Process for Producing 7-Aminocephem Compound or Salts Thereof

This invention relates to a process for producing 7-aminocephem compound or salts thereof. More particularly, the invention relates to a process for producing 7-aminocephem compound or salts thereof, a vector to be used in said process, cephalosporin compound-producing microorganisms transformed with said vector, and a DNA fragment having the promoter activity of the Acremonium chrysogenum alkaline protease gene, among others.

7-Aminocephem compound of the formula: ##STR3## wherein R is an acetoxy, hydroxy or hydrogen, is the most important starting material for the production of semisynthetic cephalosporin antibiotics and is worldwidely used in a number of pharmaceutical factories throughout the world. At present, the 7-aminocephem compound (I) is produced by a two-step process which comprises the step of cultivating in a nutrient medium a cephalosporin compound-producing strain belonging to the species Acremonium chrysogenum and capable of producing a cephalosporin compound of the formula: ##STR4## wherein R is as defined above and X is --CH(NH.sub.2)--COOH, --CO--COOH or --COOH, and recovering the cephalosporin compound (II) from the cultured broth and the step of chemically or enzymatically eliminating the acyl group at 7 position such as the .alpha.-aminoadipoyl group, from the cephalosporin compound (II). For effecting the latter step deacylation chemically, the so-called iminoetherification process (F. M. Huber et al., "Cephalosporins and Penicillins, Chemistry and Biology", page 27, Academic Press, 1972), which involves a series of chemical reactions, is used for eliminating the 7-position acyl group such as the .alpha.-aminoadipoyl group of the cephalosporin compound (II). Regrettably, however, this process requires an expensive chemical plant and complicated operations. Enzymatic deacylation also requires complicated operations as well as a separate fermentation plant.

Accordingly the present inventors made investigations in an attempt to produce 7-aminocephem compound (I) directly in one step without the above-mentioned deacylation step by causing microorganisms to fermentatively produce the 7-aminocephem compound (I). Any microorganism capable of directly producing any of the 7-aminocephem compound (I) has not yet been discovered in the natural world in spite of efforts made by researchers all over the world. No report has-been presented as yet about a success in creating an 7-aminocephem compound (I)-producing microorganism by modifying a microorganism by means of mutation treatment or using genetic engineering techniques, because the difficulty of the task is great.

Facing such difficult task, the present inventors made intensive investigations, after which a vector for producing 7-aminocephem compound (I) which contains one or more promoter(s) for Acremonium chrysogenum and gene(s) for enzyme(s) capable of converting the cephalosporin compound (II) to the corresponding 7-aminocephem compound (I), with said gene(s) connected to said promoter(s), was newly constructed and the cephalosporin compound (II)-producing microorganism belonging to the species Acremonium chrysogenum was transformed with said vector and cultivated. Surprisingly the desired 7-aminocephem compound (I) was found accumulated in the culture. Further investigations based on this fact have now led to completion of this invention.

This invention provides a process for producing 7-aminocephem compound (I) which comprises cultivating in a nutrient medium a 7-aminocephem compound (I)-producing microorganism, for example a cephalosporin compound (II)-producing microorganism belonging to the species Acremonium chrysogenum transformed with a vector for producing 7-aminocephem compounds (I) which contains one or more promoter(s) for Acremonium chrysogenum and gene(s) for enzyme(s) capable of converting the cephalosporin compound (II) to the corresponding 7-aminocephem compound. (I), with said gene(s) connected to said promoter(s), and recovering the 7-aminocephem compound (I) from the cultured broth.

FIG. 1A schematically shows the procedure followed in Example 1.

FIG. 1B shows the restriction enzyme cleavage map of an alkaline protease cDNA and that of a genomic alkaline protease DNA.

.lambda.gt-Protease2: only the insert portion of .lambda.gt11 is indicated;

.lambda.-G-Protease-1413 and .lambda.-G-Protease-0112: only the insert portions into .lambda.gtWES.cndot..lambda.B are indicated;

pGBR3-EH1 and pGBR2-EH1: only the insert portions into pBR322 are indicated.

FIG. 1C shows the restriction enzyme cleavage map of the alkaline protease cDNA and that of the genomic alkaline protease DNA.

FIGS. 1D-1F show the nucleotide sequence of the genomic alkaline protease DNA (SEQ ID NO:1). M means G or C.

FIGS. 1G and 1H show the nucleotide sequnce of the alkaline protease cDNA (SEQ ID NO:2).

FIGS. 1I and 1J show the nucleotide sequence (SEQ ID NO:3) and deduced amino acid sequence (SEQ ID NO:4) in the fused region between the alkaline protease cDNA (.lambda.gt-Protease2) and the .lambda.gt11 .beta.-galactosidase gene.

FIG. 1K shows the nucleotide sequence (SEQ ID NO:5) and deduced amino acid sequence (SEQ ID NO:6) of the open reading frame of the alkaline protease cDNA (.lambda.gt-Protease2).

FIG. 1L shows the genomic alkaline protease gene and the site of addition of a synthetic linker (SEQ ID NOS:7-9).

FIG. 1M shows the nucleotide sequence (after modification with EcoRI and BamHI) in the vicinity of the promoter of the genomic alkaline protease DNA (SEQ ID NO:10).

FIG. 1N shows the nucleotide sequence (after modification with BamHI and HindIII) in the vicinity of the terminator of the genomic alkaline protease DNA (SEQ. ID NO:11).

FIG. 1O shows construction scheme for a plasmid, pCYG-B2.

FIG. 1P shows the genomic isopenicillin N synthetase DNA and the modification scheme for a restriction enzyme cleavage site thereof (SEQ ID NOS:12-13).

FIG. 1Q shows the nucleotide sequence in the vicinity of the promoter of the genomic isopenicillin N synthetase DNA (SEQ ID NO:14).

FIG. 1R shows the restriction enzyme cleavage map in the vicinity of the terminator of the genomic isopenicillin N synthetase DNA. For the restriction enzymes BglII, PvuII, SalI, SalII and XhoI, no cleavage site was found.

FIG. 1S shows the construction scheme for a plasmid, pCYG-EB2.

.quadrature.: Promoter and terminator of the protease gene;

: Promoter and terminator of the IPNS gene;

.quadrature.: A. chrysogenum-derived ARS;

--: pBR325 DNA and Tn903 DNA.

FIG. 1T shows the nucleotide sequence in the vicinity of the terminator of the genomic isopenicillin N synthetase DNA (SEQ ID NO:15).

FIG. 2A shows the construction scheme for a plasmid, pYG-HB51 and SEQ ID NOS:16-17.

FIGS. 2B and 2C show the nucleotide sequence (SEQ ID NO:18) and deduced amino acid sequence (SEQ ID NO:19) of a hygromycin B resistance gene.

FIG. 3A schematically shows the procedure followed in Example 3.

FIG. 3B shows the route of conversion of cephalosporin C and deacetylcephalosporin C to 7ACA and 7ADCA, respectively.

FIG. 3C shows the restriction enzyme cleavage map of a plasmid, pDAO-EB101.

FIG. 3D shows the restriction enzyme cleavage map of a plasmid, pVEB104.

FIG. 3E shows the restriction enzyme cleavage map of a plasmid, pCFS315.

FIG. 3F shows the construction scheme for a plasmid, pV22B1.

FIGS. 3G and 3H show the nucleotide sequence (SEQ ID NO:20) and deduced amino acid sequence (SEQ ID NO:21) of the open reading frame of the Pseudomonas diminuta V22-derived cephalosporin C acylase gene.

FIG. 3I shows the construction scheme for a plasmid, pV22BS-A11.

FIG. 3J shows the modification scheme for a site upstream from the acylase ATG codon of the plasmid pV22B1 (SEQ ID NOS:22-23).

FIG. 3K shows the restriction enzyme cleavage map in the vicinity of the cephalosporin C acylase gene in the plasmid pV22BS-All. In the figure, .alpha. and .beta. indicate the estimable .alpha.-subunit and .beta.-subunit, respectively.

FIGS. 3L and 3M show the nucleotide sequence (SEQ ID NO:24) and deduced amino acid sequence (SEQ ID NO:25) of the coding region of the DAO gene contained in the plasmid pCFS315.

FIG. 4A shows the construction scheme for a vector for 3ACA production, pHBV1.

FIG. 4B shows the construction scheme for a vector for 7ACA production, pHDV11.

FIG. 4C shows the construction scheme for a vector for GL-7ACA production, pHBD3.

The vector for producing 7-aminocephem compound (I) in accordance with this invention contains a DNA fragment prepared by ligating at least one or more promoter(s) for Acremonium chrysogenum (when the host cells are cells of Acremonium chrysogenum, for instance) to gene(s) for enzyme(s) capable of converting the cephalosporin compound (II) to the corresponding 7-aminocephem compound (I) to each other in that order from the upstream side to the downstream side by a conventional method. A5 appropriate selective marker, autonomously replicating sequence (ARS) for Acremonium chrysogenum, terminator, translation activating sequence and so forth may be inserted into said vector at respective desired sites by a conventional method. It is convenient for vector amplification in Escherichia coli if the autonomously replicating sequence for Escherichia coli (ori) and a selective marker are inserted in the vector.

Such vector may be constructed, for example by the methods described later herein in the examples or modifications thereof.

The "promoter for Acremonium chrysogenum" means a promoter capable of enabling expression, in Acremonium chrysogenum, of the gene coding for a desired polypeptide and includes those hitherto known, such as the promoter for the Acremonium chrysogenum isopenicillin synthetase gene and the promoter for the Acremonium chrysogenum .beta.-isopropyl malate dehydrogenase gene, as well as the DNA fragment having promoter activity for alkaline protease gene which was newly isolated from the Acremonium chrysogenum chromosome by the present inventors. These promoters may contain an enhancer sequence.

As the gene for an enzyme capable of converting cephalosporin compound (II) to 7-aminocephem compound (I) may include cephalosporin C acylase gents which are genes for enzymes catalyzing one-step conversion ›e.g. Pseudomonas sp. SE83-derived cephalosporin C acylase gene ›cf. A. Matsuda et al., J. Bacteriol., 169, 5815-5826 (1987)! as well as the cephalosporin C acylase gene newly isolated from Pseudomonas diminuta V22 by the present inventors!. As the gene(s) for enzyme(s) catalyzing two-step conversion, there may be mentioned the combination of a D-amino acid oxidase (hereinafter, "DAO") gene and a GL-7ACA-acylase gene and the combination of a cephalosporin C acylase gene and a DAO gene (cephalosporin C acylases in general being also capable of converting GL-7ACA and GL-7ADCA, and keto-AD-7ACA and keto-AD-7ADCA, to 7ACA and 7ADCA, respectively). In cases where the cephalosporin compound (II) producer used as the host produces GL-7ACA, GL-7ADCA, keto-AD-7ACA and/or keto-AD-7ADCA, a cephalosporin C acylase gene or a GL-7ACA acylase gene may be used alone.

As the D-amino acid oxidase gene, there may be mentioned, for example, the Trigonopsis variabilis-derived DAO gene (cf. Japanese Kokai Tokkyo Koho No. 62-262994) and the DAO gene newly isolated from Fusarium solani M-0718. FERM-P 2688 by the present inventors (cf. European Laid-open Patent Publication No. 364,275).

As the GL-7ACA acylase gene, there may be mentioned, for example, the gene for the Pseudomonas putida ATCC 950-derived GL-7ACA acylase ›cf. Agric. Biol. Chem., 45, 1561 (1981)! as well as the cephalosporin C acylase genes specifically mentioned above (cephalosporin C acylases serving also as GL-7ACA acylases).

The enzyme gene(s) is preferably inserted each singly into the vector at a site downstream from the promoter for Acremonium chrysogenum. The vector for producing 7-aminocephem compound (I) in accordance with the invention may contain one or more genes for an enzyme or enzymes capable of converting the cephalosporin compound (II) to the 7-aminocephem compound as inserted therein.

The selective marker my be any marker utilizable in screening transformants following transformation of Acremonium chrysogenum with the vector. The marker (Hm.sup.R) which provides the hygromycin resistance phenotype, for instance, is widely used.

The autonomously replicating sequence (ARS) for Acremonium chrysogenum is, for example, the ARS of Acremonium chrysogenum (cf. Japanese Kokai Tokkyo Koho No. 61-209593). Since, however, transformation of Acremonium chrysogenum with the vector for 7-aminocephem compound production mainly results in incorporation of said vector into the genomic DNA of Acremonium chrysogenum and the subsequent replication, it is not necessary, in such a case, that the vector for 7-aminocephem compound production should contain an ARS insert. The presence of an. ARS is required only when the vector for 7-aminocephem compound production is multiplied as an extrachromosomal component in Acremonium chrysogenum. The terminator, which may contain a polyadenylation site, is, for example, the Acremonium chrysogenum genomic DNA-derived terminator used in the examples to be mentioned later herein.

Cephalosporin compound (II)-producing strains belonging to the species Acremonium chrysogenum may be transformed with the vector for 7-aminocephem compound production by a conventional method, for example by the protoplast transformation method ›cf. S. W. Queener et al.: Microbiology 1985, American Society of Microbiology, page 468 (1985)!.

The cephalosporin compound (II) producer includes, among others, A. chrysogenum ATCC 11550, ATCC 36225, etc. as cephalosporin C producers, A. chrysogenum ATCC 20371 etc. as deacetylcephalosporin C producers, A. chrysogenum ATCC 11550, ATCC 20416, etc. as deacetoxycephalosporin C producers and A. chrysogenum ATCC 20416, ATCC 20427, etc. as GL-7ACA and keto-AD-7ACA producers. A number of microorganisms other than Acremonium chrysogenum are also known to be cephalosporin compound (II) producers. It is also possible to prepare 7-aminocephem compound producers by using such microorganisms as host cells and transform them with an appropriate vector for 7-aminocephem compound production by a method conventional in the field of genetic engineering referring to the above-mentioned method, for instance.

The thus-obtained 7-aminocephem compound (I)-producing microorganism ›cephalosporin compound (II) producer microorganism transformed with the vector for the production of 7-aminocephem compound! is cultivated in a nutrient medium. In principle, this cultivation can be carried out in the same manner as in cultivating microorganisms in general. Generally, however, submerged culture using an aqueous medium is advantageous. The medium to be used in the cultivation may be a synthetic one, a semisynthetic one or a natural one. As the carbon source to be included in the medium composition, there may be mentioned, for example, glucose, sucrose, maltose, glycerin, starch, liquefied starch and the like. As the nitrogen source there may be mentioned, for example, meat extract, caseine hydrolyzate, peptone, gluten meal, corn meal, cottonseed flour, soybean meal, corn steep liquor, dried yeast, yeast extract, urea, ammonium phosphate and the like. Inorganic salts, such as disodium hydrogen phosphate, potassium dihydrogen phosphate, magnesium chloride, magnesium sulfate and calcium carbonate, may be added to the medium, if desired.

In case the medium foams severely during cultivation, an antifoam such as a vegetable oil (e.g. soybean oil or castor oil), a higher alcohol (e.g. octadecanol, tetradecanol or heptanol), or a silicone compound, may be added in an appropriate amount.

The cultivation is suitably carried out at a temperature of about 30.degree. C. When the culture volume is large, the employment of appropriate seed culture will give good results in many instances. The period of the cultivation is suitable about 100 to 170 hours and may be prolonged when a high concentration medium is used.

The cultivation conditions mentioned above may be modified depending on the characteristics of the producing strain employed so that optimum conditions can be seleced and applied.

The 7-aminocephem compound (I) formed during cultivation is generally accumulated extracellularly in the culture in most cases. Therefore the desired 7-aminocephem compound (I) can generally be isolated and purified from the filtrate (or supernatant) separated from microbial cells by centrifugation, filtration or the like means, by applying to the filtrate (supernatant) combinedly (in an appropriate order) and/or repeatedly such means as concentration under reduced pressure, solvent extraction, pH adjustment, treatment with a resin (e.g. anion exchange resin, cation exchange resin, nonionic adsorbent resin, etc.), treatment with an adsorbent (e.g. activated carbon, silicic acid, silica gel, alumina, cellulose, etc.), high-performance liquid chromatography, crystallization, recrystallization, and the like means used in producing antibiotic substances in general.

The 7-aminocephem compound (I) obtained in the free form can be converted to desired salts by reacting it with a base such as sodium hydroxide and the like.

The following examples are further illustrative of this-invention. In the examples, the following abbreviations are used.

DNA: Deoxyribonucleic acid

c-DNA: Complementary DNA

RF DNA: Replicative form DNA

RNA: Ribonucleic acid

m-RNA: Messenger RNA

dNTP: Mixture of dATP (deoxyadenosine triphosphate),

dCTP (deoxycytidine triphosphate),

dGTP (deoxyguanosine triphosphate) and

dTTP (deoxythymidine triphosphate)

bp: Base pairs

Kbp: Kilo base pairs

Ap.sup.R : Ampicillin resistance in E. coli

Ap.sup.S : Ampicillin sensitivity in E. coli

Cm.sup.R : Chloramphenicol resistance in E. coli

Cm.sup.S : Chloramphenicol sensitivity in E. coli

Km.sup.R : Kanamycin resistance in E. coli

Km.sup.S : Kanamycin sensitivity in E. coli

Tc.sup.R : Tetracycline resistance in E. coli

Tc.sup.S : Tetracycline sensitivity in E. coli

lac PO: Lactose operon promoter and operator in E. coli

tac PO: Trp-lac promoter and operator in E. coli

Acy.sup.+ : Acylase activity

Hm.sup.R : Hygromycin B resistance

G418.sup.R : Resistance to the antibiotic G418

DAO: D-Amino acid oxidase

IPNS: Isopenicillin N synthetase

CC: Cephalosporin C

CCNa: Cephalosporin C sodium

DCC: Deacetylcephalosporin C

7ACA: 7-Amino-3-acetoxymethyl-3-cephem-4-carboxylic acid

7-Amino-3-hydroxymethyl-3-cephem-4-carboxylic acid

Keto-AD-7ACA: Ketoadipyl-7ACA ›7-(5-carboxy-5-oxopentanamido)-3-acetoxlmethyl-3-cephem-4-carboxylic acid !

Keto-AD-7ADCA: Ketoadipyl-7ADCA ›7-(5-carboxy-5-oxopentanamido)-3-hydroxymethyl-3-cephem-4-carboxylic acid !

GL-7ACA: Glutaryl-7ACA ›7-(4-carboxybutanamido)-3-acetoxymethyl-3-cephem-4-carboxylic acid!

GL-7ADCA: Glutaryl-7ADCA ›7-(4-carboxybutanamido)-3-hydroxymethyl-3-cephem-4-carboxylic acid!

DTT: Dithiothreitol

Tris: Tris(hydroxmethyl)aminomethane

EDTA: Ethylenediaminetetraacetic acid

SDS: Sodium lauryl sulfate

PEG: Polyethylene glycol

IPTG: Isopropyl-.beta.-D-thiogalactopyranoside

X-gal: 5-Bromo-4-chloro-3-indolyl .beta.-galactoside

Met: Methionine

Thr: Threonine

Ala: Alanine

Gln: Glutamine

Gly: Glycine

Val: Valine

Pro: Proline

Ile: Isoleucine

Lys: Lysine

Ash: Asparagine

Glu: Glutamic acid

Phe: Phenylalanine

Leu: Leucine

Asp: Aspartic acid

Tyr: Tyrosine

Cys: Cysteine

Trp: Tryptophan

Ser: Serine

Arg: Arginine

His: Histidine

The compositions of the main buffers and media used in the examples are as follows:

TE buffer: 20 mM Tris-HCl (pH 7.5),

0.5 mM EDTA

LB agar: 10 g/liter Bacto-tryptone (DIFCO),

5 g/liter yeast extract (DIFCO),

5 g/liter NaCl,

15 g/liter agar (DIPCO),

pH 7.2

B3 agar: 10 g/liter D-Galactose,

2 g/liter Bacto-tryptone (DIFCO),

1 g/liter MgSO.sub.4 .multidot.7H.sub.2 O,

1.5 g/liter KH.sub.2 PO.sub.4,

1 g/liter NaNO.sub.3,

20 g/liter agar,

(pH about 5.6 to 6.0)

YPS medium: 20 g/liter Sucrose,

10 g/liter polypeptone (Daigo

Nutritive Chemicals),

5 G/liter powdered yeast extract

(Daigo Nutritive Chemicals),

1 G/liter K.sub.2 HPO.sub.4,

1 g/liter MgSO.sub.4 .multidot.7H.sub.2 O,

(pH 7.0)

CSl medium: 25 g/liter Sucrose,

10 g/liter glucose,

30 g/liter soybean meal,

10 g/liter corn steep liquor (CSL),

adjusted to pH 7.0 with

5 g/liter CaCO.sub.3

Main culture medium:

16 g/liter Peanut powder,

40 g/liter CSL,

8 g/liter wheat germ,

16 g/liter gluten meal,

8 g/liter (NH.sub.4).sub.2 SO.sub.4,

20 g/liter glucose,

20 g/liter sucrose,

adjusted to pH 7.0 with NaOH,

10 g/liter CaCO.sub.3,

35 ml/liter methyl oleate

1M KP buffer (pH 7.5):

1M KCl,

25 mM

10 mM MgCl.sub.2,

10 mM Tris-HCl (pH 7.5)

1M KP buffer (pH 5.8):

Buffer prepared by adjusting 1M KP

buffer (pH7.5) to pH 5.8 with HCl

0.8M NaP buffer:

0.8M NaCl,

25 mM CaCl.sub.2 .multidot.2H.sub.2 O,

10 mM MgCl.sub.2,

10 mM Tris-HCl (pH 7.5)

PDA-YE agar: 39 g/liter Potato dextrose agar medium (Nissui Pharmaceutical),

5 g/liter agar,

5 g/liter yeast extract

BRM agar medium:

›A! 2 g NaNO.sub.3,

1 g KH.sub.2 PO.sub.4,

274 g sucrose (final concentration

0.8M),

10 g brain-heart infusion (DIFCO),

7.5 g agar (Hayashi Junyaku),

(pH about 6.3 to 6.4)

›B! 20 g glucose (final concentration 2%),

1.5 g CaCl.sub.2 .multidot.2H.sub.2 O (final concentration 10 mM)

An aqueous solution (900 ml) of components ›A! and an aqueous solution (100 ml) of components ›B! are respectively sterilized by autoclaving and then mixed up, and the mixture is used for preparing plates and so on.

Unless otherwise specified, the gene manipulation techniques used in the examples are those described in T. Maniatis et al.: Molecular Cloning--A Laboratory Manual, published 1982 by Cold Spring Harbor Laboratory.

The whole procedure followed in this example is outlined in FIG. 1A.

(1) Cloning of an alkaline protease cDNA and chromosomal DNA

(1-i) Purification of alkaline protease derived from A. chrysogenum ATCC 11550 and antibody preparation:

Following the method of Yagi et al. ›J. Yagi et al., J. Ferment. Technol., 50, 5.92 (1972)!, A. chrysogenum ATCC 11550 was cultivated and alkaline protease was recovered from the culture filtrate by precipitation with 70% saturated ammonium sulfate, followed by purification using CM (carboxymethyl)-cellulose and Sephadex G75. A molecular size of about 30,000 dalton was revealed by analysis by 11% SDS-PAGE (SDS polyacrylamide gel electrophoresis). This alkaline protease was mixed with complete Freund adjuvant (DIFCO) and the mixture was injected into male New Zealand white rabbits three times at a dose of 2 mg per injection in order to cause antibody production. Then whole blood collection and serum separation were made, and the serum was treated at 56.degree. C. for 30 minutes and then with 35% saturated ammonium sulfate. The resultant precipitate was subjected to treatment with Protein A-Sepharose CL-4B (Pharmacia) for antibody purification. The whole volume of the antibody solution was made 16 ml. Ouchterlony's test confirmed that the protease protein and the antibody reacted with each other to give an immunoprecipitate. The IgG (immunoglobulin G) fraction obtained from a control rabbit (no protease injection) did not show any immunoprecipitation reaction with the protease.

(1-ii) Formation of a genomic DNA library from A. chrysogenum ATCC 11550:

The genomic DNA was extracted from A. chrysogenum ATCC 11550 by the method described in Japanese Kokai Tokkyo Koho No. 61-209593 (Isogai et al.; laid open Sep. 17, 1986). This DNA (about 25 .mu.g) was treated with the restriction enzyme EcoRI (100 units.) at 37.degree. C. for 3 hours for partial cleavage and the digest was subjected to sucrose density gradient centrifugation (5 to 20% sucrose; Hitachi ultracentrifuge rotor RPS 55T-2, 50 krpm, 4 hours) to give a purified DNA fraction (about 3 Kbp or more in size). The size-fractionated DNA was finally dissolved in 100 .mu.l of TE buffer. Separately, the .lambda.gtWES.cndot..lambda.B DNA ›Bethesda Research Laboratories (BRL)! (about 40 .mu.g) was completely cleaved with EcoEI and the digest was subjected to sucrose density gradient centrifugation (under the same conditions as mentioned above), whereby a 4.85-Kbp DNA was removed and a 21.7-Kbp left arm and a 13.8-Kbp right arm were purified. The DNA fragment obtained by partial cleavage with EcoEI (>3 Kbp, about 5 .mu.g) and the .lambda.gtWES.cndot..lambda.B-derived EcoRI arms (about 15 .mu.g) were mixed up and subjected to ligation using T4 DNA ligase. The ligation mixture was subjected to in vitro packaging using PackerGene ›.lambda. phage in vitro packaging system; Promega Biotec (imported and distributed by Seikagaku Kogyo)!. Plaque formation using E. coli DP50 supF (included in the PackerGene kit) as a host gave about 2.times.10.sup.6 plaques. Thus a genomic DNA library consisting of about 2.times.10.sup.6 clones could be constructed.

(1-iii) Extraction and purification of mRNA from A. chrysogenum ATCC 11550:

A. chrysogenum ATCC 11550 was cultured in 100 .mu.l of an aqueous medium containing 4.5% soluble starch, 3% corn steep liquor (CSL), 1.5% soybean meal and 0.35% CaCO.sub.3 at 25.degree. C. for 4 days. The cells obtained were distrupted in a mortar cooled to a low temperature with liquefied nitrogen, then suspended in 40 ml of a guanidine isothiocyanate solution ›4M guanidine isothiocyanate, 50 mM Tris-HCl (pH 7.5)., 20 mM EDTA, 2% N-lauroylsarcosine sodium, 0.17M 2-mercaptoethanol! and heated at 60.degree. C. for 5 minutes. The suspension was then centrifuged at 10,800 x g for 10 minutes and 13 mg of whole RNA was recovered from the supernatant by the guanidine-cesium chloride method (cf. Molecular Cloning, page 196, Cold Spring Harbor Laboratory, 1982).

Two purifications of the whole RNA (13 mg) using 1 g of oligo(dT)-cellulose (BRL) ›cf. Molecular Cloning (1982), page 197! gave 460 .mu.g of poly (A)-RNA (mRNA).

(1-iv) Construction of a cDNA library by the Okayama-Berg method:

For obtaining a nearly full-length cDNA, a cDNA library was constructed by the Okayama-Berg method ›cf. H. Okayama and P. Berg, Mol. Cell. Biol., 2, 161 1982)!. The mRNA (4 .mu.g) obtained in (1-iii) and 0.9 .mu.g of a primer ›3'-oligo(dT)-tailed pSV7186-derived plasmid primer (Pharmacia)! were treated with reverse transcriptase (Seikagaku Kogyo) for synthesizing 0.64 .mu.g of ss-cDNA (single-stranded cDNA). This was treated with terminal transferase for C-tailing (18 C's on the average) and then cleaved with the restriction enzyme HindIII, followed by annealing with 0.25 .mu.g of a linker ›3'-oligo(dG)-tailed pSV1932-derived HindIII linker (Pharmacia)! and treatment with E. coli DNA ligase. Then ds-cDNA (double-stranded cDNA) was synthesized by treatment of the ligase-treated mixture with RNase H (BRL), DNA polymerase I (Pharmacia) and E. coli DNA ligase (Pharmacia). This cDNA was used to transform E. coli DH1 (ATCC 33849) by the method of D. Handban ›D. Handhan, J. Mol. Biol., 166, 557 (1983)! to give 3.6.times.10.sup.4 ampicillin-resistant clones. Thus a cDNA library comprising 3.6.times.10.sup.4 clones could be constructed.

(1-v) Cloning of alkaline protease cDNA of A. chrysogenum ATCC 11550:

Plasmid DNA was isolated from the cDNA library. consisting of 3.6.times.10.sup.4 clones and recloned into .lambda.gt11 (ATCC 37194), and the protease gene was searched for by express-blot assay using the protease antibody prepared in (1-i).

The cDNA-derived DNA mixture (20 .mu.g) was cleaved with PstI, then treated with 1.2 units of Bal31 (BRL) (final volume 600 .mu.l) at 37.degree. C. for 5 minutes for rendering the ends blunt and subjected to phenol extraction. The DNA recovered by ethanol precipitation was treated with the Klenow fragment (large fragment E. coli DNA polymerase I) in the presence of dNTP, followed by phenol extraction and ethanol precipitation. The DNA recovered was treated with. EcoRI methylase (New England Biolabs) in the presence of S-adenosylmethionine for methylation of the EcoRI site to render said site unclearable with EcoRI. About 10 .mu.g of this DNA and 5 .mu.g of a pEcoRI linker ›d(pG-G-A-A-T-T-C-C); Takara Shuzo! were mixed up and ligated to each other using T4 DNA ligase. After inactivation of the enzyme by treating at 65.degree. C. for 15 minutes, the ligation mixture was cleaved with EcoRI. The pEcoRI linker was removed by 5% acrylamide gel electrophoresis. DNA was eluted from the acrylamide gel with an eluting buffer (0.5M CH.sub.3 COONH.sub.4, 10 mM magnesium acetate, 1 mM EDTA, 1% SDS) and purified on DE52 (Pharmacia) ›in 10 mM Tris-HCl (pH 7.5), 5 mM EDTA; adsorption with 0.1M MACl; desorption with 1M NaCl; a spun column used!.

Protoclone GT11 DNA (Promega Biotec, distributed by Seikagaku Kogyo; 2 .mu.g) (prepared from purified .lambda.gt11 DNA by treatment with DNA ligase, cleavage with EcoRI and treatment with alkaline phosphatase) and the cDNA mixture treated as described above were mixed up and ligated to each other using T4 DNA ligase. The resultant DNA mixture was subjected to in vitro packaging using PackerGene (.lambda. phage in vitro packaging system, Promega Biotec), followed by plaque formation with E. coli Y1090 (r.sup.- m.sup.+) (ATCC 37197) as a host. About 3.times.10.sup.5 plaques were obtained. Upon observation of the plaques in the presence of X-gal and IPTG, about 75% were white plaques and the remaining (about 25%) were blue plaques.

A mixture of a culture (0.1 ml) of E. coli Y1090 (r.sup.- m.sup.+), the above-mentioned .lambda.gt11-derived cDNA library (about 1.times.10.sup.4 clones) and a top layer agar (the same as LB agar except for the agar concentration which was 0.8%) (3 ml) was spread over 20 ml of LB agar containing 50 .mu.g/ml ampicillin as placed in a dish (about 8.5 cm in diameter) and the dish contests were allowed to stand at room temperature for 1 hour for agar solidification and then incubated at 42.degree. C. for 4 hours. A nitrocellulose filter (82.5 mm in diameter; Bio-Rad Laboratories) was immersed in 10 mM IPTG solution for 1 hour and then dried at room temperature for about 1 hour. This filter was placed on the above-mentioned dish culture and incubation was further conducted at 37.degree. C. for 2.5 hours for allowing gene expression and blotting onto the filter. From among about 7.times.10.sup.4 plaques resulting from such blotting, six positive clones were isolated by using an express-blot assay kit (Bio-Rad) and the antibody produced in (1-i). The express-blot assay (cf. T. V. Huynh et al., DNA Cloning, volume I, pages 73-75,IRL Press, 1985) was performed as described in the manual attached to the kit. Following fused protein expression and blocking of the plaque-lifted filter with gelatin, the filter was reacted with a rabbit-derived protease antibody neutralized with E. coli lysate (included in the kit) ›neutralized by adding 0.2 ml of E. coli lysate and 4 .mu.l of protease antibody to 20 ml of antibody incubation buffer (cf. the manual attached to the kit)! and then with HRP (horseraddish peroxidase)-labeled anti-rabbit IgG (included in the kit). Upon the subsequent color development, positive clones gave blue-violet spots, while negative clones and rabbit-derived control antibodies gave no such spots.

DNA was isolated from each of the above six positive clones and cleaved with restriction enzymes for comparison of cDNA inserts with respect to length. The clone having the longest insert was named. .lambda.gt-Protease2 (cf. FIG. 1B). The remaining five clones each gave a similar restriction enzyme cleavage map.

(1-vi) Cloning of the protease cDNA clone .lambda.gt-Protease2 into pBR325:

The .lambda.gt-Protease2 DNA was cleaved with EcoRI and two DNAs (about 0.5 kbp and about 2.6 Kbp) were isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution (cf. Molecular Cloning, pages 150 and 164, Cold Spring. Harbor Laboratory, 1982). Separately, the pBR325 (BRL) DNA was cleaved with EcoRI and mixed with one of the above-mentioned two DNA fargments about 0.5 Kbp and about 2.6 Kbp) and ligation was carried out using T4 DNA ligase. Each ligation mixture was used to transform E. coli DH1. From among colonies grown on an LB agar plate containing 50 .mu.g/ml ampicillin, transformants incapable of growing on an LB agar plate containing 50 .mu.g/ml chloramphenicol were recovered. Plasmid DNA was isolated from each of these transformant strains and subjected to restriction enzyme cleavage for insert confirmation. One of the plasmids containing the 0.5 Kbp DNA was named pBR-Pro2-E2 (containing a protease cDNA fragment corresponding to the neighborhood of the N terminus) and one of the plasmids with %he 5.6 Kbp DNA insert was named pBR-Pro2-E3 (containing a protease cDNA fragment corresponding to the C terminus side).

(1-vii) Cloning of genomic protease DNA clones .lambda.-G-Protease-1413 and .lambda.-G-Protease-0112:

From about 5.times.10.sup.4 clones of the A. chrysogenum ATCC 11550-derived genomic DNA library constructed in (1-ii), two genomic protease DNA clones, .lambda.-G-Protease-1413 and .lambda.-G-Protease-0112 (cf. FIG. 1-2), were cloned using pBR-Pro2-E2 and pBR-Pro2-E3 as probes.

A mixture of a culture of E. coli DP50supF (Promega Biotec) (0.1 ml), the genomic DNA library prepared in (1-ii) (about 1.times.10.sup.4 clones) and a top agar (the same as LB agar except for the agar concentration which was 0.8%) (3 ml) was spread over 20 ml of LB agar containing 10 mM MgCl.sub.2 as placed in a dish (about 8.5 cm in diameter) and the dish was incubated overnight at 37.degree. C. Phage transfer to a nitrocellulose filter (85 mm in diameter; BioRad) was performed by the method described in Advanced Bacterial Genetics, pages 162 and 174 (Cold Spring Harbor Laboratory, 1980), followed by denaturation and neutralization and by Southern hybridization using a .sup.32 P-labeled DNA mixture of pBR322-Pro2-E2 and pBR-Pro2-E3.for recovering positive clones. DNA labeling with .sup.32 P was performed by the nick translation method (Molecular Cloning, page 109, Cold Spring Harbor Laboratory, 1982) using .alpha.-.sup.32 P-dCTP.

From about 5.times.10.sup.4 plaques, nine positive clones were isolated. From these were obtained 8 clones by monoplaque isolation in the same manner as mentioned above. DNA was recovered from each of these clones, cleaved with EcoRI and subjected to agarose gel (0.8%) electrohoresis, followed by Southern hybridization by the method of Southern ›E. M. Southern, J. Mol. Biol., 98, 116 (1975)! using as a DNA probe pBR-Pro2-E2 or pBR-Pro2-E3 labeled with .sup.32 p by the method mentioned above. One of the four clones found positive with the pBR-Pro2-E3 probe was named .lambda.-G-Protease-0112 (cf. FIG. 1B), while one of the four clones found positive with the pBR-Pro2-E3 was named .lambda.-G-Protease-1413 (cf. FIG. 1B).

(1-viii) Construction of genomic protease DNAs, pGR2-EH1 and pGR3-EH1:

pGPR2-EH1 was constructed by subcloning the .lambda.-G-Protease-0112 DNA into pBR322 (BRL) (cf. FIG. 1B), and pGRB-EH1 by subcloning the .lambda.-G-Protease-1413 DNA into pBR322.

Thus, the .lambda.-G-Protease-0112 DNA was cleaved with EcoRI and HindIII and a DNA fragment of about 3.5 Kbp was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. Separately, the pBR322 DNA was cleaved with EcoRI and HindIII and the cleavage mixture was mixed with the above-mentioned DNA fragment (about 3.5 Kbp). Ligation was effected using T4 DNA ligase. The ligation mixture was used to transform E. coli DH1. From the colonies grown on an LB agar plate containing 50 .mu.g/ml of ampicillin, those transformants that could not grow on an LB agar plate containing 10 .mu.g/ml of tetracycline were recovered. The plasmid DNA was isolated from one of these transformant strains and named. pGPR2-EH1 (cf. FIG. 1B). Similarly, the .lambda.-G-Protease-1413 DNA was cleaved with EcoRI and HindIII, a DNA fragment of about 3.7 Kbp was isolated by agarose gel electrophoresis followed by electrophoretic elution and cloned into the EcoRI-HindIII fragment of pBR322. In this way, pGPR3-EH1 (cf. FIG. 1B) was constructed.

Southern hybridization carried out in the same manner as performed in (1-vii) for .lambda.-G-Protease-0112 or .lambda.-G-Protease-1413 revealed that the above plasmids pGPR2-EH1 and pGPR3-EH1 were the desired ones. (In this case, the DNAs were subjected to double digestion with EcoRI and HindIII and .sup.32 p probes corresponding to the respective inserts were used.)

(1-ix) Determination of the nucleotide sequences of the protease gene cDNA and genomic DNA derived from A. chrysogenum ATCC 11550:

The nucleotide sequences of the cDNA portion (about 1.5 Kbp) of .lambda.gt-Protease2, the XhoI-EcoRI fragment (about 1.4 Kbp) of pGPR2-EH1 and the EcoRI-PstI fragment (about 1.5 Kbp) of pGPR3-EH1 were determined by the dideoxynucleotide synthetic chain termination method ›F. Sanger et al., Proc. Natl. Acad. Sci. USA, 74, 5463 (1977)! using the M13mp10 and M13mp11 vectors (Amersham) together with .alpha.-.sup.32 P-dCTP. The results obtained are shown in FIGS. 1D-1F and FIGS. 1G and 1H. The long open reading frame of .lambda.gt-Protease2 is shown in FIG. 1K. Since this cDNA was obtained based on expression of a protein fused with .lambda.gt11 .beta.-galactosidase followed by screening with the antibody mentioned above, the cDNA and .beta.-galactosidase gene were supposed to be in the same translational reading frame and the result of translation was estimated. The result is shown in FIGS. 1I and 1J. It is evident that the two frames are just in phase. Furthermore it was found that the cDNA was joined to .lambda.gt11 via a 21-bp GC tailing portion. Considering the fact that the Okayama-Berg method was used for cDNA preparation and that the GC tailing formed in said cDNA preparation was found attaching, it was estimated that the cDNA in .lambda.gt-Protease2 was almost full-length. The result of analysis of the nucleotide sequence of this cDNA is shown in FIG. 1C A. Comparison in nucleotide sequence between the cDNA and genomic DNA indicated that portion of the genomic DNA which is transcribed into mDNA (cf. FIG. 1C B). The estimated locations of the promoter and terminator upstream and downstream respectively, from said portion are shown in FIG. 1C B.

Comparison in amino acid sequence between the preproprotein to A. chrysogenum-derived alkaline protease and the known preproprotein to Aspergillus oryzae-derived alkaline protease ›Molecular and General Genetics, 219, 33-38 (1989)! revealed that 230 out of the 402 amino acid residues in the former are identical to the corresponding ones in the latter with 57.2% homology and that the former is a serine protease. Based on this data and other -results obtained in a separate study, the primary structure of A. chrysogenum-derived mature alkaline protease was estimated to be represented by the amino acid sequence comprising the residues Nos. 1 to 285 as shown in FIG. 1K.

(2) Construction of the plasmid pCYG-B2:

(2-i) Introduction of EcoRI and BamHI synthetic DNA linkers into the A. chrysogenum ATCC 115.50-derived protease gene genomic DNA:

The pGPR2-EH1 (cf. FIG. 1B) DNA was cleaved with SmaI, a synthetic EcoRI liner ›d(CCGAATTCGG); Takara Shuzo! was added, and ligation was effected using T4 DNA ligase, followed by cleavage with EcoRI and BglII. An EcoRI-BglII fragment (about 1.02Kbp) was isolated by agorose gel (0.8%) electrophoresis followed by electrophoretic elution. This DNA fragment (about 1.02 Kbp) was mixed with an EcoRI-BamHI DNA fragment (4.0 Kbp) of the vector plasmid pBR322 and ligation was carried out using T4 DNA ligase to give the plasmid pGPR-PA3.

The pGPR-PA3 DNA was cleaved with AluI, a synthetic pBamHI linker ›d(pCGGATCCG); Takara Shuzo! was added, and ligation was conducted in the presence of T4 DNA ligase. The ligation product was then cleaved with EcoRI and BamHI and an EcoRI-BamHI fragment (about 0.59 Kbp) was isolated by polyacrylamide gel (8%) electrophoresis followed by electrophoretic elution. This DNA fragment (about 0.59 Kbp) was mixed with an EcoRI-BamHI fragment (4.0 Kbp) of the vector plasmid pBR322 and ligation was performed using T4 DNA ligase to give the plasmid pGPR-PA3Q2.

The sites of introduction of the EcoRI and BamHI linkers into the promoter region of the protease gene DNA are shown in FIG. 1L. The nucleotide sequence of the 0.59 Kbp DNA is shown in FIG. 1M.

(2-ii) Introduction of a BamHI synthetic DNA linker into the A. chrysogenum ATCC 11550-derived protease genomic DNA:

The pGPR3-EH1 (cf. FIG. 1B) DNA was cleaved with Bali, a synthetic BamHI linker ›d(CCGGATCCGG); Takara Shuzo! was added, and ligation was conducted using DNA ligase, followed by cleavage with BamHI, PstI and XhoI. A BamHI-PstI DNA fragment (about 0.92 Kbp) was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. This DNA fragment (about 0.92 Kbp) was mixed with a BamHI-PstI DNA fragment (2.7 Kbp) of the vector plasmid pUC18 and ligation was effected using T4 DNA ligase to give the plasmid pGPR-TB1.

The site of introduction of the BamHI linker into the terminator region of the protease gene DNA is shown in FIG. 1L. The nucleotide sequence comprising the 0.92-Kbp DNA is shown in FIG. 1N.

(2-iii) Cloning of a promoter and a terminator into pCEP97 (construction of pCYG-B2): (cf. FIG. 1O)

The vector plasmid pCEP97 ›cf. Isogai et al., Japanese Kokai Tokkyo Koho No. 61-209593; isolable from Escherichia coli C600 r.sup.- m.sup.- (pCEP97) ATCC 39971 by a conventional method! (Ap.sup.R Cm.sup.R) (cf. FIG. 1-11) was cleaved with EcoRI and HindIII. Separately, pGPR-PA3Q2 ›cf. (2-i); Containing a procter (about 0.59 Kbp)! (Ap.sup.R Cm.sup.S) was cleaved with EcoRI and BamHI and, further, pGPR-TB1 ›cf. (2-ii); containing a terminator (about 0.92 Kbp)!(Ap.sup.R Cm.sup.S) was cleaved with BamHI and HindIII. An EcoRI-HindIII DNA fragment (about 6.1 Kbp) of pCEP97 and a BamHI-HindIII DNA fragment (about 0.93 Kbp) of pGPR-TB1 were respectively isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution.

An EcoRI-BamHI DNA fragment (about 0.59 Kbp) of pGPR-PA3Q2 was isolated by acrylamide get (8%) electrophoresis followed by electrophoretic elution. These three DNA fragments were mixed up and ligated to one another using T4 DNA ligase, and the ligation mixture was used to transform E. coli DE1. Transformant strains capable of growing on an. LB agar plate containing 50 .mu.g/ml of ampicillin but incapable of growing on an LB agar plate containing 50 .mu.g/ml of chloramphenicol were recovered. The plasmid DNA was isolated from one of these strains and named pCYG-B2 (cf. FIG. 1O). That said DNA was the desired one which was confirmed using restriction enzymes.

(3) Cloning of the genomic IPNS gene DNA

(3-i) Extraction and purification of the genomic DNA from A. chrysogenum 3112:

A. chrysogenum 3112 was grown on a medium (100 ml) containing 1% glucose, 3% soluble starch, 3% corn steep liquor (CSL), 1.5% soybean meal and 0.5% CoCO.sub.3 (RE 6.5) at 30.degree. C. for 5 days. The cells were disrupted in a mortar under cooling with liquefied nitrogen and a buffer was added to Give the following composition: 50 mM Tris-HCl (pH 7.5), 10 mM EDTA and 1% SDS. The resultant mixture was-heated at 65.degree. C. for 10 minutes. After phenol extractions(twice), ethanol precipitation was performed. The precipitate was treated with 5 .mu.g/ml RNase A (Sigma.) and then with 100 .mu.g/ml protease K (Merck). Phenol extractions, ethanol precipitation and sucrose density gradient centrifugation (5 to 20% sucrose; Hitachi ultracentrifuge rotor SRP28; 22 krpm, 13 hours) gave a purified genomic DNA.

(3-ii) Size identification of the genomic isopenicillin N synthetase (IPNS) DNA:

Based on the DNA nucleotide sequence of the A. chrysogenum-derived IPNS gene cloned by S. M. Samson et al. ›S. M. Samson et al., Nature, 318, 191 (1985)!, the following three DNA probes (two 15-meric DNAs and one 12-meric DNA) were synthesized by the method described in Eiko Otsuka: Bunshi Idengaku Jikkenho (Experiments in Molecular Genetics), pages 298-307, Kyoritsu Shuppan, 1983.

Probe ›1!: 5'-CTATTCGGCGATGAC-3' (SEQ ID NO:26)

Probe ›2!: 5'-AAGGAGAAGAAGCTC-3' (SEQ ID NO:27)

Probe ›5!: 5'-CTCCTTGTCATC-3' (SEQ ID NO:28)

Probe ›1! and probe ›2! were respectively labeled by the method of Inglia et al. ›Inglia et al., Nucleic Acids-Res., 9, 1627 (1982)! using T4 polynucleotide. kinase (BRL) and .gamma..sup.32 P-ATP. Then a mixed DNA solution containing the probes ›1!, ›2! and ›3! was prepared, heated at 95.degree. C. for 2 minutes and cooled gradually to room temperature for effecting annealing. The annealing product was purfied using NENSORB 20 (Du Pont; imported and distributed by Daiichi Kagaku.) (according to the attached manual). Separately, the A. chrysogenum 3112-derived genomic DNA (about 5 prepared in (3-i) was cleaved with BamHI and the cleavage mixture was subjected to agarose gel (0.8%) electrophoresis, followed by transfer to a nitrocellulose filter and Southern hybridization by the method of Southern ›E. M. Southern, J. Mol. Biol., 98, 503 (1975)!. The hybridization was performed overnight at 42.degree. C. using 6 x SSC (0.9M NaCl, 0.09M sodium citrate, pH 7.0), 5 x BFP ›1 x BFP: 0.02% bovine serum albumin, 0.02% Ficoll (MW: 408,000), 0.02% Polyvinylpyrrolidone!, 0.5% SDS, 100 .mu.g/ml carrier DNA (calf thymus DNA) and the labeled DNA mentioned above. The filter was then washed once with 6 x SSC at 55.degree. C. and then twice with 2 x SSC (2 x SSC being a three-fold dilution of 6 x SSC). As a result a DNA fragment of about 3.1 Kbp was found hybridization-positive.

(3-iii) Cloning of the IPNS gene:

About 20 .mu.g of the A. chrysogenum 3112-derived genomic DNA prepared in (3-i) was cleaved with BamHI and a DNA mixture (about 2.5 to 4.4 Kbp) was separated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. Separately, the vector plasmid pBR322 was cleaved with BamHI, the digest was mixed with the above-mentioned DNA mixture (about 2.5 to 4.4 Kbp), then ligation was conducted using T4 DNA ligase, and the ligation mixture was used to transform E. coli DH1. About 2.4.times.10.sup.5 colonies appeared on an LB agar plate containing 50 .mu.g/ml ampicillin. About 5% of the colonies could not grow on an LB agar plate containing 10 .mu.g/ml tetracycline. Thus, about 1.2.times.10.sup.4 colonies contained the A. chrysogenum 3112-derived DNA. The plasmid DNA was isolated from each of about 2.4.times.10.sup.5 transformants and cleaved with EcoRI. For exclusion of the original vector plasmid free of the A. chrysogenum 3112-derived DNA, the EcoRI digest was subjected to agarose gel (0.8%) electrophoresis and a mixture of DNAs about 6.5 to 9.5 Kbp in size was separated by electrophoretic elution. This DNA mixture was subjected to ligation using T4 DNA ligase and the ligation mixture was used to transform E. coli DH1. In this way, about 2.6.times.10.sup.4 transformants resistant to 50 .mu.g/ml ampicillin were obtained.

The above-mentioned about 2.6.times.10.sup.4 genomic DNA-containing E. coli transformants were grown on an LB agar plate containing 50 .mu.g/ml ampicillin at 37.degree. C. for 9 hours to give colonies, which were transferred to a nitrocellulose filter and further incubated on an LB agar plate containing 250 .mu.g/ml chloramphenicol at 37.degree. C. for 12 hours. The filter was then treated with 0.5N NaOH-1.5M. NaCl at room temperature for 4 minutes for bacteriolysis and DNA denaturation, then neutralized by treatment with two portions of 0.5M Tris=HCl (pH 7.0)-1.5M MaCl at room temperature for 5 minutes per portion and subjected to hybridization, which was performed by the method described in (3-ii) using the .sup.32 P-labeled 30-meric DNA resulting from joining of probes ›1! and ›2!. Five hybridization-positive strains were obtained. The plasmid DNAs isolated from these strains were found quite identical to one another when compared by restriction enzyme cleavage analysis. One of the plasmids was named pIPS105 (cf. FIG. 1P).

(3-iv) Cloning of pIPS105 into M13mp10:

Amersham's M13 cloning kit was used and the procedure described in the manual attached to the kit was followed. Thus, the pIPS105 DNA was cleaved with SalI and a DNA fragment of about 0.63 Kbp was isolated by agarose gel (1.5%) electrophoresis followed by electrophoretic elution. This. DNA fragment was mixed with the SalI digest of the M13mp10 DNA and ligation was carried out using T4 DNA ligase. The ligation mixture was used to transform E. coli JM105 (Amersham) as a host. White plaques which formed in the presence of X-gal and IPTG were collected and phage multiplication was conducted using E. coli JM105 as a host. The RF (replicating form) DNA was isolated from the host cells and named M13mp10-IPS4-2. Similary, the pIPS105 DNA was cleaved with PstI and BamHI and a DNA fragment of about 1.0 Kbp was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. This DNA fragment (about 1.0 Kbp) was ligated to PstI-BamHI DNA fragments of M13mp10 using T4 DNA ligase. Cloning was carried out in the same manner as above. The phage obtained was multiplicated using E. coli JM 105 as a host and the RF DNA was isolated and named M13mp10-IPS3.

The DNA nucleotide sequence of a SalI region of about 0.63 Kbp of the ss-DNA (single-stranded DNA) of M13mp10-IPS4-2 was determined by the dideoxynucleotide synthetic chain termination method (using Amersham's sequencing kit). The 478 bp SalI-NcoI DNA fragment out of said region is illustrated in FIG. 1Q. In this way, the presence of the IPNS gene could be confirmed and the promoter region could be identified. The restriction enzyme map of the IPNS terminator of M13mp10-IPS3 and the multicloning site portion of M13mp10 is shown in FIG. 1R. The IPNS gene promoter and terminator can be readily synthesized by means of a DNA synthesizer on the basis of the DNA nucleotide sequences shown in FIG. 1Q and FIG. 1T, respectively.

(4) Construction of pCYG-EB2 (cf. FIG. 1-15)

(4-i) Deletion of the BamHI site from M13mp10-IPS3 (cf. FIG. 1P):

The RF DNA of M13mp10-IPS3 was cleaved with BamHI, then repaired using the Klenow fragment (large fragment E. coli DNA polymerase I; Amersham) in the presence of dNTP, subjected to ligation using T4 DNA ligase and again cleaved with BamHI. The cleavage mixture was used to transform E. coli JM105 and phage plaques were collected. The phage was multiplied using E. coli JM105 as a host and the RF DNA was separated from the host cells. This DNA was named M13mp10-IPS3-DS and it was confirmed that said DNA could not be cleaved with BamHI. The DNA nucleotide sequence of the terminator-containing BclI-EcoRI region. (about 1.0 Kbp) of M13mp10-IPS3-D8 was determined by the dideoxynucleotide synthetic chain termination method using .alpha.-.sup.32 P-dATP and sequenase. The sequence is shown in FIG. 1T.

(4-ii) Joining of the IPNS promoter to the G418 resistance gene:

The G418 resistance gene of Tn903 ›A. Oka et al., J. Mol. Biol., 147, 217 (1987)! (PvuII 1696 bp fragment) was purified from the PvuII-cleaved DNA of pCYG97 ›cf. Japanese Kokai Tokkyo Koho No. 61-209593; isolable from Escherichia coli C600 r-m- (pCYG97) ATCC 399770 by a conventional method!. To this 1696 bp DNA fragment was added a synthetic BamHI linker ›d(CCGGATCCGG; Takara Shuzo! was added, and ligation was carried out using T4 DNA ligase. The ligation mixture was subjected to cleavage with BamHI and ethanol precipitations (three times) to thereby remove the linker DNA remaining unbound. Separately, the pUC18 DNA (Takara Shuzo) was cleaved with BamHI, the digest was mixed with the above-mentioned BamHI linker-bound 1.7 Kbp DNA fragment, and ligation was performed using T4 DNA ligase. The resulting DNA solution was used to transform E. coli JM109 (Takara Shuzo) and transformants grown on an LB agar plate containing 20 .mu.g/ml kanamycin were obtained. The plasmid DNA was isolated from one of these transformants and named pUC-Tn903-F1. Its structure was confirmed using restriction enzymes. The pUC-Tn903-F1 DNA was cleaved with XhoI and BamHI and a DNA fragment of about 1.21 Kbp was isolated by agarose gel (0.8%). electrophoresis followed by electrophoretic elution. Separately, the IPNS gene promoter was isolated from the M13mp10=IPS4-2 RF DNA by cleavage with HindIII and NcoI followed by acrylamide gel (8%) electrophoresis and electrophoretic elution as a DNA fragment (HingIII. PstI.cndot.SalI-NcoI) of about 0.49 Kbp. ATG occurs in this NcoI site. Therefore, for allowing expression of the G4.18 resistance (kanamycin resistance) gene, the NcoI site was linked to the XhoI site by means of a synthetic DNA. Thus, two 31-meric synthetic DNAs (shown below) were synthesized using an Applied Biosystems model 381A DNA synthesizer.

5'-CATGAGCCATATTCAACGGGAAACGTCTTGC-3'(SEQ ID NO:29)

5'-TCGAGCAAGACGTTTCCCGTTGAATATGGCT-3'(SEQ ID NO:30)

The two synthetic DNAs (10 .mu.g each) were mixed up in 100 .mu.l of TE buffer, heated at 95.degree. C. for 2 minutes and then gradually cooled to room temperature for annealing. The 4.0 Kbp HindIII-BamHI fragment derived from the vector plasmid pBR322, the M13mp10-IPS4-2-derived HindIII-NcoI fragment (about 0.49 Kbp), the annealed synthetic DNA (31 bp) and the pUC-Tn903-F1-derived XhoI-BamHI fragment (about 1.21 Kbp) were mixed up and ligation was carried out using T4 DNA ligase. The ligation mixture was used to transform E. coli by the method of D. Hanahan ›D. Hanahan, J. Mol. Biol., 166, 557 (1983)! and ampicillin-resistant, tetracycline-sensitive transformants were obtained. The plasmid DNA was isolated from each of these transformants and subjected to restriction enzyme analysis for structure confirmation. One of the plasmids was named pBCG-D3.

(4-iii) Joining of the IPNS terminator to the G418 resistance gene:

The RF DNA of M13mp10-IPS3-DS as prepared in (4-i) was cleaved with PvuII and BclI and a DNA fragment of about 1.1 Kbp was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. Separately, the pBCG-D3 DNA prepared in (4-ii) was cleaved with BamHI and PvuII and a BamHI-PvuII DNA fragment of about 4.05 Kbp was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. The BclI-PvuII DNA fragment (BclI-SacI-SmaI. SacI-EcoRI-PvuII) (about 1.1 Kbp) and the BamHI-PvuII DNA fragment (about 4.05 Kbp) were mixed up and ligation was carried out using T4 DNA ligase. The ligation mixture was used to transform E. coli DH1. The plasmid DNA was isolated from each of the ampicillin-resistant transformants thus obtained and the structure thereof was confirmed by analysis with restriction enzymes. One of the plasmids was named pBCG-DT1.

(4-iv) Cloning of the G418 resistance expression unit into pcEP97 (cf. FIG. 1S):

About 10 .mu.g of the pCEP97 DNA (cf. FIG. 1S) was treated with 30 units of PvuII at 37.degree. C. for 1 hour (final volume: 200 .mu.l) for partial cleavage of the DNA. Following phenol extraction and ether extraction, DNA was precipitated with ethanol. This DNA was then completely cleaved with SalI and a DNA of about 6.0 Kbp was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. Separately, about 10 .mu.g of the pBCG-DT1 DNA prepared is (4-iii) was treated with 20 units of SmaI at 37.degree. C. for 30 minutes (final volume: 100 .mu.l). for partial cleavage. Following phenol extraction .and ether extraction, DNA was precipitated with ethanol. This DNA was then completely cleaved with SalI and a DNA fragment of about 2.7 Kbp was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. The PvuII-PvuII-SalI DNA fragment (about 6.0 Kbp) and the SalI-SmaI-SmaI DNA fragment (about 5.7 Kbp) were mixed up and ligation was carried out using T4 DNA ligase. The ligation mixture was used to transform E. coli DH1. Out of the colonies grown on an LB agar plate containing 50 .mu.g/ml ampicillin, 64 transformants capable of growing on an LB agar plate containing 35 .mu.g/ml chloramphenicol were obtained. The plasmid DNA was isolated from each of them and examined for plasmid size by agarose gel (0.8%) electrophoresis, whereupon 13 transformants were found to be the desired ones. One of them was called pCYG-E15 (cf. FIG. 1S) and the structure thereof was confirmed by analysis using restriction enzymes.

(4-v) Cloning of a protease expression unit into pCYG-E15 (construction of pCYG-EB2) (cf. FIG. 1S):

About 20 .mu.g of the pCYG-E15 DNA prepared in (4-iv) was treated with 8 units of HindIII at 37.degree. C. for 30 minutes (final volume: 200 .mu.l) for partial cleavage of the DNA. After phenol extraction and ether extraction, DNA was precipitated with ethanol. This DNA was then completely cleaved with EcoRI and a DNA fragment of about 7.4 Kbp was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. Separately, pCYG-B2 (cf. FIG. 1O) was cleaved with EcoRI and HindIII and a DNA fragment of about 1.5 Kbp was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. The HindIII-HindIII-EcoRI DNA fragment (about 7.4 Kbp) and the EcoRI-HindIII DNA fragment (about 1.5 Kbp) were mixed up and ligation was carried out using T4 DNA ligase. The ligation mixture was used to transform E. coli DE1. Out of the colonies grown on an LB agar plate containing 50 .mu.g/ml ampicillin, transformants incapable of growing on an LB agar plate containing 35 .mu.g/ml chloramphenicol were separated. The plasmid DNA was isolated from each of these transformants and analyzed using restriction enzymes for structure confirmation. One of the desired plasmids thus obtained was named pCYG-EB2 (cf. FIG. 1S).

A plasmid, DCYG-HB51, capable of expressing Hm.sup.R in A. chrysogenum was constructed by introducing the plasmid pLG9O-derived Hm.sup.R gene with its 5' side modified with a synthetic DNA into the plasmid pCYG-B2 (cf. FIG. 1-11) containing a protease gene expression unit, at the BamHI site thereof (cf. FIG. 2-1). The plasmid pLG90 and a method of constructing the same are known ›cf. P. J. M. van den Elzen et al. Plant Molecular Biology, 5, 299-302 (1985) and L. Gritz and J. Davies, Gene, 25, 179-188 (1983)!.

(1) Construction of the plasmid pCYG-HB51

(i) Construction of a plasmid, pHMP-E5:

pLG90 was cleaved with the restriction enzyme HphI and then with the restriction enzyme BamHI. An HphI-BamHI fragment (about 1.03 Kbp) was isolated by agarose gel (1.5%) electrophoresis followed by electrophoretic elution ›cf. Molecular Cloning (cited above), pages 150 and 164!.

Separately, for deleting ATG occurring just upstream from the ATG codon in the Hm.sup.R BamHI DNA fragment of pLG90 the 28-meric and 33-meric DNAs shown in FIG. 2A were Synthesized using an Applied Biosystems model 381A DNA synthesizer according to the operator's manual therefor. The DNAs synthesized were purified using an Applied Biosystems oligonucleotide purification cartridge (OPC cartridge) according to the manual therefor. The DNAs synthesized (each 10 .mu.g) were mixed up in a final volume of 100 .mu.l of TE buffer and heated at 95.degree. C. for 10 minutes. The mixture was then gradually cooled to room temperature for effecting annealing of the two DNAs.

28 mer 5'-TTTTTCATAGCTGTTTCCTGTGGATCCG-3' (SEQ. ID NO:31)

33 mer 5'-AATTCGGATCCACAGGAAACAGCTATGAAAAAG-3' (SEQ. ID NO:32)

A mixed solution containing three DNAs, namely the annealed synthetic DNA, the HphI-BamHI DNA fragment (about 1.03 Kbp) obtained above and an EcoRI-BamHI fragment (2.7 Kbp) of the plasmid pUC18 (Takara Shuzo), was prepared and ligation was carried out using T4 DNA ligase. The ligation mixture was used to transform E. coli JM109 (Takara Shuzo). Out of the colonies grown on an LB agar plate containing 50 .mu.g/ml ampicillin, transformants capable of growing on an LB agar plate containing 150 .mu.g/ml hygromycin B and 0.5 mM IPTG were separated. The plasmid isolated from one of the transformants was named pHMP-E5 (cf. FIG. 2A). Its structure was confirmed using restriction enzymes. nucleotide sequence of a synthetic DNA region-containing portion of pHMP-E5 was, determined by the dideoxynucleotide synthetic chain termination method and found to be as designed.

(ii) Construction of pCYG-HB51:

pHMP-E5 was cleaved with BamHI and a DNA about 1.06 Kbp in size was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. Separately, the expression vector in A. chrysogenum pCYG-B2 (cf. FIG. 1O) was cleaved with BamHI. Both BamHI DNA fragments were mixed up and ligated to each other using T4 DNA ligase, and the ligation mixture was used to transform E. coli JM109. Out of the colonies grown on an LB agar plate containing 50 .mu.g/ml ampicillin, transformants capable of growing on an LB agar plate containing 150 .mu.g/ml hygromycin B were collected. Since both the A. chrysogenum ATCC 11550-derived protease gene promoter and terminator show promoter activity in E. coli, the presence of an SD sequence (ribosome binding site sequence) before an ATG codon would induce translation in E. coli. Therefore, the Ap.sup.R Hm.sup.R strains might include strains containing the Hm.sup.R gene in one direction and strains containing said gene in the reverse direction. Therefore the plasmid DNA was isolated from each of these strains, and cleaved with EcoRI and the direction of the Hm.sup.R gene was judged by agarose gel (0.8%) electrophoresis. One of the plasmids in which said gene was in the same direction with that of the protease gene promoter was named pCYG-HB51 (cf. FIG. 2A). The nucleotide sequence of the hygromycin B resistance gene inserted in this plasmid pCYG-HB51 is shown in FIGS. 2B and 2C, together with the amino acid sequence encoded by said nucleotide sequence.

(1) Construction of the plasmid pDAO-EB101 (expression vector for a D-amino acid oxidase gene) and the plasmid pVEB104 (expression vector for a CC acylase gene)

For direct fermentative production of 7ACA or 7ADCA using a cephalosporin C (CC)- or deacetylcephalosporin C (DCC)-producing microorganism belonging to the species Acremonium chrysogenum, the routes shown in FIG. 3B were conceived. Thus the routes should respectively include:

(1) transformation of a CC or DCC producer With a 7ACA or 7ADCA production vector containing a CC acylase gene;

(2) transformation of a CC or DCC producer with a 7ACA or 7ADCA production vector containing a D-amino acid oxidase (DAO) gene and a CC acylase gene; and

(3) transformation of a CC or DCC producer with a 7ACA or 7ADCA production vector containing a DAO gene and a GL-7ACA acylase gene.

For the above purposes, an expression vector containing a CC acylase gene and an expression vector containing a DAO gene were prepared. It was confirmed that each enzyme could be expressed in Saccharomyces cerevisiae. The procedure followed in this Example 3 is schematically shown in FIG. 3A.

(i) Construction of pDAO-EB101:

The DAO expression plasmid pCFS315 (cf. FIG. 3E) was isolated from E. coli JM109 (pCFS315). (FERM BP-1916) by a conventional method and cleaved with BamHI, and a 1.24 Kbp DNA was isolated by agorose gel (0.8%) electrophoresis followed by electrophoretic elution. The nucleotide sequence of the DAO cDNA contained in the plasmid pCFS315 is shown in FIGS. 3-11. Separately, the vector pCYG-EB2 DNA (cf. FIGS. 1-15) having an expression unit in A. chrysogenum was cleaved with BamHI. Both the BamHI DNA fragments were mixed up and ligated to each other using T4 DNA ligase. The ligation mixture was used to transform E coli DE1 (ATCC 33849). Transformants capable of growing on an LB agar plate containing 50 .mu.g/ml ampicillin were collected. The plasmid DNA was isolated from each of these transformants, cleaved with BamHI or EcoRI+PvuII and subjected to agorose gel (1.5%) electropharesis. One of the plasmids that gave a 1.24 Kbp DNA fragment upon BamHI cleavage and a 1.33 Kbp DNA fragment upon EcoRI+PvuII cleavage was named pDAO-EB101 (cf. FIG. 3C).

(ii) Construction of pVEB104:

Pseudomonas diminuta V22 showing CC acylase activity was isolated from a soil sample and an acylase gene was cloned from the genomic DNA thereof using E. coli as a host, with CC acylase activity as a index. The N-terminal side of this gene was excised with PstI, the C-terminal side was deleted by treatment with Bal31 to reduce the gene size to about 3 Kbp, and pCPV22P (cf. FIG. 3-6) was constructed using this gene and Km.sup.R pHSG298 ›Tahara Shuzo; S. Takeshita et al., Gene, 61, 63-74 (1987)! as the vector therefor.

Starting with the CC acylase gene of pCPV22P and deleting the N-terminal side to the MluI site located upstream from ATG, pV22B1 was constructed (of FIG. 3F).

It was found that, in the vicinity of an AatII site just upstream from the ATG codon of the V22 CC acylase gene, there was another ATG belonging to a different frame. Therefore this ATG was deleted by using a synthetic DNA adapter and pV22BS-A11 (cf. FIG. 3I) was constructed. By cloning the acylase gene, with said ATG deleted, in the expression vector pCYG-EB2, pVEB104 (cf. FIG. 3D) was constructed.

(a) Construction of pV22B1 and determination of the N-terminal side nucleotide sequence:

The pCPV22P plasmid was isolated from E. coli JM109 (pCPV22P) (FERM-BP 2704) and 15 .mu.g of this DNA was treated with 30 units of the restriction enzyme, MluI (Toyoho) at 37.degree. C. for 15 minutes for partial cleavage of the DNA. Following phenol extraction and ether extraction, DNA was precipitated with ethanol and treated with the Klenow fragment (large fragment E. coli DNA polymerase I) (Takara Shuzo) in the presence of dNTP for rendering the cleavage site blunt-ended. Following phenol extraction and ether extraction, DNA was precipitated with ethanol. The DNA was cleaved with EcoRI and a DNA about 1.1 Kbp in size was isolated by agarose gel (1.5%) electrophoresis followed by electrophoretic elution. The above DNA fragment (about 1.1 Kbp) and a SmaI-EcoRI DNA fragment (2.7 Kbp) of the vector plasmid pUC19 (Takara Shuzo) were mixed up and ligated to each other using T4. DNA ligase to give a plasmid, pV22F2 (cf. FIG. 3F).

Since there was no BamHI site on the acylase C-terminus side, introduction of a BamHI site on the C terminus side was attempted by reversing the direction of the acylase gene of pCPV22P. Thus, pCPV22P was cleaved with PstI and selected to religation using T4 DNA ligase, and the religation mixture was used to transform E. coli JM109. White transformant colonies grown on an LB agar plate containing 50 .mu.g/ml, 100 .mu.g/ml X-gal and 0.5 mM IPTG were recovered (blue clones contained the vector pHSG298 having no acylase gene insert). The plasmid DNA was isolated from each of these transformants and cleaved with SalI, and pV22R6 (cf. FIG. 3F) with the acylase gene inserted therein in the reverse direction was isolated by agarose gel (0.8%) electrophoresis.

PV22F2 was cleaved with HindIII and EcoRI, and a HindIII-EcoRI fragment (about 1.2 Kbp) was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. Separately, pV22R6 was cleaved with EcoRI and BamHI and an EcoRI-BamHI fragment (about 1.4 Kbp) was isolated by agaross gel (0.8%) electrophoresis followed by electrophoretic, elution. The above-mentioned HindIII-EcoRI fragment (about 1.2 Kbp) and EcoRI-BamHI fragment (about 1.4 Kbp) and further a HindIII-BamHI fragment of the vector plasmid pHSG399 ›Takara Shuzo; plasmid containing Cm.sup.R as a marker; E. coli JM109 (pHSG399) can grow or an LB agar plate containing 30 .mu.g/ml chloramphenicol! were mixed up and ligation was carried out using T4 DNA ligase to give the plasmid pV22B1 (cf. FIG. 3F).

pV22B1 was cleaved with BamHI and EcoRI and a BamHI-EcoRI fragment (about 2.5 Kbp) was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. This DNA fragment was cloned into an EcoRI-BamHI fragment of M13mp18 and of M13mp19 and the DNA nucleotide sequence of the open reading frame of the acylase gene inserted in pV22B1. was determined (cf. FIGS. 3A and 3H). For the nucleotide sequence determination, .alpha.-.sup.32 P-dATP and seguenase ›cf. S. Tabor and C. C. Richardson, J. Biol. Chem., 264, 6447 (1989)! (United States Biochemical Corporation) were used and the dideoxynucleotide synthetic chain termination method of Sanger et al. ›F. Sanger et al., Proc. Natl. Acd. Sci. USA, 74, 5463 (1977)! was followed. Comparison of the amino acid sequences of the V22-derived acylase and the known Pseudomonas sp. SE83-derived acylase ›A. Matsuda et al., J. Bacteriol., 169, 5821 (1983)! revealed that 53 out of 774 amino acid residues were different.

(b) Construction of pV22BS-A11 (deletion of ATG upstream from ATG):

pV22B1 was cleaved with EcoRI and DNA fragments of about 3.3 Kbp and about 1.38 Kbp were isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. The 1.38 Kbp DNA fragment was used in constructing pV22BS-A11 from pV22BDS1. The 3.3 Kbp DNA fragment was subjected to ligation using T4 DNA ligase to give a plasmid, pV22BND3 (cf. FIG. 3I).

pV22BD3 was cleaved with AatII and BamHI and a DNA of about 3.3 Kbp was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. Separately, two synthetic DNAs (19-met and 11-mer) were prepared by a conventional method.

19mer 5'-GATCC GGTACC AAG GACGT-3' (SEQ ID NO:33)

11mer 5'-CCTTGGTACCG-3' (SEQ ID NO:34) The synthetic DNAs were designed such that a KpnI site could be formed for facilitating detection of clones with the same inserted therein. The synthetic DNAs (10 .mu.g each ) were mixed up in a final volume of 100 .mu.l of TE buffer, heated at 90.degree. C. for 2 minutes and cooled gradually to room temperature for effecting annealing of the two DNAs. The above-mentioned AatII-BamHI DNA fragment (about 3.3 Kbp) and the annealing product DNA were mixed up and ligated to each other using T4 DNA ligase to give a plasmid, pV22BDS1 (cf. FIG. 3I). Since this DNA contained the newly introduced KpnI site, it was easy to determine whether the synthetic DNA portion was present in a subsequent product.

pV22BDS1 was cleaved with EcoRI and then mixed with the above-mentioned. EcoRI-EcoRI DNA fragment (about 1.38 Kbp), and ligation was carried out using T4 DNA ligase. The ligation mixture was used to transform E. coli DH1 and transformants growing-on an LB agar plate containing 30 .mu.g/ml chloramphenicol were collected. The plasmid DNA was isolated from each of these transformants and cleaved with PstI. Agarose gel (0.8%) electrophoresis gave pV22BS-A11 which contained the 1.38 Kbp EcoRI DNA fragment in the desired direction (cf. FIG. 3I). Cleavage of pV22BS-A11 with PstI gave two DNA fragments about 2.5 Kbp and 2.2 Kbp in size, hence can be readily distinguished from others. The restriction enzyme cleavage map of the acylase gene region (BamHI, about 2.5 Kbp) of pV22BS-A11 is shown in FIG. 3K.

(c) Cloning of the acylase gene of pV22BS-A11 into a gene expression unit capable of functioning in chrysogenum:

pV22BS-A11 was cleaved with BamHI and a DNA of about 2.5 Kbp was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. Separately, the expression vector pCYG-EB2 (cf. FIG. 1S) for use in A. chrysogenum was cleaved with BamHI. Both the BamHI DNA fragments were mixed up and ligated to each other using T4 DNA ligase, and the ligation mixture was used to transform E. coli DE1. Transformants growing on an LB agar plate containing 50 .mu.g/ml ampicillin were recovered. The plasmid DNA was isolated from each of these transformants and cleaved with BamHI and EcoRI. One of the plasmids which gave a DNA fragment of about 2.5 Kbp upon cleavage with BamHI and a DNA fragment of about 1.67 Kbp (about 1.96 Kbp if the acylase gene was in the reverse direction) upon cleavage with EcoRI was named pVEB104 (cf. FIG. 3D).

(iii) Expression of pDAO-EB101 and pVEB104 in S. cerevisiae:

It was examined whether the A. chrysogenum ATCC 11550-derived protease gene expression unit F. solani M-0718-derived D-amino acid oxidase cDNA and P. diminuta V22-derived acylase gene could function to give enzyme activity in S. cerevisiae YNN27 ›D. T. Stinchcomb et al., Proc. Natl. Acad. Sci. USA, 77, 4559 (1980)!, one of lower eukaryotes.

Transformation of S. cerevisiae YNN7 with pDAO-EB101 or pVEB104. was performed as described in Japanese Kokai Tokkyo Koho No. 61-209593 (Isogai et., laid open Sep. 17, 1986) (in Example 2-III). Thus, the protoplast of S. cerevisiae YNN27 was mixed with 10 .mu.l (about 5 .mu.g) of each DNA and a buffer containing 20% PEG 4000 was added to the mixture for effecting transformation. Selection using about 300 .mu.g/ml of the antibiotic 0415 gate about 1.times.10.sup.4 transformants.

Each transformant was inoculated into YEPD medium (10 g/liter yeast extract, 20 g/liter peptone, 20 g/liter glucose) (5 ml) containing 10 .mu.g/ml uracil, 40 .mu.g/ml tryptophan and 300 .mu.g/ml G415 and cultivation was carried out at 30.degree. C. for 3 days. Cells were then collected by centifugations. The cells were reacted with CCNa or GL-7ACA and the supernatant separated from cells by centrifugation was subjected to high performance liquid chromatogrphay ›column: Inertsil ODS-2 (GasChro Kogyo); mobile phase: solution composed of 6.6 mM phosphate buffer (pH 7.3) and 3% methanol; detection: 254 nm! for product quantification.

In the case of pDAO-EB101-carrying transformants, 500 .mu.l of a reactant solution containing 5 mg/ml CCNa, 0.1M phosphate buffer (pH 7.5) and 14 mM NaN.sub.3 and 5 .mu.l of toluene were added to centrifuged cells and the reaction was carried out at 37.degree. C. for 3 hours with shaking. NaN.sub.3 was added so that it could inhibit catalase to thereby allow DAO-catalyzed conversion of CCNa to GL-7ACA without stopping at the stage of keto-AD-7ACA. In this way, 840 .mu.g/ml GL-7ACA was formed. In this case, it was also confirmed that transformants carrying pCYG-EB2 containing no DAO gene, which were used as controls, did not give GL-7ACA. It was thus found that the A. chrysogenum ATCC 11550-derived protease gene expression unit can function in S. cerevisiae YNN27, causing D-amino acid oxidase formation. It was also found that the F. solani M-0718-derived DAO cDNA cab be expressed in S. cerevisiae YNN27.

In the case of pVEB104-carrying transformants, 500 .mu.l of a reaction solution containing 2 mg/ml GL-7ACA and 0.1M Tris-HCl buffer (pH. 8.0) and 5 .mu.l of toluene were added to centrifuged cells and the reaction was allowed to proceed at 30.degree. C. for 3 hours with shaking, whereby 7ACA was formed in an amount of 15 .mu.g/ml. Control transformants carrying pCYG-EB2 containing no acylase gene failed to cause 7ACA formation. Further it was found that the P. diminuta V22-derived acylase gene can be expressed in S. cerevisiae YNN27, though to a lesser extent.

(1) Construction of the plasmids pEBV1 (vector for 7ACA and 7ADCA production), pEDV11 (vector for 7ACA and 7APDA production) and pHBD3 (vector for GL-7ACA and GL-7ADCA production)

In DNA introduction into Acremonium chrysogenum BC2116 (FERM-BP 2707), which is a CC and DCC producer, the hygromycin B resistance was used as a marker since selection is difficult when the G418 resistance is used as a selective marker. Thus the CC acylase gene- and Hm.sup.R gene-containing plasmid pHBV1 (cf. FIG. 4A) for 7ACA and 7ADCA production, the D-amino acid oxidase gene-, CC acylase (capable of functioning also as keto-AD-7ACA acylase and GL-7ACA acylase) gene- and Hm.sup.R gene-containing plasmid pHDV11 (cf. FIG. 4B) for 7ACA and 7ADCA production and the D-amino acid oxidase gene- and Hm.sup.R gene-containing plasmid pHBD3 (cf. FIG. 4C) for GL-7ACA and GL-7ADCA production were constructed

(i) Construction of pHBV1 (cf. FIG. 4A):

About 15 .mu.g of pCYG-EB51 was treated with 10 units of EcoRI at 37.degree. C. for 15 minutes (final volume: 100 .mu.l) for partial cleavage of the DNA. Following phenol extraction and ether extraction, DNA was precipitated with ethanol. This DNA was then completely cleaved with SmaI and a DNA fragment of about 3.1 Kbp was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. Separately, pVEB104 was cleaved with SacI and PvuII and a DNA fragment of about 5.7 Kbp was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. The above-mentioned EcoRI-SmaI fragment (about 3.1 Kbp) and PvuII-SacI DNA fragment (about 5.7 Kbp) and, further, an EcoRI-SacI DNA fragment (2.7. Kbp) of the vector plasmid pHSG298 (Tahara Shuzo) were mixed up and ligation was conducted using T4 DNA ligase. The ligation mixture was used to transform E. coli JM109 by the method of D. Hanahan ›D. Hanahan, J. Mol. Biol., 163, 557-580 (1983)! and transformants growing as white clones on an LB agar plate containing 20 .mu.g/ml kanamycin, 0.5 mM IPTG and 100 .mu.g/ml X-gal were obtained. From these transformants, those capable of growing on an LB agar plate containing 150 .mu.g/ml hygromycin but incapable of growing on an LB agar plate containing-50 .mu.g/ml ampicillin were recovered. The plasmid DNA was isolated from each of the thus-obtained Km.sup.R Hm.sup.R Ap.sup.S strains and analyzed by cleavage with restriction enzymes. One of the plasmids thus found to have the desired structure was named pHBV1 (cf. FIG. 4A).

(ii) Construction of pHDV11 (cf. FIG. 4B):

pDAO-EB101 was cleaved with ClaI and a DNA fragment of about 7.5 Kbp was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. Further, pHBV1 was cleaved with. ClaI (for cleavage at one site alone). The cleavage mixture was mixed with the above-mentioned ClaI fragment (about 7.5 Kbp) and ligation was carried out using T4 DNA ligase. The ligation mixture was used to transform E. coli JM109 and transformants growing on an LB agar plate containing 50 .mu.g/ml ampicillin were collected. From among these transformants, strains capable of growing on an LB agar plate containing 150 .mu.g/ml hygromycin but incapable of growing on an LB agar plate containing 30 .mu.g/ml kanamycin were selected. The plasmid DNA was isolated from each of the thus-obtained Ap.sup.R Hm.sup.R Km.sup.S strains. Based on the BamHI cleavage pattern, whether said plasmid was the desired one or not was judged and the direction of gene insertion was determined. One of the plasmids thus found to be as designed was named pHDV11 (cf. FIG. 4B).

(iii) Construction of pPIBD3 (cf. FIG. 4C):

About 15 .mu.g of pCYG-HB51 was treated with 10 units of EcoRI at 37.degree. C. for 15 minutes (final volume: 100 .mu.l) for partial cleavage of the DNA. Following phenol extraction and ether extraction, DNA was precipitated with ethanol. This DNA was then completely cleaved with ClaI and a DNA fragment fraction of about 4.3 Kbp was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution said fraction was a mixture of a fragment of about 4.28 Kbp and a fragment of about 4.44 Kbp. If desired, however, the necessary DNA fragment (about 4.28 Kbp) can be isolated by utilizing Hm.sup.R Ap.sup.S. Separately, about 20 .mu.g of pDAO-EB101 was treated with 50 units of PstI at 37.degree. C. for 5 minutes (final volume: 200 .mu.l). for partial cleavage of the DNA. Following phenol extraction and ether extraction, DNA was precipitated with ethanol. This DNA was then cleaved with ClaI and a DNA fragment fraction of about 6 Kbp was isolated by agarose gel (0.8%) electrophoresis followed by electrophoretic elution. This DNA fragment fraction included the following three fragments: PstI.sup.DAO PstI--ClaI (about 5.63 Kbp), PstI.sup.DAOP stI ClaI--ClaI (about 6.53 Kbp) and ClaI--PstI.sup.DAOP stI (about 5.80 Kbp). If desired, however, the last fragment (about 5.50 Kbp) can be removed by selecting Ap.sup.S transformants.

The above-mentioned first DNA fragment mixture (about 4.3 Kbp) and second DNA fragment mixture (about 6 Kbp) and, further, an EcoRI-PstI fragment (2.6 Kbp) of the vector plasmid pH0298 were mixed up and subjected to ligation using T4 DNA ligase. The ligation mixture was used to transform E. coli JM109 by the method of D. Handhah and transformants growing as white colonies on an LB agar plate containing 20 .mu.g/ml kanamycin, 0.5 mM IPTG and 100 .mu.g/ml X-gal were recovered. 800 strains out of these transformants were tested for growing on an LB agar plate containing 150 .mu.g/ml hygromycin and an LB agar plate containing 50 .mu.g/ml ampicillin. Nineteen Km.sup.R Hm.sup.R Ap.sup.S transformants were thus obtained. The plasmid DNA was isolated from each of these strains and subjected to restriction enzyme cleavage analysis for confirmation that it was the desired one. One of the plasmids thus obtained was named pHBD3 (cf. FIG. 4C).

(1) Introduction of the plasmids pHBV1, pHDV11 and pHBD3 into A. chrysogenum BC2116 and cultivation of the transformants

A. chrysogenum BC2116 was transformed with pHBV1 or pHDV11 (vector for 7ACA and 7ADCA production in A. chrysogenum) or pHBD3 (vector for GL-7ACA and GL-7ADCA production in A. chrysogenum). The thus-obtained hygromycin B-resistant transformants were cultivated in a CC production medium (without addition of hygromycin B). It was found that said transformants can produce 7ACA or GL-7ACA in the medium.

(i) Transformation of A. chrysogenum BC2116:

(a) Preparation of a uniform inoculum for cell culture:

An ampule of a cell suspension (20% glycerol) obtained by thawing, at 37.degree. C., a stock culture of A. chrysogenum BC2116 stored in liquefied nitrogen was inoculated onto ten B3 agar plates dried in advance until the surface moisture could not be observed any longer. The agar plates thus inoculated were incubated at 30.degree. C. for 6 days. The mycelial mass (inclusive of spores) covering each agar plate surface was scraped up while avoiding inclusion of the agar and suspended in 3 ml of 20% glycerol. The suspension was distributed into 5 ampules and, frozen stored in a liquefied nitrogen-cooled cell storage chamber (the ten B3 agar plates giving 50 ampules).

(b) Cell multiplication for protoplast preparation:

The frozen cell suspension (A. chrysogenum BC2116) in one ampule was thawed at 37.degree. C. and the whole amount was inoculated into 50 ml of YSP medium placed in a 250-ml shaking flask and shake culture was performed at 30.degree. C. for 4 days. A 5-ml portion of this preculture was transferred to 50 ml of fresh YPS medium and cultivation was conducted at 30.degree. C. as mentioned above for 24 hours.

(c) Protoplast preparation:

Cells were harvested from the thus-obtained 24-hour culture (200 ml) by centrifugation (3,000 rpm, 5 minutes), washed twice with sterile water (200 ml) by centrifugation and suspended in 80 ml of 10 mM Tris HCl (pH 7.5) Containing 10 mM DTT. The suspension was shaken gently at 30.degree. C. for 1 hour. Cells were then collected by centrifugation (3,000 rpm, 5 minutes) washed twice with 100 ml of 1M KCl buffer (pH 5.8) by centrifugation and suspended in 1M KCl buffer (pH 5.8) to a suspension volume of 20 ml. To this suspension (20 ml) was added 30 ml of 1 M KCl buffer (pH 5.8.) containing 16.3 mg/ml Novozyme 234 (Novo Biolabs). The mixture was shaken gently at 30.degree. C. for 30 minutes. After completion of this procedure, the resultant protoplast suspension was placed in a disposable centrifugal tube, stirred for 2 to 3 seconds, then diluted with 50 ml of 1M KP-buffer (pH. 7.5) and centrifuged (750 rpm, 2 minutes). The protoplast sedimented by centrifugation was resuspended in 50 ml of 1M KCl buffer (pH 7.5) and then harvested and at the same time washed by centrifugation (1,500 rpm, 5 minutes). This washing cycle was repeated twice. The protoplast washed was suspended in 0.8M NaP buffer to a final volume of about 5 ml.

(d) Transformation:

To 60 .mu.l (about 20 .mu.g DNA) of each plasmid DNA solution to be used for transformation was added 240 .mu.l of 1M KP buffer (pH 7.5). After mixing up, 400 .mu.l of the protoplast suspension was added. After mixing up, the mixture was allowed to stand in ice for 30 minutes, then 4 ml of 10 mM Tris-HCl (pH 7.5) containing 40% PEG, 10.8% sucrose and 50 mM CaCl.sub.2 was added and, after mixing up, the mixture was allowed to stand at room temperature for 15 minutes. Then, 10 ml of 0.8M NaP buffer was added to the mixture and, after mixing up, the whole mixture was centrifuges at 1,000 rpm for 5 minutes. The sediment was suspended in 1.2 ml of 0.8M NaP buffer.

(e) Selection of transformants resistant to the antibiotic hygromycin B:

The transformant suspension (0.2 ml) was mixed with 5 ml of BRM agar medium (48.degree. C.) and the mixture was poured onto a BRM agar medium plate (25 ml). After incubation at 20.degree. C. for 20 hours, 10 mg/ml hygromycin B (Calbiochem Corporation; imported and distributed by Nakalai Tesque) was spread over the plate to a final concentration of 25 .mu.g/ml or 50 .mu.g/ml using a spreader (three plates for each concentration). After 2 to 3 weeks of incubation at 30.degree. C., each colony that had appeared was transferred to a PDA-YE agar plate containing 50 .mu.g/ml hygromycin B. Incubation was further continued at 30.degree. C. for 7 days. In this way, hygromycin B-resistant transformants were obtained. Since abortive transformants cannot grow on fresh medium in the presence of hygromycin B in such subculture as mentioned above, stable transformants can be readily distinguished from abortive transformants. The transformants were each again spread over a PDA-YE agar plate containing 40 a .mu.g/ml hygromycin B and subcultured (30.degree. C., 5 to 7 days).

The above procedure generally gave 1 to 3 hygromycin B-resistant transformants per about 20 .mu.g of DNA. The number of colonies regenerated on the BRM agar medium plate from the final protoplast suspension obtained as described above in section (c) was about 2.times.10.sup.8 /ml (after 2 weeks of incubation at 30.degree. C.

(ii) 7ACA and GL-7ADCA production by hygromycin B-resistant A. chrysogenum BC2116 transformants:

Hygromycin B-resistant transformants (cf. Table 1) were produced by transforming A. chrysogenum BC2116 with pCYG-HB51 (Hm.sup.R alone), pHBV11 (vector for 7ACA and 7ADCA production), pPIDV11 (vector for 7ACA and 7ADCA production) or pHBD3 (vector for GL-7ACA and GL-7ADCA production) as described above in section (i). Each transformant was inoculated into 50 ml of CS1 medium placed in a 250-ml shaking flask. After 4 days of incubation at 30.degree. C., 1 ml of the resultant preculture was transferred to a 250-ml shaking flask containing 20 ml of the main culture medium and shake culture (stroke: 3 inches; 250 rpm) was conducted at 25.degree. C. for 6 to 7 days. Each main culture was filtered through a Toyo filter paper No. 2 and-the filtrate was assayed by HPLC (high-performance liquid chromatography). The HPLC conditions were as follows:

Column: Two columns connected, namely Cosmosil 5C.sub.18 column (4.6.times.150 mm) (Nacalai-Tesque) directly followed by Inertsil ODS-2 column (5.times.150 mm) (Gaskuro Kogyo)

Column temperature: 40.degree. C.

Mobile phase: 4.0. mM (0.567 g/liter) Na.sub.2 HPO.sub.4, 2.6 mM (0.36 g/liter) KH.sub.2 PO.sub.4, 4% methanol

Rate of flow: 1 ml/min

UV detection: 254 nm

Assaying under the above HPLC conditions gave the results shown in Table 1. The positions of (retention times for) 7ACA, GL-7ACA and CCNa were 17.7 minutes, 24.3 minutes and 18.9 minutes, respectively. ›Each culture filtrate was 10-fold diluted with 0.5M citrate buffer (pH 4.0) and 10 .mu.l of the dilution was subjected to HPLC.! The results indicate that while transformants obtained with the plasmid pCYG-HB51 characterized by hygromycin B resistance alone produced neither 7ACA nor GL-7ACA, transformants obtained with the vector pHBV1 for 7ACA and 7ADCA production produced. 7ACA in an amount of about 50 .mu.g/ml and transformants obtained with the vector pHBD3 for GL-7ACA and GL-7ADCA production produced GL-7ACA in an amount of about 130 .mu.g/ml. Furthermore, transformants obtained with the 7ACA and 7ADCA production vector pEVD11 containing both the D-amino acid oxidase gene and acylase gene produced 7ACA in an amount of about 150 .mu.g/mi but did not produce GL-7ACA or keto-AD-7ACA. Even GL-7ACA producers did not produce keto-AD-7ACA. This is presumably because keto-AD-7ACA was decomposed due to its instability.

For reidentification of the product 7ACA, assaying was further made under different, HPLC conditions to give similar results. The HPLC conditions used in this reassay were as follows:

Column: Cica-Merck pre-packed column for high performance liquid chromatography column (4.times.250 mm)

LiChrospher 100 RP-18(e) (5 .mu.m) (Kanto Chemical)

Column temperature: Room temperature

Mobile phase:

0.92 g/liter Sodium 1-hexanesulfonate (Tokyo Kassei),

1.32 G/liter 18-crown-6 (Nacalai Tesque),

21 g/liter citric acid,

2.47 g/liter trisodium citrate (dihydrate),

10% acetonitrile

Rate of flow: 1 ml/min.

UV detection: 254 nm

(iii) Analysis of hygromycin-resistant transformants by Southern hybridization:

Hygromycin B-resistant transformants (A. chrysogenum Hm144, Hm172, Hm155, Hm146, Hm154, Hm156, Hm161, Hm178, Hm165, Hm164, Hm168 and Hm179) were each shake-cultured in YPS medium (50 ml) containing 12.5 .mu.g/ml hygromycin B at 30.degree. C. for 5 to 7 days. Cultured cells were collected by centrifugation and stored at -20.degree. C. Cells were disrupted in a mortar cooled with liquefied nitrogen, 5 ml of 50 mM Tris-HCl buffer 7.5) containing 10 mM EDTA and 0.35 ml of 20% SDS were added, and the mixture was heated at 65.degree. C. for 20 minutes. After phenol extractions(twice), DNA was precipitated with ethanol. The precipitate was dissolved in 2.5 ml of the above-mentioned 10 mM EDTA-containing buffer, then RNase A (ribonuclease) was added to a concentration of 4 .mu.g/ml, and the mixture was maintained at 37.degree. C. for 1 hour. Protease K was then added to a concentration of 100 .mu.g/ml and the mixture was maintained at 37.degree. C. for 2 hours. Following phenol extraction, DNA precipitation was caused with ethanol and the precipitate was dissolved in about 300 .mu.l of TE buffer. The DNA solution was dialyzed against TE buffer.

Each DNA thus recovered was cleaved with the restriction enzymes BamHI and EcoRI and subjected to agarose gel (0.8%) electrophoresis, which was followed by DNA transcription from the electrophoretic gel to a nitrocellulose filter by the method of Southern (Molecular Cloning, pages 382-386, Cold Spring Harbor Laboratory, 1982). The pCPV22P DNA ›cf. FIG. 3-6; derived from the vector pHSG298 by insertion of the P. diminuta V22-derived acylase gene (about 3-Kbp)! was cleaved with the restriction enzyme PstI and then .sup.32 P-labeled using a mixed primer labeling system (Clontech; imported and distributed by Toyobo). This labeled DNA and the above-mentioned DNA-bound nitrocellulose filter were subjected to Southern hybridization (Advanced Bacterial Genetics, pages 174-177, Cold Spring Harbor Laboratory, 1980).

While no hybridization was observed between the A. chrysogenum BC2116 DNA and pCPV22P, DNAs from the twelve hygromycin B-resistant transformants showed distinct hybridization. Furthermore, BamHI-cleaved DNAs from the acylase gene-containing strains A. chrysogenum Hm172, Hm155, Hm146, Hm154, Hm156, Hm161, Hm178 and Hm165 gave an hybrid band corresponding to the acylase gene of about 2.5 Kbp. These results indicate that the DNA in question was successfully introduced into hygromycin B-resistant transformants. Furthermore, the results of Southern hybridization indicate that the DNA introduced did not occur in a plasmid form but occurred integrated in the genomic DNA.

The 7ACA producing strain Acremonium chrysogenum Hm178 was cultivated in the same manner as in Example 5. The thus-obtained culture (1,000 ml) was centrifuged at 8,000 rpm for 5 minutes. The supernatant obtained was adjusted to pH 5.0 with 1N HCl (about 10 ml) and the resultant precipitate was removed by suction filtration. The filtrate (600 ml; 7ACA content: 136 .mu.g/mt) was subjected to column chromatography using Diaion HP-20 (600 ml; Mitsubishi Kasei). After washing with 600 ml of acidified water (pH 3.5) and 600 ml of water (pH 7.0), elution was carried out with 30% aqueous isopropyl alcohol. The combined 7ACA-containing eluate fractions (1,200 ml) were concentrated at 30.degree. C. under reduced pressure and the concentrate (60 ml) was applied to a YMC reversed-phase column (ODS A60 200/60 mesh, Yamamura Kagaku Kenkyusho; 1 liter). Development was carried out with water. 7ACA (39.8 mg) was eluted in fractions 360 ml to 1960 ml. The combined 7ACA-containing eluate fractions (1,600 ml) were concentrated to 114 ml at 30.degree. C. under reduced pressure. Since this concentrate contained. CC in large excess relative to 7ACA, 5.4 ml of DAO ›cf. Japanese Patent Application No. 1-266795 (266795/1989); 182 units/ml!, 2 ml of catalase C-10 (Sigma; 10 mg/ml) and 13 ml of 1M phosphate buffer (pH 7.3) were added to the concentrate (114 ml) and the mixture was shaken at 25.degree. C. for 1 hour for decomposing CC. The reaction mixture was adjusted to pH 1.5 with 1N HCl, then washed with an equal volume of ethyl acetate, and the aqueous layer obtained (127 ml; 7ACA: 21.6 mg) was concentrated to 50 ml at 30.degree. C. under reduced pressure.

The concentrate was divided into two 25-ml portions and subjected to high performance liquid chromatography using two YMC reversed-phase columns equibrated in advance with 2% methanol-6.6 mM phosphate buffer (pH 7.3) (ODS packed column, R-354 S-15/30 .mu.m, 50.times.300 mm.times.2 columns, Yamamura Kagaku Kenkyusho; UV detection: 254 nm). Development was carried out using the same solvent system as used in column eguilibration at a flow rate of 100 ml/min. The eluate was fractionated. 7ACA-containing fractions were combined (600 ml), adjusted to pH 5.0 with 1N HCl and subjected to column chromatography using Diaion HP-20 (60 ml), development being conducted with water (1,200 ml). 7ACA was eluted in fractions 600 ml to 1,600 ml. The 7ACA-containing fractions were combined and concentrated to dryness at 30.degree. C. under reduced pressure to give 5.36 mg of 7ACA as a white powder. The .sup.1 H MNR, spectrum and IR spectrum of this product were in good agreement with those of a standard.

Hygromycin-B-resistant transformants were prepared by transforming A. chrysogenum BC2116 with pCYG-HB51 (Hm.sup.R alone), pHBV1 (vector for 7ACA and 7ADCA production), pHDV11 (vector for 7ACA and 7ADCA production) or pHBD3 (vector for GL-7ACA or GL-7ADCA production) in the same manner as described in Example 5 (cf. Table 2). These transformants Were each inoculated into 50 ml of CSl medium placed in a 250-ml shaking flask and cultured at 30.degree. C. for 4 days. One milliliter of this preculture was transferred to each of ten 250-ml shaking flasks containing 20 ml of the main culture medium. Shake culture (stroke: 3 inches; 250 rpm) was conducted at 25.degree. C. for 3, 4, 5, 6 or 7 days. The main culture (two flasks for each culture period) was filtered through a Toyo No. 2 filter paper. A 100-.mu.l portion of the filtrate was added to 900 .mu.l of 0.1M phosphate buffer (pH 6.0) for dilution and assayed by HPLC-under the following conditions:

Column: Two columns connected, namely Cosmosil

5C.sub.18 column (4.6.times.150mm) (Nacalai

Tesque) directly followed by Inertsil

ODS-2 column (5.times.150 mm) (Gaskuro Kogyo)

Column temperature: 40.degree. C.

Mobile phase: 2.2 mM tetra-n-butylammonium hydroxide,

2.82 g/liter (NH.sub.4).sub.2 HPO.sub.4

(adjusted to pH 7.3 with phosphoric

acid),

5.63% methanol

Rate of flow: 1 ml/min.

UV detection: 254 nm

Assaying under the above HPLC conditions gave the results shown in Table 2. In this assay, the positions of (retention times for) DCC, 7ADCA and GL-7ADCA were 5.5 minutes,, 6.1 minutes and 26.9 minutes, respectively. As is evident from the results, transformants obtained with the plasmid pCYG-HB51 characterized by hygromycin B resistance alone produced neither 7ADCA nor. GL-7ADCA whereas transformants obtained with the 7ACA and 7ADCA production vector pHBV1 produced 7ADCA in an amount of about 24 .mu.g/ml and transformants obtained with the GL-7ACA and GL-7ADCA production vector pHBD3 produced GL-7ACA in an amount of about 375 .mu.g/ml. Furthermore, transformants obtained with the 7ACA and 7ADCA production vector pHDV11 containing both the D-amino acid oxidase and acylase genes produced 7ADCA in an amount of about 177 .mu.g/ml and, in addition, the above-mentioned HPLC analysis revealed the simultaneous production of 7-amino-3-methyl-3-cephem-4-carboxylic acid by them (a small peak was found at a position corresonding to a retention time of 10 minutes).

The following microorganisms have been deposited, since Dec. 25, 1989, with the Fermentation Research Institute, Agency of Industrial Science and Technology, Ministry of International Trade and Industry, Japan:

Escherichia coli JM109 (pCPV22P) FERM BP-2704

Escherichia coli JM109 (pHBV1) FERM BP-2703

Escherichia coli JM109 (pHDV11) FERM BP-2706

Escherichia coli JM109 (pHBD3) FERM BP-2705

Acremonium chrysogenum BC2116 FERM BP-2707