Background of the invention: 13-Lactam antibiotics constitute the most important group of antibiotic compounds, with a long history of clinical use. Among this group the prominent 13-lactam antibiotics are the penicillins with a five-membered ring and the cephalosporins with a six-membered ring, which are condensed with a four-membered 13-lactam ring. Penicillins consist of the so-called 13-lactam nucleus, which is linked through its primary amino group in 6-position to the so-called side chain forming an amide bond. Cephalosporins carry an analogous amide bond in 7-position, but they also vary by different substituents in 3-position.

The conventionally chemical production of semi synthetic 13-lactam antibiotics is afflicted with complicated technical requirements and the use of noxious chemicals causing environmental pollution as well as impurities in the product.

The synthesis routes are long as it is necessary to introduce protection groups, before modifications can be performed. Since many steps are thus involved, the overall yields are quite low.

There is a great interest for enzymatic routes as alternatives, since enzymatically catalysed reactions are highly selective. Thus the production of many by-products and the effluent and purification problems, which result there from, are reduced or avoided. Furthermore such processes are performed in aqueous environment.

The semi-synthetic routes mostly start from fermentation products such as penicillin G (Pen G), penicillin V (Pen V) and cephalosporin C (Ceph C).

With economical and ecological benefits chemical hydrolysis of penicillin G/V forming the key intermediates 6-aminopenicillanic acid (6-APA) and 7-amino- deacetoxycephalosporanic acid (7-ADCA) was replaced by enzymatic methods using penicillin G/V amidase (PGA or PVA). The other missing key intermediate 7-aminocephalosporanic acid (7-ACA) starting from cephalosporin C is now also on the way to be replaced by a biocatalytical method using a D-amino acid oxidase (DAO) and in a second step a glutaryl-7-ACA acylase (GAC).

Up to now the amide coupling of the 13-lactam key intermediates with side chains forming semi synthetic 13-lactam antibiotics is done chemically, even though there are efforts for developing enzymatic methods.

These are based mainly on catalysis by penicillin amidase or also called penicillin acylase (EC 3.5. 1.11), which can be obtained from Acetobacter, Achromobacter, Aerobacter, Aerococcus, Aeromonas, Alcaligenes, Aphanocladium, Arthrobacter, Azotobacter, Bacillus, Beneckea, Brevibacterium, Cellulomas, Cephalosporium, Clostridium, Co7zanaonas, Corynebacterium, Erwinia ; Escherichia coli, Flavo- bacterium, Gluconobacter suboxidans, Klebsiella aerogenes, Kluyvera citrophlia, Microbacterium lacticum, Micrococcus, Mycoplana, Norcardia, Paracoccus, Paracolon bacilli, Piocianus, Protaminobacter, Proteus, Pseudobacterium, <BR> <BR> <BR> Pseudonzonas, Rhodococus rlaodochrous, Rhodopseudomonas, Sacina, Samonella, Serratia, Shigella, Soil bacterium, Spirillum, Staphylococcus, Streptomyces, Thiobacillus, Xantomonas or Yersinia species [Review: Sudhakaran V. K., Borkar P. S.; Microbial transformation of beta-lactam antibiotics: Enzymes from bacteria, sources and study-a sum up; Hindustan Antibiotics Bulletin 27 (1-4), 63-119 ; 1985]. Some approaches use a-amino acid esterase (EC 3.1. 1.43), which can be obtained from Acetobacter, Pseudomonas, Streptomyces species or a Xantomonas species [van der Does, T.; An enzymatic process for preparing 13-lactam compounds; WO 02/20819 A2; 2000]. The penicillin amidase or the a-amino acid esterase may be used as free enzymes or in any suitable immobilised form.

Penicillin G amidase is normally used for the hydrolysis of Pen G or Ceph G forming the key intermediates (ß-lactam nuclei) 6-APA or 7-ADCA for semi synthetic 13-lactam antibiotics. It is also a powerful tool for hydrolysis, e. g. of phenylacetyl protected derivatives, or synthesis, e. g. of semi synthetic 13-lactam antibiotics with side chains based on acetic acid substituted with an aromatic residue like the natural product phenylacetic acid (PAA). It may be also used for racemic resolution of amino functions: The thermodynamical equilibrium follows the charge state of the acid e. g. PAA.

A not charged side chain (e. g. by low pH or adding solvent) is a substrate for synthesis. A dissociated or protonated side chain is no substrate and only hydrolysis takes place. Therefore amino acids like p-hydroxyphenyl glycine or phenylglycine are no substrates for synthesis, since they are at any pH charged:

+H3N-CHR-COOH # H3N-CHR-COO-X H2N-CHR-COO-, but no H2N-CHR-COOH Only in activated form as ester or amide amino acids can be transferred to a not charged form suitable for synthesis: +H3N-CHR-COOR (not reactive) X H2N-CHR-COOR (reactive) +H3N-CHR-CONH2 (not reactive) ¢ H2N-CHR-CONH2 (reactive) Other acids without amino function are only suitable for synthesis at acid pH conditions, since they are there not dissociated. Activated as esters or amides these side chains are at any pH possible substrates. At least reactivity is higher and reaction time shorter due to improved Gibb's energy.

R-CH2-COOH (reactive) X R-CH2-COO- (not reactive) R-CH2-COOR (reactive) R-CH2-CONH2 (reactive) Unfortunately PGA catalyses also hydrolysis of these activated compounds: R-CH2-COOR + H2OPGA> R-CH2-COOH + ROH R-CH2-CONH2 + H2OPGA> R-CH2-COOH + RNH2 The esters or amides are also hydrolysed by the synthesis of the 13-lactam antibiotics followed by their bydrolysis, both catalysed by PGA: R-CH2-COOR + H2N-NucleusPGA> R-CH2-CONH-Nucleus + ROH R-CH2-CONH2 + H2N-Nucleus PGA> R-CH2-CONH-Nucleus + NH3 R-CH2-CONH-Nucleus + H20 -"-> R-CH2-COOH + H2N-Nucleus For the enzymatical ß-lactam synthesis it has to be distinguished between the thermodynamically and kinetically controlled approach [Review: Kasche V.; Mechanism and yields in enzyme catalysed equilibrium and kinetically controlled synthesis of 13-lactam antibiotics, peptides and other condensation products; Enzyme Mircob. Technol. 8,4-16 ; 1986].

In the thermodynamically controlled synthesis free acids are used as side chains.

The reaction is slowly, since low pH values have to be applied, where PGA is not so active. At the end the thermodynamically defined equilibrium is reached, which is conveniently for the stopping criteria. This approach works only with amidases, not with a-amino acid esterase, since this enzyme catalyses only an a-amino acyl transfer. As said before the equilibrium constant follows the charge state of the side chain, which depends on the ks of the acid and the pH value. The dissociation constant can be shifted to more favourable values by water soluble

solvents as ions are less well hydrated. In contrast the ks is reduced by high ionic strength as ions are better stabilised. Water soluble solvents have also an direct influence on the equilibrium, since they reduce the water activity and therefore water, which is a by-product of synthesis, is apparently removed. Also due to the law of mass action the yield increases by the concentration of ß-lactam nucleus and side chain. Temperature and enzyme input have mainly influence on reaction time, but less on yields. It is a common statement, that water soluble solvents increase yields, but our experiments confirmed this only for low substrate concentrations. At high substrate input, we surprisingly found, that yields are reduced by water soluble solvents.

The kinetically controlled synthesis with esters or amides is much faster, since the Gibb's energy of the activated side chains is improved and more alkaline pH values may be applied, where the amidase or a-amino acid esterase is more active. The higher Gibb's energy is also the reason for yields better than thermo- dynamic equilibrium. But these condensations have to be terminated just in time, since the formation of ! 3-lactam antibiotics reaches a maximum. The reason is the hydrolysis of activated side chain by hydrolysis or synthesis followed by 13-lactam hydrolysis. When all esters or amides are consumed, the conversion will reach at final the thermodynamic equilibrium. At the yield maximum the rate of synthesis from 13-lactam nucleus and activated side chain is equal to the rate of hydrolysis of the condensed product. Very often the effects act controversial e. g. a higher pH, higher temperature or more enzyme input increases the activity for hydrolysis and synthesis. In practice a pH around 6 to 7.5 and low temperatures e. g. 10°C are optimal for good yields. For enzyme input a compromise between reaction time and yield has to be defined e. g. 4-8 h until the yield maximum is reached works well. Also high substrate concentration will enhance synthesis followed by faster hydrolysis, but better yields will be achieved at higher load. Water soluble solvents and ionic strength will have same effect as discussed before.

The side chain has to have similarities with phenylacetic acid: - residue with s-electrons (phenyl-, pyridyl-, thienyl-, tetrazolyl-, CN-etc.) - short spacer (-CH2-,-CHOH-,-CHNH2-,-OCH2-,-SCH2-) - carboxylic function or derivative (-COOH,-COOR,-CONHR) The enzymes will convert only dissolved reactants. Therefore their solubility has a great influence on reaction rate and yields.

The history of enzymatic synthesis of 13-lactams started already in 1960, as PGA was described and its use for hydrolysis of penicillin G. Also the reverse (thermodynamically controlled) reaction starting from 6-APA and PAA was

examined [Rolinson G. N. , Batchelor F. R. , Butterworth D. , Cameron-Wood J., Cole M. , Eustace G. C. , Hart M. V. , Richards M. , Chain E. B.; Formation of 6-aminopenicillanic acid from penicillin by enzymatic hydrolysis; Nature 187, 236-237; 1960] [Claridge C. A. , Gourevitch A. , Lein J.; Bacterial penicillin amidase; Nature 187, 237-238; 1960]. The kinetically controlled synthesis was described firstly in 1969, when the amides (-NH2), N-glycine amides (-NH-CH2- COOH), N-hydroxyethyl amides (-NH-CH2-CH2-OH), thioesters of thioacetic acid (-S-CH2-COOH) and methylesters (-O-CH3) of phenylacetic acid, D-phenyl- glycine, D/L-mandelic acid, p-hydroxyphenylacetic acid and 3,4-dihydroxy- phenylacetic acid were condensed with 6-APA [Cole M.; Factors affecting the synthesis of ampicillin and hydroxypenicillins by the cell-bound penicillin acylase of Esclzerichia coli ; Biochem. J. 115,759-764 ; 1969]. Further esters (ethyl, n-/i- propyl, t-butyl and thioglycol) for activation of D-phenylglycine, p-hydroxy- phenylglycine, 4-hydroxy-3,5-dichorphenylglycine and cyclohexenylglycine were also condensed with 6-APA [Takahashi, T. , Takahashi T. , Isono M. (Takeda), Verfahren zur Herstellung von Aminopenicillinen; Deutsche Offenlegungsschrift DT 2 163 792; 1970]. The use of 7-ADCA and 7-ACA was introduced 1971 for the kinetically controlled synthesis with several activated side chains, which were already noted before except 2-thienylglycine methyl ester [Takeshi T. , Koichi K., Masao 1. (Takeda); Method for the production of cephalosporins; United States Patent US 3,816, 253; 1971]. Also 7-amino-3-chloro-3-cephem-4-carboxylic acid was condensed with phenylglycine methylester forming cefaclor [Baldaro E. M.; Effect of temperature on enzymatic synthesis of cephalosporins; Bioorg. Chem.

Healtcare Technol. , 237-240; 1991]. The only synthesis without aromatic or cyclohexenyl function at the side chain used 225 mM cyanoacetic acid methyl- ester (CNM) and 18 or 41 mM (12.5 or 5.5 fold) deacetyl-7-ACA (DA-7-ACA) giving 98 or 70 % deacetyl cefacetril [Konecny J. , Schneider A. , Sieber M.; Kinetics and mechanism of acyl transfer by penicillin acylases ; Biotechnol.

Bioeng. 25,451-467 ; 1983]. Enzymatic cefacetril synthesis itself was never described before.

The enzymatic synthesis of cephalotin was firstly described 1972. At the thermodynamically controlled synthesis using 20 mM 7-ACA and up to 160 mM thienylacetic acid (THH) at pH 6 and 37°C maximum 30 % cephalotin yield were reached. For the kinetically approach using 80 mM thienyl acetyl glycine in a 7 fold faster conversion 50 % yield were obtained [Takahashi, T. T. , Kawahara K. , Takahashi T. S. , Kato K. , Yamazaki Y. T. (Takeda); Verfahren zur Herstel- lung von Cephalosporinen ; Deutsche Offenlegungsschrift DT 2 360 265; 1972].

Yields up to 96 % crude cephalotin were described by using immobilised PGA on

a column. This was achieved by charging 26-fold thienylacetic acid methylester for the kinetically controlled synthesis, which is of cause a hindrance for economic usage [Fujii T. , Shibuya Y. (Toyo Jozo. ) ; Enzymatic production of cephalothin ; United States Patent US 3,853, 705; 1972]. By activation with low molecular polyethylene glycolesters up to 85 % yield were achieved (18 mM 7-ACA, 78 mM thienylacetic acid tetraethylene glycolester, pH 7 and 37°C using PGA on a column) [Kondo E. , Mitsugi T. , Fujiwara T. , Muneyuki R. (Shionogi) ; Enzymatisches Verfahren zur Herstellung von ß-Lactamantibiotika und Acyl- verbindungen zur Durchführung des Verfahrens; Deutsche Offenlegungsschrift DE 3035467 Al ; 1979]. At thermodynamically controlled synthesis 95 % cephalotin yield were obtained by charging a high side chain excess (15 mM 7-ACA and 400 mM THH), applying > 30 % solvents (examples with 50-65 % dioxane, 65 % acetone, 50 % butanediol, 70 % ethanol and 50-60 % DMF) and passing the solution through a column with immobilised PGA at 4-37°C and a initial pH between 5 and 8 (pH 6 for cephalotin) [Guisan J. M., Fernandez- Lafuente R. , Alvaro G. , Blanco R. M., Rosell C. M; Method for the synthesis of semi-synthetic antibiotics in thermodynamically controlled water-cosolvent organic miscible apolar systems by using penicillin G acylase; European Patent Application EP 0 458 932 A1 ; 1989]. The low 7-ACA concentration was caused by its maximum solubility. By applying pH 7.2 and reducing the pH during synthesis to pH 6.4 the 7-ACA input was enhanced to 50 mM in 50 % DMF.

While charging only 4-fold THH also 95 % cephalotin yield were achieved [Fernandez-Lafuente R. , Alvaro G. , Blanco R. M. , Guisan J. M.; Equilibrium controlled synthesis of cephalothin in water-cosolvent systems by stabilized penicillin G acylase; Appl. Biochem. Biotechnol. 27,277-290 ; 1991].

Cefoxitin is expected to be formed in an analogous reaction from THH and 3-aminocarbonyloxymethyl-7-methoxy-3-cephem-4-carbonic acid (7-MACCA), but this was never described before.

Cefamandol maybe formed by firstly 3-processing of 7-ACA with l-methyl-5- mercaptotetrazole (MMTZ) forming 7-amino-3-(1-methyl-lH-tetrazol-5-yl) thio- methyl-3-cephem-4-carboxylic acid (TZA) followed by 7-processing with D-mandelic acid (MAH) or alternatively by firstly 7-processing with MAH forming mandoyl-7-ACA (MA-7-ACA) followed by 3-processing with MMTZ.

MAH in form of its methylester (MAM, 5 fold molar ratio) was condensed with TZA forming only 8 % cefamandol, where the L-enantiomer gave 35 % yield [Fuganti C., Rosell C. M. , Rigoni R. , Servi S. , Tagliani A. , Terreni M.; Penicillin acylase mediated synthesis of formyl cefamandole; Biotechnol. Letters 14 (7),

543-546; 1992]. By converting 50 mM 7-ACA with 25 mM MAM at pH 7 and 4°C 25 % MA-7-ACA yield were reported [Fernandez-Lafuente R. , Rosell, C. M., Guisan J. M. ; The use of stabilised penicillin acylase derivatives improves the design of kinetically controlled synthesis; J. Mol. Catal. A: Chem. 101 (1), 91-97; 1995]. Applying 50 mM 7-ACA and 150 mM MAM (3 fold) at pH 6.5 and 4°C 69 % MA-7-ACA were obtained. In presence of 20 % methanol and 1 M phosphate buffer yield was improved up to 80 %. Without isolation 3-processing with MMTZ to cefamandol was performed at 65°C in 60 % yield [Terreni M., Pagani G. , Ubiali D., Fernandez-Lafuente R. , Mateo C. , Guisan J. M.; Bioorg.

Med. Chem. Lett. 11 (18), 2429-2432 (2001); Modulation of penicillin acylase properties via immobilization techniques: one-pot chemoenzymatic synthesis of cephamandole from cephalosporin C]. The thermodynamical approach with free MAH and the kinetically controlled synthesis by activation as hydroxyethylester (MAG, produced from ethylene glycol) was never described before. <BR> <BR> <BR> <BR> <P>7-p-hvdroxyphenlglycyl-amino-3- (1-methyl-1 H-tetrazole-5-yl) thiomethyl-3-ceph- em-4-carboxylic acid (HP-TZA) is an intermediate for processing with 2,3-dioxo- 4-ethyl-1-piperazin-carbonyl chloride to cefoperazon or with 4-hydroxy-6-methyl- nicotinoyl chloride to cefpiramid. It has the same substitution in 3-position, but carries D-p-hydroxyphenylglycine (HPH) in 7-position. The reaction sequence starts either via 3-processing of 7-ACA with MMTZ forming TZA followed by 7-processing with HPH to HP-TZA or via 7-processing of 7-ACA with HPH forming p-hydroxyphenylglycyl-7-ACA (HP-7-ACA) followed by 3-processing with MMTZ to HP-TZA. The thermodynamically controlled synthesis is not pos- sible as HPH is an amino acid, but only uncharged side chains are accepted by PGA as substrate. As mentioned already before most often the methylester (HPM) is used for activation of HPH. For cefprozil, cefadroxyl and amoxicillin the activation of HPH as hydroxyethylester (HPG, produced from ethylene glycol) charged in approximately a 3-fold ratio is introduced giving 90-99 % yield [Usher J. J. , Romancik G., Bristol-Meyers Squibb; PCT Int. Appl. WO 98/04732 Al (1996); Synthesis of 13-lactam antibacterials using soluble side chain esters and enzyme acylase]. No example is described for enzymatic condensation of HPH in any activated form with 7-ACA or TZA.

Cefatrizin is an other product from 7-ACA and HPH. The reaction sequence starts either via 3-processing of 7-ACA with 4-mercapto-1, 2,3-triazole (MTZ) forming 3- [ [ (1, 2,3-triazol-4-yl) thio] methyl-7-ACA (MTA) followed by 7-processing with HPH or via 7-processing of 7-ACA with HPH forming HP-7-ACA followed by

3-processing with MTZ. No example is described for enzymatic condensation of HPH in any activated form with (7-ACA or) MTA.

Cefazolin maybe formed by firstly 3-processing of 7-ACA with 2-mercapto-5- methyl-1, 3,4-thiadiazole (MMTD) forming 7-amino-3- (5-methyl-1, 3,4-thiadia- zole-5-yl) thiomethyl-3-cephem-4-carboxylic acid (TDA) followed by 7-proces- sing with tetrazolylacetic acid (TZH) or alternatively by firstly 7-processing of 7-ACA with TZH forming tetrazolyl-7-ACA (TZ-7-ACA) followed by 3-pro- cessing with MMTD. The thermodynamically controlled synthesis is not possible as TZH is a strong acid, but only uncharged side chains are accepted by PGA as substrate. For the kinetical approach using 6 mM TZH in form of the methylester (TZM) with 2 mM 6-APA, 7-ADCA, 7-ACA or 7-amino-3-pyridylmethyl-3- cephem-4-carboxylic acid at pH 6.5 and 37°C up to 80 % yield were noted [Shimizu M. , Okachi R. , Kimura K. , Nara T.; Purification and properties of penicillin acylase from Kluyvera citrophila ; Agr. Biol. Chem. 39 (8), 1655-1661; 1975]. Up to 55.7 % cefazolin yield were reported starting from tetrazolylacetic acid pentaethylene glycolester and TDA [Muneyuki R. , Mitsugi T. , Kondo E.; Water-soluble esters: useful enzyme substrates for the synthesis of 13-lactam. antibiotics; Chem. Industry 7,159-161 ; 1981]. By coupling 29 mM TDA with a 1 to 5 fold ratio of various esters at pH 6.8 and 35°C up to 63.8 % cefazolin yield were obtained. The high temperature is obviously chosen in order to dissolve TDA in such concentration at this low pH [Kostadinov M. , Nikolov A. , Tsoneva N. , Petkov N.; New tetrazole-1-acetic acid esters for enzymatic synthesis of cefazolin; Appl. Biochem. Biotechnol. 33,177-182 ; 1992]. Applying 50 mM 7-ACA and 150 mM TZM (3 fold) at pH 6.5 and 4°C 98 % tetrazolyl-7-ACA (TZ-7-ACA) were obtained. Without isolation 3-processing with MMTD to <BR> <BR> <BR> cefazolin was performed at 65°C in 70 % yield [Justiz O. H. , Fernandez-Lafuente R. , Guisan J. M. , Negri P. , Pagani G. , Pregnolato M. , Terreni M.; J. Org. Chem., 62 (26), 9099-9106 (1997); One-pot chemoenzymic synthesis of 3'-functionalized cephalosporines (cefazolin) by three consecutive biotransformations in fully aqueous medium].

Ceftezol has the same substitution as cefazolin in 7-position, but is 3-processed with 2-mercapto-1, 3,4-thiadiazol (2MTD). The reaction sequence starts either via 3-processing of 7-ACA with 2MTD forming 3- [ [ (1, 3,4-thiadiazol-2-yl) thio] methyl-7-ACA (2TDA) followed by 7-processing with TZM or via 7-processing of 7-ACA with TZM forming TZ-7-ACA followed by 3-processing with 2MTD.

No example is described for the unexpected enzymatic condensation of 2TDA with TZM.

Summary of the invention: We examined the enzymatic synthesis of cephalosporins using PGA starting from 7-ACA or 3-processed derivatives thereof.

We found much improved conditions for the kinetically controlled synthesis of TZ-7-ACA from 7-ACA with TZH preferably activated as methylester.

Kinetically controlled synthesis of cefazolin from TDA and TZH preferably activated as methylester is hindered by the low solubility of TDA at reasonably pH. At higher pH, where TDA is soluble, TZM is hydrolysed fast by PGA. We could surprisingly overcome this problem by charging preferably 50-60 mmol/1 TDA, applying a pH gradient and feeding solid TZM during synthesis.

We describe the first time the synthesis of : - ceftezol from 2TDA with TZH preferably activated as methylester - cefacetril from 7-ACA and cyanoacetic acid preferably activated as methylester We found improved conditions for the kinetically controlled synthesis of cephalotin from 7-ACA with THH preferably activated as methylester.

Description of the invention: The invention is related with an enzymatic process under kinetical control for coupling the 7-amino group of a cephalosporin nucleus with the carboxylic function of an acetic acid derivative consisting of a: - residue with n-electrons (phenyl-, pyridyl-, thienyl-, tetrazolyl-, CN-etc. ) - short spacer (-CH2-,-CHOH-, CHNH2-,-OCH2-,-SCH2-etc.) - carboxylic function (-COOH) activated as ester (-COOR), which is characterised by: using a penicillin amidase (EC 3.5. 1.11) or a-amino acid esterase (EC 3.1. 1.43) from Acetobacter, Achromobacter, Aerobacter, Aerococcus, Aeromonas, Alcaligenes, Aphanocladium, Arthrobacter, Azotobacter, <BR> <BR> <BR> Bacillus, Beneckea, Brevibacterium, Cellulomas, Cephalosporium,<BR> <BR> <BR> <BR> <BR> <BR> Clostridiuni, Comamonas, Co7ynebacterium,. Erwinia, Escherichia coli,<BR> <BR> <BR> <BR> <BR> Flavobacterium, Gluconobacter suboxidans, Klebsiella aerogenes, Kluyvera citrophlia, Microbacterium lacticum, Micrococcus, Mycoplana, Norcardia, Paracoccus, Pas-colon bacilli, Piocianus, Protaminobacter, Proteus, Pseudobacterium, Pseudomonas, Rhodococus rhodochrous, Rhodopseudomonas, Sacina, Sa7aonella, Serratia, Shigella, Soil bacteriuna,

Spi°illum, Stapl7ylococcus, Streptomyces, Thiobacillus, Xantomonas or Yersinia species as free enzyme or in any suitable immobilised form, applying at least 100 mmol/1 cephalosporin nucleus as long as it is soluble, otherwise 50-100 % of its maximum solubility is charged, applying the free acetic acid derivative activated as ester in a 1-3 fold molar ratio, adjusting pH 6.0-8. 0 by alkaline, > adjusting 0-30°C.

As described above, the process of the invention is carried out using the following reaction components: the acetic acid derivative is TZH, the cephalosporin nucleus is 7-ACA, TDA or 2TDA and the formed product is TZ-7-ACA, cefazolin or ceftezol. the acetic acid derivative is cyanoacetic acid, the cephalosporin nucleus is 7- ACA and the formed product is cefacetril. the acetic acid derivative is thienylacetic acid, the cephalosporin nucleus is 7-ACA and the formed product is cephalotin.

> the acetic acid derivative is activated as methylester. It is charged in one portion or fed during synthesis.

Thus, we describe the first time the synthesis of : - ceftezol from 2TDA with TZH preferably activated as methylester - cefacetril from 7-ACA and cyanoacetic acid preferably activated as methylester The following examples are for illustration purposes only and in no way limit the scope of this invention: Example 1: kinetically controlled synthesis of TZ-7-ACA : The experiments were performed in 1000 ml scale with 1.5 kU/1 PGA-450,150 mM 7-ACA, 262.5 mM TZM, pH 7.0 by 2 mol/1 NaOH, 10°C and 250 rpm:

Surprisingly with only a 1. 75-fold ratio of TZM to 7-ACA 94.4-96. 2 % yield are obtained. Same yield was reported before charging 3-fold the side chain (see background of invention), which is much less economic.

Example 2: influence of TDA concentration on kinetically controlled synthesis of cefazolin : The experiments were performed in 400 ml scale with 5 kU/1 PGA-450, equimolar amounts of TDA and TZM, but feeding after 0. 5,1. 0,1. 5 and 2.0 h additional half equimolar amounts of TZM resulting in a 3-fold ratio to TDA, pH 7.5 until 0.5 h, pH 7.35 until 1.0 h, pH 7.2 until 1.5 h, pH 7.1 until 2. 0 h and afterwards pH 7.0 by 2 mot/1 NaOH, 10°C and 250 rpm:

Kinetically controlled synthesis of cefazolin from TDA and TZH preferably acti- vated as methylester is hindered by the low solubility of TDA at reasonable pH.

At higher pH, where TDA is soluble, TZM is hydrolysed fast by PGA. We could surprisingly overcome this problem by charging preferably 50-62.5 mM TDA, applying a pH gradient and feeding solid TZM during synthesis. 83.7 % maxi- mum cefazolin yield is obtained after 2.5 h.

Example 3 : kineticallv controlled synthesis of ceftezol : The experiments were performed in 350 ml scale with 5 kU/1 PGA-450,100 mM 2TDA, 200 mM TZM fed in 5 parts a start, 4 parts after 0.5 h, 3 parts after 1 h, 2 parts after 1.5 h and 1 part after 2 h, pH 8 at start, adjusted afterwards to pH 7.5 by 2 mol/1 NaOH, 10°C and 250 rpm:

Similar as in the example before, solubility of 2TDA is limited at reasonable pH.

At higher pH, where 2TDA is soluble, TZM is hydrolysed fast by PGA. We could surprisingly overcome this problem by feeding solid TZM during synthesis at pH 7.5. 84.1 % maximum ceftezol yield is obtained after 2.5 h.

Example 4: kinetically controlled synthesis of cefacetril : The experiment was performed in 400 ml scale with 5 kU/1 PGA-450, 150 mM 7-ACA, 300 mM CNM, pH 7 by 2 mol/1 NaOH, IO°C and 250 rpm. As no standard was available, it was assumed that the molar extinction coefficient of cefacetril is equal to 7-ACA. Cyanoacetic acid and its methylester cannot be determined by HPLC due to a missing chromophore:

We describe the first time the synthesis of cefacetril from 7-ACA and cyanoacetic acid activated e. g. as methylester. 89. 3 % maximum cefacetril yield is obtained after 8. 2 h.

Example 5: kineticallv controlled synthesis of cephalotin: The experiment was performed in 400 ml scale with 3 kU/1 PGA-450, 100 mM 7-ACA, 175 mM THM, pH 6.75 by 2 mol/1 NaOH, 10°C and 400 rpm:

Surprisingly with only a 1. 75-fold ratio of THM to 7-ACA 84.3-86. 2 % yield are obtained. The high 7-ACA input in combination with low side chain excess disclose a more economic cephalotin synthesis at kinetical control.