WO1999062316A2

WO1999062316A2 - Identification of lactic acid bacteria genes involved in the biosynthesis of exopolysaccharides

Info

Publication number: WO1999062316A2
Application number: PCT/EP1999/002841
Authority: WO
Inventors: Francesca Stingele; Jacques Edouard Germond; Gilbert Lamothe
Original assignee: Societe Des Produits Nestle S.A.
Priority date: 1998-04-22
Filing date: 1999-04-22
Publication date: 1999-12-09
Also published as: WO1999054475A2; WO1999062316A3; AU5969699A; AU4604499A

Abstract

The invention concerns a recombined enzyme capable of being involved in exopolysaccharide biosynthesis by binding in the form of an isomer α or β between the carbon 1 bearing the aldehyde reduction function of an activated D-Galp and a lipophilic phosphate primer, in particular of the undecaprenol or protein type, particularly of the ACP type (Acyl Carrier Protein), so as to gradually ensure the formation of said exopolysaccharides including at each step the addition of a novel saccharide unit which is urged to be bound by its semi-acetal function to an alcoholic hydroxyl of another saccharide unit, which is at the end of a saccharide chain bound to said primer.

Description

LACTIOTTES BACTERIA PRODUCING EXOPOLYSACCHARIDES

Subject 1 ^"invention

The present invention relates to new enzymatic functions of food lactic acid bacteria, as well as new enzymes and genes, involved in the biosynthesis of exopolysaccharides.

State of the art

It is known that lactic acid bacteria are capable of producing in their culture medium two classes of polysaccharides, namely homopoiysaccharides such as dextrans or levans which are constituted by the repeated assembly of a single saccharide, and heteropolysaccharides commonly called exopolysaccharides or EPS (EPS is the abbreviation of the term "exopolysaccharide") constituted by the assembly of several different saccharides forming a repeating unit (Cerning J., Lactic bacteria, Vol I, of Rossart H and Luquet FM, Lorica, 309- 329, 1994).

A lactic acid bacterium producing EPS can give a stringy character and / or a smooth and creamy texture to an acidified milk (Cerning et al., FEMS Microbiol., 87, 113-130, 19/90). PSE can also present biological activities of particular interest for human or animal health, such as anti-tumor or probiotic activities, for example (Oda M. et al., Agric. Biol. Chem., 47, 1623-1625, 1983; EP-94870139.6)

In addition, the food industry is faced with a genetic instability of the biosynthesis of EPS in lactic acid bacteria. This usually results during fermentation by the loss of EPS production by all or part of the lactic acid bacteria

(see "Cerning J." above). Industrial fermented products are therefore subject to variations in their EPS content, which is not always acceptable. To remedy these problems, the industry is currently resorting to the isolation and periodic characterization of its bacteria so as to separate those which have lost their original character.

Genes of dietary lactic acid bacteria involved in the oiosynthesis of EPS are known.

Document WO92 / 02142 thus reveals the existence of the plasmid pHV67 which produces in Lactococcus lactis subsp. lacti s (mesopmle) a substance capable of increasing the viscosity of fermented milk. As the saccharide structure of this substance is not known, the different functions of the enzymes coded by pHV67 are also unknown. Document US-5,066,588 describes two plasmids originating from a strain of Streptococcus cremoris

(mesophilic) capable of giving a thickening character to a Streptococcus lactis. As the saccharide structure of this thickener is not known, the different functions of the enzymes encoded by these plasmids are also unknown.

Vescovo et al. have demonstrated a plasmid of a strain Lactobacillus caseï subsp. caseï

(mesophilic) coding for a Muc + phenotype, i.e. for functions related to the production of exocellular thickeners (Vescovo et al., Biotechnology Letters, Vol II, 709-712, 1989). As the saccharide structure of this thickener is not known, the different functions of the enzymes encoded by this plasmid are also unknown.

Document EP-A-0750043 (Nestlé Products Company) describes a gene operon of the Streptococcus ther ophilus CNCM 1-1590 strain involved in the synthesis of an EPS having the following repetitive structure:

→ 3) -β-D-Glc /? (1 → 3) -α-D-Gal / -NAc (1 → 3) -β-D-Gal / - (1 →

6

1 α -D-Galp

Van Kranenburg et al. also described a gene operon, present on a plasmid of a Lactococcus lactis strain, involved in the synthesis of an EPS having the following repetitive structure (Mol. Microbiology, 24, 387-397, 1997):

→ 4) -β-D-Gal / j (1 → 4) -β-D-Glc /? (1 → 4) -β-D-Glc / (1 →

2 3

α-D-Gal /?

Although the enzymes encoding this EPS are now known, their specificity remains to be elucidated. To date, there is still a need to also have new enzymes involved in the synthesis of EPS; in particular enzymes catalyzing the specific bonds between saccharides to improve the properties of food products, in particular milk products such as yogurts or cheeses, by incorporating such saccharides therein. The present invention aims to fulfill these needs. Another object of the present invention is to provide new saccharides, certain branches or elements of which can be advantageously used or treated to improve the organoleptic properties and the flavor (stability of the aromas of food products possibly incorporating such saccharides).

Summary of the invention The present invention relates to any recombinant enzyme, which can be purified and which is capable of being involved in the biosynthesis of EPS by a bond in the form of an α or β isomer between carbon 1 carrying the aldehyde reducing function of an activated T-Galp and a phosphate of a lipophilic primer, in particular of the genus undecaprenol, or protein, in particular of the genus ACP (Acyl Carrier Protein); for example, so as to gradually ensure the formation of these exopolysaccharides with, at each step, the addition of a new saccharide unit which attaches by its semi-acetal function to an alcoholic hydroxyl of another unit of saccharide, which esc at the end of a chain of saccharides linked to this primer.

Preferably, said recombinant enzyme is chosen from the group consisting of: l) recombinant and / or purifiable enzymes involved in the biosynthesis of one exopolysaccharide having the following repeating saccharide unit: β-D-Gal / -l

^ 3) -β-D-Glc ^ - (l → 3) -β-D-Gal / ^" - (l → 3) -α-D-Glc / 7- (l →

specifically catalyzing one of the following bonds between saccharides: - the β-1,3 osidic bond between the carbon 1 of an activated D-Galf and the carbon 3 of Glcp present at the non-reducing end of a saccharide chain in formation, in particular the α / β-D chain -Glcp-primer; - the β-1,3 osidic bond between the carbon 1 of an activated D-Glcp and the carbon 3 of Galf present at the non-reducing end of a saccharide chain in formation, in particular the β-D-Galf chain ^" (1 → 3) -α / β-D- Glcp-primer; - the β-1,6 osidic bond between carbon 1 of an activated D-Galp and carbon 6 of Glcp present at the non-terminus reducing a saccharide chain in formation, in particular the chain β-D-Glcp (1 → 3) β-D- Galf (l-3) -α / β-D-Glcp-primer; and - the osidic bond α- 1.3 between the repeating sacchaπdiques units of 1 EPS above; 2) the recombinant and / or purifiable enzymes which are likely to be involved in the synthesis of EPS having the following repeating saccharide unit: β-D-Gal / (l → 4) -β-D-Glc / 71 i

→ 5) -β-D-Gal / - (l → 3) -α-D-Galp- (l → 3) -α-D-Glcp- (l → 3) -β-D-Glcp- (l → specifically catalyzing one of the following bonds between saccharides:

- the α-1,3 osidic bond between carbon 1 of an activated D- Glcp and carbon 3 of a Glcp belonging to the non-reducing end of a saccharide chain in formation, in particular the following chain: α / β-D- Glcp-primer;

- the α-1,3 osidic bond between carbon 1 of an activated D-Galp and carbon 3 of a Glcp present at the non-reducing end of a saccharide chain in formation, in particular the following chain: α-D-Glcp- (l- »3) -β / α-D-Glcp-primer;

- the β-1,3 osidic bond between carbon 1 of an activated D-Galf and carbon 3 of a Galp present at the non-reducing end of a saccharide chain in formation, in particular the following chain: α -D-Galp- (1 → 3) -α-D-Glcp- (l- »3) -β / α-D-Glcp-primer;

- the β-1,3 osidic bond between carbon 1 of an activated D-Glcp and carbon 3 of a Galf present at the non-reducing end of a saccharide chain in formation, in particular the following chain: β -D-Galf- (l- »3) -α-D-Galp- (1 → 3) -α-D-Glcp- (1 → 3) -β-D-Glcp-primer;

- the β-1,4 osidic bond between the carbon 1 of an activated D-Galp and the carbon 4 of Glcp present at the non-reducing end of a saccharide chain in formation, in particular the following chain: β-D -Glcp- (1 → 3) -β-D-Galf- U → 3 -α-D-Galp- Η3) -α-D-Glcp- (l-3) - β-D-Glcp-primer; and

- the β-1,5 osidic bond between the repeating saccnandiσues units of 1 EPS above; and

3) the recomoined and / or purifiable enzymes which are likely to be involved in the synthesis of EPS having the following repetitive saccharide unit: β-D-Gal α-L-Rha /? β-D-Gal /?

1 1 1

I I I

4 3 3

→ 3) -β-D-Gal / 7 (1 → 4) -α-D-Gal (1 → 2) -α-D-Gal /> (1 → 3) -β-D-Glcp (1 → and specifically catalyzing one of the following bonds between saccharides:

- the β-1,3 osidic bond between carbon 1 of an activated D-Galp and carbon 3 of the saccharide present at the non-reducing end of a saccharide chain in formation, said chain being composed in particular of a main chain of 1 to 4 D-Galp and / or D-Glcp linked together by α-1,2 osidic bonds , α-1,3, β-1,4 and / or β-1,3, and said saccharide present at the non-reducing end being linked to a single other saccharide by its carbon 1;

- the β-1,4 osidic bond between the carbon 1 of an activated D-Galp and the carbon 4 of the saccharide present at the non-reducing end of a saccharide chain in formation, said chain being composed in particular of a main chain from 1 to 4 D-Galp and / or D-Glcp linked to each other by α-1,2, α-1,3, β-1,4 and / or β-1,3 osidic bonds, and said saccnaπde present at the non-reducing end being linked to a single other saccharide by its carbon 1, and

- the α-1,3 osidic bond between the carbon 1 of an activated L- Rhap and the carbon 3 of the saccharide present at the non-reducing end α 'to a saccharide chain in formation, said chain being composed in particular of a main chain from 1 to 4 D-Galp and / or D-Glcp linked to each other by α-1,2, α-1,3, β-1,4 and / or β-1,3 osidic bonds, and said saccharide present at the non-reducing end being linked to a single other saccharide by its carbon 1.

Another subject of the invention relates to any recombinant nucleotide sequence coding for the enzyme (s) of the invention, in particular any vector (including any vector of shuttle type and capable of being expressed in different cell lines), and / or any cell incorporating said nucleotide sequences and expressing the recombinant functional enzyme (s) of the invention. Another aspect of the invention relates to the process for obtaining such an EPS in which:

(1) the nucleotide sequence of the invention is cloned into a vector, in particular a DNA fragment coding for at least one of the enzymes of the invention, said vector further comprising a sequence allowing autonomous replication or l integration into one or more (shuttle vector) host cell (s),

(2) the host cell is transformed by said vector, said host cell using the enzyme (s) encoded by the vector, where appropriate in combination with other enzymes produced by the host cell, for biosynthesis said exopolysaccnaπde,

(3) the transformed host cell is then cultured under conditions suitable for obtaining said exopolysaccharide.

Another subject of the invention relates to a process for the synthesis of an exopolysaccharide which consists of the repeated assembly of a single saccharide unit, itself formed of a main chain of saccharides whose ends are involved in the polymerization units, in which several enzymes or several recombinant nucleotide sequences are used according to the invention allowing the creation on the main chain of the saccharide unit of at least one lactose branching linked to a saccharide of the main chain by its glucose.

Another object of the invention relates to a new branched (branched) polysaccharide, constituted by the repeated assembly of a single saccharide unit, itself formed of a main chain of saccharides whose ends are involved in the polymerization of the units , characterized in that the main chain comprises at least one lactose branch linked to a saccharide of the main chain by its glucose, the galactose of this branching being also linked to an N-acetylgalactosam, provided that this polysaccharide does not consist of the repetition of the following saccharide unit: β-D-Gal / _? Nac- (l → 4) -β-D-Gal / (l → 4) -β-D-Glc _{/ 7}

71

3

→ 5) -β-D-Gal- (l → 3) -α-D-G - (l → 3) -α-D-Glcp- (l → 3) -β-D-Glcp (l-

The enzymes and sequences of the invention are preferably enzymes or sequences isolated, purified or recombined from strains of food microorganisms such as Lactobacillus, Streptococcus, etc., in particular non-pathogenic strains.

The present invention also relates to food products such as yoghurts or cheeses incorporating the transformed cells of the invention and / or the saccharides obtained from such cells.

A final aspect of the present invention relates to the use of one of the enzymes or of one of the recombinant nucleotide sequences of the invention for the synthesis of an exopolysaccharide.

The present invention will be described in detail in the nonlimiting exemplary embodiments presented below with reference to the appended figures.

Brief description of the figures

FIGS. 1 to 3 represent the physical map of the operon involved in the synthesis of an exopolysaccharide of the S strains, respectively. thermophilus CNCM 1-1879, CNCM 1-1449 and CNCM 1-800. The horizontal arrows indicate the presence of potential reading frames (ORF). The names of the genes corresponding to the ORFs are indicated below the arrows.

Detailed description of the invention

In the following description, the nucleotide sequences of the genes and primers used, as well as the amino acid sequences of the recombinant enzymes of the invention are shown in the sequence list provided below.

The term "EPS" denotes a polymer formed by the condensation of a large number of saccharide molecules (= oses) of different types. This EPS is more particularly constituted by the repeated assembly of a single saccnaridic unit, itself formed of a main chain of saccnaπdes whose ends are involved in the polymerization of the units. This main chain can also comprise branches (or branches) of saccnarides which are not involved in the polymerization of the units, said branches being grafted by osidic bonds on one or more saccharides of this chain. A saccharide unit is represented by indicating:

(1) the usual acronyms of saccharides (Glc: glucose; Gai: galactose; Rha: rhamnose; etc.);

(2) the type of each sugar link in the form of an arrow between the carbon numbers of two adjacent saccharides (the carbon numbers assigned on the cycle of a saccharide are those recognized by the international saccharide nomenclature: IUPAC) ; (3) the anomerism of each osidic bond in the form of the acronym α or β, said acronym being positioned to the left of a saccharide to indicate the anomerism of the osidic bond present to the right of this saccharide; (4) the nature pyranose or furanose each saccharide ad _j anointing p letter f or just after the initials of each saccharide; and (5) the D or L series of each saccharide, by positioning the letter to the left of the saccharide concerned. In a saccharide chain in formation, the direct reaction between a semi-acetal function of a new saccharide unit and an alcoholic hydroxyl of the saccharide present at the non-reducing end of the chain does not take place spontaneously, and it is why the semi-acetal group must first be activated, that is to say made more reactive for the formation of the osidic bond. This activation is catalyzed by bacterial enzymes forming a nucleotide saccharide, and more particularly a uridme-diphosphate of D-Glcp (UDP-Glcp), D-Galp (UDP-Galp) or D-Galf (UDP-Galf) in the context of the present invention. To initiate the formation of a saccharide chain, a first activated saccharide must first be linked to the phosphate of a lipophilic or protein primer. The addition of a new activated saccharide unit on the chain will then be done by its semi-acetalic function at the level of an alcoholic hydroxyl of the saccharide present at the non-reducing end of this chain, that is to say say at the opposite end from where the primer is attached.

Studies carried out on the enzymes of the invention have made it possible to show that the biosynthesis of exopolysaccharides takes place gradually, with each step, the addition of a new saccharide unit which attaches by its semi-acetal function to an alcoholic hydroxyl of another saccharide unit, which is at the end of a chain of saccharides linked to a primer. These enzymes specifically catalyze or control the following functions:

(1) The bond in the form of an α or β isomer between carbon 1 (which carries the aldehyde reducing function) of an activated D-Galp and a phosphate of a lipophilic primer in particular of the genus undecaprenol, or protein in particular of like ACP (Acyl Carrier Protein), for example.

(2) The α-1,3 osidic bond between carbon 1 of an activated GalpNAc (UDP-GalpNAc) and carbon 3 of D-Galp linked to its primer.

(3) The β-1,3 osidic bond between carbon 1 of an activated D-Glcp (UDP-D-Glcp) and carbon 3 of the saccharide present at the non-reducing end of the β-D- chain Galp (l— »3) -α / β-D-Galp-primer. (4) The α-1,6 osidic bond between carbon 1 of an activated D-Galp and carbon 6 of the saccharide present at the non-reducing end of the β-D- Glcp (l-3) β chain -D-Galp (l → -3) -α / β-D-Galp-primer.

(5) The transport of repetitive sacchaπdic units of the EPS described above outside the bacterial cell.

(6) The β-1,3 osidic bond between the repetitive saccharide units of the above EPS.

(7) Regulation of the number of saccharide units in EPS, and therefore of its final molecular weight.

In the following description, a saccharide chain is formed, in indicating the usual acronyms of saccharides, the types of osidic bonds, the position of the primer on the chain, the nature of the cycle and of the D or L series of each saccharide. Finally, since it is not yet known which isomer is formed when the first saccharide in a chain is linked to a primer, for convenience of writing, the acronym "α / β" is thus used in order to indicate that one of the two isomers is formed.

For the purposes of the present invention, the term "homologous sequence" means any nucleotide or amino acid sequence, leading or giving an enzyme variant which catalyzes the assembly of saccharides, differing from the sequences according to the invention only by the substitution, deletion or addition of a small number of nucleotide oases with amino acids, for example 1 to 150 base pairs (bp) or 1 to 50 amino acids. These enzyme variants can retain the same assembly specificity of saccharides as the enzymes from which they are derived. The substitution, deletion or addition of a small number of amino acids can however also lead to a change in the enzyme specificity. For example, the proteins coded by the genes of the invention epsl to eps5, even though they are very close to the enzymes EpsA-D of the strain S. ther ophilus Sfι6 (EP-A-0750043) and the EpsE enzyme from S. salivarius, have different enzymatic specificities from the proteins EpsA-D and EpsE (see below).

The selection of a homologous sequence is within the reach of those skilled in the art, by implementing conventional mutagenesis methods, and in particular by adapting the methods described by Adams et al. (EP-A-0402450; Genencor), by Dunn et al. (Protein Engineering, 2, 283-291, 1988), by Greener et al. (Strategies, 7, 32-34, 1994) and / or by Deng et al. (Anal. Biochem, 200, 81, 1992), for example.

In this context, two nucleotide sequences which, due to the degeneration of the genetic code, code for the same polypeptide will be considered as homologs. Similarly, it is also considered homologous two nucleotide sequences s' stπngentes hybridizing under conditions, that is to say hybridizing tou _j bear after a step of annealing at 65 ° C for 15h in a buffer hybridization (SSPE 1, 5x: 0.225 M NaCl, 0.0015 M EPTA, 0.015 M NaH ₂ P0 ₄ , pH7; 1% SDS and 1% dehydrated milk, followed by 6 successive washing steps at 65 ° C in a buffer comprising different dilutions of SSC IGx (1.5 M NaCl, 0.15 M sodium citrate, pH7) and 0.1-s SDS, respectively for 2 times 10 min with 2x SSC, for 2 times 10 mm with SSC lx, and for 2 times 5 mm with SSC 0, lx or as described by Samorook et al. (§§ 9.47-9.51 m Molecular Cloning. A Lahora tory Manual, Cold Spr g Harbor , Laboratory Press, Cold Sprmg Harbor, New York (1989)).

One can also consider as homologous the nucleotide sequences having more than 50% of identity with the sequences according to the invention, in particular more than 70%, the identity being determined by the ratio between the number of nucleotide bases of a sequence homologous which are identical to those of a sequence according to the invention, and the total number of nucleotide bases of said sequence according to the invention, for example.

Similarly, two functional proteins which are recognized by the same antibody will be considered as homologs, the ratio of the values of intensity of recognition of the two proteins by the antibody not exceeding 1000 in an ELISA test, preferably 100, for example. The amino acid sequences which have more than 70% identity with the sequences according to the invention, in particular more than 80%, 85%, 90% or more particularly more than 95%, will be considered more particularly as homologs. In the latter case, the identity is determined by the ratio between the number of amino acids of a homologous sequence which are identical to those of a sequence according to the invention, and the total number of amino acids of said sequence according to the invention.

To isolate a nucleotide sequence of the invention, in particular a recombinant DNA, it is possible to constitute a pan with a large DNA fragments of a lactic acid bacterium producing EPS in a lactic bacterium producing no EPS, then of select the clone (s) producing EPS For this, the genomic DNA of a lactic acid bacterium producing EPS is digested by a restriction enzyme which is specific for a relatively rare restriction site (BamHI, Sali, PstI ) or by partial digestion with Sau3A, for example. The digestion product is cloned into an expression or integration plasmid which accepts large fragments (plasmid pSA3, Dao et al., Appl. Environ. Microbiol., ___, 115-119, 1985), the plasmids are introduced recomomants in the same species of lactic acid bacteria not producing EPS, at least one transformed clone producing EPS is selected, then the DNA fragment responsible for the production of EPS is identified, isolated and sequenced.

Since the nucleotide sequences of the invention, in particular DNAs, are likely to be large, and that they may contain a group of genes involved in the biosynthesis of EPS, it may be preferable to introduce the recombinant plasmids into the same strain of lactic acid bacteria from which fragments, with the difference that this strain has lost the ability to produce EPS following mutagenic treatment (UV, chemical or transposon treatment). An alternative consists in constituting a plasmid library of DNA fragments of a strain of lactic acid bacteria producing an EPS, in transforming the same strain of lactic bacteria by plasmids incapable of replicating therein, in selecting the transformants having integrated a plasmid in their genome by homologous recombination (selection by resistance to an antibiotic, for example), to select the transformants no longer producing EPS, then to isolate and sequence the genomic DNA fragments of the selected transformants which are adjacent to the integrated plasmid . For this, the chromosome of the transformants can be digested, ligated, then carry out reverse PCR using probes specific for the integrated plasmid or introduce the ligation product into a strain in which the recircularized plasmid is capable of replicating, for example Another alternative consists in transforming lactic acid bacteria producing EPS by a plasmid comprising a transposon, in subjecting the bacteria to conditions in which the transposon excises from the vector and integrates at random in the genome, in selecting the clones of bacteria having lost the capacity to produce EPS, in isolating the fragments of genomic DNA from said clones in which a transposon has integrated. This method is described in more detail by Stmgele et al. , Dev. Biol. Stand., In Genetics of Streotococci, Enterococci and Lactococci, 8_5, 487-493, 1995 (ISSN: 0301-5149).

A last alternative consists in preparing nucleotide primers derived from known genes involved in the biosynthesis of an EPS, then in carrying out a PCR on the genomic DNA of a lactic acid bacterium producing an EPS and for which it is desired to identify the genes involved in the biosynthesis of this EPS. This technique is however excessively difficult to implement, since the choice of the nucleotide sequence of the primers will be decisive in the isolation of a gene. Since the homologies between the sought-after genes and the known genes are not known in advance, this choice proceeds from an inventive step, due to the necessary selection made from a multiplicity of possible primers.

Some criteria can however be retained, such as:

(1) the selection of an enzymatic function often found during the biosynthesis of an EPS; (2) the selection of conserved nucleotide sequences by comparing genes of various bacteria coding for this enzymatic function;

(3) the selection of conserved sequences whose GC nucleotide content is as close as possible to the target organism, even if this means taking as a starting point the selection of the nucleotide sequences found in organisms phylogenically distant from l target organism (E. coli);

(4) the identification of a part in the conserved sequences which is naturally richer in GC nucleotides;

(5) the identification of a part in the conserved sequences which is slightly degenerate, that is to say coding for amino acids which are themselves coded naturally by 1 or 2 codons, such as methionm, tryptophan, glutamme or glutamic acid, for example; etc.

It should be noted that the selection methods described briefly above can be applied to all known lactic acid bacteria, in particular to mesophilic lactic acid bacteria such as ^for example Streptococcus cremoris, Streptococcus lactis, Lactobacillus casei subsp. casei and Lactobacillus sake, and thermophilic lactic acid bacteria such as, for example, Streptococcus thermophilus and Lactobacillus helveticus. For this purpose, the skilled person has transformation techniques for each species of lactic acid bacteria.

In addition, these selection methods most often make it possible to isolate only part of a gene or a group of genes involved in the biosynthesis of an EPS. Nevertheless, a person skilled in the art can easily identify the remaining part of the gene or group of genes by selecting from a genomic library, using nucleotide probes based on an isolated fragment, one or more clones containing the remaining part, by example.

Preferably, the subject of the invention relates to the nucleotide sequence SEQ ID NO: 1 or its complementary strand, obtained from the Streptococcus thermophilus strain deposited on 20 July 1997 according to the Budapest Treaty with the National Collection of Culture of Microorganisms ( CNCM), 28 rue du Docteur Roux, 75724 Paris cedex 15, France (CNCM filing number 1-1897). This Gram-positive strain presents under the microscope an appearance of non-flagellated shells forming chains, does not make spores and is anaerobic.

This sequence SEQ ID NO: 1 comprises genes coding for new enzymes involved in the synthesis of EPS having the following repeated structure:

β-D-Galσl

6 → 3) -β-D-Glc /> - (l → 3) -β-D-Gal / - (1 → 3) -α-D-Glc / - (1 → Thanks to the knowledge acquired on the biosynthesis of EPS described in the literature, in particular of the EPS described in EP-A-0750043, it has been possible to identify the new enzymatic functions involved in the biosynthesis of EPS above. The new enzymes involved in the synthesis of this EPS can specifically catalyze one of the following new bonds between saccharides:

- the β-1,3 osidic bond between carbon 1 of an activated D-Galf and carbon 3 of Glcp present at the non-reducing end of a saccharide chain in formation, in particular the α / β-D chain -Glcp-primer;

- the β-1,3 osidic bond between the caroone 1 of an activated D-Glcp and the carbon 3 of Galf present at the non-reducing end of a saccharide chain in formation, in particular the β-D-Galf chain (1—3) -α / β-D-Glcp-primer;

- the β-1,6 osidic bond between the caroone 1 of an activated D-Galp and the carbon 6 of Glcp present at the non-reducing end of a saccnarid chain in formation, in particular the β-D-Glcp chain (1 → 3) β-D-Galf (1 → 3) -α / β-D- Glcp-primer; and,

- the α-1,3 osidic bond between the repetitive saccharide units of the EPS above.

The new recombinant enzymes have a low homology with known enzymes. These new enzymes, identified from the sequence SEQ ID NO: 1, have one of the amino acid sequences chosen from the group of sequences SEQ ID NO: 7, SEQ ID NO: 8, SEQ ID NO: 9, SEQ ID NO: 10, SEQ ID NO: 11 and the sequences which are homologous to them.

This sequence SEQ ID NO: 1 contains genes involved in the synthesis of this EPS (epsl -5), coding for proteins having very strong homologies with EpsA-D proteins from S. thermophilus Sfi6 (EP-A-0750043) and the EpsE protein from S. salivarius (GenBank: No. X94980). These high degrees of homology are unexpected in view of the fact that the structure of the EPS produced by the strain CNCM 1-1897 is sufficiently different from that of the strain Sfi6, so that one can expect that the enzymes of the CNCM 1-1897 strain are different from those of the Sfi6 strain.

The proteins encoded by these genes have new enzymatic specificities making them particularly attractive for the creation of new bonds between saccharides, or for varying the length of the EPS chain, or even for regulating the production of EPS differently. One of these new enzymes is a glycosyltransferase having specifically the amino acid sequence SEQ ID NO: 6, the homologous sequences being of course excluded for reasons of novelty. The other new enzymes are involved in the regulation of EPS production and chain length, and have specifically one of the amino acid sequences chosen from the group of sequences SEQ ID NO: 2, SEQ ID NO : 3, SEQ ID NO: 4 and SEQ ID NO: 5, the homologous sequences being of course excluded for reasons of novelty. The present invention also relates to any nucleotide sequence, in particular a recombinant DNA, coding for an enzyme according to the invention, that is to say one of the enzymes chosen from the group of sequences SEQ ID NO: 2, SEQ ID NO: 3, SEQ ID NO: 4, SEQ ID NO: 5, SEQ ID NO: 6, SEQ ID NO: 7, SEQ ID NO: 8, SEQ ID NO: 9, SEQ ID NO: 10 and SEQ ID NO: 11. In particular, this DNA can comprise at least one gene having one of the nucleotide sequences chosen from the sequences delimited in the sequence SEQ ID NO: 1 by nucleotides 1-687, 690-1418, 1429-2119, 2131-2870, 2929-4244, 5933-7978, 8003-8969, 10138-11086, 11438-12852, 14220-15315 and the sequences which are ^' homologous to the sequences delimited in the sequence SEQ ID NO: l by nucleotides 5933- 7978, 8003-8969, 10138-11086, 11438-12852 and 14220-15315.

According to a second preferred embodiment of the invention, the nucleotide sequence of the invention is the nucleotide sequence SEQ ID NO: 12 or its complementary strand, obtained from the strain Lactobacillus helveti cus LH59 which was deposited on July 27, 1994 according to the Budapest Treaty with the National Collection of Culture of Microorganisms (CNCM), 28 rue du Docteur Roux, 75724 Paris cedex 15, France (CNCM deposit number 1-1449). This Gram-positive strain presents under the microscope an aspect of unflagged rods, does not make spores and is anaerobic optional.

This sequence SEQ ID NO: 12 comprises genes coding for new enzymes involved in the synthesis of EPS having the following repeated structure:

β-D-Gal / - (l → 4) -β-D-Glcp 1

3 κ5) -β-D-Gal / ^" - (l → 3) -α-D-Gal /> - (l → 3) -α-D-Glc / 7- (l → 3) -β-D- Glcp- (l-

These new enzymes involved in the synthesis of this EPS can thus specifically catalyze one of the following new bonds between saccharides:

- the α-1,3 osidic bond between carbon 1 of an activated D-Glcp and carbon 3 of a Glcp belonging to the non-reducing end of a saccharide chain in formation, in particular the following chain: α / β-D-Glcp- primer;

- the α-1,3 osidic bond between carbon 1 of an activated D-Galp and carbon 3 of a Glcp present at the end non-reducing of a saccharide chain in formation, in particular the following chain: α-D-Glcp- (l-3) -β / α-D- Glcp-primer;

- the β-1,3 osidic bond between carbon 1 of an activated D-Galf and carbon 3 of a Galp present at the non-reducing end of a saccharide chain in formation, in particular the following chain: α -D-Galp- (1 → 3) -α-D-Glcp- (1 → 3) -β / α-D-Glcp-primer;

- the β-1,3 osidic bond between carbon 1 of an activated D-Glcp and carbon 3 of a Galf present at the non-reducing end of a saccharide chain in formation, in particular the following chain: β -D-Galf- (1 → 3) -α-D-Galp- (1-> 3) -α-D-Glcp- (1 → 3) -β-D-Glcp-primer;

- the β-1,4 osidic bond between carbon 1 of an activated D-Galp and carbon 4 of Glcp present at the non-reducing end of a saccharide chain in formation, in particular the following chain: β-D -Glcp- (1 → 3) -β-D-Galf- (1 → 3) -α-D-Galp- (1 → 3) -α-D-Glcp- (1 → 3) -β-D-Glcp -bait; and, - the β-1,5 osidic bond between the repetitive saccharide units of the EPS above.

The new recombinant enzymes, identified from the sequence SEQ ID NO: 12, have one of the amino acid sequences chosen from the group of sequences SEQ ID NO: 13, SEQ ID NO: 14, SEQ ID NO: 15, SEQ ID NO: 16, SEQ ID NO: 17, SEQ ID NO: 18, SEQ ID NO: 19, and the sequences which are homologous to them.

The present invention also relates to any nucleotide sequence, in particular any recombinant DNA coding for an enzyme according to the invention, that is to say the enzymes chosen from the group of sequences SEQ ID NO: 13,

SEQ ID NO: 14, SEQ ID NO: 15, SEQ ID NO: 16, SEQ ID NO: 17, SEQ ID NO: 18, SEQ ID NO: 19. In particular, this DNA can comprise at least one gene having one of the nucleotide sequences chosen from the sequences delimited in the sequence SEQ ID NO: 12 by nucleotides 1053-1730, 1734 -2849, 2852-3943, 3930-5084, 5077-6096, 6099-7091, 7096-8259, 8284-8910 and the sequences which are homologous to them.

According to a third preferred embodiment of the invention, the nucleotide sequence of the invention is the nucleotide sequence SEQ ID NO: 20 or its complementary strand, obtained from the strain Lactobacillus delbrueckii bulgari cus Sfi5 deposited on October 4, 1988 according to Treaty of Budapest, with the National Collection of Culture of Microorganism (CNCM), 28 rue du Docteur Roux, 75724 Paris cedex 15, France (CNCM deposit number 1-800). Furthermore, this Gram-positive strain under the microscope has the appearance of non-flagellated rods, is non-sporulating and anaerobic optional.

This sequence SEQ ID NO: 20 comprises genes coding for new enzymes involved in the synthesis of EPS having the following repeated structure:

β-D-Gal /? α-L-Rha / 7 β-D-Gal 1 1 1

3 3

^> 3) -β-D-Galj-0> 4) -cx-D-Gal / - (1 → 2) -α-D-Gal />(1-> 3) -β-D-Glc /? ( l-

- the β-1,3 osidic bond between the carbon 1 of an activated D-Galp and the carbon 3 of the saccharide present at the non-reducing end of a saccharide chain in formation, said chain being composed in particular of a main chain from 1 to 4 D-Galp and / or D-Glcp linked between them by α-1,2, α-1,3, β-1,4 and / or β-1,3 osidic bonds, and said saccharide present at the non-reducing end being linked to a single other saccharide by its carbon 1; - the β-1,4 osidic bond between the carbon 1 of an activated D-Galp and the carbon 4 of the saccharide present at the non-reducing end of a saccharide chain in formation, said chain being composed in particular of a main chain from 1 to 4 D-Galp and / or D-Glcp linked to each other by α-1,2, α-1,3, β-1,4 and / or β-1,3 osidic bonds, and said saccharide present at the non-reducing end being linked to a single other saccharide by its carbon 1; - the α-1,3 osidic bond between the carPone 1 of an activated L-Rhap and the carbon 3 of the saccharide present at the non-reducing end of a saccharide chain in formation, said chain being composed in particular of a main chain from 1 to 4 D-Galp and / or D-Glcp linked to each other by α-1,2, α-1,3, β-1,4 and / or β-1,3 osidic bonds, and said saccharide present at the non-reducing end being linked to a single other saccharide by its carbon 1.

The saccharide chain in formation can be composed simply of a main chain without branching from 1 to 4 D-Galp and / or D-Glcp linked together by osidic bonds α-1,2, α-1,3, β -1.4 and / or β-1-3.

For example, this channel can be selected from tetra-sacchridiques following channels, as well as their precursors mono-, di- or tπ- saccharide (s) ^{• α-D-Gal} / 1 → 3) -β-D -Glcp (1-> 3) -β-D-Gal / X1 → 4) -α / β-D-Gal -amorce, α-D-Gal /? (1 → 2) - -O-Ga \ p ( 1 → 3) -β-D-Glc /? (1 → 3) -β / α-D-Gal / -bait, β-DG (l → 4) -α-D-Gal 7 (l → 2) -α-D-Gal / - (l-> 3) -β / α-D-Glcp-primer; and β-D-Glc (l → 3) -β-D-Gal (l → 4) -α-D-Gal (l → 2) -β / α-D-Gal /? - primer.

The saccharide chain in formation can also be composed of a main chain further comprising 1 or 2 branched saccharides, such as D-Galp and / or L-

Rhap. These branched saccharides can be linked on the main chain in β-1,3, β-1,4 or α-1,3, for example.

By way of example, this chain can be chosen from the following main tetra-sacchridic chains, as well as their mono-, di- or tri-saccharide precursors: -D-Galp (1 → 3) -β-D -Glcp (1 → 3) - [β-D-Galp (1 → 4)] - β-D-Galp (1 → 4) - [α-LRhap (1 → 3)] - α / β-D-Galp -bait; α-D-Galp (l → 2) - [β-D-Galp (l → 3)] - α-D-Galp (l → 3) -β-D-Gl (l-3) - [β-D -Galp (l → 4)] - β / α-D-Gal -amorce; β-D-Galp (l → 4) - [-Rhap (l → 3)] - -D-Gal / 7 (l → 2) - [β-D-Galp (l → 3)] - -D-Galp (l → 3) -β / α-D-Glcp-primer; and β-D-Glcp (1 → 3) - [β-D-Galp (1 → 4)] - β-D-Gal /? (1 → 4) - [-LRhap (1 → 3)] - α-D-Galp (1 → 2) - [β-D-Galp (1 → 3)] - β / α-D-Galp-primer.

Preferably, one of the new enzymes according to the invention catalyzes the α-1,3 osidic bond between the carbon 1 of an activated -Rhap and the carbon 3 of galactose present at the non-reducing end of one of the saccharide chains. above.

Preferably, one of the new enzymes according to the invention catalyzes the β-1,3 osidic bond between the carbon 1 of an activated D-Galp and the carbon 3 of the galactose present at the non-reducing end of one of the chains. saccharides above.

Preferably also, one of the new enzymes according to the invention catalyzes the β-1,4 osidic bond between the carbon 1 of an activated D-Galp and the carbon 4 of galactose present at the non-reducing end of one of the saccharide chains above. The new recombinant enzymes, identified from the DNA sequence SEQ ID NO: 20, have one of the amino acid sequences chosen from the group of sequences SEQ ID NO: 21, SEQ ID NO: 22, SEQ ID NO : 23, SEQ ID NO: 24, SEQ ID NO: 25, SEQ ID NO: 26, SEQ ID NO: 27, SEQ ID NO: 28, SEQ ID NO: 29, and the sequences which are homologous to them.

The present invention also relates to any nucleotide sequence, in particular any recombinant DNA coding for an enzyme according to the invention, that is to say one of the enzymes chosen from the group of sequences SEQ ID NO: 21, SEQ ID NO: 22, SEQ ID NO: 23, SEQ ID NO: 24, SEQ ID NO: 25, SEQ ID NO: 26, SEQ ID NO: 27, SEQ ID NO: 28 and SEQ ID NO: 29. In particular, this DNA can comprise at least one gene having one of the nucleotide sequences chosen from the sequences delimited in the sequence SEQ ID NO: 20 by nucleotides 9-353, 516-1292, 1323-1982, 2074-3174, 3246-4373, 4511- 5317. 5335-6204, 6709-7692, 7723-8784, and the sequences which are homologous to them. Unexpectedly, the inventors identified on certain sequences of the invention, in particular on the sequence SEQ ID NO. 1, IS portions arranged respectively upstream and downstream of the sequences coding for the enzymes of the invention. These IS sequences advantageously make it possible to directly use the sequence of the invention as a "transposon" and optionally to practice a coculture of microorganisms without a vector and to obtain the expression of these enzymes in cells having lost the threading character initial, which allows the synthesis of exopolysaccharides by said transformed cells.

The present invention also relates to any recombinant vector comprising the nucleotide sequence of the invention. This recombinant vector can be any fragment nucleic acid, such as a single-stranded or double-stranded DNA, linear or circular, of expression or of ^" integration", and comprising a nucleotide sequence of the invention, in particular all or part of the sequence SEQ ID NO: 1, SEQ ID NO: 12 or SEQ ID NO: 20. In the case where the method described in EP564966 is not used (see below), care must be taken that the vector expresses the DNA according to the invention by suitable nucleotide sequences (promoter; ribosome attachment site; codon preferred by the host organism, etc.), and where appropriate that it includes one or more origins of cell replication, in particular of Escheπchia coli and / or a Streptococcus, to form the shuttle vector of the invention. Thus, to carry out the biosynthesis of an EPS, it is possible to integrate all or part of the nucleotide sequence of the invention comprising at least one of the above-mentioned genes in a host cell using the method described in EP-A-0564 966, said method being incorporated by reference into the teaching of the present invention. In summary, this process allows you to:

(1) transforming the host cell with a donor plasmid which does not replicate therein, said plasmid comprising said fragment placed functionally (the reading frame is preserved) in part of an operon derived from the host cell;

(2) identify the transformants having integrated the entire plasmid;

(3) selecting transformants having only integrated into the chromosome the fragment according to the invention, the other sequences of the plasmid being excised from the chromosome; and

(4) cultivating the selected transformants under conditions suitable for the production of an EPS. It can be noted that this method makes it possible not to use functional promoter and translational activation sequences. In addition, the culture conditions suitable for the production of EPS are within the reach of those skilled in the art, who can use standard culture media, and choose the optimum pH, temperature and agitation of the medium according to the strain used.

One can also choose to clone all or part of the sequence SEQ ID N0: 1, SEQ ID NO: 12 or SEQ ID NO: 20, comprising at least one of the above genes in a self-replicating expression plasmid downstream of sequences promoter and functional translational activation, and if necessary upstream of a terminator, then transforming a host cell with the recombinant plasmid. Recombinant enzymes can also be used to modify or synthesize in-vi tro an oligosaccharide or a polysaccharide such as EPS, for example. For this, it is preferable to purify at least one of these enzymes, by classically over-expressing their gene in a cell, and by classically isolating them, by precipitation and / or chromatography (from the culture medium and / or from the branch part). cells), for example.

Another object of the present invention relates to a cell comprising, integrated into its genome or by means of a replicable plasmid, a nucleotide sequence of the invention, and expressing a functional recombinant enzyme according to the invention. This cell can be part of the group of molds, yeasts, bacteria and plants, for example. Preferably, the yeasts belong to the genera Saccharamyces, Kluyveromyces, Hansenula and Pichia; the bacteria are Gram-negative or positive belonging to the genus Escherichia, Bacillus, Lactobacillus, Lactococcus, Streptococcus and Staphylococcus; plant cells belong to legumes, and cereal and woody species; while molds are the cells traditionally used to prepare a koj 1 like Aspergillus, Rhizopus and / or Mucor. Another object of the present invention relates to a method for producing an EPS, in which:

(1) a fragment of a nucleotide sequence coding for at least one of the enzymes according to the invention is cloned into a vector, said vector further comprising a sequence allowing autonomous replication or integration into a host cell;

(2) a host cell is transformed by said vector, said host cell using the enzymes encoded by the vector, where appropriate in combination with other enzymes produced by the host cell if the vector does not provide all the enzymes necessary for the biosynthesis of an EPS, for the biosynthesis of an EPS; and

(3) the transformed host cell is then cultured under conditions suitable for the production of EPS. If the method described in EP-A-0564966 is not implemented, the vector must also comprise at least one functional promoter sequence and at least one functional translational activation sequence, allowing the expression of the genes encoding for at least one of the enzymes according to the invention.

The present invention also relates to a new process for the synthesis of an EPS, which consists of the repeated assembly of a single saccharide unit, itself formed of a main chain of saccharides whose ends are involved in the polymerization of the units, in which several recombinant enzymes or several recombinant genes capable of being involved in the biosynthesis of EPS having the repetitive saccharide unit are used: β-D-Gal /? - (l → 4) -β-D-Glc /> 1

3 → 5) -β-D-Gal ^" - (l → 3) -α-D-Galp- (l → 3) -α-D-Glcp- (l → 3) -β-D-Glcp- (l →

characterized in that these enzymes or these genes, in particular chosen from the enzymes or the genes according to the invention, allow the creation on the main chain of the saccharide unit of one or more lactose branches (β- D-Galp- ( 1— »4) -β-D-Glcp) linked to a saccharide of the main chain by their glucose.

Preferably, the main chain of the saccharide unit consists of at least 2 saccharides chosen from D-Galp, D-Galf and / or D-Glcp. In addition, the carbon 1 of the glucose of the lactose ramifications can be linked by a β sugar linkage to one of the carbons 2-6 of one of the saccharides of the main chain, for example, preferably carpone 3 or 6 d 'a D-Galf.

During the implementation of this process, an enzyme or a gene which allows the link between an N-acetyl-galactosam and the galactose of the lactose branching can be used. It is also possible to use in addition an enzyme or a supper which allows the bond between the N-acetyl-neuraminic acid and the galactose of the lactose branching, for example.

This synthesis process can be carried out in vi tro, using the recombinant enzymes and their appropriate substrates. It can also be implemented in vivo, by cloning the appropriate genes and making them express themselves so that the recombinant enzymes thus produced biosynthesize an EPS in the presence of their substrates, for example.

The present invention also relates to a new branched polysaccharide, constituted by the repeated assembly of a single saccharide unit, it- even formed from a main chain of saccharides whose ends are involved in the polymerization of the units, characterized in that the main chain comprises aα at least one lactose branching linked to a saccharide of the main chain by its glucose, the galactose of this branching being further bound to an N-acetylgalactosam.

The carbon 1 of the glucose of the lactose branching can be linked by a β-sugar linkage to one of the carbons 2-6 of one of the saccharides of the main chain, preferably the carbon 3 or 6 of a D-Galf, for example.

For reasons of novelty, this polysaccharide cannot be constituted by the repetition of the following saccharide unit:

β-D-GalpNac- (l → 4) -β-D-Gal /? - (l → 4) -β-D-Glc />

1 71 3

→ 5) -β-D-Gal- ^" (l → 3) -α-D-Gal ^ - (l → 3) -α-D-Glc /? - (l → 3) -β-D-Glcp ( l → Indeed, patent NZ-272795 (Company of

Nestlé Products) already recommends the enzymatic addition of an N-acetylgalactosam on the galactose of the branches contained in the EPS of the LH59 strain.

To obtain this new polysaccharide, it suffices for example to implement the synthesis process described above so as to obtain an EPS different from that produced by the strain LH59, then to use a β-1, 4-N- acetylgalactosammyltransferase in the presence of its substrate, in particular that described by Lutz et al. and Yamashiro et al. (J. of Biol. Chem., 269, 29227-29 231, 1994; J. of Biol. Chem., 270, 6149-6155, 1995). It is also possible to purify an EPS from a culture medium different from that produced by the strain LH59, then to add to it an N-acetylgalactosamme on the galactose of the branches. lactose. The EPS produced by the L. helveticus TY1-2 strain can thus represent a target of choice (Y. Yamamoto et al., Carbohydrate Research, 261, 67-78, 1994), for example. This EPS is constituted by the repetition of the following saccharide unit:

β-D-Gal - (1 → 4) -β-D-Glc /> 1 i 6

→ 6) -β-D-Glcp- (1 → 3) -β-D-Glc - (1 → 6) -α-D-Glc /> Nac- (1 → 3) -β-D-Galp (1 →

1

4 β-D-Gal /?

The new polysaccharide according to the invention can also comprise an N-acetyl-neurammic acid linked to the galactose of the lactose branching. For this, one can use an α-2, 6-sιalyltransferase, in particular those described by Simth et al. , Sjoberg et al. and / or Yamamoto et al. , (J. of Biol Chem., 269, 15162-15171, 1994; J. Biochem., 271, 7450-7459, 1996; Biosci. Biotech. Biochem., 6_2, 210-214, 1998), or an α- 2,3-sialyltransferase, in particular that sold by Calbiocnem® (Cat No.566218-M, USA), for example.

The present invention is described in more detail below with the aid of the additional description which follows, which refers to examples of obtaining DNA fragments, recombinant plasmids and bacteria transformed according to the invention. These examples are preceded by a description of the culture media. It goes without saying, however, that these examples are given by way of illustration of the subject of the invention of which they do not in any way constitute a limitation. DNA manipulation, cloning and transformation of bacterial cells are, in the absence of details contrary, carried out according to the protocols described in the work of Sambrook et al. (Molecular Cloning, A Laboratory Manual, Cold Spring Harbor Laboratory Press, USA, 1989). The percentages are given in weight / weight, unless otherwise indicated.

Media: (add 1.5% Bacto-agar for a solid medium)

- M17 (Difco, USA): tryptone 0.5%, soytone 0.5%, hydrolyzed meat 0.5%, yeast extract 0.25%, ascorbic acid 0.05%, magnesium sulphate 0.025%, disodium- beta-glycerophosphate 1.9% and water.

- LM17: M17 medium comprising 1% lactose. - GM17: M17 medium comprising 1% glucose.

- MSK: skimmed milk (10% reconstituted powder) comprising 0.1% yeast extract.

- HJL: tryptone 3%, beef extract 0.2%, yeast extract 1%, lactose 1% and KH ₂ P0 ₄ pH 6.5 0.5%. - Ruthenium red: yeast extract 0.5%, skimmed milk powder 10%, sucrose 1%, agar 1.5% and 0.08 g / l of ruthenium red (see FR-2632968).

- MRS: 1% peptone proteose, 1% beef extract, 0.5% yeast extract, 2% dextrose, 0.1% tween 80, 0.2% ammonium citrate, 0.5% ammonium acetate , di-potassium phosphate 0.2%, magnesium sulfate 0, lg / l, and manganese sulfate 0.05 g / 1.

- LB: 0.5% tryptone, 0.25% yeast extract, and 1% NaCl. Examples

Example I: Cloning of the operon of S. thermophilus CNCM 1-1897 involved in the synthesis of an EPS

1. Selection of an S strain. thermophil us producer of EPS

The strains of lactic acid bacteria from the Nestlé collection are cultured in an HJL liquid medium and dilutions are spread on a Ruthenium Red solid medium. The EPS producing strains remain white because the EPS prevents the dye from tinting their cell wall. On the other hand, the non-producing strains stain red due to the affinity of the dye for the peptidoglycan of their cell wall. The S. thermophilus Sfi39 strain, which has received the deposit number CNCM 1-1897 and which will be designated in the following examples by the expression "strain Sfi39", has thus been selected from the EPS-producing lactic acid bacteria.

2. Repeated structure of EPS The structure of EPS produced by the strain

Sfi39 has already been presented in patent application ΞP-97111381.6, the technical content of which is incorporated by reference in the description.

3. Cloning

The approach attempted in order to isolate the eps genes from the Sfi39 strain is based on the comparison of known eps genes (those of S. thermophilus) and their counterparts (cap, eps, rfb genes) of various species, and even phylogenically unrelated species, such as Gram-negative species. These comparisons thus make it possible to define conserved regions between the species, and to use nucleotide primers derived from these conserved regions in order to amplify by PCR an internal part of the eps operon. of the strain Sfι39. This approach unfortunately encountered the impossibility of amplifying an eps gene, whatever the primers chosen. The reasons leading to these failures can be multiple, for example linked to a weak homology between the eps genes of the strain Sfι39 and those of the conserved regions, or to inadequate PCR conditions, etc.

Faced with these failures, we decided to select, then use in a PCR, particularly degenerate nucleotide primers. The choice of these specific primers was in fact decisive for the amplification of an eps gene of the strain Sfι39, and was motivated by the following considerations:

(1) the selection of an enzymatic function often found during the biosynthesis of an EPS;

(2) selection of conserved nucleotide sequences by comparing genes of various bacteria encoding this enzyme function;

(3) the selection of conserved sequences whose GC nucleotide content is as close as possible to the target organism, even if this means taking as a starting point the selection of the nucleotide sequences found in organisms phylogenically distant from l target organism (E. coli); (4) the identification of a part in the conserved sequences which is naturally richer in GC nucleotides;

(5) the identification of a part in the conserved sequences which is not very degenerate, that is to say coding for amino acids which are themselves coded naturally by 1 or 2 codons, such as methionm, tryptophan, glutam e or glutamic acid, for example; etc. Among the primers tested, those having the nucleotide sequences SEQ ID NO: 39 and SEQ ID NO: 40, made it possible to amplify an eps gene fragment under the following PCR conditions: successively 2 min at 95 ° C., 1 min at 42 ° C, 20 s at 72 ° C, then 35 times in succession the 30 s cycle at 94 ° C, 1 min at 42 ° C and 20 s at 72 ° C. For this, the genomic DNA of the strain Sfi39 was previously isolated according to the technique of Slos et al. (Appl. Environ. Microbiol., 5_7, 1333-1339, 1991). About 100 ng of genomic DNA and Taq polymerase are used.

A 143 bp PCR fragment was thus isolated, then cloned into the linearized plasmid pGEMT (Promega, USA). The sequencing of this fragment by the dideoxynucleotide method (f-mol® DNA Sequencing System kit, Promega) indicates a sequence corresponding to nucleotides 4006 to 4149 of the sequence SEQ ID NO: 1.

This PCR fragment was used to screen a λ-ZAP Express library (Stratagene, USA) containing DNA fragments of the Sfi39 strain. For this, according to the supplier's recommendations, a DNA preparation of said mutant is partially digested with Sau3A, the fragments are separated by agarose gel electrophoresis, the bands corresponding to fragments of 5 to 12 kb are cut from the gel, eluted the DNA, then it is ligated to the λ-ZAP Express vector previously digested with BamRI. The ligation product is encapsulated using the GigagoldlII system (Stratagene), the phages are then mixed with Escheri chia coli XLIBlue (Stratagene) according to the supplier's recommendations, then the mixture is spread on a Petri dish. . The recombinant plaques are then analyzed by hybridization of their DNA transferred to a Hybond-N membrane (Amersham Life Sciences, UK) with the PCR fragment previously made radioactive (Random Pπmed DNA Label g kit, Boehr ger-Manheim).

Among the numerous recombinant plaques, it was possible to select by hybridization several positive plaques, from which the λ-ZAP Express vectors were then isolated, then excised the pBK-PCMV (pCMV) vectors containing the genomic insert (see the recommendations of the supplier Stratagene ). The genomic mserts of these pCMV vectors were then sequenced (f-mol® DNA Sequencmg System kit). By cutting the nucleotide sequences of the different genomic mserts, it was thus possible to characterize the sequence SEQ ID NO: 1 corresponding to the operon of the strain Sfι39 involved in the synthesis of an EPS (see FIG. 1). This sequence also includes, upstream and downstream of its coding sequences, IS sequences allowing the direct use of the sequences of the invention in cransfection by coculture without vector.

4. Analysis of the sequence SEQ ID NO: 1 The sequence SEQ ID NO: 1 has 1 operon eps of the strain Sfι39 and comprises 10 complete ORFs, in the same orientation, which is called epsl, eps2, eps3, eps4, eps5, epsβ, eps7, eps8, eps9 and epsl O (see Figure 1). The comparison of the amino acid sequences encoded by the ORFs with those of proteins present in the Swiss-Prot database, using the FASTA, PEPPLOT and PILEUP software from GCG-soft ear, isconsm, USA, makes it possible to deduce the function of these proteins. The results are presented below. ORFs epsl (nucleotides 2703-3389), eps2

(nucleotides 3390-4118), eps3 (nucleotides 4130-4819), eps4

(nucleotides 4832-5527) and eps5 (nucleotides 5629-6993) have more than 80% identity with the genes epsA, epsB, epsC and epsD of the S. thermophilus Sfi6 strain described in EP750043 (Genbank, n ° U40830), and with the epsΞ gene of S. salvarius (GenBank, n ° X94980). These ORFs respectively code for proteins having more than 90% identity with the proteins EpsA, EpsB, ΞpsC, EpsD and EpsE of S. thermophilus and S. salivarius. These high degrees of identity are unexpected given the fact that the structure of the EPS produced by the strain Sfi39 is sufficiently different from that of the strain Sfi6, so that it can be assumed that the enzymes of the strain Sfi39 are different of those of the strain Sfi6. The proteins encoded by the epsl - 5 genes, although very close to the EpsA-D and EpsE proteins, however have new enzyme specificities (see above) making them particularly attractive for the creation of new saccharide bonds (eps5) , or to vary the length of the EPS chain or even to regulate EPS production differently (epsl -4). The ORFs epsl, eps2, eps3, eps4 and eps5 respectively code for proteins having the amino acid sequences SEQ ID NO: 2, SEQ ID NC: 3, SEQ ID NO: 4, SEQ ID NO: 5 and SEQ ID NO: 6.

The ORF eps6 (nucleotides 8633-10678) codes for a protein (SEQ ID NO: 7) having approximately 20% identity with the protein RfbC from Klebsi ella pneumaniae (Genbank, n ° L41518). This homology confirms its role in the biosynthesis of EPS.

ORF eps7 (nucleotides 10703-11668) codes for a protein (SEQ ID NO: 8) having approximately 33% identity with the Epsl protein of S. thermophilus Sfi6.

This homology indicates that the protein is obviously a beta-glycosyltransferase.

The epsβ ORF (nucleotides 12838-13785) codes for a protein (SEQ ID NO: 9) having approximately 30% identity with the Epsl protein of S. thermophilus Sfi6. This homology indicates that the protein is obviously a beta-glycosyltransferase.

The ORF eps9 (nucleotides 14138-15553) codes for a protein (SEQ ID NO: 10) having approximately 40% identity with the protein EpsK of Lactococcus lactis

(GenBank No. U93364). This homology indicates that the protein is probably the transporter of the repeating unit in the external environment. The ORF epsl O (nucleotides 16919-18013) codes for a protein (SEQ ID NO: 11) having approximately 62% identity with the Galf protein of Escheri chia coli

(GenBank No. U03041). This homology confirms its role in the biosynthesis of EPS, by catalyzing the conversion of the pyranose form to the furanose form of a β-D-Gal.

In conclusion, the genomic mserts isolated from the pCMV vectors cover a genomic region of the strain S. thermophilus Sfι39 which is manifestly involved in the biosynthesis of EPS.

Example II: mact vation of the eps genes

The epsl -10 gene of 1 Operon is inactivated by homologous recomomaison to confirm their importance in the biosynthesis of EPS. For this, a fragment of each ORF originating from one of the pCMV vectors described in Example I above is amplified by PCR, the PCR product is ligated into the heat-sensitive plasmid pSA3 previously digested, the strain E is transformed. . coli XLl-blue by the ligation product, transformants are selected, a recombinant plasmid is isolated, then the S. thermophilus Sfi39 strain is transformed by electroporation with the recombinant plasmid using a method adapted from that described by Slos et al. . (1991). We resuspend them cells subjected to a discharge of 2, lkV, 25μF and 400Ω in 1 ml of HJL medium which are incubated for 4 h at 37 ° C (permissive temperature), the cells are spread on a LM17 solid medium supplemented with 2.5 μg / ml of erythromycm which are incubated for 16 h at 37 ° C., then the transformed colonies which survive are selected. The selected colonies are then incubated in 2 ml of HJL medium supplemented with 2.5 μg / ml of erythromycm until the optical density at 600 nm (DOgoo) of the culture reaches 0.2, the culture is subjected to 45 ° C until the OD _CQQ reaches 1.0 (the plasmid no longer replicates), then dilutions of the culture are spread on a solid LM17 medium supplemented with 2.5 μg / ml of erythromycm than l '' incubate 12 h at 45 ° C. The surviving colonies integrated the recombinant pSA3 plasmid into one of 10 eps genes. This can be verified by Soutnern-Blot of a genomic DNA preparation of the surviving colonies digested with BcoRI (cut only once in pSA3), and hy πdation of the Southern-Blot filter with the PCR product made radioactive. The colonies which have integrated the plasmid pSA3 have different bands than those obtained under the same conditions with the wild strain. In addition, the colonies which integrated the recombinant pSA3 plasmid into one of the 10 eps genes exhibit an EPS (-) phenotype on a solid red Ruthenium medium, and have lost their threading character in an MSK milk.

Example III Biosynthesis of an EPS A plasmid is prepared from the plasmid pSA3 (Dao et al., Appl. Environ. Microbiol., ___, 115-119, 1985), by adding a DNA fragment of the strain Sfi39 containing the 17.2 kb sequence described in Example I (see the sequence list), and adding the regulatory parts of the eps operon of the S. thermophilus Sfi6 strain (EP750043), so that the expression (transcription and translation) of the eps genes of the Sfi39 strain can be correctly performed in S. thermophilus. The S. thermophilus CNCM 1-1422 strain which was deposited on May 18, 1994 according to the Budapest Treaty is then transformed by electroporation with this plasmid. Any other shooting strain of S. thermophilus could also have been used. This strain presents under the microscope an appearance of non-flagellated shells forming chains, it does not make spores, it is anaerobic optional, and it produces an EPS having the composition Glc: Gal = 2: 2.

Example IV: Biosynthesis of an EPS

A plasmid is prepared from the plasmid pSA3 by adding a DNA fragment of the strain Sfi39 containing the sequence SEQ ID NO: 1 described in example I (see the sequence list), and by adding the regulatory parts of 1 epson operon of the S. thermophilus strain Sfi6 (ΞP750043), so that the expression (transcription and translation) of the eps genes of the strain Sfi39 can be carried out in S. thermophilus. The S. thermophilus CNCM 1-1351 strain which was deposited on August 5, 1993 according to the Budapest Treaty is then transformed by electroporation with this plasmid. Any other shooting strain of S. thermophilus could also have been used. This strain presents under the microscope an appearance of non-flagellated shells forming chains, it does not make spores, it is anaerobic optional, and it produces an EPS having the composition Glc: Gal: Rha = l: 3: 2 Example V: Biosynthesis of an EPS

Genomic DNA of the CNCM 1-1897 strain is isolated by the method of Slos et al. (1991). The DNA preparation is digested with PstI and BamHI, the DNA fragments are separated by electrophoresis on 0.7% agarose gel, the fragments greater than 20 kb are eluted, the DNA extracted is ligated to the vector pJIM2279 ( obtained from P. Renault, INRA, Jouy-en-Josas, Paris, France) previously digested with PstI and BamHI then dephosphorylated. The Lactococcus lactis MG1363 strain (J. Bacteriol., 154, 1-9, 1983), cultivated on GM17 medium at 30 ° C., is transformed by the method of De Vos et al. (Gene, 8d5, 169-176, 1989). The clones transformed are selected by hybridization of the genomic DNA of the clones with one of the probes derived from the sequence SEQ ID NO: 1. Among the transformants, the positive clones are selected.

Example VI: Preparation of functional genes homologous to the eps 6, eps7, eps 8, esp9 and eps 10 genes Functional derivatives of the epsβ, eps7, eps8, esp9 and epsl O genes are prepared by using a method adapted from that described by Adams et al.

(EP402450; Genencor). For this, each of the eps genes is isolated by PCR from the genomic DNA of the strain Sfi39. The primers used are chosen from the sequence SEQ ID NO: 1, so as to amplify only each gene. Each PCR product was cloned into the pFGl vector derived from the vector pSA3 (Dao et al., Appl. Environ. Microbiol., _49., 115-119, 1985) which contains, in addition to the erythromycin selection gene, the gene chloramphenicol selection and the promoter regulating the transcription and translation of the eps operon of the strain Sfi6 (EP750043). Each PCR product is cloned functionally downstream of this promoter so that the eps gene can be normally expressed in S. thermophilus. Finally, each vector is subjected to chemical mutagenesis m-vi tro with hydroxylamine, as described by Adams et al.

The E. coli XLl-blue strain is transformed with the pFGl plasmids treated with the mutagenic agent, transformants are selected, a recombinant plasmid is isolated, then the recombinant plasmid is incorporated by electroporation into a mutant of the S. thermophilus Sfι39 strain. whose eps gene which is also carried by the plasmid pFG1 was previously inactivated as described in Example II. The transformation is carried out by a method adapted from that described by Slos et al. (Appi. Environ. Microbiol., 51, 1333-1339, 1991), during which the cells are subjected to a 2.1 kV, 25 μF and 400 Ω discharge, they are resuspended in 1 ml of HJL medium which are incubated for 4 h at 37 ° C., the cells are spread on an LM17 solid medium supplemented with 2.5 μg / ml of erythromycm and 3 μg / ml of chloramphenicol which are incubated for 16 h at 37 ° C, then select the transformed colonies that survive. To select functional derivatives of the epsβ, epsl, eps8, esp9 and epsl O genes, the strains of lactic acid bacteria transformed in an HJL liquid medium are cultivated, dilutions are spread over a Ruthenium Red solid medium, the producing strains are selected. of EPS which remain white, and the rheological properties of the EPS produced by each clone are compared to that of the EPS produced by the strain Sfι39. By observing in particular the viscosity and the normal strength of the EPS of each clone, it was thus possible to identify certain clones producing an EPS having a different property. Analysis of the eps gene carried on the pFG1 plasmids of the selected clones shows that these differ from the original eps gene only by substitution, deletion or addition of a small number of nucleotide bases leading to modifications of the amino acid sequence of the glycosyltransferase.

Example VII Synthesis of Polysaccharides with the Eps6, Eps7, Eps8, Eps9 and Eps10 Glycosyltransferases

Each of the eps genes is isolated by PCR from the genomic DNA of the strain Sfi39. The primers used are chosen from the sequence SEQ ID

NO: 1, so as to amplify only each gene. Each PCR product is cloned into an expression vector pQE60

(Qiagen, US). The vectors thus treated are then introduced into the E. coli DH5α strain (Stratagen, US) which does not naturally have glycosyltransferase activity. The transformed DH5α strains are cultured in LB medium containing ampicillin up to a DOgoo of 0.8, 1 mM IPTG is added to the culture medium, the cells are incubated for 2 h, the cells are centrifuged at 6000 g for 10 min at 4 ° C., they are washed in buffer A (50 mM Tris-acetate pH8, ImM of DTT, 1 mM EDTA and 10% glycerol), they are centrifuged again, they are resuspended the cells in buffer A additionally containing 1 mM PMSF (phenylmethane sulfonyl fluoride) and the suspension is subjected to 15000 psi. The whole cells are eliminated by centrifugation at 6000 g for 15 min at 4 ° C., the cell membranes of the supernatant are recovered by ultracentrifugation at 100,000 g for 1 h, and the aliquot is resuspended in buffer A. This protocol thus generally allows '' get about 10 mg / ml protein.

In parallel, the Sfi39 strain is cultivated in a medium containing D-Glcp, D-Galp and D-Galf, subjects the cells of this culture to 20,000 psi, the saccharide chains in formation (linked to a primer) are extracted from the membranes by treating the cellular debris with chloroform / methanol (2: 1), by evaporating the extract. As a control in the tests which will follow, the procedure is the same with the strain Sfi39 cultivated in the presence of radioactive saccharides.

The specificity of the glycosyltransferases present in the membranes is then determined by a method similar to that described by Kolkman et al. (Mol. Microbiol., 2_6, 197-208, 1997). For this, the reaction is carried out for 2 h, in a reaction volume of 150 μl, 100 μl (approximately 1 mg) of membranes, 50 mM of Tris-acetate pH8, 10 mM of MgCl2, ImM of EDTA, 1 μl (approximately 25 nCi) of UDP [14C] _ Glcp, of UDP [14C] -Qalp or of UDP U4C] -Galf according to the glycosyltransferase considered, and the deposit containing the saccharide chains in formation unlabeled isolated from the strain Sfi39 (see paragraph above). To stop the reaction, 2 ml of chloroform: methanol (2: 1) are added to the reaction volume, the mixture is vigorously stirred and left to stand for 30 min. The nucleotide saccharides are eliminated by extracting the organic phase 3 times in succession in 0.4 ml of PSUP (for 1 liter: 15 ml of chloroform, 250 ml of methanol, 1.83 g of KCl and 235 ml of water). The lower phase containing all the saccharide chains linked to a primer is then dried under vacuum, resuspended in 200 μl of 1-butanol and separated in two tubes.

In the first series of tubes, the oligosaccharides are detached from their primer by carrying out a gentle hydrolysis in 50 mM of TFA (ac. Trifluoroacetic) at 90 ° C for 20 min. In the second series of tubes, the oligosaccharides are subjected to extensive hydrolysis in a 4M TFA solution for 1 h at 125 ° C.

The hydrolysates are then dried and resuspended in 5 ml of transporting saccharides in 40% isopropanol (5 mg / ml of each of the following saccharides: Glc,

Gay, GalN, maltose, maltotriose, maltotetraose). The hydrolyzed saccharides from each series of tubes are then deposited on an HPTLC silica gel 60 plate (Merck) which is subjected to an eluent composed of chloroform, acetic acid and water (respectively 6: 7: 1). The plate is then autoradiographed for 16 to 24 h with Biomax MS film

(Kodak). The transporting saccharides are visualized by diffusing on the plate an ethanoi solution comprising 5% of H 2 SO 4, and by heating the plate to 100 ° C. for 15 min. The analysis of the plates then allows the precise identification of the enzymatic functions encoded by the eps genes of the strain Sfi39.

Example VIII: Cloning of the Lactobacillus helveti cus operon CNCM 1-1449 involved in the synthesis of an EPS

1. Repeated structure of EPS

The structure of EPS produced by the strain Lactobacillus helveticus LH59 (CNCM 1-1449) has already been presented by Stingele et al. (Carbohydrate Research, 302, 197-202, 1997), and in patent application EP-A-0699689, the technical content of these documents being incorporated by reference. This strain is used in particular for the fermentation of dairy products such as cheese.

2. Cloning

Among the primers tested, those having the nucleotide sequences SEQ ID NO: 41 and SEQ ID NO: 42, have allowed to amplify an eps gene fragment under the same conditions as those of Example I. For this, the genomic DNA of the LH59 strain was previously isolated according to the technique of Slos et al. (1991). A 140 bp PCR fragment could thus be isolated and then cloned into the linearized plasmid pGEMT (Promega, USA). The sequencing of this fragment by the dideoxynucleotide method (f-mol® DNA Sequencmg System kit, Promega) indicates a sequence corresponding to nucleotides 1385 to 1526 of the sequence SEQ ID NO: 12.

This PCR fragment was used to screen a λ-ZAP Express library (Stratagene, USA) containing DNA fragments of the LH59 strain according to the method described in Example 1. By cutting the nucleotide sequences of the different genomic mserts, it was thus possible to characterize the nucleotide sequence SEQ ID NO.12 corresponding to the operon of concern LH59 involved in the synthesis of an EPS (see FIG. 2).

3. Analysis of the sequence SEQ ID NO: 12

The sequence SEQ ID NO: 12 contains 1 operon eps of the strain LH59 and includes a transposition element oriented in the opposite direction to 11 complete ORFs presented in the same orientation, which is called epsl to epsll. This transposition element is a fragment of cryptic plasmid of L. helveticus pLH3 (Pridmore et al., FEMS Microbiological Letters, 124, pp. 301-306 (1994)) (see FIG. 2). Comparison of the amino acid sequences encoded by ORFs with those of proteins present in the Swiss-Prot database, using FASTA, PEPPLOT and PILEUP software from GCG- softwear, isconsin, USA, allows to deduce the function of these proteins. The results are presented below.

The ORF epsl (nucleotides 1052-1729) codes for a protein (SEQ ID NO: 13) having approximately 59% identity with undecaprenyl-phosphate-glycosyl-1-phosphate-transferases, in particular of S. salivarius (Genbank , no. P72577). This homology confirms its role in the biosynthesis of EPS.

The ORF eps2 (nucleotides 1733-2848) codes for a protein (SEQ ID NO: 14) having approximately 36% identity with an α-glycosyltransferase from S. thermophilus (Genbank, no. Q56044), in particular at SX2EX7E motif present in all these enzymes. This homology confirms its role in the biosynthesis of EPS. The EPS3 ORF (nucleotides 2851 -3942) codes for a protein (SEQ ID NO: 15) having approximately 32% identity with glycosyltransferases, in particular from Bacillus subtilis (Genbank, no. P71055). This ORF thus probably codes for an α-glycosyltransferase. The ORF eps4 (nucleotides 3929 -5083) codes for a protein (SEQ ID NO: 16) having approximately 28% identity with glycosyltransferases, in particular of Bacillus subtilis (Genbank, n ° P71055). This ORF probably codes for an α-glycosyltransferase. The EPS5 ORF (nucleotides 5076-6095) codes for a protein (SEQ ID NO: 17) having approximately 29% identity with β-glycosyltransferases, in particular of S. thermophilus (GenBank n ° O07339). This homology confirms its role in the biosynthesis of EPS. Epsβ ORF (nucleotides ⁶⁰⁹⁸ -7090) encodes a protein (SEQ ID NO: 18) having approximately 29% identity with β-glycosyltransferases, particularly S. thermophilus (GenBank # O07339). This homology confirms its role in the biosynthesis of EPS.

The EPS7 ORF (nucleotides 7095 -8258) codes for a protein (SEQ ID NO: 19) having approximately 27% identity with an EPS polymerase from S. pneumaniae (GenBank n ° O07867). This homology confirms its role in the biosynthesis of EPS.

The epsβ ORF (nucleotides 8283 -9122) codes for a protein (SEQ ID NO: 20) having approximately 27% identity with glycogenmes (GenBank n ° Q22997; P46976; P132780). These homologies confirm its role in the biosynthesis of EPS probably as a glycosyltransferase.

The ORF eps9 (nucleotides 9125 -10263) codes for a protein (SEQ ID NO: 21) similar to UDP-galactosyl-pyranose-mutase from Ξ. coli K-12 involved in the conversion of UDP-galactopyranose to UDP-galactofuranose. The ORF eps9 must play a central role in the synthesis of exopolysaccharides, since it allows the transfer of phosphofuranose to the next repeating unit. The ORF eps10 (nucleotides 10250-11662) codes for a protein (SEQ ID NO: 22) having similarities with the EpsK of Lactococcus lactis, the CpsL of Streptococcus pneumoneia, the TrsA of Hiersma enterocol ti ca and the RfbX of E Coli (Yao and Valvano, Journal of Bacteπology, 176, pp. 4133-4143 (1994)). These homologies confirm its role as a transport unit for the polysaccharides formed.

The ORF epsll (nucleotides 1166413181) codes for a protein (SEQ ID NO: 23) having similarities with protein 0331 from E. Coli and IcaC from Staphylococcus epidermidi s (Elmann et al., Molecular Microbiology, 20 (5), pp. 1083-1091 (1996)). These homologies confirm its role as a unit for the synthesis or export of exopolysaccharides. In conclusion, the genomic mserts isolated from the pCMV vectors cover a genomic region of the L. helveticus LH59 strain which is manifestly involved in the biosynthesis of EPS.

Example IX Inactivation of eps Genes

The epsl-8 gene of the operon is inactivated by homologous recombination to confirm their importance in the biosynthesis of EPS. For this, a fragment of each ORF originating from one of the pCMV vectors described in Example I above is amplified by PCR, the PCR product is ligated into the plasmid pSA3 suitably digested beforehand, the E. coli strain is transformed. XLl-blue by the ligation product, transformants are selected, a recombinant plasmid is isolated, then the L. helveticus LH59 strain is transformed with the recombinant plasmid using the method described by Thompson et al. (Appi. Micob. Biotec, 35, 334-338, 1991). The transformed cells are spread on an MRS solid medium supplemented with 2.5 μg / ml of erythromycin, which is then incubated for 16 h at 45 ° C., then the transformed colonies which survive are selected.

The surviving colonies integrated the recombinant pSA3 plasmid into one of the 8 eps genes. This can be verified by Southern-Blot of a digested preparation of genomic DNA from the surviving colonies, and hybridization of the Southern-Blot filter with the PCR product made radioactive. The colonies having integrated the plasmid have different bands than those obtained under the same conditions with the wild strain. In addition, the colonies having integrated into one of the 8 eps genes the recombinant plasmid present have lost their spinning character in an MSK milk. Example X: Biosynthesis of an EPS

A plasmid is prepared from the plasmid pSA3 (Dao et al. (1985)), by adding a DNA fragment of the LH59 strain containing the nucleotide sequence SEQ ID NO -.12 described in Example VIII and by adding the regulatory parts of the eps operon of the S. thermophilus strain Sfi6 (EP750043), so that the expression (transcription and translation) of the eps genes of the LH59 strain can be carried out in S. thermophilus. The S. thermophilus CNCM 1-1422 strain which was deposited on May 18, 1994 according to the Budapest Treaty is then transformed by electroporation with this plasmid. Any other S. thermophilus shooting strain could also have been used. This strain presents under the microscope an appearance of non-flagellated shells forming chains, it does not make spores, it is anaerobic optional, and it produces an EPS having the composition Glc: Gal = 2: 2.

Example XI: Biosynthesis of an EPS A plasmid is prepared from the plasmid pSA3 by adding a DNA fragment of the strain LH59 containing the nucleotide sequence SEQ ID NO: 12 described in Example I and by adding the regulatory parts of the eps operon of the S. thermophilus strain Sfi6 (EP750043), so that the expression (transcription and translation) of the eps genes of the LH59 strain can be carried out in S. thermophilus. The S. thermophilus CNCM 1-1351 strain which was deposited on August 5, 1993 according to the Budapest Treaty is then transformed by electroporation with this plasmid. Any other S. thermophilus shooting strain could also have been used. This strain presents under the microscope an appearance of non-flagellated shells forming chains, it does not make spores, it is anaerobic optional, and it produces an EPS having the composition Glc: Gal: Rha = l: 3: 2

Example XII: Biosynthesis of an EPS The genomic DNA of the LH59 strain is isolated by the method of Slos et al. (1991). The DNA preparation is digested with SalI and BamHI, the DNA fragments are separated by electrophoresis on 0.7% agarose gel, the fragments greater than 5 kb are eluted, the DNA extracted is extracted with the vector pJIM2279 beforehand digested with Sali and BamHI then dephosphorylated. The strain Lactococcus lactis MG1363 (1983) is transformed as described in Example V. The clones transformed are selected by hybridization of the genomic DNA of the clones with one of the probes derived from the sequence SEQ ID NO: 12. Among the transformants, the positive clones are selected.

Example XIII: Preparation of Functional Genes Homologous to the EPS Genes Functional derivatives of the epsl-8 genes are prepared using a method adapted from that described by Adams et al. (EP-A-0402450; Genencor). For this, each of the eps genes is isolated by PCR from the genomic DNA of the LH59 strain. The primers used are chosen from the sequence SEQ ID NO: 12, so as to amplify only each gene. Each PCR product is cloned into the expression vector pQE60 (Qiagen, US) functionally downstream of an E. coli promoter regulating the transcription and translation of the eps gene. Finally, each vector is subjected to intra chemical mutagenesis with hydroxylamine, as described by Adams et al. The E. coli DH5α strain is transformed

(Stratagen, US) which does not naturally have glycosyltransferase activity, by the plasmids pQE60 treated with the mutagenic agent, then transformants are selected. To identify among the transformants eps genes coding for enzyme variants, the enzyme specificity of the enzymes expressed in each E. coli clone is analyzed, using the method described in the example.

VII. Analysis of the eps genes carried on the pQE60 plasmids of the clones thus selected shows that these differ from the original eps gene only by substitution, deletion or addition of a small number of nucleotide oases leading to modifications of the amino acid sequence of glycosyltransferase.

Example XIV ^• polysaccharide synthesis with the glycosyltransferases of LH59

Each of the eps genes coding for a glycosyltransferase is isolated by PCR from the genomic DNA of the LH59 concern. The primers used are chosen from the sequence SEQ ID NO: 12, so as to amplify only each gene according to the method described in 1 example VII.

In parallel, the LH59 strain is cultivated in a medium containing D-Glcp, D-Galp and D-Galf, the cells of this culture are subjected to 20,000 psi, the saccharide chains in formation (linked to a primer) are extracted from the membranes. by treating cell debris with chloroform / methanol and then evaporating the extract. As a control in the tests which will follow, the procedure is the same with the strain LH59 cultivated in the presence of radioactive saccharides. The specificity of the glycosyltransferases present in the membranes is then determined by a method similar to that described in Example VII.

Analysis of the plates makes it possible to confirm the function as glucosyltransferase of the operons epsl to eps4 by the functional analysis of their activity in the presence of a substrate.

Example XV: Synthesis of a New Polysaccharide 0 The EPS is purified from a culture medium of the Lactobacillus helveti cus TY1-2 strain (Y. Yamamoto et al., Carbonydrate Research, 261, 67-78, 1994). This EPS is constituted by the repetition of the following saccharide unit: 5 β-D-Gal / M l → 4) -β-D-Glc / 7

o → 6) -β-D-Glc /? - (1 → 3) -β-D-Glc / 7- (1 → 6) -α-D-Glc Nac- (1 → 3) -β-D- Gal (1 →

1

5 For this, add a volume of trichloroacetic acid (40%) to the culture medium, centrifuge

(17000 g, 20 mm), recover the supernatant, add a volume of acetone, centrifuge again (17000 g, 20 mm), recover the precipitated part which is dissolved 0 in distilled water at pH 7, it is dialyzed in water (12 h), then the insolubles are removed by ultracentπfugation (110,000 g, 1 h).

Next, an N-acetylgalactosamme is grafted onto the galactose of the lactose branches of EPS, using a β-1, 4-N-acetylgalactosammyl-transferase and its appropriate substrate, for example those described by Lutz et al. and Yamashiro et al. (J. of Biol. Chem., 269, 29227-29231, 1994; J. of Biol. Chem., 270. 6149-6155, 1995).

Example XVI: Synthesis of a New Polysaccharide 5 The EPS of Example XV is subjected to a-2, 3-sialyltransferase marketed by Calbiochem® (Cat No. 566618 -M, USA) in the presence of its appropriate substrate .

Example XVII: Synthesis of a New Polysaccharide 0 The EPS of Example XV is subjected to the α-2,6-sialyltransferase described by Sjoberg et al., Or to the α-2,6-sialyltransferase described by Yamamoto et al. , in the presence of their appropriate substrates (J. Biochem., 271, 7450-7459, 1996; Biosci. Biotech. Biochem., 62, 210-214, 1998). 5

Example XVIII: Cloning of the Operator of L. bulgaricus CNCM

1-800 involved in the synthesis of an EPS 1. Repeated structure of EPS

The structure of the EPS produced by the strain 0 L. bulgaricus Lfi5 (CNCM 1-800) was determined by magnetic resonance, according to a technique similar to those described in EP 97111379.0 and EP 97111381.6. The results show that the EPS consists of the following repeating unit: 5 β-D-Gal α-L-Rap β-D-Gal /?

1 1 1

4 3 3 o → 3) -β-D-Gal /? (1 → 4) -α-D-Galp (1 → 2) -α-D-Galp (1 → 3) -β-D-Glc /? (1 →

2. Cloning

Among the primers tested, those having the nucleotide sequences SEQ ID NO: 43 and SEQ ID NO: 44, made it possible to amplify an eps gene fragment in the same conditions as those of Example I. For this, the genomic DNA of the strain Sfι39 was previously isolated according to the technique of Slos et al. (1991).

A PCR fragment could thus be isolated and then cloned into the linearized plasmid pGEMT (Promega, USA). The sequencing of this fragment by the dideoxynucleotide method (f-mol® DNA Sequencmg System kit, Promega) indicates a sequence corresponding to nucleotides 1698 to 1841 of the sequence SEQ ID NO: 24. This PCR fragment was used to screen a λ-ZAP Express bank (Stratagene, USA) containing DNA fragments of the Lfι5 strain according to the protocol of Example 1.

It was thus possible to characterize a sequence corresponding to one operon of the Lfι5 strain involved in the synthesis of an EPS.

In order to recover the sequences bordering the initial fragment, other screens of the Lambda Zap library were carried out, but none made it possible to isolate new fragments. Certain sequences of Lb. bulgaricus containing, for example, promoters are known to be toxic and therefore unstable in E. coll. This surely explains the difficulties in isolating the adjacent DNA fragments. In order to circumvent this obstacle and to clone the regions of DNA bordering this clone, the techniques of reverse PCR and SSP-PCR (Single Specifies Primer PCR) were used. This made it possible to amplify several fragments corresponding to the 5 ′ and 3 ′ ends of the eps operon. By cutting the nucleic sequences of the different enromosomal fragments isolated, it was thus possible to characterize the sequence SEQ ID NO: 24 (see FIG. 3). 3. Analysis of the sequence SEQ ID NO. - 24 ha sequence SEQ ID NO: 24 presents the operon eps of the strain Lfι5 and includes 14 complete ORFs, in the same orientation, which is called epsA to epsN (see FIG. 3 ). The comparison of the amino acid sequences encoded by the ORFs with those of proteins present in the Swiss-Prot database, using the FASTA, PEPPLOT and PILEUP software from GCG-softwear, iscons, USA, makes it possible to deduce the function of these proteins. The results are presented below.

The ORF epsA (nucleotides 295 - 1326} codes for a protein (SEQ ID NO: 24) having 32% identity with the LytR protein from Synechocytis sp Pcc 6803 (Genbank no.

P74136). This ORF obviously codes for a transcription attenuator influencing regulation.

The EPSB ORF (nucleotides 1326-2225) codes for a protein (SEQ ID NO: 25) having approximately 26% identity with the CapA protein of Staphylococcus aureus (Genbank no.

P39850). This ORF is obviously involved in determining the length of saccharide chains.

The epsC ORF (nucleotides 2514-2858) codes for a protein (SEQ ID NO: 27) having approximately 38% identity with the EpsB protein Lactococcus lactis (GenBank n ° O06030). This ORF thus probably codes for a regulatory protein involved in the control of the molecular weight and / or the length of the polysaccharide chain.

The ORF epsD (nucleotides 3020-3796) codes for a protein (SEQ ID NO: 28) having approximately 41% identity with the protein EpsC Lactococcus lactis (GenBank n ° O06031). This homology confirms the role of this ORF in the synthesis of an EPS.

The EPSE ORF (nucleotides 3827-4483) codes for a protein (SEQ ID NO: 29) having approximately 44% identity with the protein Cpsl4E from Streptococcus pneumoniae

(Genbank, No. P72513). This ORF thus obviously codes for a galactosyi- or glucosyl -phospho-transferase catalyzing the transfer of the first saccharide onto the lipid or protein transporter.

The EPSF ORF (nucleotides 4878-5678) codes for a protein (SEQ ID NO: 30) having approximately 30% identity with the EpsF protein of Streptococcus thermophilus

(Genbank, no Q56043). This ORF thus obviously codes for an α-glycosyltransferase.

The EPSG ORF (nucleotides 5750-6877) codes for a protein (SEQ ID NO: 31) having approximately 34% identity with the EpsG protein of Streptococcus thermophilus

(Genbank, no Q56044). This ORF thus obviously codes for an α-glycosyltransferase.

The EPSH ORF (nucleotides 7015-7821) codes for a protein (SEQ ID NO: 32) having approximately 37% identity with the protein YveO of Bacillus subtilis (Genbank, n ° P71054). This ORF thus obviously codes for a β-glycosyltransferase.

The EPSJ ORF (nucleotides 7839-8708) codes for a protein (SEQ ID NO: 33) having approximately 23% identity with the protein Cpsl4K from S. pneumaniae (Genbank, n ° O07341). This ORF thus obviously codes for a glycosyltransferase.

The EPSJ ORF (nucleotides 9213-10196) codes for a protein (SEQ ID NO: 34) having approximately 33% identity with the Eps I protein of S. thermophilus (GenBank n ° Q56046). This ORF thus obviously codes for a β-glycosyltransferase.

The ORF epsK (nucleotides 10227-11288) codes for a protein (SEQ ID NO: 35) having approximately 24% identity with the protein YveQ Bacillus subtilus (GenBank P71056). This ORF thus probably codes for a protein responsible for the polymerization of repeat units.

The EPSL ORF (nucleotides 11358 -12284) codes for a protein (SEQ ID NO: 36) having approximately 20% identity with the RfaS protein of E. coli (GenBank n ° P27126). This homology demonstrates that this ORF is involved in the biosynthesis of an EPS.

The epsM ORF (nucleotides 2372-13382) codes for a protein (SEQ ID NO: 37) having approximately 38% identity with the CpsH protein of Streptococcus agaloctiae (GenBank n ° 087183). This homology shows that this ORF is involved in the biosynthesis of an EPS.

The ORF epsN (nucleotidesl4269 - 1572Ï) codes for a protein (SEQ ID NO: 38) having approximately 35% identity with the protein Cpsl9BJ of Streptococcus pneumonoiae

(GenBank # O07342). This ORF probably codes for a protein responsible for the export of EPS.

In conclusion, the observed homologies indicate that the different ORFs identified in the chromosomal inserts are involved in the biosynthesis of EPS.

Example XIX: Biosynthesis of an EPS A plasmid is prepared from the plasmid pSA3 (Dao et al. (1985)), by adding a DNA fragment of the strain Lfi5 containing the sequence SEQ ID NO: 24 and by adding the regulatory parts of the eps operon of the S. thermophilus Sfi6 strain (EP750043), so that the expression (transcription and translation) of the eps genes of the Lfi5 strain can be carried out in S. thermophilus as proposed in example X. Example XX: Biosynthesis of an EPS

A plasmid is prepared from the plasmid pSA3, by adding a DNA fragment of the Lfi5 strain containing the sequence SEQ ID NO: 24 and by adding the regulatory parts of the epson operon of the strain S. thermophilus Sfi6 (EP750043), so that the expression (transcription and translation) of the eps genes of the Lfi5 strain can be carried out in S. thermophilus as proposed in Example XI.

Example XXI: Biosynthesis of an EPS

Chromosomal DNA from the CNCM 1-800 strain is isolated, the DNA preparation is digested with Xbal, the DNA fragments are separated by electrophoresis on 0.7% agarose gel, the further fragments are eluted of 9 kb, the DNA extracted is ligated to the vector pJIM2279 (obtained from P. Renault, INRA, Jouy-en-Josas, Paris, France) previously digested with Xbal then dephosphorylated. The Lactococcus lacti s MG1363 strain is transformed as described in Example V. The clones transformed are selected by hybridization of the genomic DNA of the clones with one of the probes derived from the sequence SEQ ID NO: 24.

EXAMPLE XXII Preparation of Functional Genes Homologous to the EPS1 -9 Genes

Functional derivatives of the epsl -9 genes are prepared using a method adapted from that described by Adams et al. (EP402450; Genencor). For this, each of the eps genes is isolated by PCR from the genomic DNA of the Lfi5 strain (CNCM 1-800). The primers used are chosen from the sequence SEQ ID NO: 24, so as to amplify only each gene. Each PCR product is cloned according to the method described in Example XIII. To identify among the transformants eps genes coding for enzyme variants, the enzymatic specificity of the enzymes expressed in each E. coli clone is analyzed, using the method described in Example XVII. Analysis of the eps genes carried on the pQE60 plasmids of the clones thus selected shows that these differ from the original eps gene only by substitution, deletion or addition of a small number of nucleotide bases leading to modifications of the amino acid sequence of glycosyltransferase.

Example XXIII: Synthesis of Polysacchaπdes with Lfι5 Glycosyltransferases Each of the eps genes coding for a glycosyltransferase is isolated by PCR from the genomic DNA of the Lfι5 strain (CNCM 1-800). The primers used are chosen from the sequence SEQ ID NO: 24, so as to amplify only each gene according to the protocol described in Example VII using the expression vector pBAD (Invitrogen, USA).

At the same time, the Lfι5 strain is cultivated in a medium containing D-Glcp, D-Galp and L-Rhap, the cells of this culture are subjected to 20,000 psi, the saccharide chains in formation (linked to a primer) are extracted from the membranes. by treating cell debris with chloroform / methanol (2: 1) and evaporating the extract. As a control in the tests which will follow, the procedure is the same with the strain Lfι5 cultivated in the presence of radioactive saccharides. The specificity of the glycosyltransferases present in the membranes is then determined by a method similar to that described in Example VII.

Plate analysis then allows precise identification of different functions enzymes encoded by the eps genes of the Lfi5 strain. These constructions having incorporated in particular the epsE genes to espJ made it possible to confirm the glucosyltransferase activity of the epsE gene by a functional analysis of their activity in the presence of their glucose substrate.

Example XXIV Construction of a shuttle vector

The epsA to epsM genes were amplified by PCR and inserted into the shuttle vector pTRKH2 (0 'Sullivan et al., Gene, 137, pp. 227-231 (1993)) and cloned into heterologous hosts Lactococcus lactis MG1363 so as to determine if such a fragment allows the production of EPS by this strain. After restriction analysis of the different PCR amplicons, it was shown that no detectable deletion or rearrangement took place during the cloning steps. The transformants analyzed on the plates by means of a toothpick test show a threading phenotype by the strains not having this phenotype at the origin. This proves that the esp genes introduced are responsible for the synthesis of EPS by these transformed strains.

Claims

1. Recombinant enzyme capable of being involved in the biosynthesis of exopolysaccharides by at least one step comprising the formation of a bond in the form of an α or β isomer between carbon 1 carrying the aldehyde reducing function of a D- Activated galp and a phosphate of a lipophilic or protein primer, so as to ensure progressively the formation of these exopolysaccharides with at each step the addition of a new saccharide unit which is attached by its semi-acetal function to an alcoholic hydroxyl of another saccharide unit, which is at the end of a chain of saccnaπdes linked to this primer.

2. A recomoined enzyme according to claim 1, capable of being involved in the biosynthesis of EPS having the repetitive saccharide unit β-D-Gal l

Φ

6 → 3) -β-D-Glc / - (1 → 3) -β-D-Gal / - (1 → 3) -α-D-Glcp- (1 → specifically catalyzing one of the following bonds between saccharides:

- the β-1,3 osidic bond between carbon 1 of an activated D-Galf and carbon 3 of Glcp present at the non-reducing end of a saccharide chain in formation;

- the β-1,3 osidic bond between carbon 1 of an activated D-Glcp and carbon 3 of Galf present at the non-reducing end of a saccharide chain in formation; - the β-1, 6 osidic bond between carbon 1 of an activated D-Galp and carbon 6 of Glcp present at the non-reducing end of a saccharide chain in formation; and

3. Recombinant enzyme according to claim 2, having one of the amino acid sequences chosen from the group formed by the sequences SEQ ID NO: 7, SEQ ID NO: 8, SEQ ID NO: 9, SEQ ID NO: 10, SEQ ID NO: 11, and the sequences which are homologous to them.

4. Recombinant enzyme according to claim 2, having one of the amino acid sequences chosen from the group formed by the sequences SEQ ID NO: 2, SEQ ID NO: 3, SEQ ID NO: 4, SEQ ID NO: 5 and SEQ ID NO: 6.

5. Nucleotide sequence coding for a recombinant enzyme according to any one of the preceding claims.

6. Nucleotide sequence according to claim 5, comprising at least one gene having a nucleotide sequence chosen from the sequences delimited in the sequence SEQ ID NO: 1 by nucleotides 8633-10678, 10703-11668, 12838-13785, 14138-15553, 16919-18013, and the sequences which are homologous to them.

7. Nucleotide sequence according to claim 5, comprising at least one gene having a nucleotide sequence chosen from the sequences delimited in the sequence SEQ ID NO: 1 by nucleotides 2703-3389, 3390-4118, 4130-4819, 4832-5527 and 5629-6993.

8. Recombinant enzyme according to claim 1, capable of being involved in the biosynthesis of EPS having the repetitive saccharide unit β-D-Gal / 7- (l → 4) -β-D-Glcp 1

3 → 5) -β-D-Gal- (l → 3) -α-D-Galp- (l → 3) -α-D-Glcp- (l → 3) -β-D-Glcp- (l →

specifically catalyzing one of the following bonds between saccharides:

- the α-1,3 osidic bond between carbon 1 of an activated D-Glcp and carbon 3 of a Glcp belonging to the non-reducing end of a saccharide chain in formation;

- the α-1,3 osidic bond between the car ^p one 1 of an activated D-Galp and the carbon 3 of a Glcp present at the end

5 non-reducing of a saccharide chain in formation;

- the β-1,3 osidic bond between carbon 1 of an activated D-Galf and carbon 3 of a Galp present at the non-reducing end of a saccharide chain in formation;

- the β-1,3 osidic bond between carbon 1 of an activated D-Glcp 0 and carbon 3 of a Galf present at the non-reducing end of a saccharide chain in formation;

- the β-1,4 osidic bond between carbon 1 of an activated D-Galp and carbon 4 of Glcp present at the non-reducing end of a saccharide chain in formation; and 5 - the β-1,5 sugar link between the repetitive saccharide units of the above EPS.

9. Recombinant enzyme according to claim 8, having one of the amino acid sequences chosen from the group formed by the sequences SEQ ID NO: 13, SEQ ID 0 NO: 14, SEQ ID NO: 15, SEQ ID NO: 16 , SEQ ID NO: 17, SEQ ID NO: 18, SEQ ID NO: 19, and the sequences which are homologous to them.

10. Nucleotide sequence coding for a recombinant enzyme according to claim 8 or 9. 11. A nucleotide sequence according to claim 10, comprising at least one gene having a nucleotide sequence chosen from the sequences delimited in the sequence SEQ ID NO: 12 by nucleotides 1052-1729,1733-2848, 2851-3942, 3929-5083, 5076- _Q 6095, 6098-7090, 7095-8258, 8283-9122, 9125-10263, 10250-

11662, 11664-13181 and the sequences which are homologous to them.

12. Recombinant enzyme according to claim 1, capable of being involved in the biosynthesis of EPS having the repetitive saccharide unit β-D-Gal /? α-L-Rha /? β-D-Gal /? 1 1 1

I I I

4 3 3

→ 3) -β-D-Galp (l → 4) -α-D-Galp (l → 2) -α-D-Galp (l → 3) -β-D-Glc / 7 (l → and specifically catalyzing one of the following links between saccharides:

- the β-1,3 osidic bond between carbon 1 of an activated D-Galp and carbon 3 of the saccharide present at the non-reducing end of a saccharide chain in formation, said chain being composed in particular of a main chain from 1 to 4 D-Galp and / or D-Glcp linked to each other by α-1,2, α-1,3, β-1,4 and / or β-1,3 osidic bonds, and said saccnaride present at the non-reducing end being linked to a single other saccharide by its carbon 1; - the β-1,4 osidic bond between the carbon 1 of an activated D-Galp and the carbon 4 of the saccharide present at the non-reducing end of a saccharide chain in formation, said chain being composed in particular of a main chain from 1 to 4 D-Galp and / or D-Glcp linked to each other by α-1,2, α-1,3, β-1,4 and / or β-1,3 osidic bonds, and said saccharide present at the non-reducing end being linked to a single other saccharide by its carbon 1; and

- the α-1,3 osidic bond between the carbon 1 of an activated L-Rhap and the carbon 3 of the saccharide present at the non-reducing end of a saccharide chain in formation, said chain being composed in particular of a main chain from 1 to 4 D-Galp and / or D-Glcp linked to each other by α-1,2, α-1,3, β-1,4 osidic bonds and / or β-1,3, and said saccharide present at the non-reducing end being linked to a single other saccharide by its carbon 1.

13. Recombinant enzyme according to claim 12, having having one of the amino acid sequences chosen from the group formed by the sequences SEQ ID NO: 25, SEQ ID NO: 26, SEQ ID NO: 27, SEQ ID NO: 28 , SEQ ID NO: 29, SEQ ID NO: 30, SEQ ID NO: 31, SEQ ID NO: 32, SEQ ID NO: 33, and the sequences which are homologous to them.

14. Nucleotide sequence coding for a recombinant enzyme according to claim 12 or 13.

15. Nucleotide sequence according to claim 14, comprising at least one gene having a nucleotide sequence chosen from the sequences delimited in the sequence SEQ ID NO: 24 by nucleotides 2514-2858, 3020-3796, 3827-4483, 4878-5678, 5750-6877, 7015-7821, 7839-8708, 9213-10196, 10227-11288 and the sequences which are homologous to them.

16. Recombinant vector comprising a nucleotide sequence according to any one of claims 5 to 7, 10, 11, 14 and 15.

17. A cell comprising, integrated into its genome or by means of a replicable plasmid, a recombinant nucleotide sequence according to any one of claims 5 to 7, 10, 11, 14 and 15, and expressing at least one functional recombinant enzyme according to any one of claims 1 to 4, 8, 9, 12 and 13.

18. Cell according to claim 17, characterized in that it is a lactic acid bacterium.

19. A method of producing an EPS, in which:

(1) a DNA fragment coding for at least one of the enzymes according to any one of the clones is cloned into a vector claims 1 to 4, 8, 9, 12 and 13, said vector further comprising a sequence allowing autonomous replication or integration into one or more host cells (shuttle vector), (2) the host cell is transformed by said vector , said host cell using the enzymes encoded by the vector, where appropriate in combination with other enzymes produced by the host cell, for the biosynthesis of said exopolysaccharide, (3) the transformed host cell is then cultured under conditions suitable for 'obtaining said exopolysaccharide.

20. The method of claim 19, wherein the vector further comprises at least one functional promoter sequence and at least one functional translational activation sequence, allowing the expression of DNA encoding at least one of the enzymes according to any one of claims 1 to 4, 8, 9, 12 and 13.

21. Process for the synthesis of an EPS, which consists of the repeated assembly of a single saccharide unit, itself formed of a main chain of sugars, the ends of which are involved in the polymerization of the units, in which uses several recombinant enzymes or several recombinant genes likely to be involved in the biosynthesis of EPS with the repetitive saccharide unit β-D-Gal / - (l → 4) -β-D-Glc / - 1 3

→ 5) -β-D-Gal / - (l → 3) -α-D-Galp- (l → 3) -α-D-Glc / - (l → 3) -β-D-Glcp- ( l →

characterized in that these enzymes or these genes, in particular chosen from the enzymes claimed in claim 8 or 9, or the genes claimed in claim 10 or 11, allow the creation on the main chain of the saccharide unit of one or more lactose branches linked to a sugar of the main chain by their glucose.

22. Method according to claim 21, characterized in that the main chain of the saccharide unit consists of at least 2 sugars chosen from D-Galp, D-Galf and / or D-Glcp.

23. The method of claim 21, characterized in that the glucose of the lactose ramifications is linked by an β-1,3 osidic bond to one of the sugars of the main chain.

24. Method according to any one of claims 21 to 23, characterized in that an enzyme or a gene is also used which allows the binding between an N-acetyl-galactosamine and the galactose of the lactose branching.

25. A method according to one aca of claims 21 to 24, characterized in that one further uses an enzyme or a gene which allows the bond between the N-acetyl-neuraminic acid and the galactose of the lactose branching.

26. Method according to any one of claims 21 to 25, characterized in that the synthesis is carried out in vi tro and / or in vivo.

27. Polysaccharide capable of being synthesized according to the process claimed in claims 21 to 26, constituted by the repeated assembly of a single saccharide unit, itself formed of a main chain of saccharides whose ends are involved in the polymerization units, characterized in that the main chain comprises at least one lactose branch linked to a saccharide of the main chain by its glucose, the galactose of this branching being further linked to an N-acetylgalactosamine, provided that this polysaccharide does not consist of the repetition of the following saccharide unit β-I 3a] p> l--) - ^ ^D 3al ^ l → 4) -rGlςσ

1 71

3

→ 5) 43] al / ^" l → 3} α-IX3a] ^

28. A branched polysaccharide according to claim 27, characterized in that the galactose of the lactose branching is also linked to an N-acetyl-neuraminic acid.

29. Use of a recombinant enzyme according to any one of claims 1 to 4, 8, 9, 12 and 13 or of a nucleotide sequence according to any one of claims 5 to 7, 10, 11, 14 and 15 , for the synthesis of an EPS.