METHOD OF SELECTING A POLYPEPTIDE OF INTEREST

Abstract

The invention relates to methods for identifying polypeptides and polynucleotides of interest, be they novel or variant polypeptides and polynucleotides, by expressing a plurality of polypeptides in an obligate or facultative anaerobe that is incapable of, or displays a reduction in, the oxidation of NADH and/or NADPH under anaerobic fermentation conditions and selecting an obligate or facultative anaerobe that grows or displays a growth advantage under said conditions. The invention is also concerned with novel enzymes per se, and their use in enzymatic production processes.

Claims

1. A method of identifying a variant polypeptide of interest, or its encoding polynucleotide, the method comprising: (i) generating a plurality of variant polypeptides; (ii) expressing the plurality of variant polypeptides in an obligate or facultative anaerobe that is incapable of, or displays a reduction in, the oxidation of NADH and/or NADPH under anaerobic fermentation conditions; (iii) culturing, in growth media, the obligate or facultative anaerobe under anaerobic fermentation conditions in the presence of a substrate, wherein the polypeptide of interest enables the obligate or facultative anaerobe to oxidise, or to increase oxidation of, NADH and/or NADPH in the presence of the substrate; (iv) selecting an obligate or facultative anaerobe that grows or displays a growth advantage in the growth media; and (v) identifying the variant polypeptide of interest expressed, or its encoding polynucleotide, in the obligate or facultative anaerobe of step (iv).

2. A method according to claim 1, wherein the substrate is exogenously added to the growth media.

3. A method according to claim 1, wherein the substrate is endogenously produced by the anaerobe.

4. The method according to claim 1, wherein the variant polypeptide comprises at least one amino acid substitution, deletion or insertion compared to its wild-type counterpart, or comprises a synthetically designed polypeptide.

5. The method according to claim 1, wherein the variant polypeptide is expressed in step (ii) by the introduction of a vector comprising a polynucleotide encoding the variant polypeptide into the anaerobe.

6. The method according to claim 1, wherein the obligate or facultative anaerobe is a bacterium, yeast or fungus, optionally wherein the obligate or facultative anaerobe is Escherichia coli.

7. The method according to claim 1, wherein the obligate or facultative anaerobe is rendered incapable of, or displays a reduction in, the oxidation of NADH and/or NADPH by having at least one gene, or product thereof, associated with an NAD.sup.+ and/or NADP.sup.+ regeneration metabolic pathway, which is non-functional and/or inhibited, optionally wherein the at least one gene has been deleted, disrupted or mutated, optionally wherein the at least one gene encodes lactate dehydrogenase, alcohol dehydrogenase, soluble transhydrogenase and/or transmembrane transhydrogenase, optionally wherein the facultative anaerobe is Escherichia coli and the at least one gene encodes lactate dehydrogenase (ldhA), alcohol dehydrogenase (adhE), soluble transhydrogenase (sthA) and/or transmembrane transhydrogenase (pntA and/or pntB).

8. (canceled)

9. The method according to claim 1, wherein the obligate or facultative anaerobe is a thermophilic organism, and the obligate or facultative anaerobe is cultured in step (iii) at a temperature greater than 37° C., 40° C., 50° C., 60° C. or at least 70° C. and the variant polypeptide of interest is one which is able to provide for oxidation, or an increase in oxidation, of NADH and/or NADPH at such temperatures.

10. The method according to claim 1, wherein an obligate or facultative anaerobe that is not expressing the variant polypeptide of interest will not grow, or grow at a reduced rate, when compared to an obligate or facultative anaerobe expressing the variant polypeptide of interest, when culturing under the conditions of step (iii), enabling the selection of the obligate or facultative anaerobe expressing the variant polypeptide of interest in step (iv).

11. The method according to claim 1, wherein the identification of the variant polypeptide of interest, or it encoding polynucleotide, in step (v) comprises: i. extracting the protein and/or DNA from the obligate or facultative anaerobe; and ii. determining the variant polypeptide sequence, or the polynucleotide sequence encoding the variant polypeptide sequence.

12. The method according to claim 1, wherein the variant polypeptide is selected from the group consisting of: an enzyme, a membrane transporter, a transcription factor and a chaperone.

13. The method according to claim 1, wherein the variant polypeptide is an enzyme, optionally wherein the enzyme displays an altered specificity selected from a group consisting of: stereospecificity, thermostability, chemostability, pressure stability, substrate specificity, catalytic efficiency, oxidative stability regiospecificity, cofactor preference and/or specificity, and binding affinity for substrate and/or cofactor, optionally wherein the enzyme is an NAD(P)H-dependent oxidoreductase.

14. (canceled)

15. The method according to claim 1, wherein the variant polypeptide is a membrane transporter, optionally wherein the membrane transporter is an active transporter, a passive transporter, or a membrane channel.

16. A kit for identifying a variant polypeptide of interest, or its encoding polynucleotide, the kit comprising: i. an obligate or facultative anaerobe that is rendered incapable of, or displays a reduction in, the oxidation of NADH and/or NADPH; and ii. growth media comprising a substrate; wherein a variant polypeptide of interest will enable the obligate or facultative anaerobe to oxidise, or to increase oxidation of, NADH and/or NADPH in the presence of the substrate when grown under anaerobic fermentation conditions, the kit optionally further comprising providing a library of variant polypeptides.

17. A variant of: Clostridium beijerinckii alcohol dehydrogenase, which comprises a modification of one or more amino acids relative to the wild-type sequence of SEQ ID NO: 1, wherein the variant has altered cofactor specificity compared to its corresponding wild-type, such that it utilises NADH instead of NADPH; (ii) Myxococcus stipitatus imine reductase, which comprises a modification of one or more amino acids relative to the wild-type sequence of SEQ ID NO: 34, wherein the variant has altered cofactor specificity compared to its corresponding wild-type, such that it utilises NADH instead of NADPH, optionally wherein the variant of Myxococcus stipitatus imine reductase comprises an amino acid sequence substantially as set out in SEQ ID NO: 35, or a fragment or variant thereof or (iii) Enterobacter cloacae nitroreductase, which comprises a modification of one or more amino acids relative to the wild-type sequence of SEQ ID No: 37, wherein the variant has altered substrate specificity, such that it is able to catalyse the reduction of 2-nitrobenzoic acid (2-NBA) and/or 4-nitrobenzyl alcohol for efficiently that the wild-type nitroreductase, optionally wherein the variant of Enterobacter cloacae nitroreductase comprises an amino acid sequence substantially as set out in SEQ ID No 38 or 40, or a fragment or variant thereof.

18. The variant according to claim 17(i), comprising amino acid substitutions at positions 198, 199 and 218, optionally further comprising an amino acid substitution at position 200.

19. The variant according to claim 17(i), comprising an amino acid sequence substantially as set out in SEQ ID NO: 2, or a fragment or variant thereof.

20. The variant according to claim 17(i) encoded by a nucleic acid sequence as substantially set out in SEQ ID NO: 3, or a fragment or variant thereof.

21-31. (canceled)

Description

[0221] For a better understanding of the invention, and to show how embodiments of the same may be carried into effect, reference will now be made, by way of example, to the accompanying Figures, in which:

[0222] FIG. 1 is a schematic representation of one embodiment of the selection system of the invention. The selection system is based on a bacterial strain with impaired anaerobic fermentation. Such a strain was generated by knocking out the adhE and ldhA genes, critical for the alcoholic and lactic fermentation pathways. The strain is unable to grow under anaerobic fermentation conditions due to its inability to regenerate oxidized NAD+. If cells of this strain are cultured anaerobically in a medium supplemented with a specific oxidized substrates and transformed with a library of a NAD(H)-dependent oxidoreductases or NADP(H) dependent oxidoreductases, only the cells carrying a variant oxidoreductase which is able to oxidize the supplemented substrate will be able to grow;

[0223] FIG. 2 shows anaerobic fermentative growth recovery with adhE. FIG. 2a: Growth curve of BW25113, LS1 and LS1+pLS1 cultures. LS1 cells are unable to grow in anaerobic fermentative conditions. Transformation with pLS1 (which carries the adhE gene) allows growth recovery. FIG. 2b: HPLC-RID of fermentation broth of BW25113, LS1 and LS1+pLS1 cultures. LS1 cells transformed with pLS1 show a profile of fermentation products similar to that of BW25113, except for the absence of lactate;

[0224] FIG. 3 shows metabolic complementation with TADH. FIG. 3a: LS1 cells transformed with either pLS1 or pLS12 (carrying TADH) are unable to grow under anaerobic conditions if cyclohexanone is not added to the medium. If cyclohexanone is added to the medium, LS1 cells transformed with pLS12 (but not with pLS1) achieve growth recovery through anaerobic fermentation. FIG. 3b: Quantification by means of GC of cyclohexanone and cyclohexanol in fermentation broth of LS1+pUC19, LS1+pLS1 and LS1+pLS12 anaerobic cultures supplemented with cyclohexanone. When cells are transformed with pLS12, cyclohexanone is completely consumed, and cyclohexanol is generated. FIG. 3c: TADH is able reduce cyclohexanone, 3-methylcyclohexanone and butanal with NADH, and oxidize ethanol with NAD+;

[0225] FIG. 4 shows selection of an NAD(H)-dependent variant of CBADH. FIG. 4a: LS1 cells cultured anaerobically with acetone added to the medium and transformed with the library of variants of CBADH were able to grow faster than those transformed with pLS6. FIG. 4b: When LS1 cells were transformed with the isolated variant, anaerobic growth recovery was even more efficient than when they were transformed with the library. FIG. 4c: Characterization of the fermentation broth by means of HPLC-RID (left) and GC (right) revealed that in cells transformed with the NAD(H)-dependent variant glucose consumption did not result into ethanol or lactate production as expected. Acetone was consumed and transformed into isopropanol. FIG. 4d: Enzymatic activity assays revealed the new variant had no activity with NADP(H), but had gained activity with NAD(H);

[0226] FIG. 5 shows a DNA sequence combinatorially assembled to express an isopropanol pathway including an integrated NAD(H)-dependent variant of CBADH;

[0227] FIG. 6 is a schematic representation of one embodiment of the selection system of the invention wherein the polypeptide of interest is an enzyme that catalyses the conversion of an exogenous substrate into an intermediate product that is utilised as a substrate for a promiscuous NADH-dependent oxidoreductase;

[0228] FIG. 7 is a schematic representation of one embodiment of the selection system of the invention wherein the polypeptide of interest is a membrane transporter;

[0229] FIG. 8 is a schematic representation of one embodiment of the selection system of the invention wherein cells are transformed with a lipase and cultured in media supplemented with an ester or another molecule which the lipase can use as a substrate, yielding a molecule that can be oxidized by either an endogenous NADH-dependent oxidoreductase natively present in the cells transformed with the lipase or an exogenous NADH-dependent oxidoreductase with which the cells are also transformed

[0230] FIG. 9 is a schematic representation of one embodiment of the selection system of the invention relating to the development of thermostable enzymes utilizing a thermophilic organism;

[0231] FIG. 10 is a schematic representation of one embodiment of the selection system of the invention wherein the polypeptide of interest is an L-amino acid dehydrogenase;

[0232] FIG. 11 shows a schematic representation of the LS5 strain. LS5 strain includes further metabolic defects in addition to those already present in LS1: both sthA and pntAB transhydrogenase-encoding genes were deleted. This removes the ability to balance the redox pools of NAD and NADP, thus making the strain unable to grow anaerobically unless transformed with a strictly NADH-dependent oxidoreductase whose substrate is present in the media;

[0233] FIG. 12 shows growth curves showing that LS5 requires transformation with a strictly NADH-dependent oxidoreductase to grow anaerobically. a) Both LS1 and LS5 strains are unable to grow anaerobically in media lacking acetone, even when transformed with NAD(P)H-dependent oxidoreductases since they cannot regenerate oxidized cofactor because of the lack of the substrate. b) When the media is supplemented with acetone, LS1 strain is able to grow aerobically when transformed with either NADH-dependent or NADPH-dependent CBADH. However, LS5 strain is only able to grow anaerobically when transformed with NADH-dependent CBADH;

[0234] FIG. 13 shows a quadruple mutant strain for strict selection of NAD(H)-dependent oxidoreductases. For each E. coli mutant strain, anaerobic growth with (right panels) and without (left panels) acetone supplemented to the culture media was followed. a, AL mutant (ΔadhE ΔldhA). b, ALS mutant (ΔadhE ΔldhA ΔsthA). c, ALP mutant (ΔadhE ΔldhA ΔpntA). d, ALPS mutant (ΔadhE ΔldhA ΔpntB ΔsthA). Anaerobic growth of cells with at least one active transhydrogenase was recovered upon transformation of either an NADH or an NADPH-dependent oxidoreductase. However, in the case of ALPS cells, where both transhydrogenases were knocked-out, only the NADH-dependent enzyme restored anaerobic growth, which may indicate that metabolic complementation by NADPH-dependent enzymes is mediated by transhydrogenases;

[0235] FIG. 14 shows an overview of the metabolic complementation selection system. a, Schematic of the main steps for obtaining a novel biomolecule variant with the inventor's selection system. b, Overview of metabolic complementation within AL mutant cells. In AL mutant cells, both adhE and ldhA are knocked-out, which prevents them from regenerating oxidized NAD+ under anaerobic conditions. Since oxidized NAD+ is required for obtaining a constant supply of ATP through anaerobic fermentation of glucose, these cells are unable to grow anaerobically. Only upon transformation with a plasmid encoding a biomolecule able to restore the oxidation NADH (and addition to the media of the required substrate if necessary) cells overcome their metabolic impairment and anaerobic growth is restored. c, Recovery of anaerobic growth with endogenous adhE. AL cells transformed with adhE (Positive control) were able to grow anaerobically similarly to the WT cells, unlike untransformed AL cells (Negative control). d, Recovery of anaerobic growth with acetoin reductases from Bacillus subtilis (bdhA) and Klebsiella pneumoniae (budC). Cells transformed with either of the reductases displayed levels of anaerobic growth similar to the positive control, demonstrating that metabolic complementation can also be achieved with exogenous reductases. Untransformed cells also grew in media with acetoin, although with a much longer lag phase, possibly due to the presence of an endogenous acetoin reductase in the E. coli genome. e, NMR spectra of the fermentation broth of AL cells complemented with bhdA and budC. In both cases, the supplied acetoin was consumed and the reduced product, 2,3-butanediol, was produced. 2,3-butanediol was also observed in the fermentation broth of untransformed cells. f, Recovery of anaerobic growth with Thermus sp. ATM alcohol dehydrogenase. AL cells transformed with TADH grew anaerobically when supplemented with cyclohexanone (triangle) or 3-methylcyclohexanone (inverted triangle) (both substrates of TADH). g, NMR spectra of the fermentation broth of AL cells complemented with TADH. The alcohol corresponding to the ketone supplemented to the culture media was detected in both cases;

[0236] FIG. 15 shows the workflow of producing the variant CBADH and shows the crystal structure of pLS10 3 bound to NADH, obtaining insight into the structural basis of cofactor preference reversal;

[0237] FIG. 16 shows the evolution and characterization of a novel NADH-dependent imine reductase. a, Overimposition of a prediction of the structure of MsIRED-s (pLS133_1) (pink) and a homology model of WT MsIRED based on the crystal structure of 3ZHB (grey), with the bound NADPH cofactor displayed in orange. Structural analysis suggests that the residue substitutions present in MsIRED-s destroy the electrostatic interactions established in the WT enzyme by positively charged residues with the 2′ phosphate of NADPH. b, Comparison of anaerobic growth in media with 2-methylpyrroline for cells transformed with WT MsIRED, previously engineered NADH-dependent MsIRED variants and MsIRED-s Anaerobic growth occurred to the largest extent, and with the shortest lag phase, when cells were transformed with MsIRED-s. c, Activity assays of the best previously described NADH-dependent MsIRED variant (left) and MsIRED-s (right). MsIRED-si had a higher affinity for the substrate, as shown by its lower Km. Furthermore, the extent of substrate inhibition was reduced in MsIRED-s, which also displayed a higher activity at the optimal substrate concentration;

[0238] FIG. 17 shows a library of pathways for isopropanol production based on the combination of genes previously designed by Hanai et al (Clostridium acetobutylicum acetyl-CoA acetyltransferase (thl) and acetoacetate decarboxylase (adc), Escherichia coli acetoacetyl-CoA transferase (atoAD) and CBADH). The figure also shows a comparison of isopropanol production under aerobic conditions was compared for randomly selected variants, variants selected in plates and variants selected in plates that grew in anaerobic liquid cultures (FIG. 17b);

[0239] FIG. 18 shows the growth curve under anaerobic conditions of WT_Geobacillus thermoglucosidasius (1198) and mutant TMO236 Geobacillus thermoglucosidasius. Thus highlighting that thermophilic organisms may be utilised for the selection methods of the invention;

[0240] FIGS. 19a and 19b shows the improved kinetic properties of the variant nitroreductases described in the examples; and

[0241] FIG. 20 shows the isopropanol variant clones comprising variant promoters, RBS and terminator sequences. Sequence ID numbers are, starting from the Promoter J23100 and going down the figure to the terminator T4 (ECK120029600): SEQ ID No: 90 to 105.

[0242] Materials and Methods

[0243] Plasmid Construction

[0244] The oligonucleotides listed below in Table 1 and synthetic genes listed in Table 2 were used to construct the plasmids with reductases for metabolic complementation. Table 3 lists the plasmids that were used and generated.

TABLE-US-00014 TABLE 1 Oligonucleotides used in this project Oligo ID Sequence (5′ .fwdarw. 3′) Source Description oligoLS19 CCGTTCGCATGCAGGAGGTAC IDT adhE SphI GAACACATGGCTGTTACTAA- DNA (F) SEQ ID No: 4 oligoLS20 GCTGAAGGATCCTTAAGCGGATTTTTTCG- IDT adhE SEQ ID No: 5 DNA BamHI (R) oligoLS21 CCGTTCG IDT budC CATGCCAATCTTAATCAAATCAGACAGA DNA SphI (F) GAGAGTACAATATGAAAAAAGTCGCAC- TTGT-SEQ ID No: 6 oligoLS22 TTCAGCGGATCCTTAGTTAAACAC IDT budC Bam- CATCCCGCCGTCGAT- DNA HI(R) SEQ ID No: 7 oligoLS23 CCGTTCGCATGCAGGAGGTAC IDT bdhA SphI GAACACATGAAGGCAGCAAGATG- DNA (F) SEQ ID No: 8 oligoLS24 GCTGAAGGATCCTTAG IDT bdhA Bam- TTAGGTCTAACAAGGATTTTGACT- DNA HI (R) SEQ ID No: 9 oligoLS87 GTTCGCATGCATTCGGATCTATACAGA IDT sadh TAAGGAGAAAGAGATGAAAGGCTTT Clostridium GCCATGCT- DNA NADPH SEQ ID No: 10 SphI (F) oligoLS88 CTTCCATGGATCCTCACTATTAGAGGA IDT sadh TAACTACGGCC- Clostridium SEQ ID No: 11 DNA NADPH BamHI (R) oligoLS112 CTTGGCGGCCTCAACGCAAA IDT CBADH TAGGNNNNNNNNNGACACCAA DNA random TAATCCGACCTGC- mutagenesis SEQ ID No: 12 198,199 and 200 (R) oligoLS113 TTCTACGGCGCGAC IDT CBADH random CGACATTCTGAATNNNAAAAATGGCCATAT DNA mutagene- TGTGGAC- sis 218 (F) SEQ ID No: 13 oligoLS162 GCTGAAGGATCCTTAG IDT BDAH 6x TGGTGGTGGTGGTGGTGGTTAGGTCTAAC DNA His tag AAGGATTTTGA- C-terminus SEQ ID No: 14 oligoLS163 GCTGAAGGATCCTTAG IDT CBADH 6x TGGTGGTGGTGGTGGTGGAGGATAAC DNA His tag TACGGCCTTAATGAGA- C-terminus SEQ ID No: 15 oligoLS168 CCGTTCGCATGCAGGAGGTAC IDT ADH from GAACACATG- DNA Thermus SEQ ID No: 16 sp ATN1 SphI F oligoLS169 TTCAGCGGATCCTTATCCGCGAACTACAA IDT ADH from GCAAT- DNA Thermus SEQ ID No: 17 sp ATN1 BamHI R oligoLS170 GCTGAAGGATCCTTAG IDT ADH from TGGTGGTGGTGGTGGTGTCCGCGAAC DNA Thermus sp TACAAGCAATACCT- ATN1 6x SEQ ID No: 18 His tag C-terminus oligoLS208 TTCAGCGGATCCAATGTATCTGCATGAA IDT sthA- GCACAGACCCACCAGTTACTGG- DNA pMAK705 SEQ ID No: 19 BamHI oligoLS209 TTCAGCaagatCATTAAAC IDT sthA- CGCTCTCATCAACCATGGTCAGACCCAG DNA pMAK705 TTCG- HindIII SEQ ID NO: 20 oligoLS216 TTCAGCGGATCCGAAACGAC IDT pntA- CAGAGCCGCCAGGTTCA- DNA pMAK705 SEQ ID No: 21 BamHI oligoLS218 TTCAGCaagatCAGGAGGGTGTTCTTAA IDT pntA- GCTTCATAAAAATAATCCTTCGCCTTGCGC- DNA pMAK705 SEQ ID No: 22 HindIII oligoLS228 AAGGGGTT IDT ADC-Lvo GGTCTCATGTGGCTCTTCGATGttaaaggatgaa DNA gtaattaaacaaattagcacg- SEQ ID No: 23 oligoLS229 AAGGGGTTGGTCTCTGGTCTTAC IDT ADC-Lvo GCTCTTCATTActtaagataatcata DNA tataacttcagctctaggc- SEQ ID No: 24 oligoLS232 AAGGGGTT IDT CBADH-Lvo GGTCTCATGTGGCTCTTCGATGaaaggcttt DNA gccatgctg- SEQ ID No: 25 oligoLS233 AAGGGGTTGGTCTCTGGTCTTAC IDT CBADH-Lvo GCTCTTCATTAgaggataactacggccttaatgag- DNA SEQ ID No: 26 oligoLS234 AAGGGGTT IDT AtoD-Lvo GGTCTCATGTGGCTCTTCGATGaaaacaaaatt DNA GATGACATTACAAGACG- SEQ ID No: 27 oligoLS235 AAGGGGTTGGTCTCTGGTCTTAC IDT AtoA-Lvo GCTCTTCATTAtaaatcaccccgttgcgtattc- DNA SEQ ID No: 28 oligoLS242 AAGGGGTT IDT AtoA-Lvo GGTCTCATGTGGCTCTTCGATGGATGCGAA DNA ACAACGTATTGCGC- SEQ ID No: 29 oligoLS243 AAGGGGTTGGTCTCTGGTCTTAC IDT AtoD-Lvo GCTCTTCATTATTTGCTCTCCTGTGAAAC DNA GATGATGTG- SEQ ID No: 30 oligoLS244 TTCAGCGGATCCTGTCTGTTTT IDT GCGGTCGCCAG- DNA bamHI SEQ ID No: 31 oligoLS245 TTCAGCaagcttCAAGCAGAATCAAGTTC IDT IdhA TACCGTGC- DNA pMAK705 SEQ ID No: 32 HindIII indicates data missing or illegible when filed

TABLE-US-00015 TABLE 7 Other oligonucleotides used in this project pLS98 oligoLS294 GCAGCCATATGatgaaaggctttgccatgctgggtattaacaaattagg- SEQ ID No: 42 oligoLS295 TTATTGCTCAGCTTAgaggataactacggccttaatgagatctttaggtttatctttcatgag- SEQ ID No: 43 pLS131 oligoLS344 ACGATAATATCGCTGCGTTTAAC-SEQ ID No: 44 oligoLS345 CTGGCAAAACTGGGCGCACATC-SEQ ID No: 45 oligoLS342 CGGTTCGCTACGGGCTTTTTCATATTCCCACACCGTGGTCG- SEQ ID No: 46 oligoLS343 GGTTAATGTGATTGATTATGACACCTCTGATCAGGTTCTGCGCCAAGAC- SEQ ID No: 47 pLS132 oligoLS344 ACGATAATATCGCTGCGTTTAAC-SEQ ID No: 48 oligoLS345 CTGGCAAAACTGGGCGCACATC-SEQ ID No: 49 oligoLS343 GGTTAATGTGATTGATTATGACACCTCTGATCAGGTTCTGCGCCAAGAC- SEQ ID No: 50 oligoLS346 CGGTTCGCTCGCGGCTTTTTCATATTCCCACACCGTGGTCG- SEQ ID No: 51 pLS133 oligoLS337 GCTGAgaagaccGACCACGGTGTGGNNNNNNNNNAAAGCCNNNA GCGAACCGCTGGCAAAACTG-SEQ ID No: 52 oligoLS338 GCTGAgaagaccgtGGTCGTGTAGCCAGATTGCAGGAATGCTTTAAT CAGTGCGGAGCCCATACGGCC-SEQ ID No: 53 pLS161 oligoLS358 tctctGAAGACTCCTTAGTGGTGGTGGTGGTGGTGTTTCAGGAAGC GGGTCAGAATTGCAAAG-SEQ ID No: 54 oligoLS359 tctctGAAGACAacATGAAACCGACCCTGACCGTTATTGGC- SEQ ID NO: 55 pLS162 oligoLS358 tctctGAAGACTCCTTAGTGGTGGTGGTGGTGGTGTTTCAGGAAGC GGGTCAGAATTGCAAAG-SEQ ID No: 56 oligoLS359 tctctGAAGACAacATGAAACCGACCCTGACCGTTATTGGC- SEQ ID No: 57 pLS164 oligoLS358 tctctGAAGACTCCTTAGTGGTGGTGGTGGTGGTGTTTCAGGAAGC GGGTCAGAATTGCAAAG-SEQ ID No: 58 oligoLS359 tctctGAAGACAacATGAAACCGACCCTGACCGTTATTGGC- SEQ ID No: 59 pLS169 oligoLS363 tctctGAAGACTCGGTGCTGGCTACAATGAAGTGCCACGGCTGGGA GTTNNNNNNGGACGGGCTGTACTGC-SEQ ID No: 60 oligoLS366 ctctGAAGACCAGTGGATGGCGAAGCAGGTTTACCTGAACGTCGG- SEQ ID No: 61 oligoLS364 ctctGAAGACAGCACCGAGGAAGGAAAAGCGCGCGTGGCGAAGTC CGCTGCGGGCACCNNNGTGTTCAACGAACG-SEQ ID No: 62 oligoLS365 tctctGAAGACATCCaCTGGTCGTCATCTTTCAGATCCACGCGGTGC ATGTCGGCNNNGTAGGTGCGGCC-SEQ ID No: 63 pLS46 oligoLS230 AAGGGGTTGGTCTCATGTGCTCTTCGatgaaaaattgtgtcatcgtcagtgcg gtacg-SEQ ID No: 64 oligoLS231 AAGGGGTTGGTCTCTGGTCTTACGCTCTTCAttaattcaaccgttcaatcac catcgcaattccc-SEQ ID No: 65 pLS47 oligoLS234 AAGGGGTTGGTCTCATGTGGCTCTTCGATGaaaacaaaattgatgacatt acaagacg-SEQ ID No: 66 oligoLS243 AAGGGGTTGGTCTCTGGTCTTACGCTCTTCATTATTTGCTCTCCT GTGAAACGATGATGTG-SEQ ID No: 67 pLS48 oligoLS235 AAGGGGTTGGTCTCTGGTCTTACGCTCTTCATTAtaaatcaccccgttgc gtattc-SEQ ID No: 68 oligoLS242 AAGGGGTTGGTCTCATGTGGCTCTTCGATGGATGCGAAACAACG TATTGCGC-SEQ ID No: 69 pLS49 oligoLS228 AAGGGGTTGGTCTCATGTGGCTCTTCGATGttaaaggatgaagtaattaa acaaattagcacg-SEQ ID No: 70 oligoLS229 AAGGGGTTGGTCTCTGGTCTTACGCTCTTCATTActtaagataatcatat ataacttcagctctaggc-SEQ ID No: 71 pLS50 oligoLS232 AAGGGGTTGGTCTCATGTGGCTCTTCGATGaaaggctttgccatgctgggt attaac-SEQ ID No: 72 oligoLS233 AAGGGGTTGGTCTCTGGTCTTACGCTCTTCATTAgaggataactacggc cttaatgagatctttagg-SEQ ID No: 73 pLS63 oligoLS244 TTCAGCGGATCCTGTCTGTTTTGCGGTCGCCAG-SEQ ID No: 74 oligoLS247 CACTGGAGAAAGTCTTATGTAATCTTGCCGCTCCCCTGCATTCCAG- SEQ ID No: 75 oligoLS245 TTCAGCaagcttCAAGCAGAATCAAGTTCTACCGTGC-SEQ ID No: 76 oligoLS246 CAGGGGAGCGGCAAGATTACATAAGACTTTCTCCAGTGATGTTG AATC-SEQ ID No: 77 pLS39 oligoLS208 TTCAGCGGATCCAATGTATCTGCATGAAGCACAGACCCACCAGT TACTGG-SEQ ID No: 78 oligoLS210 AACAGGTAAGCCCTACCATGTAAAACTTTATCGAAATGGCCATC CATTCTTGCGCGG-SEQ ID No: 79 oligoLS209 TTCAGCaagcttCATTAAACCGCTCTCATCAACCATGGTCAGACCCA GTTCG-SEQ ID No: 80 oligoLS211 GCCATTTCGATAAAGTTTTACATGGTAGGGCTTACCTGTTCTTAT ACATAAAAGCAACAGAATGG-SEQ ID No: 81 pLS40 oligLS216 TTCAGCGGATCCGAAACGACCAGAGCCGCCAGGTTCA- SEQ ID No: 82 oligLS217 CCGATGGAAGGGAATATCATGTAAGGGGTAACATATGTCTGGAG GATTAGTTACAGCTGCATACATTGTTGCCGC-SEQ ID No: 83 oligoLS218 TTCAGCaagcttCAGGAGGGTGTTCTTAAGCTTCATAAAAATAATC CTTCGCCTTGCGCAAA-SEQ ID No: 84 oligoLS219 CCAGACATATGTTACCCCTTACATGATATTCCCTTCCATCGGTTT TATTGATG-SEQ ID No: 85

[0245] Dehydrogenase genes were amplified by PCR from either genomic DNA or gBlock synthetic DNA (IDT) (see Table 2) by using the corresponding oligonucleotides. The obtained PCR products were digested with SphI and BamHI restriction enzymes and then ligated with pUC19 using T4 DNA ligase. pUC19 was previously linearised by using the same restriction enzymes.

TABLE-US-00016 TABLE 2 List of Synthetic genes (gBlock) gBlock ID Sequence (5′ .fwdarw. 3′) Source Description gBlockLS3 CCGTTCG IDT SphI/BamHI- CATGCCAATCTTAATCAAATCAGACAGA DNA sadh Klebsiella GAGAGTACAATATGAAAAAAGTCGCAC pneumoniae TTGTTACCGGCGCCGGCCAGGGGATTGG TAAAGC TATCGCCCTTCGTCTGGTGAAGGATGGAT TTGCCGTGGCCATTGCCGATTATAAC GACACCACCGCCAAA GCGGTCGCCTCCGAAATCAAC CAGGCCGGCGGCCGCGCCATGGCGGTGA AAGTGGATGTCTCCGAC CGCGATCAGGTGTTT GCCGCCGTCGAACAGGCGCGCAAAAC GCTGGGCGGCTTCGAC GTCATCGTCAACAAC GCCGGCGTGGCGCCGTCCAC GCCGATCGAGTCCATTACCCCGGA GATTGTCGATAAAGTCTACAACATCAAC GTTAAAGGGGTGATCTGGGG CATTCAGGCGGCGGTCGAGGCCTTTAA GAAAGAGGGTCACGGCGG GAAAATCATCAAC GCCTGTTCCCAGGCCGGCCACGTCGG CAACCCGGAGCTGGCGGTATA TAGCTCGAGTAAATTCGCCGTAC GCGGCTTAACCCAGAC CGCCGCTCGCGACCTCGCGCCGCTGGG CATCACAGTCAACGGCTACTGCCCGGG GATTGTCAAAAC GCCAATGTGGGCCGAAATTGAC CGCCAGGTGTCCGAAGCCGCCGGTAAAC CGCTGGGTTACGGTACCGCCGAG TTCGCCAAAC GCATCACCCTCGGCCGCCTGTCCGAGCCG GAAGATGTCGCCGCCTGCGTCTCC TATCTTGCCAGCCCGGATTCTGATTA TATGACCGGTCAGTCATTGCTGATCGAC GGCGG GATGGTGTTTAACTAAGGATCCGCTGAA- SEQ ID No: 33 gBlockLS10 CCGTTCGCATGCAGGAGGTAC IDT RE: GAACACATGAAACCGACCCTGAC DNA Sphl/BamHI- CGTTATTGGCGCTGGCCGTATGGGCTCCG IREDs Myxococcus CACTGATTAAAGCATTCCTGCAATCTGGC stipitatus TACACGACCACGGTGTGGAACCGTACCAAA GCCAAAAGCGAACCGCTGG CAAAACTGGGCGCACATCTGGCTGATAC GGTGCGTGACGCCGTTAAACGCAGCGA TATTATCGTGGTTAATGTGCTGGAT TATGACACCTCTGATCAGCTGCTGCGCCAA GACGAAGTGACGCGTGAACTGCGCGG CAAACTGCTGGTTCAGCTGAC CAGCGGTTCTCCGGCAC TGGCTCGTGAACAGGAAAC GTGGGCGCGCCAACATGGCATTGAT TATCTGGACGGTGCGATCATGGCCACCCCG GAT TTTATTGGCCAGGCAGAATGCGCTCTGCTG TACAG TGGTTCCGCGGCCCTGTTCGAAAAACAC CGTGCTGTCCTGAATGTGCTGGGCGGTGCCA CCAGCCATGTCGGCGAAGATGTT GGTCATGCCTCAGCACTG GACAGCGCCCTGCTGTTTCAGATGTGGGG CACCCTGTTCGGTACGCTGCAAGCACTGGC TATTTCTCGCGCAGAAGGCATCCCGCTG GAAAAAACCACGGCGTTTATCAAACTGAC CGAACCGGTCACCCAGGGTGCCGTT GCAGATGTCCTGACCCGTGTTCAG CAAAATCGCCTGACCGCAGACGCTCAGAC GCTGGCAAGTCTGGAAGCTCATAAC GTGGCGTTCCAACAC CTGCTGGCCCTGTGTGAAGAACGTAA TATCCATCGCGGTGTTGCG GATGCCATGTACTCCGTTATTCGTGAA GCGGTCAAAGCCGGCCACGGTAAA GATGACTTT GCAATTCTGACCCGCTTCCTGAAA TAAGGATCCTTCAGC- SEQ ID No: 86 gBlockLS12 catctGAAGACAacATGGA IDT bbsl/bbsl TATCATTTCTGTCGCCCTGAAACGCCACTC DNA Nitroreductase TACCAAGGCGTTCGACGCAA Enterobacter cloacae GCAAAAAACTGACCGCGGAAGAAGCG GAAAAAATCAAAACCCTGCTGCAG TACAGCCCGTCCAGCAC CAACTCCCAGCCGTGGCACTTCATT GTAGCCAGCACCGAGGAAGGAAAA GCGCGCGTGGCGAAGTCCGCTGCGGGCAC CTATGTGTTCAACGAACGCAAAATGCTG GATGCTTCCCAC GTGGTGGTGTTCTGCGCGAAAAC CGCGATGGATGACGCCTGGCTG GAGCGCGTCGTGGATCAGGAA GAGGCCGATGGCCGTTTCAACACGCCGGAA GCCAAAGCCGCAAACCATAAGGGCCGCAC CTACTTCGCCGACATGCACCGCGTG GATCTGAAAGATGACGACCAGTG GATGGCGAAGCAGGTTTACCTGAACGTCGG CAACTTCCTGCTGGGCGTGGGCGCGATGGGT CTGGACGCGGTACCAATTGAAGGTTTCGAC GCCGCTATTCTCGACGAAGAGTTT GGCCTGAAAGAGAAAGGCTTCAC CAGCCTGGTGGTGGTACCGGTTGGGCAC CACAGCGTGGAAGATTTCAACGCCAC GCTGCCGAAATCTCGCCTGCCGCTGAGCAC GATTGTGACCGAGTGCTAAGGAGTCTTCaga ga-SEQ ID No: 87

TABLE-US-00017 TABLE 3 Plasmids used in this project Antibiotic Plasmid Description Resistance Reference pUC19 High copy expression vector (pMB1 Amp Yanisch-Perron, Vieira, & ORI) with a lacZα Messing, 1985 pCP20 Contains FLP recombinase Amp Cherepanov & Temperature-sensitive ORI Wackernagel, 1995 pMAK705 Contains pSOC1 a thermo-sensitive CatP Hamilton et al 1989 ORI pJET 1.2 Ready selection cloning vector Amp Agdanaviciute, Zakareviciene, & Lubys, 2007 (Unpublished) pStAo Combinatorial built plasmid level o Amp G. Taylor & J. Heap (Unpublished) pStA1 Combinatorial built plasmid level 1 tetR G. Taylor & J. Heap (Unpublished) pStA2 Combinatorial built plasmid level 2 KanR G. Taylor & J. Heap (Unpublished) pLS1 pUC19 with ADH NADH-dependent Amp This work from Escherichia coli (adhE) pLS2 pUC19 with sADH NADH- Amp This work dependent from Bacillus subtilis (bdhA) pLS3 pUC19 with sADH NADH- Amp This work dependent from Klebsiella pneumoniae (budC) pLS6 pUC19 with ADH NADPH- Amp This work dependent from clostridium beijerinckii pLS10 Same as pLS6 but with 4 AAs Amp This work mutated (library) pLS11 pUC19 with ADH from Thermus sp Amp This work ATN16xHis tag on the C-terminus of the TADH pLS12 pUC19 with ADH from Thermus sp Amp This work ATN1 pLS25 pUC19 with sADH NADH- Amp This work dependent from Bacillus subtilis (bdhA) but with6xHis tag on the C- terminus of bdhA pLS26 Same as pLS6 but with 6xHis tag on Amp This work the C-terminus of CBADH pLS39 pMAK705- to knock out sthA CatP This work pLS40 pMAK705- to knock out pntA CatP This work pLS46 pStAo- atoB Amp This work pLS47 pStAo- atoD Amp This work pLS48 pStAo-atoA Amp This work pLS49 pStAo-ADC Amp This work pLS50 pStAo-CBADH_WT Amp This work pLS51 pStAo- CBADH_variant Amp This work pLS53 pStA1AB- atoB with library of tetR This work promoters (Anderson promters) and RBS pLS54 pStA1BC- atoD with library of tetR This work promoters (Anderson promters) and RBS pLS55 pSt1CD- atoA with library of tetR This work promoters (Anderson promters) and RBS pLS56 pStA1DE -ADC with library of tetR This work promoters (Anderson promters) and RBS pLS57 pStA1EZ-CBADH_WT with library tetR This work of promoters (Anderson promters) and RBS pLS58 pStA1EZ- CBADH_variant with tetR This work library of promoters (Anderson promters) and RBS pLS60 pStA212- Library of the IPA KanR This work pathway with CBADH-variant pLS61 pStA212- pStA212- Library of the KanR This work IPA pathway with CBADH-WT pLS63 pMAK705- to knock out ldhA CatP This work

[0246] DH5α Escherichia coli cells were transformed with the plasmid of interest and cultured on LB agar plates. Then single colonies were picked to do 5 mL overnight cultures. Overnight cultures were spun down and the pellets were used to extract the plasmids of interest by using QIAprep Miniprep kit (Qiagen). Plasmids were sequenced by Source BioScience.

[0247] Bacterial Strains and Culture Conditions

[0248] A list of Escherichia coli strains used in the study is shown in Table 4.

TABLE-US-00018 TABLE 4 Escherichia coli strains used in this project Antibiotic Strain Description Resistance Reference DH5a F−, φ8odlacZΔM15, Δ(lacZYA- None Grant et al, argF)U169, deoR, recA1, endA1, 1990 hsdR17(rK−, mK+), phoA, supE44, λ−, thi−1, gyrA96, relA1 BW25113 F−, DE(araD-araB)567, None Datsenko & lacZ4787(del)::rrnB−3, λ−, rph−1, Wanner, DE(rhaD-rhaB)568, hsdR514 2000 LS1 F−, DE(araD-araB)567, None This work lacZ4787(del)::rrnB−3, λ−, rph−1, DE(rhaD-rhaB)568, hsdR514, ΔadhE, ΔldhA LS2 F−, DE(araD-araB)567, Kan This work lacZ4787(del)::rrnB−3, λ−, rph−1, DE(rhaD-rhaB)568, hsdR514, ΔadhE, ΔldhA LS5 F−, DE(araD-araB)567, Kan This work lacZ4787(del)::rrnB−3, λ−, rph−1, DE(rhaD-rhaB)568, hsdR514, ΔadhE, ΔldhA, ΔsthA, ΔpntA ΔldhA- F−, DE(araD-araB)567, Kan Baba et al, JW1375 lacZ4787(del)::rrnB−3, λ−, rph−1, 2006 DE(rhaD-rhaB)568, ΔadhE- F−, DE(araD-araB)567, Kan Baba et al, JW1228 lacZ4787(del)::rmB−3, λ−, rph−1, 2006 DE(rhaD-rhaB)568, hsdR514

[0249] Escherichia coli strains were grown in Luria-Bertani broth (LB) at 37° C. with shaking at 250 rpm, or on LB agar plates containing the corresponding antibiotic.

[0250] Construction of Escherichia coli Selection Strains (LSI and LS2)

[0251] Standard methods using pMAK705 (Hamilton et al 1989) and pCP20 (Cherepanov & Wackernagel, 1995) were used to construct the double mutant strains, triple mutant strains, and the quadruple mutant strain.

[0252] Metabolic Complementation

[0253] The LS1 mutant strain and the parental BW25113 strain were transformed with the desired plasmid and overnight pre-cultures were grown aerobically in 15 mL Falcon tubes with M9 media (0.4% glucose). These pre-cultures were used to inoculate 10 mL Hungate tubes with M9 medium (0.4% glucose) supplemented with 100 μg/mL ampicillin, 1 mM IPTG and with or without the specific substrate of the dehydrogenase under anaerobic conditions at 37° C. Metabolic complementation was assessed by measuring the optical density at 600 nm every two hours during daytime.

[0254] Results

EXAMPLE 1

NAD+ Regeneration Alone can Rescue Fermentative Growth of an adhE/ldhA Mutant

[0255] To design a system able to select specific enzyme variants depending on their ability to transfer electrons from NADH to a specific substrate, the inventors constructed an Escherichia coli strain, LS1, which is unable to grow under anaerobic conditions due to impaired fermentative pathways. This was achieved by deleting the genes encoding alcohol dehydrogenase (adhE) and lactate dehydrogenase (ldhA), which are essential for ethanol fermentation and lactic acid fermentation, respectively. If only adhE was deleted, cells might adapt to be able to grow. This is due to the fact that cells can in principle grow anaerobically simply by converting glucose to lactate, since the chemical stoichiometry is balanced. Only natural regulation prevents this in cells in which only adhE is mutated. Thus, deletion of ldhA in addition to adhE excludes the possibility of cells adapting to grow by lactate fermentation, in order that cells which are successfully complemented are able to grow exclusively because of the transformed oxidoreductase, and that they won't be able to grow if they are not transformed with an active variant. An article by Chang et al reported an Escherichia coli mutant which metabolizes glucose exclusively by means of lactic fermentation (Chang et al, 1999), which is a form of anaerobic fermentation known to occur naturally in other organisms. The growth of strain LS1 under aerobic conditions was unaltered from the wild type, but it was unable to grow anaerobically. Complementation by transformation with pLS1 (containing the endogenous adhE gene) resulted in restoration of ability to grow anaerobically, with cells transformed with this plasmid being able to grow as efficiently as wild-type cells under anaerobic conditions (FIG. 2a). HPLC-RID of the fermentation broth of these cultures confirmed the profile of produced metabolites was similar to that of the parental strain, except for the absence of lactate (FIG. 2b).

[0256] Without wishing to be bound to any particular theory, the inventors hypothesized that the main reason fermentative growth was impaired in strain LS1 was the lack of regeneration of oxidized NAD+, necessary for anaerobic glycolysis to continue. Thus, fermentative anaerobic growth recovery would be achievable by transforming cells with a plasmid containing any exogenous NADH-dependent oxidoreductase and culturing them in minimal M9 medium supplemented with the appropriate oxidized substrate for the exogenous enzyme. To confirm this hypothesis, the inventors tested metabolic complementation with several exogenous enzymes. First, strain LS1 cells were transformed with pLS2 and pLS3, both of them containing 2,3-butanediol dehydrogenases from Bacillus subtilis (bdhA) and Klebsiella pneumoniae (budC), respectively. Both enzymes are able to catalyse the reduction of acetoin coupled to the oxidation of NADH. When transformed cells were grown anaerobically, growth recovery was achieved if the medium was supplemented with acetoin. Interestingly, anaerobic growth recovery was also observed when mutant cells transformed with a control plasmid not containing any exogenous enzyme were supplemented with acetoin. However, growth rate of cells was much slower than when they were transformed with pLS2 and pLS3. These results suggest that the Escherichia coli genome encodes an endogenous enzyme able to catalyse the reduction of acetoin coupled to the oxidation of NADH. The longer time needed for growth recovery compared to the cells transformed with the exogenous enzymes indicate that the endogenous enzyme has a low activity towards acetoin, or alternatively its expression level is low. A potential candidate for this enzyme is YohF, a putative oxidoreductase which has been predicted to be an acetoin reductase based on sequence similarity with confirmed acetoin reductases (Reed et al, 2003).

[0257] To completely rule out that growth recovery was mostly due to the activity of an endogenous enzyme, the inventors tested another exogenous enzyme, alcohol dehydrogenase from Thermus sp. ATN1 (TADH), which is able to act on a broad range of substrates (Höllrigl et al, 2008). The inventors chose two substrates towards which no endogenous Escherichia coli enzyme was described to have any activity: cyclohexanone and 3-methylcyclohexanone. Cells transformed with pLS12 (containing TADH) were able to grow anaerobically when media was supplemented with either of the two substrates, but no growth recovery was observed when cells were transformed with a control plasmid (FIG. 3a). Enzymatic activity assays with raw cell lysate obtained from the cultures showed that the enzymes were properly expressed and able to reduce both cyclohexanone and 3-methylcyclohexanone, with the activity towards cyclohexanone being 2.5 times greater than towards 3-methylcyclohexanone (FIG. 3c). The lower activity with the latter substrate, together with its higher toxicity, explains why cultures supplemented with 3-methylcyclohexanone reach a lower maximum cell density than when supplemented with cyclohexanone. Enzymatic activity towards each substrate was confirmed by assays performed with purified enzyme. Moreover, while HPLC-RID confirmed that neither ethanol nor lactate were produced, gas chromatography revealed that cultures transformed with pLS12 consumed all of the provided cyclohexanone/3-methylcyclohexanone, and converted it into the corresponding alcohol (FIG. 3b). These findings support that growth recovery was due to regeneration of oxidized NAD+ by the activity of the exogenous enzyme. A stoichiometric conversion of the substrates into the alcohols was not detected; this is probably caused by the volatile nature of both substrates and products which facilitates losses due to evaporation.

EXAMPLE 2

Evolution of a Novel NADH-Dependent Acetone Reductase by Selection for NAD+ Regeneration

[0258] CBADH is an oxidoreductase able to oxidize isopropanol and reduce acetone characterized by its very high preference for NADP(H) over NAD(H) (Korkhin et al, 1998). In order to test the potential of the described system as a selection tool, the inventors decided to attempt to obtain a variant of CBADH with reversed cofactor specificity, based on the rationale that such a variant would allow for a more efficient growth recovery.

[0259] They first tested if wild-type CBADH was able to achieve metabolic complementation.

[0260] When LS1 strain (ΔadhE ΔldhA double mutant) cells transformed with pLS6, containing the wild type enzyme, were cultured anaerobically in media supplemented with acetone, growth recovery was only observed after 55 hours, a much longer period than LS1 cells transformed with NAD(H)-dependent enzymes. The inventors hypothesized that the very slow metabolic complementation with wild type CBADH might have been due to the activity of a transhydrogenase, which would use the accumulated pool of reduced NADH to reduce NADP+, generating NAD+ and NADPH. The genome of Escherichia coli contains two transhydrogenase genes: sthA and pntA. The inventors decided to test the effect of knocking out both genes.

[0261] The generation of the library of variants of CBADH took the available structural information as the starting point. Korkhin and collaborators (Korkhin et al, 1998) solved the crystal structure of CBADH and identified a set of 4 amino acid residues (G198, S199, R200 and Y218) potentially critical for the specificity of the enzyme towards NADP(H). All 4 residues made contacts with the 2′-phosphate oxygens of NADP(H) and were conserved in other NADP(H)-dependent alcohol dehydrogenases. The inventors made and tested the specific variant described in Korkhin et al, but found that it did not work. Thus, the inventors decided to generate a library of CBADH variants by using a standard PCR-based method to perform saturation mutagenesis of the codons corresponding to these 4 amino acid residues.

[0262] LS1 strain cells transformed with three independently-generated libraries and grown anaerobically in media supplemented with acetone required only 24 hours on average to reach exponential phase of growth, a much shorter period than the required for cells transformed with the wild type CBADH. Clones were was isolated from the three anaerobic cultures and plasmid DNA was prepared, resulting in pLS10_1, pLS10_2 and pLS10_3, respectively. Transforming LS1 strain cells with pLS10_1, pLS10_2 and pLS10_3 allowed growth recovery under anaerobic conditions in media supplemented with acetone. Furthermore, GC analysis of the fermentation broth confirmed the presence of isopropanol in cultures transformed with pLS10_1 at much higher levels than in those transformed with pLS6, which correlated with the absence of acetone.

[0263] Sequencing of pLS10_1, pLS10_2 and pLS10_3 revealed that all of them encoded the same CBADH variant, which contained 8 point mutations in the DNA sequence resulting in 3 amino acid residue substitutions at the protein level: G198D, S199Y and Y218P. Enzymatic activity assays with purified enzyme showed a 4.6-fold increase in activity for the reduction of acetone to isopropanol with NADH as the cofactor when compared to the wild type, and io-fold increase for the oxidation of isopropanol to acetone with NAD+ as the cofactor. Interestingly, the new variant showed no significant activity for both the reduction and oxidation reactions when NADP(H) was provided as the cofactor. Surprisingly, even though Korkhin et al predicted an R200G mutation to be one of the substitutions most likely to have the effect of cofactor specificity reversal, this residue remained unchanged in our NAD(H)-dependent variant. Moreover, none of the substitutions found for the other 3 residues matched those suggested in the Korkhin et al study. However, the G198D mutation has been found to switch the cofactor specificity of Thermoanaerobacter brockii and Clostridium autoethanogenum alcohol dehydrogenases from NADP(H) to NAD(H) (Maddock, Patrick & Gerth, 2015). Indeed, structure-based alignment of several NADP(H)-dependent and NAD(H)-dependent dehydrogenases revealed that the residue at position 198 is always acidic in NAD(H)-dependent dehydrogenases. In the same study, it was shown that position 218 is frequently an alanine, serine or proline in NAD(H)-dependent dehydrogenases.

[0264] Interestingly, the Cofactory server for identification of cofactor specificity of Rossmann folds based on their amino acid sequence (Geertz-Hansen et al, 2014) was not able to determine if the wild type enzyme would bind preferentially NAD(H) or NADP(H), but it predicted correctly that our NAD(H)-dependent variant had a preference for NAD(H). On the other hand, CSR-SALAD, a recently-developed tool to predict mutations to reverse nicotinamide cofactor specificity reversal (Cahn et al, 2017), correctly identified residues 198, 199 and 218 as recommended targets to attempt cofactor specificity reversal. However, none of the suggested mutations for positions 199 and 218 matched those found in our variant; only for position 198 the recommendations included a substitution for an Asp residue.

[0265] In order to try to understand why these mutations led to cofactor specificity reversal, the inventors generated a structural model of the mutated protein by using the structure of the wild type enzyme as the template with the SWISS-MODEL server. Comparison of the wild type structure with the model of the mutant enzyme revealed some information about the structural basis for the cofactor specificity reversal. The substitution of G198 by an aspartate residue placed a negatively charged sidechain in close proximity of the 2′ phosphate group of NADPH, which very likely contributes to the inability of the mutant enzyme to accept NADP(H) as the cofactor. Furthermore, the small side chain of S199 is in a position where it does not pose any impediment to the binding of NADP(H), and possibly could form a hydrogen bond with its 2′ phosphate group. In the mutant enzyme, it is replaced by a tyrosine residue, with a much bulkier sidechain which is not predicted to be placed in a position where it could form a hydrogen bond with the 2′ phosphate. Finally, the reason why the Y218P substitution contributed to cofactor specificity reversal remains unclear, since this residue is not located in the vicinity of the 2′ phosphate, but instead contacts the adenine ring moiety.

[0266] FIG. 15 summarises the workflow of producing the variant CBADH and shows the crystal structure of pLS10 3 bound to NADH, obtaining insight into the structural basis of cofactor preference reversal.

[0267] A summary of the NMR spectra confirming the formation of isopropanol both when transforming with the library or pLS10 1, pLS10_2, pLS10_3 can be seen in Table 5, and Table 6 shows that pLS10_1, pLS10_2, pLS10_3 comprising the variant gained activity with NADH and activity with NADPH had been lost (Table 6).

TABLE-US-00019 TABLE 5 1H NMR analysis of fermentation broth of anaerobic cultures δ of δ of charac- charac- Trans- teristic teristic Sub- Succin- formed signal of signal of strate Product Ethanol Lactate ate Acetate Formate plasmid/ Encoded Exogenous Resulting substrate product concen- concen- concen- concen- concen- concen- concen- library enzyme(s) substrate product (ppm) (ppm) tration tration tration tration tration tration tration pLS1 adhE — — — — — — 13.8 0 5.2 12.1 20.7 pLS2 bdhA acetoin 2,3- 1.38 (d, 3) 1.15 (d, 6) 0 8.4 0 0 0.5 6.7 3.7 butanediol pLS3 budC acetoin 2,3- 1.38 (d, 3) 1.15 (d, 6) 0.2 8.3 0 0 0.7 7.9 5 butanediol pLS6 CBADH acetone isopropanol 2.24 (s, 6) 1.18 (d, 6) 0.1 10.8 0 0 1.1 8.1 5.8 pLS10_3 CBADH- acetone isopropanol 2.24 (s, 6) 1.18 (d, 6) 0.5 12.7 0 0 1.1 13 9.9 s pLS130 MsIRED 2-methyl-1- 2- 2.43 (s, 3) 1.38 (d, 3) 5.5 6.6 0 0 1.2 4.3 2.1 pyrroline methyl- pyrrolidine pLS131 MsIRED- 2-methyl-1- 2- 2.43 (s, 3) 1.38 (d, 3) 3.45 9.2 0 0 1.7 10.7 7.9 c pyrroline methyl- pyrrolidine pLS133s1 MsIRED- 2-methyl-1- 2- 2.43 (s, 3) 1.38 (d, 3) 0.8 13.4 0 0 2 15.9 12.8 s pyrroline methyl- pyrrolidine pLS168 EntNFSB 2- ? 8.10 (d, 1) ? 2.38 (s, ?) 14.1 ? 0 0 0 0.7 0.4 nitro- benzoic acid pLS168 EntNFSB 4- ? 8.27 (d, 2) ? 2.38 (s, ?), 8.28 ? 0 0 0 0.8 0.4 nitro- 8.01 (d, ?) benzylic alcohol pLS169s1 EntNFSB- 2- ? 8.10 (d, 1) ? 2.38 (s, ?) 12.4 ? 0 0 0.3 4.4 2.9 s1 nitro- benzoic acid pLS169s2 EntNFSB- 4- ? 8.27 (d, 2) ? 2.38 (s, ?), 0 ? 0 0 1 13.2 7.9 s2 nitro- Multiple benzylic signals bet- alcohol ween 6.5 and 8 ppm

TABLE-US-00020 TABLE 6 Kinetics of evolved and parental enzymes kcat/Km Enzyme Variable Substrate (min-1 Ki concen- Enzyme substrate Cofactor Km kcat mM-1) (substrate) tration CBADH Isopropanol NADP + (1 mM)  5.80 mM 1185.6 min-1  204.6 — 110 nM CBADH Isopropanol NAD+ ND ND ND — 110 nM CBADH-s Isopropanol NADP+ ND ND ND — 110 nM CBADH-s Isopropanol NAD+ (10 mM) 17.49 mM  333 min-1 19 — 110 nM MsIRED 2-methylpyrroline NADPH (0.25 mM)  3.56 mM 89.8 min-1 25.2 18.05 mM 1.2 uM MsIRED 2-methylpyrroline NADH ND ND ND — 1.2 uM MsIRED-c 2-methylpyrroline NADH (0.25 mM) 41.79 mM 119.6 min-1  2.9  4.21 mM 1.2675 uM MsIRED-s 2-methylpyrroline NADH (0.25 mM) 19.57 mM 78.1 min-1 4 11.42 mM 1.25 uM EntNFSB 2-nitrobenzoic NADH N.D. ? N.D. N.D. N.D. 1 uM acid EntNFSB-s1 2-nitrobenzoic NADH 4.054 1.17 min-1 0.29 mM 1 uM acid EntNFSB 4-nitrobenzyl NADH  9.52 mM 8.76 min-1 0.92 mM 1 uM alcohol EntNFSB-s2 4-nitrobenzyl NADH 1.111 mM 4.35 min-1 3.92 mM 1 uM alcohol

EXAMPLE 3

Integration of NADH-Dependent CBADH into an Isopropanol Production Pathway

[0268] Hanai et al engineered a synthetic pathway for isopropanol production in Escherichia coli (Hanai, Atsumi & Liao, 2007) by expressing five genes from a combination of organisms in Escherichia coli: Escherichia coli acetyl-CoA acetyltransferase (atoB), Clostridium acetobutylicum acetoacetate decarboxylase (adc), Escherichia coli acetyl-CoA:acetoacetyl-CoA transferase (atoAD) and CBADH. The pathway is summarized in FIG. 5. The obtained yield from this recombinant strain cultivated anaerobically was 43.5% (mol/mol), exceeding the yields obtained even from native producers.

[0269] The inventors hypothesized that the yield could be increased by culturing cells under anaerobic fermentation conditions and substituting the wild type CBADH previously employed by the NAD(H)-dependent variant identified with our selection method. Under anaerobic fermentation conditions, reduced NADH cannot be used to reduce an external electron acceptor such as molecular oxygen, so a large fraction of the NADH generated by glycolysis would be used by the CBADH variant to produce isopropanol, and the yield of isopropanol obtained could approach the theoretical maximum.

EXAMPLE 4

Selection Strain Specific for NADH-Dependent Oxidoreductases

[0270] As LS1 (ΔadhE-ΔldhA double mutant) was able to grow when transformed with an NADPH-dependent oxidoreductase (wild type CBADH), the inventors generated two triple mutants where, in addition to adhE and ldhA, one transhydrogenase gene was deleted in each. Transhydrogenases catalyse the direct transfer of electrons from NADH to NADP+ and from NADPH to NAD+, in the following reaction: NADH+NADP.sup.+=NAD.sup.++NADPH. Without wishing to be bound to any particularly theory, the inventors hypothesized the activity of these transhydrogenases is what makes the system able to restore anaerobic growth when transformed with enzymes that generate NADP+. Two triple mutants, where one transhydrogenase gene was knocked out in addition to adhE and ldhA, were generated, since there are 2 transhydrogenase genes in E. coli: [0271] sthA (soluble transhydrogenase) [0272] pntA (transmembrane transhydrogenase).

[0273] When metabolic complementation was attempted with an NADPH dependent alcohol dehydrogenase (wild type CBADH) with any of the triple mutants, cells were still able to grow anaerobically.

[0274] Since in the triple mutants the non-deleted transhydrogenase could still be supporting anaerobic growth under anaerobic fermentation conditions when complemented with an NADPH-dependent oxidoreductase by generating oxidized NAD, the inventors generated a quadruple mutant where adhE, ldhA, sthA and pntA genes were deleted (LS5 strain, FIG. 11).

[0275] This strain displays the following features (FIG. 12): [0276] It is unable to grow anaerobically. [0277] When transformed with plasmid containing adhE, anaerobic growth is restored. [0278] When transformed with plasmid containing NADPH-dependent oxidoreductase (pLS6, containing wild type CBADH), anaerobic growth is not restored, independently of whether the media is supplemented with acetone or not. [0279] When transformed with plasmid containing an NADH-dependent variant of CBADH (pLS10_3), anaerobic growth is restored if the media is supplemented with acetone (the substrate of the enzyme).

[0280] This shows that the reason the double mutant is able to grow anaerobically when transformed with an enzyme that generates oxidized NADP, is the activity of transhydrogenases that use NADP to generate NAD. The LS5 strain can thus be used as a more strict selection system: to select strictly for enzymes which regenerate oxidized NAD, and not either NAD or NADP, as is the case when using LS1 strain.

[0281] The inventors tested the suitability of four Escherichia coli mutant strains for use in the selection method, and these strains were:

[0282] LS1=AL (ΔadhE ΔldhA): the main strain we use in the selection system, with metabolic defects that make them unable to grow under anaerobic fermentation conditions due to their inability to regenerate oxidized NAD+.

[0283] LS2=AL (ΔadhE:Kan ΔldhA): metabolic defects that make them unable to grow under anaerobic fermentation conditions due to their inability to regenerate oxidized NAD+.

[0284] LS3=ALS (ΔadhE ΔldhA ΔsthA): triple mutant with sthA transhydrogenase mutated.

[0285] LS4=ALP (ΔadhE ΔldhA ΔpntB): triple mutant with pntB transhydrogenase mutated.

[0286] LS5=ALPS (ΔadhE ΔldhA ΔpntB ΔsthA): quadruple mutant with both transhydrogenases mutated.

[0287] The inventors demonstrated that all four strains tested were suitable for the selection system described (FIGS. 13 and 14). However, the ALPS strain cannot grow under anaerobic fermentation conditions when transformed with a gene encoding enzymes that cause the generation of oxidized NADP+. All three of the other strains can be complemented with an NADP+regenerating enzyme, although it takes a longer time to observe anaerobic growth. Without wishing to be bound to any particular theory, the inventors conclude that the metabolic complementation observed in AL, ALS and ALP strains with NADP+ dependent enzymes is mediated by transhydrogenases (both sthA and pntB are suitable for it), which use the oxidized NADP+ to generate oxidized NAD+. In the ALPS strain, this is not possible, since both transhydrogenases are knocked out (pLS6 encodes wild type CBADH, which can only use NADPH; pLS10_3 encodes CBADH-variant, which can only use NADH).

[0288] The selection system was validated using these strains, and is summarised as follows:

[0289] Metabolic complementation was achieved when cells where transformed with the following:

[0290] Native E. coli adhE=pLS1 (FIG. 14c)

[0291] budC=pLS3 (acetoin reductase from Klebsiella pneumoniae), and acetoin was added to the culture (FIG. 14d)

[0292] bdhA=pLS2 (acetoin reductase from Bacillus subtilis), and acetoin was added to the culture (FIG. 14d)

[0293] TADH=pLS12 (alcohol dehydrogenase from Thermus sp. ATN1), and cyclohexanone was added to the culture (FIG. 14f)

[0294] TADH, and 3-methylcyclohexanone was added to the culture (FIG. 14f)

[0295] In all cases, formation of the expected reduced products was confirmed with NMR (FIGS. 14e and 14g, Table 5).

EXAMPLE 5

Generation of Imine Reductase Variants

[0296] Imine reductases (IREDs) are able to catalyse the reduction of imines and reductive amination of ketones with high enantiospecificity and regiospecificity. No naturally occurring IRED that is able to utilise NADH for catalyzing their reaction is known.

[0297] Two previous studies have obtained mutant IREDs that display activity with NADH, by means of screening methods:

[0298] 1) A variant of IRED from Streptomyces GF3587 (IR-Sgf3587), with a K40A substitution (A NADH-accepting imine reductase variant: Immobilization and cofactor regeneration by oxidative deamination, Journal of Biotechnology, vol 230, 20 Jul. 2016, pages 11-18).

[0299] 2) Several variants of Myxococcus stipitatus IRED (MsIRED) (SEQ ID No:34), with the best one containing 5 residue substitutions, reached after several rounds of mutagenesis and screening (Switching the Cofactor Specificity of an Imine Reductase, CHEMCATCHEM, Vol 10, issue 1, pages 183-187).

[0300] The inventors generated a library of MsIRED by saturation mutagenesis of residues 32, 33, 34 and 37 (FIG. 4a). AL cells were transformed with the library and grown anaerobically in media supplemented with 2-methylpyrroline (also known as 2-methyl-1-pyrroline), which contains an imine group. Growth was observed after 55 hours.

[0301] Plasmid DNA was isolated from individual colonies and sequenced, revealing all of the selected variants had the same sequence (MsIRED-s=pLS133_1) (SEQ ID No: 35), containing the following residue substitutions with respect to the wild-type: N32E, R33V, T34R and K37R.

[0302] This is a different variant than any of the obtained in previous studies and cells transformed with MsIRED-s were able to grow anaerobically in media supplemented with 2-methylpyrroline more efficiently than when transformed with the best variant identified in any previous studies (FIG. 16b).

[0303] NMR was performed to confirm the presence in the fermentation broth of 2-methylpyrrolidine, the reduced product which contains a secondary amine (Table 5).

[0304] Activity assays with MsIRED-s revealed a NADH-dependent reductase activity towards 2-methylpyrroline, whereas no activity was detected with NADPH (FIG. 16c, Table 6).

[0305] Advantageously, the kinetic parameters of MsIRED-s were better than those of the best previously identified variant, and it displayed lower substrate inhibition (FIG. 16c, Table 6).

EXAMPLE 6

Selection of Nitroreductase Variants with Altered Substrate Specificity

[0306] nsfB nitroreductase from Enterobacter cloacae (EntNFSB) (SEQ ID No: 37) is able to catalyze the reduction of several compounds with nitro groups with NADH, including 4-nitrobenzoic acid (4-NBA). The inventors sought to obtain variants with altered substrate specificity, designed to act optimally on 2-nitrobenzoic acid (2-NBA) and 4-nitrobenzyl alcohol.

[0307] A crystal structure of EntNFSB bound to 4-NBA is available. Based on it, the inventors generated a library by saturating residues 40, 41, 68 and 124.

[0308] AL cells transformed with the library were cultured anaerobically in media supplemented with 2-NBA or 4-nitrobenzyl alcohol. Anaerobic growth was observed in both cases after 6 to 8 days.

[0309] Sequencing of plasmid DNA revealed that a single different variant had been selected for 2-NBA (EntNFSB-s1=LS169_1) (SEQ ID No: 38), and a different variant was identified for cells grown with 4-nitrobenzyl alcohol (EntNFSB-s3=LS169_3) (SEQ ID

[0310] No: 40).

[0311] NMR spectra revealed that 2-NBA or 4-nitrobenzyl alcohol had been consumed in the fermentation broth of cultures transformed with the selected variants (Table 5). In both cases, unidentified products were generated. In the case of cells grown in the presence of 2-nitrobenzoic acid, cultures acquired an intense yellow colour.

EXAMPLE 7

Selection of Entire Multi-Enzymatic Metabolic Pathways

[0312] To prove the suitability of the methods of the invention to select functional variants of more complex systems, the inventors generated a library of pathways for isopropanol production based on the combination of genes previously designed by Hanai et al (Escherichia coli acetyl-CoA acetyltransferase (atoB) and, acetoacetyl-CoA transferase (atoAD) , Clostridium acetobutylicum acetoacetate decarboxylase (adc) and CBADH) (FIG. 5a). The variants of the library differed in the promoter and RBS of each of the genes of the pathway, yielding a library size of over 6 million variants.

[0313] AL cells were transformed with the library and cultured anaerobically in plates of agar M9 with gluconate as the carbon source. After 36 hours, individual colonies were visible. 10 colonies were picked and inoculated in anaerobic liquid M9 with gluconate. After 8 days, growth was observed in 2 of the cultures. Plasmidic DNA was isolated of both cultures and sequenced, resulting in variants MP-S9 and MP-S10. They were found to have the same sequence. NMR spectra of the fermentation broth revealed isopropanol was being produced. Surprisingly, the inventors also found propionate was being produced, which is a metabolite not natively produced by E. coli as a fermentation product. Finally, isopropanol production under aerobic conditions was compared for randomly selected variants, variants selected in plates and variants selected in plates that grew in anaerobic liquid cultures. FIG. 5b summarizes the isopropanol production for 10 random variants and 8 selected variants, in addition to variants MP-S9 and MP-S10. Isopropanol production was significantly higher on average for the selected variants when compared to the random variants.

[0314] Additionally, all random and selected variants were sequenced, revealing the selective pressure had acted at two levels. There was a clear trend in selected variants, where a strong preference for a reduced number of combinations of RBS and promoters was observed. On the contrary, no clear trend was observed for random variants. This indicates that specific combinations leading to levels of expression for each enzyme that maximize the production of isopropanol had been selected.

[0315] All of the selected variants had a functional copy of all of the genes involved in the pathway. However, some of the random variants had one or more absent or inactive genes. This indicates the selection pressure eliminated defective variants without a completely functional pathway.

EXAMPLE 8

Selection with Mutant Geobacillus thermoglucosidasius

[0316] The inventors looked to demonstrate the portability of the selection methods and systems of the invention to other microorganisms. To this end, the inventors extended it to a thermophilic organism, as culturing it anaerobically at high temperatures would enable it to select thermostable variant polypeptides and enzymes. The inventors used a TMO236 strain, which contains two gene deletions: formate lyase (pfl) and lactate dehydrogenase (ldhA). The mutant cells are unable to grow anaerobically, whereas the wild-type cells can grow anaerobically (in both cases at 55° C., which is not a permissive temperature for E. coli (FIG. 18). Thus, proving that it is possible to obtain a mutant of this organism such that it becomes unable to grow under anaerobic fermentation conditions because of its inability to regenerate oxidized NAD+, which behaves in a similar way to the mutant strains of E.coli the inventors developed for selection.

[0317] Discussion

[0318] The inventors have developed a novel variant polypeptide or enzyme selection method based on a double mutant Escherichia coli strain unable to grow under oxygen-limited or substantially oxygen-free conditions, i.e. anaerobic fermentation conditions. Only upon transformation with an active NAD(H)-dependent oxidoreductase able to reduce a specific substrate present in the culture medium cells are able to regenerate oxidized NAD.sup.+, and can thus grow under such conditions.

[0319] The most immediate application for such a selection system is to use it to select specific variants of NAD(P)(H)-dependent oxidoreductases by transforming cells with a library of variants of the oxidoreductase, and culturing them under anaerobic fermentation conditions in the presence of the oxidized substrate of the enzyme. The inventors have demonstrated the huge potential of the system by using it to select a variant of CBADH which uses NAD(H) as the preferred cofactor instead of NADP(H), being, to their knowledge, the first enzyme with substantial NADH-dependent acetone reductase activity. Surprisingly, neither the predictions presented in previous studies where the structure of the native enzyme was solved, nor those provided by recently developed software aimed at predicting key residues for cofactor specificity in NAD(P)(H)-dependent oxidoreductases, were totally in accordance with the mutations found in the NAD(H)-dependent variant described herein. Even though there have been several attempts in the past to find sequence patterns that determine the cofactor specificity of NAD(P)(H)-dependent oxidoreductases, the findings described herein highlight the lack of general rules that can be widely applicable to invert cofactor preference.

[0320] A number of other properties can be selected for in the final variant with the selection system without much variation in the general set-up. For example, one possibility is the selection of variants with novel substrate specificity, which would require a change in the substrate supplemented to the culture medium. Such an approach could be employed, for example, to obtain enzymes with new regiospecificity or stereospecificity. These are of particular interest for the synthesis of compounds useful for their biological activity such as pharmaceuticals or agricultural chemicals, or precursors of these, where often only one specific isomer is useful for the next synthesis step, or only one specific isomer is active and all the other isomers are inactive, or can even cause undesired effects. Alternatively, a similar methodology could be used to obtain variants with enhanced activity or binding towards a substrate metabolized with low efficiency by the native enzyme.

[0321] Furthermore, the selection system is amenable to implementation in other organisms, provided that they are dependent upon, or can be modified to be dependent upon, fermentative pathways to grow under anaerobic conditions. This widens even further the enhanced properties that can be selected. For example, by using a thermophilic facultative anaerobe microorganism, such as Geobacillus thermoglucosidasius (which, similarly to Escherichia coli, also performs mixed-acid fermentation in anaerobic fermentation conditions), enzyme variants with increased thermal stability could be selected by culturing cells at higher temperatures. This approach could yield thermostable counterparts of enzymes of mesophilic organisms.

[0322] More sophisticated variations of the basic selection system can be used to enlarge further the application scope of the method by transforming LS1 strain cells with different combinations of a gene encoding an exogenous NAD(H)-dependent oxidoreductase and another genetically encoded function, typically a gene encoding another type of protein. For example, if a substrate which could be readily reduced by the oxidoreductase but was unable to permeate the cell membrane under normal conditions was supplied, a membrane transporter (comprising one or more proteins) could be coupled to the activity of the oxidoreductase. Only with a transporter able to introduce the substrate within the cell, NAD+ regeneration could be achieved, thus allowing the selection of transporters able to act on certain substrates. Alternatively, a two-enzyme system can be devised, where the medium would not be supplemented with the direct substrate of the NAD(H)-dependent oxidoreductase, but instead with a precursor needing a one-step transformation in order to become a substrate for the NAD(H)-dependent oxidoreductase. In such a system, cells would be transformed with the NAD(H)-dependent oxidoreductase and variants of the enzyme which could potentially catalyze the conversion of the precursor into the substrate. Furthermore, these additional genetically-encoding functions could potentially be combined.

[0323] While already applicable to a class of enzymes as wide as NAD(H)-dependent oxidoreductases, the flexibility and portability of the selection system based on metabolic complementation further increase its scope. Furthermore, with only slight modifications to the global scheme, it can be tweaked to select for enhancement in different properties of the gene of interest. The inventors expect it to become a valuable tool which will help identify enzymes with novel properties which can be used to develop new synthetic pathways or be integrated into already existing ones to optimize them. The inventors have applied it to a variety of oxidoreductases, including alcohol dehydrogenases, imine reductases and nitroreductases. Furthermore, the inventors have used it to select for different properties, including cofactor specificity/preference, improvement of kinetic parameters and substrate specificity/preference.

[0324] The inventors applied the selection method to select for a different type of biomolecules other than NAD(H)-dependent oxidoreductases, namely promoter and ribosome binding site (RBS) sequences. They have demonstrated that the selection method is suitable to select an optimal combination of several of these regulatory elements leading to maximized production of a given product thanks to the combined action of a set of several enzymes, including enzymes that are not NAD(H)-dependent oxidoreductases, and not even oxidoreductases.

REFERENCES

[0325] Baba, T., Ara, T., Hasegawa, M., Takai, Y., Okumura, Y., Baba, M., Datsenko, K. A., Tomita, M., Wanner, B. L. & Mori, H. (2006) Construction of Escherichia coli K-12 in-frame, single-gene knockout mutants: the Keio collection. Molecular Systems Biology. 2 2006.0008.

[0326] Berrios-Rivera, S. J., Bennett, G. N. & San, K. Y. (2002) The effect of increasing NADH availability on the redistribution of metabolic fluxes in Escherichia coli chemostat cultures. Metabolic Engineering. 4 (3), 230-237.

[0327] Cahn, J. K., Werlang, C. A., Baumschlager, A., Brinkmann-Chen, S., Mayo, S. L. & Arnold, F. H. (2017) A General Tool for Engineering the NAD/NADP Cofactor Preference of Oxidoreductases. ACS Synthetic Biology. 6 (2), 326-333.

[0328] Cherepanov, P. P. & Wackernagel, W. (1995) Gene disruption in Escherichia coli: TcR and KmR cassettes with the option of Flp-catalyzed excision of the antibiotic-resistance determinant. Gene. 158 (1), 9-14.

[0329] Geertz-Hansen, H. M., Blom, N., Feist, A. M., Brunak, S. & Petersen, T. N. (2014) Cofactory: sequence-based prediction of cofactor specificity of Rossmann folds. Proteins. 82 (9), 1819-1828.

[0330] Hamilton, C. M., Aldea, M., Washburn, B. K., Babitzke, P. & Kushner, S. R. (1989) New method for generating deletions and gene replacements in Escherichia coli. Journal of Bacteriology. 171 (9), 4617-4622.

[0331] Hanai, T., Atsumi, S. & Liao, J. C. (2007) Engineered synthetic pathway for isopropanol production in Escherichia coli. Applied and Environmental Microbiology. 73 (24), 7814-7818.

[0332] Hollrigl, V., Hollmann, F., Kleeb, A. C., Buehler, K. & Schmid, A. (2008) TADH, the thermostable alcohol dehydrogenase from Thermus sp. ATN1: a versatile new biocatalyst for organic synthesis. Applied Microbiology and Biotechnology. 81 (2), 263-273.

[0333] Korkhin, Y., Kalb(Gilboa), A. J., Peretz, M., Bogin, O., Burstein, Y. & Frolow, F. (1998) NADP-dependent bacterial alcohol dehydrogenases: crystal structure, cofactor-binding and cofactor specificity of the ADHs of Clostridium beijerinckii and Thermoanaerobacter brockii. Journal of Molecular Biology. 278 (5), 967-981.

[0334] Maddock, D. J., Patrick, W. M. & Gerth, M. L. (2015) Substitutions at the cofactor phosphate-binding site of a clostridial alcohol dehydrogenase lead to unexpected changes in substrate specificity. Protein Engineering, Design & Selection : PEDS. 28 (8), 251-258.

[0335] Reed, J. L., Vo, T. D., Schilling, C. H. & Palsson, B. O. (2003) An expanded genome-scale model of Escherichia coli K-12 (iJR904 GSM/GPR. Genome Biology. 4 (9), R54-2003-4-9-r54. Epub 2003 Aug. 28.