Novel bacterial species for water kefir production and as a food additive

Abstract

A water kefir starter culture comprising either an isolated Bifidobacterium sp. nov, (isolated WK012_4_13) strain or an isolated Bifidobacterium sp. nov (WK041_4_12) (now assigned its official name of Bifidobacterium aquikefiricola) and one or more of a yeast, an acetic acid bacteria, and a lactic acid bacteria.

Description

[0001] Title

[0002] Novel bacterial species for water kefir production and as a food additive

[0003] Field of the Invention

[0004] The invention relates to a novel mix of starter species for consistent production of a pitched water kefir with desirable flavour profile. Specifically, the species include, for example, novel bifidobacteria species not previously identified or described, derived from a wide range of water kefirs, which can be used for water kefir or fermented beverage production and as food additives.

[0005] Background to the Invention

[0006] Fermented foods (FF) and beverages are consumed worldwide. The global FF market has grown to $575.6 billion in 2022. Over the years, researchers have explored the beneficial properties of these foods and beverages for the human body. FF have been shown to have beneficial effects on human health, and especially the human gut, e.g. 1) through the microbial production of beneficial metabolites (such as organic acids) and enzymes; 2) through the degradation of harmful or irritating compounds (e.g. FODMAPs); and 3) through the presence of the microbes themselves, as discussed in the "old friends" theory. This theory proposes that the rise in allergic and autoimmune conditions seen in recent decades is linked to reduced contact with microorganisms in industrialized societies. It is proposed that fermented foods, such as water kefir, might be key for safe microbial exposure. Water kefir, a beverage with a complex microbial community, is distinguished by its effervescence and distinctive flavour. The water kefir fermentation converts sugars into various metabolic by-products, including organic acids, carbon dioxide, low levels of alcohol, and various flavour compounds. This transformation, which contributes to the overall flavour and nutritional profile of the beverage, is facilitated by the intricate network of microorganisms within the inoculum.

[0007] In vitro and animal studies have suggested that water kefir could have potential health benefits, such as: gastroprotective effects, anti-inflammatory effects, anti-microbial effects, lower cholesterol, and benefit the immune system

[0008] WO 2023 / 072715 describes a fermented beverage composition, comprising gluconic acid and lactic acid, wherein the concentration ratio of gluconic acid to lactic acid is greater than 1 , preferably between 2 to 10, or between 3 to 9, or between 4 to 8, or between 5 to 7, or about 6. The beverage composition further comprises Zymomonas mobilis cells and one or more of lactic acid bacteria (Liquorilactobacillus, Lactiplantibacillus and combinations of these), Saccharomyces (Saccharomyces cerevisiae and / or Saccharomyces bayanus), and Bifidobacterium (Bifidobacterium aquikefiri). Flavour-associated volatiles in the water kefir were also described.

[0009] US Patent Application No. 16 / 406,843 describes a non-dairy fermented water kefir base composition comprising a) a mash, b) a starter culture (live yeast, live bacteria (Streptococcus thermophiles, Lactobacillus delbrueckii subsp. bulgaricus, Lactobcillus acidophilus, Lactobacillus paracasei subsp., and Bifidobacterium animalis), or combinations thereof), c) a medium, and d) an organic acid.

[0010] Artisanal water kefir production uses water kefir grains (often referred to as Tibicos, Water Crystals, and California Bees) that contain the microbes for the fermentation. The cultivation and isolation of bifidobacteria frequently poses considerable challenges in the production of water kefir or food products containing bifidobacteria due to their distinct growth requirements. Also, using water kefir grains for scaled up commercial production poses challenges due to inconsistency in production caused by the fluctuating water kefir microbiomes on the kefir grains, leading to inconsistent, unreliable products, which are difficult to market.

[0011] It is an object of the present invention to overcome at least one of the above-mentioned problems.

[0012] Summary of the Invention

[0013] On the basis of recent investigations, the Applicant predicted the existence of at least two novel bifidobacteria species in water kefir based on metagenome assembled genomes (MAGs) from shotgun sequencing data. Here, two novel Bifidobacterium species isolated from water kefir are described. Together, this novel blend of strains will facilitate large-scale and consistent commercial water kefir production with desirable flavour profiles, and they can also be used as valuable food additives. Firstly, as they can grow well under aerobic conditions unlike most bifidobacteria, and, secondly, bifidobacteria in general are of commercial interest due to their proven and perceived health benefits. In contrast to the typical mesophilic and obligate anaerobic nature of bifidobacteria, the novel species exhibit aerotolerance and demonstrate optimal growth at a relatively low temperature, at about 30°C. There is provided, as set out in the appended claims, an isolated Bifidobacteria strain, Bifidobacterium sp. nov (WK012_4_13) (now assigned its official name of Bifidobacterium fermentum), deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 26th December 2023 under LMG Deposit Accession No. LMG 33104 (deposited in the name of Teagasc Food Research Centre, National Dairy Products Research Centre, Fermoy, Co. Cork, Ireland), and variants thereof, wherein the variants are characterised in that they are isolated, they belong or are closely related to the species Bifidobacterium sp. nov (WK012_4_13) (now assigned its official name of Bifidobacterium fermentum), and they show F6PPK activity, and exhibit characteristics such as the ability to grow on D-arabinose and D-xylose and a NaCI (w / v%) of 0-3 (when compared to control B. aquikefiri LMG 28769; see Table 2 below) The isolated Bifidobacteria strain Bifidobacterium sp. nov (WK012_4_13) (now assigned its official name of Bifidobacterium fermentum) deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 26th December 2023 under LMG Deposit Accession No. LMG 33104 (deposited in the name of Teagasc Food Research Centre, National Dairy Products Research Centre, Fermoy, Co. Cork, Ireland) will hereafter be referred to as “isolated WK012_4_13 strain of the invention”, and the variants thereof are hereafter to as “WK012_4_13 variants”.

[0014] There is provided, as set out in the appended claims, an isolated Bifidobacteria strain, Bifidobacterium sp. nov (WK041_4_12) (now assigned its official name of Bifidobacterium aquikefiricola), deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 20thJuly 2023 under LMG Deposit Accession No. LMG 33105 (deposited in the name of Teagasc, Moorepark Food Research Centre, Fermoy, Co. Cork, Ireland), and variants thereof, wherein the variants are characterised in that they are isolated, they belong to or are closely related to the species Bifidobacterium sp. nov (WK041_4_12) (now assigned its official name of Bifidobacterium aquikefiricola), and they show F6PPK and N-acetyl-B-glucosaminidase activity, and exhibit characteristics such as the ability to grow on dulcitol but not D-mannose or D- saccharose, and a NaCI (w / v%) of 0-3 (when compared to control B. aquikefiri LMG 28769; see Table 2 below). The isolated Bifidobacteria strain sp. nov (WK041_4_12) (now assigned its official name of Bifidobacterium aquikefiricola) deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 20thJuly 2023 under LMG Deposit Accession No. LMG 33105 (deposited in the name of Teagasc Food Research Centre, National Dairy Products Research Centre, Fermoy, Co. Cork, Ireland) will hereafter be referred to as “isolated WK041_4_12 strain of the invention”, and the variants thereof are hereafter to as “WK041_4_12 variants”.

[0015] In one aspect, there is provided an isolated Bifidobacteria strain Bifidobacterium sp. nov (WK012_4_13) (now assigned its official name of Bifidobacterium fermentum) deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 26thDecember 2023 under LMG Deposit Accession No. LMG 33104.

[0016] In one aspect, the isolated Bifidobacteria strain Bifidobacterium sp. nov (WK012_4_13) (now assigned its official name of Bifidobacterium fermentum) deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 26thDecember 2023 under LMG Deposit Accession No. LMG 33104 further comprises the characteristic of having fructose-6-phosphate phosphoketolase activity, can be grown in aerobic or anaerobic conditions at between 14°C and 38°C; and can grow on D-arabinose and D-xylose and a NaCI (w / v%) concentration of 0-3, when compared to a control B. aquikefiri LMG 28769.

[0017] In one aspect, there is provided an isolated Bifidobacteria strain Bifidobacterium sp. nov (WK041_4_12) (now assigned its official name of Bifidobacterium aquikefiricola) deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 20thJuly 2023 under LMG Deposit Accession No. LMG 33105.

[0018] In one aspect, the isolated Bifidobacteria strain Bifidobacterium sp. nov (WK041_4_12) (now assigned its official name of Bifidobacterium aquikefiricola) deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 20thJuly 2023 under LMG Deposit Accession No. LMG 33105 further comprises the characteristic of having fructose-6-phosphate phosphoketolase activity and N-acetyl-B-glucosaminidase activity, can grow in aerobic or anaerobic conditions at between 14°C and 38°C, and can grow on dulcitol but not D-mannose or D-saccharose, and a NaCI (w / v%) concentration of 0-3, when compared to a control B. aquikefiri LMG 28769.

[0019] In one aspect, the isolated Bifidobacteria strains described above are in a freeze-dried form or a spray-dried form.

[0020] In one aspect, there is provided a variant of the isolated Bifidobacteria strain Bifidobacterium sp. nov (WK012_4_13) (now assigned its official name of Bifidobacterium fermentum) deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 26thDecember 2023 under LMG Deposit Accession No. LMG 33104, characterised in that it is isolated, it belongs to the Bifidobacterium sp. nov (WK012_4_13) (now assigned its official name of Bifidobacterium fermenturn), and has the characteristics of having fructose-6-phosphate phosphoketolase activity, can be grown in aerobic or anaerobic conditions at between 14°C and 38°C; and can grow on D-arabinose and D-xylose and a NaCI (w / v%) concentration of 0-3, when compared to a control B. aquikefiri LMG 28769.

[0021] In one aspect, there is provided a variant of the isolated Bifidobacteria strain Bifidobacterium sp. nov (WK041_4_12) (now assigned its official name of Bifidobacterium aquikefiricola) deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 20thJuly 2023 under LMG Deposit Accession No. LMG 33105, characterised in that it is isolated, it belongs to the Bifidobacterium sp. nov (WK041_4_12) (now assigned its official name of Bifidobacterium aquikefiricola), and has the characteristics of having fructose-6-phosphate phosphoketolase activity and N- acetyl-B-glucosaminidase activity, can grow in aerobic or anaerobic conditions at between 14°C and 38°C, and can grow on dulcitol but not D-mannose or D-saccharose, and a NaCI (w / v%) concentration of 0-3, when compared to a control B. aquikefiri LMG 28769.

[0022] In one aspect, there is provided a water kefir starter culture comprising either one of the isolated Bifidobacteria strains described above, and one or more of a yeast, an acetic acid bacteria, and a lactic acid bacteria.

[0023] In one aspect, the water kefir starter culture further comprises one or more species from the genus Bifidobacterium or Zymomonas.

[0024] In one aspect, there is provided a water kefir starter culture comprising either one of the isolated Bifidobacteria strains described above, and one or more of a yeast, an acetic acid bacteria, a lactic acid bacteria, and a Zymomonas bacteria.

[0025] In one aspect, the water kefir starter culture comprises both of the isolated Bifidobacteria strains described above.

[0026] Preferaby, the Bifidobacterium is Bifidobacterium tibiigranuli and the Zymomonas is Zymomonas mobilis. In one aspect, there is provided a water kefir starter culture comprising a yeast and optionally one or more selected from an acetic acid bacteria, a lactic acid bacterium, a Bifidobacteria strain, and a Zymomonas bacteria.

[0027] In one aspect, there is provided a water kefir starter culture comprising a yeast and an acetic acid bacteria, and optionally one or more selected from the lactic acid bacterium, the Bifidobacteria strain, and the Zymomonas bacteria.

[0028] In one aspect, the water kefir starter culture comprises the isolated Bifidobacteria strain Bifidobacterium sp. nov (WK012_4_13) (now assigned its official name of Bifidobacterium fermentum) and a yeast.

[0029] In one aspect, the water kefir starter culture comprises the isolated Bifidobacteria strain Bifidobacterium sp. nov (WK012_4_13) (now assigned its official name of Bifidobacterium fermentum), a yeast, and a lactic acid bacteria.

[0030] In one aspect, the water kefir starter culture comprises the isolated Bifidobacteria strain Bifidobacterium sp. nov (WK012_4_13) (now assigned its official name of Bifidobacterium fermentum), a yeast, and an acetic acid bacteria.

[0031] In one aspect, the water kefir starter culture comprises the isolated Bifidobacteria strain Bifidobacterium sp. nov (WK012_4_13) (now assigned its official name of Bifidobacterium fermentum) and a Zymomonas strain.

[0032] In one aspect, the water kefir starter culture comprises the isolated Bifidobacteria strain Bifidobacterium sp. nov (WK012_4_13) (now assigned its official name of Bifidobacterium fermentum), a lactic acid bacteria, and a Zymomonas strain.

[0033] In one aspect, the water kefir starter culture comprises the isolated Bifidobacteria strain Bifidobacterium sp. nov (WK012_4_13) (now assigned its official name of Bifidobacterium fermentum), an acetic acid bacteria, and a Zymomonas strain.

[0034] In one aspect, the water kefir starter culture comprises the isolated Bifidobacteria strain Bifidobacterium sp. nov (WK041_4_12) (now assigned its official name of Bifidobacterium aquikefiricola) and a yeast. In one aspect, the water kefir starter culture comprises the isolated Bifidobacteria strain Bifidobacterium sp. nov (WK041_4_12) (now assigned its official name of Bifidobacterium aquikefiricola), a yeast, and a lactic acid bacteria.

[0035] In one aspect, the water kefir starter culture comprises the isolated Bifidobacteria strain Bifidobacterium sp. nov (WK041_4_12) (now assigned its official name of Bifidobacterium aquikefiricola), a yeast, and an acetic acid bacteria.

[0036] In one aspect, the water kefir starter culture comprises the isolated Bifidobacteria strain Bifidobacterium sp. nov (WK041_4_12) (now assigned its official name of Bifidobacterium aquikefiricola) and a Zymomonas strain.

[0037] In one aspect, the water kefir starter culture comprises the isolated Bifidobacteria strain Bifidobacterium sp. nov (WK041_4_12) (now assigned its official name of Bifidobacterium aquikefiricola), a lactic acid bacteria, and a Zymomonas strain.

[0038] In one aspect, the water kefir starter culture comprises the isolated Bifidobacteria strain Bifidobacterium sp. nov (WK041_4_12) (now assigned its official name of Bifidobacterium aquikefiricola), an acetic acid bacteria, and a Zymomonas strain.

[0039] In one aspect, the yeast is selected from a Saccharomyces, a Brettanomyces, a Zygotorulaspora, a Pichia, a Lachancea, a Hanseniaspora, a Candida, a Schizosaccharomyces or a combination thereof. Preferably, the Saccharomyces yeast is selected from Saccharomyces cerevisiae, Saccharomyces bayanus, Saccharomyces uvarum, a Saccharomyces species isolated from water kefir or a combination thereof.

[0040] In one aspect, the Brettanomyces yeast is selected from Brettanomyces bruxellensis, Brettanomyces anomalus, Brettanomyces sp._HC-2020a, or a combination thereof.

[0041] In one aspect, the Zygotorulaspora yeast is Zygotorulaspora florentina.

[0042] In one aspect, the Pichia yeast is selected from Pichia fermentans, Pichia kluyveri, Pichia membranifaciens, Pichia occidentalis, or a combination thereof. In one aspect, the acetic acid bacteria is selected from Gluconacetobacter, Gluconobacter, Acetobacter, Komagataeibacter, Novacetimonas or a combination thereof.

[0043] Preferably, the Gluconobacter bacteria is selected from Gluconacetobacter liquefaciens, Gluconobacter albidus, Gluconobacter cadivus, Gluconobacter cerinus, Gluconobacter frateurii, Gluconobacter japonicus, Gluconobacter kanchanaburiensis, Gluconobacter kondonii, Gluconobacter oxydans, Gluconobacter potus, Gluconobacter roseus, Gluconobacter sp. Gdi, Gluconobacter vitians, or a combination thereof.

[0044] Preferably, the Acetobacter bacteria is selected from Acetobacter aceti, Acetobacter fabarum, Acetobacter indonesiensis, Acetobacter lovaniensis, Acetobacter malorum, Acetobacter okinawensis, Acetobacter orientalis, Acetobacter papaya, Acetobacter pasteurianus, Acetobacter persici, Acetobacter senegalensis, Acetobacter sicerae, Acetobacter sp. UBA5411, Acetobacter syzygii, Acetobacter tropicalis, an Acetobacter species isolated from water kefir, or a combination thereof.

[0045] In one aspect, the acetic acid bacteria is Gluconacetobacter dulcium, Komagataeibacter saccharivorans, or a combination thereof.

[0046] In one aspect, the lactic acid bacteria is selected from the genus Liquorilactobacillus , Lacticaseibacillus, Lentilactobacillus, Leuconostoc, Oenococcus, Schleiferilactobacillus, Sporolactobacillus or a combination thereof.

[0047] Preferably, the lactic acid bacteria is selected from Liquorilactobacillus satsumensis, Liquorilactobacillus mall, Liquorilactobacillus ghanensis, Liquorilactobacillus horde!, Liquorilactobacillus nagelii, a Liquorilactobacillus species isolated from water kefir, Lacticaseibacillus paracasei, Leuconostoc pseudomesenteroides, Lentilactobacillus hilgardii, Oenococcus oeni, Oenococcus sicerae, Oenococcus kitaharae, or a combination thereof.

[0048] More preferably, the acetic acid bacteria is selected from Acetobacter indonesiensis, Acetobacter okinawensis, Acetobacter orientalis, Gluconobacter cerinus, Gluconobacter japonicus, Gluconobacter kondonii, Gluconobacter oxydans, or a combination thereof.

[0049] In one aspect, there is provided a water kefir starter culture comprising a Bifidobacteria strain selected from an isolated Bifidobacteria strain (WK012_4_13) Bifidobacterium fermentum deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 26thDecember 2023 under LMG Deposit Accession No. LMG 33104, an isolated Bifidobacteria strain Bifidobacterium sp. nov (WK041_4_12) Bifidobacterium aquikefiricola deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 20thJuly 2023 under LMG Deposit Accession No. LMG 33105, Bifidobacterium tibiigranulil, Bifidobacterium aquikefiri, Bifidobacterium psychraerophilum, Bifidobacterium aquikefiri, Bifidobacterium psychraerophilum, or variants thereof, and, and optionally one or more selected from a yeast, an acetic acid bacteria, a lactic acid bacterium, a Zymomonas bacteria.

[0050] In one aspect, both of the isolated Bifidobacteria strains are selected. Preferably, the isolated Bifidobacteria strain comprises the characteristic of having fructose-6-phosphate phosphoketolase activity, can be grown in aerobic or anaerobic conditions at between 14°C and 38°C; and can grow on D-arabinose and D-xylose and a NaCI concentration of 0-3 (w / v%), when compared to a control B. aquikefiri LMG 28769. Preferably, the isolated Bifidobacteria strain comprises the characteristic of having fructose-6-phosphate phosphoketolase activity and N-acetyl-B-glucosaminidase activity, can be grown in aerobic or anaerobic conditions at between 14°C and 38°C, and can grow on dulcitol but not D-mannose or D-saccharose, and a NaCI concentration of 0-3 (w / v%), when compared to control B. aquikefiri LMG 28769.

[0051] In one aspect, the yeast and bacterial strains are in a freeze-dried or a spray-dried form.

[0052] In one aspect, the Zymomonas is Zymomonas mobilis. In one aspect, the Zymomonas mobilis is Zymomonas mobilis subsp. mobilis, Zymomonas mobilis subsp. pomaceae, or a combination of both.

[0053] In one aspect, there is provided a formulation comprising at least one of the isolated Bifidobacteria strains described above having their respective characteristics as described above, or a variant strain as described above. Specifically, the Bifidobacteria strain is selected from an isolated Bifidobacteria strain (WK012_4_13) Bifidobacterium fermentum deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 26thDecember 2023 under LMG Deposit Accession No. LMG 33104, an isolated Bifidobacteria strain Bifidobacterium sp. nov (WK041_4_12) Bifidobacterium aquikefiricola deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 20thJuly 2023 under LMG Deposit Accession No. LMG 33105, Bifidobacterium tibiigranulil , Bifidobacterium aquikefiri, Bifidobacterium psychraerophilum, Bifidobacterium aquikefiri, Bifidobacterium psychraerophilum, or variants thereof. In one aspect, both of the isolated Bifidobacteria strains are selected.

[0054] In one aspect, the formulation is a pharmaceutical formulation and comprises a pharmaceutically acceptable carrier.

[0055] In one aspect, the formulation is a food product; fermented dairy product; a fermented non-dairy product; a sugar-based fermented beverage; and / or a water kefir starter culture.

[0056] In one aspect, there is provided a formulation comprising an isolated strain of the invention, or a variant strain of the invention. Suitably, the formulation is a pharmaceutical formulation and additionally comprises a pharmaceutically acceptable carrier. Alternatively, the formulation may be a comestible product, for example a food product. Ideally, the food product is a fermented food, for example a fermented non-dairy product (such as water-based kefir, kombucha, tepache, beer, wine, sauerkraut, kimchi, pickles, fermented non-dairy milk alternatives (such as soy milk kefir, almond milk kefir, oat milk kefir, or coconut milk kefir), and others (miso, tamari, tempeh) or a fermented dairy product (such as a buttermilk, acidophilus milk, sour cream, cottage cheese, aged cheese, yoghurt or milk kefir) or a food product fortified with probiotics (such as unfermented or fermented beverages or foods and juices supplemented with probiotics). For example, kombucha products can contain probiotics which have been added after pasteurisation / filtration. In another embodiment, the formulation may be a starter culture that optionally comprises additional strains of bacteria.

[0057] In one aspect, there is provided an isolated strain of the invention, or a variant strain of the invention, for use in improving lipid profiles in an individual and in the treatment or prevention of cardiovascular risk factors, diabetes, and hyperlipidaemia in an individual.

[0058] In one aspect, there is provided an isolated strain of the invention, or a variant strain of the invention, for use as a probiotic culture.

[0059] In one aspect, there is provided, an isolated strain of the invention, or a variant strain of the invention, for use in the manufacture of water kefir or a dairy product (such as a yoghurt), in which the isolated strain or the variant thereof is optionally used a starter culture.

[0060] Definitions

[0061] In the specification, the term “isolated” should be considered to mean material removed from its original environment in which it naturally occurs, for example, in this instance a bacterial strain of the mammalian gut. The removed material is typically cultivated, purified and cultured separately from the environment in which it was located. Thus, the purified isolated bacterial strain in this instance ideally does not contain any significant amounts of other bacterial strains. The isolated strain or variant of the invention may be provided in a viable or non-viable form, and in a culturable or non-culturable form. The invention also relates to an isolated strain of the invention, or variant thereof, of an exopolysaccharide of the invention, in any format, for example a freeze-dried form, a suspension, a powder, or a broth.

[0062] The term “freeze-dried form” should be understood to mean that the strain of the invention, optionally together with other ingredients including, for example, preservatives, is frozen and then the ice crystals in the frozen strain are sublimated under vacuum.

[0063] The term “comestible product” should be understood to include products that are intended to be consumed by ingestion by humans or animals, such as foods and drinks. In particular, the comestible product is a food or drink product intended for consumption by humans, for example a fermented product (such as a water-based kefir, kombucha, and the like) or a dairy product, especially a fermented dairy product (such as a yoghurt, a prebiotic and / or a probiotic).

[0064] When used herein, the term “pharmaceutical composition” may comprise one or more pharmaceutically acceptable diluents, excipients or carriers. Even though the peptides and compositions of the present invention can be administered alone, they will generally be administered in admixture with a pharmaceutical carrier, excipient or diluent, particularly for human therapy. The pharmaceutical compositions may be for human or animal usage in human and veterinary medicine. Examples of such suitable excipients for the various different forms of pharmaceutical compositions described herein may be found in the “Handbook of Pharmaceutical Excipients”, 2nd Edition, (1994), edited by A Wade and PJ Weller. In particular, formulations for topical delivery are described in “Topical Drug Delivery Formulations” edited by David Osborne and Antonio Aman, Taylor & Francis, the complete contents of which are incorporated herein by reference. Acceptable carriers or diluents for therapeutic use are well known in the pharmaceutical art, and are described, for example, in Remington's Pharmaceutical Sciences, Mack Publishing Co. (A. R. Gennaro edit. 1985).

[0065] In the specification, "pharmaceutically acceptable" means approved by a regulatory agency of the Federal or a state government or listed in the U.S. Pharmacopeia or other generally recognized pharmacopeia for use in animals, and more particularly in humans.

[0066] The choice of pharmaceutical carrier, excipient or diluent can be selected with regard to the intended route of administration and standard pharmaceutical practice. The pharmaceutical compositions may comprise as, or in addition to, the carrier, excipient or diluent any suitable binder(s), lubricant(s), suspending agent(s), coating agent(s), solubilising agent(s). Examples of suitable binders include starch, gelatin, natural sugars such as glucose, anhydrous lactose, free-flow lactose, beta-lactose, corn sweeteners, natural and synthetic gums, such as acacia, tragacanth or sodium alginate, carboxymethyl cellulose and polyethylene glycol. Examples of suitable lubricants include sodium oleate, sodium stearate, magnesium stearate, sodium benzoate, sodium acetate, sodium chloride and the like. Preservatives, stabilizers, dyes and even flavouring agents may be provided in the pharmaceutical composition. Preferably, the excipient is one or more non-natural excipients.

[0067] The term "carrier" refers to a diluent, adjuvant, excipient, or vehicle with which the composition is administered. Examples of suitable carriers include lactose, starch, glucose, methyl cellulose, magnesium stearate, mannitol, sorbitol and the like. Examples of suitable diluents include ethanol, glycerol, and water. Pharmaceutical carriers can be sterile liquids, such as water and oils, including those of petroleum, animal, vegetable, or synthetic origin, such as peanut oil, soybean oil, mineral oil, sesame oil and the like. Water is a preferred carrier when the pharmaceutical composition is administered intravenously. Saline solutions and aqueous dextrose and glycerol solutions can also be employed as liquid carriers, particularly for injectable solutions. Suitable pharmaceutical excipients include starch, glucose, lactose, sucrose, gelatin, malt, rice, flour, chalk, silica gel, sodium stearate, glycerol monostearate, talc, sodium chloride, dried skim milk, glycerol, propylene glycol, water, ethanol and the like. Saline solutions and aqueous dextrose and glycerol solutions can also be employed as liquid carriers. Suitable excipients include starch, glucose, lactose, sucrose, gelatin, malt, rice, flour, chalk, silica gel, sodium stearate, glycerol monostearate, talc, sodium chloride, dried skim milk, glycerol, propylene glycol, water, ethanol and the like.

[0068] In the specification, the term “individual” or “patient” should be understood to mean all mammals, for example, a human, primates, non-human primates, farm animals (such as pigs, horses, goats, sheep, cows (including bulls, bullocks, heifers etc.), donkey, reindeer, etc), veterinary mammals (such as dogs, cats, rabbits, hamsters, guinea pigs, mice, rats, ferrets, etc.), and mammals kept in captivity (such as lions, tigers, elephants, zebras, giraffes, pandas, rhino, hippopotamus, etc.), and other mammals and higher mammals for which the use of the invention is practicable.

[0069] The term “variant” as applied to the isolated Bifidobacterium sp. nov (WK012_4_13) (now assigned its official name of Bifidobacterium fermentum), deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 26th December 2023 under LMG Deposit Accession No. LMG 33104 is one that retains the phenotypical characteristics of the bacterium as described herein. Preferably, the term variant should be understood to mean progeny (unmodified descendants), modified descendants, or derivatives of the isolated WK012_4_13 strain of the invention, for example strains which are genetically modified to alter the genotype of the bacteria, or strains which are altered by natural processes such as selection or serial passage. The variant generally has a 16S rRNA fragment gene sequence ( / .e., a partial 16S rRNA gene sequence) that is identical or substantially identical with the isolated WK012_4_13 strain of the invention, for example at least 98%, 98.5%, 99%, 99.1%, 99.2%, 99.3%, 99.4%, 99.5%, 99.6%, 99.7%, 99.8% or 99.9% identical. Sequence identity can be determined using an online algorithm “BLAST”, publicly available at http: / / www.ncbi.nlrTi.nih.gov / BL ST / , or EMBOSS Needle (http: / / www.ebi.ac.uk / tools / psa / emboss needle / ). Alternatively, the variant generally has a 16S rRNA fragment gene sequence that is identical or has a genomic average nucleotide identity (ANI) of 95% or higher with the isolated WK012_4_13 strain of the invention, for example at least 95%, 95.5%, 96%, 96.5%, 97%, 97.5%, 98%, 98.5%, 98.7%, 99%, 99.1%, 99.2%, 99.3%, 99.4%, 99.5%, 99.6%, 99.7%, 99.8% or 99.9% identical. Novel bacterial species are defined as <98.7% 16S sequence similarity to the closest related species, or an overall genome related index <95~96 % ANI or 70 % dDDH (digital DNA-DNA hybridization) to the closest related species. The reference for these numbers: DOI 10.1099 / ijsem.0.002516, e.g. https: / / qtdb.ecogenomic.org / tools / fastani could be used as tool for ANI calculations. This is the publication of another tool that calculates ANI scores: https: / / www.nature.com / articles / s41467-018-07641-9.

[0070] The term “variant” as applied to the isolated Bifidobacteria strain, Bifidobacterium sp. nov (WK041_4_12) (now assigned its official name of Bifidobacterium aquikefiricola), deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 20thJuly 2023 under LMG Deposit Accession No. LMG 33105 is one that retains the phenotypical characteristics of the bacterium as described herein. Preferably, the term variant should be understood to mean progeny (unmodified descendants), modified descendants, or derivatives of the isolated WK012_4_12 strain of the invention, for example strains which are genetically modified to alter the genotype of the bacteria, or strains which are altered by natural processes such as selection or serial passage. The variant generally has a 16S rRNA fragment gene sequence ( / .e., a partial 16S rRNA gene sequence) that is identical or substantially identical with the isolated WK012_4_13 strain of the invention, for example at least 98%, 98.5%, 99%, 99.1%, 99.2%, 99.3%, 99.4%, 99.5%, 99.6%, 99.7%, 99.8% or 99.9% identical. Sequence identity can be determined using an online algorithm “BLAST”, publicly available at http: / / www.ncbi.nlm.nih.gov / , or EMBOSS Needle

[0071] (http: / / www.ebi.ac.uk / tools / psa / emboss needle / ). Alternatively, the variant generally has a 16S rRNA fragment gene sequence that is identical or has a genomic average nucleotide identity (ANI) of 95% or higher with the isolated WK041_4_12 strain of the invention, for example at least 95%, 95.5%, 96%, 96.5%, 97%, 97.5%, 98%, 98.5%, 98.7%, 99%, 99.1%, 99.2%, 99.3%, 99.4%, 99.5%, 99.6%, 99.7%, 99.8% or 99.9% identical. Novel bacterial species are defined as <98.7% 16S sequence similarity to the closest related species, or an overall genome related index <95~96 % ANI or 70 % dDDH (digital DNA-DNA hybridization) to the closest related species.

[0072] In the specification, the term “sequence identity” should be understood mean the amount of nucleotides which match between different sequences. For example, a 16S rRNA gene sequence that shares at least 98% sequence identity with a reference sequence is one in which any 98% of aligned nucleotides of the variant are identical to the corresponding nucleotides in the reference sequence across the entire length of the sequence. Sequence identity is the amount of characters which match exactly between two different sequences. Hereby, gaps are not counted, and the measurement is relational to the shorter of the two sequences. In the specification, the “substrate” insofar as it relates to the fermentation / fermentation medium (aqueous solution) should be understood to typically contain a sugar source (such as white sugar or brown sugar) and additional nutrient sources. The nutrient sources can be fruits or vegetables, fruit or vegetable extracts, yeast extracts or other nutritional extract, or food waste streams (such as apple pomace, brewer spent grains, fruit and vegetable peels and pulp, cereal bran, and the like). The sugar and nutrient source can be the same.

[0073] Brief Description of the Drawings

[0074] The invention will be more clearly understood from the following description of an embodiment thereof, given by way of example only, with reference to the accompanying drawings, in which:-

[0075] Figure 1 illustrates a phylogenetic tree based on 16S rRNA gene sequences of strains WK012_4_13, WK013_4_14, WK048_4_13, and WK041_4_12 and its relationship to other closely related species of the genus Bifidobacterium. Tamura-Nei substitution model was used for this analysis. The tree nodes show the bootstrap values (>50%) obtained from 1000 replications. The scale bar, labelled 0.02, represents the number of substitutions per site. To ensure accuracy, all locations with gaps and missing data were excluded (complete deletion option). The analysis included a total of 29 nucleotide sequences, with 1331 positions in the final dataset. The sequence of Scardovia inopinata JCM 12537Twas used as an outgroup.

[0076] Figure 2 shows the phylogenetic tree was inferred using the de novo work flow with curated bifidobacteria genomes and 117 marker genes in GTDB-Tk using WAG+GAMMA model. Scardovia inopinata JCM 12537Tand Scardovia wiggsie F0424 was used as outgroups. The collapsed node g Bifidobacterium contains an additional 98 leaves.

[0077] Figure 3 is an image obtained by a scanning electron microscopy (SEM) showing the morphology and cell size of the four strains after incubation in MRScys at 30°C for 48 hours: (a) WK048_4_13, (b) WK013_4_14, (c) WK012_4_13, and (d) WK041_4_12.

[0078] Figure 4 shows a main spectra profile based dendrogram of the two novel species of the claimed invention, the closely related reference strain Bifidobacterium aquikefiri LMG 28769T, and 25 strains from the Bruker taxonomy library representing the genus Bifidobacterium. The distances are normalized to a maximum value of 1000 and correspond to the relative similarity of MS spectra. Figure 5 is an example composition of a water kefir rich in the isolated WK012_4_13 strain. Shades of red indicated relative metagenomic species abundance in %, calculated based on genome coverage based abundance.

[0079] Figure 6 is an example composition of a water kefir rich in the isolated WK041_4_12 strain. Shades of red indicated relative metagenomic species abundance in %, calculated based on genome coverage based abundance.

[0080] Figure 7 is a comparison of the MALDI-TOF mass spectra of strain WK012_4_13 (b) and strain WK041_4_12 (c), with the closest species Bifidobacterium aquikefiri LMG 28769 (a). The profile is shown in the range between 2,000 and 15,000 m / z and peak intensity is indicated in arbitrary units.

[0081] Figure 8 is a bar chart showing Genera detected in water kefir and the percentage of different water kefirs they were detected in.

[0082] Figure 9 is a bar chart showing species detected in water kefir and the percentage of different water kefirs they were detected in as per Figure 8 above. Novel species are indicated with a genus name followed with _WK[number]_bin. [number], such as Acetobacter_WK028_bin.4.

[0083] Figure 10 is a dot plot showing microbial influence on aroma. Fermentations were evaluated by 9 to 15 volunteers on a scale of 1 to 5, with 5 indicating the most pleasant aroma. The 256 pitched fermentations are grouped by (a) yeast species, (b) Bifidobacterium species, (c) LAB, (d) AAB, and (e) the presence or absence of Zymomonas mobilis. Water kefir (WK) grain-based fermentations were used as a reference and inoculated at 120 g L“1(red), 60 g L“1(purple), 30 g L“1(blue), or 15 g L“1(green).

[0084] Figure 11 is a series of dot plots showing microbial influence on sugars and sugar alcohols in pitched water kefir. A subset of fermentations was analysed for sucrose, fructose, glucose, glycerol, and mannitol concentrations (rows correspond to these compounds in order). Pitched fermentations are grouped by yeast species (first column), Bifidobacterium species (second column), LAB (third column), AAB (fourth column), and the presence or absence of Zymomonas mobilis (fifth column). WK grain-based fermentations were used as a reference and inoculated at 120 g L“1(red), 60 g L“1(purple), 30 g L“1(blue), or 15 g L“1(green).

[0085] Figure 12 is a series of microbial influence on ethanol and organic acids in pitched water kefir. A subset of fermentations was analysed for ethanol (first row), acetic acid (second row), and lactic acid (third row) concentrations. Pitched fermentations are grouped by yeast species (first column), Bifidobacterium species (second column), LAB (third column), AAB (fourth column), and the presence or absence of Zymomonas mobilis (fifth column). WK grain-based fermentations were included as a reference and inoculated at 120 g L“1(red), 60 g L-1(purple), 30 g L-1(blue), or 15 g L-1(green).

[0086] Figure 13 is a series of bar charts showing the sensory affective hedonic evaluation of water kefir. Mean data of 25 naive assessors with SEM (standard error of the mean).

[0087] Figure 14 is a series of bar charts showing the Optimised Descriptive Profile (ODP) of water kefir. Mean data of 25 naive assessors with SEM (standard error of the mean).

[0088] Figure 15 is a series of dot blots showing the metabolite analysis of samples consumes in water kefir taste panel. Data represent two independent fermentations, each performed with three technical replicates.

[0089] Figure 16 is a series of graphs showing how omission of AAB reduces pH acidification and dissolved oxygen (DO) consumption in pitched fermentations. Fermentations were continuously monitored using the BioLector system for pH (a, b), biomass (in relative units; RU) (c, d), and dissolved oxygen (DO) levels (e, f). The elevated biomass observed in “B.acola S.cer Lsat Z.mob" (d) was not reflected by an increase in cell concentration as measured by flow cytometry. Black lines indicate the mean of biological replicates.

[0090] Figure 17 is a series of dot blots showing that AAB influence the metabolite composition of pitched water kefir. Metabolite composition was analysed 3 days after the start of the pitched fermentations.

[0091] Figure 18 is a bar chart showing the cumulative abundance of volatile acids, alcohols, and esters in water kefir samples used forthe taste panel, along with one media control. Abundances of chemical classes are presented in relative units (RU). Data represent the mean of two biological replicates, each measured in three technical replicates.

[0092] Figure 19 is a bar chart showing the abundance of the five most abundant volatile compounds in water kefir samples used for the taste panel, along with one media control. Compound abundances are presented in relative units (RU). Data represents the mean of two biological replicates, each measured in three technical replicates.

[0093] Figure 20 shows row-scaled volatile organic compound abundances in water kefir samples used forthe taste panel, along with one media control. Data represents the mean of two biological replicates, each measured in three technical replicates. Detailed Description of the Drawings

[0094] The Applicants analysed 69 water kefir grains from 21 countries worldwide using metagenomics to identify core genera within the water kefir pangenome, and metabolomics, including volatile organic compound (VOC) analysis, to characterize the metabolic influence of individual species. The 13 water kefir core genera that were identified are: Saccharomyces, Gluconobacter, Liquorilactobacillus, Acetobacter, Lacticaseibacillus, Lentilactobacillus, Bifidobacterium, Leuconostoc, Zygotorulaspora, Pichia, Brettanomyces, Oenococcus, and Zymomonas. In addition to the core genera, the Applicants have identified novel species within the water kefir microbiome. These include two novel species of Bifidobacterium. Bifidobacteria are of particular commercial interest for their frequent use in probiotics and foods, as well as proven and perceived health benefits. The water kefir fermentations that were analysed contained on average 10.3 species and a maximum of 22 species. The insights gained by the Applicants into the water kefir pangenome and metabolome allow the Applicants to rationally design water kefir starter cultures, that are able to reproduce artisanal water kefir and are not based on unreliable water kefir grains, and can be used for commercial water kefir production.

[0095] Currently, the Applicants have isolated 30+ strains, from 23 different species, representing 13 of the 13 core genera. Particularly noteworthy is the isolation of two completely novel Bifidobacterium species:

[0096] 1) strain: WK012_4_13 isolated from a water kefir obtained from a private household in Italy and which is Bifidobacterium sp. nov deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 26th December 2023 under LMG Deposit Accession No. LMG 33104 (deposited in the name of Teagasc Food Research Centre, National Dairy Products Research Centre, Fermoy, Co. Cork, Ireland). The strain was also deposited with the Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (DSMZ) on 4th October 2023 under DSM Accession No. DSM 116073.

[0097] 2) strain: WK041_4_12 isolated from a water kefir obtained from a private household in Singapore and which is Bifidobacterium sp. nov deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 20th July 2023 under LMG Deposit Accession No. LMG 33105. The strain was also deposited with the Deutsche Sammlung von Mikroorganismen und Zellkulturen GmbH (DSMZ) on 4th October 2023 under DSM Accession No. DSM 116074. Both species are able to grow well in aerobic conditions unlike most bifidobacteria. Most bifidobacteria grow only under anaerobic conditions, posing challenges in handling and industrial processes such as food and supplement production. The unique characteristic of these novel Bifidobacterium strains to thrive in aerobic conditions simplifies their handling and allows for potential usage in environments unsuitable for other bifidobacteria strains.

[0098] One option for a starter culture will contain combinations of the following organisms from the water kefir core genera (see Table 1 for strain detail):

[0099] 1. Novel Bifidobacterium species: one or both of the novel isolated Bifidobacterium species of the claimed invention.

[0100] 2. Yeast: One or more species of yeast. Yeasts should preferentially be from the genus Saccharomyces, such as Saccharomyces cerevisiae, or from the genus Brettanomyces, such as Brettanomyces bruxellensis, or from the genus Zygotorulaspora, such as Zygotorulaspora florentina, or from the genus Lachancea, such as Lachancea cidri, or from the genus Hanseniaspora, such as Hanseniaspora valbyensis, or from the genus Pichia, such as Pichia fermentans.

[0101] 3. Acetic acid bacteria (AAB): One or more AAB, preferentially from the genus Gluconobacter or Acetobacter. For example, Acetobacter orientalis, Gluconobacter oxydans, Gluconobacter japonicus, and Acetobacter lovaniensis. AAB can also be from other genera, such as Komagataeibacter, Novacetimonas and Gluconacetobacter.

[0102] 4. Lactic acid bacteria (LAB): One or more LAB, preferentially from the genus Liquorilactobacillus, Lacticaseibacillus, Lentilactobacillus, Leuconostoc and / or Oenococcus. For example, Lacticaseibacillus paracasei, Liquorilactobacillus satsumensis, Liquorilactobacillus ghanensis, Lentilactobacillus hilgardii and Leuconostoc pseudomesenteroides.

[0103] 5. From the family of Zymomonadaceae'. Typically it is Zymomonas mobilis.

[0104] 6. Other microbes: The starter cultures can optionally also include one or more species from the genus Bifidobacterium. For example, Bifidobacterium tibiigranuli.

[0105] The water kefir starter cultures will consist of one or more species from each of the groups numbered 1 to 5 above, with an optional addition from the group numbered 6. Materials and Methods

[0106] Isolation of novel Bifidobacterium species from water kefir

[0107] As part of a microbial diversity analysis of water kefir samples obtained from different countries, two novel species belonging to the genus Bifidobacterium were identified in metagenomic sequencing data, including the four strains under study. Plates of MRS (Oxoid, England, UK) supplemented with 0.05 g / L muropicin, 0.005 g / L amphotericin B, and 0.5g / L cysteine were used as selective media to isolate colonies belonging to the Bifidobacterium genus. The plates were incubated in anaerobic conditions using Anaerocult® A gas packs (Merk, Germany). Single colonies were randomly picked, and Bifidobacterium genus-specific PCRs were used to confirm that the isolates belonged to Bifidobacterium genus. Colonies belonging to the genus Bifidobacterium were subjected to 16S rRNA gene amplification and the PCR products were subjected to Sanger sequencing (Eurofins Scientific, Luxembourg). The type strain B. aquikefiri LMG 28769Twas purchased from the BCCM / LMG Bacteria Collection and analysed with the newly isolated strains from this study.

[0108] DNA extraction, genome sequencing, and genome assembly of novel Bifidobacterium species

[0109] Genomic DNA was extracted using the GenElute™ Bacterial Genomic DNA Kit (Sigma- Aldrich, USA). Four strains, representing the two novel species, were fully genome sequenced in a hybrid approach. Illumina sequencing was carried out according to manufacturer instructions on a NextSeq 2000 P1 chip with a read length of 2 x 150 bp. Illumina read quality control was performed using MetaWRAP (v1.3.2) with hg38 for host removal. Oxford Nanopore Technologies (ONT) Sequencing was carried out using a native barcoding kit and an R10.4.1 flow cell according to manufacturer instructions at 400 bp / s. SUP mode basecalling and adapter trimming were performed using Guppy (v6.3.8; https: / / community.nanoporetech.com). ONT read quality control was performed using Filtlong (v0.2.0; https: / / github.com / rrwick / Filtlong), removing reads less than 1000 bp long and 10% of the reads with the lowest quality. Draft genomes were assembled using Flye (v2.9), followed by long-read-polishing with Medaka (v1.7.2; https: / / github.com / nanoporetech / medaka). Short-read-polishing was performed using Polypolish (v0.5.0). Additional small plasmids were assembled using Unicycler (vO.4.7). Assembly quality was checked using CheckM (v1 .0.18) and CheckM2 (vO.1.3). Genome assemblies, including the G+C content, were assessed using QUAST (v5.1 .0). Plasmids were confirmed using Platon (v1.6). Genomes were annotated using PGAP (v 6.5). ANI scores were calculated using FastANI (v1.32) against all NCBI RefSeq and GenBank genomes of the genus Bifidobacterium (as of 2024-06-11). A phylogenetic tree was inferred for the novel species, with the GTDB-Tk curated set of 115 Bifidobacterium genomes derived from type strains and reference strains, and one outgroup, using the de novo workflow in GTDB-Tk (v2.1.1). The de novo workflow used 117 marker genes and WAG+GAMMA models to infer the tree. The CGE webserver (https: / / cqe.food.dtu.dk / sei~vices / ResFinder-EFSA / ) was used to scan the genomes for potential antimicrobial resistance genes. No resistances were predicted in any of the genomes.

[0110] 16S rRNA phylogenetic tree of novel Bifidobacterium species

[0111] BLASTN analysis was conducted with the nucleotide collection in GenBank (https: / / blast.ncbi.nlm.nih.gov / Blast.cgi) using prokaryotic databases of 16S rRNA genes obtained from all type strains to determine sequence similarities. The phylogenetic study used 101 species attributed to the genus Bifidobacterium from the List of Prokaryotic Names with Standing in Nomenclature (LPSN). First, multiple sequence alignments were performed using CLUSTAL W. Gap-containing regions were removed. Maximum likelihood phylogenetic trees were generated from the aligned, gap-free sequences of approximately 900 base pairs. A bootstrap analysis of 1000 replicates was performed using MEGA 11. In addition, the 'Find the best DNA / Protein (ML) models' function, based on the value of the Bayesian information criterion (BIC), was used to determine the most appropriate nucleotide substitution model for reconstructing the phylogenetic tree.

[0112] Phenotypic characteristics of novel Bifidobacterium species

[0113] Scanning electron microscopy (SEM) was used to evaluate the cell morphology (size and shape) of the strains under study. Fresh cultures were grown in MRS broth supplemented with 0.5 g / L cysteine (MRScys), pelleted at 4000 x g for 5 min, and washed in PBS, and prepared for SEM observation by performing an overnight fixation in 2.5% glutaraldehyde (Sigma-Aldrich, USA) in a 100 mM sodium cacodylate (pH 7.3; Sigma-Aldrich, USA) solution, followed by dehydration in a 10% step increase gradient of aqueous ethanol solutions, from 30% to 100% (v / v) over a period of 4 days. The ethanol was replaced with hexamethyldisilazane (HMDS; Sigma-Aldrich, USA) by immersing the samples in a 50:50% HMDS:Ethanol solution for one hour, followed by 100% HMDS for another hour. The HMDS was left to evaporate completely. Samples were fixed to the sample holders using carbon tape and sputter-coated with gold under vacuum, using an Emitech K575X sputter coater (Quorum Technologies, UK). ZeissSupra 40VP, equipped with a 3rd generation Gemini column (ZEISS, Germany), using the secondary electron detector, was used for SEM imaging. The resulting images were measured in Imaged (v1 .53) to assess the average cell size. Growth capacity was assessed at 4, 15, 21 , 30, 37 and 45°C, pH 3, 4, 5, 6, 7, 8, 9 and 10 and NaCI concentrations of 0%, 1%, 3%, 5%, 7% and 10% under anaerobic conditions using MRScys by measuring OD6oo after 72 hours of incubation. The inoculum for each test was set starting with an overnight culture grown in MRScys broth under anaerobic conditions at 30°C, then the overnight culture was adjusted to 0.5 OD6oo and inoculum operated at 1%. The maximum growth capacity (optimum) was respectively considered to be the maximum OD6oo achieved under all conditions tested. While the maximum growth rate was only assessed at 20°C, 30°C, 37°C and evaluated by measuring ODeoo every 2 hours for 72 hours. The efficiency of growth under aerobic and anaerobic conditions was tested for 72 hours a 30°C in MRScys broth under agitation to maximize oxygen exchange and in static mode using anaerobic jars under anaerobic conditions. Growth under different conditions was determined on the basis of optical density: good growth ODeoo > 0.3; weak growth 0.1 < ODeoo < 0.3; no growth ODeoo < 0.1. Additionally, the ability to generate bacterial colonies under aerobic conditions was tested. Oxidase activity assays were carried out using the Bactident Oxidase kit (Merck Millipore) and catalase activity was verified by putting a small amount of strain biomass into two drops of 3% H2O2 according to Reiner. Motility was assessed in accordance with the method described by Tittsler et al. (The Use of Semi-solid Agar for the Detection of Bacterial Motility. Journal of Bacteriology 31 , pp. 575-580 (1936)) with MRScys supplemented with 0.3% of agar. Fructose-6-phosphate phosphoketolase (F6PPK) was assessed using a colorimetric method. The ability to ferment different substrates was assessed through API 50 CHL (bioMerieux, France). For this, two McFarland of pure culture, resuspended in API 50 CHL medium flask, was added to each API 50 CHL strip. The results were read out after 72 hours. Enzymatic activities were evaluated using the API ZYM (bioMerieux, France) following the manufacturer's instructions. For this, four McFarland cell suspensions in 0.85% NaCI were inoculated into API ZYM strips. After four hours the results were read out.

[0114] Chemotaxonomic analysis of novel Bifidobacterium species

[0115] For MALDI-TOF mass spectrometry analysis, the ethanol / formic acid extraction procedure was used following Bruker's guidelines. Briefly, the novel strains and the reference strain Bifidobacterium aquikefiri LMG 28769T, were grown anaerobically in MRS agar (VWR, Belgium) supplemented with 0.05% cysteine-HCI at 30°C for 72 h. A single colony was resuspended in an Eppendorf with 300 pL sterile distilled water and 900 pL pure ethanol were added. The suspension was centrifuged at 21 ,036 x g for 2 min and the supernatant was discarded. To the dry pellet, 10 pL 70% formic acid and 10 pL acetonitrile were added and mixed well. The suspension was centrifuged again under the same conditions and 1 pL of the supernatant was spotted onto a target polished steel BC plate (Bruker, Germany). After air-drying, 1 pL of the a-cyano-4-hydroxycinnamic acid matrix solution (HCCA, Bruker), prepared following the manufacturer’s instructions in acetonitrile / water / trifluoroacetic acid (50:47.5:2.5 v / v), was added. Once dried at room temperature, the plate with each of the samples and a Bacterial Test Standard (BTS, Bruker) was read in a MALDI-TOF MS microflex LRF mass spectrometer using flexcontrol v. 3.4 and MBT Compass v. 4.1.90. MALDI Biotyper Compass Explorer v. 4.1 .90 was employed to compare the profiles of the novel species from that of the nearest neighbours. A Main Spectra Profile dendrogram was represented including 25 spectra profiles of Bifidobactarium included in the Bruker taxonomy database and the spectrum of the reference strain B. aquikefiri LMG 28769T. Finally, flexAnalysis v. 3.4 was used to represent the spectra profiles of B. aquikefiri LMG 28769T, WK012_4_13T, and WK041_4_12T.

[0116] Sample acquisition

[0117] Water kefir (WK) grains were sourced from WK producers. Individuals were asked to provide kefir grains either dried or fresh and double bagged. Information about fermentation parameters, such as previous fermentation durations, aerobic or anaerobic setups, and substrates used, was collected. The substrate for the fermentation / fermentation medium (aqueous solution) typically contained a sugar source (such as white sugar or brown sugar) and additional nutrient sources. The nutrient sources can be fruits or vegetables, fruit or vegetable extracts, yeast extracts or other nutritional extract, or food waste streams (such as apple pomace, brewer spent grains, fruit and vegetable peels and pulp, cereal bran, and the like). The sugar and nutrient source can be the same.

[0118] Water kefir media preparation

[0119] A sterile fig extract was prepared as described in Gulitz, A. et al. (Comparative phylobiomic analysis of the bacterial community of water kefir by 16S rRNA gene amplicon sequencing and ARDRA analysis. J Appl Microbiol, 2013. 114(4): p. 1082-91) with the following modifications. 500 g of dried figs were cut into small pieces and soaked in 1 L of tap water for 45 min while shaking at 100 rpm. Fig-water was poured through a sieve and further particles were removed by centrifugation at 17 OOOxg for 1 h. The fig extract was vacuum-filtered through a 0.2 pm membrane and stored at 4°C until use. A sucrose solution was prepared by adding 88 g / L sucrose to tap water and autoclaving at 121 °C for 15 min. 15 ml of sterile fig extract were added to 135 ml of 88 g / L sucrose solution (final concentrations: 80 g / L sucrose, 10% fig extract) and 9 g (60 g / L) of kefir grains, the inoculum, were added.

[0120] Water kefir fermentation

[0121] WK fermentations were carried out under aseptic, aerobic conditions in sterile glass bottles. Glass bottles were closed with screw caps containing 0.2 pm PTFE membranes and incubated at 21 °C for 48 h. After 48 h the ferment was poured through a sterile sieve and rinsed with autoclave tap water. Grains were re-used to start a new fermentation, as described above. As recommended previously in Gulitz, A. et al. (see above), a minimum of two pre-fermentations were carried out to allow the grains to recover from transport to the research centre, before the fermentations for sample collection were carried out.

[0122] The experiments were performed in two sequential biological replicates. The first biological replicate was grown for 48h and samples were taken at 8h and 48h. After 48h, the grains were rinsed with sterile tap water and reused to start the second replicate and samples were taken again after 8h and 48h. Grain samples for shotgun sequencing were only taken after the second 48h fermentation. If there was a sufficient grain growth after the first 48h fermentation, then excess grains were taken to test the freeze tolerance of WK grains (see Supplementary Methods for additional details).

[0123] Sample collection

[0124] 15 ml of WK were collected for DNA extraction. The WK liquid was centrifuged at 4 500xg for 30 min. The supernatant was discarded, the pellet flash frozen in liquid nitrogen and stored at -20°C until DNA extraction. Samples for VOC and NMR analysis were collected in 2 ml screw cap tubes, flash frozen in liquid nitrogen, and stored at -80°C until analysis. Kefir grains were taken after the WK was sieved and rinsed with sterile tap water. Kefir grains were flash frozen in liquid nitrogen and stored at -20°C until DNA extraction.

[0125] DNA extractions

[0126] DNA extractions were performed with the DNeasy PowerSoil Pro Kit (Qiagen, 47016), according to manufacturer instructions. DNA was extracted from the pellet of 15 ml WK liquid or ~100 mg kefir grains. For the DNA extraction from kefir grains, additional 1.5 mm Zirconium beads (Merck, Z763799) were added to the PowerBead Pro Tube. Library preparations and sequencing

[0127] Libraries for sequencing were prepared with the Nextera XT DNA Library Preparation Kit (Illumina, FC-131-1096) and Nextera XT Index Kit v2 (Illumina, FC-131-200X) according to manufacturer instructions. DNA quantity was checked using the Qubit dsDNA HS Assay Kit (Invitrogen, Q33231) and DNA integrity was checked using the Agilent High Sensitivity DNA Kit (Agilent, 5067-4626). Sequencing was carried out in house on the NextSeq500 using the 300-cycle High Output v2 kit.

[0128] Read QC and MAG assembly

[0129] MetaWRAP was used for read QC, host decontamination (hg38), and MAG assembly. Bacterial high quality MAGs were obtained with >90% completeness and >5% contamination by performing a co-assembly of the five metagenomics samples of each WK. CheckM steps were disabled in MetaWRAP to obtain fungal MAGs. Fungal MAGs were identified and quality checked using BUSCO (reference databases: ascomycota_odb10 & saccharomycetes_odb10). Bacterial MAGs were assigned to species using GTDB-TK. If bacterial MAGs could not be assigned to a species using GTDB-TK, then FastANI was used with a database containing NCBI GenBank and RefSeq genomes for the relevant genus. Fungal MAGs were assigned to species using the FastANI approach from above with a database.

[0130] Taxonomic profiling

[0131] Taxonomic profiling was done using inStrain with a WK database.

[0132] Metabolite analysis

[0133] WK samples were defrosted on a roller for 40 minutes at room temperature, and were then centrifuged at 7870 g for 5 minutes at 4°C. The supernatant was collected and filtered through washed 3 kDa Ultra centrifugal filters for 35 minutes at 14480 g (Sigma- Aldrich, Merck KGaA, Darmstadt, Germany). The filtrates were then frozen at -20°C until further analysis.

[0134] On the day of analysis, samples were defrosted and 540 pL was mixed with 10 pL sodium trimethylsilyl [2,2,3,3-2H4] proprionate (TSP) (0.05g / 4mL D2O) and 60 pL deuterium oxide (D2O). Spectra were acquired from a 600-MHz Varian NMR Spectrometer (Varian Limited, Oxford, United Kingdom), using the first increment of a nuclear Overhauser enhancement spectroscopy pulse sequence at 25°C. Spectra were acquired at 16,384 complex data points and 128 scans. Water suppression was achieved during the relaxation delay (2.5 s) and the mixing time (100 ms). All 1 H-NMR sample spectra were referenced to TSP at 0.0 parts per million (ppm) and processed manually with the Chenomx NMR Suite (version 7.7) by using a line broadening of 0.2 Hz, followed by phase and baseline correction. A total of 29 metabolites were identified and quantified based on the Chenomx 600-MHz Library and The Human Metabolome Database (HMBD).

[0135] Volatile organic compound analysis

[0136] 2 g of WK liquid, 0.6 g of sodium chloride and 20 pl of internal standard (4-methyl-2- pentanol at 50ppm) was added to a 20 ml screw capped SPME vial. The SPME fibre was exposed to the headspace above the samples for 30 min at depth of 1 cm at 40°C. Sample introduction was accomplished using a Gerstel MPS Autosampler.

[0137] A single 50 / 30 pm Carboxen™ / divinylbenzene / polydimethylsiloxane (DVB / CAR / PDMS) fibre was used. The SPME fibre was exposed to the headspace above the samples for 30 min at 40°C, then retracted and injected into the GC inlet and desorbed for 3 min at 250°C. Injections were made on a Shimadzu 2010 Plus GC with an Agilent DB-624 Ul (60 m x 0.32 mm x 1.8 pm) column using a split / splitless injector in a splitless mode. A merlin microseal was used as the septum. The temperature of the column oven was set at 40°C, held for 5 min, increased at 5°C / min to 230°C then increased at 15°C / min to 260°C, held for 5 min yielding at total GC run time of 65 min. The carrier gas was helium held at a constant flow of 1.2 ml / min. The detector was a Shimadzu TQ8030 mass spectrometer detector, ran in single quad mode. The ion source temperature was 220°C and the interface temperature was set at 260°C. The MS mode was electronic ionization (70v) with the mass range scanned between 35 and 250 amu. Compounds were identified using mass spectra comparisons to the NIST 2014 mass spectral library, a commercial flavour and fragrance library (FFNSC 2, Shimadzu Corporation, Japan) and an in-house library created using authentic compounds with target and qualifier ions and linear retention indices for each compound using Kovats index. Retention indices were matched against peer reviewed publications and authentic standards where possible to confirm compound identification. Spectral deconvolution was also performed to confirm identification of compounds using AMDIS. Batch processing of samples was carried out using MetaMS. MetaMS is an open-source pipeline for GC-MS-based untargeted metabolomics. An auto-tune of the GCMS was carried out prior to the analysis to ensure optimal GCMS performance. A set of external standards (1 -butanol, dimethyl disulfide, butyl acetate and cyclohexanone at 10ppm) was run at the start and end of the sample set and abundances were compared to known amounts to ensure that both the SPME extraction and MS detection was performing within specification.

[0138] Isolation of microbial strains

[0139] All culturing and sampling procedures were performed under aseptic conditions in a laminar flow cabinet. Microbial strains were isolated from water kefir fermentations as described above. Samples were stored at -80°C prior to isolation.

[0140] Isolation of Bifidobacterium spp.

[0141] Bifidobacterium strains were isolated and described above.

[0142] Isolation of acetic acid bacteria (AAB)

[0143] Acetobacter lovaniensis 66b_G1.1 (from WK066, Colombia), Acetobacter orientalis WKS59104 (from WK059, USA), Gluconobacter japonicus 29a_G1.1 (from WK029, USA), and Gluconobacter oxydans WKS40101 (from WK040, Canada) were isolated on ABS medium supplemented with amphotericin B (5 mg l_“1). Plates were incubated aerobically at 30 °C for 3 days.

[0144] Isolation of lactic acid bacteria (LAB)

[0145] Lacticaseibacillus paracasei 59a_L1 (from WK059, USA), Lentilactobacillus hilgardii WKS59211 (from WK059, USA), Leuconostoc pseudomesenteroides WKS34141 (from WK034, Canada), and Liquorilactobacillus satsumensis WKS59208 (from WK059, USA) were isolated on de Man-Rogosa-Sharpe (MRS) agar (Difco™, BD, New Jersey, USA) supplemented with amphotericin B (5 mg L“1). Plates were incubated anaerobically at 30 °C for 3 days.

[0146] Isolation of yeasts

[0147] Brettanomyces bruxellensis WKS03116 (from WK003, Ireland) was isolated on DBDM medium and incubated aerobically at 25 °C for 12 days. Colonies were re-streaked twice on DSMZ medium 85 (10 g L“1yeast extract, 20 g L“1peptone, 10 g L“1glucose, 10 g L“1CaCO3, 15 g L“1agar) supplemented with 25 mg l_“1chloramphenicol and 25 mg l_“1kanamycin to eliminate bacterial contamination. Plates were incubated aerobically at 25 °C for 4 days.

[0148] Pichia fermentans 12-1-01 (from WK012, Italy), Hanseniaspora valbyensis WKS34255 (from WK034, Canada), and Saccharomyces cerevisiae 66b_Gr_2G (from WK066, Colombia) were isolated on DSMZ medium 186 (3 g L“1yeast extract, 3 g L“1malt extract, 5 g l_“1peptone, 10 g L-1glucose, 15 g L-1agar) with 25 mg l_“1chloramphenicol and 25 mg L“1kanamycin. Plates were incubated aerobically at 25 °C for 3 days.

[0149] Isolation of Zymomonas mobilis

[0150] Zymomonas mobilis WKS10106 was isolated from WK010 on DSMZ medium 10 (10 g l_“1peptone, 10 g L“1yeast extract, 20 g L-1glucose, 15 g L-1agar) supplemented with amphotericin B (5 mg L“1). Plates were incubated anaerobically at 25 °C for 3 days.

[0151] Culturing of microbial strains

[0152] Bifidobacterium spp. were cultured anaerobically at 37 °C on modified MRS agar supplemented with 0.5 g L“1cysteine (mMRS). Plates were incubated for 6 days to ensure colony development. Liquid pre-cultures were prepared by inoculating 10 mL of mMRS broth with a single colony and incubated anaerobically at 37 °C for 3 days.

[0153] Acetobacter lovaniensis 66b_G1.1 , Acetobacter orientalis WKS59104, Gluconobacter japonicus 29a_G1.1 , and Gluconobacter oxydans WKS40101 were cultured aerobically at 30 °C on DSMZ medium 989 (5 g L“1peptone, 5 g L“1yeast extract, 5 g L“1glucose, 1 g L“1MgSO4x 7 H2O; 15 g L“1agar if solid medium required; pH 6.6-7.0). Plates were incubated for 3 days. Liquid pre-cultures were inoculated from single colonies into 10 mL of medium and incubated aerobically at 30 °C for 2 days.

[0154] Lacticaseibacillus paracasei 59a_L1 , Lentilactobacillus hilgardii WKS59211 , Leuconostoc pseudomesenteroides WKS34141 , and Liquorilactobacillus satsumensis WKS59208 were grown anaerobically at 37 °C on MRS agar (Difco™, BD, New Jersey, USA). Plates were incubated for 3 days, except Len. hilgardii WKS59211 , which required 6 days. Pre-cultures were prepared from single colonies in 10 mL of MRS broth and incubated anaerobically for 2 days.

[0155] Brettanomyces bruxellensis WKS03116, Pichia fermentans 12-1-01 , Hanseniaspora valbyensis WKS34255, and Saccharomyces cerevisiae 66b_Gr_2G were cultured on DSMZ medium 186. Plates were incubated aerobically at 25 °C for 3 days, except for B. bruxellensis, which was incubated for 6 days. Pre-cultures were inoculated from single colonies into 10 mL of broth and incubated for 2 days at 25 °C. Zymomonas mobilis WKS10106 was cultured anaerobically at 37 °C on DSMZ medium 10 (10 g L-1peptone, 10 g L-1yeast extract, 20 g L-1glucose, 15 g L-1agar if solid medium required). Plates were incubated for 6 days. Liquid pre-cultures were inoculated from single colonies in 10 mL of medium and incubated anaerobically at 37 °C for 3 days.

[0156] All main cultures were inoculated at a 1 :10 ratio with pre-cultures and incubated for 24 h under the respective cultivation conditions.

[0157] Molecular identification of isolates

[0158] All microbial isolates were re-streaked and cultured on the appropriate media at least twice to ensure clonal purity. Genomic DNA was extracted, and taxonomic identification was performed via PCR amplification of the 16S rRNA gene (for bacteria) or the internal transcribed spacer (ITS) region (for yeasts), followed by Sanger sequencing (Eurofins Scientific, Luxembourg).

[0159] For bacterial isolates, the primers CO1 (5'-AGTTTGATCCTGGCTCAG-3' - SEQ ID NO: 1) and CO2 (5'-TACCTTGTTACGACTT-3' - SEQ ID NO: 2) were used to amplify the 16S rRNA gene. For yeast isolates, the primers ITS1 F (5'-CTTGGTCATTTAGAGGAAGTAA- 3' - SEQ ID NO: 3) and ITS4 (5'-TCCTCCGCTTATTGATATGC-3' - SEQ ID NO: 4) were used for ITS region amplification. PCR products were sequenced via Sanger sequencing, and resulting sequences were used for species-level identification via BLAST analysis.

[0160] Whole-genome sequencing of isolates

[0161] Whole-genome sequencing was performed using Oxford Nanopore Technologies (ONT) with the Native Barcoding Kit and an R10.4.1 flow cell, following the manufacturer’s instructions. Basecalling was conducted in super accuracy (SUP) mode using Guppy (v6.3.8; htps: / / community.nanoporetech.com), which also handled adapter trimming. Read quality control was performed using Filtlong (v0.2.0; https: / / qithub.com / rrwick / Filtlonq), discarding reads shorter than 1 ,000 bp and removing 10% the lowest-quality reads. Draft genome assemblies were generated using Flye (v2.9) and polished with Medaka (v1.7.2; https: / / github.com / nanoporetech / medaka). Raw sequence reads and assembled genomes are available under the BioProject number PRJNA1295428. Antimicrobial resistance gene screening

[0162] Antimicrobial resistance (AMR) genes were screened using ResFinder, with the EFSA_2023 database. No AMR genes were detected in any of the assembled genomes.

[0163] Pitched fermentations

[0164] A total of 256 pitched fermentation combinations were performed in four experimental sets. Each set included all four yeasts, lactic acid bacteria (LAB), acetic acid bacteria (AAB), and one Bifidobacterium species, with or without Zymomonas mobilis (64 fermentations per set). Additional controls included uninoculated media and fermentations using water kefir grains at 15, 30, 60, and 120 g L“1.

[0165] Fresh microbial cultures were harvested by centrifugation at 4,500 x g for 15 min. Supernatants were discarded, and pellets were resuspended in water kefir medium prepared as described above. Cell concentrations were determined using a BD Accuri C6 flow cytometer (BD, New Jersey, USA) with the LIVE / DEAD™ BacLight™ Bacterial Viability Kit (L7012; Thermo Fisher Scientific, Massachusetts, USA) according to the manufacturer’s instructions.

[0166] Fermentations were carried out in 50 mL Falcon tubes, loosely capped, containing 20 mL of water kefir medium. Strains were inoculated at 5 x 10scells mL-1each in four-strain combinations, or 4 x 105cells mL-1each in five-strain combinations, maintaining a total inoculum of 2 x 106cells mL-1. Unless stated otherwise, fermentations were incubated at 22 °C for 72 h.

[0167] Samples were collected directly after setting up the fermentations (0 h) and 72 h for downstream analyses. pH measurements were performed on 1 mL aliquots, and flow cytometry was conducted on 12.5 pL samples. Additional aliquots for sugar, organic acid, and ethanol analysis by HPLC were taken on day 3 and stored at -80 °C until analysis.

[0168] Aroma evaluations

[0169] Sensory evaluations of aroma were performed on day 4 offermentation. Blinded samples were randomized and presented to volunteers (n = 9-15), who were asked to rate the smell of each sample on a 5-point hedonic scale (1 = worst, 5 = best). Jars containing coffee beans were provided to enable olfactory neutralization between samples. Modified water kefir media (mWK media)

[0170] To enable larger-scale production, the standard water kefir medium was modified. For each litre of mWK medium, 80 g of sucrose and 10 g of dried figs (approximately one fig per litre) were added to 1 L of tap water. The medium was autoclaved at 121°C for 15 min before use.

[0171] Pressure experiment

[0172] Pitched fermentations for pressure evaluation were prepared as described above, with the following modifications. One-litre bottles containing 650 mL of mWK medium were inoculated with each strain at 5 x 105cells mL-1and capped with lids fitted with 0.2 pm membranes. Fermentations were conducted aerobically at 23°C for either 1 or 4 days prior to bottling.

[0173] Following the aerobic incubation, cultures were transferred to 750 mL swing-top bottles equipped with manometers (MMBM, Matt Mill, Germany) and sealed. Pressure (in psi) was recorded at each specified time point.

[0174] Fermentations for sensory evaluation

[0175] Two-litre fermentations using mWK medium were inoculated as described above, capped with lids containing 0.2 pm membranes, and incubated at 23°C with shaking at 100 rpm. An exception was made for the combination “ B.fer P.fer L.hil A. lov Z.mob" which acidified the medium more rapidly without agitation and was therefore incubated at 23°C without shaking. All fermentations were incubated for 96 h prior to sampling, bottling, and cold storage at 3-4°C.

[0176] To ensure microbiological safety, all pitched water kefir samples intended for sensory evaluation were screened for potential foodborne pathogens, including coagulasepositive Staphylococcus, coliforms, Listeria spp., and Salmonella spp., by Eurofins (Dungarvan, Co. Waterford, Ireland). No pathogens were detected in any of the tested samples.

[0177] Sensory Affective Evaluation

[0178] Twenty-five naive assessors (17 female, 8 female) were recruited in University College Cork, Ireland. Sensory acceptance testing was conducted using these untrained assessors. Age range of assessors was 23-59 years old. Selection criteria for these naive assessors were availability and motivation to participate on all days of the experiment and that they were kefir consumers. Assessors used the sensory Hedonic descriptors in Table 1 for six different WK.

[0179] Table 1 Sensory terms for the affective and descriptive evaluation of kefir.

[0180] Sensory analysis was carried out in panel booths conforming to international standards (ISO 8589: 2007). All samples were stored at 4°C until required, before presentation to the naive assessor panel at 4°C and coded with a randomly selected 3-digit code. Each assessor was provided with deionised water and instructed to cleanse their palates between tastings. Additionally, each assessor was asked to indicate their degree of liking on a 10 cm line scale ranging from 0 (extremely dislike) at the left to 10 (extremely like) at the right and rating subsequently scored in cm from left. The order of the presentation of all test samples was randomized to prevent first order and carryover effects. Samples were presented under white light (1000 LUX) in replicate.

[0181] Optimized descriptive profiling (ODP)

[0182] Optimized descriptive profiling (ODP) is a rapid method for obtaining sensory descriptions utilizing semi-trained judges that has the potential to quantitatively evaluate sensory attributes. These above assessors were presented with all samples simultaneously but with randomised order to prevent first order and carry-over effects. Assessors used the consensus list of sensory descriptors which were measured on a 10 cm line scale with the term “none” used as the anchor point for the 0 end of the scale and “extreme” for the 10 end of the scale (Table 1). For this study training, a consensus sensory lexicon was used that is well-known in the art. Sensory terms, which were the main sensory dimensions, were pre-selected from the sample set using an expert sensory panel (n = 10). Assessors evaluated the intensity of each attribute for each sample on the scales. All samples were prepared in the same manner as the consumer analysis study and presented in replicate.

[0183] Metagenomic sequencing and analysis of grain-based and pitched WK fermentations Metagenomic DNA was extracted from 1.8 mL aliquots of liquid fermentation samples that had been stored at -80°C. Samples were thawed, and cells were pelleted by centrifugation at 15,000 x g for 2 min. The supernatant was discarded, and the pellet was resuspended in the initial lysis solution provided with the DNeasy PowerSoil Pro Kit (QIAGEN, Germany). DNA extraction was performed according to the manufacturer’s instructions.

[0184] DNA was diluted to 0.2ng / ml. Sequencing libraries were then prepared using a miniaturised version of the Nextera XT assay (Illumina). In brief, a Labcyte Echo 525 (Beckman Coulter) was used to combine 0.5 ml diluted DNA, 1 ml TD buffer and 0.5 ml ATM in individual wells of 96 well plates. The plates were centrifuged at 280 x g for 1 min and incubated at 55°C for 5 min to “tagment” the DNA. The plate was then centrifuged again and the Labcyte Echo 525 was used to transfer 0.5 ml NT buffer to each well. The plates were centrifuged at 280 x g for 1 min and then incubated at room temperature for 5 min. The Labcyte Echo was then used to add NPM (1.5 ml) and 1 ml UDIs (supplied by Integrated DNA Technologies Ltd) and amplification of the libraries was performed as described in the Nextera XT library preparation guide (Illumina) with 14 cycles. Following amplification, 15 ml of molecular grade water was added to each well manually and the plates were centrifuged at 280 x g for 1 min. Samples were then cleaned using a 0.8X ratio of Ampure beads (Beckman Coulter) using a Biomek i7 automated liquid handler (Beckman Coulter). Following clean up, samples were quantified using the Qubit dsDNA High Sensitivity Assay (Life Technologies) and the average size of the libraries was assessed using the Agilent Bioanalyser and the High Sensitivity DNA assay (Agilent). Samples were then pooled equimolarly and the final pool was subjected to a further clean up with a 0.6 x ratio of Ampure beads to remove excess UDIs. The cleaned library was quantified by Qubit dsDNA High Sensitivity Assay and sequenced at the Teagasc Sequencing Centre on a NextSeq™ 1000 / 2000 with a P2 XLEAP-SBS™ Reagent Kit (300 Cycles) as per manufacturer’s guidelines.

[0185] Raw Illumina reads were quality filtered and processed using KneadData v0.12.0 with the hg37dec_v0.1 human genome as the host removal reference. Trimming was performed using Trimmomatic v0.39. Only paired-end reads were retained for downstream analysis. InStrain was used for taxonomic profiling (Olm, M.R., Crits- Christoph, A., Bouma-Gregson, K. et al. inStrain profiles population microdiversity from metagenomic data and sensitively detects shared microbial strains. Nat Biotechnol 39, 727-736 (2021)). Metagenomic sequencing data has been made available under BioProject number PRJNA1295435.

[0186] Metabolite analysis of pitched fermentations

[0187] Sugars and Ethanol

[0188] The samples were diluted 1 :100 and filtered using 0.22pm syringe filters (CH2213-NN, Thermo Scientific). Quantification of Sucrose, Glucose, Fructose, Glycerol, Ethanol, and Mannitol was performed by HPLC using a Waters Alliance Separations module e2695 coupled to a Waters 2414 refractive index (Rl) detector (Waters, Milford MA, USA) at 35°C. Samples or standards at a volume of 20pl were injected on to an Aminex HPX- 87P column (300 x 7.8 mm) coupled with Aminex Micro-Guard Carbo-P guard column (30 x 4.6 mm) operated at 60°C. The samples were eluted with miliQ water (18.2MQ) at a flow rate of 0.6mL min-1. Detection and quantification of each component in samples was performed with respect to a standard curve. Analytical grade standards at the concentrations 20, 50, 100, 200 pg ml-1were used to create the standard curve for each component. Organic acids

[0189] The samples were diluted 1 :100 and filtered using 0.22pm syringe filters (CH2213-NN, Thermo Scientific). Quantification of lactic acid and acetic acid was performed by HPLC using a Waters Alliance Separations module e2695 coupled to a Waters 2414 refractive index (Rl) detector (Waters, Milford MA, USA) at 35°C. Samples or standards at a volume of 20pl were injected on to an Rezex™ ROA-Organic Acid H+ (8%) column (300 x 7.8 mm) operated at 60 °C. The samples were eluted with 0.005N H2SO4(HPLC grade) at a flow rate of 0.5mL min-1. Detection and quantification of organic acids in samples was performed with respect to a standard curve. Analytical grade organic acid standards at the concentrations 20, 50, 100, 200pg ml-1were used to create the standard curve for each component.

[0190] Free amino acid analysis

[0191] Free amino acids were analysed. In brief, proteins were oxidised with performic acid containing phenol prior to microwave hydrolysis in 6M HCI at 160°C for 15 minutes. The reaction was neutralised by adjusting to pH 2.2 with 7.5N NaOH. Samples were then diluted 1 in 2 with norleucine and analysed as per the free amino acids method.

[0192] Samples were deproteinised by mixing equal volumes of 24% (w / v) trichloro-acetic acid (TCA) and sample, these were allowed to stand for 10 minutes before centrifuging at 14400 x g (Microcentaur, MSE, UK) for 10 minutes. Supernatants were removed and diluted with 0.2 M sodium citrate buffer, pH 2.2 to give approximately 250 nmol of each amino acid residue. Samples were then diluted 1 in 2 with the internal standard, norleucine, to give a final concentration of 125 nm mL-1. Amino acids were quantified using a Jeol JLC-500 / V amino acid analyser (Jeol (UK) Ltd., Garden city, Herts, UK) fitted with a Jeol Na+high performance cation exchange column.

[0193] In the JLC-500 / V amino acid analyser, specified amounts of sample, set in vials, are introduced into a column packed with cation exchange resin and by flowing buffers at different pHs through the column, each amino acid component of the sample is separated and eluted from the column.

[0194] The ninhydrin reagent is then added to the eluent, and when heated, it reacts with amino acids and creates pigments, which are measured by the flow colorimetric method. The signals detected here are sent to the data processing unit to analyse the amino acids qualitatively and quantitatively. BioLector XT Cultivation Conditions

[0195] Experiments were conducted using the BioLectorXT microbioreactor system (Beckman Coulter) in 48-well FlowerPlates (M2P-MTP-48-BOH3) with a working volume of 1 mL per well. Cultures were incubated at 23 °C with continuous orbital shaking at 800 rpm, providing effective mixing and gas-liquid transfer. Aeration was set to 50 mL / min via the system’s gassing lid, using humidified air to minimise evaporation. Biomass growth was monitored via backscatter at 620 nm. Dissolved oxygen was tracked using the DO (RF) channel, while pH was monitored via the pH (pH51) channel. All parameters were recorded every 12.8 minutes using the BioLection software. No active pH or feed control was applied during the experiment.

[0196] Statistical analysis

[0197] All data processing, statistical analysis and plotting was done in R. Relative abundance was calculated as taxonomic abundance, based on the relative coverage of a species within the sample. Plots were created using ggplot2 and pheatmap. Hierarchical clustering is done with default clustering parameters ("complete"), with the exception of correlation analysis plots, in which “ward.D” was used. Boxplots show the first and third quartiles, median, and whiskers indicate 1.5 times the interquartile range (IQR). Alpha diversity measures were calculated as per Sun, Z., et al. (Challenges in benchmarking metagenomic profilers. Nature Methods, 2021. 18(6): p. 618-626). PCoA plots were created using the Vegan package. Non-parametric tests were used due to non-normal data distributions. Differences across fermentation combinations were assessed using the Kruskal-Wallis test, followed by pairwise Wilcoxon rank-sum tests for post hoc comparisons. P-values in Spearman correlations were Benjamini Hochberg adjusted. P- values of <0.05, <0.01 , and <0.001 are flagged with one, two, and three stars (*, **, and ***), respectively. Species co-occurrence was calculated in R using the package cooccur. Multiple Factor Analysis (MFA) was calculated using the package FactorMineR. Compact letter displays (CLDs) were generated from adjusted p-values using the multcompLetters function. Bar plots depict the mean with standard error of the mean (SEM). Statistical significance in plots was indicated using CLD lettering.

[0198] Results

[0199] The Applicants investigated 69 water kefir grains from 21 different countries with shotgun metagenomics. MAG reconstruction from the shotgun metagenomics data suggested the presence of two candidate novel Bifidobacterium species. The first candidate new species (including the later isolate WK012_4_13) was predicted to be present in 12 water kefirs and the second species (the later isolate WK041_4_12T) was predicted to be present in only one water kefir. Four water kefirs were chosen and used to successfully isolate the predicted new species, resulting in four strains belonging to the two candidate new species.

[0200] General fermentation characteristics

[0201] The kefir grains received were passed through a minimum of two initial or ‘pre-’ fermentations, with the intention to allow microbial communities to re-establish after shipping, before samples were taken for analysis. The starting pH of the WK media was 6.6 and the control media, to which no grains were added, slightly increased during the 48h of incubation. In contrast, the pH decreased to an average pH of 5.3 after 8h of incubation with inoculum and further decreased to an average of 3.6 after 48h of fermentation.

[0202] The kefir grains that exhibited the strongest grain growth increased in mass by up to 236% during 48h of fermentation. Under the same conditions a loss in grain mass was observed for other grains. Grains that were received in fresh form showed a significantly stronger grain growth than grains that were dried before shipment. On average an increase of 50% grain mass was observed after 48h of fermentation.

[0203] Taxonomic Profiling

[0204] Taxonomic profiling was done using inStrain with a WK-specific database, which was used to identify the species present across the WK pangenome. To consider a species a true positive hit, strict cut off with a breadth of at least 0.35, as well a minimum ratio of 0.75 for the expected breadth to observed breadth, was required. No species were detected in the control media samples with these requirements. Ultimately, an average of 10.3 species were detected per WK sample. A total of 96 species, from 28 genera, were detected using inStrain and 76 of these species were supported with at least one MAG.

[0205] The six most commonly detected species, prevalent in more than 50% of the 69 WKs were Saccharomyces cerevisiae (85.5% prevalence), Lacticaseibacillus paracasei (75.4%), Liquorilactobacillus satsumensis (71.0%), Lentilactobacillus hilgardii (65.2%), Acetobacter orientalis (53.6%), and Liquorilactobacillus nagelii (50.7%). The most prevalent genera were Saccharomyces (85.5%), Gluconobacter (84.1%), Liquorilactobacillus (84.1%), Acetobacter (82.6%), Lacticaseibacillus (75.4%), Lentilactobacillus (69.6%), and Bifidobacterium (55.1%), being prevalent in more than 50% of the 69 WKs tested. Leuconostoc (43.5% prevalence), Zygotorulaspora (40.6%), Pichia (39.1%), Brettanomyces (36.2%), Oenococcus (34.8%), and Zymomonas (33.3%) were also commonly detected in WK.

[0206] Water kefir community states

[0207] After 48h of fermentation, hierarchical clustering revealed six WK community states (CS), with CS clustering primarily by the most abundant genera. States were dominated by Liquorilactobacillus (CS 1), Gluconobacter (CS 2), Zymomonas (CS 4), Leuconostoc (CS 5), and Acetobacter (CS 6), or were relatively balanced communities without a clearly dominant genus. Different CS were supported by genus-level beta diversity analysis, with a higher ANOSIM statistic R, indicating a higher dissimilarity between the CS, at genuslevel beta diversity analysis (ANOSIM statistic R: 0.6967) compared to the species-level beta diversity analysis (ANOSIM statistic R: 0.3807). Zymomonas dominated WKs show a separate CS on genus- and species-level beta diversity analysis.

[0208] Genomic features and phylogeny of novel Bifidobacterium species

[0209] The pairwise similarity analysis of the 16S rRNA gene, performed with blast, revealed that the greatest sequence similarities were with Bifidobacterium aquikefiri LMG 28769Twith a similarity percentage ranging from 98.82% to 98.89%. All strains were subjected to the computation of the overall genome correlation index (OGRI). The phylogenetic tree, based on the 16S rRNA gene (see Figure 1), revealed that all strains formed a cluster together with Bifidobacterium aquikefiri LMG 28769T. Notably, strains WK048_4_13, WK013_4_14, and WK012_4_13Tformed a distinct cluster, confirming their membership in the same species, while WK041_4_12 appeared to be more closely related to Bifidobacterium aquikefiri LMG 28769T. The high bootstrap values, exceeding 90, indicate a high level of confidence in the phylogenetic branches that encompass the studied strains in a resampled dataset. Based on the phylogenetic analyses, Bifidobacterium aquikefiri LMG 28769Twas chosen as the closely related species for further phenotypic comparisons.

[0210] The hybrid sequencing approach, with coverage ranging from 162x - 441x for Illumina reads and 41 Ox - 976x for ONT reads, allowed the assembly of complete, circular genomes for all strains. Complete and circular plasmids were detected in WK012_4_13 and WK048_4_13, but not in WK013_4_14 and WK041_4_12. The genome sizes (excluding plasmids) are 2.43, 2.58, 2.65, and 2.36 Mbp for WK012_4_13T, WK013_4_14, WK048_4_13, and WK041_4_12T, respectively. The G+C content ranges between 55.82 and 56.00 mol% for WK012_4_13, WK013_4_14, and WK048_4_13 and is 53.94 mol% for WK041_4_12, and therefore slightly higher than in B. aquikefiri (52.0 mol%). The genome assemblies were used for OGRI calculation. WK012_4_13 shares the highest ANI scores with B. aquikefiri (GCF_002259795.1 ; ANI = 81.46%), B. crudilactis (GCFJ322654655.1 ; ANI = 78.27%), and B. psychraerophilum (GCF_002813205.1 ; ANI = 78.11%). WK013_4_14 shares the highest ANI scores with B. psychraerophilum (GCFJ302813205.1 ; ANI = 82.31%), B. aquikefiri (GCF_002259795.1 ; ANI = 80.95%), and B. crudilactis (GCF_022654655.1 ; ANI = 78.89%). WK048_4_13 shares the highest ANI scores with B. psychraerophilum (GCF_002813205.1 ; ANI = 82.53%), B. aquikefiri (GCF_002259795.1 ; ANI = 81.77%), and B. crudilactis (GCF_000738005.1 ; ANI = 79.07%). WK041_4_12Tshares the highest ANI scores with B. aquikefiri (GCFJ302259795.1 ; ANI = 84.84%), B. psychraerophilum (GCF_002813205.1 ; ANI = 81.72%), and B. crudilactis (GCA_022649705.1 ; ANI = 80.43%). WK012_4_13, WK013_4_14, and WK048_4_13 share ANI scores >99.9%, which classifies them as the same species. WK041_4_12 shares ANI scores of 81 .69 to 80.50% with WK012_4_13, WK013_4_14, and WK048_4_13, classifying it as a separate species. The dDDH scores (see Table 2) also agree with WK012_4_13, WK013_4_14, WK048_4_13, and WK041_4_12 being classified as two novel species, as the dDDH scores are between 23.9 and 38.5%, well below the recommended < 70% dDDH for a novel species. Additional genome features are given in Table 2.

[0211] Table 2 Genomic features summary table of strains belonging of Bifidobacterium aquikefiriaerophilum, Bifidobacterium eccum and the type strain of the closely related species within the Bifidobacterium genus. Strains: 1, WK048_4_13; 2, WK013_4_14; 3, WK012_4_13T; 4, WK041_4_12T; 5, B. aquikefiri LMG 28769T

[0212] The maximum likelihood tree based on the GBTD-Tk tool showed a unique cluster composed of the four strains under study (see Figure 2). Marker based phylogenetics places WK041_4_12 closest to B. aquikefiri (now confirmed as Bifidobacterium aquikefiricola) and WK012_4_13 (now confirmed as Bifidobacterium fermentum), WK013_4_14, and WK048_4_13 on the same subclade. The sub-clusters confirmed the distribution already seen forthe phylogenetic tree obtained with the 16S rRNAgene. This result confirmed the choice of B. aquikefiri as the closely related species for the two new candidate species.

[0213] Phenotypic features of novel Bifidobacterium species

[0214] SEM image analysis revealed that the cells of the strains WK048_4_13, WK013_4_14, and WK012_4_13 had a club-shaped and bacillary shape with an average length of 1.23±0.17 m, 1.38±0.25 pm, and 1.05±0.18 pm and width of 0.38±0.04 pm, 0.40±0.06 pm, and 0.33±0.05 pm, respectively. Strain WK041_4_12Thad a square rod shape with an average length of 1.18±0.23 pm and an average width of 0.48±0.08 pm (see Figure 3). All strains were Gram-positive. The colonies were white and spherical after 48 hours of incubation in MRScys agar. The type strain of B. aquikefiri LMG 28769Twas chosen for comparison in the phenotypic tests. Strains WK048_4_13 and WK012_4_13 were able to grow between 20 and 37 °C, while WK013_4_14 showed weak growth at 15 °C. WK041_4_12 was able to grow between 15 and 37 °C, while B. aquikefiri LMG 28769Talso showed weak growth at 4 °C and 45°. The growth optimum based on the maximum growth capacity was 30 °C for all strains studied. Considering the maximum growth rate, strain WK041_4_12 showed the maximum growth rate at 37 °C. All strains grew in a pH range between 4 and 8, except for strain WK048_4_13, which was able to grow weakly at pH 9. All strains had a higher maximum growth capacity at pH 6 (see Table 3). The growth in the presence of increasing NaCI concentrations showed that the strains WK048_4_13, WK041_4_12, and WK012_4_13 grow in a range of 0-3% while WK013_4_14 and B. aquikefiri LMG 28769Tshowed a growth range of 0-1%. Strains tested were able to grow well under aerobic conditions in MRScys agar and broth, including the reference strain B. aquikefiri LMG 28769T, except strain WK012_4_13, which showed weak growth. Species of the genus Bifidobacterium, isolated from water kefir, demonstrated the ability to tolerate aerobic environments, such as the species B. aquikefiri, Bifidobacterium tibigranuli, and Bifidobacterium psychraerophilum. The backslopping process of water kefir grains creates aerobic conditions, making it essential for the colonization of water kefir grains to be more resistant to oxidative stress.

[0215] All strains tested negative in catalase, oxidase, and motility tests. Fructose-6-phosphate phosphoketolase (F6PPK) was tested as a taxonomic marker commonly used to identify the genus Bifidobacterium. All strains showed F6PPK activity. Based on the API 50 CHL strip, it was possible to identify the ability to produce acid with different carbohydrates and derivatives. All strains, including the reference strain B. aquikefiri LMG 28769Twere able to produce acid from D-ribose, D-fructose, esculin ferric citrate, D-maltose, D- melibiose and D-raffinose, but not from glycerol, erythritol, L-xylose, D-adonitol, methyl- B-D-xylopyranoside, L-sorbose, L-ramnose, inositol, D-mannitol, D-sorbitol, methyl-a-D- mannopyranoside, D-trehalose, inulin, D-melezitose, starch, glycogen, xylitol, D- tagatose, D-fucose, L-fucose, D-arabitol, L-arabitol, potassium 2-ketoglucose, and D- lactose. Table 3: Differential characteristics of strains S. aquikefiriaerophilum WK012_4_13, WK013_4_14, WK048_4_13, B eccum WK041_4_1C and the type strain B. aquikefiri LMG 28769, the closest related species within the Bifidobacterium genus. Strains: 1, WK012_4_13; 2, WK013_4_14; 3, WK048_4_13; 4, WK041_4_12; 5, fl. aquikefiri LMG 28769 (control). Data were generated in this study, unless otherwise stated. Symbols: +, positive; -, negative; w, weak positive reaction; nd, not determined

[0216] a: weak growth ; b: good growth ; *: Resulting optimum based on maximum growth rate

[0217] Table 3 shows the distinguishing characteristics between strains WK048_4_13, WK013_4_14, WK012_4_13, WK041_4_12, and B. aquikefiri LMG 28769T. In particular, the strain WK041_4_12Tproduced acid from Dulcitol, but not from D-arabinose, D- xylose, D-galactose, D-mannose, D-Saccharose, potassium gluconate and D4yxose compared to strains WK048_4_13, WK013_4_14, and WK012_4_13.

[0218] The carbohydrate fermentation profile observed in strain WK013_4_14 shows the ability to produce acid, unlike the other strains, using Amygdalin, Arbutin, Salicin and D- Cellobiose. The latter difference in carbohydrate fermentation in strain WK013_4_14 suggested a strain-specific ability to ferment carbohydrates and derivatives.

[0219] In addition, the API ZYM system was used to test the enzymatic production of the strains. Strain WK041_4_12Texclusively produced N-acetyl-B-glucosaminidase and weakly valine arylamidase. In contrast, it did not produce a-galactosidase and p-galactosidase.

[0220] Chemotaxonomic analysis of novel Bifidobacterium species

[0221] The two mass spectra profiles from the candidate novel species were compared with that of B. aquikefiri LMG 28769T(see Figure 7). In addition, they were evaluated with spectra from other 25 known Bifidobacterium spp. strains included in the Bruker Taxonomy database (see Figure 4). The proposed novel species did not match to any of the species found in the database, suggesting possible new species. In the MSP dendrogram, B. aquikefiri LMG 28769Tis the closest species based on the distance matrix to WK012_4_13 and WK041_4_12, especially to the latter one. Their clustering at a distance level of >500 underlines the possibility of being novel species. MALDI-TOF results are thus in agreement with results obtained in the 16S rRNA and GTDB-TK phylogenetic trees.

[0222] Metabolite analysis

[0223] A selection of 15 WK fermentations, representing a range of microbial compositions, were further investigated by NMR metabolomic and GC-MS volatile organic compound (VOC) analysis at 8h and 48h. Metabolomics detected 29 compounds in the WK liquid, with the majority of samples clearly separating on the basis of whether they were collected at 8h or 48h. There was an overall pattern of decreased sugars throughout the fermentation, while alcohols and organic acids increased during the fermentation. Out of the 29 metabolites detected, 22 differed significantly throughout the fermentation. Acetoin, citrate, ethanol, fructose, isoleucine, lactate, methylamine, propylene glycol, pyruvate, pyruvatoxime, and valerate accumulated during the fermentation while choline, formate, fumarate, 4-aminobutyrate (GABA), malate, proline, succinate, and sucrose decreased. Acetate was elevated at 8h, but then decreased again after 48h of fermentation). Alanine and leucine concentrations decreased during the first 8h of the fermentation and then increases again towards 48h of fermentation, to not significant levels compared the media controls. No significant differences between the media control, the 8h and 48h fermentations was detected for glucose, glutamine, glycylproline, homocitrulline, malonate, methanol, and valine. It should be noted though that some WKs showed the production or degradation of some of these compounds while not being overall statistically different. A. orientalis showed significant positive correlations with isoleucine and leucine. Gluconobacter oxydans showed positive correlations with pyruvate and Lach, fermentatl showed positive correlations with homocitrulline. Li. satsumensis and Lact. paracasei showed nearly significant correlations with lactate (adj. p-value 0.064 and 0.094, respectively). None of the negative correlations were statistically significant after p-value adjustment.

[0224] Flavour analysis

[0225] Volatile organic compound (VOC) analysis, i.e., flavour analysis, detected 84 VOCs from WK. Media samples clustered separately from the WK samples and a grouping of the WK samples by time point was discernible. At 8h of fermentation volatile alcohols increased most, while volatile acids and esters increased by 48h of fermentation. A Zym. mobilis dominated WK (WK042) clustered separately from other WK flavour profiles. Out of the 84 detected VOCs, 48 VOCs showed a significant difference during the fermentations, while 25 VOCs did not significantly change throughout the fermentation. Eleven compounds were not statistically tested, as two of the conditions did not contain detectable quantities of the VOC. Eight of the eleven not statistically tested compounds were present only in the 48h fermentations and three were only detected in the 8h fermentations.

[0226] Correlation analysis showed 40 significant positive and 9 significant negative correlations between VOCs and species that were detected through shotgun metagenomics. Br. bruxellensis and Brettanomyces sp. HC-2020a formed a cluster with strong positive correlations with several esters, such as ethyl 2-methylbutanoatec, and ethyl tetradecanoate, isobutyl octanoate (Br. sp. HC-2020a only), methyl dodecanoate (Br. bruxellensis only), methyl decanoate (Br. bruxellensis only), ethyl dodecanoate (Br. bruxellensis only;. Many of these esters are described as fruity and sweet. Pi. fermentans is the only species that showed significant negative correlations with this cluster of esters. Overall, Pi. fermentans contributes to the development of the VOC profile of WK in a particular way, as it showed eight significant negative and six significant positive correlations with VOC flavour compounds. Pi. fermentans had positive correlations with acetone, 2-methylfuran, ethyl ether, 2-butanone, 2-methylpropanal, and hexane. Sa. cerevisiae showed positive correlations with styrene and a negative correlation with acetone. Zyg. florentina showed strong positive correlations with 2-nonanone. Zym. mobilis showed significant positive correlations with methyl dodecanoate and methyl isobutyl ketone. Multi-factor analysis supports these trends. Positive correlations between volatile esters and A. indonesiensis, A. orientalis, A. WK045_bin.3, Bi. WK041_bin.7, Lach, fermentati, Pseudoclavibacter A WK041_bin.2 and Zym. mobilis were observed as well.

[0227] Description of Bifidobacterium sp. nov. (WK012 4 13) (now assigned its official name of Bifidobacterium fermentum)

[0228] Cells are Gram-positive-staining, non-motile, not spore-forming, F6PPK-positive, catalase and oxidase negative, bacillary club shape with an average length of 1.23±0.17 pm. The colonies are white and spherical after 48 hours of incubation in MRScys. A good growth occurs under anaerobic, in the temperature interval 20-37°C, in the pH range 4- 8 and in NaCI range 0-1%. Optimal conditions for growth occur at 30°C and pH 6. Strong growth occurs under anaerobic conditions for all three strains analysed and aerophilic conditions for two of the three strains analysed. WK012_4_13Tshowed weak growth under aerobic conditions. Fermentation occurred from D-ribose, D-fructose, esculin ferric citrate, D-maltose, D-melibiose, D-raffinose, D-arabinose, D-xylose, D-mannose, D- saccharose and Methyl-a-D-glucopyranoside, but not from glycerol, erythritol, L-xylose, D-adonitol, methyl-B-D-xylopyranoside, L-sorbose, L-rhamnose, inositol, D-mannitol, D- sorbitol, methyl-a-D-mannopyranoside, D-trehalose, inulin, D-melezitose, starch, glycogen, xylitol, D-tagatose, D-fucose, L-fucose, D-arabitol, L-arabitol, potassium 2- ketogluconate, D-lactose, L-arabinose, dulcitol, N-acetylglucosamine, potassium 5- ketogluconate and D-turanose. A weak fermentation occurs from D-galactose, potassium gluconate, D-glucose, gentiobiose and D-lyxose. Amygdalin, arbutin, salicin and D- cellobiose provided positive results for one of three strains tested. Enzymatic activities present are F6PPK, a-glucosidase, and B-galactosidase, but not lipase C14, a- chymotrypsin, B-glucuronidase, a-mannodisade, a-fucosidase, esterase lipase C8, esterase C4, naphthol-AS-BI-phosphohydrolase, alkaline phosphatase, trypsin, valine arylamidase and N-acetyl-B-glucosaminidase. Weak enzymatic activities are B- glucosidase, leucine arylamidase, cystine arylamidase, acid phosphatase, and a- galactosidase.

[0229] The WK012_4_13 (= LMG 33104T= DSM 1 16073T), designated as the type strain, was obtained from a water kefir sample. The G+C DNA content of this strain is 56%.

[0230] Description of Bifidobacterium sp. nov. (WK041 4 12) (now assigned its official name of Bifidobacterium aquikefiricola)

[0231] Cells are Gram-positive-staining, non-motile, not spore-forming, F6PPK-positive, catalase- and oxidase-negative, square rod shaped with an average length of 1.18±0.23 pm and an average width of 0.48±0.08 pm. The colonies are white and spherical after 48 hours of incubation in MRScys. Growth occurs under anaerobic and aerophilic conditions. The strain grows in the temperature interval 15-37°C, the pH range 4-8 and NaCI range 0-3%. Optimal conditions for growth occur at 30°C and pH 6. Fermentation occurred from D-ribose, D-fructose, esculin ferric citrate, D-maltose, D-melibiose, D- raffinose, D-glucose and dulcitol but not from glycerol, erythritol, L-xylose, D-adonitol, methyl-B-D-xylopyranoside, L-sorbose, L-rhamnose, inositol, D-mannitol, D-sorbitol, methyl-a-D-mannopyranoside, D-trehalose, inulin, D-melezitose, starch, glycogen, xylitol, D-tagatose, D-fucose, L-fucose, D-arabitol, L-arabitol, potassium 2- ketogluconate, D-lactose, D-arabinose, D-xylose, L-arabinose, D-galactose, potassium gluconate, D-mannose, D-saccharose, N-scetylglucosamine, potassium 5- ketogluconate, amygdalin, arbutin, salicin, D-cellobiose, gentiobiose, D-turanose and D- lyxose. A weak fermentation occurs from methyl-a-D-glucopyranoside. Enzymatic activities are F6PPK, a-glucosidase and N-acetyl-B-glucosaminidase but not C14, a- chymotrypsin, B-glucuronidase, a-mannodisade, a-fucosidase, esterase lipase C8, esterase C4, naphthol-AS-BI-phosphohydrolase, alkaline phosphatase, trypsin, B- galactosidase and a-galactosidase. Weak enzymatic activities are valine arylamidase, B-glucosidase, leucine arylamidase, cystine arylamidase and acid phosphatase.

[0232] The WK041_4_12 (= LMG 33105T= DSM 1 16074T), designated as the type strain, was obtained from a water kefir sample. The G+C DNA content of this strain is 53.9%.

[0233] All Pitched Results

[0234] In this study, the inventors isolated representative species from the most frequently detected genera within the core microbiome. Selected strains included the yeasts Brettanomyces bruxellensis, Hanseniaspora valbyensis, Pichia fermentans, and Saccharomyces cerevisiae; the LAB Lacticaseibacillus paracasei, Lentilactobacillus hilgardii, Leuconostoc pseudomesenteroides, and Liquorilactobacillus satsumensis; the AAB Acetobacter lovaniensis, Acetobacter orientalis, Gluconobacter japonicus, and Gluconobacter oxydans the bifidobacteria Bifidobacterium aquikefiricola and Bifidobacterium fermentum, and the ethanologenic bacterium Zymomonas mobilis.

[0235] To systematically recreate WK fermentations, all possible combinations consisting of one yeast, one LAB, one AAB, one Bifidobacterium species, and with or without Z. mobilis were generated, resulting in a total of 256 pitched fermentations. For practical handling, the fermentations were carried out in four experimental blocks: 1) B. aquikefiricola with Z. mobilis, 2) B. aquikefiricola without Z. mobilis, 3) B. fermentum with Z. mobilis, and 4) B. fermentum without Z. mobilis. WK grain-based fermentations were included as a reference and inoculated at four different grain concentrations: 15, 30, 60, and 120 g L“1.

[0236] Inoculating fermentations with increasing concentrations of WK grains resulted in a stronger initial pH reduction. In contrast, the addition of defined starter cultures had minimal effect on the initial pH. After 72 h of fermentation, all pitched combinations successfully reduced the pH to below 4.4. WK grain-based fermentations reached a similar final pH, irrespective of the initial inoculum concentration. Among the tested microbial groups, AAB had the strongest influence on acidification. Fermentations containing Gluconobacter spp. achieved significantly lower pH values compared to those with Acetobacter spp. and reached levels comparable to the WK grain-based fermentations. The yeast species also influenced the final pH. Fermentations inoculated with Pichia fermentans exhibited significantly higher pH values compared to those with other yeast species.

[0237] All pitched fermentations were inoculated at a total concentration of 2 * 10scells mL“1. Within a few hours, initial microbial growth was detectable across all pitched fermentations, and cell concentrations were comparable to those observed in grainbased fermentations inoculated with 15-30 g L1of WK grains. After 72 h of fermentation, all cultures exhibited robust growth, with final cell densities similar to those of grain-based fermentations. No significant differences in cell concentration were observed between the grain-based fermentations inoculated with 15, 30, 60, or 120 g L“1of grains. When analysed by yeast species, fermentations with P. fermentans exhibited significantly lower cell densities, while those with Br. bruxellensis showed significantly higher cell densities than other pitched fermentations and the grain-based controls. Analysis by LAB revealed that fermentations containing Len. hilgardii had significantly fewer cells mL-1, whereas those with Leu. pseudomesenteroides had significantly more, compared to all other combinations and grain-based fermentations.

[0238] On day 4, fermentations were evaluated for aroma using a 5-point hedonic scale (Figure 10). Fermentations containing Br. bruxellensis were rated as significantly less pleasant than other pitched fermentations or the grain-based controls (Figure 10a). In contrast, combinations containing B. fermentum (Figure 10b) or lacking Z. mobilis (Figure 10e) were rated more pleasant. Among LAB, fermentations with Liq. satsumensis were perceived as significantly more pleasant than those with Lac. Paracasei (Figure 10c). Regarding AAB, fermentations with G. oxydans were rated significantly more pleasant than those containing A. orientalis (Figure 10d).

[0239] A subset of fermentations was selected for metabolite profiling by HPLC (Figure 11 and Figure 12). Most pitched fermentations hydrolysed less sucrose than the WK grainbased controls (Figure 11a-e). Among the yeasts, S. cerevisiae fermentations showed the lowest residual sucrose levels (Figure 11a), while the presence or absence of Z. mobilis had no significant effect on sucrose hydrolysis (Figure 11e). Fructose and glucose levels were significantly lower in fermentations containing B. bruxellensis, H. valbyensis, or P. fermentans compared to those with S. cerevisiae or WK grains (Figure 11f,k). S. cerevisiae fermentations retained the highest concentrations of both monosaccharides. The inclusion of Z. mobilis led to significantly reduced levels of fructose and glucose(Figure 11j,o). Glycerol concentrations were highest in fermentations with P. fermentans and WK grains (Figure 11 p), whereas fermentations with B. bruxellensis produced the lowest levels. The addition of Z. mobilis significantly decreased glycerol production (Figure 11t). Mannitol accumulation was significantly higher in fermentations with Leu. pseudomesenteroides compared to all other conditions (Figure 11w).

[0240] The highest ethanol concentrations were observed in fermentations with WK grains and those containing S. cerevisiae (Figure 12a). The addition of Z. mobilis also significantly increased ethanol production (Figure 12e). Despite lower sucrose consumption, fermentations with Br. bruxellensis and H. valbyensis resulted in elevated levels of acetic acid (Figure 12f). Elevated acetic acid concentrations were also detected in fermentations containing G. japonicus and Z. mobilis (Figure 12i J, respectively). WK grain-based fermentations produced the highest lactic acid concentrations overall (Figure 12k-o). Among the pitched fermentations, those containing Leu. pseudomesenteroides yielded the highest lactic acid levels.

[0241] Pure cultures showed pressure build-up in Falcon tubes for some of the yeasts, and to a lesser extent for some of the LAB. To test this effect in defined combinations, yeast cultures or Z. mobilis were grown together with B. aquikefiricola, Liq. satsumensis, and A. lovaniensis. and measured the resulting pressure build-up. The combination with Z. mobilis led to the fastest pressure increase, followed by fermentations with WK grains and S. cerevisiae. Fermentations with H. valbyensis and B. bruxellensis produced the lowest pressures. Pressure build-up with P. fermentans depended on the duration of the agitated pre-fermentation.

[0242] Sensory evaluation of pitched water kefir

[0243] Based on the results of the pitched fermentations, five combinations were selected for scale-up and evaluated in a WK sensory panel. The combination with B. bruxellensis, “B.acola B.bru L.satG.japZ.mob", had received favourable scores in the smell evaluation but was not perceived as pleasant after scale-up and was therefore not pursued further.

[0244] Grain-based fermentation contained AAB (74.3 % relative taxonomic abundance), LAB (0.3 %), Bifidobacterium spp. (8.3 %) and yeasts (16.9 %), but no Zymomonas mobilis (Table 4). The pitched fermentations contained the species as expected, but H. valbyensis in “B.fer H.val L.par G.jap Z.mob" and Liq. satsumensis in “B.acola S.cerL.sat G.jap" and “B.acola S.cer L.sat G.jap Z.mob" had fallen below the limit of detection (requiring at least 35% breadth and a minimum ratio of 0.75 for the expected breadth to observed breadth).

[0245] Table 4: Species-level composition of WK grain-based fermentations and pitched fermentations from the WK taste panel.

[0246] Metagenomics-based taxonomic composition of grain-based and pitched fermentations from the WK taste panel, along with one grain-based fermentation from the screening of pitched fermentations (first column). **Species detected in grain-based fermentations. ***Species inoculated in pitched fermentations. *Species that were inoculated in pitched fermentations but were no longer reliably detected after applying stringent inStrain filters. In the WK sensory acceptance test (Figure 13), WK grain-based fermentations received the highest appearance ratings, significantly outperforming “B.fer P.fer L.hil A.lov Z.mob" (Figure 13a). The highest liking scores for aroma and flavour were observed in pitched fermentations containing S. cerevisiae, including “B.fer S.cer Lhil G.oxy", “B.acola S.cer Lsat G.jap Z.mob", and “B.acola S.cer L.sat G.jap", as well as in the WK grain-based samples (Figure 13b, c). In contrast, fermentations with H. valbyensis 'B.fer H.val L.par G.jap Z.mob") and P. fermentans (‘B.fer P.fer Lhil A.lov Z.mob") scored significantly lower than other combinations in aroma, flavour, and overall acceptability (Figure 13b, c,e). No significant differences were observed in viscosity ratings across the tested combinations (Figure 13d). Some combinations with Leu. pseudomesenteroides had shown a gelling of the ferment in the initial screening, most likely due to exopolysaccharide production.

[0247] In the Optimised Descriptive Profile (ODP; Figure 14), “B.fer H.val L.par G.jap Z.mob" ranked highest for both off-flavour and acetone flavour (Figure 14a and I, respectively). Fermentations containing Z. mobilis (‘B.fer H.val L.par G.jap Z.mob", “B.fer P.fer L.hil A.lov Z.mob", and “B.acola S.cer L.sat G.jap Z.mob") received the highest scores for yeasty flavour (e), apple cider flavour (i), vinegar flavour (k), sourtaste (n), effervescence (o), and alcohol aftertaste (p), often significantly exceeding those of the other pitched combinations or the WK grain-based reference. The combination “B.fer S.cer Lhil G.oxy" scored highest in fruity estery flavour (f), berry flavour (g), and pear flavour (j), and lowest in off- flavour (a), medicinal flavour (d), yeasty flavour (e), vinegar flavour (k), and acetone flavour (I), indicating a cleaner and more pleasant flavour profile. Fermentations containing S. cerevisiae and no Z. mobilis (‘B.fer S.cer L.hil G.oxy", “B.acola S.cer Lsat G.jap") were rated significantly sweeter than combinations that included Z. mobilis (m).

[0248] The fermentations tested in the WK taste panel were analysed for metabolite composition by HPLC (Figure 15). The combinations “B.acola S.cer L.sat G.jap Z.mob" and “B.acola S.cer L.sat G.jap" showed significantly lower residual sucrose levels compared to WK grain-based fermentations, while “B.fer H.val L.par G.jap Z.mob" had significantly higher remaining sucrose. The combinations “B.fer S.cer L.hil G.oxy” and “B.fer P.fer L.hil A.lov Z.mob" did not differ significantly from the WK grain-based reference (Figure 15a). From the initial 80 g L“1of sucrose, between 25 and 66 g L“1remained after fermentation. “B.fer S.cer Lhil G.oxy” and “B.acola S.cer L.sat G.jap" contained significantly higher concentrations of glucose and fructose than WK grain-based fermentations, while “B.fer P.fer L.hil A.lov Z.mob" and “B.fer H.val L.par G.Jap Z.mob" had significantly lower levels (Figure 15b, c). Additionally, “B.acola S.cer L.sat G.jap Z.mob" contained significantly less glucose than the WK grain-based control (Figure 15b)). “B.fer P.fer L.hil A.lov Z.mob" and “B.acola S.cer L.sat G.jap Z.mob" produced significantly higher ethanol levels than the WK grain-based fermentations, while “B.fer S.cer L.hil G.oxy” yielded the lowest ethanol content (0.28% ABV; Figure 15d). In comparison, the WK grain-based fermentation reached 0.61 % ABV. “B.acola S.cer Lsat G.jap" and WK grain-based fermentations produced the highest levels of glycerol (Figure 15e). Lactic acid concentrations did not differ significantly between pitched and grain-based fermentations (Figure 15f). However, pitched fermentations containing Z. mobilis exhibited significantly higher acetic acid levels than WK grain-based fermentations (Figure 15g). Notably, the combinations “B.acola S.cer Lsat G.jap Z.mob" and “B.acola S.cer L.sat G.jap" displayed considerable replicate-to-replicate variation in metabolite concentrations.

[0249] The most abundant volatile organic compounds (VOCs) in the pitched fermentations were acids, alcohols, and esters (Figure 18). The five most abundant VOCs were 1- butanol, 2-methyl- (described as malty, roasted, winey, fruity, fusel, alcoholic), 1-butanol, 3-methyl- (whiskey, fusel, alcoholic, fruity, banana), 1-propanol, 2-methyl- (ethereal, winey), acetic acid (pungent, vinegar-like), and ethanol (wine-like, alcoholic);(Figure 19)). Hierarchical clustering of VOC profiles revealed compounds characteristic of the pitched fermentations (Figure 20). For example, compounds in the cluster containing compounds such as methyl vinyl ketone and octyl aldehyde were typical of the WK medium and were not detected in the fermented products. Esters in the cluster with ethyl nonanoate (fatty, fruity, brandy), isobutyl acetate (fruity, floral), and propanoic acid, 2- methyl-, ethyl ester (sweet, fruity, slightly pineapple) were characteristic of WK grainbased fermentations. The “B.fer H.val Lpar G.jap Z.mob" fermentation was characterised by a relatively high abundance of methyl isobutyl ketone (sharp, solvent, green, herbal, fruity, dairy, spicy), acetic acid, butyl ester (pear, ethereal, green), and ethyl propanoate (pineapple, solvent, fruity), all potentially contributing to what was described as an acetone flavour (Figure 141).

[0250] The importance of AAB in water kefir fermentations

[0251] The contribution of AAB (G. oxydans and G. japonicus) was evaluated in two pitched combinations: “B.fer S.cer L.hil G.oxy” and “B.acola S.cer L.sat G.jap Z.mob". Inclusion of G. oxydans in “B.fer S.cer L.hil G.oxy’ resulted in faster acidification, reaching pH <4.4 after 9.4 h, compared to 44.6 h in the corresponding combination lacking G. oxydans (‘B.fer S.cer L.hil". Similarly, “B.acola S.cer L.sat G.jap Z.mob" reached pH < 4.4 after5 h, whereas the version without G.japonicus (“B.acola S.cer L.sat Z.mob") required 24.1 h. Biomass production was comparable between “B.fer S.cer L.hil G.oxy” and “B.fer S.cer Lhil". In contrast, “B.acola S.cer L.sat Z.mob" showed increased biomass accumulation compared to “B.acola S.cer L.sat G.jap Z.mob"', however, this increase was not reflected in a higher number of cells mL“1. Fermentations lacking G. oxydans or G. japonicus exhibited reduced consumption of dissolved oxygen (DO).

[0252] The omission of AAB (G. oxydans and G. japonicus) in the pitched fermentations “B.fer S.cer L.hil G.oxy” and “B.acola S.cer L.sat G.jap Z.mob" did not significantly affect sucrose hydrolysis, but resulted in a significant increase in glucose concentrations. Fructose levels were unchanged between “B.fer S.cer Lhil G.ox ’ and “B.fer S.cer Lhil”, while “B.acola S.cer Lsat Z.mob" exhibited significantly higher fructose concentrations than “B.acola S.cer L.sat G.jap Z.mob". The omission of AAB led to significantly elevated levels of ethanol, glycerol, and lactic acid. In contrast, acetic acid concentrations were drastically reduced in the absence of G. oxydans or G. japonicus.

[0253] The importance of bifidobacteria in water kefir fermentations

[0254] The role of bifidobacteria in pitched fermentations was evaluated using two combinations: “S.cer L.hil G.oxy” and “S.cer L.sat Z.mob G.jap", each tested with the addition of B. aquikefiricola, B. aquikefiri, B. fermentum, or no Bifidobacterium species. The inclusion or omission of Bifidobacterium spp. did not significantly affect the rate of pH reduction or biomass accumulation in either fermentation combination. However, some influence on dissolved oxygen (DO) consumption was observed. Fermentations containing B. fermentum, specifically “B.fer S.cer L.hil G.ox ’, and “S.cer L.sat Z.mob G.jap", showed the strongest DO depletion.

[0255] The addition of Bifidobacterium spp. to “S.cer Lhil G.oxy’ did not significantly alter concentrations of sucrose, glucose, fructose, ethanol, glycerol, or acetic acid. However, “B.fer S.cer Lhil G.oxy’ and “B.aqui S.cer L.hil G.oxy’ showed significantly higher levels of lactic acid compared to “S.cer L.hil G.oxy’ without Bifidobacterium. In the combination “S.cer Lsat G.jap Z.mob", the addition of Bifidobacterium spp. did not significantly affect concentrations of glucose, fructose, glycerol, lactic acid, or acetic acid. Minor differences were observed in sucrose consumption and ethanol levels. Several free amino acids are depleted during fermentation

[0256] During fermentation, the concentration of free amino acids (FAA) decreased from 829 nmol mL-1in the mWK medium to 77-252 nmol mL-1in the pitched fermentations. The use or omission of different Bifidobacterium species resulted in comparable FAA profiles. The omission of AAB (G. oxydans or G. japonicus) led to further depletion of FAA, and proline was undetectable in these fermentations.

[0257] Alanine, aspartic acid, cysteine, GABA, glycine, proline, serine, taurine, threonine, and valine were strongly depleted during fermentation. GABA was present at 63 nmol mL-1in the medium and decreased to ~4 nmol mL-1in fermentations containing Z. mobilis, and to <1 nmol mL-1in those lacking Z. mobilis. Proline, the most abundant amino acid in the medium (387 nmol mL-1), decreased to 168-188 nmol mL-1in most pitched fermentations, but was not detectable when AAB were omitted. Arginine, histidine, leucine, and lysine concentrations remained mostly unchanged during fermentation. Methionine sulfone, methionine, phenylalanine, tryptophan, tyrosine, and isoleucine were either undetectable or present only at low concentrations (<5 nmol mL-1) in both the medium and fermentations. Cysteic acid and glutamic acid tended to increase during fermentation but decreased in the presence of Z. mobilis. The omission of G. Japonicus in the combination “B.acola S.cer L.sat Z.mob" compared to “B.acola S.cer Lsat G.jap Z.mob" resulted in a significant increase in free glutamic acid levels.

[0258] Discussion

[0259] WK is a fermented beverage produced with a diverse set of fermentation practices and a diverse set of microbes. The Applicants obtained 69 WK grains from 21 different countries to study WK in an unprecedented depth. Endpoint samples of WK fermentations were collected at 48h, as well as another sample representative of the early stages of the fermentation (8h). There was a relatively even split between aerobic and anaerobic fermentations among the participants. This is an interesting feature of WK as other fermented beverages typically have specific aerobic or anaerobic needs, e.g., kombucha and vinegar fermentations are performed under aerobic conditions, while wine, beer, and cider fermentations mostly require anaerobic conditions to allow Sa. cerevisiae to produce ethanol. The Applicants set out to validate their predictions of novel species and successfully isolated the two novel Bifidobacterium species described herein, extending the number of bifidobacteria that have been isolated from WK.

[0260] To the knowledge of the Applicant, 68 of the 96 species detected with inStrain, have not been isolated from WK before. Considering culture dependent and independent methods, 41 / 96 species have not been previously detected in WK and 18 of these were completely novel species at the time of analysis (see Figure 8 and Figure 9).

[0261] Most WK studies have analysed the microbial composition of a single WK community or a small number of communities. As these studies have used different methods for species detection, such as culture based and culture independent methods with different protocols, it is difficult to compare the results and to estimate the prevalence of species within the different WK communities. In this study, the Applicant has analysed 69 WK communities with the same experimental protocols, allowing unprecedented insights into prevalent taxa within WK and with this providing a foundation for defining a WK core microbiome. Core microbiomes are commonly defined either by abundance or occupancy of taxa, or through combinations of these two. The Applicants propose a minimum 30% occupancy for defining the WK core microbiome. Following this definition, the WK core microbiome comprises the following key members: Bifidobacterium, yeast (such as Saccharomyces, Zygotorulaspora, Pichia, and Brettanomyces), LAB (such as Liquorilactobacillus, Lacticaseibacillus, Lentilactobacillus, Leuconostoc and Oenococcus), AAB (such as Gluconobacter and Acetobacter) and Zymomonas. Genera, such as Schleiferilactobacillus, Lachancea, Gluconacetobacter, Komagataeibacter, Sporolactobacillus, Ethanoligenens, Novacetimonas, Pseudoclavibacter, and Hanseniaspora, are less frequently detected in water kefir.

[0262] The definition of the WK core microbiome and the average of 10.3 species per sample, observed in this study, provides a foundation for defining what a synthetic or pitched WK community, for WK production at industrial scale, should look like and will contribute towards a regulatory definition of WK. This WK core microbiome makes up the water kefir base. Other elements that can be added to the WK core microbiome afterwards to add more flavour, for example, fruits, flavours, etc

[0263] Volatile organic compound (VOC) detection can be used to assess food quality and safety, and furthermore can be used to monitor the progression of food fermentations. In WK, an increase in volatile alcohols was observed after 8h of fermentation and the volatile profile was dominated by alcohols, acids, and esters after 48h of fermentation, leading to a clear separation of the volatile profiles and an increased accumulation of VOCs with time. In general, yeasts and LABs are key producers of esters (created from an acid and an alcohol), while yeasts are also able to produce complex alcohols, and volatile fatty acids can be produced by microbes, such as Acetobacter, Clostridium and Propionibacterium spp.

[0264] Volatile esters are highly important aroma compounds in fermented foods, such as wine and beer. The highest number of positive correlations between esters and species are with Brettanomyces spp., highlighting Brettanomyces spp. as key flavour producers in WK. The role of Brettanomyces spp. in fermented beverages has been discussed before and the role of Brettanomyces spp. can range from essential for good flavours (e.g., in lambic beers) to a source for off-flavours (e.g., in wine).

[0265] Successful reconstruction of water kefir

[0266] Reconstructing the complex microbial community of water kefir using defined starter cultures remains challenging due to the intricate interactions between bacteria, yeasts, and their physicochemical environment. Previous efforts using selected microbial consortia have yielded only partial success.

[0267] The approach described herein involved the systematic reconstruction of 256 fermentation combinations derived from 13 water kefir isolates, representing 11 of the 13 core genera previously identified. These combinations included representatives from the five major functional groups in WK: yeasts, LAB, AAB, Bifidobacterium spp. and Z. mobilis. In previous studies, fermentations were inoculated with 6 x 10sto 1.7 x 107bacterial cells mL“1(or CFU mL-1) and 3.67 x5to 1 x 10syeast cells mL“1(or CFU mL-1). In the experiments described herein, each strain was added at 5 x 10scells mL-1in four-strain communities or 4 x 10scells mL-1in five-strain communities, maintaining a total inoculum of 2 x 10scells mL-1, comparable to WK grain-based fermentations inoculated at 15-30 g L-1. Despite this lower per-strain inoculation, the pitched fermentations reached final cell densities of approximately 108cells mL-1, similar to those observed in WK grain-based fermentations and in previous studes.

[0268] The pitched fermentations described in previous studies did not include AAB. In contrast, the Applicant identified Gluconobacter and Acetobacter as the second and fourth most frequently detected genera in water kefir, present in over 75% of samples and classified as part of the WK core microbiome. Based on this, AAB were included in all pitched fermentations in the present study. All combinations acidified the medium to below pH 4.4 within 72 h, consistent with the acidification observed in grain-based fermentations. Achieving a final pH below 4.4 is relevant from a food safety perspective, as it can inhibit the growth of foodborne pathogens such as Clostridium botulinum.

[0269] The broad screening of pitched fermentations enabled the identification of general trends. Yeast species had a strong influence on aroma, with fermentations containing S. cerevisiae tending to receive the highest aroma ratings, while those with Br. bruxellensis were less favourably perceived. In addition to sensory effects, species-specific patterns in metabolite profiles were observed across the different combinations. The role of specific taxa in WK production is discussed in more detail below.

[0270] Sensory outcomes and consumer acceptance

[0271] To evaluate the acceptability of the rationally designed water kefir fermentations, five pitched combinations were selected based on aroma ratings and metabolite profiles and assessed by a naive consumer panel. The hedonic evaluation showed that fermentations containing S. cerevisiae “B.fer S.cer L.hil G.oxy", “B.acola S.cerL.sat G.jap Z.mob", and “B.acola S.cerL.sat G.jap") were on par with water kefir produced using grains, receiving the highest scores in appearance, aroma, flavour, and overall acceptability. In contrast, fermentations containing H. valbyensis or P. fermentans received significantly lower ratings across multiple sensory dimensions, particularly aroma and flavour. To further characterise sensory attributes, the same panel conducted an Optimised Descriptive Profile (ODP) analysis. The most favourable profile was associated with “B.fer S.cerL.hil G.oxy", marked by strong fruity, estery, and berry notes, along with a pear-like flavour and minimal off-flavour, medicinal, or acetone-like characteristics. Pitched fermentations containing Z. mobilis scored higher in descriptors such as yeasty flavour, sourness, vinegar and apple cider flavour, and alcohol aftertaste, indicating an association with less favourable sensory notes. In contrast, combinations with S. cerevisiae and lacking Z. mobilis were perceived as significantly sweeter and contained less ethanol.

[0272] The Applicant has shown previously that Br. bruxellensis showed strong positive correlations with various esters, suggesting a potential role in water kefir flavour development. This hypothesis was tested here; however, fermentations containing Br. bruxellensis tended to be perceived as less pleasant. Notably, the combination “B.acola B.bru L.sat G.jap Z.mob", which initially scored well in aroma evaluation, performed poorly when scaled up, highlighting the importance of assessing fermentation performance under conditions that closely mimic consumer-relevant formats.

[0273] Taxa specific functional roles

[0274] Yeasts play a key role in water kefir fermentations by expressing invertases that hydrolyse the disaccharide sucrose into the monosaccharides glucose and fructose, which in turn become more readily available to the bacterial community. Among the yeast strains tested here, S. cerevisiae was the most efficient at sucrose hydrolysis. Yeasts also act as producers of ethanol during fermentation. Interestingly, however, the superior sucrose hydrolysis capacity of S. cerevisiae did not directly translate into higher ethanol concentrations. Fermentations with Br. bruxellensis yielded the lowest ethanol levels but relatively high concentrations of acetic and lactic acids, alongside low glycerol production.

[0275] Acetic acid bacteria are commonly found in a range of fermented foods, including water kefir, milk kefir, kombucha and vinegar, as well as diverse environmental niches. WK harbours a variety of AABs, which are likely to contribute different organic acids during fermentation. In this study, the Applicant has demonstrated that the inclusion of AAB, particularly G. japonicus and G. oxydans, markedly increased acetic acid concentrations while simultaneously reducing ethanol and glucose levels. Although G. oxydans produced the lowest acetic acid levels among the pitched fermentations, it likely converted glucose into gluconic acid instead. Another key role of AAB in WK fermentations is their capacity to rapidly reduce pH, thereby enhancing food safety. Here, it was observed that while pitched fermentations without AAB still achieved a final pH below 4.4, they required substantially more time to do so compared with fermentations containing AAB.

[0276] This study demonstrates that rationally designed starter cultures can reproduce key sensory and metabolic features of traditional water kefir, while offering greater control over the fermentation. The results highlight the functional contributions of individual species, such as the desirable impact of S. cerevisiae and Gluconobacter spp.. Together, these findings provide a basis for the development of stable, defined consortia tailored to deliver safe, consistent, and appealing non-alcoholic fermented beverages. In the specification the terms "comprise, comprises, comprised and comprising" or any variation thereof and the terms “include, includes, included and including" or any variation thereof are considered to be totally interchangeable and they should all be afforded the widest possible interpretation and vice versa.

[0277] The invention is not limited to the embodiments hereinbefore described but may be varied in both construction and detail.

[0278] (Original in Electronic Form)

[0279] (This sheet is not part of and does not count as a sheet of the international application)

[0280] (Original in Electronic Form) (This sheet is not part of and does not count as a sheet of the international application)

[0281] FOR RECEIVING OFFICE USE ONLY

[0282] FOR INTERNATIONAL BUREAU USE ONLY

Claims

Claims1 . A water kefir starter culture comprising a yeast and optionally one or more selected from an acetic acid bacteria, a lactic acid bacteria, a Bifidobacteria strain, and a Zymomonas bacteria.

2. A water kefir starter culture according to Claim 1 comprising a yeast and an acetic acid bacteria, and optionally one or more selected from a lactic acid bacteria, a Bifidobacteria strain, and a Zymomonas bacteria.

3. The water kefir starter culture according to Claim 1 or Claim 2, wherein the yeast is selected from a Saccharomyces, a Brettanomyces, a Zygotorulaspora, a Pichia, a Lachancea, a Hanseniaspora, a Candida, a Schizosaccharomyces or a combination thereof.

4. The water kefir starter culture according to Claim 3, wherein the Saccharomyces yeast is selected from Saccharomyces cerevisiae, Saccharomyces bayanus, Saccharomyces uvarum, a Saccharomyces species isolated from water kefir or a combination thereof.

5. The water kefir starter culture according to Claim 3 or Claim 4, wherein the Brettanomyces yeast is selected from Brettanomyces bruxellensis, Brettanomyces anomalus, Brettanomyces sp._HC-2020a, or a combination thereof.

6. The water kefir starter culture according to any one of Claim 3 to 5, wherein the Zygotorulaspora yeast is Zygotorulaspora florentina.

7. The water kefir starter culture according to any one of Claims 3 to 6, wherein the Pichia yeast is selected from Pichia fermentans, Pichia kluyveri, Pichia membranifaciens, Pichia occidentalis, or a combination thereof.

8. The water kefir starter culture according to any one of the preceding Claims, wherein the acetic acid bacteria is selected from Gluconacetobacter, Gluconobacter, Acetobacter, Komagataeibacter, Novacetimonas or a combination thereof.

9. The water kefir starter culture according to Claim 8, wherein the Gluconobacter bacteria is selected from Gluconacetobacter liquefaciens, Gluconobacter albidus, Gluconobacter cadivus, Gluconobacter cerinus, Gluconobacter frateurii, Gluconobacterjaponicus, Gluconobacter kanchanaburiensis, Gluconobacter kondonii, Gluconobacter oxydans, Gluconobacter potus, Gluconobacter roseus, Gluconobacter sp. Gdi, Gluconobacter vitians, or a combination thereof.

10. The water kefir starter culture according to Claim 8 or Claim 9, wherein the Acetobacter bacteria is selected from Acetobacter aceti, Acetobacter fabarum, Acetobacter indonesiensis, Acetobacter lovaniensis, Acetobacter malorum, Acetobacter okinawensis, Acetobacter orientalis, Acetobacter papaya, Acetobacter pasteurianus, Acetobacter persici, Acetobacter senegalensis, Acetobacter sicerae, Acetobacter sp. UBA5411, Acetobacter syzygii, Acetobacter tropicalis, an Acetobacter species isolated from water kefir, or a combination thereof.11 . The water kefir starter culture according to any one of Claims 8 to 10, wherein the acetic acid bacteria is Gluconacetobacter dulcium, Komagataeibacter saccharivorans, or a combination thereof.

12. The water kefir starter culture according to any one of the preceding Claims, wherein the lactic acid bacteria is selected from the genus Liquorilactobacillus, Lacticaseibacillus, Lentilactobacillus, Leuconostoc, Oenococcus, Schleiferilactobacillus, Sporolactobacillus or a combination thereof.

13. The water kefir starter culture according to Claim 12, wherein the lactic acid bacteria is selected from Liquorilactobacillus satsumensis, Liquorilactobacillus mali, Liquorilactobacillus ghanensis, Liquorilactobacillus hordei, Liquorilactobacillus nagelii, a Liquorilactobacillus species isolated from water kefir, Lacticaseibacillus paracasei, Leuconostoc pseudomesenteroides, Lentilactobacillus hilgardii, Oenococcus oeni, Oenococcus sicerae, Oenococcus kitaharae, Lentilactobacillus diolivorans, Leuconostoc mesenteroides, Schleiferilactobacillus harbinensis, or a combination thereof.

14. The water kefir starter culture according to any one of the preceding claims, wherein the Bifidobacteria strain is selected from an isolated Bifidobacteria strain (WK012_4_13) Bifidobacterium fermentum deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 26thDecember 2023 under LMG Deposit Accession No. LMG 33104, an isolated Bifidobacteria strain Bifidobacterium sp. nov (WK041_4_12) Bifidobacterium aquikefiricola deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 20thJuly 2023 under LMG Deposit Accession No. LMG 33105, Bifidobacterium tibiigranulil , Bifidobacterium aquikefiri, Bifidobacteriumpsychraerophilum, Bifidobacterium aquikefiri, Bifidobacterium psychraerophilum, or variants thereof.

15. The water kefir starter culture according to Claim 14, wherein both of the isolated Bifidobacteria strains are selected.

16. The water kefir starter culture according to Claim 14 or Claim 15, wherein the isolated Bifidobacteria strain comprises the characteristic of having fructose-6-phosphate phosphoketolase activity, can be grown in aerobic or anaerobic conditions at between 14°C and 38°C; and can grow on D-arabinose and D-xylose and a NaCI concentration of 0-3 (w / v%), when compared to a control B. aquikefiri LMG 28769.

17. The water kefir starter culture according to Claim 14 or Claim 15, wherein the isolated Bifidobacteria strain comprises the characteristic of having fructose-6-phosphate phosphoketolase activity and N-acetyl-B-glucosaminidase activity, can be grown in aerobic or anaerobic conditions at between 14°C and 38°C, and can grow on dulcitol but not D-mannose or D-saccharose, and a NaCI concentration of 0-3 (w / v%), when compared to control B. aquikefiri LMG 28769.

18. The water kefir starter culture according to any one of one of the preceding claims, wherein the yeast and bacterial strains are in a freeze-dried or a spray-dried form.

19. The water kefir starter culture according to any one of the preceding claims, wherein the Zymomonas is Zymomonas mobilis.

20. The water kefir starter culture according to Claim 19, wherein the Zymomonas mobilis is Zymomonas mobilis subsp. mobilis, Zymomonas mobilis subsp. pomaceae, or a combination of both.

21. A water kefir starter culture comprising a Bifidobacteria strain selected from an isolated Bifidobacteria strain (WK012_4_13) Bifidobacterium fermentum deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 26thDecember 2023 under LMG Deposit Accession No. LMG 33104, an isolated Bifidobacteria strain Bifidobacterium sp. nov (WK041_4_12) Bifidobacterium aquikefiricola deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 20thJuly 2023 under LMG Deposit Accession No. LMG 33105, Bifidobacterium tibiigranulil, Bifidobacterium aquikefiri, Bifidobacterium psychraerophilum, Bifidobacterium aquikefiri,Bifidobacterium psychraerophilum, or variants thereof, and, and optionally one or more selected from a yeast, an acetic acid bacteria, a lactic acid bacterium, a Zymomonas bacteria.

22. An isolated Bifidobacteria strain, Bifidobacterium sp. nov (WK041_4_12) Bifidobacterium aquikefiricola, deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 20thJuly 2023 under LMG Deposit Accession No. LMG 33105, and variants thereof, wherein the variants are characterised in that they are isolated, they belong to or are closely related to the strain Bifidobacterium sp. nov (WK041_4_12) Bifidobacterium aquikefiricola, and they show fructose-6-phosphate phosphoketolase and N-acetyl-B-glucosaminidase activity, and exhibit characteristics such as the ability to grow on dulcitol but not D-mannose or D-saccharose, and a NaCI (w / v%) of 0-3 when compared to control B. aquikefiri LMG 28769.

23. An isolated Bifidobacteria strain Bifidobacterium sp. nov (WK012_4_13) Bifidobacterium fermentum deposited with the Belgian Coordinated Collections of Microorganisms (BCCM-LMG) on 26thDecember 2023 under LMG Deposit Accession No. LMG 33104, or variants thereof, wherein the variants are characterised in that they are isolated, they belong to or are closely related to the strain Bifidobacterium sp. nov (WK012_4_13) Bifidobacterium fermentum, and they have fructose-6-phosphate phosphoketolase activity, can be grown in aerobic or anaerobic conditions at between 14°C and 38°C; and can grow on D-arabinose and D-xylose and a NaCI (w / v%) concentration of 0-3, when compared to a control B. aquikefiri LMG 28769.

24. A formulation comprising the isolated strain of Claim 22 or Claim 23.

25. A formulation comprising the water kefir starter culture according to any one of Claims 1 to 21.

26. The formulation of Claim 24 or Claim 25 which is a pharmaceutical formulation and comprises a pharmaceutically acceptable carrier.

27. The formulation of Claim 24 or Claim 25 which is a comestible product.

28. The formulation according to Claim 27, wherein the comestible product is a food product.

29. The formulation according to Claim 27, wherein the food product is a fermented food selected from a water-based kefir, kombucha, tepache, beer, wine, sauerkraut, kimchi, pickles, soy milk kefir, almond milk kefir, oat milk kefir, coconut milk kefir, miso, tamari, tempeh, buttermilk, acidophilus milk, sour cream, cottage cheese, aged cheese, yoghurt, milk kefir, or a food product fortified with probiotics.

Images