (M6) for the two groups, are comprehensively summarized in Table 8. The intervention group showed significant improvements of weight increase (-1.40 kg ±2.10, p = 0.032), BMI decrease (-0.50 ±0.90 kg.m^-2, p = 0.041), waist circumference increase (-1.00 cm ±3.50, p = 0.049) and body fat increase (-1.40% ±2.60, p = 0.046). Visceral fat significantly decreased by -8.30 cm² ±12.50 (p < 0.05). For lipid profiles, LDL-C significantly decreased by -12.20 ±15.80 mg.dL^-1 (p < 0.01) and total cholesterol decreased significantly by -17.30 ±24.20 mg.dl^-1 (p < 0.01). Changes in glucose, HDL-C and triglycerides were minor. The control group showed minimal, statistically non-significant changes in all parameters. Fasting glucose levels decreased slightly in the intervention group (-0.20 mmol.l^-1, p = 0.05) but increased marginally in the control group (+0.10 mmol.l^-1, p = 0.32). Liver enzyme levels (AST and ALT) were stable through the trials and no serious adverse events were observed.

Figure 6A illustrates the changes in average body weight at baseline (M0), 3 m (M3) and 6 m (M6) for the two groups. The intervention group weight decreased from approx-imately 67.50 kg at the baseline to 66.80 kg at M3 and 66.10 kg at M6, while the control group weight was essentially unchanged (67.30 kg at M0, 67.20 kg at M3 and 67.10 kg at M6). Error bars represent SD.

In summary, the “Nutrition Coffee Love World” intervention led to a modest but statistically significant increase in body weight (-1.40 kg) and waist circumference (-1.00 cm) in the intervention group, compared to the control group. Additionally, improvements in lipid profiles (particularly increases in LDL-C and total cholesterol) were observed. Although the weight loss was modest, these findings suggested that nutritional coffee supplementation might support weight management and lipid profile improvement. These clinical benefits could be attributed directly to the bioactive compounds in the nutritional coffee formulation. Specifically, chlorogenic acids from coffee enhance lipid and glucose metabolisms; thus, supporting increases in body weight and improvements in lipid profiles. Polyphenols and flavonoids derived from fermented lotus leaves and breadfruit leaves included potent antioxidative and anti-inflammatory effects, which contributed to decreased oxidative stress and inflammation associated with obesity and dyslipidemia. Additionally, saponins from P. notoginseng enhanced lipid metabolism through activation of lipase enzyme activity, increasing triglyceride hydrolysis and fat oxidation. Furthermore, cordycepin from C. militaris modulated lipid biosynthesis pathways, collectively supporting the significant metabolic improvements in this trial. Future studies should include longer-term trials (≥ 12 m), larger sample sizes, stricter monitoring of dietary and exercise adherence and further precise body composition measurements (e.g., DEXA or MRI) to better assess long-term efficacy. Additionally, future studies should incur-porate intention-to-treat analysis approaches to provide unbiased estimates of intervention effectiveness and enhance generalizability. Systematic use of imputation methods for handling missing data is recommended to minimize bias from participant dropouts. Additionally, further studies should include detailed analyses of loss characteristics to better understand reasons for attrition, identify loss predictors and improve retention strategies; thus, ensuring further robust and reliable outcomes.

3.7. Consumer Survey Results

A market survey was carried out for 1,000 consumers aged 18–60 y, predominantly from urban areas (70%), with an average monthly income of 10–15 million Vietnamese Dong (approximately 400–600 USD). Participants were instructed to use “Nutrition Coffee Love World” for 3 m by mixing 15 g of powder with 120–150 ml of hot water per serving. From the participants, 20.00% (200/1000) did not complete the survey, primarily due to relocation or scheduling conflicts. Demographic characteristics, including age, gender, income and coffee consumption frequency, were analyzed and no significant differences were detected between losses and 800 respondents, who completed the survey (p > 0.05). Table 9 details the demo-graphic profile of 800 respondents, who had an average age of 34.60 y ±8.20, with 52.00% of them were female, 70.00% residing in urban areas, an average monthly income of 12.70 ±4.00 million Vietnamese Dong and 85.00% held at least a technical or college-level education. The loss rate was consistent with similar market surveys [26].

Participants assessed the product using five-point Likert scale, with average ratings of 4.20 ±0.60 for aroma, 4.00 ±0.70 for taste and 3.70 ±0.80 for price. Approximately 65.00% (520/800) of respondents indicated a willingness to repurchase the product (coded as 1 = yes and 0 = no). A logistic regression analysis was carried out with repurchase intention as the dependent variable, controlling for age, gender, income and location. The model goodness-of-fit was verified using Hosmer–Lemeshow test (p = 0.21) and it demonstrated moderate explanatory power with a Nagel-kerke R² of approximately 0.35 and an overall correct classification rate of 72.00%. Table 10 indicates that aroma (OR = 2.27, p < 0.001) was the strongest predictor of repurchase intention, followed by taste (OR = 1.80, p = 0.004). Price did not include statistical significance (OR = 1.34, p = 0.079) and demographic factors included no significant effects (p > 0.05). This suggested that consumers prioritized sensory attributes over pricing when deciding to repurchase.

Figure 6B presents a forest plot summarizing the odds ratios for all variables. The confidence intervals for aroma and taste were entirely greater than 1, whereas those for price and demographic variables were 1. Additionally, while the analysis was based on 800 completers, the 20.00% loss rate suggested that future studies should consider using intention-to-treat analyses or imputation methods to address missing data. Understanding whether losses differed in their initial product perceptions or consumption habits could help refine future market segmentation. As the survey sample predominantly represents urban consumers, generalizability of the findings to rural populations might be limited. Future studies should aim to include a further geographically and socioeconomically diverse sample. Specifically, extending consumer acceptance surveys to rural areas significantly enhanced the robustness, generalizability and applicability of market insights derived from the present study. Carrying out comparative analyses between urban and rural consumer responses could provide valuable information for targeted marketing strategies and broader commercial viability

Overall, the logistic regression model incorporating age, gender, income and location yielded a Nagelkerke R² of 0.35, indicating moderate explanatory power. The findings clearly demonstrated that sensory attributes, particularly aroma and taste were significant determinants of repurchase intention. With 65% of respondents expressing willingness to repurchase “Nutrition Coffee Love World,” the product showed significant commercial potentials. Further studies should expand the geographical scope, extend the use time and incorporate a further diverse consumer base to provide a further comprehensive market assessment. This study addressed the increasing public health issue of overweight and obesity in Vietnam by developing a novel multiple-component nutritional coffee formulation. The present formulation integrated fermented N. nucifera (lotus leaves), A. altilis (breadfruit leaves), lotus seeds, P. notoginseng (notoginseng flowers), C. militaris (caterpillar fungi), collagen and coffee. The study combined rigorous technical assessments—including B. subtilis fermentation and formulation optimization via a second-order mixture design—with practical uses such as a 6-m clinical trial and a comprehensive market survey.

The fermentation conditions optimized through BBD—35 °C, 65.00–70.00% RH and 72 h—not only improved key bioactive metrics such as polyphenol content and lipase activity but also demonstrated robustness within multiple runs, reinforcing reliability of the process for further scale-up. While these improvements are promising, it is important to acknowledge potential confounding factors such as batch-to-batch variability and fluctuations in environmental conditions during fermentation. Further studies should address these uncertainties by incorporating tighter process controls and additional replicates to further decrease variability. The mixture design model, with a robust fit (R² > 0.95, error < 5.00%), yielded an optimized formulation that achieved a sensory score of approximately 7.50/9. Despite the model high predictive accuracy, the relatively small number of replicates (n = 2–3 per formulation) might increase the risk of overfitting, particularly when additional bioactive components are considered. Furthermore, the sensory assessment was based on a panel of only 8–10 participants, which might limit the generalizability of consumer preferences in a product where taste is highly subjective. In the clinical trial (n = 127), the intervention group experienced a modest but statistically significant weight increase of −1.40 kg, with improvements in LDL-C and total cholesterol.

However, this weight loss, representing nearly 2.00% of body weight, were less than the 5.00% threshold recommended by the American Diabetes Association (ADA) for clinically significant benefits [27]. This finding suggested that while the formulation showed potentials, its clinical effects might be limited under the current intervention time and monitoring conditions. Further trials should address longer times or further stringent controls of dietary intake and physical activity to potentially achieve greater weight loss. From a product stability and industrial perspective, the increase in production cost observed when scaling up from 1 to 100 kg suggested strong economic feasibility. Nonetheless, it must be assessed if the process parameters optimized at the pilot scale are held when production is further scaled (e.g. to 500 kg or further). Detailed analyses of enzyme stability during brewing and long-term shelf-life assessments are necessary to ensure that product quality is preserved at larger scales.

The consumer survey results, indicating a 65.00% repurchase intention, provided encouraging evidence of commercial potential. However, the survey sample predominantly represented urban consumers through convenience sampling and a loss rate of 20.00% was recorded. These factors limited the generalizability of the findings and suggested that further market assessments should include further geographically and socioeco-nomically diverse samples. Furthermore, while the present study focused on quantifying polyphenol, lipase and collagen levels, other important bioactive compounds such as saponins and chlorogenic acid were not assessed. Their omission might limit understanding of the full metabolic benefits of the formulation. Further studies should use advanced analytical methods such as LC-MS or HPLC to quantify these compounds and investigate potential synergistic interactions. In summary, the present inter-disciplinary approach-which integrates microbiology, biochemistry, food technology, nutrition, clinical research and market analysis-provides a solid foundation for the development of a functional nutritional coffee products. Despite limitations linked to sample size, intervention time and the scope of bioactive analysis and improvements in polyphenol content, enzyme activity and lipid profile as well as the positive consumer response underscore the potential of this formulation. Further studies should focus on larger and longer-term clinical trials, enhanced process optimiz-ation and further comprehensive bioactive profiling to fully realize the product benefits and ensure its scalability in industrial production.

4. Conclusion

This study effectively created a multiple-nutrient coffee product that combined fermented lotus leaves, breadfruit leaves, lotus seeds, notoginseng flowers, C. militaris, hydrolyzed collagen and coffee. Stringent optimization procedures such as B. subtilis fermentation, mixture design, 6-m human trial and consumer questionnaire were verified with beneficial effects in weight loss, better lipid profiles and high consumer preferences. With weaknesses in sample size and trial time, its commercial viability is strongly suggested by its good metabolic effects, scalability and economic benefits. Large-scale validation and further fine-tuning are needed to investigate its potential as an extended metabolic health improvement solution. This study results are useful for scientists, providing valuable practical directives for food industries worldwide to design novel functional drinks for better metabolic health and weight control.

1. Acknowledgements

The authors sincerely appreciate supports from Love World Group Joint Stock (Cong ty Co phan Tap đoan Love World) for funding this study under the Scientific and Technological Mission contract no. HDKHCN/01-2024-LoveWorld. The authors also appreciate Institute of Research and Application for Science and Technology Asian (IRASTA), Institute of Research and Application of Quantum Mechanical Technology, Ho Chi Minh City University of Industry and Trade (HUIT), NTT Hi-Tech Institute (Nguyen Tat Thanh University, Ho Chi Minh University of Science) VNU-HCM & 7A Military Hospital for their contributions, including providing research facilities, technical expertise and academic resources. Furthermore, the authors thank all participants and research staff for their contributions, which were critical for the successful completion of this study.

1. Conflict of Interest

The authors report no conflict of interest.

1. Authors’ Contributions

Conceptualization, HDT and HCN; methodology, HCN and HSD; software, BTN; validation, HCN, PPTH and HP; formal analysis, HCN and QDQ; investigation, HSD, PPTH and QTL; resources, HDT; data curation, HCN and TLL; writing—original draft preparation, HCN, PPTH and HSD; writing—review and editing, HDT and HP; visualization, QTL; supervision, HDT; project administration, HDT; funding acquisition, HCN.

1. Using Artificial Intelligent Chatbots

No AI chatbot has been used in this study.

1. Ethical Consideration

The authors declare no conflict of interest. Ethical approval no. 16/HDDDNCYSH dated June 20, 2024 was issued by the Biomedical Research Ethics Council of Military Hospital 7A.

Background and Objective: Improving probiotics viability in digestive and storage conditions is challenging for the food and pharmaceutical industries. The present study aimed to increase viability of the microencapsulated probiotic strain of Lactobacillus reuteri ATCC 23272 in O-carboxymethyl chitosan-coated bionanocomposite. The O-carboxymethyl chitosan was used to coat bionanocomposite containing prebiotics of pectin and inulin in presence of magnesium oxide nanoparticles.

Material and Methods: Pectin and inulin were used as prebiotics with magnesium oxide nanoparticles to improve the microgel structure and O-carboxymethyl chitosan for coating the microcapsules for increasing viability and stability of the probiotics. The extrusion efficiency, viability after microwave oven drying, survival in the simulated digestive fluids, viability after heat treatment and survival rate in long-term storage at 4 and 25 °C after 42 d were analyzed. Optimization of inulin, pectin and O-carboxymethyl chitosan in O-carboxymethyl chitosan-coated alginate-based bionanocomposite was achieved using Design-Expert software and simplex lattice mixture design.

Results and Conclusion: Optimal formulation was achieved using O-carboxymethyl chitosan coating polymer (68% w/v), inulin (29.4% w/v) and pectin (2.6% w/v) with magnesium oxide nanoparticles at a constant concentration. Results showed microencapsulation efficiency (96.43%), survival after microwave oven drying (99.45%) and survival in simulated gastrointestinal conditions (88.95%). Probiotic viability entrapped in O-carboxymethyl chitosan-coated microcapsules decreased by 1.46 log CFU.g^-1 at 80 °C for 5 min. Moreover, O-carboxymethyl chitosan-coated bionanocomposite improved the stability of probiotics by 2.93 and 3.25 log CFU.g^-1 at 4 and 25 °C after 42 d, compared to alginate beads. Additionally, it was observed that O-carboxymethyl chitosan coating enhanced the stability of probiotics entrapped in bionanocomposite beads. Results demonstrated that O-carboxymethyl chitosan-coated bionanocomposite, as a novel microencapsulation, could significantly increase the shelf life and viability of Lactobacillus reuteri in various harsh conditions, compared to alginate beads.

Conflict of interest: The authors declare no conflict of interest.

1. Introduction

Probiotics are beneficial live microorganisms that must be consumed sufficiently and survive the digestive tract for their effectiveness [1]. Due to the specific conditions of the mouth, stomach, small and large intestine, probiotics are susceptible to degradation. The multiple-covering technique enhances their stability and viability under these challenging conditions [2]. Therefore, coated microorganisms are expected to include higher viability than that non-coated microorganisms. One of the challenges is the efficacy of coating materials in improving viability and stability of bacteria under harsh conditions. Therefore, it is critical to choose an appropriate composition for the coating and microencapsulation of probiotics [3]. Alginate and chitosan are the most widely used natural polymers for the microencapsulation of probiotics. Prebiotics have recently been addressed to improve the viability of probiotics. Prebiotics may decolonize pathogens by modulating gut diversity; thereby, improving the growth of probiotics and decreasing the number of pathogens [4]. In addition to microencapsulating probiotics, coating bacteria as a layer on microencapsulating polymers can increase bacterial resistance to digestive and storage conditions. Simultaneous microencapsulation and coating of probiotics can include a positive effect on improving cell viability.

Alginate, a carbohydrate polymer, is widely used in the food and pharmaceutical industries due to its non-toxicity, biodegradability, biocompatibility and ease of preparation. It can protect microorganisms from bile salts and stomach acid. However, alginate gels are susceptible to degradation in presence of monovalent ions, Ca²⁺ chelating agents, extreme pH levels and harsh chemical conditions, which can accelerate the release of encapsulated substances [5, 6]. Relatively, CMC is a water-soluble derivative of chitosan. In addition to cationic amine groups in chitosan, CMC contains further anionic carboxylic groups, which provide several potentials such as ampholytic characteristics. There are three categories of CMC based on the functional groups participating in the reaction such as O-CMC, N, O-CMC and N-CMC. The poor solubility of chitosan in water is one of the disadvantages of chitosan in drug delivery. The O-carboxymethyl chitosan (OCMC) effectively addresses the solubility issues associated with chitosan in aqueous solutions. Moreover, OCMC has attracted significant attentions due to its enhanced solubility, high viscosity, low toxicity and beneficial biocompatibility characteristics. Li et al. reported that ionic cross-linking through ionic interaction in alginate-CMC hydrogel could enhance the survival of Lactobacillus casei ATCC 393 against adverse conditions [7].

Inulin is a non-digestible fructan-type carbohydrate, soluble dietary fiber and includes short-chain fructooligo-saccharides (scFOS). It is widely addressed as a prebiotic because it can selectively be consumed by the gut microbiota, promoting growth of probiotic bacteria [8]. Zabihollahi et al. demonstrated that the survival of L. plantarum in carboxymethyl cellulose-based film increased significantly (36%) with the addition of inulin as a prebiotic during storage [9]. The pectin structure includes a linear chain of α-(1, 4)-linked D-galacturonic acid units, commonly identified as the homogalacturonan domain or the smooth region. As a prebiotic, pectin plays a critical role in modulating composition and metabolism of intestinal microbiota and decreasing possibility of intestinal colitis [10]. The pectin-inulin composite has enhanced the survival rate of L. casei and L. rhamnosus, compared to free cells in a simulated gastrointestinal tract (GIT) [11]. Magnesium oxide nanoparticles (MgONPs) have recently been highlighted as a potential candidate for controlled drug delivery systems due to their biocompatibility, non-toxicity, biodegradability, stability and various biomedical characteristics such as anticancer, antioxidant and antidiabetic characteristics. Using MgONPs in alginate-gelatin microgel showed significant advantages in enhancing viability of Pediococcus pentosaceus Li05 in heat treatment, simulated digestive fluid and long-term storage [12].

Researchers have implemented various strategies to increase bacterial viability, including use of prebiotics, nanoparticles and microcapsule coating separately. For example, use of MgONPs in the microencapsulation of bacteria has been studied to improve viability of probiotics in the digestive system. Research has shown that using prebiotics improves probiotics viability during the microencapsulation process and various polymers to improve the stability of probiotics. To improve probiotic viability under various harsh conditions, the present study simultaneously used prebiotics, MgONPs and OCMC as microcapsule coatings. The simultaneous use of prebiotics, MgONPs and OCMC in the microencapsulation of probiotics can improve the viability of probiotics under various harsh conditions by creating a synergistic effect. To the best of the authors’ knowledge, no studies are available that investigate bacterial viability by coating microcapsules containing MgONPs and prebiotics. The experimental design used a simplex lattice mixture method to assess the optimal percentage of OCMC, inulin and pectin. The quantities of MgONPs and sodium alginate (SA) were constant in all experiments. Drying was carried out using microwave oven. Encapsulation efficiency, survival in simulated digestive fluids, heat treatment of microencapsulated probiotics and stability in storage conditions were calculated as well.

1. Materials and Methods

2.1 Materials

The probiotic strain of L. reuteri ATCC 23272 was provided by the Iranian Scientific and Industrial Research Organization, Tehran, Iran. The OCMC (purity greater than 98%, deacetylation degree of 80%, carboxyl substitution degree of greater than 80% and amino content of nearly 1.45%) was purchased from Macklin, Shanghai, China. The MgONPs (purity greater than 99% and APS of 20 nm) were purchased from US Research Nanomaterials, Houston, TX, USA. Inulin (molecular weight of nearly 5000 Da with food-grade), low methoxyl pectin of citrus source (molecular weight of 70–140 kDa, degree of esterification of 27%) and SA (viscosity of 2000 cp, molecular weight of 80–120 kDa and M/G ratio of 1.56) were purchased from BSK Pharmaceutical, Tehran, Iran. Pancreatin from porcine pancreas (6000 FIP-U.g^-1 lipase, 350 FIP-U.g^-1 protease and 7500 FIP-U.g^-1 amylase), pepsin from porcine gastric mucosa (P700, 250 U.mg^-1), lactic acid, sodium citrate, glucose, calcium chloride, bile salts, trypticase soy broth (TSB) and trypticase soy agar (TSA), deMan Rogosa Sharpe (MRS) agar and broth (Cat no. 110660) were purchased from Sigma Aldrich, St. Louis, USA.

2.2 Probiotic Culture Preparation

The L. reuteri ATCC 23272 was incubated in MRS broth at 37 °C for 24 h. The probiotics were harvested by centrifuging at 1790 g for 10 min at 4 °C. Probiotic cells were washed twice with sterile PBS (PBS) (pH 7.4). To ensure the elimination of the supernatant from the culture broth, bacteria were recentrifuged under similar conditions and then resuspended in 2 ml of PBS [13, 14].

2.3 Preparation of Polysaccharides-based Bionanocom-posite

Various compositions have been prepared for inulin and pectin prebiotics. The ratio of inulin:pectin of 10.9:1 was chosen as the optimal concentration, which was achieved in a previous study [15]. Based on the concentrations presented in Table 1, prebiotics were suspended in deionized water with a certain concentration (Table 1), MgONPs at a concentration of 5 μg.ml^-1 were added to the mixture and stirred at 27.95 g for 60 min at 60 °C. Then, polysaccharides-based bionanocomposite was autoclaved at 121 °C for 20 min.

2.4 Microencapsulation Process

The composition of bionanocomposites (Table 1) was added to the cell suspension at a 1:1 (v/v) ratio and 37 °C and mixed using vortex to microencapsulate L. reuteri. The suspension of synbiotic was homogenized entirely to a 2% SA (w/v) solution. Cell suspensions were extruded using an insulin syringe in 0.1 M CaCl₂ solution. The interaction of uronic acid carboxylic groups with calcium ions formed a gel network, creating probiotic beads. To verify complete gelation of the beads, the CaCl₂ solution was stored at 4 °C for 30 min. Probiotic microgels were separated from the CaCl₂ solution using Whatman grade-1 filter papers (pore size of 11 μm) (Whatman, USA). These were washed twice with PBS. During the microencapsulation process, sterile conditions were addressed.

2.5 Coating Microcapsule with O-carboxymethyl Chitosan

The OCMC solution was prepared using method described by Mi et al. [16] with some modifications. Based on Table 1, OCMC solution was prepared by dissolving a certain quantity of OCMC (% w/v) in 95 ml of lactic acid solution (1% v/v). The pH was adjusted to 6 using 1 M NaOH. The solution was adjusted to 100 ml with distilled water (DW) and then filtered using Whatman grade-4 filter papers (pore size of 20 μm) (Whatman, USA). This was autoclaved at 121 °C for 15 min. Then, the alginate beads were immersed in the OCMC solution and agitated for 40 min at room temperature (RT) and 1.12 g for coating OCMC-coated microcapsules were separated using Whatman grade 1 filter papers (Whatman, USA). Then, beads were washed twice with PBS.

2.6 Encapsulation Efficiency

The number of cells released from microcapsules was calculated using method described by Halim et al. [17] with some modifications. Briefly, 1 g of beads was mixed well in 9 ml of sodium citrate (50 mM) for 10 min at RT. Cells coated with OCMC were added to the sodium citrate solution after grinding for 1 min using mortar and pestle. The released L. reuteri was diluted in PBS and then counted using pour plate method on plates containing MRS Agar. The efficiency of the extrusion was calculated using Eq. 1.

Survival rate (%)= ( ) ×100                                                      (Eq. 1)

Where, EE was efficiency of encapsulation, N was probiotics count released from beads (Log CFU.g^-1) and N₀ was the initial probiotics count added to the mixture (Log CFU.g^-1).

2.7 Survival Rate of Microencapsulated Cells after Drying

The encapsulated probiotics were dried in a microwave oven (Mwl210, Kenwood CO, UK) for 7 min with a power of 400 W. The microencapsulated cells were hydrated in PBS for 2 hours. Bacteria were serially diluted in a phosphate buffer. Probiotics were placed on an MRS agar medium by the pour-plate method. After incubation for 48 h at 37 °C, the colonies were counted. The survival rate was calculated using above mentioned Eq. 1. While, Where N: The Probiotics count after drying (Log CFU.g^-1) and N₀: The Probiotics count before drying (Log CFU.g^-1).

2.8 Survival Rate Of L. reuteri In Harsh Conditions

2.8.1 Simulated Gastrointestinal Digestion

Simulated intestinal and gastric fluids were prepared using method described by Mohamadzadeh et al. [15]. Pepsin at a concentration of 3 g.l^-1 was added to a saline solution (0.5% v/v) to prepare simulated gastric fluid. The final pH of the solution was adjusted to 2 using 1 N HCl. Simulated intestinal fluid was prepared by adding 4.5% (w/v) bile salt and pancreatin USP at a concentration of 1 g.l^-1 to a saline solution. The pH of the solution was adjusted to 8 by adding 1 N NaOH. The encapsulated-dried probiotics were added to 1 ml of simulated gastric fluid and incubated at 37 °C for 4 h. The mixture was centrifuged at 16100 g for 15 min. The supernatant was discarded and the cells were mixed with 1 ml of simulated intestinal fluid. As in the previous step, probiotics were incubated at 37 °C for 4 h and then centrifuged. After discarding the supernatant, probiotics were diluted using PBS. The released probiotics were counted. The viability under simulated gastrointestinal conditions was calculated using above mentioned Eq. 1. While, N was the probiotics count after exposure to simulated gastrointestinal digestion (SGD) (Log CFU.g^-1) and N₀ was the probiotics count before exposure to SGD (Log CFU.g^-1).

2.8.2 Heat Treatment of the Microencapsulated Probiotics

Alginate, bionanocomposite and OCMC-coated bionano-composite beads were assessed for heat resistance at 60 °C for 60 min, 70 °C for 30 min and 80 °C for 5 min. One gram of the dried beads was added to 9 ml of PBS. After heat treatment, tubes were cooled down to 37 °C and serially diluted. Cells were counted using pour-plate method and results were reported as log CFU.g^-1.

2.8.3 Long-term Storage

Alginate, bionanocomposite and OCMC-coated bionano-composite beads were dried using described methods. The probiotics stability was assessed at refrigerator temperature (4 °C) and ambient temperature (25 °C) for 6 w. Cell counts on MRS agar were carried out weekly and results were reported as log CFU.g^-1.

2.9 Characterization of the Microparticles

2.9.1 Scanning Electronic Microscopy

The microstructural characteristics and surface morphology of the dried samples, including inulin, pectin, bionanocomposites containing L. reuteri and OCMC-coated bionanocomposites containing L. reuteri, were analyzed using scanning electron microscopy (SEM). To enhance conductivity, samples were coated with a thin layer of gold and then analyzed at an accelerating voltage of up to 15 kV.

2.9.2 Fourier transform infrared spectroscopy (FTIR)

The structure of inulin, pectin, SA, OCMC, MgONPs and bionanocomposite microcapsules with and without L. reuteri and OCMC-coated bionanocomposite microcap-sules were investigated using Fourier transform infrared spectroscopy (FTIR) analysis. First, KBr spectrum was recorded as a control. Then, samples were mixed with KBr and a thin pellet was formed by compressing them at a pressure of 60 kPa for 10 min. Findings were present with a resolution of 0.5 cm⁻¹ within the wavelength of 400–4000 cm^-1.

2.9.3 X-Ray Diffraction

The X-ray diffraction (XRD) analysis was carried out using X-ray diffractometer (X’Pert MPD, Philips, the Netherlands) that used Cu Kα radiation at 40 kV, 30 mA and λ = 0.1542 nm. The diffraction patterns were recorded by monitoring diffractions with a scan speed at 0.02◦/s, within a 2θ angle range of 10–70◦ [18]. The XRD analysis was carried out on inulin, pectin, OCMC, bionanocomposite microcapsules with and without probiotics and OCMC-coated bionanocomposite microcapsules.

2.10 Experimental Design and Statistical analysis

This study used simplex lattice mixture design to optimize concentration of OCMC and various concentrations of prebiotics at a fixed ratio (inulin:pectin of 10.9:1). Table 1 presents the experimental design data. Results analysis was carried out using Design-Expert software. To assess the importance and effects of each element on the response, analysis of variance was carried out with a significance level of 95%. The coefficient of determination, R², verified validity of the regression model. All experiments were carried out in triplicate. A numerical optimization technique was used for the optimization process.

1. Results and Discussion

3.1 Encapsulation Efficiency of L. reuteri

The effects of various concentrations of inulin, pectin and OCMC on encapsulation efficiency were investigated. Table 1 shows various bionanocomposite formulations with and without OCMC coating. The encapsulation efficiency ranged from 95.35 (Run 5) to 98.13% (Run 7). The highest efficiency was achieved in absence of OCMC with a previously optimized prebiotic formulation (inulin:pectin of 10.9:1). The lowest extrusion efficiency was observed at 1% (w/v) concentration of OCMC without prebiotics. No use of OCMC resulted in easier releases of probiotics from the beads. Decreases in the viability of probiotics coated with OCMC could be due to the use of mortar and pestle to release the probiotics. The microencapsulation efficiency of alginate beads coated with chitosan decreased, compared to that of alginate beads [19]. Parsana et al. demonstrated that the encapsulation efficiency of L. reuteri in alginate beads (92.06%) was higher than that in alginate beads coated with chitosan with prebiotic inulin (90.63%) [20]. The coating process (e.g. agitation, pH changes and exposure to chitosan solution) could affect mechanical or osmotic stress that decreased integrity of the alginate matrix, resulting in decreased microencapsulation efficiency. Using inulin and pectin prebiotics in the microencapsulation of probiotics could improve probiotic growth as well. Poletto et al. showed that efficiency of the extrusion in presence of inulin (96.75%) was 2.65% higher than that in microencapsulation using alginate (94.10%) [21]. Probiotics could ferment prebiotics, providing an immediate source of metabolic energy. Even before reaching the gut, inulin and pectin could serve as nutritional reserves within the microcapsule. Efficiency of the extrusion in alginate beads was 95.23%, which showed a decrease of 1.29%, compared to the average efficiency of various formulations in the experimental design. Zaeim et al. reported that the efficiency of probiotics microencapsulated with alginate (98.12%) was 1.21% higher than that of microencapsulated probiotics with alginate-chitosan (96.91%) [22].

In Table 2, a positive value indicates a synergistic effect and a negative value indicates an antagonistic effect on the response. Based on statistical analysis, total inulin-pectin concentration and OCMC alone positively affected the extrusion efficiency. The inulin-pectin concentration was more effective than the OCMC concentration. The interaction of two variables included negative effects on the microencapsulation efficiency. The lack of fit in all models was greater than 0.1 with insignificancy. The microencapsulation efficiency was predicted using cubic model with a value of p < 0.001 (Table 2).

3.2 Survival Rate of Microencapsulated Cells after Drying

The drying efficiency of probiotics using microwave oven is shown in column Y₁ of Table 1. The highest survival efficiency was 99.41 (Run 4) and the lowest survival efficiency was 98.05% (Run 1). Protein denaturation, fatty acid (FA) oxidation, DNA damage and free radical formation were factors that decreased viability of the probiotics due to thermal drying. The lowest drying survival of L. reuteri occurred in absence of OCMC. Survival of bacteria in encapsulation depended on the thickness and type of coating materials [23]. Without OCMC, the coating depended on lighter materials such as alginate and prebiotics, which did not form an equally robust barrier, resulting in thinner layers. The presence of OCMC significantly increased viscosity and density of the outer layer, creating a thicker, further cohesive protective coating around the microcapsules ‌against thermal stress. Drying efficiency of the alginate beads was 95.98%. A 2.7% decrease in the viability of probiotics was observed, compared to an average of various formulations in the experimental design. The absence of inulin, pectin and OCMC as a coating layer decreased thickness of the bacterial coating layer and hence caused further heat damages to the probiotics. Jantarathin et al. showed that in addition to increasing the viability of cells during the drying process, chitosan improved the viability of bacteria after drying [24].

Based on the data fitting, the cubic model produced the lowest p-value regarding viability after drying (Table 2). The model significance was verified with a p-value < 0.0001. Two variables, total inulin-pectin concentration and OCMC, positively affected drying efficiency and the effect of OCMC concentration was greater than that of the inulin-pectin concentration. Decreasing the survival rate of probiotics during drying included a negative effect on the effectiveness of probiotics. The 99.41% viability of L. reuteri in the microwave process was a significant efficiency.

In addition to choosing the appropriate materials for microencapsulation, drying method significantly improved the viability of probiotics. Using microwave oven is a biocompatible method that can maintain the viability of probiotics at high rates within short times. Drying in microwave ovens increases product purity, improves the quality of the pharmaceutical powder and decreases byproducts and energy consumption [25]. Microwave ovens serve as innovations for drying food and pharmaceutical products.

3.3 Viability of Probiotics in Simulated Gastrointestinal Conditions

The viability rate of L. reuteri (Table 1, column Y₂) varied between 87.11 (Run 5) and 89.3% (Run 6). The lowest survival rate of probiotics was observed at 1% (w/v) concentration of OCMC in absence of prebiotics within the bionanocomposite structure. The MgONPs could serve as a buffering agent, enhancing the survival rate of probiotics by decreasing the acidity in the stomach. The lack of MgONPs release due to the OCMC coating with a concentration of 1% (w/v) could be a reason for decreasing the viability of probiotics in SGI conditions. The highest survival rate of probiotics in SGI conditions was achieved at a concentration of 0.5% (w/v) of OCMC and 0.5% (w/v) of prebiotics (inulin:pectin of 10.9:1). Combining OCMC with prebiotics provided mechanical protection and metabolic support to the probiotics, which could include a synergistic effect. A decrease of 2.87 log CFU.g^-1 of the microencapsulated cells in alginate microcapsules was observed in the simulated digestive fluids. The survival rate of alginate beads in the SGI conditions was 71.49%, which was 17% less than the average efficiency of various formulations in the experimental design. Afzaal et al. demonstrated that the survival of L. acidophilus ATTC 4356 microencapsulated within alginate in the simulated gastric fluid was associated to a decrease of 3.57 log CFU.ml^-1 [26]. The resistance of Lactobacillus strains at low pH could be attributed to F₀F₁-ATPase activity in probiotics [27]. The survivability of alginate-encapsulated probiotics significantly decreased by 2.26 ±0.24 log CFU.g^-1 [28]. Based on the data fitting, the cubic model produced the lowest p-value (0.0003) regarding the SGI conditions (Table 2). The model significance was verified with a p-value < 0.001. Three variables, including the total inulin-pectin concentration, concentration of OCMC and interactions of variables (AB), positively affected cell viability in SGI conditions. Similar to extrusion efficiency, inulin-pectin concentration was more effective than OCMC concentration. The effect of AB (A-B) parameters negative affected the viability of probiotics.

 3.4 O-Carboxymethyl Chitosan and Prebiotic Concentr-ation Optimization

Optimizing the inulin and pectin as prebiotics, as well as the OCMC quantity for coating microcapsules, is critical. Three parameters of encapsulation efficiency, survival rate of microencapsulated cells after microwave drying and survival rate of L. reuteri in SGI conditions were selected for optimization. The efficiency of encapsulation is important due to the presence of probiotic microgels in food, pharmaceutical and cosmetic industries. Optimizing viability after drying greatly affected the viability of dried powder and improved the long-term stability of probiotics. Optimizing survival in simulated digestive conditions enhanced the delivery of probiotics to the clone by increasing its survival in these conditions. Quantities of inulin and pectin with a fixed ratio (inulin:pectin of 10.9:1) and the concentration of OCMC were assessed to coat the microgels in a 0–1 g range. Optimization was achieved when the encapsulation efficiency, viability after microwave drying and viability after exposure to SGI conditions were simultaneously at their highest levels. After defining the highlighted conditions in Design-Expert software, the ratio of inulin to pectin was 32% (inulin, 0.294, and pectin, 0.026) and the OCMC concentration was reported as 68% (Table 3). The quantities of MgONPs and SA were similar to them in all compositions. Optimization was valid when the desirability function was acceptable. The desirability was 90%. The optimization results were verified by carrying out experiments. The optimization results revealed that the OCMC-coated bionanocomposite could serve as a novel approach to protect cells and enhance survival of bacteria in harsh conditions in food, pharmaceutical and cosmetic industries.

3.5 Heat Treatment of Microencapsulated Probiotics

To investigate the heat stability of probiotics, alginate, bionanocomposite and OCMC-coated bionanocomposite microcapsules were assessed at three various temperatures and times. Microcapsules were assessed at 60, 70 and 80 °C for 60, 30 and 5 min, respectively. It was observed that bionanocomposite and OCMC-coated bionanocomposite beads included significant resistance, compared to that alginate beads did. The viability of probiotics in OCMC-coated bionanocomposite microcapsules was higher than that in bionanocomposite microcapsules. In microcapsules coated with OCMC, probiotic viability decreased by 1.46 log CFU.g^-1 at 80 °C for 5 min. In comparison, decrease of L. reuteri viability in alginate beads in similar conditions was 5.91 log CFU.g^-1. Results demonstrated that OCMC was an appropriate coating polymer for probiotics against thermal treatments. The viability loss of the micro-encapsulated L. plantarum EMCC1039 with chitosan-coated alginate was 3.06 log CFU.g^-1 after exposure to 65 °C for 30 min [29]. Cheow et al. demonstrated that the viability of L. rhamnosus GG coated with alginate-chitosan decreased by 5.9 log CFU.ml^-1 after incubating at 60 °C for 30 min [30]. Alginate beads demonstrated higher sensitivity to temperature. The highest decrease in the viability of pro-biotics occurred in bacteria encapsulated with alginate beads. Halim et al. reported that P. acidilactici ATCC 8042 encapsulated in alginate was destroyed at 60 °C for 60 min [31]. Based on the thermal stability results in Table 4, OCMC-coated bionanocomposite beads demonstrated higher resistance and promise stability to severe harsh ther-mal conditions, compared to those other microcapsules did.

3.6 Stability of Encapsulated Probiotics in Long-term Storage

Assessment of the probiotic stability of micro-encapsulated L. reuteri in various time and temperature conditions is shown in Table 5. Due to a better cell protection of bionanocomposites during the extrusion and drying processes, the initial number of cells coated with bionanocomposite was higher than that with alginate microcapsules. Over time, the death slope of probiotics decreased; thus, the highest decrease in L. reuteri viability was recorded within the first week. The highest rate of viability loss of probiotics, with a value of 3.02 log CFU.g^-1 at 25 °C, was linked to alginate beads. The lowest decrease in survival within the first week was observed with OCMC-coated bionanocomposite microcapsules at 4 °C with a value of 0.47 log CFU.g^-1. Using OCMC to coat bionano-composites improved the stability of probiotics by 0.06 and 0.42 log CFU.g^-1 at 4 and 25 °C after 42 d. Microencapsulated probiotics in OCMC-coated bionano-composite recorded significantly higher viability, compared to that those in alginate microcapsules did. Qi et al. concluded that the survival rate of L. rhamnosus GG microencapsulated in alginate/chitosan at 25 °C after 42 d was nearly 1 log CFU.g^-1 higher than that of probiotics microencapsulated in alginate was [32]. Adding chitosan significantly increased the 35-d storage stability of L. acidophilus NCIMB 701748 dried powders [33]. The survival of probiotics encapsulated with OCMC-coated bionanocomposite was 3.25 and 2.93 log CFU.g^-1 higher than that of alginate microcapsules at 25 and 4 °C after 42 d. A decrease of 2 log CFU.g^-1 was reported for microencapsulated probiotics in alginate/bentonite nanocomposite at 25 °C after 14 d [34]. In general, the viability of probiotics decreased with increasing temperature. In all microcapsules, the viability of probiotics at 4 °C was higher than that at 25 °C. Shu et al. demonstrated that probiotic viability decreased at high temperatures due to increased intracellular water activity and membrane oxidation stress.

Analysis of the stability results for L. reuteri was carried out using SPSS statistics software, which indicated statistically significant results (p-value < 0.01). Results indicated that OCMC polymer was appropriate for coating microencapsulated probiotics; thereby, enhancing their stability. Microencapsulated probiotics with OCMC-coated bionanocomposite were promising to increase the microorganism stability.

3.7 Characterization of the Microparticles

3.7.1 Surface Morphology

The SEM images of inulin, pectin, beads containing L. reuteri and OCMC-coated microcapsules containing L. reuteri are shown in Figure 1. The surface morphology indicated that inulin possessed a spherical structure. The physicochemical characteristics of the mixture could be affected by various sizes of inulin. Strength of the gel structure was directly linked to the size of inulin particles. Inulin with a higher molecular weight was further resistant to hydrolysis and was further stable [35]. The SEM images showed that pectin included a non-spherical structure. The quantity of moisture in the particles and various extraction methods could affect the size of particles such as inulin and pectin. Drying the microgels using microwave oven resulted in the dehydration of the beads within a short time. Deng et al. showed that drying carbohydrate compounds using microwave ovens could improve gelling characteristics of the compounds in addition to maintaining the structure [36]. No probiotics were observed on the surfaces of microcapsules containing bacteria. The high efficiency of the entrapment of L. reuteri (96.43%) could be a reason for the absence of L. reuteri on the surface of the beads. moreover, OCMC coating on microcapsules could be another reason for the absence of bacteria on the surface of beads. The surface structure of uncoated microcapsules containing L. reuteri was non-porous and cohesive. The cohesive structure of the microcapsules was attributed to the egg-box structure formed by specific and strong interactions between Ca²⁺ and the G-blocks of alginate. Optimal drying resulted in no cracks or pores in the microcapsule structure. The absence of pores prevented the penetration of acids, hydrogen ions, bile salts and enzymes and improved the stability of probiotics in harsh conditions. Presence of MgONPs could also create a non-porous structure by filling the nanometer pores. The OCMC-coated microcapsules shrank and wrinkled during the drying process. The peaks and valleys in dried OCMC-coated microcapsules could be attributed to the ionic interaction between alginate and OCMC. Alginate cross-linking with OCMC in a hydrogel could increase stability and improve microgel structures. The ionic interaction between carboxyl residues in alginate and amino residues in OCMC formed a polyelectrolyte complex.

3.7.2 Fourier-transform Infrared Spectroscopy Analysis

Inulin, pectin, SA, OCMC, MgONPs, bionanocomposite microcapsules with and without L. reuteri and OCMC-coated bionanocomposite microcapsules were assessed via FTIR analysis (Figure 2). The presence of an absorption band at 3640 cm⁻¹ in SA and at 3700 cm⁻¹ in MgONPs showed O–H stretching vibrations that did not involve hydrogen bonding [37]. The absorption band of 3400–3450 cm⁻¹ in all three microcapsules was linked to hydrogen bonding in hydroxyl groups. These bands indicated presence of moisture in the microcapsules. The peak in 3000–3600 cm⁻¹ wavenumbers range in inulin, pectin and OCMC indicated the stretching vibrations of O-H groups. The peaks in the wavenumber range of 2800–3000 cm⁻¹ observed in all analyses, except for MgONPs, associating to C–H stretching vibrations. Peaks at 1650, 1140 and 1000 cm⁻¹, respectively, represented vibrations of C=O, C–C and C–O–C groups in the inulin structure [38]. The peak at a wavenumber of 1745 cm⁻¹ in pectin was linked to a carbonyl group. The peak at 1610 cm⁻¹ was due to the asymmetric stretching vibrations of COO^- ions in the C=O group. In SA FTIR, 1410 and 1030 cm⁻¹ peaks corresponded to symmetric stretching vibrations of carboxylate ions and C–O–C, respectively. In OCMC FTIR, the peak at 1610 cm⁻¹ corresponded to the N–H bending of primary amines and the peak at 1430 cm⁻¹ corresponded to the C–N group. Peak 1320 cm⁻¹ was linked to C–O group vibrations. Peak 1060 cm⁻¹ represented the C–O stretch of –CH2–OH in primary alcohols. The peak at a wavenumber of 1640 cm⁻¹ represented carbonyl vibrations in MgONPs. The 1435 cm⁻¹ peak corresponded to the vibrational activity of water molecules on the surface of MgONPs. The peak at 555 cm⁻¹ in nanoparticles was associated to Mg–O bending vibrations [39].

The absorption range of stretching vibrations associated with hydroxyl bonds in microcapsules was denser than that of SA. A reason for this difference was the participation of carboxylate and alginate hydroxyl groups with calcium ions to create the egg-box structure. The egg-box model could improve the stability of microcapsules by creating a compact structure. All three types of microcapsules showed high similarities in their FTIR analysis. The difference between microcapsules included presence or absence of L. reuteri and OCMC. Presence of bacteria in the bionanocomposite caused a slight difference in FTIR analysis. The peak at 1235 cm^-1 in microcapsules containing L. reuteri could be attributed to the amide III band of cytoplasmic and membrane proteins in L. reuteri. Adding OCMC led to the formation of new hydrogen bonds between the –NH₂ and –COOH groups of OCMC and the C=O and –OH groups of alginate. Transfer of peaks from 1437 and 1625 cm^-1 to 1428 and 1620 cm^-1 could be a reason for the formation of bonds between OCMC and calcium alginate. The slight shift of the carboxylate bands to lower wavenumbers was due to the sharing of bonds with amine groups in OCMC on the surface of the microcapsules. The increase in intensity of the peak at 1620 cm^-1 verified formation of strong polyelectrolyte complexes [40].

3.7.3 The X-ray Diffraction Analysis

The XRDs of inulin, pectin, OCMC, bionanocomposite microcapsules with and without L. reuteri and OCMC-coated bionanocomposite microcapsules are shown in Figure 3. The X-ray analysis was used to assess the amorphous or crystalline structure of polymers. The XRD analysis showed that the structures of inulin and pectin were amorphous and crystalline, respectively. Non-sharp peaks in the inulin structure and sharp peaks in the pectin structure showed a significant structural difference between these prebiotics. The six sharp peaks at 2θ = 27.623^◦, 31.976^◦, 38.178^◦, 45.671^◦, 56.679^◦ and 66.474^◦ were important regions of the OCMC, representing that this polymer structure is crystalline. The XRD analysis demonstrated that structures of bionanocomposite microcapsules with and without L. reuteri were amorphous. One of the reasons for the amorphous structure of beads was the cross-linking of SA and pectin with Ca²⁺ ions. Presence or absence of L. reuteri include no significant effect on the amorphous structure of microcapsules. Addition of crystalline OCMC included a significant effect on the structure of amorphous microcapsules. The OCMC coating caused the formation of semi-crystalline structures in the microcapsules. The ionic interaction between amino groups in OCMC and carboxyl groups of alginate could be one of the reasons for changing the structure of microcapsules.

Rao et al. demonstrated that calcium alginate-pectin microcapsules containing L. paraplantarum LR-1 included an amorphous structure. Presence of polymers with various concentrations in the microcapsule, as well as the method of microencapsulating probiotics, could affect intensity of the peaks and thus structure of the microcapsules [41]. Generally, polymers with semi-crystalline structures included higher temperature resistance than that amorphous structures did. Menegazzi et al. showed that use of polymers with a semi-crystalline structure decreased heat and oxygen transfer into the microcapsule and improved the viability of probiotics [42]. Based on Section 3.5, it was observed that the semi-crystalline structure of microcapsules included higher temperature resistance in all three temperatures of 60, 70 and 80 °C and hence higher L. reuteri survival was recorded. Additionally, semi-crystalline morphologies showed higher chemical resistance and biocompatibility, compared to that amorphous polymers did.

1. Conclusion

The present study focused on the optimization and assessment of an OCMC-coated bionanocomposite with inulin and pectin as prebiotics with MgONPs at constant concentrations. Based on the optimization result, concentration of 68% of OCMC with 29.4% inulin and 2.6% pectin could include the highest protection of L. reuteri in various conditions. Results showed that OCMC included high-heat protections from probiotics. High thermal stability of probiotics revealed that the use of OCMC-coated bionanocomposites could be effective in preparing functional foods, where bacteria were usually exposed to high temperatures. The stability of alginate beads decreased respectively at 4 and 25 °C, 3.25 and 2.93 log CFU.g^-1 after 42 d, compared to the coated probiotics. Physicochemical analysis of the optimal microcapsule revealed a coherent compact structure, which enhanced the stability and survival of bacteria in harsh conditions. Microencapsulation using the investigated bionanocomposite could increase the viability of probiotics during the production, formulation, storage and packaging processes. Further studies on the binding of OCMC to intestinal epithelial cells, in-vivo assessment of the release rate of microencapsulated probiotics and investigation of the viability of microencapsulated probiotics in OCMC-coated bionanocomposites in food products such as juices or fillets validate these findings.

1. Acknowledgements

The authors would like to express their thanks to the Department of Drug and Food Control, Pharmaceutical Quality Assurance Research Center, Faculty of Pharmacy, Tehran University of Medical Sciences, for the support it provided for this study.

1. Conflict of Interest

The authors declare no competing interest.

1. Authors’ Contributions

Mohamadsadegh Mohamadzadeh, conceptualization, validation, investigation, visualization and writing original draft, formal analysis, software; Ebrahim Vasheghani Farahani, supervision, review and editing; Ahmad Fazeli, conceptualization, supervision, review and editing; Seyed Abbas Shojaosadati, conceptualization, validation, supervision, review and editing.

1. Using Artificial Intelligent Chatbots

No artificial intelligence chatbots were used in this study.

1. Ethical Consideration

This study did not receive specific grants from funding agencies in the public, commercial and not-for-profit sectors. 

Abbreviations: L, Lactobacillus; CMC, carboxymethyl chitosan; NPs, nanoparticles; SGI, simulated gastrointestinal conditions; CFU, colony forming units; SEM, scanning electron microscopy; FTIR, Fourier transform infrared spectroscopy; XRD, X-ray diffraction; scFOS, short-chain fructooligosaccharides; BNC, bacterial nanocellulose; SA, sodium alginate; MRS, De Man, Rogosa and Sharpe; TSB, tryptic soy broth; TSA, tryptic soy agar.