Cloning of the bioA, bioD, bioF, bioC and BioH genes of bacillus spraericus, vectors and transformed cells

Abstract

The present invention relates, in particular, to a DNA sequence corresponding to one of the following genes involved in the chain of biotin biosynthesis in bacteria: bioA gene, bioD gene, bioF, bioC and bioH gene.Vectors containing these sequences enable other microorganisms to be transformed in order to improve the production of biotin.

Description

The present invention relates to the preparation of biotin by fermentation using the recombinant DNA technique.

Biotin is a vitamin which is necessary for man, animals and plants, and for some microorganisms.

It has been isolated from egg yolk and is found in brewer's yeast, cereals and various organs, in free form or combined with proteins.

This vitamin has been, in particular, proposed as a regulator of cutaneous metabolism, in particular in the treatment of seborrheic dermatitis in man.

Apart from the different sources which have been recalled above, biotin is synthesized by certain microorganisms, especially microorganisms of the genus Bacillus such as Bacillus sphaericus.

FIG. 1 shows schematically the chain of biotin biosynthesis from pimelic acid in this type of microorganism. This biosynthetic chain comprises 5 different enzymatic stages in which the genes employed are designated successively bioC, bioF, bioA, bioD and bioB.

The systematic study of the production of biotin from pimelic acid during industrial fermentations has shown that certain microorganisms, especially those belonging to the genus Bacillus sphaericus, hyperproduce biotin vitamers as well as biotin (the different intermediates leading to biotin will be referred to hereinafter as "vitamers").

For these bacteria, among biotin precursors, DTB (desthiobiotin) represents the predominant compound which is produced in much larger amounts than the final molecule, namely biotin.

There is a strong control of transcription repressing biotin synthesis which depends on the amount of biotin present in the culture medium; this repression takes place in E. coli as well as in Bacillus sphaericus.

However, no feedback inhibition regulates biotin biosynthesis in E. coli and a control of this kind has never been described in Bacillus sphaericus.

The full explanation for the production of a very large amount of DTB (for a small amount of biotin) from pimelic acid in Bacillus sphaericus still requires a very substantial molecular biological study to be carried out on the organization and regulation of the pathway of biotin biosynthesis in this bacterium.

Preliminary studies of some of the extracted and semi-purified biosynthetic enzymes from strains of B. sphaericus producing DTB do not reveal significant differences from the biotin enzymes that are well known in E. coli.

Nevertheless, the production of biotin by industrial fermentation of such strains of Bacillus sphaericus (IFO 3525, NCIB 9370) could become competitive with the existing chemical processes if the yield of biotin was improved. In order to achieve this object, it would be advantageous to obtain a constitutive hyperexpression of all the biosynthetic genes for biotin in these strains.

The selection of derepressed strains of E. coli, either by their resistance to biotin analogs such as alphadehydrobiotin or by the selection of mutants that hypersecrete biotin, has of course been described. "Cis-dominant" mutations which act pleiotropically on the synthesis of all the biotin genes organized into a bipolar operon have also been described. Trans-acting mutations have also been mentioned although, in addition to their ability to abolish the control of the transcription of the biotin genes, they frequently have pleiotropic properties which may affect the general physiology of the cell.

These methods of traditional microbiological selection may also be applied to the strains of Bacillus sphaericus producing DTB if the assumption is made that the molecular control of biotin biosynthesis and the general organization of the biotin genes are the same as in E. coli.

Over and above all these assumptions, the use of mutant strains in large-scale industrial fermentations and in continuous culture requires the selection of mutants that show a very low level of reversion, and this represents a difficult task given the absence at present of methods of fine genetic analysis in Bacillus sphaericus.

Another approach could consist in cloning all the genes involved in the chain of biotin biosynthesis from a strain of Bacillus sphaericus producing DTB, and then in modifying in vitro the 5' control region so as to abolish the control of transcription.

In addition, the cloning and manipulation of these genes in vitro would enable their general organization to be studied and their in vivo transcription to be improved, using genetic engineering techniques which are known, involving especially the use of strong promoters which are known to be functional in strains of Bacillus sphaericus.

The reintroduction of all these hyperexpressed genes which would no longer be regulated by biotin (but which would be regulated in an inducible manner using, for example, induction by an increase in temperature or by a low-cost substrate) into the original strain of Bacillus sphaericus could then be carried out using the known techniques of transformation, transduction or conjugation-mobilization.

It is also possible to envisage the introduction of these genes into different microorganisms which are known to be acceptable in the foodstuffs industry and which are permeable to pimelic acid, for example Bacillus subtilis, Saccharomyces cerevisiae or strains of Pseudomonas.

Finally, the stabilization of this genetic information in the microorganisms could be achieved by a directed integration in the chromosomes or by a self-selection of plasmids as described, for example, in French Patent No. 84/12,598.

The cloning and characterization of the bioB gene of Bacillus sphaericus have already been described in Japanese Patents A-166,992/85 of 29th July 1985, A-66,532/86 of 25th Mar. 1986 and A-95,724/86 of 24th Apr. 1986.

More especially, the present invention relates to the DNA sequences coding for the enzyme produced by one of the following genes involved in the chain of biotin biosynthesis in bacteria:

bioA gene bioD gene bioF gene bioC gene bioH gene.

Some these DNA sequences according to the invention are linked in the form of a cluster with the bioB gene which had previously been identified, and the whole group hence covers most of the chain of biotin biosynthesis.

Among the DNA sequences of interest, the DNA sequences which code for the enzymes produced by the following genes should be mentioned:

bioB and bioD bioB, bioD and bioA bioB, bioD, bioA and bioF.

This type of sequence used in suitable vectors enables biotin to be prepared from its different vitamers.

The latter can, as will be described below, also be prepared by fermentation using vectors which express the DNA sequences which code for the enzyme or enzymes produced by the following genes:

bioF and bioC bioF and bioH bioF, bioC and bioH, as well as the sequences which code, in addition to the above sequences, for the products of the following genes: bioA bioA and bioD, or bioA, bioB and bioD.

Although the abovementioned DNA sequences may be of various origins, it is preferable to use the sequences originating from a strain of Bacillus, especially a strain of Bacillus sphaericus.

As stated above, it is especially advantageous that these DNA sequences should be devoid of the natural sequences providing for the control of the transcription of the enzymes involved in the pathway of biotin biosynthesis in the original bacterium, in order to abolish the natural regulation due to biotin and to place these DNA sequences under the control of chosen elements which will provide for their efficient transcription in the host strain.

The invention relates, in particular, to all or part of the sequences which are shown in FIGS. 4 to 8 and 17 to 19 and which code for one of the genes mentioned above.

The DNA sequences according to the present invention may be used in different ways.

Preferably, the DNA sequences in question will be carried by a plasmid vector capable of providing for the transformation of a bacterium and containing all the elements providing for the expression of the corresponding genes.

This plasmid may, as has been stated, be of the autonomous and self-replicating type or alternatively, on the other hand, it may be arranged so as to provide for its integration in the chromosome of the host strain.

For this purpose, the different techniques to be employed are known or will be described in the examples.

Thus, in the case of an integration vector, the latter should contain at least one sequence homologous with a sequence present in the genome of the strain to be transformed. thereby enabling chromosomal integration to be ensured. Either a homologous sequence corresponding to a natural genomic sequence or a homologous sequence introduced by another plasmid integration vector may obviously be used.

When the vector is of the autonomous and self-replicating type, it will contain an origin of replication that is effective in the host cell.

Similarly, these different plasmid vectors may contain elements providing for selection, such as a gene for resistance to an antibiotic and/or a marker gene, under the control of a promoter of the strain to be transformed.

Among cells which may be used as a host strain, there should be mentioned, more especially, bacteria, in particular of the genera Escherichia, Bacillus and Pseudomonas, as well as yeasts, especially yeasts of the genus Saccharomyces.

Among host cells that are especially advantageous, Bacillus sphaericus, Bacillus subtilis and Excherichia coli should be mentioned.

Finally, the most especially advantageous strains for transformation by the DNA sequences according to the present invention are the strains which have already been transformed by vectors providing for the expression of other genes involved in this biosynthetic pathway, that is to say the F, A, D and B genes as described in the present application.

Under these conditions, the introduction of the bioC and bioH genes into the bacterium enables this bacterium to synthesize biotin from the first vitamer of the chain, namely pimelate, and this offers considerable economic advantage at the industrial level.

In the context of the present invention, the plasmid integration vector is more especially plasmid pTG475, which will be described below and which contains, in particular, an inducible promoter.

When the vector contains an autonomous origin of replication, the DNA sequences coding for the enzymes mentioned above will preferably be flanked by control elements providing for their expression in the host strain; these will comprise, in particular, at the 5' end, a strong promoter that is effective in the said strain and, where appropriate, other elements such as a termination sequence when the host strain is a yeast or any other microorganism in which such a sequence is necessary.

The present invention also relates to the cells transformed by the vector plasmids according to the invention. Among these cells, there should be mentioned, more especially, bacteria, in particular of the genera Bacillus, Escherichia or Pseudomonas, but also yeasts, especially of the genus Saccharomyces.

Among the strains which may be transformed by these vectors, strains which already produce biotin or one of its vitamers should be mentioned.

Finally, the present invention relates to a method for preparing biotin, wherein a growth medium containing at least pimelic acid, or one of the vitamers of biotin, is fermented with cells as described above which are permeable either to pimelic acid or to the said vitamers of biotin, and wherein the biotin produced is recovered.

The method according to the invention can be carried out in the form of different variants.

In particular, it is possible to prepare the vitamer in situ using cells transformed with a vector according to the invention; in particular a transformed strain containing the bioF, bioH and bioC genes may be capable of producing KAPA from pimelic acid, the conversion of this KAPA to biotin being accomplished by a strain carrying the complementary genes D, A, B for example.

It is possible to arrange for two successive fermentations, or alternatively a co-fermentation if the strains are mutually suited thereto.

It is also possible to arrange for the complementation of a strain which possesses only part of the genes in question, or alternatively to transform a strain which already produces biotin so as to make it overproductive.

Other characteristics and advantages of the present invention will emerge on reading the examples described below with reference to the figures, wherein:

FIG. 1 shows the chain of biotin biosynthesis,

FIG. 2 shows schematically plasmid pTG1400,

FIG. 3 shows schematically plasmid pTG475,

FIG. 4 shows the non-coding sequence upstream from the No. 1 LORF sequence,

FIGS. 5, 5-1 and 5-2 show the No. 1 LORF sequence corresponding to the bioD gene,

FIGS.6, 6-1, 6-2, and 6-3 show the No. 2 LORF sequence corresponding to the bioA gene,

FIGS. 7 and 7(cont'd) show the No. 3 LORF sequence corresponding to the Y gene,

FIGS. 8, 8-1, 8-2 and 8-3 show the No. 4 LORF sequence corresponding to the bioB gene,

FIG. 9 show schematically the study of complementation of pTG1400,

FIG. 10 shows schematically the study of complementation between different plasmids,

FIG. 11 shows schematically the structure of plasmid pTG1418,

FIG. 12 shows schematically the test of complementation of pTG1418,

FIG. 13 shows the restriction map of the insert of B. sphaericus used in the following plasmids:

FIGS. 14A and B show schematically plasmids pTG1418 and pTG1420,

FIGS. 15A and B show schematically plasmids pTG1422 and pTG1435,

FIGS. 16A and B show schematically plasmids pTG1436 and pTG1437,

FIGS. 17A, 17A(cont'd), and 17B show the LORF X sequence,

FIGS. 18A and 18B show the LORF W sequence,

FIGS. 19A, 19A(cont'd), 19B, and 19C show the LORF F sequence,

FIG. 20 shows schematically plasmid pTG1440,

FIG. 21 shows the restriction map of the insert of plasmid pTG1418,

FIGS. 22 and 23 show the different plasmids derived from pTG1418 that are used in the complementation tests.

For reasons of simplification, the DNA sequences and the structures of the plasmids have been shown in the attached drawings; it is nevertheless understood that they are to be considered as forming an integral part of the present description.

EXAMPLE 1

Cloning of the bioA and BioD genes of Bacillus sphaericus IFO 3525 by complementation and demonstration of their linkage with bioB

a) In E. coli

Bacillus sphaericus IFO 3525 is cultured in 200 ml of PAB culture medium (DIFCO Bacto antibiotic medium 3, 17.5 g/l) at 37.degree. C. for 17 hours. The bacteria are recovered by centrifugation and the whole DNA is then extracted from the cells by Saito's method (Saito et al. BBA 1963, 72, 619-629). A quantity of 450 .mu.g of pure DNA is obtained.

20 .mu.g of the whole DNA is completely restricted with HindIII (3 U/.mu.g of DNA). pBR322 is treated with alkaline phosphatase after being completely digested with HindIII.

The hybrid recombinant plasmids are obtained by mixing the genomic DNA, digested with HindIII (2 .mu.g), and pBR322, treated as above (1 .mu.g), with 2 units of T.sub.4 ligase (Boehringer Mannheim) in 50 .mu.l of reaction buffer containing 30 mM NaCl, 30 mM Tris-HCl pH 7.5, 10mM MgCl.sub.2, 0.2 mM EDTA pH 8, 2 mM DTT, 0.5 mM ATP pH 7 and BSA 0.1 mg/ml. The incubation is performed at 14.degree. C. for 16 hours. Aliquots of the ligation mixture are then added in a transformation experiment [Cohen et al. (1972) PNAS 69, 2110-2114], using E. coli strain C600 r.sub.K -m.sub.K + and selecting the strains for their resistance to ampicillin (100 .mu.g/ml) on LB medium.

4 different pools of plasmid DNA are then extracted, each corresponding to an average of 10.sup.4 individual clones on the transformation dishes.

Different bio mutants of E. coli are then transformed with these DNA pools and the transformants are selected either in the presence of ampicillin (100 .mu.g/ml) on LB medium or for resistance to this antibiotic, and for prototrophy for biotin at the same time (LB medium +ampicillin 100 .mu.g/ml+avidin 0.2 U/ml).

The results observed are collated in Table 1:

TABLE 1______________________________________ Selection on Amp Amp. selection medium + avidinGenotype of the Transformants/.mu.g 0.2 U/ml Trans-E. coli strain of DNA formants/.mu.g of DNA______________________________________C268* .DELTA. bioA, his >10.sup.3 3 (pool No. 4) (for each pool) 1 (pool No. 1)C173* .DELTA. bioD, his >10.sup.3 2 (pool No. 4) (for each pool)______________________________________ *Cleary and Campbell (1972) J. Bacteriol. 112, 830.

The plasmids are isolated from the clones selected on ampicillin+avidin and analyzed using restriction enzymes. Three plasmids (2 originating from the strain C268 and 1 from the strain C173) contain a 4.3-kb HindIII insert with a BglII site and 2 SphI sites and without a BamHI, SalI, PstI, EcoRV, PvuII or AvaI site.

With one of these plasmids, designated pTG1400, the restriction map of which is shown in FIG. 2, it is possible to retransform, by selecting for resistance to ampicillin and prototrophy for biotin, equally well the strains C268 (.DELTA.bioA), C173(.DELTA.bioD), R877(bioD19) (Cleary and Campbell, 1972) or C162(bioB) with an average frequency corresponding to that obtained for the selection with ampicillin alone (more than 10.sup.3 per .mu.g of DNA). No complementation of the auxotrophy for biotin of the strains R878 (bioC23) or R901 (.DELTA.bioA-D) (Cleary and Campbell, 1972) can be obtained using pTG1400.

b) In Bacillus subtilis

Complementation of the bio mutants of Bacillus subtilis could prove difficult since it is known that deletions can occur at a very high frequency in foreign inserts cloned into the usual replicative plasmids of Bacillus subtilis. A new strategy, based on complementation tests using a non-replicative plasmid, was thus developed. The integration of non-replicative plasmids in the genomic DNA of Bacillus subtilis takes place at a fairly high frequency (approximately 10.sup.4 transformants/pg of DNA), using naturally competent Bacillus subtilis cells, if there are homologous regions between the genomic DNA and the plasmid.

The first stage consists in integrating plasmid pTG475, whose structure is shown in FIG. 3, in different bio mutants of Bacillus subtilis.

Plasmid pTG475 contains the XylE gene which codes for the enzyme C 2,3-oxygenase [Zukowski et al. (1983) PNAS USA, 80, 1101-1105], which may be used as a chromogenic (yellow) marker and which is expressed under the control of the inducible promoter of the levansucrase gene of Bacillus subtilis. This plasmid also contains a CAT gene conferring resistance to chloramphenicol; the CAT and XylE genes are inserted into pBR322.

This plasmid is integrated in the chromosome of the following strains of Bacillus subtilis:

strain bioA: JKB 3173 [bioA 173, aro G932; CH Pai (1975) J. Bacteriol. 121, 1-8], strain bioB: BGSC1A92 [bioB 141, aro G932 Sac A 321, Arg A2; CH Pai (1975) J. Bacteriol. 121 1-8], strain bio112: JKB 3112 [bio 112; CH Pai (1975) J. Bacteriol. 121, 1-8], by the competent cell transformation technique of R.J. Boyland (1972), J. Bacteriol. 110, 281-290, the selection being performed on TBAB (DIFCO Blood tryptose agar base) plus chloramphenicol 3 .mu.g/ml.

Various checks are performed on the transformed clones: they turn yellow when induced with sucrose and, in addition, a check by Southern hybridization for the strains bioA, bioB and bio112 shows that integration of pTG475 by simple recombination in the promoter of the levansucrase gene has taken place. These transformed strains of Bacillus subtilis are referred to as bioA TG1, bioB TG2 and bio112 TG3. Since plasmid pTG475 has carried pBR322 sequences into the genome of Bacillus subtilis, it becomes possible to integrate, by homologous recombination, any foreign plasmid containing these same sequences.

In a second stage, plasmid pTG1400, previously cloned by complementation into E. coli, was used for transforming Bacillus subtilis strains bioA TG1, bioB TG2 and bio112 TG3. The selection is performed on LB medium +avidin 0.2 U/ml +chloramphenicol 10 .mu.g/ml.

It is observed that it is possible to select, at a very high frequency, transformants which are prototrophic for biotin from strains bioA TG1 and bioB TG2 but not with the strain bio112 TG3. pTG1400 hence does not complement the bio112 mutation.

c) Final characterization of the HindIII insert . of pTG1400

Southern hybridization experiments were performed so as to detect the same HindIII fragment in the genomic DNA of Bacillus sphaericus IFO 3525, corresponding to the pTG1400 insert.

Under drastic hybridization conditions (50% formamide, 0.6% Denhardt's solution, 0.1% SDS, 3.times.SSC at 42.degree. C.), and using a plasmid pTG1400 labelled with .sup.32 P by incorporation of radioactive nucleotides by in vitro polymerization (nick translation) (2.5.times.10.sup.7 cpm/.mu.g of DNA), a single 4.3-kb HindIII band could be visualized in the genomic DNA of Bacillus sphaericus IFO 3523 after 6 hours' autoradiography at -80.degree. C. It was verified that, under these hybridization conditions, pBR322 did not give a cross reaction with the genomic DNA of Bacillus sphaericus IFO 3525. Under these same conditions, no positive reaction, with pTG1400 as probe, could be detected in the HindIII-treated genomic DNA of Bacillus subtilus BGSC1A289 or in the HindIIl treated genomic DNA of E. coli C600.

The whole DNA sequence of the 4.3-kb HindIII fragment was analyzed using the "Shotgun" method, that is to say the systematic cloning of the fragments obtained by sonification, or using "cyclone deletions" or elongation with oligonucleotide primers.

For the "shotgun" method, plasmid pTG1400 was broken up by ultrasound treatment. After treatment of the DNA segments with phage T.sub.4 DNA polymerase, the blunt-ended fragments are allowed to migrate on a low-melting point agarose gel and fragments approximately 300 bp in size are isolated.

The cloning of these fragments is carried out in M13 vectors digested with SmaI and treated with phosphatase.

The clear plaques which do not give cross hybridization with pBR322 are screened and 100 clones are sequenced. The results were analyzed by computer.

A recent method for producing a series of overlapping clones (the cylone system) was used for sequencing the DNA [R.M.K. Dale, B.A. Mc Clure, J.P. Houchins (1985) Plasmid 13, 31-40]. This method was conducted in parallel with the "shotgun" method so as to confirm the results. In addition, this method produces defined deleted plasmids containing groups of bio genes or isolated bio genes.

The complete sequence of the HindIII fragment of pTG1400is shown in FIGS. 4, 5, 6, 7 and 8.

Computer analysis of this sequence reveals that the fragment has the capacity to code for four long open reading-frames (LORF). The possible translation initiation sites and the Shine-Dalgarno regions are underlined. A palindromic region is underlined at the 5' end of the sequence which might represent a transcription termination site.

Detailed complementation analysis shows that the first LORF region (FIG. 5) (with 3 possible translation initiation sites) corresponds to the bioD gene.

Experiments known under the name of "maxicells" were carried out with E. coli strain CSR 603 (recA1, phr-1, uvrA6, thr-1, leu-6, thi-1, argE3, lacY1, galK2, ara14, xy115, mt11, proA2, str-31, tsx-33, supE44, F.sup.-, lambda.sup.-); they show that this region codes for a polypeptide with an apparent molecular weight of approximately 25 kd.

The second LORF region (FIG. 6) (with 4 possible translation initiation sites) corresponds to the bioA gene. A "maxicell" experiment with E. coli CSR603 reveals a polypeptide having an apparent molecular weight of approximately 40 kd which corresponds to the product of the bioA gene.

It was not possible to determine the function of the third LORF sequence, referred to as the Y gene (FIG. 7). A very great hydrophobicity and the presence of a probable signal sequence in the coding region suggest that this LORF, if it is transcribed and translated, codes for a protein that interacts with the membrane. The fact that this Y gene is in a cluster with the other bio genes (A, D and B) suggests that it codes for another function involved in biotin metabolism.

The fourth LORF region (FIG. 8) (with 3 possible translation initiation sites) has already been identified: it is a region coding for the bioB gene. It is demonstrated here that this gene is linked in a cluster with the bioA and bioD genes.

A complementation analysis with plasmids containing subcloned regions of the 4.3-kb HindIII insert of pTG1400 shows clearly, as is seen in FIG. 9, that the first LORF region corresponds to the product of the D gene and that the second LORF region corresponds to the product of the A gene.

EXAMPLE 2

Cloning by complementation of the bioF gene of Bacillus sphaericus IFO 3525

The HindIII genomic DNA library of Bacillus sphaericus IFO 3525, described above, was used for transforming a mutant of E. coli, bioF 12, following the methodology described above.

The results obtained are collated in Table 2:

TABLE 2______________________________________ Selection on Amp Amp. selection medium + avidinGenotype of the Transformants/.mu.g 0.2 U/ml Trans-E. coli strain of DNA formants/.mu.g of DNA______________________________________R874* bioF 12, his >10.sup.4 2 (pool No. 1) (for each pool) 2 (pool No. 3) 4 (pool No. 4)______________________________________ *Cleary and Campbell (1972) J. Bacteriol. 112, 830

The plasmids are isolated from clones selected on medium containing ampicillin and avidin, and analyzed using restriction enzymes.

Two of these plasmids contain two approximately 4.5- and 0.6-kb HindIII inserts with SphI, KpnI, HpaI, NdeI, AvaI, ClaI, BglI, XmnI, PvuII, ScaI and StuI sites and without a BglII, XbaI, SmaI, PstI, SalI, NruI, BamHI, PvuI, EcoRI, HindII, EcoRV, FspI, ApaI, BalI or AatII, etc., site.

With one of these plasmids, Ptg1418, it was possible to retransform, by selecting for resistance to amoicillin and prototrophy for biotin, E. coli strain R874 (bioF12, his) with an average frequency corresponding to that obtained for the selection of the clones on ampicillin (more than 10.sup.4 /.mu.g of DNA).

Subcloning experiments showed that only the 4.5-kb HindIII fragment codes for the KAPA synthetase function providing for the complementation of the bioF12 mutation of E. coli strain R874.

"Southern" experiments were performed so as to detect the 4.5-kb HindIII fragment in the genomic DNA of Bacillus sphaericus IFO 3525 compared with the 4.5-kb insert of pTG1418. Using very drastic hybridization conditions (50% formamide, 3.times.SSC, 0.1% SDS, 0.6% Denhardt's solution, 42.degree. C.) and using M13TG1425 [phage M13 containing the 4.5-kb HindIII insert labelled with .sup.32 P by incorporation of radioactive nucleotides by in vitro polymerization (nick translation) with 2.times.10.sup.6 cpm/.mu.g of DNA], an approximately 4.5-kb HindIII band may be visualized in the genomic DNA of Bacillus sphaericus IFO 3525 after 7 hours' autoradiography at -80.degree. C.

Under the same conditions, no positive reaction can be detected either on the HindIII-treated genomic DNA of Bacillus subtilis BGSC1A289 or on the HindIII-treated genomic DNA of E. coli C600.

No cross reaction can be detected between the pTG1400 insert and pTG1418 inserts either with an 8.3-kb SphI fragment of pTG1406 overlapping the 3' end of pTG1400 or with an 8.2-kb MboI fragment of pSB01 overlapping the 5' end of pTG1400 (see FIG. 10).

The restriction map of the pTG1418 inserts was analyzed and is shown in FIG. 11.

Complementation studies and "Southern" analysis demonstrate that the bioF gene of Bacillus sphaericus IFO 3525 is not linked to the bioA, bioD and bioB genes of the same microorganism (FIG. 12).

EXAMPLE 3

Complementation of B. subtilis mutant bio112TG3 (derived from JBK 3112, bioC/F-112 aroG932) by integrative transformation with pTG1418 and different derivatives of the latter

B. subtilis mutant bio112 was identified by nutritional tests as being affected in the bioF or bioC function (C.H. Pai, 1975, J. Bacteriol 121, 1-8). In this mutant, plasmid pTG475 has been integrated at the level of the sacR - sacB locus following the methodology described above, the new strain thereby obtained being designated bio112 TG3.

Transformation of the strain bio112 TG3 by different plasmids (pTG1418, 1420, 1422, 1435, 1436 and 1437, shown in FIGS. 13 to 16) was performed, followed by selection on LB medium +chloramphenicol 10 .mu.g/ml+avidin 500 U/1. The results of complementation are summarized in Table 3. Since the cross test with E. coli mutant R874(bioF, his) gives the same result, it is probable that B. subtilis mutant bio112 is affected in the bioF gene. By analyzing the complementations obtained in terms of the plasmids used, it is possible to localize the bioF gene on the pTG1418 insert at the level of a fragment bounded by the XmnI and NcoI restriction sites (see FIGS. 21 to 23).

TABLE 3______________________________________Integrative transformation ofB. subtilis strain bio112 TG3 Complementation on LB + 500 U/l avidin + 10 .mu.g/ml chloram-Plasmid Insert phenicol______________________________________pTG1418 5.1-kb HindIII insert ++pTG1422 3.1-kb ClaI insert con- ++ taining bioF and part of LORF WpTG1420 2-kb ClaI-HindIII insert - containing X and part of LORF WpTG1436 1.3-kb XmnI-NcoI insert ++ containing the bioF genepTG1437 1.3-kb NcoI-XmnI insert ++ containing the bioF genepTG1435 0.6-kb HpaI-PvuII insert - containing LORF X______________________________________

EXAMPLE 4

Nucleotide sequence of the 4.53-kb HindIII insert of plasmid pTG1418

The HindIII insert was cloned into M13TG131 at the level of the corresponding site of the polylinker. The so-called "cylone" method was applied to this plasmid M13TG1425 and the results were analyzed by computer. The few reading differences between the two complementary strands were elucidated on sequencing using specific oligonucleotides.

The sequence is given in detail in FIGS. 17 to 19. It is seen that this insert contains three open reading-frames coding, respectively, for proteins (subunits) of molecular weight 18,462, 28,048 and 42,940 daltons. The last gene, with respect to the reading direction, is localized in the region identified as bioF, the molecular weight of this B. sphaericus protein being of the same order of magnitude as that of E. coli KAPA synthetase (M.A. Eisenberg, 1973, adv. Enzymol. 38, 317-372).

The juxtaposition of these three open reading-frames might be, as in the case of the insert of plasmid pTG1400, typical of an operon structure. It should be noted that, upstream from the first gene, a 15-base pair sequence (underlined in FIG. 17) also present upstream from the first gene (bioD) of the insert of plasmid pTG1400 can be identified. This significant characteristic might indicate that the two groups of biotin genes of B. sphaericus are subjected at least to common regulation. The latter might correspond to control by biotin (or a derivative of the latter) as has already been described for the KAPA synthetase (bioF) of B. sphaericus (Y. Yzumi, K. Sato, Y. Tani and K. Ogata, 1973, Agric. Biol. Chem. 37, 1335).

On the 3' side of the last gene of the sequenced insert, a sequence possessing the characteristics of a transcription terminator may be identified (underlined in FIG. 18).

EXAMPLE 5

Complementation of E. coli mutants R878 (bioC, his) and C261 (.DELTA.bioFCD, his), respectively, using plasmids pTG1418 (1433) and pTG1440

The traditional complementation methodology for bio mutants of E. coli are supplied using plasmid pTG1418 and different derivatives of the latter.

When competent cells of E. coli mutant R878 (bioC, his) are transformed with plasmids pTG1418 and pTG1433 (see Table 4), and then plated on LB medium+ampicillin 100 .mu.g/ml+avidin 200 U/l growth is detected after 36 h of incubation at 37.degree. C. The frequency of appearance of the transformed clones on this medium is of the same order of magnitude as that measured on LB medium +ampicillin 100 .mu.g/ml.

TABLE 4______________________________________Transformation of E. coli mutant R878 bioCNumber of transformants per .mu.g of DNA Selection on LB Selection on LB medium + medium + ampicillin ampicillin + avidinPlasmid (100 .mu.g/ml) (100 .mu.g/ml) (200 U/l)______________________________________pTG1418 10.sup.3 10.sup.3 smallpTG1433 10.sup.3 10.sup.3 smallpBR322 10.sup.3 0 (after 36 h of incuba- tion at 37.degree. C.)______________________________________

The growth of the transformed clones, which are normal on LB+ampicillin, is retarded in the absence of biotin (LB medium+ampicillin+avidin; minimum plus casamino acids, devoid of biotin). This complementation of the biotin auxotrophy of the mutant R878 bioC is altogether significant, given the total absence of residual growth of this same mutant when it is transformed by various plasmids derived from pBR322 on medium devoid of biotin.

The two inserts of plasmids pTG1400 and pTG1418 were cloned into pBR322 to give plasmid pTG1440 (FIG. 20). This plasmid pTG1440, when introduced into E. coli mutant C261 (.DELTA.bioFCD, his) enables clones to be selected on LB medium+ampicillin 100 .mu.g/ml+avidin 200 U/1. The frequency of transformation obtained is directly comparable to that measured on LB+ampicillin. Again, the growth of these recombinant clones, which is normal on LB medium+ampicillin, is retarded in the absence of biotin. From these two results (complementation of the biotin auxotrophy of the mutant bioC and bio .DELTA.FCD), it emerges clearly that the insert of plasmid pTG1418 also contains the bioC gene of B. sphaericus. The different subclonings derived from the pTG1418 insert (FIGS. 21 to 23) do not enable complementation of the bioC mutation of E. coli to be obtained. Only the inserts possessing all three genes confer effective complementation of the bioC mutation of E. coli.

EXAMPLE 6

Complementation of E. coli mutant bioH (PA505 MA.DELTA.108, argH, metA, bioH, malA, str.sup.r)

This mutant was originally described as bioB (D. Hatfield, M. Hofnung and M. Schwartz, 1969, J. Bacteriol. 98? 559-567). It was then characterized as not excreting any vitamer and as being capable of growth on inorganic medium in the presence of KAPA, DAPA, DTB or biotin. Eisenberg (1985, Annals New York Academy of Sciences 447, 335-349) then proposed that this gene codes for a subunit of pimeloyl-CoA synthetase (bioH). It should be noted that the E. coli mutants which are overproductive of biotin (selected either by a level of excretion of vitamin permitting the growth of a bioB auxotroph of E. coli, or by resistance to alpha-dehydrobiotin) have all been identified genetically as affected at the bioR locus. This locus codes for a multifunctional protein (repressor of the synthesis of the messenger RNAs of the bioABFCD operon and synthetase holoenzyme binding biotin to a lysine residue of different apoenzymes having a carboxylase function).

The fact that all the E. coli mutants which are overproductive of biotin identified to date are localized in the gene coding for the trans-active repressor and never in the operator of the bioABFCD operon suggests that another gene involved in biotin biosynthesis is subject to this regulation. From a review of the literature, the best candidate is the bioH gene.

Since the pTG1418 insert contains the bioF and bioC genes of B. sphaericus, an investigation was performed as to whether the third gene corresponded to bioH.

Complementation of the bioH mutant of E. coli was effectively obtained using plasmid pTG1433. Once again, the growth obtained on this medium is retarded, but is altogether significant compared with the controls (see Table 5).

TABLE 5______________________________________Transformation of E. coli mutant bioH PA505 MA.DELTA.08Number of transformations per .mu.g of DNA Selection on LB Selection on LB medium + medium + ampicillin ampicillin + avidinPlasmid (100 .mu.g/ml) (100 .mu.g/ml) (200 U/l)______________________________________pBR322 10.sup.4 0pTG1433 10.sup.3 10.sup.3 small (after 36 h of incuba- tion at 37.degree. C.)______________________________________

As in the case of the complementation of the bioC mutant, a necessary and sufficient condition for complementation of the bioH mutant is the simultaneous presence of the three genes of the pTG1418 insert on the plasmids introduced into the strain (FIGS. 21 to 23).

In distinction to the bioF mutants of E. coli and B. subtilis, which are capable of being complemented by a single gene of B. sphaericus, growth of the bioC and bioH mutants of E. coli in the absence of biotin can hence be obtained only when recombinant plasmids carrying the three genes of the HindIII insert of pTG1418 are introduced. This might reflect, inter alia, differences in enzymatic properties between the pimeloyl-CoA synthetase of B. sphaericus and the corresponding enzyme of E. coli (difference in affinity towards the substrates, multienzyme edifice, in particular) or a reduced synthesis of pimeloyl-CoA synthetase of B. sphaericus in E. coli, limiting the metabolic flux of pimelate to KAPA.

EXAMPLE 7

The functional test of the bio FCH genes

The tests of complementation of the bioF, bioH and bioC mutants of E. coli using recombinant plasmids carrying inserts derived from the 4.5-kb HindIII fragement isolated from B. sphaericus may be recognized as evidence of the presence on this DNA of the bioF, bioH and bioC genes.

Nevertheless, it is useful to complement this information by a test of activity of the products of these genes cloned from B. sphaericus. The product of the bioF gene of E. coli was characterized as catalyzing the conversion of pimeloyl-CoA to KAPA (Eisenberg, 1968). The products of the two genes bioC and bioH are not clearly identified at the present time; a hypothesis mentioned in the literature (Eisenberg, 1985) suggests that they code for the proteins involved in the biosynthesis of pimeloylCoA. A test was performed to find out whether the bioF, C, H sequence provides specifically for the conversion of pimelate to KAPA.

The insert carrying the FCH coding portion of B. sphaericus (recovered in the form of an EcoRV-SphI fragment of pTG1434 in combination with the SphI-SphI fragment of pTG1436, ligated together) was fused to the promoter of the gene conferring resistance to tetracycline of pBR322 to give plasmid pTG1446, in order to obtain a significant level of expression of the proteins associated with the bioF, C and H genes.

This plasmid was then introduced into competent cells of E. coli strain bioH. This recombinant strain is then cultured at 37.degree. C. for 48 hours in GP medium (for 1 liter: glycerol, 20 g; proteose peptone, 30 g; vitaminfree casamino acids, 5 g; K.sub.2 HPO.sub.4, 1 g; KCl, 0.5 g; MgSO.sub.4.7H.sub.2 O, 0.5 g; FeSO.sub.4.7H.sub.2 O, 0.01 g; MnSO.sub.4.4H.sub.2 O, 0.001 g, pH 6.8-7; thiamine-HCl, 20 .mu.g) to which ampicillin 100 .mu.g/ml and pimelate (pH 7.5) 0.5 mg/ml are added.

An aliquot of the supernatant (5 .mu.l) is then chromatographed on a silica plate [chromato.solvent: n-butanol (60)/acetic acid (15)/water (25), vol/vol; migration of the solvent: 10 cm]so as to separate the vitamers and the biotin that are produced.

After migration specific to each vitamer, the KAPA is measured quantitatively by biological assay using a Saccharomyces cerevisiae collection strain, namely ATCC 7754.

Under these conditions, a quantity of 85 .mu.g (expressed as biotin equivalents) of KAPA per ml of culture supernatant can be measured reproducibly.

The control in this experiment is the same strain containing plasmid pBR322, cultured under the same conditions. In this case, no significant detection of KAPA can be demonstrated.

It is hence shown here that the pTG1418 insert codes for the enzyme functions which are necessary and sufficient in the conversion of pimelate to KAPA, namely the products of the bioC, H and F genes.

Deposition of strains that are representative of the invention

E. coli strains C600 pTG1400 and R874 pTG1418 were deposited at the Collection Nationale de Cultures de Microorganismes (National Collection of Microorganisms Cultures) of the Institut Pasteur, 28 rue du Docteur-Roux - 75724 Paris Cedex 15, on 26 September 1986 under Nos. I-608 and I-609.

REFERENCES

P.P. Cleary and A. Campbell (1972) J. Bacteriol. 112, 830-839. M.M. Zukowski, D.F. Gaffney, D. Speck, M. Kauffmann, A. Findeli, A. Wisecup and J.P. Lecocq (1983) PNAS USA 80, 1101-1105. C.H. Pai (1975) J. Bacteriol. 121, 1-8. R.J. Boyland, N.H. Mendelson, D. Brooks and F.E. Young (1972) J. Bacteriol. 110, 281-290. R.M.K. Dale, B.A. Mc Clure and J.P. Houchins (1986) Plamid 13, 31-40. S.N. Cohen, A.C.Y. Cheng and L. Hsu (1972) PNAS USA 69, 2110-2114. H. Saito and K.I. Miura (1963) BBA 72, 619-629. M.A. Eisenberg (1985) Regulation of the biotin operon. Annals New York Academy of Sciences, 447, 335-349. M.A. Eisenberg and C. Star (1968) J. Bacteriol., 96, 1291-1297.

Claims

1. A recombinant DNA molecule comprising a sequence encoding a gene product of B. sphaericus, wherein said sequence is selected from the group consisting of: (i) a sequence encoding the product of gene bio A, (ii) a sequence encoding the product of gene bioD, (iii) a sequence encoding the product of gene bio F, (iv) a sequence encoding the product of gene bio H, (v)( a sequence encoding the product of gene bio X, (vi) a sequence encoding the product of gene bio Y, and (viii) a sequence encoding the product of gene bio W.

2. A recombinant DNA molecule according to claim 1 wherein said molecule is devoid of natural sequences controlling the expression of said bio A, bio D, bio F, bio H, bio X, bio Y and bio W genes.

3. A cell comprising said recombinant DNA molecule according to claim 2.

4. The cell according to claim 3 wherein said cell is of the E. coli species.

5. The cell according to claim 3 wherein said cell is of the B. sphaericus species.

6. The recombinant molecule according to claim 1 wherein said molecule is a plasmid.

7. A cell of the E. coli species comprising said recombinant DNA molecule according to claim 1.

8. A cell of the B. sphaericus species comprising more than one copy of said recombinant DNA molecule according to claim 1.

9. A recombinant DNA molecule comprising, in any order, (i) a first open reading frame (ORF) encoding the product of gene bio D of B. sphaericus, (ii) a second ORF encoding the product of gene bio A of B. sphaericus, (iii) a third ORF encoding the product of gene bio Y of B. sphaericus and (iv) a fourth ORF encoding the product of gene bio B of B. sphaericus.

10. The recombinant DNA molecule according to claim 9 wherein said molecule comprises the 4.3 kb Hind III fragment of pTG14000.

11. A cell of the E. coli species comprising said recombinant DNA molecule according to claim 10.

12. The recombinant DNA molecule according to claim 9 wherein said first, second, third and fourth ORF's are devoid of natural sequences controlling the expression of said ORF's.

13. The recombinant molecule according to claim 9 wherein said molecule is a plasmid.

14. A cell of the E. coli species comprising said recombinant DNA molecule according to claim 9.

15. A recombinant DNA molecule comprising, in any order, (i) a first ORF encoding the product of gene bio X of B. sphaericus, (ii) a second ORF encoding the product of gene bio W of B. sphaericus, and (iii) a third ORF encoding the product of gene bio F of B. sphaericus.

16. The recombinant DNA molecule according to claim 15, wherein said molecule comprises the 4.5 kb Hind III fragment of pTG1418.

17. A cell of the E. coli species comprising said recombinant DNA molecule according to claim 16.

18. A cell of the E. coli species comprising said recombinant DNA molecule according to claim 15.

19. The recombinant DNA molecule according to claim 15 wherein said first, second and third ORF's are devoid of natural sequences controlling expression of said ORF's.

20. The recombinant molecule according to claim 15 wherein said molecule is a plasmid.

21. A plasmid selected from pTG1400 and pTG1418.