WO1994026882A1 - Proline iminopeptidase, process for its preparation and its use in the flavouring of food compositions - Google Patents

Abstract

A novel proline iminopeptidase which is a metal dependent serine peptidase, which is obtainable from Propionibacterium shermanii ATCC 9617 and which has a calculated molecular mass of 45 kDa. The proline iminopeptidase has an amino acid sequence according to the unique sequence of Figure 3. Genetic variants having a homology exceeding 60 % and having the same functionality are comprised by the invention. The proline iminopeptidase can be prepared in large quantities with the use of a genetically modified microorganism. It is used for modifying the flavour of food products, and in particular for rendering the product less bitter.

Description

PROLINE IMINOPEPTIDASE, PROCESS FOR ITS PREPARATION AND ITS USE IN THE FLAVOURING OF FOOD COMPOSITIONS

The present invention relates to proline iminopeptidase, to a process for its preparation and to its use in the flavour modification of food compositions.

Background of the invention

Proline iminopeptidase (EC 3.4.11.5) is a general name used for enzymes which share the property that they are able to catalyse the liberation of an aminoterminal proline residue from a protein or a polypeptide chain. By catalysis is meant here that the proline removal is not observed in the absence of the enzyme and that the enzyme is not changed during the process. The proline iminopeptidases are found in various living matter: animals, plants (e.g. mushrooms, apricot seeds) and micro-organisms, e.g. in bacteria such as Escherichia , Bacillus and Propionibacterium but also in mammals (swines kidney) . In the original sources proline iminopeptidase is produced at levels which are too low to be of any practical use. Propionibacterium is an important industrial bacterium which is used for the manufacturing of Swiss cheese.

(A) Lait 1990, 0_, 439-452 describes the purification and characterisation of a proline iminopeptidase obtained from the cell extract of Propionibacterium shermanii strain 13673 (private collection) . M_R is 61 kDa, optimum activity at 40°C and pH 8.0.

(B) J Dairy Sci 1978 61:303-308 discloses that the Propionibacterium shermanii P-59 strain when present in a peptide containing medium, produces large amounts of proline.

(C) Japanese patent application 02/113,887 describes a process for the preparation of a proline iminopeptidase by culturing an Escherichia coli strain, which had been subjected to a homologous transformation with a clone bank of DNA fragments originating from the strain Escherichia coli HB 101, a strain which is able to produce proline iminopeptidase. The selected proline iminopeptidase producing transformant produced considerably more proline iminopeptidase than the original strain.

(D) Japanese patent application 03/108,483 describes a proline iminopeptidase, found in mushrooms, with a M_R of 150 kDa (dimer) and an isoelectric point at about pH 4.3 The optimum activity is at about pH 7.2 The enzyme is not inhibited by phenyl ethylsulfonyl fluoride (PMSF) . This reference tells that the bitter taste of protein hydrolysate is believed to be caused by proline containing peptides and that the removal of proline is considered to be effective. The mushroom enzyme removes aminoterminal proline. It is stated that bitter peptides were decomposed and the bitter taste of the processed protein removed.

(E) J. Bacteriology 1992, vol. 174. p. 7919-7925 discloses the cloning, sequencing and expression of the gene of a proline iminopeptidase (M_R 33 kDa) from B. coagulanε . The host organism is an E. coli strain.

The presence of proline iminopeptidase in food compositions has been associated with flavour modifications. The enzyme is said to cause the slightly sweet flavour in Swiss cheese, particularly in Emmentaler cheese. Moreover, it may take away bitter taste in processed protein. The various proline iminopeptidases, however, are different substances, which only share a common enzymatic activity. Additionally they have their individual properties including a characteristic organoleptic dynamic profile. When added to food each proline iminopeptidase may in its own way contribute to the flavour of food, which is a process quite similar to cheese flavouring caused by peptidases which are secreted by micro-organisms responsible for cheese ripening.

Therefore a need exists for novel proline iminopeptidases and for processes to make these proline iminopeptidases available in sufficiently large quantities.

STATEMENT OF THE INVENTION

The invention provides a novel polypeptide having proline iminopeptidase activity of a metal dependant serine peptidase comprising at least the amino acid sequence responsible for proline iminopeptidase activity of a proline iminopeptidase, obtainable from Propionibacterium shermanii ATCC 9617 with a calculated molecular mass of 45 kDa or said polypeptide comprising a derivative of said amino acid sequence still having proline iminopeptidase activity.

The invention further provides a polypeptide having proline iminopeptidase activity which comprises an amino acid sequence which essentially corresponds to the unique amino acid sequence of Figure 3 and Sequence id. no. 3, or a derivative thereof wherein the extent of homology between the amino acid sequence of the derivative and the unique sequence exceeds 60%, preferably 75%, more preferably 90%. The polypeptide according to the invention is preferably a food grade polypeptide suitable for use in food. The polypeptide which corresponds to the amino acid sequence of Figure 3 or a derivative thereof exhibiting more than 60% homology is preferably also a metal dependant serine 5 peptidase. In particular a polypeptide according to the invention is sensitive to phenylmethyl sulphonyl fluoride. The invention also comprises nucleotide sequences encoding such polypeptide having proline iminopeptidase activity (denoted as pip gene in the present specification) .

10 The invention comprises too a genetically modified micro¬ organism capable of producing such polypeptide having proline iminopeptidase activity and the progeny obtained from such a micro-organism containing a nucleotide sequence encoding the above polypeptide having proline

15 iminopeptidase activity.

Another embodiment of the invention is a process for the preparation of a polypeptide having proline iminopeptidase activity comprising growing the above micro-organism in a suitable growth medium

20 and collecting the enzyme from the fermentation broth, optionally after disrupting the micro-organism cells.

Still another embodiment of the invention is a process for modifying the flavour of a food composition which process

25 comprises subjecting the food composition or a component thereof to the proline iminopeptidase activity of the polypeptide of the invention or to the enzymatic activity of a food-grade micro-organism which produces such a polypeptide having proline iminopeptidase activity.

30 Other embodiments of the invention are food compositions which themselves or of which components have been exposed to the activity of the polypeptide having proline iminopeptidase activity of the invention or in which the polypeptide having proline iminopeptidase activity or a

35.genetically modified micro-organism of the invention has been incorporated.

DETAILS OF THE INVENTION

5 The found proline iminopeptidase is an intracellular enzyme which can only be obtained in minor quantities from the original source organism Propionibacterium shermanii ATCC 9617. The invention provides a process which makes use of genetically modified, preferably food-grade icro-

10 organisms, and which makes the instant proline iminopeptidase available in amounts large enough for practical purposes. The invention is also directed at polypeptides that are functional equivalents of the proline iminopeptidase, i.e. a polypeptide comprising at least the

15 amino acid sequence responsible for the iminopeptidase activity of the found proline iminopeptidase.

The found enzyme has a calculated molecular mass of 45 kDa and is characterised by a unique amino acid sequence which

20 is depicted in Figure 3 and Sequence id No. 3. The enzyme has its optimum activity at pH 7.5 and 40-45°C. It is a metal dependant serine peptidase as its activity is reduced by metal chelating compounds and phenylmethylsulfonyl fluoride. A polypeptide having proline iminopeptidase

25 activity comprising an amino acid sequence which essentially corresponds to the unique amino acid sequence of Figure 3 and Sequence id No. 3, or a derivative thereof wherein the extent of homology between the amino acid sequence of the derivative and the unique sequence exceeds

30 60%, preferably 75%, more preferably 90% also falls within the scope of the invention.

Figure 3 and Sequence id No. 3 also show the isolated and 35 established DNA sequence encoding the proline iminopeptidase obtained from Propionibacterium shermanii ATCC 9617.

The invention comprises all DNA sequences encoding the polypeptide havaing proline iminopeptidase activity of the invention as mentioned above.

It will be appreciated by a skilled person that it may be advantageous to adapt the codons of a heterologous gene to the preferred codon usage of a host cell. Therefore DNA sequences with alternative codon usage are also included within the scope of the invention.

"An amino acid sequence essentially corresponding to the sequence" is understood to include genetic variants, which retain the functional properties of the enzyme according to the unique sequence. Genetic variants are based on the introduction of modifications of the amino acid sequence represented in Figure 3 which have no substantial adverse effect on the functionality, i.e. proline iminopeptidase activity, of the polypeptide. Also it includes the possibility to improve the functionality of the polypeptide by modifications in the amino acid sequence. Thus the present invention not only covers nucleotide sequences encoding the unique amino acid sequence of Figure 3, but also nucleotide sequences coding for a different amino acid sequence which still is capable of proline iminopeptidase activity, i.e. a polypeptide as defined above.

A widespread method for establishing the extent of homology (percentage similarity of amino acid sequences) between proline iminopeptidase amino acid sequences and the unique sequence depicted in Figure 3 uses the standard algorithm of Needleman and Wunsch (J. Mol . Biol . 48: 443-445, 1970). For the calculations of the similarity percentages suitably a computer program named "Gap" can be used. This program forms part of a sequence analysis software package (version 6.0) issued by G.C.G (see also Devereux et al. Nucleic Acids Res . 12: 387-395 (1984)). The invention comprise^ proline iminopeptidases which have an extent of homology with the unique amino acid sequence depicted in Figure 3 and Sequence id No. 3 exceeding 60%, more preferably 75%, still more preferably 85%, and most preferably 90%.

Another aspect of the present invention relates to a genetically modified micro-organism capable of producing a polypeptide having proline iminopeptidase activity, which micro-organism contains a nucleotide sequence coding for the polypeptide having proline iminopeptidase activity of the invention. The encoding DNA sequence is present either in an expression vector, e.g. a plasmid or is integrated in the chromosome. The nucleotide sequence coding for the polypeptide having proline iminopeptidase activity may be flanked on one or both sides by nucleotide sequences which enable or promote the expression of the gene in the host organism. Generally the nucleotide sequence is a double- stranded DNA (ds-DNA) coding for proline iminopeptidase and the 5'-end of the coding region of the plus strand of the ds-DNA is preceded by and in phase with a translation initiation codon and an expression regulon situated upstream of the translation initiation codon, while the 3 *-end of the coding region of the plus strand of the ds-DNA is followed by a translation stop codon, optionally followed by a transcription termination sequence. Preferably the proline iminopeptidase encoding nucleotide sequence in the micro-organism is preceded by a sequence which encodes a signal peptide which is effective in secreting the proline iminopeptidase in the medium in which the micro-organism is grown. The retrieval of the product is thus simplified and the cells can continuously produce the desired polypeptide. A number of sequences for secreting polypeptides are known to a person skilled in the art. For inserting and expressing the proline iminopeptidase gene in the cells of the host organism the coding sequence may be inserted into a suitable vector, a so-called recombinant expression vector. Then the recombinant vector, comprising a nucleotide sequence coding for said proline iminopeptidase, is transferred into a host micro-organism which becomes capable of expression of said proline iminopeptidase.

The vector may be in the form of a plasmid which remains as a separate genetic entity in the host cell. Alternatively the present coding sequence after entering the cell may be advantageously integrated into the chromosomal DNA of the micro-organism. This integration of the coding sequence has the effect that the transformed micro-organism will produce generations of progeny which are equally capable of forming proline iminopeptidase because the proline iminopeptidase encoding gene is not easily lost. It is therefore advantageous that the DNA flanking the proline iminopeptidase encoding nucleotide sequence comprises a nucleotide sequence which facilitates the integration of the proline iminopeptidase encoding nucleotide sequence into the host genome. An example of such a sequence are fragments of the gene encoding X-prolyl-dipeptidyl aminopeptidase (pepxp) , described by Mayo et al. (1991) Appl.Environ.MicrobioL 57. 38-44, in which case integrants easily can be detected in a plate assay using the chromogenic substrate Gly-Pro-BNA (Leenhouts et al. (1991), J. Bacteriology 123. 4794-4798). Optionally the vector comprises at least one marker gene, which may be a protein conferring to the micro-organism resistance to an antibiotic or, preferably, to a food-grade natural bacteriocin. More preferably an auxotrophic marker sequence is used. Suitable examples of auxotrophic marker sequences comprise DNA sequences coding for enzymes involved in carbohydrate metabolism, more particularly in the metabolism of sugars such as lactose, sucrose and raffinose, e.g. phospho-/3-galactosidase and α-galactosidase. Such DNA sequences are preferably obtained from food-grade micro-organisms. Suitable examples of such auxotrophic marker genes are described in EP 0 355 036.

The invention includes not only the directly transformed micro-organisms but also their progeny, provided that the gene encoding the proline iminopeptidase of the invention has been inherited by the progeny. The transformation process can be used to transform micro-organisms previously unable to make proline iminopeptidase into proline iminopeptidase producing micro-organisms. Alternatively the process may be used to increase the proline iminopeptidase activity of micro-organisms already capable of producing proline iminopeptidase. When expression of the present gene in a proper micro-organism has been effected, said micro- organism can normally be reproduced using conventional fermentation techniques.

The present invention comprises too micro-organisms capable of producing the instant polypeptide having proline iminopeptidase activity. Such micro-organisms comprise host micro-organisms in which a recombinant vector as hereinbefore defined has been introduced and the progeny obtained from such host organisms and which include bacteria, fungi and yeasts. A suitable micro-organism is Escherichia coli . A food grade micro-organism, particularly a lactic acid bacterium is preferred. Preferably the present micro-organism is selected from the genera Lactococcuε, Streptococcus, Lactobacillus , Leuconostoc, Pediococcus, Bacillus, Bifido-bacterium, Brevibacterium, Micrococcuε, Propionibacterium, Staphylococcus, Streptococcus , Gluconobacter, Acetobacter, Vibrio, Corynebacterium, Aspergillus and Zymomonas . According to a more preferred embodiment of the invention the micro-organism is selected from the genera Lactococcus, Streptococcus, Lactobacillus , Leuconostoc, Bifidobacterium, Brevibacterium, Aspergillus and Propionibacterium. Lactococcus lactis is a most preferred species. According to another preferred embodiment the present micro-organism is a yeast, in particular a Saccharomyceε cerevisiae strain.

When the transformed host is grown under proper conditions in a growth medium, the proline iminopeptidase gene will be expressed. The amount of proline iminopeptidase produced is substantially affected by temperature, composition and aeration of the growth medium. Generally the growth medium may be a common one, mainly consisting of water, carbon sources, nitrogen sources, inorganic ions and, if necessary, amino acids. The medium may be chosen such that it is suited for optimal growth, reduced growth or for maintenance only of the micro-organisms, depending on what is needed for optimal induction of expression of the proline iminopeptidase (pip) gene.

An enzyme containing solution may be obtained, by e.g. disrupting the host cells (e.g. by sonification) and centrifuging the cell debris.

The proline iminopeptidase activity of the supernatant can be assessed by measuring its ability to hydrolyse proline- p-nitroanilide according to a method described in example 1 and 4. Table I shows the supernatant activities of the original source organism Propionibacterium shermanii ATCC 9617 and of the transformed E. coli strain of the invention, while in Table II the supernatant activities of the transformed Lactococcuε lactis strain of the invention are shown. It is apparent that the source organism is a very ineffective proline iminopeptidase producer yielding only one seventh of the proline iminopeptidase activity of the micro-organisms of the invention. Moreover, the source organism growth is very slow, needing 3-4 days to reach its stationary phase, in contrast to the transformed E . coli or L . lactis strains of the invention, which reach full growth in 16 hours or less.

According to a more preferred embodiment a micro-organism is used which is capable of secreting the enzyme directly into the growth medium due to the presence of an effective signal sequence in the DNA preceding the structural proline iminopeptidase DNA.

As desired the enzyme may be more or less purified by one or more purification steps according to methods well known in the art. Alternatively, the crude supernatant or the recombinant micro-organism can be used as such.

The proline iminopeptidase of the invention is primarily intended for treating food compositions. Hitherto the micro-organisms which produce a proline iminopeptidase either are not food-grade or produce proline iminopeptidase very inefficiently. Substantial amounts of proline iminopeptidase can now be obtained with a food-grade process.

Food compositions are understood to be products which are suited and intended for human consumption comprising products which are used for their preparation.

According to a further aspect of the invention the proline iminopeptidase may be used for the modification of the flavour of food compositions. This is achieved by subjecting the food composition or a component thereof to the polypeptide or to the enzymatic activity of the intact or lysed cells of a food-grade micro-organism of the invention.

More specifically, when an edible composition is subjected to the present proline iminopeptidase under proper conditions, a bitter taste which often is present in food containing processed protein and which originates from proteolysed proteins or polypeptides having an N-terminal proline residue may be removed or at least reduced. With advantage food compositions containing an added protein hydrolysate, such as infant food, sport food and diet food can be made less bitter. Alternatively, the protein hydrolysate itself may be made less bitter. In the absence of organoleptic assessments of the bitterness reducing properties of the present proline iminopeptidase and with a view to the earlier mentioned relationship between the removal of aminoterminal proline residues and reduction of bitterness, measurements of the aminoterminal proline release are considered to be a suitable substitute.

The proline iminopeptidase of the invention being new, after being incorporated into a food composition it may impart unexpected organoleptic properties to that composition.

The invention may be especially used when preparing with the help of a micro-organism casein or caseinate containing products, such as cheese or quark, in which the micro¬ organism contains the proline iminopeptidase gene of the invention.

The invention comprises food compositions which have been exposed to the activity of the proline iminopeptidase or in which the proline iminopeptidase has been incorporated or in which a food-grade micro-organism according to the invention, preferably a lactic acid bacterium, has been incorporated.

The pleasant flavour of fresh baked wheat products like 5 bread is due to the formation of volatile flavour components during baking. These flavour components are formed during the baking process from different precursors which have been characterized as amino acids and free sugars (Spicher G. and Nierle W. (1988) , Appl. MicrobioL

10 Biotechnol. , 2j3, 487-492). During the baking process, these amino acids and reducing sugars interact as precursors to form different types of Maillard intermediates, that can have characteristic bread flavours. Increasing levels of free amino acids produced during sour dough fermentation

15 contributes to bread flavour. Proline has been identified as one of the key amino acids in this respect. Reaction of proline with glucose leads to the formation of 2-acetyl-l- pyrroline which has been identified (Schieberle P. and Grosch W. (1985) , Z. Lebensm. Unters. Forsch. 180 , 474-

20 478), as the most important volatile component of white bread and bread crust. Increased levels of proline in model bread-making systems resulted in improved crust colour and bread flavour (Fadel H.H.M. and Hegazy N.A. (1993) , Die Nahrung 37, 386-394) . Increasing the amount of free proline

25 in wheat dough by the action of proline iminopeptidase could therefore improve and or enhance the flavour of bread and other dough fermentation products.

Therefore, according to an alternative embodiment a proline 30 iminopeptidase may be applied for the enhancement of the flavour of food compositions which are made from a dough, e.g. bread. This can be accomplished by leavening the dough with a yeast, particularly a Saccharomyces cereviεiae strain, in which the proline iminopeptidase encoding 35 sequence is incorporated and which is capable of expressing proline iminopeptidase. Obviously other food microorganisms of the invention could also be used but as yeast is already used in dough this would appear to be an extremely suitable microorganism.

Description of the drawings

Fig. 1 and Sequence id No. 1 show the initial amino¬ terminal amino acid sequence of purified proline iminopeptidase isolated from Propionibacterium shermanii ATCC 9617, as determined on a gas phase sequencer

Fig. 2 and Sequence id No. 2 show the mixed DNA probe, derived from the amino-terminal amino acid sequence of the proline iminopeptidase, used to isolate the gene encoding this enzyme

Fig. 3 and Sequence id. No. 3 show the nucleotide sequence with the corresponding amino acid sequence of the proline iminopeptidase and the flanking nucleotide sequences isolated from Propionibacterium shermanii ATCC 9617. Relevant restriction sites are indicated as are the start and the end (*) of the pip gene.

Fig. 4 shows the restriction enzyme map of plasmid pUR5302. ■■■■ is chromosomal DNA of Propionibacterium shermanii ATCC 9617.

Fig. 5 shows the restriction enzyme map of plasmid pMMB67EH

Fig. 6 shows the restriction enzyme map of plasmid pUR5303. ■■■■ is chromosomal DNA of Propionibacterium εhermanii ATCC 9617.

Fig. 7A shows the restriction enzyme map of plasmid pMG36E Fig. 7B shows the restriction enzyme map of plasmid pMG36E and the sequence of the multiple cloning site (sequence id. No. 7) . T is the transcriptional terminator of prtP

Fig. 8 shows the restriction enzyme map of plasmid pUG37. T is the transcriptional terminator of prtP.

Fig. 9 shows the restriction enzyme map of plasmid pUG38. T is the transcriptional terminator of prtP.

Fig. 10 shows the restriction enzyme map of plasmid pUG39. T is the transcriptional terminator of prtP.

Fig. 11 shows the restriction enzyme map of plasmid pUG40. T is the transcriptional terminator of prtP.

Fig. 12 shows the restriction enzyme map of plasmid pUG41. T is the transcriptional terminator of prtP.

Fig. 13 shows the restriction enzyme map of plasmid pGKV223D

Fig. 14 shows the restriction enzyme map of plasmid pUG42.

Fig. 15 shows the restriction enzyme map of plasmid pKL15A

Fig. 16 shows the restriction enzyme map of plasmid pUK39.

Fig. 17 shows the restriction enzyme map of plasmid pORI28. 0RI+ is the origin of replication of the lactococcal plasmid pWVOl.

Fig. 18 shows the restriction enzyme map of plasmid pORI22. ORI+ is the origin of replication of the lactococcal plasmid pWVOl. Fig. 19 shows the restriction enzyme map of plasmid pINT124. ORI+ is the origin of replication of the lactococcal plasmid pWVOl.

Fig. 20 shows the restriction enzyme map of plasmid pINT125. 0RI+ is the origin of replication of the lactococcal plasmid pWVOl.

Fig. 21 shows the restriction enzyme map of plasmid pINT29. ORI+ is the origin of replication of the lactococcal plasmid pWVOl.

Fig. 22 shows the restriction enzyme map of plasmid pINT51. ORI+ is the origin of replication of the lactococcal plasmid pWVOl. T is the transcription terminator of prtP.

Fig. 23 shows the restriction enzyme map of plasmid pINT61. ORI+ is the origin of replication of the lactococcal plasmid pWVOl. T is the transcription terminator of prtP.

Fig. 24 shows the restriction enzyme map of plasmid pORI280. ORI+ is the origin of replication of the lactococcal plasmid pWVOl. T is the transcription terminator of prtP.

Fig. 25 shows the restriction enzyme map of plasmid pUG139. 0RI+ is the origin of replication of the lactococcal plasmid pWVOl. T is the transcription terminator of prtP.

Fig. 26 shows the restriction enzyme map of plasmid pUG139R. ORI+ is the origin of replication of the lactococcal plasmid pWVOl. T is the transcription terminator of prtP.

Fig. 27 shows the restriction enzyme map of plasmid pUG239. ORI+ is the origin of replication of the lactococcal plasmid pWVOl. T is the transcription terminator of prtP.

Fig. 28 shows the restriction enzyme map of plasmid pUG339. ORI+ is the origin of replication of the lactococcal plasmid pWVOl. T is the transcription terminator of prtP.

Fig. 29 shows the restriction enzyme map of plasmid pUG439. ORI+ is the origin of replication of the lactococcal plasmid pWVOl. T is the transcription terminator of prtP.

Fig. 30 shows the restriction map of plasmid pINT250. ORI+ is the origin of replication of the _rlactococcal plasmid pWVOl.

Fig. 31 shows the restriction map of plasmid pUG539. ORI+ is the origin of replication of the lactococcal plasmid pWVOl. T is the transcription terminator pf prtP.

Fig. 32 shows the restriction map of plasmid pUR2778

Fig. 33 shows the restriction map of plasmid pUR5305

Fig. 34 shows the restriction map of plasmid pUR5306 The invention is illustrated by means of the following examples.

All standard molecular biological procedures were performed as described by Maniatis et al. (1982) Molecular Cloning, a Laboratory Manual. Cold Spring Harbor Laboratory, Cold Spring Harbor, NY, USA. Enzyme incubations were done as recommended by the supplier, Amersham International PLC, UK.

All percentages are wt.% unless indicated otherwise.

Example 1

Isolation of proline iminopeptidase from Propionibacterium εhermanii ATCC 9617

To isolate the proline iminopeptidase enzyme, cells of Propionibacterium εhermanii ATCC 9617 were grown anaerobically in 150 1 Sodium Lactate Broth (1 % sodium lactate, 1 % trypticase-peptone, 0.025 % K₂HP0₄, 0.0005 % MnS0₄, 1 % yeast extract; pH adjusted to 7.0). Cells were collected by centrifugation, washed three times with 0.5 M NaCl and resuspended in 20 % sucrose at 4 °C with mechanical stirring for 2 hours. Cells were recovered by centrifugation and resuspended in 3 1 of lysozyme buffer (0.5 M sucrose, 0.3 M NaCl, 50 mM MgS0₄, 10 mM KC1, 10 mM Tris-HCl and 2 mg lysozyme/ml at pH 7.8) at room temperature and stirred for 8 hours. Spheroplasts were recovered by centrifugation and lysed in 5 1 of 5 mM sodiumphosphate pH 7.5 by stirring for 4 hours at room temperature. The crude extract was ultracentrifuged to remove unlysed cells and cell debris. (NH₄)₂S0₄ was added to the cell extract to reach a degree of saturation of 15 %. The precipitate was removed by centrifugation and additional (NH₄)₂S0₄ was added to the supernatant to reach 45 % saturation. The precipitate was collected by centrifugation, dissolved in 20 mM Tris-HCl, pH 7.0 and dialysed against this same buffer. The sample was applied to a Q-Sepharose™ Fast Flow anion exchange column

(Pharmacia) equilibrated in the Tris-HCl buffer. Bound material was eluted with a linear gradient of 0 to 0.5 M NaCl. Eluted fractions were tested for their ability to hydrolyse proline-p-nitroanilide in 10 mM sodiumphosphate buffer at pH 7.0. The release of p-nitroaniline was spectrofotometrically measured at 405 nm. Active fractions were concentrated on an Amicon concentration device with a PM 10 membrane. The concentrated sample was applied to a Superdex 2001™ Biopilot™ column in the same Tris-HCl buffer and fractions were tested for hydrolysing activity. Active fractions were again dialysed and concentrated. The concentrate was applied to a Mono Q™ anion exchange column in the Tris-HCl buffer. Bound material was eluted with a linear NaCl gradient (0 to 0.5 M) in the same buffer. The fractions containing hydrolysing activity were pooled. The aminoterminal sequence was determined with a gasphase sequencer (Applied Biosystems 470A™ equipped with an on¬ line PTH-analyzer type 120 A) . The initial part is indicated in Fig. 1 and Sequence id. No. 1.

Example 2

Cloning of the gene encoding proline iminopeptidase in E. coli

To isolate the chromosomal DNA, cells of Propionibacterium εhermanii ATCC 9617, were grown anaerobically for 36 hours at 30°C in 100 ml Sodium Lactate Broth (1% sodium lactate, 1% trypticase-peptone, 0.025% K₂HP0₄, 0.0005% MnS0₄, 1% yeast extract; pH adjusted to 7.0). Cells were collected by centrifugation and resuspended in 12 ml Tris-Sucrose buffer (6.7% sucrose, 50 mM Tris-HCl, 1 mM EDTA, pH 8.0) and heated for 2 minutes at 37°C. 3 ml Lysozyme Buffer (25 mM Tris-HCl pH 8.0 and 100 mg/ml lysozyme) was added and the mixture was incubated for 30 minutes at 37°C. Subsequently 750 μl of a solution of 10 mg/ml pronase E™ in water, preincubated for 30 minutes at 37°C was added and the mixture was incubated for another 10 minutes at 37°C. 1.5 ml of a solution of 50 mM Tris-HCl, 0.25 M EDTA, pH 8.0 was added, the mixture was gently mixed by inversion and 900 μl 20% SDS was added. The mixture was mixed once more by inversion and incubated for 10 minutes at 37°C. The rather viscous solution was extracted with a phenol/chloroform (1:1) mixture, equilibrated with TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 7.5), and the waterphase, containing the nucleic acids, was separated by centrifugation. The waterphase was collected and extracted once again. Subsequently, the nucleic acids were precipitated from the waterphase by the addition of one volume cold (-20°C) ethanol. The DNA clot was collected on a glass rod and dissolved in TE buffer.

To make a clonebank of the chromosomal DNA obtained, 150 μg of the DNA was partially digested with the restriction enzyme Sau3A, in an end volume of 500 μl, to give DNA fragments having a mean molecular weight of 5 kb. The digest was extracted with phenol/chloroform 1:1, precipitated with 2 volumes of ethanol, dissolved in 150 μl TE buffer, and size separated on a 10-40% sucrose gradient in a buffer of 1 M NaCl, 20 mM Tris-HCl and 5 mM EDTA, pH 8.0. The gradient was centrifuged for 20 hours at 22000 rp in a Beckman SW25™ rotor. 0.5 ml fractions were collected, 20 μl was tested on size on an agarose gel and those fractions, having a molecular weight between 5 and 10 kb, were pooled, precipitated with 2 volumes of ethanol and dissolved in TE buffer. 3 μg of the DNA thus obtained, was ligated into Ba HI cut pBR322, and the ligation mixture was transformed into competent E. coli 294 cells (endo I^", BI~, r_k ^", m_k ⁺; Backman et al. (1976) Proc. Natl. Acad. Sci. USA, 73., 4174-4178) . A clonebank of 4200 amp^r colonies was obtained.

To screen this clonebank for colonies containing the proline iminopeptidase gene, a mixed DNA probe of 20 nucleotides in length, based on the first 7 aminoacids of the N-terminal proline iminopeptidase sequence (Example 1, see also Fig. 2 and Sequence id No. 2) , was synthesized on a DNA synthesizer (Applied Biosystems, model 380A™) . To test if the proline iminopeptidase gene could be detected with this probe, the radioactively labelled probe was hybridized with a Southern blot of the chromosomal DNA, digested with the restriction enzymes EcoRI, Hindlll and BamHI. In all lanes a hybridizing band could be detected, when the hybridization was performed in a mixture of 0.2% bovine serum albumin, 0.2% polyvinyl pyrrolidon, 0.2% Ficoll™, 50 mM Tris-HCl pH 7.4, 10 mM EDTA, 1 M NaCl, 0.5% SDS and 0.1% sodium pyrophosphate at 37°C for 16 hours. Using this probe, 1500 of the 4200 colonies were screened with a colony hybridization procedure. Six of these colonies hybridized with the probe and the plasmid DNA was isolated. Example 3

Nucleotide sequence analysis of the proline iminopeptidase gene

A preliminary restriction map analysis of the 6 clones obtained in Example 2, showed that they all contain a 1 kb BamHI fragment, which hybridizes with the mixed DNA probe, under the conditions, specified in example 2. Therefore, this fragment was subcloned in the BamHI site of the plasmid pEMBL9, resulting in pUR5301. The DNA sequence of the 1 kb insert was established by the Sanger dideoxy chain termination procedure (Sanger et al. (1977) Proc. Natl. Acad. Sci. USA 24_., 5463-5467) , with the modifications of Biggin et al. (1983, Proc. Natl. Acad. Sci. USA 8), 3963- 3965) , using the universal M13 primers, Klenow enzyme and α-³⁵S-ATP (2000 Ci/mmol) . This sequence analysis revealed that the proline iminopeptidase gene had been cloned indeed. The BamHI fragment actually contains a fragment which runs from 240 bp before the startcodon of the proline iminopeptidase gene, until 680 bp after the start codon (Fig. 3 and Sequence id No. 3) . The complete sequence of the proline iminopeptidase gene and its flanking sequences, subsequently has been determined on one of the six original clones (pUR5302, Fig. 4 and Sequence id No. 4), using the double stranded plasmid as a template, following the procedure of Hattori and Sakakis (1986, Anal Biochem. 152- 232-238) . Primers were synthesized as described in Example 2.

For the DNA sequence analysis, the sequence alignments and sequence database searching, two different software programmes were used: a) The "Staden" sequencing programme, as described by Staden (1986, Nucleic Acids Res. .14., 217-231) b) The Sequence Analysis Software Package, version 6.1, licenced from the Genetics Computer Group (University of Wisconsin, Madison, Devereux et al. 1984, Nucleic Acids Res. 12, 387-395)

The analysis of the nucleotide sequence, presented in Fig. 3 and Sequence id No. 3 shows an open reading frame of 1242 bp, encoding a protein of 414 aminoacids having a calculated molecular weight of 45.376 kDa (Sequence id No. 3 and 4) . In the region before the start codon ATG at position 504, no clear promoter region or ribosome binding site, similar to such regions in E. coli , were found. However, the extremely high GC content of the Propionibacterium genome, in relation to the E. coli genome, might explain this phenomenon.

Example 4

Expression of the proline iminopeptidase gene in E . coli

To bring the proline iminopeptidase gene to expression in E. coli , it is preferably placed behind a strong E. coli regulon. Therefore, the gene was introduced into the expression plasmid pMMB67EH, containing the tac promoter and an E. coli ribosome binding site (Fig. 5, Furste et al. (1986) Gene .48_., 119-131) . To clone the proline iminopeptidase gene into this vector, first a PCR fragment was made on pUR5302, as described by Sambrook et al. (Molecular Cloning, a Laboratory Manual, Second Edition (1989) . Cold Spring Harbor Laboratory Press) , using the primers: A: 5'-GGAATTCATGACATGGCAGCACAGTAT 3'

ECORI B: 5'-CCAGGGAGCGGATCCGTGAGGGCC 3' BamHI

Primer A (Sequence id No. 5) is identical to the start of the proline iminopeptidase gene (pos. 503-523) and is extended with an EcoRI site. Primer B (Sequence id No. 6) is complementary to the region around the BamHI site within the proline iminopeptidase gene (pos. 1174-1197) . The resulting 700 bp fragment was digested with EcoRI and Pvul and the resulting 600 bp fragment was isolated on an agarose gel. Subsequently pUR5302 was digested with Pvul and PvuII and the 910 bp long PvuI-PvuII fragment (Fig. 3) , containing the remaining part of the proline iminopeptidase gene, was isolated on an agarose gel. Both fragments were ligated together into the plasmid pMMB67EH, cut with the restriction enzymes EcoRI and S al (Fig. 5) . The ligation mixture was transformed to E . coli JM109 and an amp^r transformant was picked up, containing the expected plasmid pUR5303 (Fig . 6 ) .

To determine if the proline iminopeptidase gene was expressed under control of the tac promoter, E. coli JM109 (pUR5303) was grown in 100 ml Luria Broth to an OD660 of 1.5. The tac promoter was induced by the addition of 0.4 mM IPTG (isopropyl-β-D-thiogalactopyranoside) . Cells were grown for another 1.5 hours, harvested by centrifugation, washed in 20 ml 10 mM potassium phosphate buffer pH 7.0 and resuspended in 20 ml of the same buffer. To 5 ml of this suspension 3 g glass beads (0.15 mm diameter) were added and the mixture was sonicated six times for 30 seconds in a Branson B12 Sonifier™, using a small tip. After cell disruption, the suspension was centrifuged for 10 minutes at 15.000 * g. The resulting supernatant was assayed for proline iminopeptidase activity using proline-p- nitroanilide as substrate (1 mMol p-nitroanilide in 10 mMol Sodiumphosphate at pH 7.0 and 20°C ) . Release of p- nitroaniline was measured spectrofotometrically at 405 nm. Activity was expressed as Units per gramme of protein. One Unit produces 1 mole of p-nitroaniline (e ₀₅ = 9620 M~ ¹.cm^~1) in one minute at 20°C in phosphate buffer at pH 7.0. Protein was determined using the Biorad protein Assay using Bovine Serum Albumin as reference protein. As can be seen from Table I, activity of 224 U/g protein was found, which is seven times as high as found in the original

Propionibacterium εhermanii ATCC 9617 strain. E. coli JM109 (pMMB67EH) is a reference organism which does not contain the proline iminopeptidase gene. Table I

Enzyme activity on proline-p-nitroanilide of bacterial lysates

micro-organism specific proline iminopeptidase activity

E. coli JM109 (pMMB67EH) 0 U/g E. coli JM109 (pUR5303) 224 U/g P. εhermanii ATCC9617 35 U/g

Activity is expressed as units per gramme of protein. One unit produces 1 μmole of p-nitroaniline

(e₄₀₅ = 9620 M^-1.cm^_1) in one minute at 20 °C in phosphate- buffer at pH 7.0.

Example 5

The expression of the pip gene in Lactococcuε lactis

5

To bring the pip gene to expression in L . lactis , it is preferably put behind a strong lactococcal regulon. Several strong lactococcal promoters p32 and p23 have been isolated (van der Vossen et al . (1987) Appl. Environ. MicrobioL 53.

10 2452-2457) and are present in derivatives based on the lactococcal broad-host-range plasmid pWVOl (Leenhouts et al . (1991) Plasmid .2jS,55-66). These derivatives were used to achieve optimal expression of the pip gene. All plasmids were constructed in E. coli or L . lactis .

15 1. Van de Guchte et al . (1992) Appl.

MicrobioL Biotechnol. 37,216-224 and FEMS MicrobioL Rev. 8_8_,73-92 described that efficient expression of genes can be obtained by making use of translational coupling of the gene of interest to the efficiently transcribed truncated

20 ORF32 present in pMG36e (Fig. 7A + B) . Several vectors were constructed to achieve translational coupling of ORF32 and the pip gene. First, the C-terminal part of the pip gene was cloned as a S'acI-XJ al fragment (963 bp) from plasmid PUR5303 (Fig. 6) into the Sacl-Xbal sites of pMG36e, 5 resulting in plasmid pUG37 (Fig. 8) . A small region downstream of the pip gene was removed by digesting pUG37 with Styl and Hindlll, klenov enzyme treatment to make blunt ends and religating, to remove a number of restriction enzyme sites. The resulting plasmid was denoted 0 pUG38 (Fig. 9) . To couple the translation of the pip gene to the truncated ORF32 a PCR fragment was made on pUR5303 using primers:

5 pip210 :

Xho RBS start pip

ATA AGA GCT CGA GAA TAT TCG GAG GAA TTT TGA AAT GAC ATG GCA GCA CAG TAT Sad Sspl stop ORF32 CC

pip203 (identical to primer B, example 4, Sequence id No. 6):

CCA GGG AGC GGA TCC GTG AGG GCC BamHI

Primer pip210 (Sequence id No. 8) contains 22 nucleotides of the N-terminus of the pip gene including its start codon. It is coupled to the translational initiation signals of ORF32 (21 nucleotides) , including a Sspl site, and is extended with an Xhol and a SacI site at the 5'-end. Primer pip203 surrounds the BamHI site in the pip gene (Fig. 3 and 6) . The PCR fragment was digested with SacI and the 552 bp fragment was isolated and ligated into the vector pUG38, cut with the same restriction enzyme.

Transformants containing an active pip gene were identified using the chromogenic substrate Pro-rBNA. In the resulting plasmid pUG39 (Fig. 10) the translation of the preceding truncated ORF32 stops 1 bp upstream of the pip start codon. 2. Translational coupling of the E. coli lacZ gene to the truncated ORF32 was found to be most optimal in a configuration in which the stop codon of ORF32 overlapped the start codon of lacZ (van de Guchte et al . (1991) Mol. Gen. Genet. 227,65-71) . To achieve a similar configuration for the expression of the pip gene a PCR fragment was made on pUR5303, using as primers pip203 (as in 1.) and: pip211:

Xhol RBS start pip ATA AGA GCT CGA GAA ATA TTC GGA GGA ATT TTG AAA TGA CAT GGC AGC ACA GTA SacI Sspl stop ORF32

TCC Primer pip211 (Sequence id No. 9) contains one nucleotide mutation as compared to primer pip210 (Sequence id No. 8) . This mutation consists of an insertion of an A between the Xhol and Sspl site. The resulting PCR fragment was digested with SacI , the 554 bp fragment was isolated and ligated into vector pUG38, cut with the same restriction enzyme. Transformants containing an active pip gene were identified using the chromogenic substrate Pro-βNA. In the resulting plasmid pUG40 (Fig. 11) the stop of the truncated ORF32 overlaps with the start codon of the pip gene in a ATGA configuration.

3. The pip gene was put under direct control of p32, thus without translational coupling, by the removal of the truncated ORF32 from pUG39 by partial Sspl digestion (Fig. 10; deletion of 74 bp) , religation and digestion with Xhol. A construct was picked up with the expected configuration: pUG41 (Fig. 12) .

4. The isolated strong lactococcal promoter p23 lacks translation initiation signals. Therefore, the pip gene with the translation initiation signals of ORF32 was placed under control promoter p23 by making use of vector pGKV223D (Fig. 13) . This vector is derived from the vectors pGKV210 and pGKV223 (v.d. Vossen et al. , 1987. Appl. Environ. MicrobioL 53_., 2452-2457) . pGKV223 contains a 680 random lactococcal insert which carries promoter P23, a ribosome binding site (RBS) and the start of an open reading frame, upstream of the cat-68 chloramphenicol resistance gene. A 407 bp -EcoRV-Haelll fragment of pGKV223, containing the P23 and RBS was subcloned into the Sma site of pGKV210, resulting in the vector pGKV223D. The 1543 bp Sspl fragment carrying the pip gene of pUG39 (Fig. 10) was isolated and ligated to BamHI cut pGKV223D, which had been made blunt by klenov enzyme treatment. A plasmid containing the pip gene inserted in the correct orientation was picked up and was designated pUG42 (Fig. 14) . The final constructs pUG39, 40, 41 and 42 were transformed to L . lactiε MG1363 and LL108 by electrotransformation (Holo and Nes (1989) Appl. Environ. MicrobioL 55.3119- 3123) and proline iminopeptidase activity was determined in cell free extracts as described in Example 1. Strain LL108 was obtained as follows. The repA gene of pWVOl under control of the lactococcal promoter p23 was transferred from plasmid pUC23rep3 (K. Leenhouts et al . 1991. Appl. Environ. MicrobioL 57:2562-2567) to a deletion derivative of the pBR322-based integration plasmid pHV60A (K.

Leenhouts et al . 1989. Appl. Environ. MicrobioL 55:394- 400) , resulting in plasmid pKL15A (Fig. 15) . This vector was integrated by Campbell-type recombination in the chromosome of MG1363 by selection for chloramphenicol resistant transformants. The resulting Cm^r strain was designated LL108. As shown before (K. Leenhouts et al . 1989. Appl. Environ. MicrobioL 55:394-400) this type of selection results at the chromosomal location used in an amplification of approximately 15 copies of the integration plasmid in the chromosome. As a consequence of this Ori⁺- based plasmids have a high copy number in this strain (approximately 10 times higher than pWVOl-derived plasmids normally have in MG1363) . In MG1363 pUG39, 40 and 41 showed a high level of plasmid instability, whereas in LL108 the plasmids were stable. To determine the activity of proline iminopeptidase produced by different genetically modified micro-organisms, the activity was assayed as described in Example 4. The proline iminopeptidase activity levels in strain LL108 carrying either pUG39, 40 or 41 were similar as shown in Table II. The proline iminopeptidase activity in LL108 carrying pUG42 was significantly lower. Therefore the expression cassette as present in pUG39 was chosen for stabilization on the lactococcal chromosome. To facilitate easy transfer of the cassette to the various integration vectors the EcoRI-XmnI fragment (1840 bp) of pUG39 was transferred into the EcoRI-Smal sites of pUK21 (J. Vieira and J. Messing [1991] Gene 100:189-194), resulting in plasmid pUK39 (Fig. 16) .

Table II

Enzyme activity on proline-p-nitroanilide of bacterial lysates

micro-organism specific proline iminopeptidase activity

E. coli JM109 (pUR5303) 60 U/g L. lactiε LL108 0 U/g

L . lactiε LL108 (pUG39) 143 U/g

L . lactiε LL108 (pUG40) 137 U/g

L. lactiε LL108 (pUG4l) 219 U/g

P. shermanii ATCC9617 17 U/g

Example 6

The stabilization of the pip gene onto the chromosome of Lactococcus lactis __

To stabilize the pip gene onto the lactococcal chromosome use was made of a plasmid integration system based on the lactococcal plasmid pWVOl (Leenhouts et al . (1991) Appl. Environ. MicrobioL 5_7,2562-2567; European Patent application EP 0 487 159) . This system renders the genetically modified strains free of heterologous plasmid DNA and antibiotic resistance markers. Two types of integration vectors were used. The first type of vectors integrated entirely into the chromosome through a single cross-over event and gave rise to stable head-to-tail arrangements of 20-30 copies of the vector containing the pip gene. The second type of vectors integrated through a double cross-over event on a single copy of the pip containing part of the integration plasmid. All integration vectors were constructed in an L. lactis Rep⁺ helper strain LL108 (see Example 5) . The correct constructs were isolated from this strain and transferred to the Rep- L. lactiε strain MG1363 to allow integration of the vectors into the chromosome.

Conεtruction Or i⁺ -integration vectorε

The basis of all Ori⁺-integration vectors consists of the

601-bp TaqI fragment (1) of pWVOl containing the plus origin of replication (Leenhouts et al . (1991), Plasmid 26:55-66) and the 121-bp Spel fragment of pUK21 containing a multiple cloning site (Vieira and Messing (1991) , Gene 100: 189-194) with in the Xhol site a 1-kb erythromycin resistance gene of pE194 (2) (Horinouchi and Weisblum (1982), J. Bacteriol. 150:804-814) or the same fragment from pUK21 but with the erythromycin resistance gene in the BamHI site (3) . All fragments were isolated and blunt ends were generated by Klenov enzyme treatment. Fragments (1) and (2) were ligated resulting in pORI28 (Fig. 17) . Fragments (1) and (3) were ligated resulting in pORI22 (Fig. 18).

SCO integration vectors

Plasmid pORI22 was used to construct the Campbell-type integration vectors pINT124 (Fig. 19) and pINT125 (Fig.

20) . Plasmid pINT124 contains a 785-bp XhoII-PvuII internal fragment of the L . lactiε chromosomal pepXP gene (Mayo et al . (1991) Appl. Environ. MicrobioL 57:38-44). Plasmid pINT125 contains a 1504-bp X al fragment carrying the 3'- end of the pepXP gene. Both plasmids contain a 4.6-kb Bgrlll-Sall fragment of plasmid pSRQl of Pediococcuε pentoεaceuε PPEl.0 carrying the sucrose genes εcrA and scrB (corresponding to nt 13285 to 17897 in the Genome Sequence DataBase accession number L32093) . The erythromycin resistance gene was excised from both construct by BamHI digestion and religation.

DCO integration vectors

Plasmid pORI28 was used to construct the replacement-type vectors pINT29 (Fig. 21), pINT51 (Fig. 22) and pINT61 (Fig. 23) . Plasmid pINT29 contains two chromosomal fragments of the pepXP gene region: an Xjal fragment of 1504 bp containing the 3'-end of the pepXP gene and a 1479-bp Spel- Mlul fragment containing the 5'-end of the pepXP gene and the complete ORF1 and its terminator (Mayo et al . (1991) Appl. Environ. MicrobioL 57:38-44). The two fragments are interrupted by the multiple cloning site in pORI28, therefore, integration of pINT29 into the chromosome results in the inactivation of pepXP. For the construction of pINT51 and pINT61 first pORI280 (Fig. 24) was constructed. The E . coli lacZ reporter gene under control of the lactococcal promoter p32 was isolated from pMG57 (v.d. Guchte et ai. (1991) Mol. Gen. Genet. 227:65-71) as an _5coRI-XmnI fragment. Blunt ends were generated using Klenov enzyme and the fragment was ligated in the StuI site of pORI28 resulting in plasmid pORI280. Plasmid pINT61 contains a 1479-bp Mlul-Spel fragment carrying the 5'-end of the pepXP gene and the complete ORF1 with its terminator and an 1.7-kb Spel-Xbal fragment from plasmid pBM330 (Mayo et al . (1991) Appl. Environ. MicrobioL 57:38-44) consisting of chromosomal DNA downstream of ORF1. The multiple cloning site in pINT61 is located just downstream of the terminator of ORF1, therefore, replacement integration into the chromosome does not result in inactivation of genes.

Plasmid pINT51 contains a 655-bp Fεpl-Spel fragment carrying the 3'-end of ORF1 and its terminator, in addition it contains a 506-bp Xbal-Scal internal fragment of the laαR gene of pMG820 (Maeda and Gasson, (1986) J. Gen. MicrobioL 132:331-340; v. Rooijen and de Vos, (1990) J. Biol. Chem. 265:18499-18503). The two fragments are interrupted by a multiple cloning site. Replacement integration of these sites in the chromosome (of strain LB250) does not inactivate genes.

1. Single croεε-over vectorε containing the pip gene . A. The integration plasmid pINT124 (Fig. 19) carries a 4.5 kb DNA fragment containing the sucrose metabolizing genes isolated from the lactic acid bacterium Pediococcuε pentoεaceuε PPEl.O (European patent application 0 487 159 and Gonzalez G.F. and Kunka, B.S. (1986) Appl. Environ. MicrobioL 5JL, 105-109) and which function as selectable marker in L . lactiε . Furthermore, the plasmid contains a 780 bp internal DNA fragment of the lactococcal pepXP gene (Mayo et al. (1991) Appl. Environ. MicrobioL 57.38-44) which provides homology for integration. The integration event disrupts the pepXP gene and this event is easily detected in a plate assay using the chromogenic substrate Gly-Pro-BNA (Leenhouts et al . (1991) J. Bacteriol. 5 173,4794-4798, .

The pip gene from pUK39 (Fig. 16) was isolated as Spel-Xbal fragment and ligated into the X al site of pINT124. Vectors pUG139 and pUG139R were isolated with a configuration as depicted in Fig. 25 and 26, respectively. 0 B__^_ Although the inactivation of the pepXP gene does not influence the growth of L . lactiε in milk, the effect of the mutation on the organoleptic composition of the fermentation products is unknown. To avoid inactivation of pepXP use was made of plasmid pINT125 (Fig. 20) . This 5 vector differs from pINT124 in that it carries the 3'-end of the pepXP gene as homologous region for integration. Integration through this chromosomal fragment does not inactivate pepXP. The Bglll-Wlul fragment of pUK39 was ligated in the Ba HI-WIuI sites of pINT125, which resulted 0 in pUG239 (Fig.27).

2. Double croεε-over vectorε containing the pip gene .

A. Vector pINT29 (Fig. 21) is specific for the delivery of a single copy of a gene of interest in the pepXP gene. The pip gene of pUK39 was isolated as an Bgrlll-MIuI fragment and was ligated with BamHI-MIuI digested pINT29. The resulting plasmid was designated pUG339 (Fig. 28) .

B. Vector pINT6l (Fig. 23) is specific for the delivery of a single copy of a gene of interest downstream 0RF1 in the pepXP gene region, this event does not inactivate pepXP. The pip gene fragment was isolated as an EcoRI-Bglll fragment from pUK39 and was ligated with EcoRI-Bglll digested pINT61. The resulting plasmid was designated pUG439 (Fig. 29) . C. Vector pINT51 (Fig. 22) is specific for delivery of a single copy of a gene of interest downstream of 0RF1 and lacR in strain LB250. Strain LB250 was constructed using pINT250 (Fig. 30), a derivative of pINT61 in which the 9.1- kb iac-operon of L . lactiε was cloned in the BamHI-Wotl sites. The pip gene was inserted as an .EcoRI-BglII fragment from pUK39 into the JBcoRI-Bglll sites of pINT51, resulting in pUG539 (Fig. 31) . The vectors pUG139R, 339 and 439 were integrated in the chromosome of L. lactiε MG1363(pLP712) resulting in strains UG139, 339 and 439, respectively. Plasmid pUG239 was integrated in the chromosome of MG1363 and pUG539 was integrated in LB250, resulting in strains UG239 and 2539, respectively. Strains UG139 and 239 were isolated by direct selection on agar media containing sucrose. Strains UG339, 439 and 539 were selected following a two step procedure as described by Leenhouts and Venema (1992) Med. Fac. Landbouww. Univ. Gent, 57/4b,2031-2043. The proline iminopeptidase activity in the cell free extracts of all strains was determined as described in Example 5 and are depicted in Table III.

Table III

Enzyme activities on proline-p-nitroanilide of bacterial lysates

micro-organism Specific proline iminopeptidase activity

E. coli JM109 (pUR5303) 152 U/g

L . lactiε LL108 0 U/g

L . lactiε LL108 (pUG39) 143 U/g

L . lactiε UG139 33 U/g

L . lactis UG239 31 U/g Functional activity of proline iminopeptidase from different genetically modified micro-organisms was determined using different N-terminal proline containing peptides. Substrates were dissolved (0.2-0.4 mMol) in 0.1 M sodiumphosphate pH 7.0 and incubated with proline iminopeptidase activity solutions at 25-50 Units/1 and incubated at 37°C. To determine the release of free proline from peptide substrates using proline iminopeptidase, samples were taken in time intervals during incubation and analyzed by amino acid analysis using an Alpha Plus II (Pharmacia LKB) amino acid analyzer. After addition of Norleucine (100 nmol/ml) as an internal standard, the samples were adjusted to pH 2.2 using 0.2 M Sodium-citrate and applied on an Ultrapac 7 cation exchange resin. Amino acid analysis was accomplished using a Sodium- citrate/borate elution system using an increasing pH gradient. Detection took place using the Ninhydrin containing method of the amino acids followed by UV- detection at 570 nm for the standard amino acids and 440 nm specifically for proline. Results were compared with results from similar experiments using extracts from corresponding micro-organisms without the genetically modified proline iminopeptidase gen. As shown in Table IV, the release of proline from an example (poly-proline- glycine-proline) substrate is much higher in the extracts recovered from the proline iminopeptidase containing micro¬ organisms than the control sample. Table IV

Release of proline as determined by aminoacid analyses from 1 hour incubation of supernatants from different micro¬ organisms using poly-P-G-P as substrate and equal added activity.

micro-organism amount of released Proline in mg / kg sample

E. COli JM109 (PUR5303) 28 L. lactiε LL108 10

L . Lactiε LL108(pUG39) 30

L . lactiε UG139 47

L . lactiε UG239 28

P. shermanii ATCC9617 39

Example 7

Expression of Proline iminopeptidase by yeast.

The yeast Saccharomyces cerevisiae is widely used in bakery applications and is partly responsible for the flavour generation in dough based products. Expression of proline iminopeptidase in such a yeast would allow faster flavour development and shorter proofing times for these products. To create a yeast strain capable of secreting proline iminopeptidase a PCR procedure could be employed using plasmid pUR5302 as template and primers pip212 (Sequence id No. 10) and pip213 (Sequence id No. 11) :

5* Primer pip212:

Bsal

GCG CGG TCT CC ! G GCC ACA TGG CAG CAC AGT ATC CC

3' Primer pip213:

HindiII

CGC GAT A ! AG CTT TCA CCT GGC GAT CTC G

This would generate an approximately 1269 bp fragment which could subsequently be digested with Bsal and iϊindlll to give a fragment of approximately 1247 bp with 5* overhanging single stranded ends compatable with ligation with an EagI /Hindlll digested vector. The fragment so obtained could subsequently be ligated into the approximately 7378 bp fragment of the yeast expression vector pUR2778 ( Fig 32, Van Gorcom, R.F.M., Hessing, J.G. , Maat, J. , Verbakel, J.M.A. (1991) International patent Application W091/19782 ) , formed by digestion of this plasmid with Hindlll and partial digestion with Eagl . This vector is capable of stable, multi-copy, integration into the yeast genome at the rDNA locus. The resulting plasmid, pUR5305 (Fig. 33), contains the PIP protein fused, in frame, with the secretion signal from the yeast invertase (SUC2) gene product, under the control of the yeast GAL7 promoter. The cloning procedure leads to a replacement of the N terminal methionine with an alanine residue.

Replacement of the galactose inducible GA 7 promoter with a constitutive promoter is required for PIP expression in dough systems where the main carbon sources are maltose or glucose. The constitutive TDH3 (glyceraldehyde 3 phosphate dehydrogenase GAPDH) promoter was isolated as a

Hpal/Hindlll fragment from pgap491 (Holland, M.J. and Holland, J.P., (1979) J. of Biol. Chem. 254. 5466-5474) and introduced into Hindi /Hindi11 digested vector pTZ19U (Mead, D.A. , Szczesna-Skorupa, E. , and Kemper, B. (1986) Protein Engineering 1, 67-74) . By introduction of the nucleotides GAGCT before the C position at -18 relative to the ATG start codon by site directed mutagenesis, a SacI site was created. The resulting plasmid was used as a template for PCR with the primers:

GAP 01 (Sequence id No. 12)

Bgiττ

5' GGAATTCAGATCTTAACTGGTTCAGACGC

and

GAP 02 (Sequence id No. 13) 5 • CCCAAGCTTGAAGACGTCATTTTGTTTTATTTATGTGTG

The PCR fragment so obtained was digested with BglII and SacI. The ligation of the resulting fragment with the approximately 8414 bp vector fragment obtained by partial digestion of pUR5305 with BamHI and SacI gives pUR5306 (Fig. 34) .

Introduction of pUR5306, linearised by digestion with Hpal, into a leu2 deficient S. cerevisiae strain by electroporation in a manner known per se and selection for leucine prototrophy will give a yeast capable of secreting proline iminopeptidase in dough systems.

Claims

1. A polypeptide having proline iminopeptidase activity of a metal dependant serine peptidase, said polypeptide comprising at least the amino acid sequence responsible for the proline iminopeptidase activity of a proline iminopeptidase obtainable from Propionibacterium εhermanii ATCC 9617, with a calculated molecular mass of 45 kDa or said polypeptide comprising a derivative of said amino acid sequence exhibiting proline iminopeptidase activity.

2. A polypeptide having proline iminopeptidase activity comprising an amino acid sequence which essentially corresponds to the unique amino acid sequence of Figure 3 and Sequence id No. 3 or a derivative thereof wherein the extent of homology between the amino acid sequence of the derivative and the unique sequence exceeds 60%, preferably 75%, more preferably 90%.

3. A polypeptide according to claim 1 or 2, said polypeptide being sensitive to phenylmethyl sulphonyl luoride.

4. A polypeptide according to any of the preceeding claims, said polypeptide being food-grade.

5. Nucleotide sequence encoding the polypeptide having proline iminopeptidase activity of any of claims 1-4.

6. Genetically modified micro-organism capable of producing a polypeptide having proline iminopeptidase activity and the progeny obtained from such a micro-organism, said microorganism containing a nucleotide sequence according to claim 5.

7. Micro-organism according to claim 6, wherein the nucleotide sequence encoding the polypeptide having proline iminopeptidase activity is double-stranded DNA (ds-DNA), the 5'-end of the coding region of the plus strand of the ds-DNA being preceded by and in phase with a translation initiation codon and an expression regulon situated upstream of the translation initiation codon, with the 3'-end of the coding region of the plus strand of the ds-DNA being followed by a translation stop codon, optionally followed by a transcription termination sequence.

8. Micro-organism according to claim 6 or 7, wherein the nucleotide sequence encoding the polypeptide having proline iminopeptidase activity is preceded by nucleic acid comprising a sequence which encodes a signal peptide effective in secreting the proline iminopeptidase in the medium in which the micro-organism is grown.

9. Micro-organism according to any one of claims 6-8,wherein the nucleotide sequence encoding a polypeptide having proline iminopeptidase activity is flanked by DNA comprising a nucleotide sequence which facilitates the integration of the proline iminopeptidase encoding nucleotide sequence into the host genome.

10. Micro-organism according to any one of claims 6-9,selected from the genera Lactococcus, Streptococcuε, Lactobacillus, Leuconostoc, Pediococcus, Bacillus, Bifidobacterium, Brevibacterium, Micrococcuε, Propionibacterium, Staphylococcuε, Streptococcus, Gluconobacter, Saccharomyces, Acetobacter, Vibrio, Corynebacterium, Zymomonaε and Aεpergilluε and which preferably is a food-grade micro-organism, selected from the genera Lactococcuε, Streptococcuε, Lactobacillus , Leuconostoc, Bifidobacterium, Brevibacterium, Propionibacterium, Aspergillus and Saccharomyces .

11. A micro-organism according to claim 10, wherein the micro-organism is selected from the group comprising Escherichia coli , Lactococcus lactiε and Saccharomyceε cereviεiae.

12. Process for the preparation of a proline iminopeptidase comprising growing a genetically modified micro-organism according to any one of claims 6-11 in a suitable growth medium and collecting the enzyme from the fermentation broth, optionally after "disrupting the micro¬ organism cells.

13. Process for modifying the flavour of a food composition comprising subjecting the food composition or a component thereof to the polypeptide having proline iminopeptidase activity according to any of claims 1-4.

13. Process for modifying the flavour of a food composition comprising subjecting the food composition or a component thereof to the enzymatic activity of a food- grade micro-organism according to any of claims 6-11.

15. Process according to any of claims 12-14, wherein the modification comprises reduction of bitterness.

16. Food composition which has been exposed to the activity of the polypeptide having proline iminopeptidase activity according to any of claims 1-4.

17. Food composition in which the polypeptide having proline iminopeptidase activity according to any of claims 1-4 has been incorporated.

18. Food composition in which a food-grade micro-organism, preferably a lactic acid bacterium or a yeast according to any one of claims 6-11 has been incorporated.

19. Food composition according to any one of claims 16-18, which contains casein or caseinate and which preferably is cheese.

20. Food composition, preferably dough, which has been exposed to the enzymatic activity of a Saccharomyceε cerevisiae strain producing a polypeptide with proline iminopeptidase activity.

21. Food composition according to any one of claims 16-20which composition contains free proline residues.