EP2217080A1

EP2217080A1 - Fermented soy-based beverage

Info

Publication number: EP2217080A1
Application number: EP08851809A
Authority: EP
Inventors: Christoph Hendrik Beckmann; Jan Roelf Klooster; Michel Mellema; Jan Willem Sanders; Leendert Hendrik Wesdorp
Original assignee: Unilever PLC; Unilever NV
Current assignee: Coca Cola Co
Priority date: 2007-11-23
Filing date: 2008-11-03
Publication date: 2010-08-18
Also published as: CN101932250A; MX2010005304A; BRPI0819037A2; CN101932250B; ZA201002995B; WO2009065724A1; AR069383A1

Abstract

The present invention relates to a method of producing a beverage by: - providing a pasteurised or sterilised aqueous liquid containing 0.5-8 wt. % of dissolved soy protein and 0-0.2 wt.% of dairy protein; - inoculating the pasteurised or sterilised liquid with a thermophilic lactic acid bacterium containing starter culture; - fermenting the inoculated aqueous liquid by incubation at a temperature in the range of 40-48°C for 0.5-24 hours to obtain a fermented product having a viscosity at a temperature of 7°C of less than 50 mPa.s at 100 s-1; wherein in total less than 6% disaccharides by weight of the fermented product are added before, during or after fermentation and wherein during fermentation the following changes in concentrations of flavour compounds occur: • lactate concentration increases with at least 500 ppm; • diacetyl concentration increases with at least 0.3 ppm and/or acetaldehyde concentration increases with at least 0.05 ppm; • concentration of n-hexanal decreases by at least 60%; • at least two of n-pentanal concentration, n-heptanal concentration, n-octanal concentration and n-nonanal concentration decrease by at least 50%. The present method enables the effective removal of soy off-notes as well as the preparation of fermented substrates with a desirable flavour profile that is similar to that of fermented dairy products.

Description

FERMENTED SOY-BASED BEVERAGE

Field of the Invention

The present invention relates to the field of fermented (or cultured) soy-containing products, especially fermented soy- based beverages.

Background of the invention

Consumers have become more knowledgeable about protein and its role in a healthy diet. This new understanding has had a profound effect, stimulating consumer interest and demand for healthier beverages that are fortified with protein. Because beverages are a convenient way to incorporate protein into the diet, manufacturers continue to formulate new products in an effort to make protein more accessible to a wider group of consumers . The two most popular beverage proteins are milk protein and soy protein, and their various isolate derivatives. According to the U.S. Food and Drug Administration, the consumption of food products rich in soy protein can reduce cholesterol, enhance athletic performance, and even aid in the battle against diabetes. In addition, interest for milk replacement by soy protein has increased in view of, on the one hand, issues in relation to over-sensitivity and/or intolerance towards milk constituents experienced by growing numbers of consumers and, on the other hand, elevated milk protein prices and supply issues that some manufacturers are experiencing relative to this commodity. Soy proteins have been proposed to replace milk proteins, either partially or totally, depending on the system, and dairy-like products have been developed based entirely on soy protein. In view of soy protein' s documented health benefits it is desirable to incorporate substantial quantities of soy protein in beverages. However, incorporation of soy protein into e.g. beverages presents several challenges. The incorporation of soy protein in beverages is known to impart a noticeable aftertaste and a distinctive "beany" taste. Different types of processes to isolate soy protein have been proposed which aim to remove these undesirable (off-) flavour-notes. Unfortunately, however, it is almost impossible to completely remove the typical soy "off-notes" from sources of soy protein such as soy concentrates and soy isolates. In addition, in most product applications the intensity of soy off-notes increases during processing and storage, probably as a result of the formation of off-flavour compounds from precursor molecules. US 3,364,034 describes a method for removing characteristic flavour and/or odour from vegetable protein materials to provide a substantially bland product, comprising: inoculating said protein materials with bacteria selected from the group consisting of Lactobacillus lactis, Lactobacillus bulgaricus, Lactobacillus acidophilus, Leuconostoc citrovorum, Pediococcus cerivisiae, Pseudomonas ovalis, Pseudomonae fragi, Aerobacter aerogenes, Streptococcus lactis, incubating for 16-144 hours under conditions conducive to bacterial growth; and terminating said bacterial growth after said material is rendered substantially bland.

US 3,937,843 describes a method for elimination of a bean odour from a soy milk which comprises subjecting the soy milk containing added saccharide to lactic fermentation and then distilling the resulting fermented soy milk at a temperature of 10-85 ⁰C, under a pressure of not more than 500 mmHg.

US 4,664,919 describes a process for producing yogurt-like food, comprising fermenting soy milk with Streptococcus sojalactis . It is observed in the US patent that the yogurt- like food so obtained does not have peculiar 'green' smell of soy milk and that it has a good taste. It is further stated that the aforementioned Streptococcus strain is capable of removing the 'green' smell of soybeans and that the amount of diacetyl and acetone formed is larger than that of other lactic acid bacteria. Data provided about the Streptococcus sojalactis show that the micro-organism is capable of growing at temperatures in the range of 30-40 ⁰C, but not at 20 ⁰C or less or at temperatures of 45 ⁰C or higher. US 6,599,543 relates to process for preparing a fermented soybean milk comprising: contacting dehulled and dehypocotyl whole soybeans with warm or hot water; removing warm or hot water-soluble component from the soybeans; pulverizing the soybeans to make a slurry; removing insoluble component from the slurry to make a soybean milk, inoculating a lactic acid bacterium of the genus Bifidobacterium, Lactobacillus bulgaricus and one strain selected from the group consisting of Lactobacillus acidophilus and Lactobacillus easel into the soybean milk, adding one or more saccharides which can be utilized by the lactic acid bacterium to the soybean milk, and fermenting the soybean milk to produce the fermented soybean milk. The examples of the US patent suggest that it is essential to employ a Bifidobacterium strain to produce a fermented soybean milk with acceptable taste. Furthermore, the removal of hypocotyl from the soybeans as taught by US 6,599,543 is laborious and costly.

JP-A-2004-261003 describes a method of preparing fermented soybean milk by fermenting soybean milk and by adding palatinose (isomaltulose) before, during or after fermentation. In the Japanese application it is observed that the grass smell inherent to soybean milk can be reduced to a certain extent by fermenting soybean milk with a combination of a lactic acid bacterium and a bifidobacterium, but that satisfactory results cannot always be established, in particular due to the taste and smell of acetic acid, which is metabolic product of the fermentation. Palatinose is incorporated in the fermented soybean milk to suppress the unpleasant taste, fermentation smell, harshness and acetic acid smell generated in the course of lactic acid fermentation or acetic acid fermentation. The examples of the Japanese patent application describe the preparation of fermented drink yogurts from a soybean milk to which isomaltulose and sucrose are added prior and/or after fermentation. In comparative examples the preparation of fermented drink yogurts is described in which large amounts of sucrose are added before and/or after fermentation. In the aforementioned examples a starter culture is used that comprises Lactobacillus delbrueckii subsp. Bulgaricus and Streptoccus thermophilus . The drink yogurts disclosed in the Japanese application are extremely sweet as they contain more than 8 wt . % of added disaccharide (isomaltulose and/or sucrose) . The use of large amounts of sweetener helps to mask soy off-flavour notes, but is often deemed undesirable. Murti TW et al . , Journal of Food Science (1993), 58(1), pages 153-157 describe experiments in which set yogurts were prepared by inoculating soymilk and cow milk with Streptococci, Lactobacilli in the absence of presence of Bifidobacteria and wherein the concentration of a number of volatile compounds was monitored during fermentation. The results show that during fermentation of soy milk at 42 ⁰C the concentration of n- hexanal decreases and the concentrations of lactic acid and diacetyl increases. The yoghurt-type products described by Murti et al . develop a viscosity of 100 mPa.s or more after 4 hours of fermentation. Summary of the Invention

The inventors have developed a method of manufacturing a beverage by fermenting a soy protein containing substrate that effectively removes off-flavour notes originating from the soy component and additionally delivers a very pleasant taste, without using high amounts of sweetener (notably sucrose and/or isomaltulose) . The method of the present invention utilises a culture containing a thermophilic lactic acid bacterium (LAB) strain, which LAB strain is highly effective in metabolising n- hexanal and other n-aldehydes and that additionally is capable of producing lactic acid and at least one of diacetyl and acetaldehyde . The substrate employed in the present method is a pasteurised or sterilised aqueous liquid containing 0.5-8 wt . % of dissolved soy protein and 0-0.2 wt . % of dairy protein. In the present method, after inoculation, the substrate is incubated at a temperature of 40-48 ⁰C for 0.5-24 hours to yield a non-viscous fermented product with excellent taste. Compared to set soy-based yogurts, which are substantially more viscous than the present soy-based beverages, the beverages obtained by the present method exhibit a significantly more intense dairy flavour. Thus, the beverages obtained by the present method offer the advantage that they have a pronounced dairy flavour and that undesirable soy flavour notes are virtually absent.

C5-C9 n-alkanals (including n-hexanal) and trans-2-hexenal are flavour molecules that are believed to be largely responsible for the ^λbeany' off-flavour often found in soy based products. The flavour molecules diacetyl and acetaldehyde are often associated with fermented dairy products such as yogurt, butter and butter milk. The inventors have discovered that by fermenting the present soy protein containing substrate with a suitable thermophilic LAB strain, it is possible to remove typical soy-related off-notes and at the same time to introduce pleasant flavour notes that enhance the quality of the final beverage in which the fermented substrate is applied. In order to obtain a fermented substrate with significantly reduced off- flavour notes and a pleasant flavour, the fermentation conditions must be controlled so that: i. the concentration of n-hexanal decreases by at least 60%; ii. the concentration of lactate (lactic acid and salts thereof) increases with at least 500 ppm; and iii. the concentration of diacetyl increases with at least 0.3 ppm and/or the concentration of acetaldehyde increases with at least 0.05 ppm. As mentioned before, the method of the present invention offers the advantage that it enables the effective removal of soy off- notes as well as the preparation of fermented substrates with a desirable flavour profile that is similar to that of fermented dairy products, such as yogurt and fermented dairy drinks. Unlike the process taught by US 6,599,543, the present method does not require that the soy protein in the substrate is derived from soybeans from which the hypocotyl has been removed.

Detailed Description of the Invention

Accordingly, the present invention relates to a method of producing a beverage by fermenting a soy protein containing substrate, said method comprising:

- providing a pasteurised or sterilised aqueous liquid containing 0.5-8 wt . % of dissolved soy protein and 0-0.2 wt .% of dairy protein, said soy protein being derived from soybeans that have not been dehypocotylized; - inoculating the pasteurised or sterilised liquid with a thermophilic lactic acid bacterium containing starter culture, said thermophilic lactic acid bacterium having an optimum growth temperature in excess of 35 ⁰C, preferably in excess of 38 ⁰C;

- fermenting the inoculated aqueous liquid by incubation at a temperature in the range of 40-48 ⁰C for 0.5-24 hours to obtain a fermented product having a viscosity at a temperature of 7 ⁰C of less than 50 mPa.s at 100 s^'1; wherein in total less than 6% disaccharides by weight of the fermented product are added before, during or after fermentation and wherein during fermentation the following changes in concentrations of flavour compounds occur:

• lactate concentration increases with at least 500 ppm; • diacetyl concentration increases with at least 0.3 ppm and/or acetaldehyde concentration increases with at least 0.05 ppm;

• concentration of n-hexanal decreases by at least 60%, preferably by at least 70%; • at least two of n-pentanal concentration, n-heptanal concentration, n-octanal concentration and n-nonanal concentration decrease by at least 50%.

The term "dehypocotylized soybeans" as used herein refers to soybeans from which the hypocotyl has been removed. Hypocotyl is a botanical term for a part of a germinating seedling of a seed plant. As the plant embryo grows at germination, it sends out a shoot called a radicle that becomes the primary root and penetrates down into the soil. After emergence of the radicle, the hypocotyl emerges and lifts the growing tip above the ground, bearing the embryonic leaves (called cotyledons) and the plumule that gives rise to the first true leaves. The hypocotyl is the primary organ of extension of the young plant and develops into the stem.

The term "lactic acid bacterium" as used herein refers to acid tolerant, non-sporulating, non-respiring rod-shaped Gram positive bacilli or cocci that produce lactic acid as the major metabolic endproduct of carbohydrate fermentation. As used herein, the term "lactic acid bacterium" does not encompass Bifidobacterium . The term "lactate" as used herein encompasses lactic acid as well as edible salts of lactic acid.

A decrease in concentration of a particular substance by X% means that if the starting concentration was Y ppm, the decreased concentration equals Y(IOO-X) /100 ppm. The lactate concentration in the fermented product is suitably determined by the methodology described in the examples.

The diacetyl, acetaldehyde, alkanal and alkenal concentrations referred to in this document are determined by the analytical methods described in the examples. As explained herein before, in order to produce a fermented product without soy off-flavour notes and a pleasant, for instance, yogurt-like or buttery flavour, the fermentation conditions in the present method are controlled so as to promote metabolisation of n-hexanal and simultaneous production of considerable quantities lactate, and diacetyl and/or acetaldehyde. Using the aforementioned analytical methods it is well within the skill of a person skilled in the art of food fermentation to select suitable thermophilic LAB strains and to optimise the process conditions in such a way that the targeted profile (substantial reduction of hexanal and increase in lactate, and diacetyl and/or acetaldehyde) is realised. In the present method, during fermentation, diacetyl concentration advantageously increases with at least 0.5 ppm, most preferably by at least 1 ppm. Likewise, acetaldehyde concentration preferably increases with at least 0.1 ppm, most preferably by at least 0.2 ppm.

The amount of lactate produced during fermentation can easily exceed 1 gram per litre. Hence, in a preferred embodiment, during fermentation the lactate concentration increases by at least 1.0 g/kg (1000 ppm). Even more preferably lactate concentration increases by at least 2.0 g/kg and most preferably by at least 3.0 g/kg. In the present method the microbial production of considerable quantities of lactate advantageously causes a substantial pH decrease. Typically, during fermentation the pH of the aqueous liquid decreases by at least 1.5 pH units, even more preferably by at least 2.0 pH units. Typically pH does not decrease by more than 3.5 pH units. The pH of the fermented product obtained in the present method usually does not exceed 6.0, more preferably it does not exceed 5.5. During fermentation, the pH of the aqueous liquid typically does not decrease below 3.8, more preferably it does not decrease below 4.0. The present invention also encompasses an embodiment in which, for instance, lactate is added to the fermented product in order to further decrease pH. However, according to a particularly preferred embodiment of the present method, no lactate, diacetyl or acetaldehyde is added before, during or after fermentation. In other words, in accordance with this preferred embodiment all of the lactate, diacetyl and acetaldehyde present in the fermented product is produced during fermentation or was naturally present in one of or more of the ingredients present in the pasteurised or sterilised aqueous liquid that is used as substrate in the present fermentation .

The fermented product obtained in the present method typically contains at least 0.5 ppm, more preferably at least 1 ppm of diacetyl. The diacetyl concentration in the fermented product usually does not exceed 40 ppm.

The amount of acetaldehyde in the fermented product advantageously is at least 0.1 ppm, most preferably at least 0.2 ppm. The acetaldehyde content of the fermented product typically does not exceed 10 ppm.

The inventors have found that it is feasible to achieve very significant reductions in n-aldehyde levels during fermentation. According to a preferred embodiment, during fermentation at least one, more preferably at least two and most preferably at least three of n-pentanal concentration, n- heptanal concentration, n-octanal concentration and n-nonanal concentration decrease by at least 50%, more preferably by at least 60% and most preferably by at least 70%. According to one particularly preferred embodiment, during fermentation the n-hexanal concentration decreases by at least 75% and most preferably by at least 80%. According to another preferred embodiment, during fermentation trans-2-hexenal concentration decreases by at least 50%, preferably by at least 60%, most preferably by at least 70%.

The inventors have found that the present method can advantageously be employed to substantially reduce the level of 2-methylbutanal and 3-methylbutanal . Sources of soy protein usually contain 2-methylbutanal and 3-methylbutanal in concentrations that are well above the so called flavour threshold level. In many soy products, the typical "cereal" and/or "malty" flavour contribution of 2-methylbutanal and 3- methylbutanal is highly undesirable. Consequently, in accordance with a particularly preferred embodiment of the present method, during fermentation concentration of 2- methylbutanal and/or 3-methylbutanal decreases by at least 25%, more preferably by at least 40% and most preferably by at least 60%. The optimum growth temperature of the thermophilic LAB strain employed in the present method preferably lies in the range of 40-50 ⁰C, most preferably in the range of 41-48 ⁰C. The present method may suitably employ a variety of thermophilic LAB strains, provided these strains are capable of producing lactate as well as diacetyl and/or acetaldehyde . Furthermore, these strains must exhibit the capability to metabolise one or more of the undesired aldehydes described herein before. Examples of thermophilic LAB strains that may advantageously be employed in the present method include Streptococcus thermophilics, Lactobacillus delbrueckii (including its subspecies Lactobacillus delbrueckii subsp. bulgaricus and Lactobacillus delbrueckii subsp. lactis) , Lactobacillus helveticus, Lactobacillus acidophilus, Lactobacillus casei, Lactobacillus reuteri, Lactobacillus rhamnosus and combinations thereof. Most preferably, the thermophilic LAB strain employed in the present method is selected from the group consisting of: Streptococcus thermophilus, Lactobacillus delbrueckii, Lactobacillus helveticus, Lactobacillus acidophilus and combinations thereof. The fermented product obtained by the present method preferably has a water content of at least 85 wt%, most preferably 87 wt%, based on the total weight of the fermented product. According to another preferred embodiment, the soy protein content of the soy-based fermented beverage of the invention is within the range of 1-7 wt%, more preferably in the range of 2- 6 wt .% and most preferably of 2.5-5.5 wt%, based on the total weight of the fermented soy-based beverage. "Soy protein content", as used herein, refers to the total amount of soy protein and soy protein derived peptides contained in the fermented beverage. The fermented beverage may be based on soy protein, on soy protein hydrolysate or combinations thereof. As will be understood by the skilled person (enzymatic) hydrolysis of peptide bonds may occur during fermentation .

The substrate that is fermented in the present method, i.e. the aqueous liquid containing 0.5-8 wt . % of dissolved soy protein is preferably prepared from a soy protein source selected from the group consisting of soy isolate, soy concentrate, soy flour and combinations thereof. According to a particularly preferred embodiment, the latter soy protein sources were obtained from soy beans exhibiting very low lipoxygenase activity, e.g. a lipoxygenase activity of less than 15 kU/mg. Even more preferably the soy protein source is derived from soy beans exhibiting a lipoxygenase activity of less than 10 kU/mg, most preferably of less than 8 kU/mg. Likewise, it is advantageous to prepare the present aqueous liquid from a soy protein source having a lipoxygenase activity of less than 5 kU per gram of soy protein. Most preferably the soy protein source has a lipoxygenase activity of less than 1 kU per gram of soy protein . Lipoxygenase activity is suitably determined spectrophotometrically by using a dye-coupling assay specific for hydroperoxides generated from linoleic acid, as fully described by Anthon and Barrett (J. Agric. Food Chem. 2001, 49, 32-37) . Typically, the soy protein employed in the preparation of the aqueous liquid containing 0.5-8 wt . % of dissolved soy protein has protein content of at least 30 wt . % and fat content of less than 40 wt . % .

The present method employs soy protein derived from soybeans that have not been dehypocotylized. Accordingly, the aforementioned soy isolate, soy concentrate and/or soy flour are advantageously derived from optionally dehulled soybeans that still comprise the hypocotyl . Most preferably the latter sources of soy protein sources are derived from dehulled soybeans that still comprise the hypocotyls.

Because the present method utilises a thermophilic LAB strain, it is advantageous to carry out the fermentation at elevated temperatures of at least 40 ⁰C. The fermentation temperature advantageously does not exceed 46 ⁰C, most preferably it does not exceed 45 ⁰C.

The duration of the fermentation step in the present method advantageously does not exceed 12 hours. Even more preferably fermentation is discontinued after not more than 10 hours, most preferably after not more than 8 hours. As shown in the examples, it is possible to produce a fermented product with substantially reduced off-flavour and a substantial level of diacetyl and/or acetaldehyde using a fermentation time of only 4 hours .

The beverage obtained by the present method has a rheology that is quite different from that of e.g. set yogurt. The fermented product obtained in the present method has a relatively low viscosity at a temperature of 7 ⁰C of less than 50 mPa.s at 100 s^'1. Preferably the fermented product has a viscosity at 7 ⁰C of less than 25 mPa.s at 100 s^"1. A viscosity of 50 mPa.s at 100 s^~ ¹ means that the product is 50 times more viscous than water and about 25 times more viscous than milk. Viscosity is suitably measured with the help of a rheometer ARlOOO (TA Instruments, Etten-Leur, the Netherlands), using a 40-mm diameter, 2% angle cone measuring system. A steady-state shear process should be used, increasing shear rate from 0.01 to 250/s. The measuring temperature is 7 ⁰C and only the data point at 100 s^'1 is used. In order to ensure that fermentation times can be reduced to e.g. less than 12 hours, or even less than 8 hours, it is preferred to inoculate the pasteurised or sterilised aqueous liquid with at least 10⁵, preferably at least 10⁶ and most preferably at least 10⁷ viable cells of the thermophilic LAB per ml. Typically, the amount of viable cells of the thermophilic LAB employed for inoculation does not exceed 10⁸ viable cells per ml. The fermented product obtained at the end of fermentation typically contains at least 10⁵, preferably at least 10⁶ and most preferably at least 10⁷ viable cells per ml of the thermophilic LAB employed for inoculation. In one embodiment of the present invention the pasteurised or sterilised aqueous liquid is fermented in a fermentation vessel, following which the fermented product is filled into containers that are subsequently sealed. In accordance with a particularly preferred embodiment, the fermented product is pasteurised or sterilised before filling or, alternatively, it is sterilised once it is filled into a container.

According to another preferred embodiment, edible acid is added to the fermented product before it is filled into containers that are subsequently sealed, and said fermented product is not pasteurised or sterilised. The inventors have unexpectedly found that post-acidification of the fermented product yields a microbiologically stable product that is substantially better tasting than a pasteurised or sterilized product. More in particular, the inventors have found that post-acidification yields a beverage product having much stronger dairy flavour notes (e.g. yoghurt and creamy notes) .

Accordingly, in accordance with a particularly preferred embodiment the present method comprises filling the fermented product into containers and subsequently sealing the filled containers, wherein edible acid is added to the fermented product prior to the filling into the containers so as to adjust the pH to less than 4.5, and wherein the fermented product is not subjected to pasteurisation or sterilisation prior, during or after the filling into the containers. The inventors have also observed that the aforementioned post- acidification produces a better tasting product than in case the same end-pH is achieved by prolonged fermentation. In other words, the inventors have found that it is advantageous to start post-acidification before the fermentation has proceeded to the stage at which product inhibition prevents further pH reduction by microbial lactic acid production. Thus, in another preferred embodiment, the edible acid is added to the fermented product before the fermentation has reached the point at which product inhibition prevents further production of lactic acid by lactic acid bacteria

Preferably, the addition of edible acid to the fermented product causes a pH decrease of at least 0.3 units, preferably of at least 0.5 units. It is preferred to add the edible acid prior to the filling of the fermented product into the containers .

Examples of edible acids that may suitably be employed to adjust the pH of the fermented product include lactic acid, citric acid, malic acid, tannic acid, tartaric acid, phosphoric acid, acetic acid, maleic acid, succinic acid, gluconic acid, adipic acid, ascorbic acid and fumaric acid. Preferably, the edible acid is selected from the group of lactic acid, citric acid and combinations thereof. As explained herein before, the present method enables the preparation of a pleasant tasting beverage without using high amounts of sweetener to mask soy off-flavour notes. According to a particularly preferred embodiment, the fermented product in the sealed container contains less than 5 wt . % of disaccharides, most preferably less than 4 wt . % of disaccharides. According to yet another preferred embodiment, the packaged fermented product contains less than 8 wt . % of mono- and/or disaccharides, most preferably less than 5 wt . % of mono- and/or disaccharides. According to another embodiment, the inoculated pasteurised or sterilised liquid is filled into a container which is subsequently sealed and the fermentation actually occurs within the container. The fermented product that is filled into the containers preferably is a liquid containing not more than 6 wt . % of protein, most preferably containing 1.0-5.0 wt . % of protein. The pasteurised or sterilised aqueous liquid that is used as a substrate in the present method advantageously contains an added carbohydrate that can be metabolised by the thermophilic LAB strain. Examples of suitable carbohydrates include: glucose, fructose, galactose, sucrose, raffinose, stachyose, lactose and combinations thereof. Typically, these fermentable carbohydrates are added in an amount of 0.2-100 g/1, most preferably of 0.5-30 g/1.

Besides a considerable amount of soy protein, the pasteurised or sterilised aqueous liquid of the present invention may contain a small quantity of milk protein, such as whey protein or casein. Preferably, the pasteurised or sterilised aqueous liquid contains no milk protein.

In the present method a variety of food ingredients and/or additives may be incorporated in the fermented product by introducing these components to the aqueous liquid prior to or after pasteurisation or sterilisation, during fermentation or after fermentation. Examples of food ingredients and additives that may suitably be incorporated include: fruit pieces; fruit preparations; sweeteners, including artificial sweeteners; oil; flavours, colourings, vitamins, minerals and fibres. As explained herein before, JP-A-2004-261003 describes the preparation of a fermented soy-based product wherein palatinose (isomaltulose) is added prior, during or after fermentation. In the present method the pasteurised or sterilised aqueous liquid preferably does not contain isomaltulose and no isomaltulose is added prior to, during or after fermentation.

The invention is further illustrated by means of the following non-limiting examples:

EXAMPLES

Analytical methods

Analysis of off-flavour volatiles by SPME followed by GC-MS: 2 g of sample were put in a 20 ml headspace vial and sealed with an airtight cap. Samples were analyzed by means of solid phase micro extraction. Fiber used: Carboxen/PDMS 85 μm ex. Supelco.

Analyses were carried out on an Agilent GC/MS, equipped with a Gerstel CIS-4 injector and a Gerstel MPS-2 autosampler with SPME option. Column: VF-5; 50m * 0.2 mm * 0.33 μm

GC program:

• 40°C(2 min) - (3° /min) ->160 °C (0 min) - (20 ° /min) ->250⁰C (2 min)

• Gas: Helium

• Flow: 1 ml/min, constant flow

SPME Sampling time: 35 min at 40⁰C Desorption: 40 minutes at 170⁰C Split-less time: 2 min.

Quantitative analysis of diacetyl and acetaldehyde : 2 g of sample were put in a 20 ml headspace vial and sealed with an airtight cap. All samples were analyzed in duplicate. External calibration levels were constructed by adding acetaldehyde and diacetyl to 2 g of soy base prepared according to example 1, at levels of 0, 1, 2, 4 and 10 μg/g. Calibration lines were constructed by plotting the peak areas of ion 44 (for acetaldehyde) resp. 86 (for diacetyl) against the amount added. These calibration lines were then used to determine the amount of acetaldehyde and diacetyl in the fermented soy samples.

Samples were analyzed by means of solid phase micro extraction. Fiber used: Carboxen/PDMS 85 μm ex. Supelco.

GC program:

• -20⁰C(IO min) - (1.5° /min) ->40⁰C (0 min) - (20 ° /min) ->250⁰C (2 min)

• Gas : Helium • Flow: 1 ml/min, constant flow

Lipoxygenase activity:

Lipoxygenase activity was determined spectrophotometrically by using a dye-coupling assay specific for hydroperoxides generated from linoleic acid, as fully described by Anthon and Barrett (J. Agric. Food Chem. 2001, 49, 32-37) .

Enzyme extracts were obtained by homogenising 100 mg of defatted soybean grindings in 20 ml of 100 mM pH 6.0 Na₂HPO₄ buffer + 1% w/v NaCl then centrifuging at 15,000 rpm at 4°C for 30 minutes and filtering the supernatant through a 0.2 μm filter.

To 500 μl of a solution of 10 mM 3- (dimethylamino) benzoic acid in 100 mM pH 6.0 Na₂HPO₄ was added 20 μl of 27 mM linoleic acid dispersed in 1.4% w/v Tween 20, and 10 μl of enzyme extract. After 5 minutes, 500 μl of a solution containing 0.2 mM 3- methyl-2-benzothiazolinone hydrazone and 0.1 mg/ml bovine haemoglobin was added. After a further 5 minutes, 500 μl of 1% w/v sodium lauryl sulphate was added to terminate the reaction. The absorbance at 598 nm was then measured. The above actions can be carried out in a single 4.5 ml 1 cm path cuvette.

A standard curve was produced by reacting dilutions of soybean lipoxygenase from Sigma-Aldrich, of certificated activity. This was used to calculate the activities of the extracts from the soybeans. Blank readings were subtracted, obtained using denatured enzyme extracts heated at 95°C for 30 minutes. One unit of activity is the increase of 0.001 absorbance units at 598 nm per minute from the hydroperoxidation of linoleic acid.

Lactate concentrations :

Lactate concentrations were determined using a commercially available reflectometrical test (Lactic Acid Test, Merck KGaA, 64271 Darmstadt, Germany) . Samples were analysed in duplicate and mean values were taken.

Viable counts:

The number of live bacteria in the cultures was determined by plate counting appropriate dilutions of samples containing T- 071016, using M17 agar and aerobic incubation for 3 days at

30⁰C. L .bulgaricus was enumerated using MRS agar and anaerobic incubation for 3 days at 37⁰C. Numbers are expressed as Colony forming units per ml of product (Cfu/ml) .

Example 1

Nine commercially available thermophilic LAB cultures were screened for their ability to metabolise n-hexanal during fermentation of a fresh soy milk product. The LAB cultures tested contained the following LAB species:

1. Mixed culture of Streptococcus thermophilus strains

2. L . Helveticus

3. Mixed culture containing strains of Lactobacillus acidophilus, Bifidobacterium animalis and Streptococcus thermophilus

4. L . bulgaricus

5. Mixed culture containing strains of Lactobacillus delbrueckii and Streptococcus thermophilus

6. Mixed culture containing strains of Lactobacillus helveticus and Streptococcus thermophilus

7. L . Rhamnosus

8. L. acidophilus

9. L. Plantarum The analyses showed that two cultures, i.e. cultures 2 and 4 achieved a significant reduction of the hexanal concentration during fermentation. Further experiments were conducted with the culture that yielded the strongest reduction in hexanal concentration, i.e. culture number 4. The results of these experiments are described in Example 4.

Example 2

A soybase was prepared by mixing 5.6% soy bean powder (Soy Supreme Fibre Reduced soy bean powder, provided by SunOpta

Grains and Foods group, Hope, MN, USA) in hot water (75-85°C) . The powder had a lipoxygenase activity of less than 5 kU per milligram of soy powder and was hydrated for at least 15 minutes before addition of 2% sucrose, 1% glucose and 0.35% HM pectin. The mixture was then pasteurised for 20sec at 72°C and homogenized at 150 bar. The soybase was kept at 5 ⁰C until further use.

500 ml of the soybase was inoculated with 0.02 % of a commercially available frozen yoghurt culture concentrate T- 071016 (defined mixed culture of Streptococcus thermophilus strains, provided by Chr Hansen, Hørsholm, Denmark) at 43 ⁰C. The mixture was incubated for a period of 4 hrs at the same temperature during which the pH decreased from 6.7 to 5.3. The fermentation process was stopped by rapidly cooling the mixture to 4 ⁰C.

During the fermentation process, viable counts increased from 4.4xlO⁷ to 9.2xlO⁷ and lactic acid levels increased to 1.77 g/1. An aliquot of the fermented sample and of the original soybase were subjected to SPME followed by GC-MS analysis. It was found that during fermentation the peak area for pentanal, hexanal, heptanal, octanal, nonanal, trans-2-hexenal, 2-methyl butanal and 3-methyl butanal had decreased considerably as shown in Table 1: Table 1

Decrease in peak area (%) n-pentanal 71 n-hexanal 92 n-heptanal 90 n-octanal 58 n-nonanal 59 trans-2- 37 hexenal

2-methyl 25 butanal

3-methyl 95 butanal

The analyses further showed that during fermentation the concentrations of acetaldehyde and diacetyl had increased by 1 ppm and 6 ppm respectively, whereas these compounds were virtually absent in the starting material.

In line with the analytical data, tasting of the fermented sample by an expert panel indicated a significant reduction not only of "green" notes but also of "cereal" and other soy related off-notes.

Example 3

500 ml of soybase prepared according to example 2 was inoculated with 0.02 % of a commercially available frozen yoghurt culture concentrate T-071016 (defined mixed culture of Streptococcus thermophilics strains, provided by Chr Hansen, Hørsholm, Denmark) at 43 ⁰C. The mixture was incubated for a period of 20 hrs at the same temperature during which the pH decreased from 6.7 to 4.2. The fermentation process was stopped by rapidly cooling the mixture to 4 ⁰C. Viable counts increased from 4.4xlO⁷ to 1.2xlO⁸ Cfu/ml over 20 hrs. During that same period the lactic acid level increased to 4.2 g/1. An aliquot of the fermented sample and of the original soybase were subjected to SPME followed by GC-MS analysis. It was found that during fermentation the peak area for pentanal, hexanal, heptanal, octanal, nonanal, trans-2-hexenal, 2-methyl butanal and 3-methyl butanal had decreased considerably as shown in Table 2:

Table 2

Decrease in peak area (%) n-pentanal 49 n-hexanal 90 n-heptanal 89 n-octanal 88 n-nonanal 77 trans-2- 40 hexenal

2-methyl 78 butanal

3-methyl 93 butanal

The analyses further showed that during fermentation the concentrations of acetaldehyde and diacetyl had increased by 1 ppm and 6 ppm respectively, whereas these compounds were virtually absent in the starting material. In line with the analytical data, tasting of the fermented sample by an expert panel indicated a significant reduction not only of "green" notes but also of "cereal" and other soy related off-notes.

Example 4 500 ml soy base (prepared according to example 2) was inoculated by adding 2% of a washed overnight preculture in MRS broth of L.bulgaricus Lb291 (Wiesby/Danisco, Niebull, Germany). The mixture was incubated for 20 hrs at 43 ⁰C. After 7 hrs of fermentation, a part of the fermentate was removed and the fermentation process was stopped by rapid cooling to 4 ⁰C.

After 20 hrs of fermentation the remainder of the fermentate was treated in the same way to stop the fermentation process.

In a sample obtained from the fermentate after 7 hrs of fermentation the pH had decreased from 6.7 to 5.6 and viable counts increased from 1.15xlO⁷ to 4.2xlO⁷ Cfu/ml. The lactic acid level in this sample was found to have increased to 0.7 g/1 during fermentation.

In a sample obtained after 20 hrs of fermentation the pH had decreased from 6.7 to 4.0 and viable counts increased from

1.15xlO⁷ to 1.5xlO⁸ Cfu/ml. The lactic acid level was found to have increased to 5.1 g/1 during fermentation.

Aliquots of the fermented samples and of the original soybase were subjected to SPME followed by GC-MS analysis. The effect of fermentation on the concentrations of pentanal, hexanal, heptanal, octanal, nonanal, trans-2-hexenal, 2-methyl butanal and 3-methyl butanal are depicted in Table 3:

Table 3

Decrease in

After 7 hrs After 20 hrs n-pentanal 71 8 n-hexanal 88 52 n-heptanal 85 39 n-octanal >90 >90 n-nonanal 81 68 trans-2- 78 13 hexenal

2-methyl 90 94 butanal

3-methyl 95 65 bbuuttaannaall

The analyses further showed that after 7 hrs of fermentation the concentration of acetaldehyde had increased by 0.1 ppm and that no significant production of diacetyl had occurred. In line with the analytical data, testing of the sample obtained after 7 hrs of fermentation by an expert panel indicated a significant reduction not only of "green" notes but also of "cereal" and other soy related off-notes. The analyses of the 20 hrs sample showed that during fermentation the concentrations of acetaldehyde and diacetyl had increased by 2 ppm and 0.4 ppm respectively.

From the data in Tables 3 it is apparent that concentrations of some aldehydes, notably n-pentanal, n-heptanal, n-octanal and trans-2-hexenal, increase again if fermentation is continued for up to 20 hours. It is believed that this is due to the fact that at some stage the rate at which these aldehydes are formed by lipid oxidation exceeds the rate with which these aldehydes are metabolised by the lactic acid bacteria. Clearly, this shows that it may be preferable to use relatively short fermentation times (e.g. 8-10 hours) in case the substrate is easily oxidized under fermentation conditions.

Example 5 A soybase was prepared by mixing 7.8% soy bean powder (Soy

Supreme Fibre Reduced soy bean powder SunOpta Grains and Food Group, Hope, MN, USA) in water (room temperature) . The powder had a lipoxygenase activity of less than 5 kU per milligram of soy powder and was hydrated for at least 15 minutes. The mix was then pasteurised for 20 sec. at 72°C and homogenized at 180 bar. The soybase was kept at 5 ⁰C until further use. Part of this soy base was inoculated with a commercially available yoghurt culture concentrate Yo-Mix 305 (defined mixed culture of Streptococcus thermophilics and Lactobacillus delbrueckii subsp bulgaricus strains, provided by Danisco, Niebull, Germany) according to instructions of the manufacturer. The same batch of soy base was simultaneously inoculated with 0.02 % of a frozen Bifidobacterium lactis Bb-12 preparation, provided by Chr Hansen, Hørsholm, Denmark) at 43 ⁰C.

Small portions of the inoculated soy base were fermented in closed tubs and left untouched to obtain a set type of product (Product A) .

A larger batch was fermented in a vessel without stirring. In both cases the mixture was incubated for a period of 7 hrs at 43°C during which the pH decreased from 6.5 to 4.8. A part of the larger batch was subsequently homogenised at 180 bar to obtain a soy yoghurt drink. The fermentation process was then stopped by rapidly cooling the mixture to 4 ⁰C. (Product B) . Another part was first heat treated for 20 sec. at 72°C and then homogenized at 180 bar to obtain a pasteurised soy yoghurt drink (Product C) . The viscosity of Products B and C was about 20 fold lower than that of non-homogenized Product A.

The three products were evaluated in duplo for flavour attributes by a sensory panel consisting of 11 trained panellists . Clear differences were observed. Product B had significantly stronger yoghurt and creamy flavour notes than Products A and C. This result shows that the lower viscosity of product B is associated with stronger dairy flavour notes. Product C, unlike products A and B, had a clearly perceptible soy odour and soy taste. These soy-related flavour notes were not liked by the panel members.

Claims

1. A method of manufacturing a beverage by fermenting a soy protein containing substrate, said method comprising:

- providing a pasteurised or sterilised aqueous liquid containing 0.5-8 wt . % of dissolved soy protein and 0-0.2 wt .% of dairy protein, said soy protein being derived from soybeans that have not been dehypocotylized;

- inoculating the pasteurised or sterilised liquid with a thermophilic lactic acid bacterium containing starter culture, said thermophilic lactic acid bacterium having an optimum growth temperature in excess of 35 ⁰C, preferably in excess of 38 ⁰C;

• lactate concentration increases with at least 500 ppm;

• diacetyl concentration increases with at least 0.3 ppm and/or acetaldehyde concentration increases with at least 0.05 ppm;

• concentration of n-hexanal decreases by at least 60%, preferably by at least 70%;

• at least two of n-pentanal concentration, n-heptanal concentration, n-octanal concentration and n-nonanal concentration decrease by at least 50%.

2. Method according to claim 1, wherein during fermentation concentration of 2-methylbutanal decreases by at least 25%, preferably by at least 40%.

3. Method according to claim 1 or 2, wherein during fermentation concentration of 3-methylbutanal decreases by at least 25%, preferably by at least 40%.

4. Method according to any one of the preceding claims, wherein during fermentation trans-2-hexenal concentration decreases by at least 50%, preferably by at least 60%.

5. Method according to any one of the preceding claims, wherein during fermentation at least two, preferably at least three of n-pentanal concentration, n-heptanal concentration, n- octanal concentration and n-nonanal concentration decrease by at least 60%.

6. Method according to any one of the preceding claims, wherein during fermentation pH decreases by at least 1.5 pH units.

7. Method according to any one of the preceding claims, wherein the fermented product contains at least 0.5 ppm of diacetyl.

8. Method according to any one of the preceding claims, wherein the fermented product contains at least 0.1 ppm of acetaldehyde .

9. Method according to any one of the preceding claims, comprising filling the fermented product into containers and subsequently sealing the filled containers, wherein edible acid is added to the fermented product prior to the filling into the containers so as to adjust the pH to less than 4.5 and wherein the fermented product is not subjected to pasteurisation or sterilisation prior, during or after the filling into the containers.

10. Method according to claim 9, wherein the edible acid is added to the fermented product before the fermentation has reached the point at which product inhibition prevents further production of lactic acid by lactic acid bacteria.

11. Method according to claim 9 or 10, wherein the addition of edible acid causes a pH decrease of at least 0.3 units, preferably of at least 0.5 units.

12. Method according to any one of claims 9-11, wherein the edible acid is selected from lactic acid, citric acid, malic acid, tannic acid, tartaric acid, phosphoric acid, acetic acid, maleic acid, succinic acid, gluconic acid, adipic acid, ascorbic acid, fumaric acid and combinations thereof .

13. Method according to claim 12, wherein the edible acid is selected from the group of lactic acid, citric acid and combinations thereof.

14. Method according to any one of the preceding claims, wherein the thermophilic lactic acid bacterium is selected from the group consisting of: Streptococcus thermophilics, Lactobacillus delbrueckii, Lactobacillus helveticus, Lactobacillus acidophilus, Lactobacillus casei, Lactobacillus reuteri, Lactobacillus rhamnosus and combinations thereof.

15. Method according to any one of the preceding claims, wherein the aqueous liquid containing 0.5-8 wt . % of dissolved soy protein is prepared from a soy protein source selected from the group consisting of soy isolate, soy concentrate, soy flour and combinations thereof, said soy protein source being derived from soy beans exhibiting a lipoxygenase activity of less than 15 kU/mg, more preferably of less than 10 kU/mg.

16. Method according to any one of the preceding claims, wherein the fermentation is carried out at a temperature within the range of 40-46 ⁰C, preferably of 40-45 ⁰C.

17. Method according to any one of the preceding claims, wherein at a temperature of 7 ⁰C the fermented product has a viscosity of less than 25 mPa.s at 100 s^'1.