The present invention also relates to novel DNA fragments or combination of DNA fragments encoding a new Asp-Phe dipeptide synthetase, micro- organisms containing such DNA fragments, as well as to the new Asp-Phe dipeptide synthetase itself.

As meant herein the term"enzymatical production"is understood to encompass all biochemical/microbiological methods, in the broadest meaning, involving a dipeptide synthetase, for instance by using whole cells-which may be living cells, or permeabilised cells, etc.-comprising such synthetase, or by using other systems, e. g. in vitro methods using dead cell material or isolated or immobilised enzymes, where such synthetase is present.

Background of the invention a-L-Aspartyl-L-phenylalanine (hereinafter also referred to as

Asp-Phe) is an important dipeptide, inter alia used for the production of a-L-aspartyl-L- phenylalanine methyl ester (hereinafter also referred to as APM). APM is known to be a high intensity artificial sweetener, having a sweetness which is about 200x as potent as the sweetness of sucrose. The ß-form of APSI, as well as the stereoisom@rs of APM wherein one or both of the amino acids are in the D-configuration, do not have sweet properties. APM is used for the sweetening of various edible materials.

Various production methods of APM exist ; present routes may be divided into chemical and biochemical/microbiological (in particular, enzymatic) routes. In the ways of producing APM by using known peptide synthesis techniques tedious and expensive processes have to be performed in order to achieve selective a-L, L-coupling, involving intensive protecting and deprotecting of a-amino or carboxyl groups.

Fermentative routes, on the other hand, in general are cheap and intrinsically they display enantio-and regioselectivity. Therefore, fermentative routes have been considered to be promising alternatives for the above-mentioned chemical and biochemical synthesis routes. As can be seen from EP-A-0036258, it has so far been deemed unsuited to produce the dipeptide Asp-Phe in a micro-organism as part of the micro-organism's own protein producing processes; theoretically such production might be achieved by inserting in the DNA of a micro-organism the nucleotide base sequences GAC or GAT (being known to be a codon for L-Asp) and TTT or TTC (being known to be a codon for L-Phe), preceded and followed by appropriate processing or termination codons in the correct reading frame, and under appropriate control. It therefore has been attempted in EP-A- 0036258 to achieve the synthesis of Asp-Phe indirectly through prior production of protein segments of the formula (Asp-Phe) n, where n is a large number; this has been done by inserting into a cloning vehicle a synthesised DNA fragment coding for such poly- (Asp- Phe) protein. However, such ribosomal fermentative route is still tedious and economically unattractive. Major drawbacks are lying in the recovery of the Asp-Phe dipeptide from the polypeptide. Similar drawbacks can be attributed to a method, as described by Choi, S.-Y. et al. in J. Microbiol. Biotechnol., 2, 1992, p. 1-6, wherein a polypeptide comprising segments of the tripeptide sequence Asp-Phe-Lys is synthesised.

In a recent patent application, WO/0058478 (the contents whereof are herewith incorporated by reference), it has been shown that Asp-Phe can be produced enzymatically, in a direct fermentative route, from the substrates L-Asp and L-Phe with the help of an Asp-Phe dipeptide synthetase. This is done by contacting the substrates, in the presence of an effective amount of adenosine-triphosphate (ATP), with a non-

ribosomal dipeptide synthetase, said synthetase comprising a first and a second minimal module connected by one condensation domain, each of the minimal modules being composed of an adenylation domain and a 4'-phosphopantetheinyl cofactor containing thiolation domain, the first (N-terminal) minimal module recognising L- aspartic acid and the second (-terminal) minimal module recognising L-phenyl lanine and being covalently bound at its N-terminal end to the condensation domain, and said synthetase having a thioesterase releasing factor as an integrated domain at the C- terminus of the second minimal module.

The present inventors, however, have found that the methods as described in WO/0058478 do suffer from the same disadvantage as is generally observed in chemical syntheses of Asp-Phe, namely that the coupling reaction does not lead to very selective formation of the a-Asp-Phe coupling product, but that also a substantial amount, of about 20%, of the Asp-Phe formed is being obtained in the (3-form.

This is disadvantageous, not only because of lowering the yield of the a-Asp-Phe, but also because separation of the a-and p-forms is laborious.

Therefore still a need exists for finding an improved direct fermentative route to Asp-Phe. Direct fermentation of Asp-Phe while obtaining Asp-Phe at an improved oc, (3-ratio (i. e. much better than about 80/20), most preferably while obtaining substantially pure a-Asp-Phe, containing at most 5 wt. % of P-Asp-Phe is hitherto unknown.

Description of the invention improved method for the production of Asp-Phe : Surprisingly, inventors now found new, and promising alternative enzymatical methods for the production-according to WO/0058478-of oc-L-aspartyl-L- phenylalanine (Asp-Phe) from the substrates L-aspartic acid (L-Asp) and L-phenylalanine (L-Phe) by contacting the substrates, in the presence of an effective amount of adenosine-triphosphate (ATP), with a non-ribosomal dipeptide synthetase, said synthetase comprising a first and a second minimal module connected by one condensation domain, each of the minimal modules being composed of an adenylation domain and a 4'-phosphopantetheinyl cofactor containing thiolation domain, the first (N-terminal) minimal module recognising L-aspartic acid and the second (C-terminal) minimal module recognising L-phenylalanine and being covalently bound at its

N-terminal end to the condensation domain, and said synthetase having a thioesterase releasing factor as an integrated domain at the C-terminus of the second minimal module. An important first embodiment of these new and inventive methods is characterised in that > the condensation domain and the adenylation domain of the second minimal module originate from one minimal elongation module recognising L-phenylalanine, and that > the thioesterase releasing factor and the thiolation domain of the second minimal module originate from one C-terminal minimal elongation module, and that > the L-aspartyl-L-phenylalanine (Asp-Phe) formed is being recovered.

This new method thus provides an improved enzymatical process for direct fermentation of Asp-Phe, which in a subsequent methylation step may be converted into the intense sweetener aspartame, substantially free of its (3-isomer. The Asp-Phe recovered in the method according to the invention has an a, (3-ratio much better than 80/20, usually better than 90/10. Accordingly, the Asp-Phe recovered is usually at least for 90% in the a-form. Preferably, the Asp-Phe recovered is at least for 95% in the a-form.

Following the nomenclature as described in WO/0058478 the non- ribosomal dipeptide synthetases according to the present invention, are also indicated herein as Asp-Phe dipeptide synthetases or as Asp-Phe synthetases. It is known (for instance, from P. Zuber et al., in"Bacillus subtilis and other Gram-positive bacteria", Sonenshein et al. (Eds. ), Am. Soc. Microbiol., Washington, DC, 1993, p. 897-916) that micro-organisms can produce bioactive peptides through ribosomal and non-ribosomal mechanisms. The bioactive peptides that, before WO/0058478, were known to be synthesised non-ribosomally, are produced by a number of soil bacteria and fungi.

These bioactive peptides can range from 2 to 48 residue, and are structurally diverse.

They may show a broad spectrum of biological properties including antimicrobial, antiviral or antitumor activities, or immunosuppressive or enzyme-inhibiting activities.

As such, these non-ribosomally synthesised bioactive peptides form a class of peptide secondary metabolites that has found widespread use in medicine, agriculture, and biological research. Already more than 300 different residue thus far have been found to be incorporated into these peptide secondary metabolites. However, until the invention of WO/0058478 not a single non-ribosomally formed peptide had been

identified having (as a part of its peptide sequence) the dipeptide Asp-Phe in it; neither had the dipeptide Asp-Phe itself been identified as a non-ribosomally synthesised product. It is to be noticed, that the Asp-Phe dipeptide synthetases according to the present invention are different from, and can easily be distinguished from, those as constructed in WO/0058478.

According to the present invention a-Asp-Phe that is substantially free of its p-isomer, can now be produced non-ribosomally in an improved way, and novel non-ribosomal Asp-Phe synthetases can be used for the synthesis of such a-Asp-Phe.

Hereinafter, in the part of the specification dealing with the DNA fragments encoding the novel, improved, Asp-Phe synthetases, it will be elucidated in more detail how these novel Asp-Phe synthetases can be obtained and have been made available in the context of the present invention. For better understanding of the present invention, first, however, some general background as to non-ribosomal peptide synthesis is presented.

In non-ribosomal synthesis of peptides generally a multiple carrier thiotemplate mechanism is involved (T. Stein et al., J. Biol. Chem. 271,1996, p. 15428- 15435). According to this model, peptide bond formation takes place on multi-enzyme complexes which are named peptide synthetases and which comprise a sequence of amino acid recognising modules. On the peptide synthetases a series of enzymatic reactions take place which ultimately lead to the formation of a peptide by sequential building-in of amino acids, in an order predetermined by the order of modules recognising the cognate amino acids, into the peptide. This series of enzymatic reactions includes, schematically : 1. recognition of the amino acid substrates ; 2. activation of said recognised amino acid substrates to their aminoacyl- adenylates (that is, the aminoacyl adenosine-monophosphate; aa-AMP) at the expense of Mg2+-ATP (adenylation) ; the adenylation domain (A-domain) is involved in steps 1 and 2; 3. binding of the aminoacyl-adenylates in the form of their more stable thioester to the cysteamin group of the enzyme-bound 4'-phosphopantetheinyl (4'-PP) cofactors (thiolation). The ATP consumed in the adenylation reaction is hereby released in the monophosphate form (AMP) ; generally a thiolation domain (T-domain) is involved in step 3; such T-domain is also referred to in the literature as Peptidyl Carrier Protein (PCP);

4. depending of the peptide to be synthesised non-ribosomally, the thiol-activated substrates may be modified (e. g. by epimerisation or N-methylation) ; various additional domains (which, however, are irrelevant for the present invention) may be involved in step 4 ; 5. formation of the peptide product by N to C stepwise integration of the thioesterified substrate amino acids (modified, as the case may be) into the growing peptide ; the condensation domain (C-domain) is involved in step 5; 6. releasing the peptide formed non-ribosomally from the template ; generally, and specifically for the synthesis of Asp-Phe, a thioesterase domain (Te-domain) is involved in step 6.

Assuming this general scheme also to be correct for the novel non- ribosomal synthesis of a-Asp-Phe according to the present invention, this means that this synthesis involves the subsequent steps of (i) recognition of L-Asp and L-Phe, (ii) formation of an L-aspartyl-and an L-phenylalaninyl-acyladenylate, (iii) binding thereof to the cysteamin group of the 4'-PP cofactor in the respective thiolation domains, (iv) formation of the Asp-Phe dipeptide by transfer of the thioester-activated carboxyl group of L-Asp to the amino group of L-Phe, while the condensation product remains covalently attached to the multi-enzyme complex via the 4'-PP cofactor in the thiolation domain of the Phe-recognising module, and (v) release of the Asp-Phe formed.

According to the present invention the substrates L-Asp and L-Phe are contacted with a non-ribosomal Asp-Phe dipeptide synthetase, in the presence of an effective amount of ATP. An effective amount of ATP as meant herein is an amount of ATP which ensures that the dipeptide formation takes place at a suitable rate. In order to enable an economically attractive process the ATP consumed by the peptide synthesis reaction is preferably regenerated.

The contacting of the substrates L-Asp and L-Phe with the non- ribosomal Asp-Phe dipeptide synthetase may be done in any suitable way ; for instance "if the Asp-Phe dipeptide synthetase is present in a micro-organism-L-Asp and L-Phe may be fed into the culture medium containing said micro-organism. Alternatively micro-organisms may be used which are capable of overproducing L-Asp and/or L-Phe (e. g. from glucose), with separately feeding to the micro-organism of the amino acid (L-Asp or L-Phe) which is not produced by the micro-organism. All these methods may be called in vivo methods. ATP may be regenerated io vivo in the Asp-Phe producing micro-organism, at the expense of a carbon source.

The contacting of the substrates L-Asp and L-Phe with the non-ribo- somal Asp-Phe dipeptide synthetase also may be done by using the synthetase in its isolated form, that is by an in vitro method. In such in vitro methods ATP-regeneration is to be taken care of separately. This may be done by applying an ATP-regeneration system. ATP-regeneration systems are readily available to the skilled man.

Protein chemical studies and recent progress in cloning and sequencing of genes encoding peptide synthetases of bacterial and fungal origin have made it clear that the known peptide synthetases have a highly conserved and ordered structure composed of so-called modules. These modules have been defined as semi- autonomous units within peptide synthetases that carry all information needed for recognition, activation, and modification of one substrate. Although the modules in principle can act independently, it is generally assumed that they have to work in concert, in a template-based mode of action to achieve peptide elongation.

In general, the modules of peptide synthetases, each module being about 1000-1400 amino acids in length (i. e. , the modules have molecular weights in the range of 120-160 kDa), are themselves composed as a linear arrangement of conserved domains specifically representing the enzyme activities involved in substrate recognition, activation, (and, optionally, as the case may be, modification) and condensation (i. e. peptide bond formation). Two of such distinct domains, the adenylation and thiolation domains (A-domain and T-domain), together form the smallest part of a module that retains all catalytic activities for specific activation and covalent binding of the amino acid substrate. Stachelhaus et al. have designated this core fragment of the modules as a"minimal module" (T. Stachelhaus et al., J. Biol.

Chem. , 270,1995, p. 6163-6169).

As mentioned above, an important first embodiment of the methods according to the invention is characterised (in terminology which focuses on keeping together certain domains) in that > the condensation domain and the adenylation domain of the second minimal module originate from one minimal elongation module recognising L-phenylalanine, and that he thioesterase releasing factor and the thiolation domain of the second minimal module originate from one C-terminal minimal elongation module, and that > the L-aspartyl-L-phenylalanine (Asp-Phe) formed is being recovered.

Accordingly, it is not only required that the condensation domain and the adenylation domain of the second minimal module (i. e. the minimal module recognising L-phenylalanine in the Asp-Phe dipeptide synthetase) originate from one minimal elongation module, but more specifically, the condensation domain and the adenylation domain of the second minimal module should originate from one minimal elongation module recognising L-phenylalanine.

As used herein, the term"minimal elongation module"is normally intended to represent a combination-in the N-to C-terminal order as mentioned, and <BR> <BR> <BR> <BR> covalently bound to each other-of condensation, adenylation and thiolation domains (as can be abbreviated by"CAT"), but for a C-terminal module it is specifically intended to represent a combination-also in the N-to C-terminal order as mentioned, and covalently bound to each other-of condensation, adenylation, thiolation and thioesterase domains (as can be abbreviated by"CATTe"). This terminology is also explained in Mootz, H. D. et al., PNAS, 97, 2000, p. 5848-5853. A C-terminal minimal elongation module can alternatively be referred to as a"termination module".

Another way of phrasing this first embodiment of the present invention (in terminology which focuses on the way the various domains are fused together) would be by stating that these methods are being characterised in that the non-ribosomal dipeptide synthetase contains a first fusion site between the thiolation domain of the first minimal module and the condensation domain of the second minimal module, which condensation domain is covalently joined to the adenylation domain of the second minimal module, and > contains a second fusion site between the adenylation and thiolation domains of the second minimal module recognising L-phenylalanine, and that > the L-aspartyl-L-phenylalanine (Asp-Phe) formed is being recovered.

For the purposes of this application, this way of phrasing is considered to be equivalent to that of the wording of claim 1. The term"fusion site"as meant herein, represents a site at which two molecules (protein or DNA, as the case may be) are joined by a covalent bond.

In the context of this invention, it is preferred that the thioesterase releasing factor at the C-terminus of the second minimal module has at least 60% identity, more preferably at least 75% identity, and most preferably at least 90% identity, with the thioesterase releasing factor at the C-terrinus of TycC6.

TycC6 is the sixth module of the 724 kDa TycC, the third polypeptide of the Tyrocidine synthetase. Tyrocidine is a cyclic decapeptide produced by Bacillus brevis ATCC 8185. ! t is further noted that amino acid specificity (or editing function) of condensation domains has been the subject of recent studies, for instance those reported in Mootz, H. D. et Ell*, PNAS, 97, 2000, p. 5848-5853, and in Belshaw, P. J. et al., Science, 284, 1999, p. 486-489.

As used herein, the term"identity" (in percentage figures) indicates the percent identity of two amino acid sequences. These sequences are aligned for optimal comparison purposes (e. g., gaps can be introduced in the sequence of a first amino acid sequence for optimal alignment with a second amino acid sequence). The amino acid residue at corresponding amino acid positions are then compared. When a position in the first sequence is occupied by the same amino acid residue as the corresponding position in the second sequence, then the molecules are identical at that position. The percent identity between the two sequences is a function of the number of identical positions shared by the sequences (i. e. , % identity = number of identical positions/total number of positions (i. e. overlapping positions) x 100). Preferably, the two sequences are the same length.

The skilled person will be aware of the fact that several different computer programs are available to determine the identity between two sequences. For instance, a comparison of sequences and determination of percent identity between two sequences can be accomplished using a mathematical algorithm. In a preferred embodiment, the percent identity between two amino acid sequences is determined using the Needleman and Wunsch (J. Mol. Biol. (48) : 444-453 (1970) ) algorithm which has been incorporated into the GAP program in the GCG software package (available at http ://www. gcg. com), using either a Blossom 62 matrix or a PAM250 matrix, and a gap weight of 16, 14, 12,10, 8,6, or 4 and a length weight of 1,2, 3,4, 5, or 6. The skilled person will appreciate that all these different parameters will yield slightly different results but that the overall percentage identity of two sequences is not significantly altered when using different algorithms.

The nucleic acid sequences of the present invention can further be used as a"query sequence"to perform a search against public databases to, for example, identify other family members or related sequences. Such searches can be performed using the NBLAST and XBLAST programs (version 2.0) of Altschul, et al.

(1990) J. Mol. Biol. 215: 403-10. BLAST nucleotide searches can be performed with

the NBLAST program, score = 100, wordlength = 12 to obtain nucleotide sequences homo) ogous to QWERTY nudeic acid moiecuies of the invention. To obtain gapped alignments for comparison purposes, Gapped BLAST can be utilised as described in Altschul et al., (1997) Nucleic Acids Res. 25(17):3389-3402. When utilising BLAST and Gapped BLAST programs, the default parameters of the respective programs (e. g., XBLAST and NBLAST) can be used. See http ://www. ncbi. nlm. nih. gov.

Most suitable results are achieved when the thioesterase releasing factor at the C-terminus of the second minimal module is the thioesterase releasing factor present at the C-terminus of TycC6. In this case identity is 100%.

The present inventors, surprisingly, have found that a strong correlation exists between the Asp-Phe a, (3-ratio and the initial turn-over number for the non-ribosomal synthesis of Asp-Phe. In particular, the inventors have found, that such initial turn-over number for obtaining substantially pure a-Asp-Phe, containing at most 5 wt. % of p-Asp-Phe, should be higher than a certain minimum value.

It is to be noticed that, although WO/0058478 suggests that usually the rate of the direct fermentative route with the help of an Asp-Phe dipeptide synthetase could be at least one turn-over per minute, i. e. a turn-over number (t. o. n. , or k,,, t) of 1 per minute, and that preferably kCat would be even at least 10 per minute, such rates haven't been reached by far in the methods disclosed according to said document. At most a t. o. n. of about 0,1 min~1 could be achieved (as can be seen in Comparative Example A of the present application. According to the methods of the present invention, however, an improved direct fermentative route to Asp-Phe is provided at a much higher rate than can be realised according to WO/0058478.

Another important, second, embodiment of these methods is, thus, characterised in that the dipeptide synthetase has an initial turn-over number (t. o. n.; also represented as kCat) of at least 0.25 per minute, and that the L-aspartyl-L- phenylalanine (Asp-Phe) formed is being recovered. In this way, as it has been demonstrated, the Asp-Phe recovered is at least for 90% in the a-form. In more preferred embodiments according to the invention the Asp_Phe recovered is even at least for 95% in the a-form.

The initial turn-over number, as used herein, can easily and without undue experimentation be determined in vitro for any Asp-Phe dipeptide synthetase as is to be used in the present invention with the purified, C-terminal His6-tagged synthetase, under the following standard reaction conditions: 500 nM synthetase; pH =

7.8 ; temperature 37 °C ; 20 mM MgCl2 ; 1 mM of each of the substrates L-Asp and L- Phe ; 2 mM ATP; 60 minutes of reaction time. The initial t. o. n. is thus defined as the t. o. n. for the first 60 minutes of reaction under such standard conditions.

Hiss-tagged refers to the iM-or C-terminal extension of a protein, which extension occasional may form part of some larger extension, as is being used by the skilled man in order to facilitate purification of proteins. Within the scope of the present application, instead of Hiss-tagging also tagging with other suitable extensions may be used equivalently. Also maltose, or chitin binding proteins or the like may be used.

According to the present invention, the dipeptide synthetase preferably has an initial turn-over number (t. o. n. ) of at least 0.5 per minute, more preferably of at least 0.65 per minute.

It will be evident that further developments of the present Asp-Phe synthetases as are used in the methods of this invention may lead to further increase in initial t. o. n. due to mutation of the wild-type gene sequences of the (modules of the) Asp-Phe dipeptide synthetase, for instance via directed evolution, or by insertion of or exchange for other wild-type partial sequences, leading to hybrid domains. Alternatively one or more codons might be deleted from the current fragment without substantially affecting the functioning thereof. Moreover, improvements also can be achieved by directed mutagenesis based on structural analysis of the proteins. In future also other types of improvements may be foreseen, which lead to functional expression of heterologously produced peptide synthetases in micro-organisms used for production, more specifically for production of Asp-Phe.

The term"recognising" (e. g. in Asp-recognising or in Phe- recognising) as used for specific modules in the present application, in general means that such module preferentially-i. e. at a higher proportion-recognises (and binds) a specific amino acid (e. g. L-Asp or L-Phe) from a mixture of amino acids present.

Instead of"recognising"also the term"specificity"may be used. Of course, also the concentration of each individual amino acid present near the module may play a role.

If, for instance, the concentration of a specific amino acid is much higher than that of (most of) the other amino acids, the requirements for specificity may be somewhat iess strict. Preferably, such recognising properties for the modules are at least 3 times more efficient in recognising (meaning : in wild-type in vivo function) a specific amino acid (for instance, L-aspartic acid).

In the methods for the production of Asp-Phe according to the present invention preferably also a non-integrated protein with thioesterase Type-ll-like activity is present together with the dipeptide synthetase. As meant herein proteins having thioesterase Type-ll-lilce activity are proteins with strong sequence similarities to type-11 fatty acid thioesterases of vertebrate origin. Such non-integrated protein with thioesterase Type-ll-like activity is differení from the integrated thioesterase (Te-domain). Recent work (Schneider et al., Arch. i\Aicrobiol., 169, 1998, p. 404-410) has shown that deletion of a gene encoding such non-integrated protein with thioesterase Type-ii-like activity from the surfactin synthase operon leads to an 84% reduction of peptide production. It is suggested that the non-integrated protein with thioesterase Type-li-like activity enhances production of non-ribosomal peptides, possibly by reactivation through liberation of mischarged modules that are blocked with an incorrect aminoacyl group or an undesired acyl group at the 4'-PP cofactor.

Genes coding for the non-integrated proteins with thioesterase Type- Il-like activity can be positioned at the 5'-or 3'-end of the peptide synthetase encoding operon. These proteins have molecular masses of 25-29 kDa, are about 220-340 amino acid residues in length, and carry the sequence GxSxG which is presumed to form the active site. It is noticed that in almost all of the prokaryotic peptide synthetase coding operons known so far, such distinct genes have been detected.

In the methods for the production of Asp-Phe according to the present invention, the dipeptide synthetase used is preferably present in living cell material of a micro-organism, and a carbon source (most preferably glucose) and/or L- aspartic acid and/or L-phenylalanine are being fed. The enzymatical production of Asp- Phe according to the invention can be carried out in any suitable type of enzyme reactor, for instance in a vessel or in a fermentor.

As used herein, the term"living cell material"represents cell material capable of self-reproduction. It will be clear, however, that the synthesis according to the invention-when using, instead of the living cell material, an enzyme preparation of the Asp-Phe-synthetase (including such preparation in immobilised form) or permeabilised cells containing such Asp-Phe-synthetase-will not be carried out in a fermentor, but in another suitable type of enzyme reactor.

Feeding of the carbon source, and/or L-aspartic acid and/or L- phenylalanine to the bioreactor (for instance, to the fermentor) can be done by any method known to the skilled man. It can be done simultaneously, intermittently,

continuously or batch-wise, with or without specific control of the feeding rate. Of course, controlled feeding is preferred. The carbon source used most preferably is glucose, but any other energy source, necessary for regeneration of ATP in the living cell material and for the maintenance energy required for said living cell material, might be used instead. Feeding of any of L-aspartic acid and/or L-phenylalanine may be omitted in case a micro-organism is being used which is also capable of fermentative production of L-aspartic acid and/or L-phenylalanine.

The skilled man, of course, will be aware that the feeding of glucose, L-Asp and/or L-Phe is to be done under appropriate conditions of temperature and pH, including as required the presence of an appropriate nitrogen source, salts, trace elements, and other organic growth factors as vitamins and amino acids, etc. to the fermentor or other type of (enzyme) reactor which is used for the production of Asp- Phe. The Asp-Phe formed is recovered. Such recovery may take place during the process or at the end thereof.

In the method for the production of Asp-Phe according to the present invention it is preferred that the micro-organism is first grown, for instance in a fermentor, to reach a predetermined cell density before the expression of the Asp-Phe dipeptide synthetase is switched on and feeding of the glucose, and/or L-Asp and/or L- Phe for the synthesis of the Asp-Phe dipeptide is started.

The skilled man can easily determine the growth of the micro- organism, e. g. by measuring its optical density (O. D. ), and find the most appropriate level of cell density. To prevent any negative effect on the growth of the micro- organism, growth phase and Asp-Phe synthetase production phase are preferably uncoupled. Such uncoupling can be achieved by expressing the gene for the Asp-Phe synthetase from an inducible, tightly regulable, promoter. The expression of the Asp- Phe dipeptide synthetase is preferably switched on by addition of a specific chemical component (inducer) or by depletion of a specific chemical component (repressor), or changing the physical conditions, e. g. the temperature, pH or dissolved oxygen pressure, after a predetermined level of cell density has been reached. The expression is assumed to be switched-on as compared to the non-induced state, if the expression level of the Asp-Phe dipeptide synthetase is raised at least by a factor of 10.

Then also the feeding of substrates, etc. in amounts as required, is started, and production of Asp-Phe starts.

In case the production of Asp-Phe takes place in a fermentor, the micro-organism used is most suitably first grown to reach a predetermined cell density before the expression of the Asp-Phe dipeptide synthetase is switched on (as, for instance, can be done by induction with IPTG or dep) etion oftryptophane) and then feeding of the carbon source, preferably glucose, and/or L-Asp and/or L-Phe for the synthesis of the Asp-Phe dipeptide is started.

As at present already many micro-organisms are being known which are capable of fermentative production of L-Phe, the production of Asp-Phe according to the invention is preferably carried out in an L-phenylalanine producing micro- organism; in such case only the carbon source, preferably glucose, and L-Asp need being fed.

Suitable micro-organisms are, for instance, micro-organisms which (a) are producing peptides via non-ribosomal synthesis, for instance, bacteria as Streptomyces species, Bacillus species, Actinomyces species, Micrococcus species, Nocardia species, or fungal species as Tolypocladium species, Fusarium species, Penicillium species, Aspergillus species, and Cochliobolus species; or (b) are capable of producing amino acids, in particular L-Asp and/or L-Phe, preferably on industrial scale, for instance, Escherichia species, e. g. E. coli, and Corynebacterium species, e. g. C. glutamicum.

The production of Asp-Phe according to the invention is preferably carried out in an Escherichia, Corynebacterium or Bacillus species. In order to prevent decomposition of the Asp-Phe formed, it is preferred that the micro-organism used is a strain with reduced protease activity for Asp-Phe or is lacking such activity towards Asp-Phe. Such strains can be easily constructed by the skilled man, for instance by using protein purification techniques to identify the responsible Asp-Phe degrading enzymes in the respective micro-organism, followed by knocking-out of the genes encoding such Asp-Phe degrading enzymes.

Instead of carrying out the production of Asp-Phe in a fermentor, using living cell material, said production can also suitably be carried out in vitro in an enzyme reactor, while ATP is supplied, and L-Asp and L-Phe are being fed ; also in this embodiment of the invention, the Asp-Phe formed is recovered. The supply of ATP is in such case most suitably provided at least in part by an in situ ATP-regenerating system. The ATP-regenerating system is preferably present in a permeabilised micro- organism.

Various ATP regenerating systems (which in the literature are also being referred to as ATP generating systems) are known to the skilled man. As ATP regenerating systems both whole cell systems (e. g. yeast glycolysis systems) or isolated ATP regenerating enzymes, for instance adenylate kinase combined with acetate I irrase, may be used. A very elegant ATP regeneration system has been described by T. Fujio et al. (Biosci., Biotechnol., Biochem., 61, 1997, p. 840-845). They have shown the use of permeabilised Corynebacterium ammoniagenes cells for regeneration of ATP from the corresponding monophosphate (AMP) coupled to an ATP-requiring reaction in permeabilised E coli cells. In this elegant way (cheap) glucose can be supplied as an energy source instead of most of the ATP.

DNA fragments encoding an Asp-Phe dipeptide synthetase, etc.

The present invention also relates to novel DNA fragments or a combination of DNA fragments encoding an Asp-Phe dipeptide synthetase.

These novel DNA fragments or a combination of DNA fragments code for a non-ribosomal Asp-Phe dipeptide synthetase, which synthetase comprises a first and a second minimal module connected by one condensation domain, each of the minimal modules being composed of an adenylation domain and a 4'- phosphopantetheinyl cofactor containing thiolation domain, the first (N-terminal) minimal module recognising L-aspartic acid and the second (C-terminal) minimal module recognising L-phenylalanine and being covalently bound at its N-terminal end to the condensation domain, and said synthetase having a thioesterase releasing factor as an integrated domain at the C-terminus of the second minimal module. In particular these novel DNA fragments or novel combination of DNA fragments are characterised in that (a) the DNA fragment encoding the condensation domain and the adenylation domain of the second minimal module is derived from one DNA fragment encoding a minimal elongation module recognising L-phenylalanine, and that (b) the DNA fragment encoding the thioesterase releasing factor and the thiolation domain of the second minimal module is derived from one DNA fragment encoding a C-terminal minimal elongation module.

The term"DNA fragment" (whether used as such or in"combination of DNA fragments") as used herein is understood to have its broadest possible meaning. The term first of all relates to the composite biological material (on one or more DNA fragments) as mentioned herein-above and coding for the minimal modules

for Asp and Phe in the correct order and for the condensation domain, each coding sequence being surrounded by any transcription and translation control sequences (e. g. promoters, transcription terminators) and the like which may be suitable for the expression of the Asp-Phe dipeptide synthesising activity. The control sequences may be homologous or heterologous, and the promoter (s) present in the DMA may be constitutive or inducible.

The term"DNA fragment"as used herein is further understood to code, in addition to coding for the Asp and Phe minimal modules and the condensation domain, for the activities of the other domains, e. g. Te-domains. Furthermore, these fragments may code for activities which are not located on the Asp-Phe dipeptide synthetase polypeptide itself, such as non-integrated thioesterase Type-))-iike proteins, and other activities co-operating concertedly with the Asp and Phe minimal modules.

The term"DNA fragment"as used herein is also understood to comprise gene structures comprising DNA fragments as described herein-above. More precisely, a gene structure is to be understood as being a gene and any other nucleotide sequence which carries the DNA fragments according to the invention.

Appropriate nucleotide sequences can, for example, be plasmids, vectors, chromosomes or phages. The gene structures may exist either as (part of) an autonomously replicating vector in single or multicopy situation, or integrated into the chromosome in single or multicopy situation.

The gene structure is also to be understood as being a combination of the above-mentioned gene carriers, such as vectors, chromosomes or phages, on which the DNA fragments according to the invention are distributed. For example, the Asp-Phe dipeptide synthetase encoding DNA fragment can be introduced into the cell on a vector and the non-integrated thioesterase Type-ii-like protein encoding DNA fragment can be inserted into the chromosome. In addition, a further DNA fragment can, for example, be introduced into the cell using a phage. These examples are not intended to exclude other combinations of DNA fragment distributions from the invention. The DNA fragments according to the invention may be introduced into the micro-organism at a sufficiently high copy number, for instance of up to 50 copies.

A detailed discussion of the Asp-Phe dipeptide synthetase, all relevant domains present therein (A-, C-, T-and Te-domains), and the two minimal modules comprised therein already has been given in the preceding parts of this patent application, and is further being discussed in detail in the recent patent application,

WO/0058478 (especially at pages 24-27), the contents whereof-as has been stated earlier in this application-are deemed to be incorporated by reference.

In the DNA fragments or combination of DNA fragments (coding for an Asp-Phe dipeptide synthetase) according to the present invention, preferably the DNA fragment encoding the condensation domain and the adenylation domain of the second minimal module is fused in-frame to the DNA fragment encoding the first minimal module recognising L-aspartic acid. The term"fused in-frame"means that the fused DNA fragments, due to covalent bonding, form part of one open reading frame.

As already mentioned earlier in this application, the term "recognising" (e. g. in Asp-recognising or in Phe-recognising) as used for specific modules encoded by the DNA fragments, in general means that such module preferentially-i. e. at a higher proportion-recognises (and binds) a specific amino acid (e. g. L-Asp or L-Phe) from a mixture of amino acids present. Instead of"recognising" also the term"specificity"may be used. Of course, also the concentration of each individual amino acid present near the module may play a role. If, for instance, the concentration of a specific amino acid is much higher than that of (most of) the other amino acids, the requirements for specificity may be somewhat less strict. Preferably, such modules are at least 3 times more efficient in recognising (in their wild-type in vivo function) a specific amino acid (for instance L-aspartic acid) than in recognising any other amino acid.

More preferably, the DNA fragments or combination of DNA fragments coding for an Asp-Phe dipeptide synthetase according to the invention are characterised in that the DNA fragment encoding the thioesterase releasing factor and the thiolation domain of the second minimal module encodes for a thioesterase- releasing factor with at least 60% identity, more preferably at least 75% identity, and still more preferably at least 90% identity, with the thioesterase releasing factor at the C-terminus of TycC6.

The term"identity" (in percentage figures) as used herein has the same meaning as has been indicated hereinabove.

Most preferably, the DNA fragments or combination of DNA fragments encoding the thioesterase releasing factor and the thiolation domain of the second minimal module encodes for the thioesterase releasing factor as is present at the C-terminus of TycC6.

In a particularly preferred embodiment of the DNA fragments or combination of DNA fragments according to the invention, such fragments also code for a non-integrated protein with thiossterase Type-ll-like activity.

The invention further relates to micro-organisms containing a DNA fragment or combination of DNA fragments according to the invention, and in particular to such micro-organisms which are capable of producing L-Asp and/or L-Phe. In particular, the micro-organism is an Escherichia, Corynebacterium or Bacillus species.

The present invention finally also relates to novel Asp-Phe dipeptide synthetases. The terms and expressions used hereinafter with respect to the Asp-Phe dipeptide synthetase all have the same meaning as explained herein-above.

The non-ribosomal Asp-Phe dipeptide synthetases according to the present invention comprise a first and a second minimal module connected by one condensation domain, each of the minimal modules being composed of an adenylation domain and a 4'-phosphopantetheinyl cofactor containing thiolation domain, the first (N-terminal) minimal module recognising L-aspartic acid and the second (C-terminal) minimal module recognising L-phenylalanine and being covalently bound at its N- terminal end to the condensation domain, and said synthetase having a thioesterase releasing factor as an integrated domain at the C-terminus of the second minimal module, and are particularly characterised in that the condensation domain and the adenylation domain of the second minimal module originate from one minimal elongation module recognising L-phenylalanine, and that > the thioesterase releasing factor and the thiolation domain of the second minimal module originate from one C-terminal minimal elongation module.

In a particularly preferred embodiment of the invention, the thioesterase releasing factor of the second minimal module has at least 60% identity, more preferably at least 75% identity, and still more preferably at least 90% identity, with the thioesterase releasing factor at the C-terminus of TycC6.

It will be clear, that all kinds of particular embodiments as have been claimed in method claims 1-15, or in the DNA fragment claims 16-20, or in the recombinant micro-organism claims 21-23, but have not explicitly been claimed in the Asp-Phe dipeptide synthetase claims 24 and 25, are deemed to be specifically claimed as well in further subclaims dependent from claims 24 and 25, with all amendments as would be necessary because of change to a different category of claims.

Experiments) part GENERAL PROCEDURES Standard molecular cloning techniques such as DNA isolation, gel electrophoresis, enzymatic restriction modifications of nucleic acids, E. coli transformation etc., were performed as described by Sambrook et al., 1989,"Molecular Cloning : a laboratory manual", Cold Spring Harbor Laboratories, Cold Spring Harbor, New York and Innis et al., 1990,"PCR protocols, a guide to methods and applications", Academic Press, San Diego. Synthetic oligo deoxynucleotides were obtained from MWG-Biotech AG, Ebersberg, Germany. DNA sequence analyses were performed on an Applied Biosystems ABI 310 genetic analyzer, according to supplier's instructions.

Sequencing reactions were carried out by the chain termination method with dye- labelled dideoxy terminators from the PRISM ready Reaction DyeDeoxy Terminator cycle sequencing kit with AmpliTaq FS polymerase (Applied Biosystems).

CONSTRUCTION OF PLASMIDS In the following parts the construction of all plasmids is described, thereby first describing the construction of pATCsrfB2-AtycA-TTesrfC-His6, hereinafter also referred to as construct &num 1, in the same way as the construction of this plasmid has been described in the experimental section of patent application WO/0058478; in that application, this plasmid was designated pasp-phe-TE-His6. For convenience of comparison with said earlier patent application the terminology of said earlier application has been retained for the construction of construct &num 1, which has been used in Comparative Example A, hereinafter.

In the sequence listings hereinafter shown (for [SEQ ID : No. 1] to [SEQ ID : No. 29] ), each time the relevant restriction sites are marked by underlining.

Construction of pasp-phe-TE-His6 (construct #1) occurred in a number of steps : A. Construction of plasmid pasp-leu-Hisfj A 4919 bp fragment comprising regions from the srfB locus from chromosomal Bacillus suM/sATCC 21332 (accession number X72672) DNA was 1bp number corrected (in comparison to bp number in WO/0058478) according to new accession number X72672 instead of X70356

amplified (PCR) using the following primers: 5' - TAAGCATGCTGCTTTCATCTGCAGAAAC - 3' (5' asp-leu-Sph1-srfB2) [SEQ ID: No. 1] 5' - AATGGATCCTTCGGCACGCTCTAC - 3' (3' asp-leu-BamHI-snB3)2 [SEQ ID: No.2].

Correct size of the amplified fragment was confirmed by agarose gel electrophoresis.

The fragment (20 µg) was digested with 1 unit of the enzymes BamHI/Sphl (37°C, 16 h) to generate terminal restriction sites.

Plasmid pQE70 (provided by Qiagen, D-Hilden) (10 µg) was digested with the same enzymes and subsequently incubated for 1 hour with 1 unit Alkaline Phosphatase (37°C). Complete digestion was confirmed by transforming 1 LL of the linearised plasmid DNA into competent cells of E coli XL1 blue. The two fragments were subsequently ligated in a ligation reaction (10 IlL) in a vector/insert ratio of 1: 3 with 1 unit of T4-DNA-ligase enzyme (16°C, 16 h).

L of the ligation mixture was used to transform 40 pi. competent cells of E coliXL1 blue (Stratagene, D-Heidelberg) by electroporation. The transformants were selected on 2x YT agar plates containing Ampicillin (100 ug/mL). Analysis of 48 transformants resistant to ampicillin revealed that 4 of them had inserted a ca. 5000 bp fragment. Correct insertion was confirmed using restriction enzyme digestion analysis and terminal sequencing of the insert. A correct clone designated pasp-leu-His6 was used for further investigations.

B. Construction of plasmid pasp-phe-His6 Plasmid pasp-phe-His6 was constructed from plasmid pasp-leu-His6 as follows.

A 1895 bp chromosomal DNA-fragment3 from Bacillus brevis ATCC 8185 (accession number AF004835) DNA was amplified (PCR) using the following primers: 5'-ATTTGGTCACCAATCTCATCGACAA-3' (5'BstEll-TycA-NLID) SEQ ID : No. 3] in the nucleotide sequence of this primer 3'-was changed into 5'- (clerical error in previous application WO/0058478) 3bp number corrected (in comparison to bp number in WO/0058478) according to recalculation [for correction of obvious error] from accession numberAF004835

5' - ATAGGATCCTGTATTCGTAAAGTTTTTC - 3' (3'-PheAT-BamHI) [SEQ ID : No. 4].

Correct size of the fragment was confirmed using agarose gel electrophoresis.

The fragment was digested with 1 unit of enzyme BamHI and incubated at 30°C for 4 hours. Subsequently 1 unit of enzyme BstEll was added and incubated for another 4 hours at 60°C.

Plasmid pasp-leu-His6 was digested in the same way and subsequently incubated for 1 hour with 1 unit of Alkaline phosphatase. The vector portion (ca. 6,5 kb) was separated from other DNA fragments by agarose gel electrophoresis and repurified.

Complete digestion was confirmed as before with linearised pasp-leu-His6. The two fragments were ligated in an equimolar ratio for 5 hours at 16°C using 1 unit of T4-ligase enzyme. 1 IlL of the ligation mixture was used for electroporation of E. coli XL1 blue competent cells. Transformants were selected on 2x YT agar containing Ampicillin (100 pg/mL). Analysis of transformants revealed that 1 out of 90 clones had inserted a fragment of ca. 2000 bp. Correct insertion was confirmed using restriction enzyme digestion analysis and terminal sequencing of the insert. The correct clone was designated pasp-phe-His6.

C. Construction of plasmid pasp-phe-TE-His6 Plasmid pasp-phe-TE-His6 was constructed from plasmid pasp-phe- His6.

A 913 bp chromosomal DNA-fragment4 from Bacillus subtilis ATCC 21332 (accession number X70356) DNA was amplified (PCR) using the following primers: 5'-ATAATCGATAATCGCACAAATATGGTC-3' (5'TE-srfC1-Clal) [SEQ ID : No. 5] 5'-ATAAGATCTAACAACCGTTACGGTTTGTGT-3' (3'int TE-srfC1-Bglll) [SEQ ID : No. 6].

Correct size of the fragment was confirmed using agarose gel electrophoresis.

The fragment was digested with 1 unit of enzyme Coal for 4 hours at 37°C, before adjusting buffer conditions and digesting with 1 unit of enzyme BglII (4 4bp number corrected (in comparison to bp number in WO/0058478) according to recalculation [for correction of obvious error] from accession number X70356

hours, 37°C).

Plasmid pasp-phe-His6 was digested with enzyme Clal (4 h, 37°C) and subsequently with BamHl (4 h, 37°C) before the linearised plasmid was incubated for one hour with 1 unit of Alkaline phosphatase. The vector portion (ca. 8 kb) was separated from other DNA-fragments by agarose gel electrophoresis and repurified.

Control of complete digestion, ligation, electroporation and selection of transformants was established as described before.

Two of the analysed transformants were shown to contain the desired DNA-fragment. Correct insertion of the 900 bp fragment was confirmed by restriction enzyme analysis and terminal sequencing of the insert. A correct clone was designated pasp-phe-TE-His6. Hereinafter this will be designated as pATCsrfB2-AtycA-TTesrfC-His6 (or as construct &num 1).

Construction of plasmid has been used in Example 1 hereinafter, via intermediate constructs a-e Construction of plasmid pATsrfB2-CAtycB2-TTetycC6-His6 (construct #2) was performed via five intermediate plasmid constructs (a-e): intermediate construct a: pATEeA-Hise, intermediate construct b: pAbacA1-TEtycA-His6, intermediate construct c: pAbacA-TCATTetyc6-His6, intermediate construct d : pAbacA1-TTetycC6-His6, and intermediate construct e: pAbacA1-TCAtycB2-TTetycC6-His6.

Construction of intermediate construct a Construction of intermediate construct a (plasmid pATECA-His6) was started from plasmid pQE60 (Qiagen), which was digested with 1 unit of each of the enzymes BamHI and Ncol (37°C, 16h) to generate terminal restriction sites. The pQE60 vector fragment that has a size of 3431 bp was subsequently purified by silica gel using the QlAquick Spin Purification Kit (Qiagen) according to the instructions of the manufacturer. The 3263 bp ATE cA-fragment was amplified from the chromosomal DNA of Bacillus brevis ATCC 8185 (accession number AF004835) using the following primers :

5'-TATCCATGGTAGCAAATCAGGCCA-3' (5 tycA-A-Ncol) [SEQ ID : No. 7] 5' - ATAGGATCCAAGCAATTCGAAGATATC - 3' (3' tycA-E-BamHI) [SEQ ID : No¢8] The PCR fragment was digested with 1 unit each of the enzymes BamHI and Ncol (37°C, 16h) and subsequently purified by silica gel using the QlAquick Spin Purification Kit (Qiagen) according to the instructions of the manufacturer.

The two purified fragments (vector containing fragment and ATEtycA fragment) were subsequently ligated in a ligation mix (total volume 1 0p1) containing 1 unit of T4-DNA-ligase enzyme (16°C, 20h). 1, ul of the ligation mixture was used to transform 40 ul competent cells of E. coli XL1 blue via electroporation (Stratagene, Heidelberg, Germany). The transformants were selected on LB agar plates containing ampici)) in (100ug/mL). A clone containing the correct ATECA insert fragment as demonstrated by restriction enzyme digestion pattern analysis and terminal sequence determination was designated pATE «, CA-His6 and used for further construction work.

Construction of intermediate construct b Construction of the intermediate construct b (plasmid pAbacA1-TEcA- His6) started from plasmid pATEtycA-His6 (intermediate construct a). A 5120 bp fragment comprising the pQE vector and the TECA gene was amplified by PCR from plasmid pATE «, CA-His6 (intermediate construct a; see above) using the following primers: 5' - AGCCTGCAGGCCTACCATCCTCCGAG - 3' (5'tycA-T-Pstl) [SEQ ID : No. 9] 5' - TGGACCCATGGTTAATTTCTCCTCT - 3' (3' T5-Ncol) [SEQ ID: No. 10] The 5120 bp PCR fragment was digested with 1 unit each of the enzymes Pstl and Moot (37°C, 16h) to generate terminal restriction sites and subsequently purified by silica gel using the QlAquick Spin Purification Kit (Qiagen) according to the instructions of the manufacturer.

The 1610 bp AbacA1-fragment was amplified from the chromosomal DNA of B. licheniformis ATCC 10176 (accession number AF007865) using the following primers : 5' - TTTCCATGGCTAAACATTCATTAGA - 3' (5' bacA1-A-Ncol) [SEQ ID : No. 11] 5' - TTCCTGCAGCGCCCCCGCCGTTCTG - 3' (3' bacA1-A-Pstl) [SEQ ID : No. 12] The PCR fragment was digested with 1 unit each of the enzymes Pstl and Ncol (37°C, 16h) to generate terminal restriction sites and subsequently purified by silica gel using the QlAquick Spin Purification Kit (Qiagen) according to the instructions of the manufacturer. Both the ligation of the fragments (vector containing fragment and AbacA1 fragment) and the transformation were performed as described for intermediate construct a. A clone containing the correct insert fragment as demonstrated by restriction pattern analysis and terminal nucleotide sequence determination, was designated pAbacA1-TEtycA-His6 and used for further construction work.

Construction of intermediate construct c Construction of intermediate construct c (plasmid pAbacA1- TCATTeece-Hise) started from intermediate construct b (pAbacA1-TEtycA-His6), which was digested with 1 unit of each of Pstl and BamHl (37°C, 16h). The 5025 bp fragment comprising vector pQE60 and the AbacA1 gene was subsequently purified by electrophoresis.

The 4131 bp TtycC5-CATTetycC6-fragment was amplified from the chromosomal DNA of Bacillus brevis ATCC 8185 (accession number AF004835) using the following primers : 5' - ATACTGCAGGAGTATGTAGCGCCGC - 3' (5' tycC5-T-Pstl) [SEQ ID : No. 13] 5' - TATGGATCCTTTCAGGATGAACAGTTCTTG - 3' (3'cC6-Te-6amHf) [SEQ ID : No. 14]

The PCR fragment was digested with 1 unit of each of the enzymes Pstl and BamHI (37°C, 16h) to generate terminal restriction sites and subsequently purified by silica gel using the QlAquiclc Spin Purification Kit (Qiagen) according to the instructions of the manufacturer. The iigation of the vector containing fragment and the TtycC5-CATTetycC6 fragment and subsequent transformation were performed as described for intermediate construct a. A clone containing the correct insert fragment as demonstrated by restriction pattern analysis and terminal nucleotide sequence determination, was designated pAbacA1-TCATTetycC6-His6 and used for further construction work.

Construction of intermediate construct d The 6032 bp fragment comprising vector pQE, the AbacA1 gene and the TTecc6 gene, was amplified by PCR from plasmid pAbacA1-TCATTetycC6-His6 (intermediate construct c) using the following primers: 5'-ACCGTTAACGAATACGTGGCCCCGAG-3' (5'tycC6-T-Hpal) [SEQ ID : No. 15] 5'-AATGTTAACCTCCTGCAGCGCCCC-3' (3 bacA1-A-Hpal/Pstl) [SEQ ID : No. 16] The desired PCR fragment was digested with 1 unit of the enzyme Hpal (37°C, 16h) to generate terminal restriction sites and subsequently purified by silica gel using the QlAquick Spin Purification Kit (Qiagen) according to the instructions of the manufacturer. Self-ligation of this fragment and subsequent transformation were performed as described for intermediate construct a. A clone containing the correct insert fragment as demonstrated by restriction pattern analysis and terminal nucleotide sequence determination, was designated pAbacA1-TTetycC6-His6 and used for further construction work.

Construction of intermediate construct e Construction of intermediate construct e (pAbacAi-TCA-TTetycce- His6) started from intermediate construct d, which was digested with 1 unit each of the enzymes Pstl and Hpal to generate terminal restriction sites. Then the 6013 bp fragment comprising vector pQE, the AbacA1 gene and the TTetycC6 gene was

subsequently purified by silica gel using the QlAquick Spin Purification Kit (Qiagen) according to the instructions of the manufacturer.

The 3117 bp 7 bp TtycB1-CAtycB2-fragment was amplified from the chromosomal DNA of Bacillus brewis ATGC 8185 (accession number AF004835) using the following primers : 5' - ACGCTGCAGGATTACGTCGCCCCGA - 3' (5' tycB1-T-Pstl) [SEQ ID : No. 17] 5' - AGCGTTAACTGTTGCAGGCTTTCCTTC - 3' (3' tycB2-A-Hpal) [SEQ ID : No. 18] The TtycB1-CAtycB2-fragment was digested with 1 unit of each of the enzymes Pstl and Hpal (37°C, 16h) to generate terminal restriction sites and subsequently purified by silica gel using the QlAquick Spin Purification Kit (Qiagen) according to the instructions of the manufacturer. Ligation of the vector containing fragment and the TtycB1-CAtycB2-fragment as well as subsequent transformation were performed as described for intermediate construct a. A clone containing the correct insert fragment as demonstrated by restriction pattern analysis and terminal nucleotide sequence determination, was designated pAbacA1-TcA, cB2-TTetycc6-His6 and used for further construction work.

Construction of plasmid pATsrfB2-CAtycB2-TTetycC6-His6 (construct #2) (This construct is used in Example 1, hereinafter) Construction of plasmid pATsrfB2-CAtycB2-TTetycC6-His6 (construct #2) started from intermediate construct e (plasmid pAbacA1-TcAcB2-TTetyCc6-His6) A 7298 bp fragment comprising vector pQE, the CACB2 gene and the TTecce gene was amplified by PCR from intermediate construct e using the following primers : 5' - ATAGATATCGAGGAAAGCGCGTATCTCG - 3' (5' tycB2-C-EcoRV) [SEQ ID : No. 19] 5' - TGGACCCATGGTTAATTTCTCCTCT - 3' tv3e T5-Ncol) [SEQ ID : No. 10]

The vector-containing fragment was digested with 1 unit of each of the enzymes EcoRV and Ncol (37°C, 16h) to generate terminal restriction sites and subsequently purified by silica gel using the QIAquick Spin Purification Kit (Qiagen) according to the instructions of the manufacturer.

The 1826 bp ATsrfB2-fragment was amplified from the chromosomal DNA of Bacillus subtilis ATCC 21332 (accession number X72672) using the following primers : 5'-TAACCATGGTGCTTTCATCTGCAGAAAC-3' (5'srfB2-A-Ncol) [SEQ ID : No. 20] 5'-TATGATATCCTCCATATAAGCCGC-3' (3'srfB2-T-EcoRV) [SEQ ID : No. 21] The ATs B2-fragment was digested with 1 unit of each of the enzymes EcoRV and Ncol (37°C, 16h) to generate terminal restriction sites and subsequently purified by silica gel using the QlAquick Spin Purification Kit (Qiagen) according to the instructions of the manufacturer. Ligation of the vector containing fragment and the ATsBz-fragment and subsequent transformation was performed as described for intermediate construct a. A clone containing the correct insert fragment as demonstrated by restriction pattern analysis and terminal nucleotide sequence determination, was designated pATsrfB2-CAtycB2-TTetycC6-His6 and used for further investigations Construction of plasmid pATsrfB2-CAtycB2-TTesrfC-His6 (constructt #3) (This construct is used in Example 2, hereinafter) Plasmid pATsrfB2-CAtycB2-TTesrfC-His6 is based on plasmid pATsrfB2- CAcB2-TTecC6-HiSe (construct &num 2, see above). The 945 bp TTesrfC-fragment was amplified from the chromosomal DNA of Bacillus subtilis ATCC 21332 (Accession number X70356) using the following primers : 5' - TATGTTAACTGGATTGGACCGCGGAAC - 3' (5'srfC-T-Hpal) [SEQ ID : No. 22]

5' - TATGGATCCTGAAACCGTTACGGTTTGTG - 3' (3' srfC-Te-BamHI) [SEQ ID : No. 23] The TTesrfC-fragment was digested with 1 unit of each of the enzymes Hpal and BamHl (37°C, 16h) and subsequently purified by silica gel using the QlAquick Spin Purification Kit (Qiagen) according to the instructions of the manufacturer. Plasmid pATsrfB2-CAtycB2-TTetycC6-His6 (construct #2) was digested in the same way. The vector containing fragment (size 8100 bp) was subsequently purified by gel electrophoresis.

Ligation of the he vector containing fragment and the TTesrfC-fragment and the subsequent transformation were performed as mentioned before (see construction of intermediate construct a). A clone containing the correct insert fragment as demonstrated by restriction enzyme digestion pattern and terminal nucleotide sequence determination, was designated pATsrfB2-CAtycB2-TTesrfC-His6, and used for further investigations.

Construction of plasmid pATsrfB2-CAtycB2-TtyC6-His (construct #4) (This construct is used in Comparative Example B, hereinafter) Plasmid pATsrfB2-CAtycB2-TtycC6-His6 is based on plasmid pAT,, m2- CAtycB2-TTetycC6-His6 (construct #2, see above). The 240 bp Tcc6-fragmeni : was amplified from the chromosomal DNA of Bacillus brevis ATCC 8185 (Accession number AF004835) using the following primers: 5'-TATGTTAACGAATACGTGGCCCCGAG-3' (5'tycC6-T-Hpal (2) ) [SEQ ID : No. 24] 5'-TATGGATCCGAAATCGGCCACCTTTTCG-3' (3'tycC6-T-BamHl) [SEQ ID : No. 25] The desired PCR fragment was digested with 1 unit of each of the enzymes Hpal and BamHi (37°C, 16h) and subsequently purified by silica gel using the QlAquick Spin Purification Kit (Qiagen) according to the instructions of the manufacturer. The plasmid pATsrfB2-CAtycB2-TTetycC6-His6 (construct #2) was digested in the same way. The vector-containing fragment (size 8110 bp) was subsequently purified by gel electrophoresis. Ligation of the vector containing fragment and the T cce- fragment and subsequent transformation were performed as described for intermediate construct a. A clone containing the correct insert fragment as demonstrated by

restriction enzyme digestion pattern and terminal nucleotide sequence determination, was designated pATsrfB2-CAtycB2-TtycC6-His6, and used for further investigations.

Construction of plasmid pATsrfB2-(CATtycB2-TetyC6-His6 (construct #5) (This construct is used in Comparative Example C, hereinafter) The 1826 bp ATsrfB2-fragment was amplified from the chromosomal DNA of Bacillus subtilis ATCC 21332 (Accession number X727672) using the following primers : 5' - TAACCATGGTGCTTTCATCTGCAGAAAC - 3' (5' srfB2-A-Ncol) [SEQ ID : No. 20] 5'-TATGGATCCCTCCATATAAGCCGC-3' (3' srfB2-T-BamHI) [SEQ ID : No. 26] The desired PCR fragment was digested with 1 unit of each of the enzymes Ncol and BamHI (37°C, 16h) subsequently purified by silica gel using the QlAquick Spin Purification Kit (Qiagen) according to the instructions of the manufacturer. Vector pQE60 (Qiagen) was digested and subsequently purified in the same way. Ligation of the vector fragment and the ATssz-fragment and subsequent transformation were performed as described for intermediate construct a, resulting in the plasmid pATs The 3117 bp CATtycB2-fragment was amplified from the chromosomal DNA 8185 (Accession number AF004835) using the following primers : 5' - ATTAGATCTGAGGAAAGCGCGTATCTCG -3' (5' tycB2-C-Bg/II) [SEQ ID : No. 27] 5' - AATAGATCTTTCGATCAAGCGGGCCAAG -3' (3' tycB2-T-Bg/II) [SEQ ID : No. 28] The desired PCR fragment was digested with 1 unit of the enzyme Bg/II (37°C, 16th) and subsequently purified by silica using the QIAquick Spin Purification Kit (Qiagen) according to the instructions of the manufacturer. Plasmid

pATss2-His6 was digested with BamHl and Bglll (37°C, 16h) and subsequently purified by silica gel using the QlAquick Spin Purification Kit (Qiagen) according to the instructions of the manufacturer. Ligation of the 5231 bp vector containing fragment and the CATcB2-fragment and subsequent transfrrnation were carried out as described for intermediate construct a, resulting in plasmid pATsrfB2-CATtycB2-His6.

The 744 bp bp TetycC6-fragment was amplified from the chromosomal DNIA from Bacillus brevis ATCC 8185 (Accession number AF004835) using the following primers : 5' - TAAAGATCTGCCATTTTGTTAAATCAG - 3' (5' tycC56-Te-Bg/II) [SEQ ID : No. 29] 5'-TATGGATCCTTTCAGGATGAACAGTTCTTG-3' (3'tycC6-Te-BamHl) [SEQ ID : No. 14] The desired PCR fragment was digested with 1 unit of each of the enzymes Bglll and BamHl (37°C, 16h) and subsequently purified by silica gel using the QlAquick Spin Purification Kit (Qiagen) according to the instructions of the manufacturer. Plasmid pATsfB2-CATcB2-His6 was digested with Bglll and subsequently purified by silica gel using the QlAquick Spin Purification Kit (Qiagen) according to the instructions of the manufacturer. Ligation of the vector containing fragment and the Tehees-fragment and subsequent transformation were performed as described for intermediate construct a. A clone containing the correct insert fragment in the correct orientation as demonstrated by restriction enzyme digestion pattern and terminal nucleotide sequence determination, was designated pATsrfB2-CATtycB2-TetycC6-His6, and used for further investigations.

Expression of the peptide synthetases 1 ; j. L of each constructed plasmid were transformed in E coli BL21/pgsp competent cells. Strain BL21 kDE3 was obtained from Stratagene, D- Heidelberg. Plasmid pgsp, which is based on plasmid pREP4 (obtained from Qiagen, D- Hilden), contains the gsp gene (the 4'-PP transferase gene from the Gramicidin S- biosynthesis operon from Bacillus brevis ATCC 9999) under control of the T7 promoter.

Transformants were selected on 2x YT agar plates containing ampicillin (100 lig/mL) and Kanamycin (25 lig/mL). Several colonies were used to inoculate 4 mL of

2x YT liquid medium (containing in addition 10 mM MgCI2) and incubated at 37°C for 16 hours. These 4 mL cultures were subsequently used to inoculate 400 mL of the same medium. Cells were grown at 30°C in a waterbath shaker (250 rpm) After 3-4 hours the cells reached an optical density of 0, 7 (ODeoonm) and were induced by the addition of 200 , M IPTG. Cells were incubated for an additional 1, 5 hours before being harvested.

Expression of recombinant proteins was confirmed by SDS-PAGE comparing protein samples taken at the time of induction and 1, 5 hours later.

From cultures expressing the correct recombinant proteins glycerol stocks were prepared and stored at-80°C.

Purification of the recombinant peptide synthetases 800 mL cultures of all recombinant strains, obtained as described in "Expression of the peptide synthetases...", were centrifuged at 5000 rpm for 5 minutes and resuspended in 30 mL/L culture of buffer A (50 mM HEPES, 300 mM NaCI, pH 8, 0).

Cell suspensions were used directly or were stored at-20 °C till usage. Cell lysis was established using two French press passages at a working pressure of 12000 psi.

Directly after cell lysis PMSF was added to a final concentration of 1 mM. After centrifugation of the cell lysates at 10000 rpm for 30 minutes, the supernatant was combined with 1% (v/v) buffer B (50 mM HEPES, 300 mM NaCI, 250 mM Imidazol, pH 8,0). Protein solutions were applied on a Ni2+-NTA-agarose column (Qiagen, D-Hilden) previously equilibrated with 1% (v/v) buffer B. Flow rate was 0,75 mL/min. After the non- His6-tagged proteins had passed through the column, it was washed with 1% buffer B for another 10 min before a linear gradient was applied (30 min to 30% B, an additional 10 min to 100% B). All peptide synthetases eluted at a concentration of about 5% buffer B (15 mM Imidazol) and were collected as 2 mL fractions.

Fractions containing the recombinant peptide synthetases were detected using the Bradford reagent, by the absorption at 595 nm. These fractions were pooled. The imidazol in the pooled fractions for constructs #1-5 was removed with Hi-trap desalting columns (Pharmacia) using a buffer containing 50 mM HEPES, 20 mM MgCl2 and 2 mM DTT. The flow rate was 5mUmin. The protein solution was collected in 2mL fractions, and the amount of enzyme in the fractions was determined using the Bradford reagent, by the absorption at 595 nm. Fractions containing the recombinant peptide synthetases were pooled and protein concentrations were determined again.

Till further usage proteins were stored at-20°C after addition of glycerol

to 10% (v/v). Grade of purification in all cases was estimated to be 95% by SDS-PAGE.

HPLC-MS analysis Assay conditions : 50 mM HEPES (pH 8, 0) 20 mM MgCl2 500 nM purified Asp-Phe synthetase from respective constructs &num 1 to 9-5 2 mM ATP 1 mM Asp 1 mM Phe The reaction mixture was incubated at 37°C. Samples of 100 uL were taken at certain time points. The reactions in the samples each time were quenched by addition of 1 00pi n-butanol and the prectipitated protein was removed. The remaining clear solutions were then dried and the pellets were resuspended in 1 OOPL 10% methanol.

HPLC analysis was performed using HPLC-MS techniques (1100 HPLC-Systems and HP series 1100 MSD, Hew/ett Packard) by injecting 10 uL of the sample onto a C 250/3 Nucleosil 120-3 C18 3 column (Macherey & Nagel).

HPLC-program: 0 min 10% buffer B 25 min 60% buffer B 30 min 95% buffer B 34 min 95% buffer B 36 min 10% buffer B 40 min 10% buffer B with flow rate: 0.3 mL/min, buffer A : demineralised H20, with 0. 1% trifluoro acetic acid (TFA), and buffer B: Methanol, containing 0.1% TFA.

Detection was done using electrospray inonization in the positive mode. Asp-Phe was identified in the single ion mode (S) M) scanning for 175, 221, 235, 281, 303 and 319 amu (atomic mass units) with a dwell time of 95 msec. The retention time of a-Asp-Phe was 18, 0 minutes, whereas that of ß-Asp-Phe was 17, 2 minutes.

The amount of Asp-Phe was determined by integration of the HPLC- MS signals and calculated by comparison to solutions (dilution series) of a chemical

Asp-Phe standard with different concentrations.

The a : P-Asp-Phe ratios shown in table 1 were calculated by comparison of the integrals of the HPLC-MS signals for the respective components.

Both the turn-ower numbers and the aç sp-Phe ratios as listed in table 1 were calculated for samples taken after incubation of 60 minutes.

Results of the assays (Examples and Comparative Examples) The results of the assays (Examples 1 and 2, carried out with constructs &num 2 and &num 3) and (Comparative Examples A, B and C, carried out with constructs #1, #4 and #5are shown in table 1.

Table 1 Example/Construct Construct Turn-over a : p-Asp-Phe Comp. Ex. No. number ratio 1 #2 ATsrfB2CAtycB2TTetycC6 0, 7 min-1 99:1 2 #3 ATsrfB2CAtycB2TTesrfC 0, 27 mini 92 : 8 A #1 ATCsrfB2AtycATTesrfC 0,09 min-1 82 : 18 [pasp-phe-TE-His6] 5 B #4 ATsrfB2CAtycB2TtycC6 0, 01 min~ 81 : 19 C #5 ATsrfB2CATtycB2TetycC6 0,08 min-1 75 : 25 5 nomenclature according to WO/0058478

SEQUENCE LISTING <110> Holland Sweetener Company V. O. F. <120> MICROBIOLOGICAL PRODUCTION METHOD FOR alpha-L-ASPARTYL-L-PHENYLALANINE <130> 21455 <160> 29 <170> PatentIn version 3. 1 <210> 1 <211> <212>'DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 1 taagcatgct gctttcatct gcagaaac 28 <210> 2 <211> 24 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 2 aatggatcct tcggcacgct ctac 24 <210> 3 <211> 25 <212> DNA

<213> Artificial Sequence <220> <223> Synthetic DNA <400> 3 atttggtcac caatctcatc gacaa 25 <210> 4 <211> 28 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> ataggatcct gtattcgtaa agtttttc <210> 5 <211> 27 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 5 ataatcgata atcgcacaaa tatggtc 27 <210> 6 <211> 30 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 6

ataagatcta acaaccgtta cggtttgtgt 30 <210> 7 <211> 24 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 7 tatccatggt agcaaatcag gcca 24 <210> 8 <211> 27 <212> DNA <213> Sequence <220> <223> Synthetic DNA <400> 8 ataggatcca agcaattcga agatatc 27 <210> 9 <211> 26 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 9 agcctgcagg cctaccatcc tccgag 26 <210> 10 <211> 25 <212> DNA <213> Artificial Sequence

<220> <223> Synthetic DNA <400> 10 tggacccatg gttaatttct cctct 25 <210> 11 <211> 25 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 11 tttccatggc taaacattca ttaga 25 <210> 12 <211> 25 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 12 ttcctgcagc gcccccgccg ttctg 25 <210> 13 <211> 25 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 13 atactgcagg agtatgtagc gccgc 25

<210> 14 <211> 30 <212> <213> Artificial Sequence <220> <223> Synthetic DNA <400> 14 tatggatcct ttcaggatga acagttcttg 30 <210> 15 <211> 26 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 15 accgttaacg aatacgtggc cccgag 26 <210> 16 <211> 24 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 16 aatgttaacc tcctgcagcg cccc 24 <210> 17 <211> 25 <212> DNA <213> Artificial Sequence

<220> <223> Synthetic DNA <400> 17 acgctgcagg attacgtcgc cccga 25 <210> 18 <211> 27 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 18 agcgttaact gttgcaggct ttccttc 27 <210>'19 <211> 28 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 19 atagatatcg aggaaagcgc gtatctcg 28 <210> 20 <211> 28 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 20 taaccatggt gctttcatct gcagaaac 28 <210> 21

<211> 24 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 21 tatgatatcc tccatataag ccgc 24 <210> 22 <211> 27 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 22 tatgttaact ggattggacc gcggaac <210> 23 <211> 29 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 23 tatggatcct gaaaccgtta cggtttgtg 29 <210> 24 <211> 26 <212> DNA <213> Artificial Sequence <220>

<223> Synthetic DNA <400> 24 tatgttaacg aatacgtggc cccgag 26 <210> 25 <211> 28 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 25 tatggatccg aaatcggcca ccttttcg 28 <210> 26 <211> 24 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 26 tatggatccc tccatataag ccgc 24 <210> 27 <211> 28 <212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 27 attagatctg aggaaagcgc gtatctcg 28 <210> 28 <211> 28

<212> DNA <213> Artificial Sequence <220> <223> Synthetic DNA <400> 28 aatagatctt tcgatcaagc gggccaag <210> 29 <211> 27 <212> DNA <213> Artificial Sequence <220> <223> DNA <400> 29 taaagatctg ccattttgtt aaatcag 27