EP4665166A1 - Bacteria cultures for plantbased applications - Google Patents

Abstract

The present invention relates to fermented plant-based foodstuff products and methods of producing them through fermentation. Specific bacteria are disclosed, which are useful for commercial manufacturing of plant-based products.

Description

BACTERIA CULTURES FOR PLANT-BASED APPLICATIONS

FIELD

The present disclosure generally relates to fermented plant-based products and methods of producing them through fermentation.

BACKGROUND

The production of dairy cheese is an ancient practice that started as a way of preserving milk and turned into the development of fermented products with a broad range of flavors and interesting textures that are nowadays widely consumed on a regular basis. The characteristic organoleptic properties of cheese are directly linked to the production processes, the performance of different microorganisms, and most importantly, the nature of dairy milk. This colloidal dispersion of fat globules stabilized by casein micelles presents extremely particular behavior upon heat treatment or acidification, and its replication with plant raw materials a difficult challenge to overcome.

In the production of cheese, milk is heat-treated to ensure safety and inactivate endogenous microorganisms that could be present in the raw product. During this process, proteins denature and unfold. Microorganisms, e.g. lactic acid bacteria, are then inoculated and start fermenting the milk, namely transforming lactose into lactic acid, and thus acidifying the media. This pH drop complements the action of rennet, encouraging proteins to interact with each other and start forming a protein network known as curd, that entraps fat globules in its pores and conforms the firm texture of cheese. Milk casein is responsible for the formation of this three-dimensional network, and its characteristic molecular structure is responsible for its versatility to confer liquid but also gel-like textures. For these reasons, knowledge on plant protein behavior under different processing conditions is required to develop high quality products.

The demand for plant-based alternatives to dairy cheese is exponentially growing, mostly due to an increasing awareness of sustainability and the large amount of resources that animal protein production requires, and consumers are looking for products that can provide a similar experience to eating dairy cheese but are based on plant-based raw materials.

Two of the most important challenges in plant-based cheese production are texture and flavor development, and producers mostly rely on functional ingredients such as texturizers and flavoring agents for the development of these attributes. However, the currently commercialized plant-based cheeses do not meet consumer satisfaction (Short, Kinchia and Nolden, 2021) and therefore there is still room for improvement in this category. Fermentation has great potential to modify and improve the physicochemical and sensory properties of proteins, thus creating pleasant textures and flavors that might improve the quality of plant-based cheese.

Currently available plant-based cheeses in the market, and also previous studies on plant protein gels, include coconut oil (Grasberger, Gregersen, Jensen, Sanggaard and Corredig, 2021) or/and hydrocolloids such as agar agar, carrageenans (Nunes, Raymundo and Sousa, 2006; Ferawati, Hefni, Ostbring and Witthoft, 2021), or gums such as guar, xanthan or arabic gum (Ferawati et al., 2021 ; Shen and Li, 2021), to achieve a firm texture. 74% of the plant-based cheeses in the UK market are coconut oil-based (Nicolas Saraco and Blaxland,

2020). This fat is preferred because of its high melting point, which makes it remain solid at ambient temperature and therefore, provide firmer textures. Furthermore, it can melt when the temperature increases, namely in the mouth or during cooking. In addition, when adding agents such as carrageenans and other hydrocolloids, it is expected to obtain gel-like structures when including them in a liquid plant protein matrix that undergoes hydration and further heat treatment. The behavior of these ingredients is already well known, and still the textural properties of most plant-based cheeses are not fully satisfactory.

Therefore, there is an increasing interest in new ways of producing gels where the plant proteins are the main structural units of the gel structure. This is something that can be achieved through fermentation. When lactic acid bacteria are grown in a plant protein matrix, they produce lactic acid, thus decreasing the pH to values near the isoelectric point of the plant proteins. The protein net charges are then close to zero and they start aggregating and forming a curd that retains water and oil inside its pores. This three-dimensional structure is able to remain firm without the need for solid fats or texturizing agents.

The flavor profile of the currently commercialized plant-based cheeses is achieved by cheese flavoring agents or by the characteristic flavor of the plant raw material (Short et al.,

2021). The latter one implies the presence of beany, nutty, or earthy aftertaste with low acceptance levels among consumers. The potential of fermentation in flavor development is double: it can reduce the off-flavor from the raw material and it can boost the intensity of acid and dairy like notes reminiscent of animal milk products. The flavor profile of dairy cheese is formed by a complex mixture of very diverse volatile organic compounds (VOC) and it depends on factors such as VOC and precursors already present in milk and the metabolism of the bacteria that are fermenting it, among others. Therefore, replicating such a complex VOC mixture with different starting raw materials is a challenge since the precursors of the typical VOC of dairy cheese might not be found in plant protein matrices. VOC products of the carbohydrate metabolism are easier to obtain through plant protein matrices. Supplementation of those with simple sugars enables bacteria to convert them into diacetyl or acetoin (Marilley and Casey, 2004) the same way they do with lactose. VOC products of protein metabolism are produced after protein hydrolysis occurring during cheese maturation, and, as well as VOC products of fat metabolism, they are directly linked to the nature of those proteins and lipids that are present in animal milk. Multiple studies have investigated the potential of different bacteria to modify the volatile compounds in different legume proteins, many of them focusing on pea protein, and have reported improved flavor profiles (Schindler, Zelena, Krings, Bez, Eisner and Berger, 2012; Shi, Singh, Kitts and Pratap-Singh, 2021 ; Youssef, Bonnarme, Fraud, Peron, Helinck and Landaud, 2020; Ben-Harb, Saint-Eve, Panouille, Souchon, Bonnarme, Dugat-Bony and Irlinger, 2019a).

Thus, it would be advantageous to provide bacterial cultures and fermentation processes resulting in plant-based cheese with both improved texture properties and flavor profile.

SUMMARY

The present disclosure provides different blends of bacteria starter cultures and optionally bacteria adjunct cultures for texture development, reduction of beany flavor, and production of dairy-like volatile organic compounds in plant-based protein matrices derived from pea, and method of making foodstuff by fermentation.

According to a first aspect, a method for producing a fermented plant-based product is provided, the method comprising a) adding a starter culture comprising Streptococcus thermophilus and Lactobacillus bulgaricus to a plant base, b) optionally adding an adjunct culture comprising Lactobacillus helveticus to the plant base, c) fermenting the plant base for a period of time until a target pH is reached, thereby producing the fermented plantbased product; wherein the plant base is a pea base.

In one embodiment, the starter culture further comprises Lactobacillus acidophilus, Lactobacillus paracasei and Bifidobacterium.

In one embodiment, the adjunct culture further comprises Pediococcus and Lactobacillus plantarum.

In one embodiment, the adjunct culture is added at the start of the fermentation period. In one embodiment, the plant base is a liquid base with plant protein content of 5-15% (w/w), such as 5, 6, 7, 8, 9, 10, 11 , 12, 13, 14 or 15 % (w/w).

In one embodiment, the plant protein content is 10 % (w/w).

According to a second aspect, a plant-based product obtainable by the method according to the first aspect, is provided.

According to a third aspect, a fermented plant-based product is provided, which has a pea protein content of 5-15% (w/w), such as 5, 6, 7, 8, 9, 10, 11 , 12, 13, 14 or 15 % (w/w) or higher than 15% and which comprises Streptococcus thermophilus and Lactobacillus bulgaricus, and optionally Lactobacillus helveticus.

In one embodiment, the fermented plant-based product further comprises Lactobacillus acidophilus, Bifidobacterium lactis, and Lactobacillus paracasei.

In one embodiment, the plant-based product further comprises Pediococcus and Lactobacillus plantarum.

DEFINITIONS

As used herein, Vega™, NU-TRISH®, L.CASEI 431®, and LGG® are trademarks of Chr. Hansen A/S.

The products Vega™ Premium, Vega™ Vibe, Vega™ Harmony, Vega™ Classic, Vega™ Boost HA, Vega™ NU-TRISH®, Vega™ Boost PA, Vega™ NU-TRISH® L.CASEI 431®, Vega™ Boost LP and LGG® are well defined, quality controlled cultures, readily available as commercial products from Chr. Hansen A/S.

In the present disclosure the term "starter culture" is a culture which is a preparation (composition) of one or more bacterial strains (such as lactic acid bacteria strains) to assist the beginning of the fermentation process in preparation of fermented products such as various foods, feeds and beverages.

The term "adjunct culture" is a culture which is a preparation (composition) of one or more bacterial strains (such as lactic acid bacteria strains) to further assist the fermentation process in preparation of fermented products such as various foods, feeds and beverages.

The term "lactic acid bacteria" ("LAB") designates food-grade bacteria producing lactic acid as the major metabolic end-product of carbohydrate fermentation. These bacteria are related by their common metabolic and physiological characteristics and are usually Gram positive, low-GC, acid tolerant, non-sporulating, non-respiring, rod-shaped bacilli or cocci. During the fermentation stage, the consumption of carbohydrate by these bacteria causes the formation of lactic acid, reducing the pH and leading to the formation of a protein coagulum. These bacteria are thus generally responsible for the acidification of milk and for the texture of various dairy products. The industrially most useful lactic acid bacteria are found within the order "Lactobacillales" which includes Lactococcus spp., Streptococcus spp., Lactobacillus spp., Leuconostoc spp., Pediococcus spp. and Propionibacterium spp. These are frequently used as food cultures alone or in combination with other lactic acid bacteria.

The term "milk” or ’’dairy” is to be understood as the lacteal secretion obtained by milking any mammal, such as cows, sheep, goats, buffaloes or camels. In a preferred embodiment, the milk is cow's milk. The term milk also includes protein/fat solutions made partly or exclusively of plant materials.

The term "dairy analogue" or “plant-based” product as used herein is meant to refers to dairy-like products, which are products used as culinary replacements for dairy products, prepared where one or more milk constituents have been replaced with other ingredients and the resulting food resembles the original product. The milk constituents are replaced completely or substantially with plant material, for example, using planted-based milks derived from legumes (such as soybeans, pea, lentils or chickpeas), nuts (such as coconut), cereals (such as oat).

The term “plant base”, is used to describe the plant material used as a base for fermentation.

BRIEF DESCRIPTION OF THE FIGURES

In the following detailed description, references are being made to the following figures.

Figure 1 is a graph showing acidification of fermented Pea Protein Isolate (PPI) gels with different bacterial blends and their acidification slopes.

Figure 2 is a graph showing gel firmness of samples fermented with the same culture/adjunct combinations. Results are displayed as force values (g).

Figure 3 shows spectra of volatile compounds related to beany flavor detected in fermented PPI samples whose production was significantly affected by the different bacterial blends. Results are displayed as signal-to-noise (S/N).

Figure 4 shows spectra of Volatile compounds found in dairy cheeses and also detected in fermented PPI samples whose production was significantly affected by the different bacterial blends. Results are displayed as S/N. DETAILED DESCRIPTION OF THE INVENTION

This disclosure relates to plant-based fermented foodstuff, and in particular plant-based cheese analog products.

Texture and flavor development in plant-based cheese analog products generally relies on the use of texturizing and flavoring agents. However, fermentation can be a powerful tool to improve those sensorial attributes.

Acidification performance of bacteria used for fermentation is an important criterion to determine whether they are commercially favorable or not. Furthermore, how the bacteria contribute to gel hardness of the plant base, and their ability to mask off-flavors and develop dairy-like flavors are important advantages, which can be monitored through pH measurement, compression tests, and volatile compounds analysis, respectively.

EXAMPLES

EXAMPLE 1

Twenty-four different bacterial blends with four different starter cultures and six different adjunct combinations were designed to ferment a pea protein emulsion. Their acidification performance, contribution to gel hardness, and their ability to mask off-flavors and develop dairy-like flavors were evaluated through pH measurement, compression tests, and volatile compounds analysis, respectively.

Materials and Methods

Pea protein isolate (PPI) ProFam®580 (ADM, Chicago, Illinois, United States), sunflower oil (Ollineo, Budapest, Hungary), sucrose, and dextrose (Sigma Aldrich, S0borg, Denmark) were used for the preparation of the PPI matrix. All bacterial cultures used in this example are from Chr. Hansen A/S (H0rsholm, Denmark). Table 1 shows an overview of the strains included in each culture and each adjunct combination.

Table 1. Starter cultures and adjunct combinations for bacterial blends.

Matrix preparation

The pea protein isolate (PPI) matrix was prepared as previously described ((Masia, Jensen, Petersen and Buldo, 2022)). 10% pea protein isolate was suspended in a 1 % glucose and 1% sucrose water solution at 8,100 rpm with a L5M Laboratory Mixer (Silverson, Chesham,

United Kingdom) for 2 min. The protein suspension was emulsified with 10% sunflower oil under the same mixing conditions and homogenized with high pressure in a GEA Lab Homogenizer PandaPLUS 2000 (GEA, Parma, Italy) at two stages (150; 50 bars) in one pass. The matrix was pasteurized at 90°C for 20 minutes and cooled down to 43°C for further inoculation.

Matrix fermentation

Four starter cultures and five adjunct cultures (plus a negative control) were combined in a total of 24 different blends (Table 2). The samples were inoculated with 0.02% starter culture and a total of 0.02% adjunct cultures, regardless of the amount of adjunct cultures per combination. The inoculated samples were incubated at 43°C for 8 hours and further stored under refrigeration. pH was measured every 5 min using an iCinac (AMS S.R.L., KPM Analytics, Rome, Italy) and the slope of the acidification curves was calculated with Matlab (Math- Works, Natick, Massachusetts, United States) and used for statistical analysis (Table 3).

Table 2. Overview of blends

Table 3. Acidification slopes as shown in Fig. 1

Texture analysis The texture of the fermented gels was analyzed after 7 days of storage under refrigeration with a compression test in a Texture Analyzer (Stable Micro Systems, Surrey, United Kingdom). Each sample was cut in cylinder shape of 2 cm x 2 cm and compressed 5 mm with a plate geometry of 40 mm using a 5 kg load cell at 1 mm/s. The trigger was set to 15 g and the data acquisition rate was 500 pps. No oil was applied on the probe or in the sample prior to compression.

Identification of volatile compounds

The volatile compounds produced after fermentation and 7 days of storage under refrigeration were analyzed by head space solid phase microextraction gas chromatography coupled to mass spectrometry (HS-SPME-GC-MS). The instrument was a Multi Purpose Sampler (Gerstel, MSCI, Skovlunde, Denmark), with a 7890B GC (Agilent Technologies, Denmark) and a 5977A MS (Agilent Technologies, Denmark). VOCs were extracted by SPME using a DVB/Car/PDMS-fiber (Supelco#57299, VWR, Denmark) for 20 min. at 60°C, desorbed splitless at 270°C onto a TenaxTA-filled liner (Gerstel#012438, MSCI, Skovlunde, Denmark) kept at -30°C. After fiber desorption, the TenaxTA-filled liner were heated to 300°C and the trapped VOCs transferred splitless and separated on a DB-5MS Ul column 30m x 0.25mm x 1 pm (Agilent#122-5533UI, Agilent Technologies, Denmark) at 170 kPa constant pressure using helium as carrier gas. Oven temperature program was as follows: starting at 32°C/2min - increased to 102°C@10°C/min - further increased to 145°C@5°C/min - further increased to 200°C@15°C/min - further increased to 200°C@15°C/min - further increased to 280°C@20°C/min - hold at 280°C for 5 min. The mass spectrometer operated in electron impact mode at -70eV and the analyzer was scanning from 29-209 amu.

NIST 17 library search and Retention Indexes were used for identification of VOCs. Feature extraction was done using MassHunter Quantitative Analysis (Version 10.2, Build 10.2.733.8, Agilent Technologies, Denmark) and results calculated as peak height divided by baseline noise (signal-to-noise, S/N). Samples were analyzed in biological duplicates after fermenting 3 grams of pea protein matrix directly in a 20 ml headspace vial. A removal ratio was calculated for the beany off-flavors comparing the detected S/N values of those in the fermented samples to those in the non-fermented PPI matrix.

Statistical analysis

Acidification, gel firmness, and chromatography results were analyzed with an analysis of variance with a factorial to degree design, where the effect of the starter cultures, the effect of the adjuncts cultures combinations and the interaction between both were evaluated. Tukey tests were performed to define homogeneous subsets showing significance. The software used for the statistical analysis was JMP Pro 16 (SAS Institute, Cary, North Carolina, United States), and p values of less than 0.05 were interpreted as significant differences. All texture analysis were performed in biological triplicates and the analysis of volatile compounds was performed in duplicates.

Results and Discussion

The following is an overview of the results according to Example 1 , with reference to the figures. Acidification profiles

All bacterial combinations showed the same acidification trend, with a pH drop between the third and fourth incubation hours (Fig 1). The slopes of the acidification curves were considered to compare the performance of the blends. Fast acidifying combinations were preferred for food safety reasons. The faster the pH drop, the lower the chances of encouraging background growth, since a lower pH will not be a favorable condition for endogenous flora of e.g. Bacillus cereous to proliferate. Samples fermented with blends including Vega™ Harmony and Vega™ Classic showed significantly faster acidification. This is certainly influenced by the strains present in the blend. Samples containing Vega™ Vibe acidified significantly slower than the other three starter cultures. Regarding the adjunct combinations, samples without any adjunct cultures, namely those fermented just with starter cultures, showed significantly faster acidification than samples fermented with adjunct combination C4 (L helveticus, Pediococcus, L plantarum, and L casei).

A potential explanation could be that the adjuncts have a slower acidification capacity and since they would consume part of the substrate that is available for the strains present in the starter culture, this one has less substrate to grow and produce more acid, causing slower pH drops. Amongst the three aspects studied in this example, fast acidification was prioritized, since plant-based raw materials are microbially active and it is important to ensure safety by avoiding the growth of undesired microorganisms. Moreover, from an industrial application point of view, the faster the acidification, the more efficient the production of the fermented product. Therefore, blends including Vega™ Harmony and Vega™ Classic would be considered as acidification leads in this example.

Effect of different bacterial blends on gel hardness

All samples gelled upon fermentation with all the different bacterial combinations, confirming that fermentation enables pea protein gel formation sustained by a protein network without the need of solid fats or texturizing agents, and also that all the blends produced enough lactic acid for the formation of fermentation-induced PPI gels. The proteins in the pea matrix are from a commercial protein isolate, which are mostly denatured during the purification process. Moreover, the heat treatment applied prior to fermentation would ensure further denaturation in case they would initially be only partially denatured. This protein unfolding exposes the hydrophobic domains (O’Kane, Happe, Vereijken, Gruppen and Van Boekel, 2004) that are normally hidden inside the protein structure. With a gradual acidification occurring during fermentation, the electrostatic repulsion between proteins is reduced (Mession, Chihi, Sok and Saurel, 2015), and these start interacting with each other forming a protein network. The protein gels obtained in this example presented a similar texture to that of a fresh cheese after one week of storage under refrigeration. As mentioned in the methods section, the bacterial combinations are labelled as CO in case of no adjunct cultures present in the blend, and C1 to C5 for the different adjunct culture combinations described in Table 1. No significant effect of the starter cultures on the force values was observed (Fig 2). Also, there was no interaction between the effect of the starter culture and the effect of the adjunct culture combinations (Fig 2). This would suggest that all the starter cultures would be equally suitable for the production of fermentation-induced PPI gels. Regarding the adjunct combinations, CO and C1 belong to the same homogeneous group, according to the Tukey test results. This shows that the adjunct cultures in C1 , namely L helveticus and L casei, did not improve gel hardness when supplemented together and in absence of other adjunct cultures. Samples fermented with combination C3 were firmer than those fermented with C4 and those with C5. All three combinations share L helveticus, Pediococcus, and L plantarum, but C4 and C5 also contain L casei, or L casei with L rhamnosus, respectively. For a better understanding, the effect of the single strains present in each adjunct combination was evaluated. Pediococcus did not show any significant effect on gel hardness, which is also reflected in C1 and C2 not being significantly different. Samples fermented with C2, which does not contain L plantarum, showed higher force values than C4 (same adjunct strains without L plantarum), reflecting a potential negative impact on gel firmness. L rhamnosus did not show significant effects on gel firmness. When comparing C3 vs C4, it is possible to observe that the presence of L casei weakens the gel. From these results, it is possible to conclude that the presence of L plantarum and L casei would potentially reduce the firmness of the fermentation-induced gels produced in this example. L helveticus previously showed texturizing abilities in yogurt by producing exopolysaccharides (EPS) that interact with proteins and fill the pores in a yogurt gel. However, this example focuses on cheese-like texture, and for instance, it has been reported that cheeses fermented with EPS-producing cultures presented lower hardness and consistency, among other textural parameters (Ahmed, El Soda, Hassan and Frank, 2005; Awad, Hassan and Muthukumarappan, 2005). Moreover, fresh and aged dairy cheeses containing EPS producing strains previously presented less compact protein matrices (Dabour, Kheadr, Benhamou, Fliss and LaPointe, 2006). Since all the samples with an adjunct combination contained L helveticus, it was difficult to elucidate its effect on hardness in comparison with other strains, but the statistical analysis reported a significant positive effect with a p value of 0.0011 , which could be unrelated to EPS production.

Volatile compounds after fermentation

The VOC analysis with SPME-GC-MS detected 60 volatile compounds, from which 27 were selected for their contribution to beany flavor and cheese flavor for further statistical analysis. This selection was based on internal knowledge on off-flavor removal, especially green and beany notes, and formation of dairy notes compounds such as cheesy and buttery, as well as on existing studies on volatile compounds profile of fermented and unfermented pea matrices and dairy cheese products (Fischer, Cayot and Cachon, 2022; Ben-Harb et al., 2019a; Morales, Feliu and Ferna, 2004; Youssef et al., 2020)). These selected compounds comprised aldehydes, ketones, esters, furans, and sulfur derivative compounds. According to Fisher et al., beany flavor results from a combination of different sensorial attributes, as combination of mold, earthy, green, and fresh pea (Fischer et al., 2022). The beany, green and/or grassy flavor of pea protein is mostly characterized by the aldehydes hexanal, 2- hexenal, heptanal, 2-heptenal, octanal, 2-octenal, pentanal, and 2,4-decadienal (Ben-Harb et al., 2019a; Trikusuma, Paravisini and Peterson, 2020a). Therefore, their degradation through fermentation is of great interest for plant-based cheese production. Vega™ Classic and Vega™ Harmony removed significantly greater levels of all these beany compounds, except for 2,4-decadienal, where no significant differences were observed with different blends, and these results are shown in Fig 3. Increasing levels of 2,4-decadienal upon UHT treatment of pea beverages were previously reported and related it to the oxidation of linoleic acid, the main fatty acid in pea (Trikusuma, Paravisini and Peterson, 2020b). This might have occurred also in the samples fermented in this example, and potentially due to further oxidation of the lipids in the sunflower oil. All the blends containing adjuncts were significantly more efficient at removing hexanal and octanal, two relevant volatile compounds providing green flavor in pea (Youssef et al., 2020), than the starter cultures by themselves. Moreover, larger hexanal, heptanal, and pentanal removal ratios were observed when L helveticus was present in the blends.

The blend containing Vega™ Classic + C5 was the one that degraded the highest levels of hexanal with a 73% removal ratio, whereas Vega™ Premium without adjuncts was the least efficient with 24% removal ratio. Furthermore, this latter blend degraded significantly less hexanal than any combination containing Vega™ Harmony and Vega™ Classic. In the case of heptanal, octanal, and pentanal, also Vega™ Classic + C5 was the most efficient with 76%, 78%, and 70% removal ratios, respectively. These results support the idea of combining the starter cultures with adjuncts to remove the off-flavors of legume raw materials such as pea. Blends containing Vega™ Harmony and Vega™ Classic as starter cultures, and furthermore blends containing adjunct combinations C5, C3, or C4 would be ideal to reduce the beany flavor in pea protein gels. 3-methylbutanal, an aldehyde derived from isoleucine and leucine providing malty and nutty notes (Utz, Spaccasassi, Kreissl, Stark, Tanger, Kulozik, Hofmann and Dawid, 2022; Ben-Harb, Saint-Eve, Panouille, Souchon, Bonnarme, Dugat-Bony and I rlinger, 2019b), was previously found in pea protein isolate (Utz et al., 2022). It is unclear if its presence is desired and it is potentially productdependent. It is a compound with a low taste threshold (Gadaga, Viljoen and Narvhus, 2007) therefore its presence could be relevant for flavor perception regardless of its concentration. It was detected in the unfermented matrix and it was reduced to significantly lower levels in the samples fermented with Vega™ Harmony and Vega™ Classic (94 to 97% reduction ratios, respectively) in comparison to those fermented with Vega™ Premium and Vega™ Vibe (69 to 94% reduction ratios). 3-methylbutanal was previously found in pea protein samples fermented with different lactic acid bacteria and yeasts (Ben-harb, I rlinger, Saint- eve, Panouille, Souchon and Bonnarme, 2020), and it was reported as a key flavor compound in numerous cheese varieties (Afzal, Ariceaga, Boulahya, Jacquot, Delaunay and Cailliez-Grimal, 2017). No single effect of the starter cultures or the adjunct combinations was observed in the production of acetaldehyde, 2-butenal, and 2,4-dimethyl benzaldehyde, although blends including L rhamnosus produced greater levels of the latter. Benzaldehyde can be found in semi-hard cheeses (Molimard and Spinnler, 1996) and its production was associated to phenylalanine degradation in lactic acid bacteria (Nierop Groot and de Bont, 1998). In this example, its presence was only affected by the starter cultures. Samples fermented with Vega™ Premium and Vega™ Vibe presented significantly higher levels of this aldehyde in comparison to Vega™ Harmony and Vega™ Classic. Masja et al. found that oxygen concentration, storage temperature, and pH could have an impact on the production of this aldehyde, but most important, the activity of aminotransferrases in LAB (Nierop Groot and de Bont, 1998), which could have been higher in these two starter cultures in comparison to the others.

Ketones were detected in all fermented PPI gels. These compounds are common in cheese, typically increasing during ripening, and bringing fruity-floral notes to the cheese as well as green, blue cheese (2-heptanone), and hot milk and musty (2-nonanone) aromas. Among them, 2,3-pentanedione was detected in all samples (Fig 4), with an increase of 95-99% in comparison to the unfermented base. It is a compound found in cheeses such as Mozzarella (Natrella, Faccia, Lorenzo, De Palo and Gambacorta, 2020) that can impart butter, creamy, nutty and cheese notes. It was produced in significantly higher amounts in samples fermented with Vega™ Classic, followed by those fermented with Vega™ Harmony and Vega™ Premium, and lastly those with Vega™ Vibe. The most remarkable effect of the adjunct combinations among each of the starter cultures was that samples including no adjuncts contained higher levels of 2,3-pentanedione in comparison to those with adjuncts in their blends. 2-Pentanone, in contrast, was present in greater levels in samples fermented with adjunct combination C5 in comparison to the rest of the combinations, and the presence of L rhamnosus increased those. 3-octen-2-one, 2-butanone, 2-heptanone, 2- octanone, 2-nonanone, diacetyl, and acetoin did not reflect significant differences among any of the fermented samples. However, its presence in the fermented samples is of interest since they are compounds that confer dairy and cheesy notes. Diacetyl and acetoin, two characteristic compounds in butter, buttermilk and sour cream and also found in dairy cheese (Morales et al., 2004), can be produced through glucose (present in the matrix used in this example) or through citrate metabolism. Thermophilic strains are usually citratenegative (Cogan and Hill, 1993), therefore these volatile compounds might have been produced through the glucose pathway in the case of S. thermophilus. Both compounds have been previously identified in studies where pea protein matrices have been fermented (Garcia Arteaga, Leffler, Muranyi, Eisner and Schweiggert-Weisz, 2021). Among the esters, ethyl hexanoate was the only one showing significant differences among bacterial blends. Youssef et al. previously found this compound in a fermented pea protein-based product and correlated it with a sweet, waxy and fatty odor (Youssef et al., 2020). In this example, it was detected in significantly greater levels in samples fermented by Vega™ Premium, and higher ethyl acetate S/N levels were found in samples fermented with Vega™ Vibe in comparison to those fermented with Vega™ Harmony and Vega™ Classic. 2-pentylfuran, a volatile compound previously reported in pea protein matrices (Xiang, Zhu, Jiang, Chen, Zhou and Zhong, 2023), was detected in all fermented PPI samples but no significant differences in their production caused by the starter cultures or the adjunct combinations were detected. This furan was probably a product of lipid oxidation (Ben-harb et al., 2020; Trikusuma et al., 2020a) and has been reported to be responsible for an earthy taste. The sulfur compound dimethyl-disulfide was found in high levels in Cheddar cheese (Burbank and Qian, 2005) and it is a product of the metabolization of sulfur-containing amino acids. In this example, it was present in all samples, and the presence of L plantarum in the blends decreased its levels. Conclusion

Twenty-four different bacterial blends (Table 1) with four different starter cultures were evaluated for their ability to ferment a pea protein matrix. The evaluation was based on acidification, texture, and volatile profile, including removal of off-flavor and production of dairy-like compounds. Fast acidification was prioritized for food safety reasons. Blends containing Vega™ Classic and Vega™ Harmony acidified significantly faster. The texture of the gels was not affected by the starter cultures, but the adjunct combination C3 (L helveticus, Pediococcus, and L plantarum) provided significantly higher firmness than C4 (L helveticus, Pediococcus, L plantarum and L casei) and C5 (L helveticus, Pediococcus, L plantarum, L casei and L rhamnosus). Also blends containing Vega™ Classic and Vega™ Harmony removed greater levels of beany compounds, and blends containing adjunct cultures were more efficient at the reduction of these VOCs than those containing only starter cultures. In contrast, blends without adjuncts produced more dairy-like compounds such as 2,3-pentanedione than those with adjuncts, being bacterial blends including Vega™ Classic the most efficient at its production. Other desired VOCs contributing to cheese flavor such as diacetyl and acetoin were present in all fermented samples but no significant differences in their production by the different blends was observed. The results of this example highlight the potential of fermentation to create gels with increased gel firmness and to the vastly improved aroma profile of plant-based cheeses.

The present invention has been described with reference to various embodiments, aspects, examples, or the like. It is not intended that these elements be read in isolation from one another. Thus, the present disclosure provides for the combination of two or more of the embodiments, aspects, examples, or the like.

All embodiments described herein are intended to be within the scope of the invention disclosed. These and other embodiments of the present invention will become readily apparent to those skilled in the art from the following detailed description of the preferred embodiments having reference to the whole description, the invention not being limited to any particular preferred embodiment(s) disclosed.

REFERENCES

Afzal, M.I., Ariceaga, C.C.G., Boulahya, K.A., Jacquot, M., Delaunay, S., Cailliez-Grimal, C., 2017. Biosynthesis and role of 3-methylbutanal in cheese by lactic acid bacteria: Major metabolic pathways, enzymes involved, and strategies for control. Critical Reviews in Food Science and Nutrition 57, 399-406. Ahmed, N.H., El Soda, M., Hassan, A.N., Frank, J., 2005. Improving the textural properties of an acid-coagulated (Karish) cheese using exopolysaccharide producing cultures. LWT - Food Science and Technology 38, 843-847.

Awad, S., Hassan, A.N., Muthukumarappan, K., 2005. Application of exopolysaccharide- producing cultures in reduced-fat Cheddar cheese: Texture and melting properties. Journal of Dairy Science 88, 4204-4213.

Ben-harb, S., Irlinger, F., Saint-eve, A., Panouille, M., Souchon, I., Bonnarme, P., 2020. Versatility of microbial consortia and sensory properties induced by the composition of different milk and pea protein-based gels. LWT - Food Science and Technology 118, 108720.

Ben-Harb, S., Saint-Eve, A., Panouille, M., Souchon, I., Bonnarme, P., Dugat-Bony, E., Irlinger, F., 2019a. Design of microbial consortia for the fermentation of pea-protein- enriched emulsions. International Journal of Food Microbiology 293, 124-136.

Ben-Harb, S., Saint-Eve, A., Panouille, M., Souchon, I., Bonnarme, P., Dugat-Bony, E., Irlinger, F., 2019b. Design of microbial consortia for the fermentation of pea-protein- enriched emulsions. International Journal of Food Microbiology 293, 124-136.

Burbank, H.M., Qian, M.C., 2005. Volatile sulfur compounds in Cheddar cheese determined by headspace solid-phase microextraction and gas chromatograph-pulsed flame photometric detection. Journal of Chromatography A 1066, 149-157.

Cogan, T.M., Hill, C., 1993. Cheese Starter Cultures, in: Cheese: Chemistry, Physics and Microbiology, pp. 193-255.

Dabour, N., Kheadr, E., Benhamou, N., Fliss, I., LaPointe, G., 2006. Improvement of texture and structure of reduced-fat Cheddar cheese by exopolysaccharide-producing lactococci. Journal of Dairy Science 89, 95-110.

Ferawati, F., Hefni, M., Ostbring, K., Witthoft, C., 2021. The application of pulse flours in the development of plant-based cheese analogues: Proximate composition, color, and texture properties. Foods 10.

Fischer, E., Cayot, N., Cachon, R., 2022. Potential of Microorganisms to Decrease the "Beany" Off-Flavor: A Review. Journal of Agricultural and Food Chemistry 70, 4493-4508.

Gadaga, T.H., Viljoen, B.C., Narvhus, J. A., 2007. Volatile organic compounds in naturally fermented milk and milk fermented using yeasts, lactic acid bacteria and their combinations as starter cultures. Food Technology and Biotechnology 45, 195-200. Garcia Arteaga, V., Leffler, S., Muranyi, I., Eisner, P., Schweiggert-Weisz, U., 2021. Sensory profile, functional properties and molecular weight distribution of fermented pea protein isolate. Current Research in Food Science 4, 1-10.

Grasberger, K.F., Gregersen, S.B., Jensen, H.B., Sanggaard, K.W., Corredig, M., 2021. Plant-dairy protein blends: gelation behaviour in a filled particle matrix. Food Structure 29, 100198.

Marilley, L, Casey, M.G., 2004. Flavours of cheese products: Metabolic pathways, analytical tools and identification of producing strains. International Journal of Food Microbiology 90, 139-159.

Masia, C., Jensen, P.E., Petersen, I.L., Buldo, P., 2022. Design of a Functional Pea Protein Matrix for Fermented Plant-Based Cheese. Foods 11 , 1-13.

Mession, J.L., Chihi, M.L., Sok, N., Saurel, R., 2015. Effect of globular pea proteins fractionation on their heat-induced aggregation and acid coldest gelation. Food Hydrocolloids 46, 233-243.

Molimard, P., Spinnler, H.E., 1996. Review: Compounds Involved in the Flavor of Surface Mold-Ripened Cheeses: Origins and Properties. Journal of Dairy Science 79, 169-184.

Morales, P., Feliu, I., Ferna, E., 2004. Volatile Compounds Produced in cheese by Enterobacteriaceae. Journal of Food Protection 67, 567-573.

Natrella, G., Faccia, M., Lorenzo, J.M., De Palo, P., Gambacorta, G., 2020. Short communication: Sensory characteristics and volatile organic compound profile of high- moisture mozzarella made by traditional and direct acidification technology. Journal of Dairy Science 103, 2089-2097.

Nicolas Saraco, M., Blaxland, J., 2020. Dairy-free imitation cheese: is further development required? British Food Journal 122, 3727-3740.

Nierop Groot, M.N., de Bont, J.A., 1998. Conversion of phenylalanine to benzaldehyde initiated by an aminotransferase in lactobacillus plantarum. Applied and environmental microbiology 64, 3009-13.

Nunes, M.C., Raymundo, A., Sousa, I., 2006. Gelled vegetable desserts containing pea protein, w-carrageenan and starch. European Food Research and Technology 222, 622- 628.

O’Kane, F.E., Happe, R.P., Vereijken, J.M., Gruppen, H., Van Boekel, M.A., 2004. Characterization of Pea Vicilin. 2. Consequences of Compositional Heterogeneity on Heat- Induced Gelation Behavior. Journal of Agricultural and Food Chemistry 52, 3149-3154. Schindler, S., Zelena, K., Krings, U., Bez, J., Eisner, P., Berger, R.G., 2012. Improvement of the Aroma of Pea (Pisum sativum) Protein Extracts by Lactic Acid Fermentation. Food Biotechnology 26, 58-74.

Shen, Y., Li, Y., 2021. Acylation modification and/or guar gum conjugation enhanced functional properties of pea protein isolate. Food Hydrocolloids 117, 106686.

Shi, Y., Singh, A., Kitts, D.D., Pratap-Singh, A., 2021. Lactic acid fermentation: A novel approach to eliminate unpleasant aroma in pea protein isolates. Lwt 150, 111927.

Short, E.C., Kinchia, A. J., Nolden, A.A., 2021. Plant-based cheeses: A systematic reviewof sensory evaluation studies and strategies to increase consumer acceptance. Foods 10.

Trikusuma, M., Paravisini, L, Peterson, D.G., 2020a. Identification of aroma compounds in pea protein UHT beverages. Food Chemistry 312.

Trikusuma, M., Paravisini, L, Peterson, D.G., 2020b. Identification of aroma compounds in pea protein UHT beverages.

Utz, F., Spaccasassi, A., Kreissl, J., Stark, T.D., Tanger, C., Kulozik, U., Hofmann, T., Dawid, C., 2022. Sensomics-Assisted Aroma Decoding of Pea Protein Isolates (Pisum sativum L.). Foods 11 , 1-15.

Xiang, L, Zhu, W., Jiang, B., Chen, J., Zhou, L., Zhong, F., 2023. Volatile compounds analysis and biodegradation strategy of beany flavor in pea protein. Food Chemistry 402, 134275.

Youssef, C.E., Bonnarme, P., Fraud, S., Peron, A.C., Helinck, S., Landaud, S., 2020. Sensory improvement of a pea protein-based product using microbial co-cultures of lactic acid bacteria and yeasts. Foods 9.

Claims

1. A method for producing a fermented plant-based product, the method comprising a) adding a starter culture comprising Streptococcus thermophilus and Lactobacillus bulgaricus to a plant base, b) optionally adding an adjunct culture comprising Lactobacillus helveticus to the plant base, c) fermenting the plant base for a period of time until a target pH is reached, thereby producing the fermented plant-based product; wherein the plant base is a pea base.

2. The method according to claim 1, wherein the starter culture further comprises Lactobacillus acidophilus, Lactobacillus paracasei and Bifidobacterium.

3. The method according to any of claims 1 or 2, wherein the adjunct culture further comprises Pediococcus and Lactobacillus plantarum.

4. The method according to claim 1, wherein the starter culture is Vega™ Classic or Vega™ Harmony.

5. The method according to claim 1, wherein the adjunct culture is a mix of Vega™ Boost HA, Vega™ Boost PA and Vega™ Boost LP.

6. The method according to any of the preceding claims, wherein the adjunct culture is added at the start of the fermentation period.

7. The method according to any of the preceding claims, wherein the plant base is a liquid base with plant protein content of 5-15% (w/w), such as 5, 6, 7, 8, 9, 10, 11 , 12, 13, 14 or 15 % (w/w).

8. The method of claim 7 wherein the plant protein content is 10 % (w/w).

9. A fermented plant-based product obtainable by the method according to any of the preceding claims.

10. A fermented plant-based product which has a pea protein content of 5, 6, 7, 8, 9, 10, 11 , 12, 13, 14 or 15 % (w/w) or higher and which comprises Streptococcus thermophilus and Lactobacillus bulgaricus, and optionally Lactobacillus helveticus.

11. The fermented plant-based product according to claim 10, which further comprises Lactobacillus acidophilus, Bifidobacterium lactis, and Lactobacillus paracasei.

12. The fermented plant-based product according to claim 10 or 11 , which further comprises Pediococcus and Lactobacillus plantarum.

13. The fermented plant-based product according to claim 10, wherein the starter culture is Vega™ Classic or Vega™ Harmony.

14. The fermented plant-based product according to claim 10, wherein the adjunct culture is a mix of Vega™ Boost HA, Vega™ Boost PA and Vega™ Boost LP.