NO314267B1 - Use of Expression Cassette in Multigen Systems, Expression Cassette, Recombinant Host Cell, Method of Preparation of Single Mixture of Exogenous Proteins, Method of Selective Biochemical Oxidation, and Method of Oxidase - Google Patents

Abstract

Genetically engineered host cells containing new expression cassettes are provided which are able to carry out biochemical oxidations of steroids. In particular the oxidation is carried out with cells into which DNA has been introduced which encodes protein involved in the biological pathway of cholesterol to hydrocortisone. Suited host cells comprise species of $i(Bacillus), $i(Saccharomyces) or $i(Kluyveromyces). The new host cells are suited for microbiological oxidations of cholesterol, pregnenolone, progesterone, 17$g(a)-hydroxy-progesterone, and cortexolone, which are intermediates in said biological pathway. The new expression cassettes are also useful in the ultimate production of a multigenic system for a one-step conversion of cholesterol into hydrocortisone.

Description

The invention relates to a method for biochemical oxidation

for the manufacture of pharmaceutically usable steroids. The invention includes the use of expression cassettes in multigene systems, expression cassettes, recombinant host cells, methods for producing a mixture of exogenous proteins, methods for selective biochemical oxidation, and methods for oxidation. 11 p,17a,21-trihydroxy-4-pregnene-3,20-dione (hydrocortisone) is an important pharmaceutical steroid, which is used because its pharmacological properties as a corticosteroid and as a starting compound for the preparation of numerous useful steroids, especially other corticosteroids. Hydrocortisone is made in the adrenal cortex of vertebrates and was originally produced, only in small amounts, by a labor-intensive extraction from adrenal cortex tissue. Only after the structure had been clarified were new manufacturing methods developed, characterized by a combination of chemical synthesis and microbiological transformations. Only because the starting materials used such as sterols, bile acids and sapogenins (detergent) are numerous and cheap, the current methods allow a less expensive product, but these are still rather complicated. Several possibilities were proposed to improve the present methods, and biochemical methods have also been tried.

One attempt aimed to have a suitable starting steroid transferred in an in vitro biochemical system using isolated adrenocortical proteins known to be responsible for the in vivo enzymatic conversion of steroids to hydrocortisone. However, the difficult isolation of the proteins and the high price of the necessary cofactors seemed to be an obstacle to an economically attractive large-scale procedure. Another approach was to keep the catalytic proteins in their natural environment and have adrenal cortex cells make the desired hydrocortisone in a cell culture. But due to low production in the cells, it proved impossible in practice to make such a biochemical method economically attractive.

The in vivo process in the adrenal cortex of mammals and other vertebrates consists of a biochemical reaction pathway that starts with cholesterol and, via various intermediates, finally produces hydrocortisone (see Figure 1). Eight proteins are directly involved in this reaction pathway, five of which are enzymes, of which four are cytochrome P4so enzymes, and the other three are electron-transferring proteins.

The first step in the conversion of cholesterol to 3β-hydroxy-5-pregnene-20-ene (pregnenolone). In this transfer, a mono-oxygenase reaction, three proteins are involved: side-chain-cleaving enzyme (P45oSCC, a heme-Fe-containing protein), adrenodoxin (ADX, an FE2S2-containing protein) and adrenodoxin reductase (ADR, a FAD-containing protein ). In addition to cholesterol as a substrate, the reaction also requires molecular oxygen and NADPH. Pregnenolone is then transferred by dehydrogenation/isomerization to 4-pregnene-3,20-dione (progesterone). This reaction, catalyzed by the protein 3|3-hydroxysteroid dehydrogenase/isomerase (3|3-HSD), requires pregnenolone and

NAD+.

To obtain hydrocortisone, progesterone is then hydroxylated in three positions whose transfers are catalyzed by mono-oxygenases. In the transfer of progesterone to 17a-hydroxyprogesterone, two proteins are involved: steroid 17a-hydroxylase (P45o17a, a heme-Fe-containing protein) and NADPH cytochrome P4so-reductase (RED, an FAD- and FMN-containing protein). The reaction uses progesterone, molecular oxygen and NADPH.

For the transfer of 17α-hydroxyprogesterone to 17α,21-dihydroxy-4-pregnen-3,20-dione (cortexolone), two proteins are also necessary: steroid 21-hydroxylase (P4soC21, a heme-Fe-containing protein) and the the aforementioned protein RED. The reaction consumes 17a-hydroxyprogesterone, molecular oxygen and

NADPH.

In the transfer of cortexolone to hydrocortisone, three proteins are involved: steroid 11p-hydroxylase (P450IIP), a heme-Fe-containing protein, and the above-mentioned proteins ADX and ADR.

As described above, cytochrome P4so proteins are enzymes essential for the biochemical transfer of cholesterol to hydrocortisone. These enzymes belong to a larger group of cytochrome P450 proteins (or P4so proteins for short). They are present in prokaryotes (various bacteria) and eukaryotes (yeast, fungi, plants and animals). In mammals, high levels of P45o proteins are found in the adrenal cortex, ovary, testis and liver. Many of these proteins have been purified and are now well characterized. Their specific activity has been determined. Recently, a number of review articles have been published on this topic, such as K. Ruckpaul and H. Rein (eds), "Cytochrome P450" and P.R. Ortiz de Motellano (ed.) "Cytochrome P450, structure, mechanism and biochemistry". Cytochrome P45o proteins are characterized by their specific absorption maximum at 450 nm after reduction with carbon monoxide. In prokaryotic organisms, the P45o proteins are either membrane-bound or cytoplasmic. As far as the bacterial P45o proteins have been studied in detail (eg P45omeg and P45ocam) it has been shown that a ferredoxin and a ferredoxin reductase are involved in the hydroxylation activity. For eukaryotic organisms, two types of P45o proteins, I and II have been described. Their two differences consist in: 1. subcellular location, type I is located in the microsomal fraction and type II is located in the inner membrane of mitochondria; 2. the way the electrons are transferred from the P45o protein. Type I is reduced by NADPH via a P45o reductase, while type II is reduced by NADPH via a ferredoxin reductase (e.g. adrenodoxin reductase) and a ferredoxin (e.g. adrenodoxin).

According to EP-A-0281245, cytochrome P450 enzymes can be made from Sfreptomyces species and used for the hydroxylation of chemical compounds.

The enzymes are used in isolated form, which is a rather cumbersome and expensive method.

JP-A-62236485 (Derwent 87-331234) mentions that it is possible to introduce into Saccharomyces cerevisiae the genes for liver cytochrome P4so enzymes and to express them as enzymes which can be used for their oxidation activity.

However, in the above-mentioned references there is no indication of the use of cytochrome P45o enzymes to make steroid compounds.

The invention relates to a series of expression cassettes (units) for the production of proteins necessary in the construction of a multigene system for the one-step transfer of cheap steroid starting compounds to more rare and expensive end products, where such transfer is carried out in a natural system via a series of enzyme-catalyzed and cofactor-mediated transfers, such as the production of hydrocortisone from cholesterol. The expression cassettes of the invention are applicable to the final production in multigene systems to perform these multistep transformations.

Accordingly, in one aspect, the invention is directed to a use of an expression cassette in multigene systems for performing multistep transformations as depicted in Figure 1 wherein the expression cassette is operable in a non-mammalian recombinant host, characterized in that the expression cassette comprises a heterologous coding DNA sequence which encodes a protein that is functional, alone or together with one or more accessory proteins, in catalyzing an oxidation step in the biological reaction pathway for the conversion of cholesterol to hydrocortisone, which step is selected from the group consisting of: conversion of cholesterol to pregnenolone;

conversion of pregnenolone to progesterone;

conversion of progesterone to 17cc-hydroxyprogesterone;

conversion of 17α-hydroxyprogesterone to cortexolone; and conversion of cortexolone to hydrocortisone,

and the corresponding control sequences effective in said host.

The expression cassettes of the invention, operable in a non-mammalian

recombinant host comprising a heterologous coding DNA sequence encoding a protein, which is functional, alone or together with one or more additional proteins, in catalyzing an oxidation step in the biological reaction pathway for the conversion of cholesterol to hydrocortisone, which step is selected from the group consisting of of: conversion of cholesterol to pregnenolone;

conversion of pregnenolone to progesterone;

conversion of progesterone to 17α-hydroxyprogesterone;

conversion of 17α-hydroxyprogesterone to cortexolone; and conversion of cortexolone to hydrocortisone,

and the corresponding control sequences effective in said host, characterized in that the protein is selected from the group consisting of: side chain-cleaving enzyme (P45oSCC); adrenodoxin (ADX); adrenodoxin reductase (ADR); 3β-hydroxysteroid dehydrogenase/isomerase (3β-HSD); steroid 17α-hydroxylase (<p>45o17cc); NADPH cytochrome P45o reductase (RED); steroid 21-hydroxylase P450C21); and steroid-11 |3-hydroxylase (P4so11p), and that the heterologous DNA codes for at least one additional protein of the aforementioned group of proteins.

In other respects, the invention is directed to recombinant host cells comprising cells of microorganisms, plants or non-mammalian animals containing an expression cassette of heterologous DNA, characterized in that the expression cassette is one defined in any of claims 10-16, and to methods of producing the above-mentioned exogenous proteins (enzymes) and for using these enzymes for oxidation, and for methods for using said host cells for specific oxidations in a culture mixture.

Brief description of the figures

Abbreviations used in all figures: Ri, EcoRI: H, HindIII: Sc, Seal: P, Pstl; K, Kpnl: St, Stul; Sp, Sphl; X, Xbal: N, Ndel: S, Narrow: Ss, Sstl; Rv, EcoRV: Si, Sac1; B, BamHI: Sn, Sadl: Sal, SaM; Xh, XhoJ; Pv, Pvull: Bg, BgMI and M, Mlul. Figure 1 shows a schematic overview of the proteins involved in the subsequent steps in the transfer of cholesterol to hydrocortisone as it occurs in the adrenal cortex in mammals. Figure 2 shows the construction of plasmid pGBSCC-1. The P450SCC sequences are indicated in shaded area (\^/// A ). Figure 3 shows the insertion of a synthetically derived Pstl/HindIII fragment containing the 5'-P45oSCC sequences into the plasmid pTZ18R to obtain the plasmid pTZ "synlead" (synthetic sequence).

Figure 4 shows the construction of a full-length P450SCCCDNA

of synthetic ) and by cDNA cloning from P45oSCC<a>sequences in pTZ18R to obtain pGBSCC-2.

Figure 5 shows the complete nucleotide sequence of plasmid pBHA-1.

Figure 6 is a schematic representation of the construction of pGBSCC-3. P45oSCCcDNA sequences from plasmid pGBSCC-2 were introduced into Bacillus/ E. coli shuttle (common) plasmid pBHA-1. Filled fields are as indicated in the text of Figure 4. Figure 7 shows the introduction of an Ndel restriction site in combination with an ATG start codon before the P45oSCC maturation site in pGBSCC-3 to obtain pGBSCC-4. Figure 8 shows a physical map of pGBSCC-5 obtained by removing E coji sequences from the plasmid pGBSCC-4. Figure 9 shows a Westem blot analyzed with antibodies against P450SCC, and shows P450SCC expression of plasmid pGBSCC-5 introduced into B. subtilis (lane c) and B. licheniformis (lane f). Control extracts from B. subtilis and B. licheniformis are shown respectively. in lanes a and d. For comparison, purified adrenal cortex P450SCC (30 ng) was also added to these control extracts (lanes b and e, respectively). Figure 10 is a schematic representation of the construction of pGBSCC-17. The coding P450SCC DNA sequences from plasmid pGBSCC-4 were introduced into the E. coli expression vector pTZ18RN. The P45oSCC sequences are indicated in a box

Figure 11 shows P4SoSCC production of pGBSCC-17 in E. coli JM101.

(a) SDS/PAGE and Coomassie brilliant blue staining of the cellular protein fractions (20 µl) prepared from the E. coli control species (lane 3) and E. coli transformed cells SCC-301 and 302 (lanes 1 and 2 respectively ). 400 ng purified bovine P45oSCC (lane 4) is shown for comparison. (b) Western blot analysis probed with antibodies against P45oSCC in cellular protein fractions (5 (µl) prepared from the control strain E. coli JM101 (lane 2) and from E. coli transformants SCC-301 (lane 3) and SCC-302 ( lane 4).100 ng of purified bovine P450SCC (lane 1) is shown for comparison.

Figure 12 shows the construction of plasmid pUCG418.

Figure 13 shows the construction of the yeast expression vector pGB950 by inserting the promoter and terminator with multiple cloning sites (pm) from lactase into pUCG418. To obtain pGBSCC-6, a synthetic SalI/XhoI fragment containing an ATG start codon and the codons for the first eight amino acids of P450SCC is inserted into pGB950. Figure 14 is a schematic representation showing the construction of the yeast P450SCC expression cassette pGBSCC-7. Figure 15 shows a Western blot analyzed with antibodies specific for the protein P450SCC.

Blot A contains extracts obtained from Saccharomyces cerevisiae 237-10B transformed with pGBSCC-10 (lane 1); from S. cerevisiae 273-1 OB as control (lane 2); from Kluvveromvces lactis CBS 2360 transformed with pGBSCC-7 (lane 3) and from K. lactis CBS 2360 as a control (lane 4).

Blot B contains extracts obtained from K. lactis CBS 2360 as a control (lane 1) and K. lactis CBS 2360 transformed with pGBSCC-15 (lane 2), with pGBSCC-12 (lane 3) or with pGBSCC-7 (lane 4) .

Blot C contains extracts obtained from S. cerevisiae 273-1 OB as a control (lane 1), transformed with pGBSCC-16 (lane 2) or with pGBSCC-13 (lane 3). Figure 16 is a schematic representation of the construct from the gjasr expression vector pGBSCC-9 containing the isocytochrome C1 (cyc-1) promoter from S. cerevisiae.

Figure 17 shows a construction diagram of P45oSCCcDNA which

contains the expression vector pGBSCC-10 for S. cerevisiae.

Figure 18 shows the construction of the P45oSCC expression vector pGBSCC-12 in which a synthetically produced DNA fragment encoding the pre-P45oSCC sequence ( tøfiM ) is inserted 5' of the coding sequence of mature P450SCC. Figure 19 shows the construction of pGBSCC-13. This P450SCC expression cassette for S. cerevisiae contains the pre-P4soSCCcDNA sequence located 3' of the cyc-1 promoter of S. cerevisiae. Figure 20 shows a schematic representation of the construction of the plasmids pGBSCC-14 and pGBSCC-15. The latter contains the P450SCC coding sequence in reading frame with the cytochrome oxidase VI pre-sequence ( \/////%\. Figure 21 shows the construction of the plasmid pGBSCC-16. In this plasmid, the cytochrome oxidase VI pre-sequence (\$ y // fo av S. cerevisiae bound to the P450SCC coding sequence located 3' of the cyc-1 promoter Figure 22 shows the physical maps of plasmids pGB17a-1 (A) and pGB17v-2

(B) which contains respectively the 3' 1.4 kb fragment and the 5' 345 bp fragment \ W ////// A to P45o17acDNA. In pGB17a-3 (C) containing the full length of

P45o17acDNA sequence, the position of the ATG start codon is indicated.

Figure 23 shows the mutation of pGB17a-3 by in vitro mutagenesis. The resulting plasmid pGB17a-4 contains a SalI restriction site followed by optimal yeast translation signals just upstream of the ATG start codon. Figure 24 is a schematic representation of the construction of the yeast P45017a expression cassette pGB17a-5. Figure 25 shows the mutation of pGB17a-3 by in vitro mutagenesis. The resulting plasmid pGB17a-6 contains an Ndel restriction site at the ATG start codon. Figure 26 is a schematic representation of the construction of pGB17a-7. P45o17acDNA sequences from plasmid pGB17ot-6 were introduced into Bacillus/ E. coli shuttle (common) plasmid pBHA-1. Figure 27 shows a physical map of pGB17a-8 obtained by removing E. coli sequences from the plasmid pGB17cc-7. Figure 28 shows physical maps of pGBC21-1 and 2, which respectively contain 1.53 Kb 3'-P450C21cDNA and a 540 bp 5'-P45oC21cDNA EcoRI fragment, in the EcoRI site of the cloning vector pTZ18R. Figure 29 shows the in vitro polymerase chain reaction (PCR) mutagenesis in pGBC21-2 to introduce EcoRV and Ndel restriction sites upstream of the P450C21 ATG start codon, followed by molecular cloning in the cloning vector pSP73 to obtain pGBC21-3. Figure 30 is a schematic representation of the construction of pGBC21-4, which contains the full-length P450C21 cDNA sequence. Figure 31 is a schematic representation of the construct pGBC21-5. The p45oC21 cDNA sequence from plasmid pGBC21-4 was inserted into Bacillus/ E. coli shuttle (common) plasmid pBHA-1. Figure 32 shows a physical map of pGBC21-6 obtained by removing E. coli sequences from the plasmid pGBC21-5. Figure 33 shows the mutation of pGBC21-2 by in vitro mutagenesis. The resulting plasmid pGBC21-7 contains a SalI restriction site followed by optimal yeast translation signals just upstream of the ATG start codon. Figure 34 shows the construction of pGBC21-8, which contains a full-length P45oC21 cDNA with modified flanking restriction sites suitable for cloning into the yeast expression vector. Figure 35 is a schematic representation showing the construction of the yeast P45oC21 expression cassette pGBC21 -9. Figure 36 shows the in vitro polymerase chain reaction mutagenesis in pGB11p-1 to introduce appropriate flanking restriction sites and an ATG start codon to the full-length P45o1ipcDNA sequence, followed by molecular cloning into the Bacillus/ E. coli shuttle vector pBHA- 1 to obtain the plasmid pGB1ip-2. Figure 37 shows the in vitro polymerase chain reaction mutagenesis in pGB11p-1 to introduce appropriate flanking restriction sites and an ATG start codon to the full-length P45o11pcDNA sequence, followed by molecular cloning into the yeast expression vector pGB950 to obtain plasmid pGB11p-4. Figure 38 is a schematic representation of the molecular cloning of

The ADXcDNA sequence from a bovine adrenal polyA<+>RNA/cDNA mixture by the polymerase chain reaction method. The cDNA sequence encoding the mature ADX protein was inserted into the appropriate sites of the yeast expression vector pGB950 to obtain the plasmid pGBADX-1.

Figure 39 shows a Western blot analyzed with antibodies against ADX, and shows

ADX production in plasmid pGBADX-1 in K. lactis CBS 2360 transformants ADX-101 and 102 (lanes 4 and 5, respectively). Extract of controlled K. lactis CBS 2360 is shown in lane 3. For comparison, purified adrenal cortex ADX (100 ng) is also applied to the gel in lane 1.

Figure 40 shows the in vitro polymerase chain reaction mutagenesis of pGBADR-1 to introduce appropriate flanking restriction sites and an ATG start codon to the full-length ADRcDNA sequence, followed by molecular cloning into the yeast expression vector pGB950 to obtain pGBADR-2 .

The invention relates to the production and cultivation of cells suitable for use in large-scale biochemical production reactors and the use of these cells for the oxidation of compounds and in particular for the production of steroids, shown in Figure 1. An interchange of steps in a multi-step reaction is included in the invention. Microorganisms are preferred hosts, but other cells may also be used, such as cells from plants or animals, optionally used in a cell culture or in the tissues of living transgenic plants or animals. The cells according to this invention are obtained by means of genetic transformation (transformation) of suitable receptor cells, preferably cells from suitable microorganisms, with vectors containing DNA sequences encoding the proteins that transfer cholesterol to hydrocortisone, including side chain cleaving enzyme (P450SCC), adrenodoxin ( ADX), adreno-doxin reductase (ADR), 3p-hydroxysteroid dehydrogenase/- isomerase (3p-HSD), steroid-17a-hydroxylase (P45o17a), NADPH cytochrome P450-reductase (RED), steroid-21-hydroxylase (P45oC21) and steroid-11 p-hydroxylase (P45011P)- Certain host cells can already produce on their own one or more of the necessary proteins in sufficient quantity and therefore need to be transformed only with the necessary complementary DNA sequences. Such possible own proteins are ferredoxin, ferredoxin reductase, P45o-reductase, and 3p-hydroxy-steroid dehydrogenase/isomerase.

In order to recover the sequences coding for proteins responsible for the transfer of cholesterol to hydrocortisone, appropriate DNA sources have been selected. A suitable source for obtaining DNA encoding all the proteins that transfer cholesterol to hydrocortisone is adrenal cortical tissue from vertebrates e.g. bovine adrenal cortex tissue. The relevant DNA can also be obtained from different microorganisms, e.g. from Pseudomonas testosteroni, Streptomyces qriseocarneus or Brevibacterium sterolicum for DNA encoding 3B-hydroxysteroid dehydrogenase/isomerase and from Curvularia lunata or Cunnin<g>hamella blakesleeana for DNA encoding proteins that carry out 11 (5-hydroxylation of cortexolon. The DNA sequences encoding for the proteins bovine P450SCC, bovine P45011P or a microbiologically equivalent protein, bovine adrenodoxin, bovine adrenodoxin reductase, 3phydroxysteroid dehydrogenase/isomerase of bovine or microbiological origin, bovine P45o17a, bovine P450C2I and NADPH cytochrome P4so reductase of bovine or microbiological origin, were isolated according to the following steps:

1. Eukaryotic sequences fcDNA)

a. Total RNA was prepared from appropriate tissues.

b. PolyA<+->containing RNA was transcribed into double-stranded cDNA and

ligated into bacteriophage vectors.

c. The obtained cDNA library was screened by <32>P-labeled oligomers specific for the desired cDNA or by screening an isopropyl-β-D-thiogalactopyranoside (IPTG)-induced Iambda-gt11 cDNA library using a specific (12 <5>l-labelled) antibody. d. cDNA inserts in positive plaque-forming units (pfu's) were inserted into appropriate vectors to determine:

- the full length of the cDNA by nucleotide sequencing.

2. Prokaryotes genes a. Genomic DNA was made from a suitable microorganism.

b. To obtain a DNA library, DNA fragments were cloned into

appropriate vectors and transformed into an appropriate E. coli host.

c. The DNA library was probed with <32>P-labeled oligomers specific for the gene of interest or by probing an IPTG-induced Iambda-gt11 DNA library using a specific (<1><ø>l-labeled) antibody. d. Plasmids from positive colonies were isolated and inserted DNA fragments subcloned into appropriate vectors to determine:

- the complete length of the gene.

Note: According to an improved method, the particular cDNA (eukaryotic sequences) or gene (prokaryotic sequences) using two specific oligomers by the method known as the polymerase chain reaction (PCR) (Saiki et al., Science, 239, 487-491 , 1988). The prepared cDNA or DNA was then inserted into the appropriate vectors.

According to one part of the invention, suitable expression cassettes were made in which the heterologous DNA isolated according to the previous method and encoding an enzyme catalyzing an oxidation step of the reaction pathway in Figure 1 and at least one additional protein in said reaction pathway, were placed between suitable control sequences for transcription and translation, which enables the DNA to be expressed in the cellular environment of a suitable host, allowing the formation of the desired protein or proteins. If desired, the initiation control sequences may be followed by a secretion signal sequence.

Appropriate control sequences have been introduced together with the structural DNA by means of said expression cassettes. Production is enabled by transformation of an appropriate host cell with a vector containing control sequences compatible with the relevant host and operably linked to the coding sequences whose expression is desired.

Alternatively, appropriate control sequences present in the host genome are used. Production is made possible by transformation of a suitable host cell with a vector containing coding sequences for the desired protein flanked by host sequences that enable homologous recombination with the host genome in such a way that the host control sequences properly control the expression of the introduced DNA.

As a general understanding, the term control sequences is characterized by all DNA segments that are necessary for the correct regulation of the expression of the coding sequence to which it is operatively linked, such as operons (operators), enhancers (amplifiers) and, in particular, promoters and sequences which controls the translation.

The promoter may or may not be controlled via regulation of its environment. Suitable promoters for rokaryotes include e.g. the trp promoter (stimulable by tryptophan deficiency), the lac promoter (stimulable by the galactose analogue IPTG), the B-lactamase promoter, and the phage-derived P|_ promoter (stimulable by temperature variation). In addition, especially for expression in Bacillus. useful promoters include those for alpha-amylase, protease, Spo2, spac and 0105 and synthetic promoter sequences. A preferred promoter is that depicted in Figure 5 and designated "Hpall". Suitable promoters for expression in yeast include the 3-phospho-glycerate kinase promoter and those for other glycolytic enzymes, as well as promoters for alcohol dehydrogenase and yeast phosphatase. Also suitable are the transcription elongation factor (TEF) and lactase promoters. Mammalian expression systems generally use promoters derived from viruses such as adenovirus promoters and the SV40 promoter, but they also include regulatable promoters such as the metallothionein promoter, which is controlled by heavy metal or glucocorticoid concentration. Current virus-based insect cell expression systems are also suitable, as well as expression systems based on plant cell promoters such as nopaline synthetase promoters.

Translational control sequences include a ribosome binding site (RBS) in prokaryotic systems, while in eukaryotic systems translation may be controlled by a nucleotide sequence containing a start codon such as AUG.

In addition to the required promoter and translational control sequences, a variety of other control sequences, including those that regulate termination, may

(eg, resulting in polyadenylation sequences in eukaryotic systems)

be used to control expression. Certain systems contain enhancer elements which are desirable, but usually not mandatory, to initiate expression.

A group of vectors designated pGBSCC-n, where "n" is any number from 1 to 17, has been specifically developed for DNA encoding the P4soSCC enzyme.

Another group of vectors designated pGB17a-n, where "n" is any number from 1 to 5, is specifically designed for DNA encoding the P45017<x enzyme.

A further group of vectors designated pGBC21-n, where "n" is any number from 1 to 9, has been specifically designed for DNA encoding the P450C2I enzyme.

Yet another group of vectors designated pGB1 ip-n, where "n" is any number from 1 to 4, is specifically designed for DNA encoding the P45o11 p-enzyme.

According to a further aspect of the invention, suitable host cells have been selected which accept vectors described in the invention and allow the inserted DNA to be expressed. When the transformed host cells are cultured, the proteins responsible for the transfer of cholesterol to hydrocortisone will appear inside the cell. The presence of the desired DNA can be detected by DNA hybridization methods, their transcription by RNA hybridization, their expression by immunological methods and their activity by measuring the presence of oxidized products after incubation with the starting compound in vitro or in vivo.

Transformed microorganisms are preferred host cells, particularly bacteria (Escerichia coli and Bacillus and Stre<p>tomvces species are more preferred) and yeast (such as Saccharomvces and Kluvveromvces). Other suitable host organisms have been found among plants and animals, including insects, where the isolated cells have been used in cell culture, such as COS cells, Ci27 cells, CHO cells, and Spodoptera frugiperda (Sfg) cells. Alternatively, a transgenic plant or animal can be used

A special type of recombinant host cells are those in which either two or more expression cassettes according to the invention have been introduced or which have been transformed by an expression cassette that codes for at least two heterologous proteins, which enables the cell to make at least two proteins that are part of the reaction pathway in figure 1.

A main feature of the invention is that the created new cells are not only able to produce the proteins responsible for the oxidative transfer of steroids which ultimately results in hydrocortisone, but also to use these proteins on the spot in the desired oxidative transformation of the corresponding substrate compound added to the culture fluid. Steroids are preferred substrates. The cells transformed with the heterologous DNA are particularly well suited to be cultured with the steroids mentioned in figure 1, including other sterols such as p-sitosterol. As a result, oxidized steroids are made.

Depending on the presence in the host cell of a number of heterologous DNAs that code for proteins participating in the reaction pathway in Figure 1, the result is several biochemical transformations characterized by side chain cleavage of a sterol and/or oxidative modifications on C11, C17, C3 and C21. Therefore, the expression cassettes according to the invention can be used to construct a multigene system that can carry out a series of intracellular transformations characterized by the many steps in the sequence depicted in figure 1. It may be necessary to introduce into the desired host, expression cassettes that code in their entirety for the necessary proteins. In certain cases, one or more of the proteins in the pathway may already be present in the host as a native protein that exhibits the same activity. For example, ferredoxin, ferredoxin reductase and P45o reductase may already be present in the host. Under these conditions, only the remaining enzymes must be transferred by recombinant transformation.

As an alternative to biochemical transfers in vivo, the proteins responsible for the transfer of cholesterol to hydrocortisone are collected, purified as far as necessary, and used for in vitro transfer of steroids in a cell-free system, e.g. immobilized on a column. Alternatively, the more or less purified mixture containing one or more enzymes in the reaction pathway is used as such for steroid conversion. For example a host containing DNA encoding two heterologous proteins namely the enzyme P450SCC and the protein ADX necessary for the production of pregnenolone. Compared to a host that has only P450SCC DNA, the yield of pregnenolone in a cell-free extract after addition of ADR, NADPH and cholesterol is significantly improved.

The present invention provides expression cassettes necessary for the construction of a one-step production process for several useful steroids. By starting from cheap and abundantly available starting compounds, it is particularly suitable for the production of hydrocortisone and intermediate compounds. The invention makes traditional expensive chemical reactions old-fashioned. Intermediate compounds do not need to be isolated. Apart from new host cells, the methods themselves used to grow these cells from steroid conversions are analogous to the biotechnological methods well known in the field.

The invention is further illustrated by the following examples.

Example 1

Molecular cloning of a full-length cDNA encoding the bovine cytochrome P450 side-chain cleaving enzyme (P450SCC)

General cloning techniques as well as DNA and RNA analyzes were used as described in the handbook of T. Maniatis et al., Molecular Cloning, Cold Spring Harbor Laboratory, 1982. Unless otherwise stated, all DNA-modifying enzymes, molecular cloning tools and E. coli strains obtained from commercial sources and used according to the manufacturers' instructions. Materials and apparatus for DNA and RNA separation and purification were used according to the manufacturers' instructions.

Bovine adrenocortical tissue was prepared from freshly excised bovine (bovine) kidneys, which were snap-frozen in liquid nitrogen and stored at -80°C.

From frozen bovine adrenal cortices, total cellular RNA was prepared as described by Auffrey and Rougeon (Eur. J. Biochem., 107, 303-314, 1980). Adrenal poly A+ RNA was obtained by heating the total RNA sample to 65°C prior to the polyA selection on oligo(dT) chromatography.

DNAs complementary to polyA<+->RNA from bovine adrenal cortex were prepared as follows: 10 µg of polyA<+->RNA, treated with methylmercury hydroxide was neutralized with beta-mercaptoethanol. This mixture was adjusted to 50 mM Tris/HCl (pH 8.3 at 42°C), 40 mM KCl, 6 mM MgCl 2 , 10 mM DTT 3000 U RNasin/ml, 4 mM Na4P2C>7, 50 µg actinomycin D/ ml, 0.1 mg oligo(dT12-i8)/ml,

0.5 mM dGTP, 0.5 mM dATP, 0.5 mM dTTP, 0.25 mM dCTP and 400 nCi alpha <32>P-dCTP/ml, all in a final volume of 100 µl. The mixture was placed on ice for 10 min., heated for 2 min. at 42°C and the synthesis was started by the addition of 150 U of AMV reverse transcriptase (Anglian Biotechnology Ltd.); the incubation was carried out for 1 hour at 42°C.

Second chain synthesis was performed by adding DNA polymerase and RNase H according to Gubler and Hoffman (Gene, 25, 263-269, 1983). After treatment of ds DNA with T4 DNA polymerase (BRL) to obtain equal ends, decameric EcoR linkers (helper sequences) (Biolabs Inc.) were ligated to the ds DNA fragments. After digestion with EcoRI (Boehringer), double-stranded cDNA fragments were separated from excess EcoRI linkers by Biogel A15 m (BioRad) chromatography. Approximately 200 ng of EcoRI linker-containing double-stranded cDNA was ligated with 10 µg of EooRI digested and calf intestinal phosphatase (Boehringer) treated with Iambda-gt11 vector DNA (Promega) by T4-DNA ligase (Boehringer) as described by Huynh

et al. (In: "DNA donating techniques: A practical approach", pp. 49-78, Oxford IRL-press, 1985). Phage obtained after in vitro packaging of the ligation mixture were used to infect the E. coli Y1090 host (Promega).

From this cDNA library, approximately 10<6> plaque-forming units (pfu) were probed with a <32>P-end-labeled synthetic oligomer SCC-1 (5'-GGC TGA CGA AGT CCT GAG ACA CTG GAT TCA GCA CTGG-3 '), specific for bovine P450SCC DNA sequences as described by Morohashi et al. (Proe. Nati. Acad. Sei. USA, 81, 4647-4651, 1984. Six hybridizing pfu's were obtained and further purified by two additional rounds of infection, smearing, and hybridization. P45oSCCcDNA EcoRI inserts were subcloned into the EcoRI site of pTZ18R (Pharmacia) Clone pGBSCC-1 (Figure 2), which contained the largest EcoRI insert (1.4 kb), obtained from clone Iambda-gt11 SCC-54 was further analyzed by restriction enzyme mapping and sequencing.

Sequence data showed that the pGBSCC-1 EcoRI insert was identical to the nucleotide sequence of the SCCcDNA between positions 251 and 1824 of the P45oSCCcDNA map as described by Morohashi et al.

The remaining 5'-P45oSCCcDNA nucleotides were synthetically made by cloning a 177 bp Pst/Hindi I1 fragment into the correct sites of pTZ18R, resulting in pTZTsynlead" as shown in Figure 3, which contained in addition to the nucleotides that encodes the mature P45oSCC protein from position 118 to 273 as published by Morohashi et al., restriction sites for Seal, Avrll and Stu I without displaying the putative amino acid sequence of the P45oSCC protein.

Full-length P450SCCcDNA was constructed by cloning into E. coli JM101 (ATCC 33876) a ligation mixture containing the 1372 bp HindIII/Kpn1 pGBSCC-1 fragment, in which the 177 bp Pst/H indl 11 pTZ/"svnlead" fragment and pTZ19R DNA was digested with Pst and KpnI.

The resulting plasmid pGBSCC-2, which contained all the nucleotide sequences encoding the mature bovine P45o side chain cleavage protein is shown in Figure 4.

Example 2

Construction, transformation and expression of P450SCC in the bacterial host Bacillus subtilis

To achieve expression of cytochrome P450SCC in Bacillus host, P45oSCCcDNA sequences were transferred into an E. coli/ Bacillus common vector pBHA-1.

Figure 5 shows the nucleotide sequence of the common plasmid pBHA-1. The plasmid consists of positions 11-105 and 121-215: bacteriophage FD terminator (double); positions 221-307: part of plasmid pBR322 (that is positions 2069-2153); positions 313-768: bacteriophage Fl, origin of replication (these are positions 5482-5943); positions 772-2571: part of plasmid pBR322, the origin of replication and the B-lactamase gene; positions 2572-2685: transposon TN903, complete genome; positions 2719-2772: tryptophan terminator (double); positions 2773-3729: transposon Tn9, the chloramphenicol acetyltransferase gene. The nucleotides at positions 3005 (A), 3038 (C), 3302 (A) and 3409 (A) differ from the wild-type cat coding sequence. These mutations were introduced to eliminate Ncol. Ball. The EcoRI and Pvull sites: positions 3730-3804: multiple cloning site; positions 3807-7264: part of plasmid pUB110 containing the Bacillus "HpaH" promoter. replication function and kanamycin resistance gene (EcoRI-PyuM fragment)

(McKenzie et al., Plasmid 15, 93-103, 1986 and McKenzie et al., Plasmid 17, 83-85, 1987); positions 7267-7331: multiple cloning site. The fragments were assembled using known cloning techniques, e.g. filling in overhangs with Klenow, adapter cloning, etc. All results were obtained from Genbank<R>, National Nucleic Acid Sequence Data Bank, NIH, USA.

pGBSCC-3 was obtained by molecular cloning in E. coli JM101 of the Kpnl/Sph P450SCCCDNA insert of pGBSCC-2 (described in Example 1) into the appropriate sites in pBHA-1 as indicated in Figure 6.

By means of cloning in E. coli JM101, the methionine start codon was inserted by exchanging the Stul/SphI fragment in pGBSCC-3 with a synthetically engineered Sphl/SphI fragment.

which contains an Ndel site at the ATG start codon. The obtained plasmid pGBSCC-4 is shown in Figure 7. The "Hpa II" Bacillus promoter was inserted upstream of the P45oSCCcDNA sequences by cleaving pGBSCC-4 with the restriction enzyme Ndel. separation of the E. coli portion of the common (shuttle) plasmid by agarose gel electrophoresis and subsequent ligation and transformation into Bacillus subtilis 1A40 (BGSC 1A40) competent cells. Neomycin resistant colonies were analyzed and the plasmid pGBSCC-5 (Figure 8) was isolated. Expression of bovine P450SCC was studied by preparing a cellular protein fraction from an overnight culture at 37°C

in TSB medium (Gibco) containing 10 µg/ml neomycin. Cells from 100 µl culture, containing approximately 5.10<6> cells, were harvested by centrifugation and resuspended in 10 mM Tris/HCl pH 7.5. Lysis was performed by adding lysozyme (1 mg/ml) and

incubation for 15 min. at 37°C. After treatment with 0.2 mg DNase/ml within 5 min. at 37°C, the mixture was adjusted to 1x SB buffer, as described by Laemmli, Nature 227, 680-685, 1970, in a final volume of 200 µl. After heating for 5 min. to 100°C, 15 µl of the mixture was run on a 7.5% SDS/polyacrylamide gel electrophoresis. As shown in Figure 9 (lane c), a 53 kDa band could be detected after immunoblotting the gel analyzed with P4soSCC-specific antibodies.

Specific bovine P450SCC antibodies were obtained by immunizing rabbits with purified P450SCC protein isolated from bovine adrenal cortex tissue.

Example 3

Expression of P450SCC in the bacterial host Bacillus licheniformis

Expression of bovine P4soSCC in B. licheniformis was performed by transformation of plasmid pGBSCC-5 into the appropriate host strain B. licheniformis T5(CBS 470.83). A cellular protein fraction prepared as described in Example 2, from an overnight culture at 37°C in Trypton Soy Broth (TSB) medium (Oxoid) containing 10 µg/ml neomycin, was analyzed by SDS/PAGE and Western blotting. As shown in figure 9 (track f), a 53 kDa protein band was visualized after incubation of the nitrocellulose filter with antibodies specific for bovine P450SCC. A transformant was further analyzed in vivo with regard to activity of P450SCC (see example 1).

Example 4

Expression of P450SCC in the bacterial host Escherichia coli

(a) Construction of the expression cassette.

To create a suitable expression vector in the E. coli host for bovine P450SCC, pTZ18R was mutated by site-specific mutagenesis as described by Zoller and Smith (Methods in Enzymology 100, 468-500, 1983); Zoller and Smith (Methods in Enzymology 154, 329-350, 1987) and Kramer and Fritz (Methods in Enzymology 154, 350-367, 1987). Plasmids and strains for the in vitro mutagenesis experiments were obtained from Pharmacia Inc.

A synthetic oligomer with the sequence:

was used to create an Ndel restriction site at the ATG start codon of lac Z-genetiTZ18R.

The resulting plasmid pTZ18RN was digested with Ndel and K<p>nl and the Ndel/ Kpnl DNA fragment from pGBSCC-4, containing full-length SCCcDNA, was inserted by molecular cloning as indicated in Figure 10.

The transcription of P45oSCCcDNA sequences in the resulting plasmid pGBSCC-17 will be driven by the E. coli lac promoter.

(b) Expression of P450SCC in the host E. coli JM101

pGBSCC-17 was introduced into E. coli JM101 competent cells by selecting for ampicillin-resistant colonies. Expression of cytochrome P450SCC was studied by preparing a cellular protein fraction (described in Example 2) from transformants SCC-301 and 302 from an overnight culture at 37°C in 2xTY medium (containing

per liter of de-ionized water: Bacto tryptone (Difco), 16 g; yeast extract (Difco), 10 g and NaCl, 5 g) which contained 50 µg/ml ampicillin.

Protein fractions were analyzed by SDS/PAGE stained with Coomassie brilliant blue (Figure 11A) or by Western blot and analyzed with antibodies specific for bovine P450SCC (Figure 11B). Both analyzes show a protein of the expected length (figure 11A, tracks 1 and 2 and in figure 11B, tracks 3 and 4) for respectively transformants SCC-301 and SCC-302, which are absent in the E. coli JM101 control strain (Figure 11A, lane 3 and Figure 11B, lane 2).

Example 5

Construction, transformation and expression of P450SCC i

the yeast Kluweromvces lactis

(a) Introduction of the geneticin resistance marker into pUC19

A DNA fragment comprising the Tn5 gene (Reiss et al, EMBO J., 3, 3317-3322, 1984) and confers resistance to geneticin under the control of the alcohol dehydrogenase I (ADHI) promoter from S. cerevisiae. similar to that described by Bennetzen and Hall (J. Biol. Chem., 257, 3018-3025, 1982) was inserted into the Smal site of pUC19 (Yanisch-Perron et al., Gene, 33, 103-119, 1985). The resulting plasmid pUCG418 is shown in Figure 12.

E. coli containing pUCG418 was deposited at the Centraal Bureau voor Schimmelcultures under CBS 872.87.

(b) Construction of the expression cassette

A vector was constructed, which consisted of pUCG418 (for description see Example 5(a)) cut with XbaI and HindIII. The XbaI-SalI fragment from pGB901 which contained the lactase promoter (see J. A. van den Berg et al., Continuation-in-part of US Patent Application No. 572,414: Kluyveromyves as a host strain) and synthetic DNA comprising parts of the 3' non-coding region of the lactase gene from K. lactis. This plasmid, pGB950, is shown in Figure 13. pGB950 was cut with SaJi and XhoI and synthetic DNA was inserted:

resulting in plasmid pGBSCC-6 as shown in Figure 13.

The Stul-EcoRI fragment from pGBSCC-2 (see Example 1) containing the P450SCC coding region was isolated and the overhang filled in, using Klenow DNA polymerase. This fragment was inserted into pGBSCC-6 cut with Stu I. The plasmid containing the fragment in the correct orientation was called pGBSCC-7 (see Figure 14).

(c) Transformation of K. lactis

K. tactis strain CBS 2360 was grown in 100 ml of YEPD medium (1% yeast extract, 2% peptone, 2% glucose monohydrate) containing 2.5 ml of a 6.7%

(w/w) dinitrogen base (Difco laboratories) solution to an ODs-io of about 7. From 10 ml of the culture, the cells were collected by centrifugation, washed with TE buffer (10 mM Tris-HCl pH 7.5; 0, 1 mM EDTA) and resuspended in 1 ml TE buffer. An equal volume of 0.2 M lithium acetate was added and the mixture incubated for 1 hour at 30°C in a shaking water bath. 15 µg pGBSCC-7 was cut at the rare SacII site in the lactase promoter, ethanol precipitated and resuspended in 15 µl TE buffer. This DNA preparation was added to 100 µl of preincubated cells, and the incubation was prolonged for 30 min. Then an equal volume of 70% PEG4000 was added, and the mixture incubated for 1 hour at the same temperature, followed by a heat shock of 5 min. at 42°C. Then 1 ml of YEPD medium was added and the cells were incubated for 1.5 hours in a shaking water bath at 30°C. Finally, the cells were collected by centrifugation, resuspended in 300 µl YEPD and plated on agar plates containing 15 ml of YEPD agar with 300 µg/ml geneticin and were overlaid 1 hour before use with 15 ml EPD agar without G418. Colonies were grown for 3 days at 30°C.

(d) Analysis of the transformants

Transformants and the control strain CBS 2360 were grown in YEPD medium for about 64 hours at 30°C. The cells were collected by centrifugation, resuspended in physiological saline at an OD610 of 300 and disrupted by shaking with glass beads for 3 min. on a Vortex shaker at maximum speed. Cell debris was removed by centrifugation for 10 min. at 4500 rpm in a Hearaeus Christ minifuge GL. From the supernatants, 40 µl samples were taken for analysis on immunoblot (see Figure 15A, lane 3 and Figure 15B, lane 4).

The results show that a protein of the expected length is expressed in K. lactis cells transformed with pGBSCC-7. The transformant was named K. lactis SCC-101.

Example 6

Construction, transformation and expression of P450SCC i

the yeast Saccharomyces cerevisiae

(a) Construction of the expression cassette

To remove the lactase promoter, pGB950 (see Example 4(b)) was cut with XbaI and SalI, the overhangs were filled in using Klenow DNA polymerase and then ligated (joined). In the resulting plasmid, pGBSCC-8, the XbaI site is destroyed but the SalI site is retained.

The SalI fragment from pGB161 (see J.A. van den Berg et al., EP 96430) containing the isocytochrome C1 (cyc 1) promoter from S. cerevisiae was isolated and partially digested with XhoI. The 670 bp XhoI-SalI fragment was isolated and cloned into the SalI site of pGBSCC-8. In the selected plasmid, pGBSCC-9, the SalI site between the cyc 1 promoter and the 3' non-coding region of the lactase gene is retained (Figure 16) (HindIII partially digested).

The SalI-HindIII fragment from pGBSCC-7 containing the P450SCC coding region was inserted into pGBSCC-9 cut with SalI and HindIII. In the resulting plasmid, pGBSCC-10, the P45oSCC coding region is downstream of the cyc 1 promoter (Figure 17).

(b) Transformation of S. cerevisiae

S. cerevisiae strain D273-10B (ATCC 25657) was grown in 100 ml YEPD overnight at 30°C, then diluted (1:10000) in fresh medium and grown to an ODe-io of 6. The cells from 10 ml of the culture was collected by centrifugation and suspended in 5 ml of TE buffer. Again, the cells were collected by centrifugation, suspended in 1 ml of the TE buffer and 1 ml of 0.2 M lithium acetate added. The cells were incubated for 1 hour in a shaking water bath at 30°C. 15 µg pGBSCC-10 was cut at the rare MluI site in the cyc 1 promoter, ethanol precipitated and resuspended in 15 µl TE. This DNA preparation was added to 100 µl of pre-incubated yeast cells and incubated (shaking) for 30 min. at 30°C. After addition of 115 uJ of 70% PEG4000 solution, the incubation was continued for 60 min., without shaking. Then a heat shock of 5 min. at 42°C given to the cells, 1 ml of YEPD medium was added, followed by 1 1/2 hour incubation at 30°C in a shaking water bath. Finally, the cells were collected by centrifugation, resuspended in 300 µl YEPD and spread on YEPD agar plates containing geneticin (300 µg/ml). Colonies were grown for 3 days at 30°C.

(c) Analysis of the transformants

Transformants and the control strain were grown in YEPL medium (1% yeast extract, 2% bactopeptone, 3.48% K2HP04 and 2.2% of a 90% L-(+)-lactic acid solution; before sterilization, the pH was adjusted to 6, 0 using 25% ammonia solution) for 64 hours at 30°C. Further analysis was carried out as described in Example 5(d).

The immunoblot analysis shows expression of P450SCC in S. cerevisiae (Figure 15A, lane 1).

Example 7

Construction, transformation and expression of pre-P45oSSC coding DNA in the yeast Klyveromyces lactis

(a) Construction of the expression cassette

Plasmid pGB950 (see Example 5(b)) was cut with SalI and XhoI and synthetic DNA was inserted:

resulting in plasmid pGBSCC-11 (Figure 18). Analogous to that described in Example 5(b), the P45oSCC coding region of pGBSCC-2 inserted into pGBSCC-11 was cut with Stul. The plasmid containing the fragment in the correct orientation was named pGBSCC<a>12 (Figure 18).

(b) Transformation of K. lactis and analysis of the transformants

Transformation of K. lactis with pGBSCC-12 was performed as described in Example 5(c). The transformants were analyzed as described in Example 5(d). The analysis shows production of P450SCC by K. lactis (Figure 15B, lane 3).

Example 8

Construction, transformation and expression of pre-<p>45oSCC-encoding DNA in the yeast Saccharomvces cerevisiae

(a) Construction of the expression cassette

SalI-HindIII (HindIII partially digested) fragment from pGBSCC-12, containing the pre-P4soSCC coding region was inserted into pGBSCC-9 cut with SalI and HindIII. The resulting plasmid was named pGBSCC-13 (Figure 19).

(b) Transformation of S. cerevisiae and analysis of the trans-formants

S. cerevisiae strain D273-10B was transformed with pGBSCC-13 as described in Example 6(b). The transformants were anlayed as described in example 5(c). The results shown in Figure 15C (lane 3) show the expression of P450SCC by S. cerevisiae. One transformant, SCC-105, was further analyzed for in vitro activity with respect to P450SCC (see Example 12).

Example 9

Construction, transformation and expression in Kluweromvces lactis of P45oSCC sequences linked to the pre-region of cytochrome oxidase VI

from Saccharomyces cerevisiae

(a) Construction of the expression cassette

Plasmid pGB950 (see Example 6(b)) was cut with SaH and Xhol and synthetic DNA inserted:

resulting in plasmid pGBSCC-14.

The amino acid sequence of the cytochrome oxidase VI (COX VI) pre-sequence was taken from the article of Wright et al. (J. Biol. Chem., 259, 15401-15407, 1984). The synthetic DNA was made using the preferred yeast codons. The P450SCC coding region of pGBSCC-2 was inserted into pGBSCC-14 cut with Stul.

in the same way as described in example 5(b). The plasmid containing the P450SCC coding sequence in reading frame with the COX VI pre-sequence was named pGBSCC-15 (Figure 20).

(b) Transformation of K. lactis and analysis of the transformants

Transformation of K. lactis with pGBSCC-15 was performed as described in Example 5(c). The transformants were analyzed as described in Example 5(d). The result (Figure 15B, lane 2) shows that P450SCC is produced.

Example 10

Construction, transformation and expression in Saccharomyes cerevisiae of P45oSCC sequences linked to the pre-region of cytochrome oxidase VI from Saccharomyces cerevisiae

(a) Construction of the expression cassette

SalI-HindIII (HindIII partially cleaved) fragment from pGBSCC-15, containing the coding region for P450SCC linked to the COX VI pre-sequence, was inserted into pGBSCC-9 cut with SalI and HindIII. The resulting plasmid was named pGBSCC-16 (Figure 21).

(b) Transformation of S. cerevisiae and analysis of the transformants

S. cerevisiae strain D273-10B was transformed with pGBSCC-16 as described in Example 6(b). The transformants were analyzed as described in Example 6(c). The result shown in Figure 15C (lane 2) demonstrates expression of P450SCC in S. cerevisiae.

Example 11

In vivo activity of P450SCC in Bacillus licheniformis SCC-201

B. licheniformis SCC-201 was obtained as described in Example 3. The organism was inoculated into 100 ml of medium A.

After sterilization and cooling to 30°C to complete the medium, 60 g maltose monohydrate was dissolved in 200 ml distilled water (sterilized 20 min., 120°C), 200 ml 1M potassium phosphate buffer (pH 6.8; sterilized 20 min ., 120°C), 1.7 g of yeast nitrogen base (Difco) dissolved in 100 ml of distilled water (sterilized by membrane filtration) added to the medium.

The culture was grown for 64 hours at 37°C and then 2 ml of this culture was added as inoculum to 100 ml of medium A containing 10 mg of cholesterol. Cholesterol was added as a solution containing cholesterol 10 mg; Tergitol™/ethanol (1:1, v/v), 0.75 ml and Tween 80™, 20 µl. The culture was grown for 48 hours at 37°C, then the culture was extracted with 100 ml of dichloromethane. The mixture was separated by centrifugation and the organic solution layer collected. The extraction procedure was repeated twice and 3 x 100 mL dichloromethane fractions were collected. Dichloromethane was evaporated by vacuum distillation and the dried extract (about 450 mg) was analyzed for pregnenlone using a gas chromatography-mass spectrometer combination.

GC-MS analysis

From the dried extract, a defined amount was taken and silylated by adding a mixture of pyridine bis-(tri-methylsilyl)-trifluoroacetamide and trimethyl-chlorosilane. The silylated sample was analyzed by a GL-MS-DS combination (Carlo Erba MEGA 5160-Finnigan MAT 311A-Kratos DS 90) in a selected ion mode. Gas chromatography was performed under the following conditions: injection moving needle at 300°C; column M. cpsil 29 0.25 inner diameter df 0.2 µm operated at 300°C isothermal; direct entry into MS source.

Samples were analyzed by measuring ions m/z 298 from pregnenolone with a resolution of 800. From the measurements it is clear that in the case of the host strain B. licheniformis T5 no pregnenolone could be detected (detection limit 1 picogram), while in the case of B .licheniformis SCC-201 production of pregnenolone could be easily measured.

Example 12

In vitro activity of P450SCC obtained from Saccharom<y>es cerevisiae SCC-105

S. cerevisiae SCC-105 obtained as described in example 8 was inoculated into 100 ml medium B. Medium B contained per In distilled water:

This culture was grown for 48 hours at 30°C and then this culture was used as an inoculum for a fermenter containing medium C.

The pH was adjusted to pH = 6.0 with ammonia (25%) and the fermenter including medium was sterilized (1 hour, 120°C).

After cooling, 2.4 g of geneticin was dissolved in 25 ml of distilled water and sterilized by membrane filtration and added to the medium. The inoculated mixture was grown in the stirred reactor (800 rpm) at 30°C, while sterile air was passed through the medium at a rate of 300 l/h and the pH was automatically maintained at 6.0 with 4N H2SO4 and 5% NH4OH ( 5% NH4OH in distilled water; sterilized by membrane filtration). After 48 hours, a feed of lactic acid (90%, sterilized by membrane filtration) was started at a rate of 20 g/hour. The fermentation is then resumed for 40 hours, while the cells are collected by centrifugation (4000xg, 15 min.).

The precipitate was washed with 0.9% (w/w) NaCl, followed by centrifugation (4000xg, 15 min.); the precipitate washed with phosphate buffer (50 mM, pH = 7.0) and the cells collected by centrifugation (4000xg, 15 min.). The precipitate was taken up in phosphate buffer (50 mM, pH = 7.0) resulting in a suspension containing 0.5 g wet weight/ml. This suspension was processed in a Dyno<R->mill (Willy A. Bachofen Maschinen-fabrik, Basel, Switzerland). Unbroken cells were removed by centrifugation (4000xg, 15 min.). The cell-free extract (2250 ml, 15-20 mg protein/ml) was stored at -20°C.

P450SCC was coarsely cleaned by the following procedure. From 50 ml of thawed cell-free extract, an impure membrane fraction was centrifuged by ultracentrifugation (125,000xg, 30 min.) and resuspended in 50 ml of 75 mM potassium phosphate solution (pH 7.0), containing 1% of sodium cholate. This disperse solution was gently stirred for 1 hour at 0°C, and then centrifuged (125,000xg, 60 min.). To the supernatant thus obtained, which contained soluble membrane proteins, (NH 4 ) 2 SO 4 was added (30% w/v), while the pH was maintained at 7.0 by the addition of small amounts of NhUOH solution (6N). The suspension was then stirred for 20 min. at 0°C, after which the fraction of precipitated proteins was collected by centrifugation (15000xg, 10 min.). The precipitate was resuspended in 2.5 ml of 100 mM potassium phosphate buffer (pH 7.0), containing 0.1 mM dithiothreitol and 0.1 mM

EDTA. This suspension was eluted through a gel filtration column (PD10, Pharmacia), yielding 3.5 ml of desalted protein fraction (6 mg/ml), which was assayed for P45oSCC activity.

P45oSCC activity was determined by a method based essentially on a method of Doering (Methods Enzymology, 15, 591-596, 1969). The measurement mixture consisted of the following solutions:

Solution A (natural P450SCC electron-donating system): a

10 mM potassium phosphate buffer (pH 7.0), which contained 3 mM EDTA,

3 mM phenylmethylsulfonyl fluoride (PMSF), 20 nM adrenodoxin and

1 |*M adrenodoxin reductase (electron carriers; both purified from bovine adrenal cortex),

1 mM NADPH (electron donor) and

15 mM gtucose-6-phosphate and

8 units/ml glucose-6-phosphate dehydrogenase (NADPH regenerating system).

Solution B (substrate): a micelle solution of 37.5 µM cholesterol (double radioactively labeled with [26,27-<14>C]cholesterol (40 Ci/mol) and [7 alpha-<3>H]cholesterol (400 Ci/mol)) in 10% (v/v) Tergitol™ NP40/ethanol (1:1, v/v).

The measurement was started by mixing 75 µl of solution A with 50 µl of solution B and 125 µl of the impure purified P45oSCC fraction (or buffer as a reference). The mixture was stirred gently at 30°C. The samples (50 nine) were drawn after 0, 30 and 180 min. and diluted with 100 ul of water. Methanol (100) and chloroform (150 µl) were added to the diluted sample. After extraction and centrifugation (5000xg, 2 min.), the chloroform layer was collected and dried. The dry precipitate was dissolved in 50 µl of acetone, which contained 0.5 mg of a steroid mixture (cholesterol, pregnenolone and progesterone (1:1:1, w/w/w)) and then 110 ni of concentrated formic acid was added. The suspension was heated for 15 min at 120°C. Then <14>C /<3>H ratio determined by doubly labeled liquid scintillation counting This ratio is a direct measure of the side chain cleavage reaction, because labeled side chain is volatilized from the mixture as isocaprylic acid during the heating procedure.

Using this method, it was found that the P45oSCC fraction, crudely purified from S. cerevisiae SCC-105, showed side chain cleavage activity. Within 3 hours of incubation, 45% of the cholesterol had been converted. By means of thin-layer chromatography, the reaction product was identified as pregnenolone.

Example 13

Molecular cloning of a full-length cDNA encoding the bovine cytochrome <p>45o-steroid 17a-hydroxylase (P4501 7a).

Approximately 10<6> pfu of the bovine adrenal cortex cDNA library described in Example 1 was screened for P45o17acDNA sequences using two <32>P-end-labeled synthetic oligomers specific for P4so17acDNA. Oligomer 17ct-1 (5"-AGT GGC CAC TTTGGG ACG CCC AGA GAA TTC-3') and oligomer 17a-2 (5-

GAG GCT CCT GGG GTA CTT GGC ACC AGA GTG CTT GGT-3') is

complementary to the bovine P45o17acDNA sequence as described by Zuber et al.

(J. Biol. Chem., 261, 2475-2482, 1986) from respectively positions 349 to 320 and 139 to 104.

Selection with oligomer 17a-1 showed ± 1500 hybridizing pfu<*>s. Several hybridizing pfu's were selected, purified and scaled up for preparative isolation of phage DNA. The EcoRI inserts of the recombinant Iambda-gt11 DNAs were subcloned into the EcoRI site of pTZ18R. One clone, pGB17a-1, was further characterized by restriction endonuclease mapping and DNA sequencing. Plasmid pGB17a-1 contains a 1.4 kb EooRI insert complementary to the 3" portion of P4so17a from the EcoRI site at position 320 to the polyadenylation site at position 1721 as described by Zuber et al. A map of pGB17a-1 is shown in Figure 22A .

Eight hybridizing pfu's were obtained by screening the cDNA library of oligomers 17a-2. After purification, scale-up of recombinant phage and isolation of rec Iambda-gt11 DNAs, the EcoRI inserts were subcloned into the EcoRI site of pTZ18R. The EcoRI inserts ranged in length from 270 bp to

1.5 kbp. Only one clone, pGB17a-2 containing a 345 bp EcoRI fragment, was further examined by nucleotide sequencing and compared with the published P45o17acDNA sequence results of Zuber et al. As shown in Figure 22B, the P45017acDNA sequence in pGB17a-2 started 72 bp upstream of the putative AUG start codon at position 47 and shows complete homology with the 5' portion of P4so17acDNA to the EcoRI site at position 320 as described by Zuber et al.

A full-length bovine P4so17acDNA was constructed by molecular cloning in E. coli JM101 using a ligation mixture containing a partial EcoRI cleaved pGB17a-1 and the 345 bp EcoRI fragment of pGB17a-2. The obtained clone pGB17a-3 contains a full-length bovine P4so17acDNA and is shown in Figure 22C.

Example 14

Construction and transformation of a full-length P450I7ctcDNA clone in the yeast Kluyveromvces lactis

(a) Construction of the expression vector

To obtain a suitable expression vector in yeast hosts for bovine P45017a, pGB17a-3 was mutated by site-specific mutagenesis as described by Zoller and Smith, (Methods in Enzymol., 100,468-500,1983); Zoller and Smith, (Methods in Enzymol., 154, 329-350, 1987) and Kramer and Fritz, (Methods in Enzymol., 154, 350-367, 1987). Plasmids and strains for in vitro mutagenesis experiments were obtained from Pharmacia Inc.

As shown in Figure 23, 9 bp just upstream of the ATG start codon were changed to obtain a SalI restriction site and optimal yeast translation signals used using the synthetic oligomer 17a-3

The resulting plasmid pGB17a-4 was digested with SalI and SmaI; The DNA fragment containing full-length P45o17acDNA was separated by gel electrophoresis, isolated and transferred by cloning into E. coli JM101 in the pGB950 vector (see Example 5) which was first cleaved with XhoI. overhangs filled in with Klenow NA polymerase and then digested with SalI, resulting in plasmid pGB17a-5 as shown in Figure 24.

(b) Transformation of K. lactis

15 ng of pGb17a-5, cut in the rare SacII site of the lactase promoter, was used to transform K. lactis strain CBS 2360 as shown in Example 5. Transformants were analyzed for the presence of integrated pGB17a-5 sequences in the host genome by Southern analysis. One transformant 17a-101, which contained at least three copies of pGB17a-5 in the genomic host DNA, was further analyzed for in vivo activity of P4so17a (see Example 16).

Example 15

Construction and transformation of P45017a in the bacterial hosts Bacillus subtilis and Bacillus licheniformis

(a) Construction of the expression vector

To obtain a suitable expression vector in the Bacillus hosts for bovine P45o17a, pGB17cc-3 was mutated by site-specific mutagenesis as described in Example 14.

As indicated in Figure 25, an Ndel restriction site was introduced at the ATG start codon using the synthetic oligomer 17a-4:

The resulting plasmid pGB 17a-6 was partially digested with EcoRI: the DNA fragment containing full-length P^^acDNA was separated by gel electrophoresis, isolated and ligated to EcoRI-digested pBHA-1 DNA as shown in Figure 26. The ligated DNA was cloned by transferring the mixture to E. coli JM101 pre-GB17a-7.

(b) Transformation of B. subtilis and B. licheniformis

The "Hpall" Bacillus promoter was introduced upstream of the P45o17acDNA sequences by digesting pGB17a-6 with the restriction enzyme NdeI. separation of the E. coli portion of the shared plasmid by agarose gel electrophoresis and then ligation and transformation of B. subtilis 1A40 (BGSC 1A40) competent cells. Neomycin resistant colonies were analyzed and the plasmid pGB17a-8 (Figure 27) was obtained.

Transformation of the host B. licheniformis T5 (CBS 470.83) was also performed with pGB17a-8. The plasmid remains stable in the appropriate Bacillus host as shown by restriction analysis of pGB17a-8 even after many generations.

Example 16

In vivo activity of P45o17a in Kluweromvces lactis 17a-101

K. lactis 17a-101 was obtained as described in example 14. The organism was inoculated in 100 ml medium D. Medium D contained per liter of distilled water:

After sterilization and cooling to 30°C, 2.68 g of yeast nitrogen base (Difco) dissolved in 40 ml of distilled water (sterilized by membrane filtration) and 50 mg of neomycin dissolved in 1 ml of distilled water (sterilized by membrane filtration) were added to the medium. Then 50 mg of progesterone was dissolved in 1.5 ml of dimethylformamide and 100 ml of medium was added. The culture was grown for 120 hours at 30°C and then 50 ml of culture liquid was extracted with 50 ml of dichloromethane. The mixture was centrifuged and the organic solution layer separated. Dichloromethane was evaporated by vacuum distillation and the dried extract (about 200 mg) was taken up in 0.5 ml of chloroform. This extract contained 17α-hydroxyprogesterone as shown by thin layer chromatography. The structure of the compound was confirmed by H-NMR and <13>C-NMR. NMR analysis also showed that the ratio 17a-hydroxyprogesterone/progesterone

in the extract was about 0.3.

Example 17

Molecular cloning of full-length cDNA encoding bovine cytochrome

P45o-steroid-21-hydroxylase (P450C2I)

Approximately 10<6> Pfu of the bovine adrenal cortex cDNA library, made as described in Example 1, was hybridized with a <32>P-end-labeled oligo C21-1. This oligo, which contained the sequence 5'- GAT GAT GCT GCA GGT AAG CAG AGA GAA TCC-3', is a specific marker for the bovine P450C2I gene located downstream of the EcoRI site in the P45oC21 cDNA sequence as described by Yoshioka et al. (J. Biol. Chem., 261, 4106-4109, 1986). From the analysis was

one hybridizing pfu obtained. The EcoRI insert in this recombinant lambda-gt11 DNA was subcloned into the EcoRI site of pTZ18R resulting in a

construct called pGBC21-1. As shown in Figure 28, this plasmid contains a 1.53 kb EcoRI insert complementary to the P450C21 cDNA sequences from the EcoRI site at position 489 to the polyadenylation site as described by Yoshioka et al., as shown by nucleotide sequencing.

To isolate the remaining 5' portion (490 bp) of P450C2ICDNA, new bovine adrenal cortex cDNA library was made according to the method described in Example 1 with only one modification. As a primer for the first strand cDNA synthesis, an additional oligomer C21-2 was added. Oligomer C21-2 with the nucleotide sequence 5'- AAG CAG AGA GAA TTC-3' is located downstream of the EcoRI site of P450C21 cDNA from position 504 to 490.

Probing this cDNA library with a <32>P-end-labeled oligomer C21-3, which contains the P450C21 specific sequence 5'-CTT CCA CCG GCC CGA TAG CAG GTG AGC GCC ACT GAG-3* (positions 72 to 37) showed approximately 100 hybridizing pfu's. The EcoRI insert of only one recombinant Iambda-gt11 DNA was subcloned into the EcoRI site of pTZ18R resulting in a construct named pGBC21-2.

This plasmid (Figure 28) contains a 540 bp insert complementary to P450C21 cDNA sequences from position -50 to the EcoRI site at position 489 as shown by nucleotide sequencing.

Example 18

Construction of the P450C21 cDNA Bacillus expression vector and transformation of the bacterial hosts Bacillus subtilis and Bacillus licheniformis

(a) Construction of the expression vector

To construct a full-length P45oC21cDNA with flanking sequences specific for the Bacillus expression vector pBHA-1, the 5' part of the P450C21 gene was first modified by the polymerase chain reaction (PCR) method with pGBC21-2 as template and two specific P4soC21 oligomers as primers (start DNA). Oligomer C21-4 (5'-CTG ACT GAT ATC CAT ATG GTC CTC GCAGGG CTG CTG-3') contains 21 nucleotides complementary to the C21 sequences from positions 1 to 21 and 18 additional bases to provide an EcoRV restriction site and an Ndel restriction site at the ATG stat codon.

Oligomers C21-5 (5'-AGC TCA GAA TTC CTT CTG GAT GGT

CAC-3') is 21 bases complementary to the minus strand upstream of the EcoRI site at position 489.

PCR was performed as described by Saiki et al (Science 239, 487^91, 1988) with minor modifications. PCR was performed in a volume of 100 µl containing: 50 mM KCl, 10 mM Tris-HCL pH 8.3, 1.5 mM MgCl2, 0.01% (w/v) gelatin, 200 nM each dNTP, 1 ^ M each C21 primer and 10 ng pGBC21-2 template. After denaturation (7' at 100°C) and addition of 2 U Taq polymerase (Cetus), the reaction mixture was subjected to 25 amplification cycles (each: 2 min. at 55°C, 3 min. at 72°C , 1 min at 94°C) in a DNA amplifier (Perkin-Elmer).

In the last cycle, the denaturation step was omitted. A schematic view of this P45oC21cDNA amplification is shown in Figure 29.

The amplified fragment was digested with EcoRV and EcoRI and inserted by cloning into the appropriate sites of pSP73 (Promega). The resulting plasmid is called pGBC21-3. As shown in Figure 30, the 3' P^0C21 - EcoRI fragment of pGBC21-1 was inserted in the correct orientation into the EcoRI site of pGBC21-3. The obtained vector pGBC21-4 was digested with EcoRV and K<p>nl ( Kpnl is located in the multiple cloning site of pSP73) and the fragment containing the full-length P450C21 cDNA was isolated by gel electrophoresis and inserted into the appropriate sites of pBHA- 1 by molecular cloning. The resulting plasmid pGBC21-5 is illustrated in Figure 31.

(b) Transformation of Bacillus

"Hpall" Bacillus promoter was introduced upstream of the P450C21cDNA gene by cleaving pGBC21-5 with the restriction enzyme Ndel, separation of the E. coli part of the common plasmid by agarose gel electrophoresis and then ligation and transformation of B. subtilis 1 A40 (BGSC 1 A40) competent cells. Neomycin resistant colonies were analyzed to obtain pGBC21-6 (Figure 32).

Transformation of the host B. licheniformis T5 (CBS 470.83) was also performed with pGBC21-6. The plasmid remains stable in both Bacillus hosts as judged by restriction analysis.

Example 19

Construction of a P450C2ICDNA yeast expression vector and transformation of the yeast host Kluweromvces lactis

(a) Construction of the expression vector

To obtain a suitable restriction vector in yeast hosts for bovine P45oC21, pGBC21-2 was mutated by site-specific mutagenesis as described in Example 14. For the mutation, oligomer C21-6 (5-CCT CTG CCT GGG TCG ACA AAA ATG GTC CTC GCA GGG<a >3') used to create a Sall restriction site and optimal yeast translation signals upstream of the ATG start codon as indicated in Figure 33.

The SalI/EcoRI DNA fragment of the obtained plasmid pGBC21-7 was ligated to the 3' P450C21 - EcoRI fragment of pGBC21-1 and inserted by cloning into the appropriate sites of pSP73 as indicated in Figure 34. The obtained pGBC21-8 was cut with SalI and EcoRV (the EcoRV site is located in the multiple cloning site of pSP73) and the DNA fragment containing the full-length P45oC21cDNA was inserted into the yeast expression vector pGB950. The obtained pGBC21-9 is depicted in Figure 35.

(b) Transformation of K. lactis

15 ng of pGBC21-9 was digested with Sadl and transformation of K. lactis CBS 2360 was performed as described in Example 5(c).

Example 20

Molecular cloning of full-length cDNA encoding the bovine cytochrome P450 steroid 11 β-hydroxylase (P45011B) Bovine adrenal cortex cDNA library was made as described in Example 1 with one modification. An additional P4so11 β-specific primer (oligomer 1 ip-1) with the nucleotide sequence 5-GGC AGT GTG CTG ACA CGA-3* was added to the reaction mixture of the first strand cDNA synthesis. Oligomer 1 ip-1 is located just downstream of the translational stop codon from position 1530 to 1513. The nucleotide sequences and map positions of said P4so11 p-oligomers are all taken from the P450I ipcDNA sequence results described by Morohashi et al. (J. Biochem. 102 (3), 559-568, 1987). The cDNA library was probed with a <32>P-labeled oligomer 11B-2(5'-CCG CAC CCT GGC CTT TGC CCACAG TGC CAT-3') and located at the 5' end of P45o11 pcDNA from position 36 to 1.

Examination with oligomer 11B-2 showed 6 hybridizing pfu's. These were further purified and analyzed with oligomer 11 p-3 (5'-CAG CTC AAA GAG AGT CAT CAG CAA GGG GAA GGC TGT-3', positions 990 to 955). Two of six showed a positive hybridization signal with <32>P-labeled oligomer 11 p-3.

The EcoRI inserts of both 1 ip-lambda-gt11 recombinants were subcloned into the EcoRI site of pTZ18R. One clone with a 2.2 kb EcoRI insert (pGB11 p-1) was further analyzed by restriction enzyme mapping and is shown in Figure 36. pGB11 p-1 contains all coding P45o11 pcDNA sequences as determined by Morohashi et al.

Example 21

Construction of P45oC21cDNA Bacillus expression vector and transformation of the bacterial hosts Bacillus subtilis and Bacillus licheniformis

(a) Construction of the expression vector

A full-length P45o11pcDNA with modified flanking sequences of the Bacillus expression vector pBHA-1 was obtained by the PCR method (described in Example 18) with pGB1ip-1 as template and two specific P4so11 p-oligomers as primers.

Oligomer 11p-4(5'-TTT GAT ATC GAA TTC CAT ATG GGC ACA AGA GGT GCT GCA GCC-3') contains 21 bases complementary to the mature P4so11 pcDNA sequence from position 72 to 93 and 21 bases to make EcoRV. EcoRI and Ndel restriction sites and ATG start codon.

Oligomer 11 p-5(5*-TAA CGA TAT CCT CGA GGG TAC CTACTG GAT GGC CCG GAA GGT-3) contains 21 bases complementary to the minus P4so11 PcDNA strand upstream of the translational stop codon at position 1511 and 21 bases to create the restriction sites for EcoRV. Xhol and K<p>nl.

After PCR amplification (amplification) with the above-mentioned template and P45011 p-primers, the amplified fragment (1.45 kb), was digested with EcoRI and

KpnI and inserted by cloning into the Bacillus expression vector pBHA-1 cut with EcoRI and K<p>nI to obtain the vector pGB11 p-2 (see Figure 36).

(b) Transformation of Bacillus

The "HpaH" Bacillus promoter was introduced upstream of the P450I IpcDNA sequences by digesting pGB1ip-2 with Ndel, separation of the E. coli portion of the shared plasmid by agarose gel electrophoresis and then ligation (as described in Example 18) and transformation of B .subtilis 1A40 (BGSC 1A40) competent cells. Neomycin-resistant colonies were assayed, and the plasmid pGB11 p-3 was obtained. The obtained plasmid pGB11 p-3 was also transferred to the B. licheniformis host strain T5 (CBS 470.83).

Example 22

Construction of the P45011 pcDNA yeast expression vector and transformation of the yeast host Kluvveromvces lactis

(a) Construction of the expression cassette

A full-length P450I IpcDNA with modified flanking sequences of the yeast expression vector pGB950 was obtained by the PCR method (described in Example 18) with pGB11 p-1 as template and two specific P45011 p-oligomers as primers.

Oligomer 11 p-6(5'-CTT CAG TCG ACA AAA ATG GGC ACA AGA GGT GCT GCA GCC-3') contains 21 bases complementary to the mature P450I IpcDNA sequence from position 72 to 93 and 18 additional bases to create a Sali restriction site, an optimal yeast translation signal and an ATG start codon.

Oligomer 11 p-5 is described in example 21 (a). After PCR amplification with the above-mentioned template and P45o11 p primers, the amplified fragment (1.45 kb), digested with SaH and Xhol and inserted by cloning into the yeast expression vector pGB950 cut with SaH to obtain the vector pGB11 p- 4 (figure 37).

(b) Transformation of K. lactis

15 ng of pGB11B-4 was cut at the rare SacII site in the lactase promoter and transformation of K. lactis CBS 2360 was performed as described in Example 5(c).

Example 23

Molecular cloning and construction of full-length coding cDNA

for the bovine adrenodoxin (ADX), and then transformation and expression of ADXcDNA in the yeast Kluweromvces lactis

(a) Molecular cloning of ADX

Full-length ADXcDNA, with 5* and 3' flanking sequences modified into the yeast expression vector pGB950, was directly obtained from bovine adrenal cortex mRNA/cDNA mixture (for detailed description see Example 1) by amplification using the PCR method (see Example 18) .

For ADXcDNA amplification, two synthetic oligomer primers were synthesized.

Oligomer ADX-1 (5'-CTT CAG TCG ACA AAA ATG AGC AGC TCA GAA GAT AAA ATA-3") which contained 21 bases complementary to the <5->end of the mature ADXcDNA sequence as described by Okamura et al (Proe . Nat. Acad. Sci. USA 82, 5705-5709, 1985) from positions 173 to 194. The oligomer ADX-1 contains at the 5-end 18 additional nucleotides to create a SalI restriction site, an optimal yeast translation signal and an ATG start codon.

The oligomer ADX-2 (5'-TGT AAG GTA CCC GGG ATC CTT ATT CTA TCT TTG AGG AGT T-3') is complementary to the 3-end of the minus strand of ADXcDNA from position 561 to 540 and contains additional nucleotides to create restriction sites for BamHI, SmaI and KpnI.

PCR was performed as described in Example 18 with 1 µM each of ADX primers and 10 ni mRNA/cDNA mixture (as described in Example 1) as template.

A schematic overview of this ADXcDNA amplification is shown in figure 38.

The amplified fragment contains full-length ADXcDNA sequence with modified flanks, which was characterized by restriction site analysis and nucleotide sequencing.

(b) Construction of the expression vector

The amplified ADX cDNA fragment was digested with SalI and SmaI and inserted by cloning into the yeast expression vector pGB950 cut with SalI and EcoRV. The resulting plasmid pGBADX-1 is depicted in Figure 38.

(c) Transformation of K. lactis

15 µg of pGBADX-1 was cut at the rare SacII site in the lactase promoter and transformation of K. lactis CBS 2360 was performed as described in Example 5(c).

(d) Analysis of the transformants

Two transformants, ADX-101 and ADX-102 and the control strain CBS 2360 were selected for further analysis. The strains were grown in YEPD medium for approximately 64 hours at 30°C. Total cellular protein was isolated as described in Example 5(d). From the supernatants, 8 µl samples were taken for analysis on immunoblot (see Figure 39, lanes 3,4 and 5).

The results showed that a protein of the expected length (14 kDa) is expressed in K. lactis cells transformed with pGBADX-1.

The in vitro ADX activity of transformant ADX-102 is described in Example 24.

Example 24

In vitro activity of adrenodoxin obtained from

Kluyveromyces lactis ADX-102

K. lactis ADX-102, obtained as described in Example 23, and control strain K. lactis CBS 2360 were grown in 100 ml YEPD medium (1% yeast extract, 2% peptone, 2% glucose monohydrate) containing 2.5 ml 6, 7% (w/w) yeast nitrogen base (Difco laboratories) solution and 100 mg 1~<1> of geneticin (G418 surfate; Gibco Ltd.), for 56 hours at 30°C. The cells were collected by centrifugation (4000xg, 15 min.), resuspended in a physiological saline solution and washed with phosphate buffer (pH 7.0, 50 mM). After centrifugation (4000xg, 15 min.), the precipitate was dissolved in phosphate buffer (pH 7.0, 50 mM), resulting in a suspension containing 0.5 g cell wet weight/ml. The cells were disrupted using a Braun MSK homogenizer (6x15 seconds, 0.45 - 0.50 mm glass beads). Unbroken cells were removed by centrifugation (4000xg, 15 min.). Cell-free extracts (40 mg protein/ml) were stored at -20°C.

ADX activity, i.e. electron transfer capacity from adrenodoxin reductase to cytochrome P450SCC, in cell-free extracts was determined by P4soSCC activity measurement. The measurement mixture consisted of the following solutions: Solution A (natural P450SCC electron-donating system with the exception of ADX): a 50 mM potassium phosphate buffer (pH 7.0), containing 3 mM EDTA, 2 µM adrenodoxin reductase (purified from bovine adrenal cortex), 1 mM NADPH electron donor), 15 mM glucose-6-phosphate and 16 units/ml glucose-6-phosphate dehydrogenase (NADPH regenerating system).

Solution B (substrate and enzyme): a micelle solution of 75 nM cholesterol (double radiolabeled with [26,27-<14>C]-cholesterol (40 Ci/mol) and [7a-<3>H] cholesterol (400 Ci/ mol)) and 1.5 µM P450SCC (purified from bovine adrenal cortex) in 10% (v/v) Tergitol™ NP 40/ethanol (1:1, v/v).

The assay was started by mixing 75 µl of solution A with 50 µl of solution B and 125 µl of cell-free extract or 125 µl of potassium phosphate buffer (50 mM, pH 7.0) containing 10 nM ADX (purified from bovine adrenal cortex). The mixture was stirred gently at 30°C. The samples were drawn after 15 min. incubation and diluted with 100 ni water. From one sample, substrate and product(s) were extracted with 100 n' methanol and 150 n' chloroform. After centrifugation (5000xg, 2 min.), the chloroform layer was collected and dried. The dry precipitate was dissolved in 50 ml of acetone, containing 0.5 mg of steroid mixture (cholesterol, pregnenolone and progesterone (1:1:1, w/w/w)) and then 110 ml of concentrated formic acid was added. The suspension was heated for 15 min. at 120°C. Then the <14>C/<3>H ratio was determined by doubly labeled liquid scintillation counting. The ratio is a direct measure of the side chain cleavage reaction, because the <14>C-labeled side chain is evaporated from the mixture as isocaprylic acid during the heating procedure.

Using this method, ADX electron transfer activity could be readily demonstrated in a cell-free extract of K. lactis ADX102. In the measurements with cell-free extract of K. lactis ADX-102 or with purified ADX, the side chain of cholesterol was cleaved within 15 min. with a yield of 50%, while in the method with cell-free extract of the control strain K. lactis CBS 2360 no side chain cleavage could be detected.

Example 25

Molecular cloning and construction of a full-length cDNA which

codes for the bovine adrenodoxin oxidoreductase (ADR), and subsequent transformation of ADRcDNA in the yeast Kluweromvces lactis

(a) Cloning of adrenodoxin oxidoreductase

Bovine adrenocortical cell cDNA library was made as described in Example 1 except for one change. An additional ADR-specific primer (oligomeric ADR-1) with the nucleotide sequence 5-GGC TGG GAT CTA GGC-3' was added to the reaction mixture for first-strand cDNA synthesis. Oligomeric ADR-1 is located just downstream of the translational stop codon from position 1494 to 1480. The nucleotide sequences and map positions of said ADR oligomers are all taken from the ADRcDNA sequence results described by Nonaka et al, Biochem. Biophys. Res. Comm. 145(3), 1239-1247, 1987.

The obtained cDNA library was probed with a <32>P-labeled oligomer

ADR-2 (5'-CAC CAC ACA GAT CTG GGG GGT CTG CTC CTG TGG GGA-3"). 4 hybridizing pfu were identified and subsequently purified. However, only 1 pfu also showed a positive signal with oligomeric ADR-3 ( 5'-TTC CAT CAG CCG CTT CCT CGG GCG AGC GGC CTC CCT-3'), which is located in the middle of ADRcCDNA (position 840 to 805). The ADRcDNA insert (approximately 2 kb) was cloned into the EcoRI site of pTZ18R.

The resulting plasmid pGBADR-1 contains a full-length ADRcDNA as shown by restriction enzyme mapping and nucleotide sequencing. The physical map of pGBADR-1 is illustrated in Figure 40.

(b) Construction of the expression cassette

Full-length ADRcDNA with modified flanking sequences of the yeast expression vector pGB950 was obtained by the PCR method (see Example 18) with pGBADR-1 as template and two specific ADR oligomers as primers.

Oligomer ADR-4 (5"-CGA GTG TCG ACA AAA ATG TCC ACA CAG GAG CAG ACC-3') contains 18 bases complementary to the mature ADRcDNA sequences from position 96 to 114 and 18 bases to introduce a SalI restriction site, an optimal yeast translation signal, and an ATG start codon.

Oligomer ADR-5 (5'-CGT GCT CGA GGT ACC TCA GTG CCC CAG CAG CCG CAG-3") contains 18 bases complementary to the minus strand of ADRcDNA upstream of the translational stop codon at position 1479 and 15 bases to create Kpnl and Xhol residues sion sites for cloning into different expression vectors.

After amplification with the above-mentioned template and ADR primers, the amplified fragment (1.4 kb) was digested with SalI and XhoI and inserted by cloning into the yeast expression vector pGB950 which had been cut with SalI and XhoI.

The resulting plasmid pGBADR-2 is illustrated in Figure 40.

(c) Transformation of K. lactis

15 ng of pGBADR-2 was cut at the rare SacII site in the lactase promoter and transformation of K. lactis CBS 2360 was performed as described in Example 5(c).

Example 26

Cloning of full-length cDNA encoding bovine NADPH cytochrome

P450 reductase (RED)

Bovine adrenal cortex cDNA library described in Example 1 was probed with a <32>P-labeled synthetic oligomer 5'-TGC CAG TTC GTA GAG CAC ATT GGT GCG TGG CGG GTT AGT GAT GTC CAG GT-3" specific for a conserved amino acid region in rat -, pig and rabbit RED as described by Katagari et al (J. Biochem., 100, 945-954, 1986) and Murakami et al (DNA, 5, 1-10, 1986).

Five hybridizing pfu's were obtained and further characterized by restriction enzyme mapping and nucleotide sequencing. Full-length REDcDNA was inserted into expression vectors and transformed into appropriate host cells as mentioned in Examples 2, 3 and 6.

Example 27

Construction, transformation and expression of an expression cassette encoding the proteins P450SCC and ADX in the yeast Kluweromvces lactis

(a) Construction of the expression cassette

The expression cassette pGBADX-1 (see Example 23) was digested with Sadl and HindIII (partially) and overhangs were filled in using Klenow DNA polymerase. The DNA fragment that included part of the lactase promoter (but still functional), the ADX coding sequence and the lactase terminator

was separated and isolated by agarose gel electrophoresis and then inserted into pGBSCC-7, which was first linearized by XbaI digestion (see Example 5(b)) and overhangs filled in using Klenow DNA polymerase. The construction was set up in such a way that a unique restriction seat (Saddle) is achieved, which is

required to transfer the plasmid to K. lactis.

This rare Sac1 restriction site is located in the lactase promoter sequence flanking the SCC sequence, since the Sac1 restriction site in the lactase promoter flanking the ADX sequence is destroyed by the complementation reaction.

The resulting expression cassette pGBSCC/ADX-1 contains the coding sequence for SCC as well as for ADX, each driven by the lactase promoter.

(b) Transformation of K. lactis

Transformation of K. lactis CBS 2360 was performed as described in Example 5(c) with 15 µg of pGBSCC/ADX-1, which was linearized at the rare SacII restriction site. One transformant (SCC/ADX-101) was selected for SCC and ADX expression studies.

(c) Analysis of the transformant K. lactis SCC/ADX-101

Cellular protein fractions were prepared from cultures of SCC/ADX-101 and control strain CBS 2360 as described in Example 5(d) and analyzed by SDS/PAGE and Western blotting. The blot was examined using antibodies that were specific for the respective SCC and ADX. Compared to the control strain, the cellular protein fraction of transformant SCC/ADX-101 showed two additional bands of expected length (53 and 14 kDa, respectively) indicating production of both SCC and ADX proteins. Production levels of both proteins in transformant SCC/ADX-101 are comparable to the levels observed in the transformants that produced only one protein (for SCC see Figure 15A, lane 3, and for ADX Figure 39, lane 5).

In vitro SCC and ADX activity in transformant SCC/ADX-101 is described in Example 28.

Example 28

In vitro activity of P450SCC and adrenodoxin obtained from Kluweromvces lactis SCC/ADX-101

K. lactis SCC/ADX-101 obtained as described in Example 27 and control strain K. lactis SCC-101 as described in Example 5(d) were grown in 1 I YEPD medium (1% yeast extract, 2% peptone, 2% glucose monohydrate) containing 100 mg 1"<1> of geneticin (G418 sulfate; Gibco Ltd.), for 72 hours at 30°C. The cells were collected by centrifugation (4000xg, 15 min.), resuspended in physiological saline and washed with phosphate buffer (pH 7.5, 75 mM). After centrifugation (4000xg, 15 min.), the pellet was resuspended in phosphate buffer (pH 7.5, 75 mM) resulting in a suspension containing 0.5 g cell wet weight/ml. The cells were disrupted using a Braun MSK homogenizer (6x15 seconds, 0.45 - 0.50 mm glass beads). Undisrupted cells were removed by centrifugation (4000xg, 15 min.).

In the cell-free extracts, the activity of the protein complex P45oSCC/ADX was measured, by determining the cholesterol side chain cleavage reaction in the presence of NADPH and ADR. The test mixture contained the following solutions: Solution A (natural P450SCC electron-donating system excluding ADX): 50 mM potassium phosphate buffer (pH 7.0), containing 3 mM EDTA, 2 µM adrenodoxin reductase (purified from bovine adrenal cortex), 1 mM NADPH (electron donor ), 15 mM glucose-6-phosphate and 16 units/ml glucose-6-phosphate dehydrogenase (NADPH regenerating system).

Solution B (substrate): micelle solution of 37.5 |iM cholesterol (double radioactively labeled with [26,27-<14>C] cholesterol (40 Ci/mol) and [7<x-<3>H] cholesterol (400 Ci/mol)) in 10% (v/v) TergitolTM NP 40/ethanol (1:1, v/v).

The measurement was started by mixing 75 µl of solution A with 50 µl of solution B and 125 µl of cell-free extract. The mixture was stirred carefully at 30°C. The samples were withdrawn after 60 min. incubation and diluted with 100 µl of water. From one sample was substrate and product(s) extracted with 100 µl methanol and 150 µl chloroform. After centrifugation (5000xg, 2 min.), the chloroform layer was collected and dried. The dry precipitate was dissolved in 50 µl acetone, containing 0.5 mg steroid mixture (cholesterol, pregnenolone and progesterone (1:1:1, w/w/w)) and then 110 ni concentrated formic acid was added.

The suspension was heated for 15 min. at 120°C. Hereafter, the 14C/3H ratio was determined by doubly labeled liquid scintillation counting. The ratio is a direct measure of the side chain cleavage reaction, because the <14>C-labeled side chain is evaporated from the mixture as isocaprylic acid during the heating procedure.

Using this method, cholesterol side chain cleavage activity was shown in the cell-free extract of K. lactis SCC/ADX-101, while no activity was measured in the cell-free extract from K. lactis SCC101.

Using HPLC analysis, the reaction product made from a cell-free extract from K. lactis SCC/ADX-101 was identified as pregnenolone.

Claims (27)

1. Use of an expression cassette in multigene systems for performing multi-step transformations as shown in Figure 1 where the expression cassette is operable in a non-mammalian recombinant host, characterized in that the expression cassette comprises a heterologous coding DNA sequence which codes for a protein which is functional, alone or together with one or more additional proteins, in catalyzing an oxidation step in the biological reaction pathway for the conversion of cholesterol to hydrocortisone, which step is selected from the group consisting of: conversion of cholesterol to pregnenolone; conversion of pregnenolone to progesterone; conversion of progesterone to 17α-hydroxyprogesterone; conversion of 17α-hydroxyprogesterone to cortexolone; and conversion of cortexolone to hydrocortisone, and the corresponding control sequences effective in said host. 2. Use of an expression cassette according to claim 1, characterized in that the protein is selected from the group consisting of: side chain-cleaving enzyme (P450SCC); adrenodoxin (ADX); adrenodoxin reductase (ADR); 3B-hydroxysteroid dehydrogenase/isomerase (3B-HSD); steroid 17α-hydroxylase (P45o17cc); NADPH cytochrome P45o reductase (RED); steroid-21 -hydroxylase (P45oC21); and steroid-11 B-hydroxylase (P45011B). 3. Use of an expression cassette according to claim 2, characterized in that the heterologous coding DNA sequence originates from bovine species. 4. Use of an expression cassette pGBSCC-5 according to claim 3, characterized in that the heterologous DNA sequence codes for the enzyme P450SCC and that the expression cassette reproduced in Figure 8 is operative in Bacillus. 5. Use of the expression cassette pGBSCC-17 according to claim 3, characterized in that the heterologous DNA sequence codes for the enzyme P450SCC and that the expression cassette reproduced in Figure 10 is operative in E. coli. 6. Use of the expression cassette according to claim 5, characterized in that the heterologous DNA codes for the enzyme P450SCC and that the expression cassette, which is operative in Kluyveromyces lactis, is taken from the group consisting of pGBSCC-7 reproduced in Figure 14, pGBSCC-12 reproduced in Figure 18 and pGBSCC-15 reproduced in Figure 20, or that the expression cassette operative in Saccharomyces cerevisiae is taken from the group consisting of pGBSCC-10 reproduced in Figure 17, pGBSCC-13 reproduced in Figure 19 and pGBSCC-16 reproduced in Figure 21. 7. Use of the expression cassette pGB17a-5 according to claim 3, characterized in that the heterologous DNA codes for the enzyme P45017a and that the expression cassette reproduced in Figure 24 is operative in Kluyveromyces lactis. 8. Use of an expression cassette pGB17a-8 according to claim 3, characterized in that the heterologous DNA codes for an enzyme P45017a and that the expression cassette reproduced in figure 27 is operative in species of Bacillus. 9. Use of an expression cassette pGB11p-4 according to claim 3, characterized in that the heterologous DNA codes for an enzyme P45011P and that the expression cassette reproduced in figure 37 is operative in Kluyveromyces lactis. 10. Expression cassette, operable in a non-mammalian recombinant host comprising a heterologous coding DNA sequence encoding a protein, which is functional, alone or together with one or more additional proteins, in catalyzing an oxidation step in the biological pathway of conversion of cholesterol to hydrocortisone, which step is selected from the group consisting of: conversion of cholesterol to pregnenolone; conversion of pregnenolone to progesterone; conversion of progesterone to 17α-hydroxyprogesterone; conversion of 17α-hydroxyprogesterone to cortexolone; and conversion of cortexolone to hydrocortisone, and the corresponding control sequences effective in said host, characterized in that the protein is selected from the group consisting of: side chain-cleaving enzyme (P450SCC); adrenodoxin (ADX); adrenodoxin reductase (ADR); 3β-hydroxysteroid dehydrogenase/isomerase (3β-HSD); steroid 17ct-hydroxylase (P4so17a); NADPH cytochrome P45o reductase (RED); steroid 21 -hydroxylase (P450C2I); and steroid-11 β-hydroxylase (P45011P). and that the heterologous DNA codes for at least one additional protein of the said group of proteins. 11. Expression cassette according to claim 10, operable in a non-mammalian recombinant host comprising a heterologous coding DNA sequence encoding a protein, which is functional alone or together with one or more additional proteins in catalyzing an oxidation step in the biological reaction pathway for the conversion of cholesterol to hydrocortisone, which step is selected from the group consisting of: conversion of cholesterol to pregnenolone; conversion of pregnenolone to progesterone; conversion of progesterone to 17α-hydroxyprogesterone; conversion of 17α-hydroxyprogesterone to cortexolone; and conversion of cortexolone to hydrocortisone, and the corresponding control sequences which are effective in said host, characterized in that the protein is selected from the group consisting of: side chain-cleaving enzyme (P450SCC); adrenodoxin (ADX); adrenodoxin reductase (ADR); 3β-hydroxysteroid dehydrogenase/isomerase (3B-HSD); steroid 17α-hydroxylase (P45o17a); NADPH cytochrome P45o reductase (RED); steroid-21 -hydroxylase (P45oC21); and steroid-11 β-hydroxylase (P45011P), and where the heterologous DNA which codes for said additional protein is an additional heterologous DNA with its own effective control sequences which codes for a protein from said group of proteins. 12. Expression cassette according to claims 10 and 11, characterized in that the heterologous DNA codes for bovine P450SCC and bovine ADX. 13. Expression cassette operable in a non-mammalian recombinant host characterized in that it comprises a heterologous coding DNA sequence encoding at least two proteins, which are functional, alone or together with one or more additional proteins, in catalyzing at least two oxidation steps selected from the group consisting of: conversion of cholesterol to pregnenolone; conversion of pregnenolone to progesterone; conversion of progesterone to 17<x-hydroxyprogesterone; conversion of 17α-hydroxyprogesterone to cortexolone; and conversion of cortexolone to hydrocortisone, and the corresponding control sequences effective in said host. 14. Expression cassette according to claim 13, operable in a non-mammalian recombinant host comprising a heterologous coding DNA sequence encoding at least two proteins, which are functional alone or together with one or more additional proteins, in catalyzing an oxidation step selected from the group consisting of: conversion of cholesterol to pregnenolone; conversion of pregnenolone to progesterone; conversion of progesterone to 17cc-hydroxyprogesterone; conversion of 17α-hydroxyprogesterone to cortexolone; and conversion of cortexolone to hydrocortisone, characterized in that the heterologous DNA encoding the second protein has its own effective control sequences. 15. Expression cassette according to claims 13 or 14, characterized in that the protein is selected from the group consisting of: side chain-cleaving enzyme (P450SCC); adrenodoxin (ADX); adrenodoxin reductase (ADR); 3B-hydroxysteroid dehydrogenase/isomerase (3B-HSD); steroid 17α-hydroxylase (P45017a); NADPH cytochrome P450 reductase (RED); steroid 21 -hydroxylase (P450C2I); and steroid-11 β-hydroxylase (P45011P)- 16. Expression cassette according to claim 15, characterized in that the heterologous coding DNA sequence originates from bovine species. 17. Recombinant host cell comprising cells of microorganisms, plants or non-mammalian animals containing an expression cassette with heterologous DNA, characterized in that the expression cassette is one defined in any of claims 10-16. 18. Recombinant host cell according to claim 17, characterized in that the host is a microorganism. 19. Recombinant host cell according to claim 18, characterized in that the host is a species of Saccharomyces, Kluyveromyces or Bacillus or is Escherichia coli. 20. Recombinant host cell according to any one of claims 17-19 and containing at least two expression cassettes, characterized in that they are as defined in any of claims 10-16. 21. Method for producing a mixture of exogenous proteins of a recombinant cell in a nutrient medium under conditions that enable the enzymes to form and accumulate in the culture, characterized in that the recombinant cell is a recombinant host cell as defined in claim 20. 22. Method for selective biochemical oxidation in vitro where the process comprises: incubation of the compound to be oxidized in the presence of proteins under conditions that allow oxidation and accumulation of oxidized compound in the liquid culture, followed by recovery of oxidized compound, characterized in that the proteins have been produced by the method in claim 21. 23. Method for oxidation of a selected compound, wherein the method comprises: culturing recombinant cells in the presence of said compound under conditions where the desired oxidation occurs and the oxidized compound accumulates in the liquid culture, followed by recovery of the oxidized compound, characterized in that the recombinant cells are the recombinant host cells according to any of claims 17-20. 24. Method according to claims 22 or 23, characterized in that the oxidation is one selected from the group consisting of: cleavage of the side chain of a sterol compound in pregnenolone; conversion of pregnenolone ti) progesterone; conversion of progesterone to 17α-hydroxyprogesterone; conversion of 17α-hydroxyprogesterone to cortexolone; and conversion of cortexolone to hydrocortisone, 25. Method according to claim 24, characterized in that the oxidation is cleavage of the side chain of a sterol compound resulting in pregnenolone. 26. Method according to claim 24, characterized in that the oxidation is 17<x-hydroxylation to progesterone. 27. Method according to claim 24, characterized in that at least two oxidations from said group are carried out on the same substrate molecule in one step.