Production of riboflavin

Abstract

The present invention provides an improved biotechnological production of riboflavin (also referred herein as vitamin B2) through modification in the operon containing the riboflavin biosynthetic genes (rib operon), in particular modifications of/in the leader sequences (rib leader) upstream of the corresponding riboflavin biosynthetic genes (rib operon). Furthermore, the present invention relates to genetically engineered microorganisms carrying said modified sequences, processes to generate said modified sequences/microorganisms and the use thereof for production of riboflavin.

Claims

1. A mutated Bacillus subtilis rib leader sequence of the riboflavin biosynthesis operon operatively linked to at least one downstream rib gene, wherein said mutated Bacillus subtilis rib leader sequence does not comprise a Bacillus subtilis rib leader terminator sequence and does comprise at least one ribO mutation, wherein said Bacillus subtilis rib leader terminator sequence consists of a contiguous sequence of nucleotides 231-263 of SEQ ID NO: 42, and said at least one ribO mutation is selected from the group consisting of a mutation of one of the following positions of SEQ ID NO:42: T31, G39, G40, G41, C55, C85, C86, G88, C93, A116, G121 and C128.

2. A mutated Bacillus subtilis rib leader sequence of the riboflavin biosynthesis operon operatively linked to at least one downstream rib gene and a constitutive promoter, wherein said mutated Bacillus subtilis rib leader sequence operably linked to a constitutive promoter comprises the nucleotide sequence of SEQ ID NO:67, 68, 69, 70 or 71.

3. The mutated Bacillus subtilis rib leader sequence of claim 1 fused to a constitutive promoter.

4. A riboflavin-producing microorganism genetically engineered with a mutated Bacillus subtilis rib leader sequence of claim 1.

5. A riboflavin-producing microorganism according to claim 4 capable of producing at least 5% more riboflavin from a given carbon source compared to production of riboflavin using a wild-type microorganism.

6. A riboflavin-producing microorganism according to claim 4 wherein the accumulation of intact, full-length riboflavin mRNA is improved compared to a wild-type microorganism.

7. A method for production of riboflavin comprising culturing a microorganism with a carbon source under conditions whereby riboflavin is produced by said microorganism, said microorganism comprising the mutated Bacillus subtilis rib leader sequence of claim 1.

8. A process for the production of a microorganism according to claim 4 comprising the steps of: (a) providing a microorganism capable of riboflavin production comprising a rib operon including leader sequences; and (b) genetically engineering said microorganism with a polynucleotide.

9. A process for the production of riboflavin comprising the use of a microorganism according to claim 4 and optionally isolating and/or purifying the produced riboflavin from the reaction mixture.

10. A process according to claim 9 wherein the microorganism is incubated in an aqueous medium under conditions that allow the production of riboflavin from a given substrate.

11. A process for the production of full-length mRNA transcripts from riboflavin biosynthetic genes in a riboflavin-producing microorganism comprising introducing into said microorganism a polynucleotide according to claim 1.

12. The modified polynucleotide sequence according to claim 3 wherein the constitutive promoter is selected from P.sub.veg or P.sub.Spo15.

13. A mutated Bacillus subtilis rib leader sequence of the riboflavin biosynthesis operon operatively linked to at least one downstream rib gene, wherein said mutated Bacillus subtilis rib leader sequence does not comprise a Bacillus subtilis rib leader terminator sequence and does comprise at least one ribO mutation, wherein said Bacillus subtilis rib leader terminator sequence consists of a contiguous sequence of nucleotides 231-263 of SEQ ID NO:42, and said at least one ribO mutation is selected from the group consisting of a mutation of one of the following positions of SEQ ID NO:42: T31, G39, G40, G41, C55, C85, C86, G88, C93, A116, G121 and C128, and wherein said mutated rib leader sequence leads to relaxation or reduction of a repressive effect of a non-modified Bacillus subtilis rib leader sequence on expression of a downstream rib gene when said mutated rib leader sequence is expressed in a riboflavin biosynthesis operon as compared to expression of said non-modified Bacillus subtilis rib leader sequence in the riboflavin biosynthesis operon.

14. A mutated Bacillus subtilis rib leader sequence of the riboflavin biosynthesis operon operatively linked to at least one downstream rib gene, wherein said mutated Bacillus subtilis rib leader sequence does not comprise a Bacillus subtilis rib leader terminator sequence and does comprise at least one ribO mutation, wherein said Bacillus subtilis rib leader terminator sequence consists of a contiguous sequence of nucleotides 231-263 of SEQ ID NO:42, and said at least one ribO mutation is selected from the group consisting of a mutation of one of the following positions of SEQ ID NO:42: G39, G40, G41, C85 and G121.

Description

BRIEF DESCRIPTION OF THE DRAWINGS

(1) FIG. 1 shows the rib leader of Bacillus subtilis including 5 and 3 flanking regions thereof. The rib leader is located between nucleotide positions 562 to 857. (A) wild-type rib operon wherein 5 regulatory elements are indicated in bold (ttgcgt and tataat, respectively) as well as the transcription start (A) corresponding to the first nucleotide of the rib leader as shown in SEQ ID NO:42. (B) rib leader wherein the positions of ribO mutations such as triple ribO, RK41 and RK1a are indicated in bold. (C) to (I) rib leader wherein the position of flank-right, terminator, stem loop-right, stem loop-left, SWITCH deletion, del mro175, and the entire rib leader, respectively, is indicated through underlining. For more explanation see text, in particular the examples. FIGS. 1A-B disclose SEQ ID NOS 93-94, respectively and FIGS. 1C-I each disclose SEQ ID NO: 93.

(2) FIG. 2 shows the different steps for constructing a genetically altered microorganism carrying an mRNA stabilizing element (St) downstream of a strong constitutive P.sub.spo15 promoter via LFH-PCR which is located at the 5 end of the rib operon comprising the genes ribD, ribE, ribA. For selection of recombinant microorganisms, the chloramphenicol (Cm) antibiotic resistance cassette is used. (A) Replacement of the native promoter with a strong promoter using LFH-PCR. (B) Introduction of mRNA stabilizing elements. For more explanation see the examples.

(3) The following examples are illustrative only and are not intended to limit the scope of the invention in any way. The contents of all references, patent applications, patents and published patent applications, cited throughout this application are hereby incorporated by reference, in particular EP 405370, WO 04/106557, WO 07/051,552 and EP1186664.

EXAMPLES

(4) The following media as referred to in the examples are described in WO 04/106557: Tryptose Blood Agar Broth (TBAB) medium, Veal infusion-Yeast Extract broth (VY) medium, 10 Spizizen salts and Minimal Medium (MM). Additionally, the following media have been used:

(5) 100 Trace elements solution A: 12.5 g MgCl.sub.2.6H.sub.2O; 0.55 g CaCl.sub.2; 1.35 g FeCl.sub.2.6H.sub.2O; 0.1 g MnCl.sub.2.4H.sub.2O; 0.17 g/l ZnCl.sub.2; 0.043 g CuCl.sub.2.2H.sub.2O; 0.06 g CoCl.sub.2.6H.sub.2O; 0.06 g Na.sub.2MoO.sub.4.2H.sub.2O; ad 11H.sub.2O, autoclaved.

(6) 5 Minimal Salt Solution: 0.057 M K.sub.2SO.sub.4; 0.31 M K.sub.2HPO.sub.4.5H.sub.2O; 0.22 M KH.sub.2PO4; 0.017 M Na-citrate.7H.sub.2O; 0.004 M MgSO.sub.4.H.sub.2O, pH 7.0, autoclaved.

(7) 100 Trace elements solution B: 0.55 g CaCl.sub.2; 0.17 g ZnCl.sub.2; 0.043 g CuCl.sub.2.2H.sub.2O; 0.06 CoCl.sub.2.6H.sub.2O; 0.06 g Na.sub.2MoO.sub.4.2H.sub.2O; ad 11H.sub.2O, autoclaved.

(8) Riboflavin screening medium (RSM): 200 ml 10 Spizizen salts; 10 ml 100 Trace elements solution A; 2 ml 50% glucose; 36 ml 25% raffinose; 10 ml 10% yeast extract; ad 11H.sub.2O.

(9) Spizizen Minimal Medium (SMM): 100 ml 10 Spizizen salts; 10 ml 50% glucose; 1 ml 40% sodium glutamate; 10 ml trace element solution A; ad 11H.sub.2O.

(10) Riboflavin production in shake flasks were performed as follows: strains were inoculated from frozen glycerol stocks in 5 ml VY rich medium and grown overnight at 37 C. with an agitation of 280 rpm. Cells were collected by centrifugation and resuspended in 1 ml RSM (see above). 250 l of the cell suspension was used for inoculation of 25 ml RSM in 250 ml baffled shake flasks. After 48 h incubation at 39 C. with an agitation of 220 rpm, 500 l culture were taken and 35 l 4 N NaOH were mixed with the sample for 1 min at room temperature allowing to dissolve the riboflavin crystals. Samples were neutralized by the addition of 465 l 1 M potassium phosphate buffer (pH 6.8) and pelleted by centrifugation (5 min, 13200 rpm). The supernatant was used for HPLC determination of the concentrations of riboflavin and two side products: 6,7-dimethyl-8-ribityllumazine (DMRL) and oxolumazine. In addition, a second culture sample was taken and after centrifugation (5 min, 13200 rpm) the supernatant was used for the determination of the concentrations of the residual glucose and raffinose in the medium for calculation of the riboflavin yield on carbon source.

(11) Samples from shake flask cultures were analyzed by HPLC. Chromatography was carried out on an Agilent 1100 HPLC system equipped with a thermostatted autosampler, a diode array and a fluorescence detector. The separation was performed on a Supelcosil LC-8 DB-5 column (150 mm4.6 mm) equipped with a 4 mm LC-8 DB guard column. A mixture of 0.1 M acetic acid and methanol was used as mobile phase. Gradient elution was applied starting at 2% methanol (constant for 5 min) and going up to 50% methanol in 15 min. The column was kept at 20 C. The signal was recorded by UV at 280 nm. Riboflavin was well separated from the impurities (e.g. side products: DMRL and oxolumazine) and eluted at 15.2 min. The calibration is based on reference material obtained from Fluka. The method is calibrated from 10 g/ml to 1 mg/ml riboflavin.

(12) Additionally, the concentration of glucose and raffinose in the culture broth was analyzed by an Agilent 1100 series HPLC system using a quaternary pump, an autosampler a UVand a refractive index detector. The separation was achieved on a CAPCELL PAK NH2 UG80 column (4.6 mm250 mm, 5) (Shiseido). The optimal column temperature was 35 C. The mobile phase was a mixture of acetonitrile and DI water at a 65/35 ratio. The flow rate was 1.0 ml/min and the injection volume set to 5 l. The refractive index signal was monitored and used for detection. The calibration range for each compound is from 0.5 mg/ml to 30 mg/ml.

Example 1

Generation of Riboflavin Auxotrophic Strains

(13) For engineering of the original leader and promoter sequence of the riboflavin operon of B. subtilis, the riboflavin promoter, the 5 leader sequence and the 5 part of ribD (ribG) coding for the deaminase domain of RibD was replaced by a neomycin resistance (neo) cassette obtained from plasmid pUB110 (Itaya et al., 1989, Nucleic Acid Res. 17:4410) generating the riboflavin-auxotrophic strain B. subtilis BS3813. Genomic DNA derived from B. subtilis strain 1A747 (SP.sup.c, prototroph), which is a derivative of B. subtilis 168 (trpC2), has been obtained from the Bacillus Genetic Stock Center, The Ohio State University, Columbus, Ohio 43210 USA.

(14) For strain construction, Long Flanking Homology Polymerase Chain Reaction (LFH-PCR) was used to generate DNA fragments containing the 1236 bp neo resistance cassette flanked with the 526 bp upstream region of the native P.sub.rib promoter (flank 5) and the 502 bp 3 end of ribD gene (flank 3). Therefore, 3 DNA fragments flank 5, the neo resistance cassette and flank 3 were first amplified by PCR. DNA fragments flank 5 and flank 3 were generated as follows: 1 l of a 100 M solution of primers p50 (SEQ ID NO:6) together with p51 (SEQ ID NO:7) or primers p44 (SEQ ID NO:4) together with p45 (SEQ ID NO:5) were added to 0.1 g B. subtilis 1A747 chromosomal DNA in a 50 l reaction volume containing 1 l of 10 mM dNTP's, 5 l of 10 buffer and 0.5 l Pfu polymerase (Stratagene). For generating DNA fragment containing the neo resistance cassette, 1 l of a 100 M solution of primers p9 (SEQ ID NO:2) together with p10 (SEQ ID NO:3) were added to 0.05 g pUB 110 DNA containing the neo resistance cassette in a 50 l reaction volume containing 1 l of 10 mM dNTP's, 5 l of 10 buffer and 0.5 l Pfu polymerase (Stratagene). The PCR reactions were performed in 35 cycles of three sequential steps: (i) denaturing step at 94 C. for 30 sec; (ii) annealing step at 52 C. for 30 sec; (iii) elongation step at 72 C. for 1 min. The PCR cycles were preceded by a denaturation step at 95 C. for 3 min. The three PCR products were separated by agarose gel electrophoresis and extracted from the gel using the MinElute Gel Extraction Kit (Qiagen). In the final LFH-PCR reaction, the three purified PCR products (flank 5, neo resistance cassette and flank 3) were assembled: 1 l of a 100 M solution of primers p45 together with p51, 1 l flank 5 PCR product (50 ng), 1 l flank 3 PCR product (50 ng) and 1 l neo resistance cassette (100 ng) were added to give a final reaction volume of 50 l containing 1 l of 10 mM dNTP's, 5 l of 10 buffer and 0.5 l Pfu polymerase (Stratagene). The LFH-PCR reaction was performed in 35 cycles of three sequential steps: (i) denaturing step at 94 C. for 30 sec; (ii) annealing step at 52 C. for 30 sec; (iii) elongation step at 72 C. for 2.5 min. The PCR cycles were preceded by a denaturation step at 95 C. for 3 min. The assembled LFH-PCR product was purified by using the QiaQuick PCR purification kit (Qiagen). The purified LFH-PCR product (2 g) was used for transformation of competent B. subtilis 1A747 cells. Neomycin-resistant (Nm.sup.r) transformants were selected on TBAB plates containing 2 mg/l neomycin and 100 mg/l riboflavin. The correct genotype of the resulting riboflavin-auxotrophic and Nm.sup.r BS3813 strain was confirmed by two PCR reactions using primers p45 together with p10, and primers p51 together with p9, and chromosomal DNA of the transformants as template DNA. The PCR reactions were performed using standard reaction conditions as described above for the generation of DNA fragments flank 5 and flank 3. In addition, the sequence of ribD from BS3813 was confirmed by sequencing.

(15) Transduction of the deletion construct was performed with PBS-1 phage according to the method described in WO 07/051,552 (see Example 6), wherein a lysate of BS3813 was used for transducing the riboflavin-overproducing strain B. subtilis BS3534 carrying a mutation in the transketolase gene (for construction of BS3534 see WO 07/051,552). BS3534 is based on the riboflavin-overproducing strain B. subtilis RB50, which is described in detail in EP 405370 and available under the deposition number NRRL B-18502. Nm.sup.r transductants were selected on TBAB agar plates containing 2 mg/l neomycin and 100 mg/l riboflavin. The genotype of isolated transductants was confirmed by PCR as described above. The resulting strain was designated BS3798.

Example 2

Generation of Strains Carrying Modified Rib Leader Sequences

(16) Two types of mutations were introduced into the rib leader to deregulate the transcription of the rib genes wherein type I refers to ribO mutations and type II refers to (partial) deletions of the rib leader (see FIG. 1).

(17) Three ribO mutations were generated, i.e. C85T named RK41, G121A named RK1a and a triple combination G39A-G40A-G41A named triple ribO. For the type II mutants, the following parts were deleted from the rib leader, wherein the numbering of nucleotides refers to the rib leader sequence shown in SEQ ID NO:42 isolated from B. subtilis: deletion of nucleotides 250 to 257 (SEQ ID NO:36) named del stem loop-right, deletion of nucleotides 231 to 238 (SEQ ID NO:37) named del stem loop-left, deletion of nucleotides 239 to 263 (SEQ ID NO:38) named del flank-right, deletion of nucleotides 231 to 263 (SEQ ID NO:39) named del terminator, deletion of nucleotides 166 to 263 (SEQ ID NO:40) named SWITCH deletion, deletion of nucleotides 135 to 263 (SEQ ID NO:41) named del mro175 and a deletion of the complete leader, i.e. nucleotides 1 to 263 (SEQ ID NO:42) named leader deletion. In the case of the mro175 deletion, also an insertion of four nucleotides took place (5-ATGG-3).

(18) Construction of strains carrying modified rib leader sequences together with the rib promoter (SEQ ID NOs:45-54) were basically performed via two PCR reactions according to the protocol/conditions outlined in Example 1 wherein in a first PCR reaction, the two DNA fragments designated flank 5 and flank 3 were generated using chromosomal DNA from B. subtilis 1A747. In a second PCR reaction according to the protocol/conditions as described in Example 1, these two PCR-fragments were assembled using primers p45 and p51. The respective primer-pairs for the first PCR reactions, i.e. primers for generation of flank 5 and primers for generation of flank 3 resulting in the desired mutations/deletions (see above) are listed in Table 1, column 2 and 3, respectively. After purification using the QiaQuick PCR purification kit (Qiagen), 2 g of the purified full-length PCR product was used for transformation of competent B. subtilis BS3813 cells (see Example 1). The cells were plated onto SMM plates. Riboflavin-prototroph transformants were suspended in 1 ml 0.9% NaCl solution. 100 l of a 500-fold dilution of the original cell suspension was plated onto TBAB agar plates. Single colonies were transferred onto fresh TBAB agar plates and onto TBAB agar plates supplemented with 2 mg/l Nm and 100 mg/l riboflavin. Correct transformants were sensitive to neomycin and thus grew only on TBAB agar plates and did not grow on plates supplemented with neomycin. In addition, the genotype was confirmed by sequencing of the newly introduced rib promoter and ribD. The newly generated B. subtilis strains containing the respective mutations/deletions in the rib leader were designated as shown in Table 1, column 4.

(19) PBS-1 lysates from the newly generated strains were prepared and used for transduction of BS3798 (generated in Example 1). Transduced cells were selected on SMM plates. The riboflavin-prototrophic B. subtilis transductants were suspended in 1 ml 0.9% NaCl solution. 100 l of a 500-fold dilution of the original cell suspension was plated onto TBAB agar plates. Single colonies were transferred onto fresh TBAB agar plates and TBAB agar plates supplemented with 2 mg/l Nm and 100 mg/l riboflavin. Correct transductants grew only on TBAB agar plates and were therefore neomycin-sensitive. The newly transduced strains were named as indicated in Table 1, column 5.

(20) TABLE-US-00001 TABLE 1 Primer pairs used for construction of flank 5 and flank 3 fragments, respectively, resulting in modified rib leader sequences and the designation of the resulting strains either transformed with said PCR-fragments or transduced with the respective lysates (for more explanation see text). Flank 3 Flank 5 Designation primer pair primer pair Trans- Trans- of rib leader (SEQ ID (SEQ ID formed duced mutation NOs) NOs) strains strains RK1a p74/p51 p45/p75 BS3833 BS3839 (20/7) (5/21) RK41 p72/p51 p45/p73 BS3958 BS3987, (16/7) (5/18) BS3988, BS3989 triple ribO p72a/p51 p45/p73a BStriple_ (17/7) (5/19) ribO del flank-right p56/p51 p45/p57 BS3814 BS3832 (8/7) (5/9) del terminator p58/p51 p45/p59 BS3815 BS3821 (10/7) (5/11) del stem loop-right p76/p51 p45/p77 BS3842 BS3846 (22/7) (5/23) del stem loop-left p79/p51 p45/p78 BS3847 BS3859 (25/7) (5/24) SWITCH deletion p80/p51 p45/p81 BS3867 BS3900, (26/7) (5/28) BS3916 del mro175 p80a/p51 p45/p81a BSmro175 (27/7) (5/29) Leader deletion p96/p51 p45/p95 BSleader (31/7) (5/30) Transformed strains: designation of B. subtilis strains after transformation of B. subtilis BS3813 (neo-resistant; B2-auxotroph based on the wt-strain B. subtilis 1A747) with the respective PCR-fragment; Transduced strains: designation of B. subtilis strains after transduction of B. subtilis BS3798 (neo-resistant, B2-auxotroph based on B. subtilis RB50) with lysate of the respective strains according to column 4. For more explanation see text.

(21) The newly generated strains were tested for riboflavin production in shake flask screening as described above. After 48 h, the riboflavin of a 500 l sample was dissolved by addition of 4 N NaOH, neutralized and after centrifugation, the riboflavin concentration of the processed sample was determined by HPLC together with the concentration of DMRL and oxolumazine, a degradation product of DMRL. For calculation of the riboflavin yield on carbon source, the starting and residual concentration of all carbon sources were determined by HPLC.

(22) The results are presented in Table 2. While the native rib operon in a wild-type strain (1A747) background basically does not secrete any riboflavin into the medium, the same rib operon in a strain background selected for riboflavin overproduction (RB50) secreted a measurable amount of riboflavin. The best results were obtained with the ribO mutant RK41 in both the wt (BS3958) and the RB50 background (BS3987), respectively. Deletion of the terminator did not result in the expected results, in particular in the wt background (BS3815). While in the wt background the second best results were obtained with RK1a ribO mutants (BS3833), the second best results in the RB50 background were achieved with the SWITCH deletion (BS3900 and BS3916), which showed even a higher yield in riboflavin than the leader deletion strain.

(23) TABLE-US-00002 TABLE 2A Riboflavin production with transformed B. subtilis strains as indicated based on the wt strain B. subtilis 1A747 carrying the native rib leader. The yield is given as g riboflavin per g carbon source (for more explanation see text). Strain Rib leader mutation Yield [%] 1A747 wild-type 0.0004 BS3833 RK1a 0.0570 BS3958 RK41 0.1100 BS3814 del flank-right 0.0090 BS3815 del terminator 0.0103 BS3867 SWITCH deletion 0.0139

(24) TABLE-US-00003 TABLE 2B Riboflavin production with transduced B. subtilis strains as indicated based on the riboflavin-overproducing strain B. subtilis RB50 carrying the native rib leader. The yield is given as g riboflavin per g carbon source (for more explanation see text). Strain Rib leader mutation Yield [%] RB50 wild-type rib operon 0.46 BS3839 RK1a 1.29 BS3987 RK41 1.64 BS3832 del flank-right 0.33 BS3821 del terminator 0.74 BS3900, SWITCH deletion 1.48 BS3916 BS3846 del stem loop-right 0.11 BS3859 del stem loop-left 0.52

Example 3

Replacing the Native Rib Promoter by Strong Constitutive Promoters

(25) In order to evaluate a possible synergistic effect of rib leader mutations combined with a strong promoter, the original rib promoter of the newly generated constructs (see Example 2) was replaced either by the strong constitutive promoter P.sub.veg or by P.sub.Spo15. The way of construction closely resembled the approach of Example 2. For generation of the DNA fragments flank 3 and flank 5, 1 l of a 100 l solution of primer p60 (SEQ ID NO:12) together with p51 (for construction of P.sub.veg) or primer p62 (SEQ ID NO:14) together with p51 (for construction of P.sub.Spo15) and primer pair p45 together with p61 (SEQ ID NO:13) (for construction of P.sub.veg) or primer p45 together with p63 (SEQ ID NO:15) (for construction of P.sub.Spo15) were added to 0.1 g 1A747 chromosomal DNA in a 50 l reaction volume containing 1 l of 10 mM dNTP's, 5 l of 10 buffer and 0.5 l Pfu polymerase (Stratagene). The PCR reactions were performed for 35 cycles of three sequential steps: (i) denaturing step at 94 C. for 30 sec; (ii) annealing step at 52 C. for 30 sec; (iii) elongation step at 72 C. for 1 min. The cycles were preceded by a DNA denaturation step at 95 C. for 3 min. The two PCR products were separated by agarose gel electrophoresis and extracted from the gel using the MinElute Gel Extraction Kit (Qiagen). In the final PCR reaction, the two purified PCR products (flank 5 and flank 3) were assembled: 1 l of a 100 M solution of primers p45 and p51, 1.0 l flank 5 PCR product (50 ng) and 1.0 l flank 3 PCR product (50 ng) were mixed in a 50 l final reaction volume containing 1 l of 10 mM dNTP's, 5 l of 10 buffer and 0.5 l Pfu polymerase (Stratagene). The PCR reaction was performed for 35 cycles of three sequential steps: (i) denaturing step at 94 C. for 30 sec; (ii) annealing step at 52 C. for 30 sec; (iii) elongation step at 72 C. for 2.5 min. The PCR cycles were preceded by a denaturation step at 95 C. for 3 min. The assembled PCR product was purified by using the QiaQuick PCR purification kit (Qiagen). The purified full-length PCR product (2 g) was used for transformation of competent B. subtilis BS3813 cells. The cells were plated onto SMM plates. Riboflavin-prototrophic Bacillus transformants were suspended in 1 ml 0.9% NaCl solution. 100 l of the 500-fold dilution of the original cell suspension was plated on TBAB agar plate. Single colonies were transferred onto fresh TBAB agar plates and onto TBAB agar plates supplemented with 2 mg/l Nm and 100 mg/l riboflavin. Correct transformants grew only on TBAB agar plates and therefore were neomycin-sensitive. In addition, the genotype was confirmed by sequencing of the newly integrated promoter, rib leader and ribD. The strain with a P.sub.veg driven rib operon was called BS3811, the strain with the P.sub.Spo15 driven rib operon was called BS3817.

(26) In order to combine the strong promoters with the leader modifications, the gDNA of strain 1A747 was replaced by gDNA of strain BS3811 (P.sub.veg promoter) and BS3817 (P.sub.Spo15 promoter), respectively, in all described PCR reactions of Example 2. All other conditions of each PCR reaction including the used primer pairs were kept identical. The final PCR products were transformed into BS3813 and obtained transformants were verified as described (Example 2). In the case of the construct P.sub.Spo15_leader deletion, the primer pairs p45/p95 for the flank 5 PCR product and p96/p51 for the flank 3 PCR product were used as indicated in Table 4. All other steps were performed as described above. The following constructs were made, with the designation of the resulting B. subtilis strain transformed with said constructs in brackets: P.sub.veg_del flank-right (BS3840), P.sub.SPo15_del flank-right (BS3831), P.sub.veg_del terminator (BS3844), P.sub.Spo15_del terminator (BS3871), P.sub.Spo15_SWITCH deletion (B53874), P.sub.Spo15_leader deletion (B53944), P.sub.veg_RK41 (BS3887), P.sub.veg_RK1a (BS3953), P.sub.Spo15_RK1a (B53884), P.sub.veg_triple ribO (BS3912).

(27) PBS-1 lysates were prepared from the strains mentioned above and transduced into BS3798. Transduced cells were selected on SMM plates. The riboflavin-prototrophic B. subtilis transformants were suspended in 1 ml 0.9% NaCl solution. 100 l of the 500-fold dilution of the original cell suspension was plated onto TBAB agar plates. Single colonies were transferred onto fresh TBAB agar plates and TBAB agar plates supplemented with 2 mg/l Nm and 100 mg/l riboflavin. Correct transformants grew only on TBAB agar plates and were therefore neomycin-sensitive. The following strains were generated: BS3970 and BS3971 transduced with PBS-1 lysate from BS3953; BS3905 and BS3907 transduced with PBS-1 lysate from BS3884; BS3903 and BS3914 transduced with PBS-1 lysate from BS3887; BS3981-83 transduced with PBS-1 lysate from BS3912; BS3890 transduced with PBS-1 lysate from BS3840; BS3880 and BS2882 transduced with PBS-1 lysate from BS3844; BS3850 and BS3851 transduced with PBS-1 lysate from BS3817; BS5026 and BS5041 transduced with PBS-1 lysate from BS3944; BS3853 transduced with PBS-1 lysate from BS3831; BS3897 and BS3956 transduced with lysate from BS3874.

(28) Most of the strains were tested for riboflavin production in shake flasks as described above. After 48 h, riboflavin of a 500 l sample was dissolved by addition of NaOH, neutralized and after centrifugation the riboflavin concentration of the processed samples was determined by HPLC together with the concentration of DMRL and oxolumazine. For calculation of the riboflavin yield on carbon source, the starting and the residual concentration of all carbon sources were determined by HPLC. The results are shown in Table 3.

(29) TABLE-US-00004 TABLE 3A Riboflavin production with transformed B. subtilis strains as indicated based on the wt strain B. subtilis 1A747 carrying the native rib leader. The yield is given as g riboflavin per g carbon source (for more explanation see text). Strain Rib leader mutation Yield [%] 1A747 wild-type 0.0004 BS3811 P.sub.veg 0.0120 BS3817 P.sub.Spol5 0.0163 BS3953 P.sub.veg RK1a 0.1600 BS3884 P.sub.Spol5 RK1a 0.7706 BS3887 P.sub.veg RK41 0.3757 BS3912 P.sub.veg triple ribO 1.0900 BS3840 P.sub.veg del flank-right 0.0748 BS3831 P.sub.Spol5 del flank-right 0.0702 BS3844 P.sub.veg del terminator 0.0666 BS3871 P.sub.Spol5 del terminator 0.0359 BS3874 P.sub.Spol5 SWITCH deletion 0.2823 BS3944 P.sub.Spol5 leader deletion 0.1150

(30) TABLE-US-00005 TABLE 3B Riboflavin production with transduced B. subtilis strains as indicated based on the riboflavin-overproducing strain B. subtilis RB50 carrying the native rib leader. The yield is given as g riboflavin per g carbon source (for more explanation see text). Strain Rib leader mutation Yield [%] RB50 wild-type rib operon 0.46 BS3849 P.sub.veg 1.33 BS3850, P.sub.Spol5 1.57 BS3851 BS3970, P.sub.veg RK1a 2.88 BS3971 BS3905, P.sub.Spol5 RK1a 3.31 BS3907 BS3903, P.sub.veg RK41 4.18 BS3914 BS3981 P.sub.veg triple ribO 2.52 BS3890 P.sub.veg del flank-right 1.14 BS3853 P.sub.Spol5 del flank-right 1.63 BS3880 P.sub.veg del terminator 1.80 BS3875 P.sub.Spol5 del terminator 2.38 BS3897 P.sub.Spol5 SWITCH deletion 3.76 BS5026 P.sub.Spol5 leader deletion 2.16

(31) All manipulation done to the rib leader or to the rib promoter led to an increased riboflavin production. In accordance to the results described in Example 2 (see Table 2), the ribO mutation RK41 was also the most effective one when replacing the native promoter by a stronger, constitutive one. By using a strong promoter a yield of up to 4.2% was reached with the combination of P.sub.veg with RK41. Surprisingly, the combination of P.sub.Spo15 with SWITCH deletion showed the next best yield with 3.8%, which is much better than the combination P.sub.Spo15 with del terminator that resulted in a yield of 2.4%.

(32) All manipulation done to the rib leader or to the rib promoter led to an increased riboflavin production. In accordance to the results described in Example 2 (see Table 2), the ribO mutation RK41 was also the most effective one when replacing the native promoter by a stronger, constitutive one. By using a strong promoter a yield of up to 4.2% was reached with the combination of P.sub.veg with RK41. Surprisingly, the combination of P.sub.Spo15 with SWITCH deletion showed the next best yield with 3.8%, which is much better than the combination P.sub.Spo15 with del terminator that resulted in a yield of 2.4%.

Example 4

Combination of ribO Mutations with Leader Deletions and Replacement of the Native Rib Promoter by Strong Constitutive Promoters

(33) To see whether the combination of a typical deregulating ribO mutation with leader deletions is able to increase riboflavin production, some of the mutations generated in Example 2 and 3 were combined. The construction followed the protocol outlined above. Templates and primer pairs for the flank 5 and 3 PCRs are shown in Table 4 (for more information see also Example 1). Assembling PCR and transformation into BS3813 were done as described in Example 2 and 3. Sequencing revealed that two additional mutations were present in constructs P.sub.Spo15_triple ribO_del mro175 and P.sub.veg_triple ribO_del mro175, namely T25G and C101T, wherein the numbering relates to SEQ ID NO:42. The designation of the new strains carrying the newly generated constructs is shown in Table 4, column 4.

(34) PBS-1 lysates from the newly generated strains were prepared and used for transduction of BS3798 (generated in Example 1). Selection of transduced cells on SMM plates were performed as described in Example 2. The newly transduced strains were named as indicated in Table 4, column 5.

(35) TABLE-US-00006 TABLE 4 Primer pairs used for construction of flank 5 and flank 3 fragments, respectively, resulting in modified rib leader sequences combined with the respective constitutive promoters and the designation of the resulting strains either transformed with said PCR-fragments or transduced with the respective lysates (for more explanation see text). Flank 3 Flank 5 Designation of primer pair primer pair Trans- Trans- rib leader mutation (SEQ ID (SEQ ID formed duced (Template for PCR) NOs) NOs) strains strains P.sub.veg triple ribO_del p45/p81a p80a/p51 BS3889 BS3908 mro175 (BS3912) (5/29) (27/7) P.sub.Spol5 triple p45/p63 p62/p51 BS3923 BS3922 ribO_del mro175 (5/15) (14/7) (BS3889) P.sub.Spol5 RK41_del p45/p73 p72/p51 BS3954 BS3984 terminator (BS3871) (5/18) (16/7) P.sub.Spol5 RK41_SWITCH p45/p73 p72/p51 BS3915 BS3964 deletion (BS3874) (5/18) (16/7) P.sub.veg RK41_SWITCH p45/p81 p80/p51 BS3920 BS3899 deletion (BS3887) (5/28) (26/7)

(36) Testing of the newly generated strains for riboflavin production was performed via shake flask screening a described above. The results are presented in Table 5 showing a further increase in riboflavin production compared to the results presented in Example 3 (see Table 4). The combination of ribO mutations RK41 and triple ribO, respectively, with a strong promoter and a rib leader deletion resulted in riboflavin yields of more than 7.6% in a shake-flask screening compared to 4.2% achieved with the combination of the P.sub.veg promoter with ribO mutation RK41 (see Table 4), the best combination without a leader deletion. The specific leader deletion as described herein did nearly double the yield of deregulated (ribO mutation, constitutive promoter) rib leader-promoter combinations. Besides the SWITCH deletion also del mro175 was unexpectedly able to improve the riboflavin yield in a RB50 background over a construct without deletion. The deletion of the terminator (del terminator) had also a positive but less pronounced effect. The leader deletion construct in which P.sub.spo15 is put directly in front of the Shine-Dalgarno sequence of ribD showed a 3.6-times smaller yield than the best triple constructs (BS5026, see Table 4 and 6). These results suggest that the rib leader is not only required for regulation but also for stabilization of the full-length transcript.

(37) TABLE-US-00007 TABLE 5 Riboflavin production with transduced B. subtilis strains as indicated based on the riboflavin-overproducing strain B. subtilis RB50 carrying the native rib leader. The yield is given as g riboflavin per g carbon source (for more explanation see text). Strain Rib leader mutation Yield [%] RB50 wild-type rib operon 0.46 BS3908 P.sub.veg triple ribO_del mro175 6.79 BS3922 P.sub.Spol5 triple ribO_del mro175 6.47 BS3984 P.sub.Spol5 RK41_del terminator 5.24 BS3964 P.sub.Spol5 RK41_SWITCH deletion 7.80 BS3899 P.sub.veg RK41_SWITCH deletion 7.67 BS5026 P.sub.Spol5 leader deletion 2.16

Example 5

Replacing the Native Rib Leader by an mRNA Stabilizing Element

(38) The DNA sequence of the aprE mRNA stabilizing element was distributed over two PCR products. For amplifying PCR product 1 containing the 5 region of the rib operon at the 5 end of the aprE mRNA stabilizing element, the primers p45 together with p143 (SEQ ID NO:33) and the chromosomal DNA from strain BS3817 as template was used under standard PCR conditions. For amplifying PCR product 2 containing the ribD at the 3 end of the aprE mRNA stabilizing element, the primers p51 together with p142 (SEQ ID NO:32) and the chromosomal DNA from strain BS3817 as template were used under standard PCR conditions. In the standard LFH-PCR reaction, the gel-purified PCR products 1 and 2 were assembled into one DNA fragment as described before. The same method was applied to the grpE mRNA stabilizing element with chromosomal DNA from strain BS3817 as template using primer pair p45/p145 (SEQ ID NO:35) and primer pair p51/p144 (SEQ ID NO:34) for the first two PCRs followed by the assembling PCR using the primer pair p45/p51 under the conditions as described above. The purified LFH-PCR products were transformed again into competent cells of the riboflavin-auxotroph B. subtilis BS3813, in which the riboflavin promoter region and the 5 part of ribD was replaced by a neomycin resistance cassette. Riboflavin-prototroph transformants were selected on SMM plates. Isolated transformants were suspended in 1 ml 0.9% NaCl solution and 100 l of the 500-fold dilution of the original cell suspension was plated on TBAB agar plates. Single colonies were transferred onto fresh TBAB agar plates and TBAB agar plates supplemented with 2 mg/l Nm and 100 mg/l riboflavin. The correct transformants grew only on TBAB agar plates and were therefore neomycin-sensitive. In addition, the genotype was confirmed by sequencing of the newly introduced stretch of DNA. The resulting strains were designated as BS5193 (carrying the construct P.sub.15aprEribDEAHT) and BS5196 (carrying the construct P.sub.15grpEribDEAHT), respectively.

(39) PBS-1 lysates from strains BS5193 and BS5196 were prepared and used for transduction of BS3798 (generated in Example 1). Selection of transduced cells on SMM plates were performed as described in Example 2. The newly transduced strains based on strain BS5193 were named BS5260 and BS5262, respectively, the newly transduced strain based on strain BS5196 was named BS5244.

(40) Riboflavin production was tested in shake flask experiments as described above. The results are shown in Table 6.

(41) TABLE-US-00008 TABLE 6A Riboflavin production with transformed B. subtilis strains as indicated compared to the wt B. subtilis strain 1A747 carrying the native rib leader. The yield is given as g riboflavin per g carbon source (for more explanation see text). Strain Rib leader mutation Yield [%] 1A747 wild-type 0.0004 BS3817 P.sub.Spol5 0.0163 BS3944 P.sub.Spol5 leader deletion 0.1150 BS5193 P.sub.Spol5 aprE 0.1000 BS5196 P.sub.Spol5 grpE 0.3700

(42) TABLE-US-00009 TABLE 6B Riboflavin production of transduced B. subtilis strains as indicated compared to the riboflavin-overproducing strain B. subtilis RB50 carrying the native rib leader. The yield is given as g riboflavin per g carbon source (for more explanation see text). Strain Rib leader mutation Yield [%] RB50 wild-type rib operon 0.46 BS3850 P.sub.Spol5 1.57 BS5026 P.sub.Spol5 leader deletion 2.16 BS5260 P.sub.Spol5 aprE 2.07 BS5244 P.sub.Spol5 grpE 2.99

(43) Replacing the rib promoter by a non-regulated constitutive promoter like P.sub.Spo15 increased the riboflavin yield in the shake-flask screening compared to a regulated wild-type rib operon 41-fold in a wild-type background and 3.4-fold in a RB50 background. When the rib leader was removed and the P.sub.Spo15 promoter was put directly in front of ribD, the riboflavin yield was again 7-fold increased in a wild-type host strain and 1.4-fold in the RB50 background. Replacing the rib leader by the aprE mRNA stabilizing element showed no influence on the yield in the two strain backgrounds used above. However, by replacing the rib leader with the grpE mRNA stabilizing element, the riboflavin yield was 3.7-fold increased in a wild-type host and 1.44-fold in the RB50 background.

(44) In order to check the influence of the host strain with regards to the performance of a modified rib operon, the lysates from B. subtilis 1A747 (wild-type rib operon), BS5193, BS5196 and BS3944 were also transduced into another RB50 variant named BS5178 (spo0A.sup., rib::neo, ribC1, bpr::cam, tkt.sup.mut). Shake flask screenings were performed as described above. The results including the names of the newly constructed strains are shown in Table 7.

(45) TABLE-US-00010 TABLE 7 Riboflavin production of transduced B. subtilis strains as indicated above. The yield is given as g riboflavin per g carbon source and compared to the yield of BS5240. Strain Rib leader mutation Yield [%] BS5240 P.sub.Spol5 leader deletion 100 BS5191 wild-type rib operon 40 BS5237 P.sub.Spol5 aprE 140 BS5238 P.sub.Spol5 grpE 270

(46) In the new strain background the positive effect of the stabilizing elements was more pronounced when tested in the shake flask format. Now, the aprE mRNA stabilizing element, too, increased the riboflavin yield by 40% compared to a strain carrying the rib operon without a leader. The grpE mRNA stabilizing element even showed a 2.7-fold increased yield compared to the construct without a rib leader.

(47) In the new strain background the positive effect of the stabilizing elements was more pronounced when tested in the shake flask format. Now, the aprE mRNA stabilizing element, too, increased the riboflavin yield by 40% compared to a strain carrying the rib operon without a leader. The grpE mRNA stabilizing element even showed a 2.7-fold increased yield compared to the construct without a rib leader.

Example 6

Combination of Additional ribO Mutations with Leader Deletions and Strong Constitutive Promoters

(48) In the same way as the ribO mutations RK41 (RK61a), RK1a and triple ribO, which combines the ribO mutations RK4, RK8, RK5 (RK2) of Kil et al., 1992, were introduced into the rib leader, additional ribO mutation described in Kil et al., 1992, can be used for the generation of alternative optimized rib leaders. Numbering of the mutations refers to the rib leader sequence shown in SEQ ID NO:42. The described mutations RK111a (G59A), RK116a (G56A), RK62a (G60A; identical to RK82a), RK93a (C87T), and RK27a (C128T) are combined with del terminator, the SWITCH deletion and del mro175 (see above) together with a strong constitutive promoter P.sub.Spo15 or P.sub.veg depending on the template used for generation of flank 5 and flank 3. Any mutation effective as ribO mutation can be combined in this way with the described promoters and leader deletions.

(49) The following primer pairs are used for construction of the ribO mutations: primer pair p111a_f/p 111a_r (SEQ ID NO:83 and 84) for construction of RK111a, primer pair p116a_f/p116a_r (SEQ ID NO:85 and 86) for construction of RK116a, primer pair p62a_f/p62a_r (SEQ ID NO:87 and 88) for construction of RK62a, primer pair p93a_f/p93a_r (SEQ ID NO:89 and 90) for construction of RK93a, and primer pair p27a_f/p27a_r (SEQ ID NO:91 and 92 for construction of RK27a. These primers are used in PCR-reactions as described in Example 2. For generation of flank 5, primer p45 is applied together with the antisense primer. For generation of flank 3, the sense primer is applied together with primer p51. Flank 5 and 3 are assembled in a third PCR using primers p45 and p51 (see Example 2). Transformation and transduction of strains are performed as described in Example 2. For the replacement of the native promoter the instructions according to Example 3, for combination of ribO mutations with leader mutations the instructions according to Example 4 are followed.

(50) TABLE-US-00011 TABLE 9 Templates required for generation of the ribO constructs by the PCR-based method as described above in the previous examples (for more explanation see text). Designation of Template for Template for rib leader mutation flank 5 flank 3 P.sub.Spol5 RK111a BS3817 BS3817 P.sub.veg RK111a BS3811 BS3811 P.sub.Spol5 RK111a_ SWITCH deletion BS3874 BS3874 P.sub.veg RK111a_ SWITCH deletion BS3811 BS3867 P.sub.Spol5 RK111a_del terminator BS3871 BS3871 P.sub.veg RK111a_del terminator BS3844 BS3844 P.sub.Spol5 RK111a_del mro175 BS3817 BSmro175 P.sub.veg RK111a_del mro175 BS3811 BSmro175

(51) The generated PCR products for all constructs listed in Table 9 are transformed into BS3813. Selection takes place as described in Example 2 and lysates of the confirmed strains are used for transduction of BS3798. The yield of riboflavin obtained in flask shake experiments as described above are in the range as for constructs RK41_SWITCH deletion or triple ribO_del mro175 (see Table 5). These amounts are even increased when using another strain background, as e.g. described in Example 5.

Example 7

Generation of Strains Other than B. subtilis Carrying Modified Rib Leader Sequences

(52) The constructs as described in the Examples above can be used to identify/generate corresponding modifications in rib leader sequences from other strains which are known to have a riboswitch in place and which are suitable host strains for riboflavin production.

(53) Corresponding parts of the non-modified rib leaders are identified in other organisms according to the alignment depicted in FIG. 2 of Vitreschak et al., Nucleic Acid Res 30, 3141-3151, 2002. Deletion mutations are generated as described above and optionally combined with ribO mutations (homologous to the ones identified in B. subtilis). The constructs can be furthermore combined with strong promoters or other known modification of the host strain as described above in order to increase riboflavin production under suitable culture conditions which are known to the skilled person.