WO2009070822A2 - Recombinant pichia pastoris cell - Google Patents

Abstract

The present invention relates to a recombinant Pichia pastoris cell comprising at least one mutation in the dihydroxyacetone synthase gene 2 (DAS2) resulting in a reduced dihydroxyacetone synthase activity compared to a wild-type Pichia pastoris cell, wherein the DAS2 gene comprises a nucleotide sequence having at least 80% identity with SEQ ID No. 1.

Description

Recombinant Pichia pastoris cell

The present invention relates to a recombinant Pichia pastoris cell.

Biocatalysis emerged as a mature technology for industrial organic syntheses. Whole-cell catalysts are frequently employed where cofactor-recycling and catalyst stability are essential. While baker' s yeast was the most popular microorganism for such bioreductions in the past, more recently recombinant "designer bugs" are used for biotransformations. Some were redesigned to strengthen substrate to product conversion by elimination of competing side-reactions or product metabolisation or by overexpressing the redox catalyst, others by overexpressing additional enzymes for cofactor-regeneration or by any combination of such modifications. Both, resting cells and growing cells can be applied for whole-cell conversions.

While at the beginning the use of biocatalysts was restricted to pharmaceutical intermediates with relatively high commercial value, biocatalysis is now also a competitive technology in the production of commodity chemicals. Nevertheless, the main use of biocatalysts is, and will be, in the production of chiral intermediates for bioactive compounds, whose sale increase by 7-8% annually.

P. pastoris has been established as efficient protein production host demonstrated by a large number of proteins expressed in this yeast. Although less frequently used for wholecell biotransformations as e.g. E. coli or S. cerevisiae several examples already describe applications of P. pastoris in bioconversions. Up to now cofactor regeneration in P. pastoris was performed by co-expression of glucose dehydrogenase or by feeding formate as substrate for the endogenous formate dehydrogenase .

It is an object of the present invention to provide a Pichia pastoris cell which shows an increased NADH and/or NADPH regeneration capacity compared to a wild-type Pichia pastoris strain.

The present invention relates to a recombinant Pichia pastoris cell comprising at least one mutation in the dihydroxyacetone synthase gene 2 (DAS2) resulting in a reduced dihydroxyacetone synthase activity compared to a wild-type Pichia pastoris cell, wherein the DAS2 gene comprises a nucleotide sequence hav ing at least 80% identity with SEQ ID No. 1. The cell of the present invention is modified to strengthen the cell's own metabolism for redox cofactor regeneration instead of introducing additional metabolic burden. Cell growth during the production phase was reduced to a necessary minimum. Knocking out of dihydroxyacetone synthase gene 1 (DASl) of Pichia pastoris resulted in readily growing cells in glucose but reduced the metabolism to the essential task for whole-cell reductions, i.e. NADH regeneration and a remaining minimal growth for catalyst regeneration. Thus, a whole-cell system located between the resting cell and the growing cell taking the advantages of both systems was established.

In the present invention a new strategy for cofactor dependent whole-cell bioconversion by reducing the flux through the major metabolic pathways of Pichia pastoris thereby strengthening the already existing pathway for NADH regeneration is presented (see Fig. 1) . In this minimal metabolism designed for NADH dependent bioreductions 2 molecules of NADH are generated by the complete (and irreversible) oxidation of methanol to CO₂. Methanol consumption in P. pastoris is also characterised by a specific flux of ~18 mmol g^'1 h^'1 comparable to the specific glucose consumption rate of S. cerevisiae of ~18 mmol g^-1 h^-1. Therefore a high NADH regeneration rate with only CO₂ as byproduct can be obtained using this irreversible pathway. Methanol as cosubstrate for Pichia pastoris conversion adds some more advantages. The endogenous methanol utilisation pathway and therefore the NADH regeneration system is highly induced. Using strong promoters, such as the alcohol oxidase 1 promoter (P_AOX1) , for the expression of the biocatalyst one can achieve high expression as well. Therefore methanol acts as the inducer for gene expression of the bioreduction enzyme and the NADH regeneration system. In addition methanol is used as co-substrate. Due to the oxidation of methanol to CO₂ this reaction series is irreversible thereby driving the reaction towards the reduction. Since it is preferred to exploit a highly expressed endogenous system rather than introducing new enzymes for cofactor regeneration, the host system is highly adapted for handling the co-substrate. And finally, if needed, methanol can act as a solvent for the substrate .

When formaldehyde is fed into cellular metabolism by the di hydroxyacetone synthase (DHAS) , it is not available for oxidation to CO2 and therefore for NADH regeneration. To increase the carbon efficiency for the cofactor regeneration system, it is aimed at reducing this enzyme function by deleting the corresponding gene(s). Using this strategy a minimal cell metabolism can be applied, highly effective for NADH regeneration just by switching the carbon source to methanol which also induces the biocatalyst/cofactor regeneration system. This system can also be used to separate cell growth, enzyme production and bioconversion just by switching the major carbon source. Using this approach a system with reduced growth was generated therefore not redirecting too much carbon atoms (and cofactor) to cell growth but not generating resting cells with zero growth. Due to the remaining carbon flux into cellular metabolism the cell is still able to produce enzyme which is expected to increase the time where high conversion could be obtained with the whole-cell biocatalyst .

It was surprisingly found that Pichia pastoris cells comprise next to the DASl gene a further gene coding for a second dihydroxyacetone synthase, namely DAS2. Mutating or knocking out said gene lead to recombinant Pichia pastoris cell which exhibits a very low growth rate when these cells are incubated with methanol. As discussed above methanol is converted in Pichia pastoris by an alcohol oxidase to formaldehyde which normally is used by the cell for the carbon metabolism and biomass production and for the regeneration of NADH via formaldehyde dehydrogenase and formate dehydrogenase. A mutation in the DAS2 gene resulting in a reduced activity of its translated product shifts the metabolism of the cell from biomass production towards NADH regeneration. This is a very surprising and unexpected effect because a mutation within the well-known DASl gene results only in a slightly reduced growth rate in a medium comprising methanol as carbon source (instead of glucose) whereby a mutation within the DAS2 gene results in a Pichia pastoris cell having a much lower growth rate on methanol compared to a wild-type Pichia pastoris cell or a Pichia pastoris cell having a mutation within the DASl gene (see Table 1 of example 1) . If both DAS genes are mutated or even knocked out, the growth rate of Pichia cells on methanol can further be decreased. Therefore in a preferred embodiment of the present invention the cell further com prises a mutation within the dihydroxyacetone synthase gene 1 (DASl) .

The advantage of the modified Pichia pastoris cells of the present invention is, that these cells are able to produce a high amount of NADH and NADPH rather than biomass which can be used in other biochemical reactions. Thus, the cells of the present invention can be considered as NADH and NADPH factories. The produced NADH and NADPH can be used in other biochemical reactions for producing other metabolites. Therefore the cells of the present invention are perfectly suited to be used in biocatalysis .

The fact that Pichia pastoris is able to grow an methanol although DASl and/or DAS2 have been mutated (e.g. knocked out) is surprising. This is in contrast to the teachings of the prior art (Sakai Y et al., J. Bacteriol, 180 (1998) : 5885-5890), from which it is known that the deletion of the dasl gene in Candida boidinii leads to cells which can not be grown on methanol.

The nucleotide sequence of the DAS2 gene of a Pichia pastoris cell has at least 80%, preferably at least 90%, more preferably at least 95%, identity with SEQ ID No. 1 in the overlapping region of the gene with SEQ ID No. 1 when aligned:

5'gttgacaacactatctatgctattgttggtgatgcttgtttgcaagagggacctgctttggaat cgatttccttagctggtcacttggccttggacaaccttattgtgatctacgacaacaaccaggt ttgttgtgatggttccgtcgatgttaacaacaccgaagacatttctgctaagtttagagctcag aactggaatgtcattgaagtcgagaatggttctagagatgttgctacccttgtcaaggccatcg aatgggccaaggctgagaatgagagaccaactctgatcaacgttagaactgaaattggacagga ttctgctttcggtaaccaccacgctgctcacggttctgctcttggtgaggaaggtatccgggag ttgaaggccaagtacggtttcgatgtcgctagaaagttctggttcccacaggaggtctatgatt tctttgctgaaaaaccagccgagggtgatcaactagttgctaactggaagaaacttttggatga gtacgttaagaactatcctcaagaaggtgaggaattaaaggcccgtattagaggtgaacttcca aagaactggaagagtttcattccacaggacaaaccaaccgagccaactgctaccagaacctctg ctagagaaattgttagatctctgggacaaaaccttcctcaggttattgctggttctggtgactt gtccgtgtccattcttttgaactggggaggagttaagtacttcttcaaccctaagttacaaact ttctgtggattgggtggtgactactctggtagatatattgagtttggtatcagagaacactcta tgtgtgctattgccaatggtttggctgcatacaacaagggtactttcttgcctattacctcaac tttctacatgttctacctgtatgcagcacctgccttgcgtatggctgcacttcaagagttgaaa gcaattcacattgctacacacgactccatcggagctggtgaagatggtccaacgcaccagccta ttgctttgtcttcattattcagagctatgcccaacttctactacattagaccagccgatgctac cgaggttgcagctctgtttgaagtagctgttgagctcgagcactccaccttgttctctctgtcc agacacgaggttgagcaatacccaggtaagacttcggctgagggagccaaaagaggtggttacg tcgttgaagactgtgagggtaagccagacgtccaattaattggtgctggttccgaattggagtt tgccgtcaaaactgctcgtttgctaagacaacagaagggatggaaggtcagagttctgtcattc ccatgtcagagactgtttgaccaacaatccctggcatacagacgttctgtccttagaagaggag aggttccaactgtcgttgttgaggcctatgtcgcatacggatgggagagatacgccactgctgg ttacaccatgaacacctttggtaagtctcttcctgttgaggatgtctacaaatacttcggatac actcctgagaagattggtgagaaggttgctgcatacgtcaactctattaaggctagtcctcaaa tcctttacgaattcaccgatttgaagggaaaaccaaagcacgacaaactataa-3 ' (SEQ ID No. 1)

The Pichia pastoris gene coding for Dasl has the following nucleotide sequence:

5¹atggctagaattccaaaagcagtatcgacacaagatgacattcatgaattggtcatcaaaacct tccgttgttacgttctcgacttagtcga acagtatggtggtggtcaccctggttctgccatgggtatggtcgccattggtatcgctctgtgg aagtaccagatgaagtacgctccaaat gatccagactacttcaacagagatcgttttgtcttgtcaaacggtcacgtctgtctgttccaat acttgttccagcacttaactggtttgaagg agatgactgtcaagcaacttcaatcttaccactcttccgattatcactcattgactcctggaca ccctgaaattgagaaccctgctgttgag gttaccactggtcccctgggacaaggtatctctaacgctgtcggtatggccattggttcaaaga acctggccgctacttacaacagacct ggcttccctgtcgttgacaacactatctatgctattgttggtgatgcttgtttgcaagagggac ctgctttggaatcgatttccttagccggtc acttggccttggacaaccttattgtgatctacgacaacaaccaggtttgttgtgatggttccgt cgatgttaacaacaccgaagacatctcc gcaaagttcagagctcagaactggaatgttatcgacattgtagacggttctagagatgtcgcta ccattgtcaaggctatcgattgggcca aggctgagactgagagaccaactctgatcaacgttagaactgaaattggacaggattctgcttt cggtaaccaccacgctgctcacggtt ctgctctaggtgaggaaggtatccgggagttgaagactaagtacggttttaaccctgcccaaaa gttctggttccctaaagaagtatacg acttctttgctgagaaaccagctaaaggtgacgagttagtaaagaactggaaaaagttagttga tagctatgtcaaagagtaccctcgtg agggacaagagttcctttctcgtgttagaggtgagcttccaaagaactggagaacttacattcc tcaagacaagcctaccgaaccaacc gccaccagaacctctgctagagaaattgttagggcccttggaaagaaccttcctcaagttattg ccggttccggtgacttatctgtctcaat tcttttgaactgggacggagtgaagtacttcttcaaccctaagttacagactttctgtggatta ggtggtgactactctggtagatatattgag tttggtatcagagaacactctatgtgtgctattgccaacggtttggctgcatacaacaagggta ctttcttgcctattacctctaccttctacat gttctacctgtatgcagcacctgccttgcgtatggctgctcttcaagagttgaaagcgattcac attgctacacacgactctattggagctg gtgaagatggtccaacccaccagcctattgctttgtcttcattattcagagctatgcccaactt ctactacatgagaccagccgatgctacc gaagttgcagctctgtttgaagtggctgttgagcttgaacactccacattgctttctctgtcca gacacgaggttgaccaatacccaggtaa gacttctgcccaaggagccaaaagaggtggttacgttgttgaagactgcgaaggaaagccagat gtgcaactgatcggaactggttcc gagttggaattcgctattaagactgctcgtttgctaagacaacagaagggatggaaggtcagag ttctgtcattcccatgtcagagattgtt tgacgagcagtctattacttacagacgttccgtccttagaagaggagaagttccaactgtcgtt gttgaggcctatgtcgcatacggatgg -3'

A polynucleotide has a certain percent "sequence identity" to another polynucleotide, meaning that, when aligned, that percentage of bases or amino acids are the same, and in the same relative position, when comparing the two sequences. To determine sequence identity, sequences can be aligned using the methods and computer programs, including BLAST, available over the world wide web at ncbi.nlm.nih.gov/BLAST. See, e.g., Altschul et al. (1990), J. MoI. Biol. 215:403-10. Another alignment algorithm is FASTA, available in the Genetics Computing Group (GCG) package, from Madison, USA, a wholly owned subsidiary of Oxford Molecular Group, Inc. Other techniques for alignment are described in Methods in Enzymology, vol. 266: Computer Methods for Macromolecular Sequence Analysis (1996), ed. Doolittle, Academic Press, Inc., a division of Harcourt Brace & Co., San Diego, Calif., USA. Of particular interest are alignment programs that permit gaps in the sequence. The Smith-Waterman is one type of algorithm that permits gaps in sequence alignments. See Meth. MoI. Biol. 70: 173-187 (1997) . Also, the GAP program using the Needleman and Wunsch alignment method can be utilized to align sequences. See J. MoI. Biol. 48: 443-453 (1970) .

The mutation of the DAS gene(s) can be of any kind provided that the mutation leads to a reduced DAS activity within the Pichia pastoris cell. Hence, the mutation is preferably a deletion, an insertion or a substitution (see e.g. Lin Cereghino GP, Gene 263 (2001) : 159-169) . In a particular preferred embodiment of the present invention at least 20%, more preferably at least 50%, even more preferably at least 80%, most preferably 100%, of one or both DAS genes is removed from the genome of the cells resulting in knockout mutants.

The DAS activity within the cell can not only be reduced by mutating the DAS genes, but also by introducing mutation within the regulatory sequences of said genes. For instance, the promoter can be modified so that no polymerase can bind to the regulatory sequence in the upstream region of the DAS genes. Respective methods are well known to the person skilled in the art .

In order to enhance even more the regeneration of NADH and NADPH within the Pichia pastoris cell said cell comprises an increased formaldehyde dehydrogenase (FLD, e.g. Genbank accession No. AF066054) enzyme activity and/or formate dehydrogenase (FDH; see e.g. US 7,087,418 and Fig. 6) enzyme activity compared to a wild-type Pichia pastoris cell.

As discussed and shown in the experimental section herein an increased enzymatic activity of FLD and FDH results in an even more increased generation of NADH and NADPH. FLD catalyses the reaction of formaldehyde to formic acid thus leading to the formation of NADH. The formic acid generated in this first step is catalysed to CO₂ by FDH. Also in this latter step NADH is formed. Due to the reduced DAS activity within the cell formaldehyde is accumulated within the cell when fed with methanol. A high activity of FLD and FDH supports the conversion towards CO₂.

According to a preferred embodiment of the present invention the increased enzyme activity is that of an enzyme, endogenous to said cell, encoded by a nucleic acid coding sequence operably linked to at least one regulatory sequence not natively associated with said nucleic acid coding sequence, whose expression is increased as compared to the expression of the enzyme activity when said nucleic acid coding sequence is associated with its native regulatory sequence or of an enzyme, exogenous to said cell, encoded by a nucleic acid coding sequence, operably linked to at least one regulatory sequence.

The enzymatic activity of FLD and FDH can be increased in various ways. Since both enzymes are naturally present in Pichia pastoris cells, their respective regulatory sequence (including their promoter) on the genome may be modified in order to enhance its expression rate (see e.g. Nevoigt E. et al., Appl . Environ Microbiol, 72 (2006) : 5266-73, La C et al, Appl. Environ Microbiol, 73 (2007 ) : 6072-7) . For instance, the native regulatory sequences of FLD and/or FDH may be exchanged with an alternative promoter like the AOXl promoter or variants thereof (see e.g. WO 2006/089329) . The use of such a promoter has the advantage that this promoter is mainly regulated by methanol. Consequently the addition of methanol to a medium or as the sole carbon source leads next to the induction of the expression of alcohol oxidases also to an increase in the biosynthesis of FLD and/or FDH. The expression of FLD and/or FDH may also be increased by introducing into the genome of a Pichia pastoris cell further copies of the respective genes (the resulting cell comprises at least 2 of said genes; multi-copy strains) . Of course the additional genes may be part of the genome of the cells or present on plasmids.

The term "promoter" or "regulatory sequence" includes within its scope inducible, repressible and constitutive promoters. An inducible promoter is a promoter that is positively regulated; that is the promoter is activated in the presence of an inducer molecule or system, either directly or indirectly.

By "operably linked" a linkage of polynucleotide elements in a functional relationship is meant. Thus, a promoter is operably linked to a transcribable polynucleotide coding region if it affects the transcription of the polynucleotide, the polynucleotide being located so as to be under the regulatory control of the promoter, which then controls the transcription and optionally translation of that polynucleotide. In the construction of heterologous promoter/structural gene combinations, it is gener^¬ ally preferred to position a promoter or variant thereof at a distance from the transcription start site of the transcribable polynucleotide, which is approximately the same as the distance between that promoter and the gene it controls in its natural setting; i.e.: the gene from which the promoter is derived. As is known in the art, some variation in this distance can be accommodated without loss of function and in some instances variations in this distance can enhance expression. As mentioned above the cells according to the present invention can be used, e.g., in biocatalysis . Therefore said cell comprises at least one exogenous nucleic acid molecule encoding at least one NADH or NADPH dependent enzyme operably linked to at least one regulatory sequence.

In order to catalyse biochemical reactions which require NADH or NADPH respective nucleic acid molecules encoding enzymes have to be introduced into the cell. These nucleic acid molecules may be integrated into the genome of Pichia pastoris cells. Of course, it is also possible not to integrate these nucleic acid molecules and to use extrachromosal plasmids or other vetcors in the cell. Next to these exogenous enzymes the cell may also comprise other nucleic acid molecules encoding other exogenous enzymes which are also required for the chemical synthesis of biocatalytics .

The at least one regulatory sequence is preferably a promoter, more preferably an alcohol oxygenase promoter, even more preferably an alcohol oxygenase promoter from Pichia pastoris, most preferably AOXl or a variant thereof (AOXl variants are disclosed e.g. in WO 2006/089329) . The AOXl promoter or derivatives thereof are particularly advantageous because these promoters can be activated by adding methanol to a culture medium. Methanol serves not only as promoter activating agent or carbon source but also substrate for the NADH and/or NADPH production.

According to the present invention all kind of nucleotide sequences encoding NADH and NADPH dependent enzymes can be introduced into the Pichia pastoris cell described herein. NADH and NADPH dependent enzymes include oxidoreductases (e.g. dehydrogenases) etc.. Particularly preferred the at least one NADH or NADPH dependent enzyme is selected from the group consisting of butanediol dehydrogenase (EC 1.1.1.4), xylose reductase (EC 1.1.1.21), alcohol dehydrogenase (EC 1.1.1.1) enoate reductase (EC 1.3.1.31) and NADPH-kenoprotein reductase (EC 1.6.2.4) .

The NADH and NADPH dependent enzyme to be expressed in the Pichia pastoris cell of the present invention (this means a Pichia pastoris cell carries a nucleic acid molecule encoding such enzymes, preferably operably linked to a regulatory sequence) is preferably an oxidoreductase (EC 1.6; particularly EC 1.6.1, EC 1.6.2, EC 1.6.5 and EC 1.6.99) and selected from the group consisting of EC 1.1.1, EC 1.2.1, EC 1.3.1, EC 1.4.1, EC 1.5.1, EC 1.6.2, EC 1.7.1, EC 1.8.1, EC 1.14, EC 1.17.1 and EC 1.6. Cells capable to produce one or more of these enzymes can be used in biocatalysis in order to produce chemicals such as 2Butanol, 2-Penthanol, 6-Methylhept-5-en-2-ol, 2-Oktanol, 1,3Butandiol, 2, 5-Hexandiol, 1, 1, 1-Trifluoro-2-propanol, 1,4Dichloro-2-butanol, l-Chloro-2-propanol, 1, l-Dichloro-2propanol-Phenylethanol, 1- (3-chlorophenyl) ethan-1-ol, 2-Chloro1-phenylethan-1-ol, 2-Chloro-1- (3-chlorophenyl) ethan-1-ol, 1¬

[3, 5-Bis (trifluormethyl) phenyl] ethan-1-ol, 2-chloro-1- (3hydroxyphenyl) ethan-1-ol, 3-Chloro-1-phenylpropanol, 2-chloro-1¬

(4-fluorophenyl) ethan-1-ol, 2-chloro-1- (3-pyridyl) ethan-1olethyl-3hydroxybutanoate, methyl-3-hydroxybutanoate, ethyl-4chloro-3-hydroxybutanoate, ethyl-4-bromo-3-hydroxybutanoate, methyl-3-Hydroxypentanoate, ethyl-lactate, ethyl-2-hydroxy-4phenylbutanoate, ethyl-mandelate etc.. In the course of such a biocatalytic process educts required are added to a cell culture comprising the cells of the present invention. After a defined incubation time the products formed by such a process can be isolated by various methods (e.g. chromatography, extraction etc.) from the supernatant of the cell culture.

According to a preferred embodiment of the present invention the butanediol dehydrogenase is of Saccharomyces cerevisiae

(Ace. No. NP009341) , xylose reductase of Candida tenuis (Ace. No. AAC25601), alcohol dehydrogenase of Candida magnoliae.

The xylose reductase of Candida tenuis comprises preferably a K274R and/or N276D mutation (see Petschacher B et al., Microb. Cell Fact, 7 (2008) :9) .

The introduction of one or both of said mutations results in an enzyme having an increased activity with NADH, although the wild type enzyme has a higher activity with NADPH.

Glycerol is required for the biosynthesis of various substances which can be produced with a cell of the present invention. In order to reduce the glycerol assimilation of Pichia pastoris cells, which in said cells is regularly used for producing biomass, the glycerol kinase gene of said cell comprises at least one mutation resulting in a reduced glycerol kinase activity compared to a wild-type Pichia pastoris cell. The nucleotide sequence coding for said glycerol kinase is: 5 ' ¬

ATGGGAAAAGACTATACACCACTAGTTGCTACCATCGATATTGGTACTACCTCCACCAGAGCTA

TTCTTTTTGACTACCACGGTCAGGAAGTGGCCAAGCACCAGATCGAGTACTCTACCTCTGCTCA

GGATGATATCAAAAGAAAGCGTTCTCAGATCATCTCTTCCGAAGGTATTTCCCTGACAGTTTCT

GACGACTTGGAAGTTGAGTCCGTTGACAATAAGGCTGGTCCAACTTTGCAATTTCCTCAGCCAG

GCTGGGTTGAATGTCGTCCAAGTCACATCTTGGCCAACGCCGTTCAGTGTCTTGCTGCTTGTTT

AGTCACCATGGAGAACAAGAACTTGGACCGAGATGAAAAAAATAAATACAAGCTCATATCTATC

GGTGTTGCCAACATGAGAGAGACCACGGTTGTTTGGTCCAAGAAGACAGGAAAGCCTCTTTACA

ACGGTATTGTGTGGAACGATACCAGAAACAACGATATTGTTGACGAGTACACCGCCAAGTACTC

TGAGAAGGAGAGAGAGGAAATGCGAACCTTGTGTGGTTGTCCAATCTCCACCTACTTTTCTGCA

ACCAAGTTCAGATGGTTACTGAAGCATGTTCCTGAGGTCAAACAGGCCTATGACAATGCTGATG

GGGATCTAATGTTTGGTACTATTGACTCTTGGTTGATTTACCACTTGACTAACGAAAAATCCCA

CGTCACTGATGTTACCAATGCCTCCAGAACCAACTTCATGAACATTGAAACCAACAAATATGAC

GACAGACTTTTGAAATTCTGGGACGTCGATACTTCCAAAGTCATCCTTCCAGAAATCAGATCTT

CCGCAGAAGTCTACGGACACTTCAAGGTCCCACACTTGGAGTCTATTGGATATGTTGAGTCTTA

CCTTACTGATGACGCCCTTGCTCTTCTGGAAACTATTGAAGGTGCTCCTTTGGCTGGATGTCTG

GGTGACCAATCTGCCTCTTTGGTTGGACAGTTGGCTGTCAGAAAGGGTGATGCCAAATGTACAT

ATGGTACCGGTGCTTTCTTGCTGTACAACACTGGTGATCAGACTTTGATTTCTGAGCACGGTGC

TTTGACCACTGTGGGATATTGGTTCCCAGGTTTGGATGAGTCCGAAGATGGCAAACACTCTTCT

AAGCCACAGTATGCTTTGGAGGGATCGATTGCTGTCGCTGGATCTGTCGTGCAATGGTTAAGAG

ATAACCTTCGTTTGATTTCCAAGGCTCAGGACGTCGGACCATTGGCTTCTCAAGTTGACAACTC

TGGAGGTGTGGTATTTGTTCCAGCATTTTCAGGATTGTTTGCCCCTTACTGGGATTCCAACTCC

AGAGGAACCATTTTCGGTCTGACCCAATACACCTCAGCTTCTCATATTGCTAGAGCTGCTTTGG

AAGGTGTCTGTTTCCAAACTAGAGCCATTTTGAAGGCCATGATCAGCGATGCAGGAGCTTCTGC

TGACTTTTTGGAGGAATCATCCAAGGCCACTGGCCACAACCCTCTGTCAGTTCTTGCCGTGGAC

GGAGGTATGTCCAAATCAGACGAGATGATGCAGATCCAAGCTGATATTTTGGGTCCATGTGTCA

CTGTTAGACGTTCCATCAACCCTGAATGTACTGCACTGGGAGCTGCCATTGCTGCCGGTTTTGG

TGTCCCTAAGGAAGATAGAATTTGGGGTTCCTTGAAGGAATGTACCGAGGCCATTCTTGAGGGT

AACAAGATGTACTTGGCTGCAGGGAACACTTCTTTGGACTTCAAGGCCACATTGAGCGACGAGG

TCAGAAGAAAGGAATGGAGATTGTGGGAAAATGCCATTGCAAAGGCAAAGGGCTGGCTTAAGGA

CACTGCTTAA -3'

Another aspect of the present invention relates to the use of a cell according to the present invention in whole cell biocatalysis .

Biocatalytic processes for producing enantiomerically pure pharmaceutical intermediates or active ingredients are of growing importance. Recent advances in molecular biological methods such as recombinant enzyme expression, high-throughput DNA sequencing, and enzyme-evolution technologies, make biocatalysis a viable option for producing single enantiomers. Even two- and three-step biotransformations can be accomplished by combining and adapting enzymes from different sources. Recombinant microbial whole-cell biocatalysts, or so-called "designer cells" which provide each enzyme at the optimum amount, is a particularly efficient approach. Under this approach, all of the required enzymes can be produced by one fermentation. Cell disruption, clarification, and concentration of the enzyme solution are dispensable. The separation of the biocatalyst after biotransformation can easily be done by flocculation and filtration of the biomass. Substantial substrate and product-transfer limitations through the cell membranes have not been observed using frozen or dried biocatalysts. Many biochemical reactions require NADH and/or NADPH as co-substrates. Therefore the optimal cell to be used in biocatalysis should be able to produce NADH and/or NADPH in a much higher amount than wild type Pichia pastoris cells. This is achieved by reducing the activity of the enzyme encoded by the DAS2 gene.

Yet another aspect of the present invention relates to a method for catalysing at least one educt to at least one product comprising a) cultivating a cell according to the present invention comprising at least one enzymatic activity, b) incubating an educt with the culture of a) , and c) isolating the synthesized product.

The cells of the present invention can also be immobilized on a solid carrier or it can be entrapped within particles or between membranes. The immobilisation of cells has the advantage that the medium comprising the educts and products can be removed easily from the culture (see e.g. Carvajal-Vallejos P. et al., 64(529) (2007) : 384-389; Kawakami et al., Appl . Biochem. Biotech 67(1-2) (1997) : 23-31; Duff J. B. et al., Biotechnol . Bioeng., 31(8) (1988) : 790-5; Narasu, M. et al., BioTechnology (Rajkot, India), 1(1) (2007) : 5-8; Norouzian et al., Enzyme Microbial Techn., 33(2-3) (2003) : 150-153 ; Tripathi C. K. M. et al., Indian J. Exp. Biol., 35(8) (1997) : 886-889) .

The present invention is further illustrated by the following figures and examples, however, without being restricted thereto .

Fig. 1 shows a general reaction scheme of the minimal cell regeneration concept of the present invention using the Pichia pastoris methanol metabolism. 1: alcohol oxidase, 2: formaldehyde dehydrogenase, 3: formate dehydrogenase, 4: catalase, 5: dihydroxyacetone synthase.

Fig. 2 shows a multiple sequence alignment of P. pastoris Daslp and Das2p with the H. polymorpha DAS protein. Amino acid residues formatted in bold and underlined indicate residues belonging to transketolase signature 1 and 2, respectively. Hp_X02424: H. polymorpha Das protein sequence (Genbank ID X02424), PpDASl: P. pastoris Daslp, PpDAS2: P. pastoris Das2p. Daslp showed 68% identity (81% similarity) and Das2p showed 61% identity (76% similarity) to the amino acid sequence of the H. polymorpha DHAS .

Fig. 3 shows the vector map of the E. coli / P. pastoris shuttle vector Kan3,2-8.

Fig. 4 shows the DNA-Sequence of the E. coli / P. pastoris shuttle vector Kan3,2-8.

Fig. 5 shows the increased FLD-activity in Pichia pastoris strains with additional FLD-expression cassettes (transformants

- Series 1; control strains used for transformation - CBS7435, Δdasldas2 WT and BDH WT (= CBS7435 + BDH), respectively; sterile control - Blank) .

Fig. 6 shows the DNA-sequence of the FDH gene from Pichia pastoris CBS7435. Compared to the published Pichia pastoris FDH gene, two nucleotide exchanges were found: the cytidine at position 443 is exchanged with an adenine leading to the amino acid exchange P148H and the thymidine at position 486 is replaced by a cytidine leading to a silent mutation at position 1161. The mutated nucleotides are underlined and highlighted in bold.

Fig. 7' shows the increased FDH-activity in Pichia pastoris strains with additional FDH-expression cassettes (transformants

- Series 1; control strains used for transformation - CBS7435, Δdasldas2 WT and BDH WT (= CBS7435 + BDH) , respectively; sterile control - Blank) .

Fig. 8 shows FDH+FLD-cassette representing the oePCR product which was used for the ligation into the E. coli / P. pastoris shuttle vector 'Kan3,2-8' to result in 'Kan3, 2-8-FDH+FLD' .

Fig. 9 shows the increased FDH- and FLD-activity in Pichia pastoris strains with additional FDH+FLD-expression cassettes (transformants - Series 1; control strain used for transforma- tion - CBS7435; sterile control - Blank) .

Fig. 10 shows the increased FDH- and FLD-activity in Pichia pastoris strains with additional FDH+FLD-expression cassettes (transformants - Series 1; control strain used for transformation - BDH WT (= CBS7435 + BDH), Δdasldas2 WT + BDH; sterile control - Blank) .

Fig. 11 shows the effect of acetoin on the catalytic activity of P. pastoris -S.c.-BDH.

Fig. 12 shows the effect of methanol on the catalytic activity of P. pastoris -S.c.-BDH.

Fig. 13 shows increasing product concentrations during repetitive biotransformation batch processes with P. pastorisS. c. -BDH.

Fig. 14a shows batch biotransformation processes applying P. pastoris BDH 742 D4 and P. pastoris BDH 742 D4 with overexpression of FLD.

Figure 14b shows the whole cell reduction of acetoin applying different engineered P. pastoris strains.

Fig 14c shows a comparison of biocatalytic reduction of diacetyl in batch and fed-batch mode (see example 5) .

Fig. 15 shows vector map of the E. coli / P. pastoris shuttle vector T2.

Fig. 16 shows the DNA sequence of the E. coli I P. pastoris shuttle vector T2.

Fig. 17 shows the xylose reductase activity comparing the initial strains P. pastoris CBS7435 and P. pastoris D12-B1H (zldasldas2) used for the transformation with resulting -CtXRwt and -CtXRmut transformants (Series 1) . The sterile control is depicted as Blank.

Fig. 18 shows the xylose reductase activity comparing the initial strain P. pastoris CBS7435-FDH+FLD used for the transformation with resulting -CtXRwt transformants (Series 1) . In addition, FLD and FDH-activity was detected and most of the transformants showed equal or slightly increased activities compared to the initial strain P. pastoris CBS7435-FDH+FLD (transformants - Series 1; P. pastoris CBS7435-FDH+FLD - FLD+FDH WT; sterile control - Blank) .

Fig. 19 shows the xylose reductase activity comparing the initial strain P. pastoris CBS7435-FDH+FLD used for the transformation with resulting -CtXRmut transformants (Series 1) . In addition, FLD and FDH-activity was detected and most of the transforraants showed equal or slightly increased activities compared to the initial strain P. pastoris CBS7435-FDH+FLD (transformants - Series 1; P. pastoris CBS7435-FDH+FLD - FLD+FDH WT; sterile control - Blank) .

Fig. 20 shows the xylose reductase activity comparing the initial strain P. pastoris CBS7435-FDH used for the transformation with resulting -CtXRwt transformants (Series 1) . In addition, remaining FDH-activity was detected and most of the transformants showed slightly decreased activities compared to the initial strain P. pastoris CBS7435-FDH (transformants - Series 1; P. pastoris CBS7435-FDH - WT FDH; sterile control - Blank) .

Fig. 21 shows the xylose reductase activity comparing the initial strain P. pastoris CBS7435-FDH used for the transformation with resulting -CtXRmut transformants (Series 1) . In addition, remaining FDH-activity was detected and most of the transformants showed equal or slightly decreased FDH-activities compared to the initial strain P. pastoris CBS7435-FDH (transformants - Series 1; P. pastoris CBS7435-FDH - WT FDH; sterile control - Blank) .

Fig. 22 shows the xylose reductase activity comparing the initial strain P. pastoris CBS7435-FLD used for the transformation with resulting -CtXRwt transformants (Series 1) . In addition, remaining FLD-activity was detected and most of the transformants showed equal FLD-activities compared to the initial strain P. pastoris CBS7435-FLD (transformants - Series 1; P. pastoris CBS7435-FLD - WT FLD; sterile control - Blank) .

Fig. 23 shows the xylose reductase activity comparing the initial strain P. pastoris CBS7435-FLD used for the transformation with resulting -CtXRmut transformants (Series 1). In addition, remaining FLD-activity was detected and most of the transformants showed a similar FLD-activity compared to the initial strain_P-._ pastoris CBS-7-435--FLD- (-transformants - Series-l; 'P. pastoris CBS7435-FLD - WT FLD; sterile control - Blank) .

Fig. 24 shows the DNA-Sequence of the designed, synthetic gene CtXRsyn_wt including the restriction recognition sites for EcoRI and Notl (in italic) . The start and stop codon is highlighted in bold.

Fig. 25 shows the xylose reductase activity comparing the initial strains P. pastoris CBS7435 and P. pastoris D12-B1H (Δdas1das2 ) used for the transformation with resulting - CtXRsyn_wt and -CtXRsyn_mut transformants . The sterile control is depicted as Blank. Most of the transformants show significantly elevated xylose reductase activity compared to the initial wild type strains .

Fig. 26 shows the xylose reductase activity of the initial strain P. pastoris CBS7435, P. pastoris CBS7435+XRsyn_mut which was used for the transformation and the resulting -FLD cotransformants . In addition, the introduced, additional FLDactivity was detected and most of the transformants showed significantly increased activities compared to the initial strain P. pastoris CBS7435+XRsyn_mut . The sterile control is depicted as Blank.

Fig. 27 shows the xylose reductase activity of the strain P. pastoris CBS7435+XRsyn_mut which was used for the transformation and the resulting -FDH co-transformants . In addition, the introduced, additional FDH-activity was detected and most of the transformants showed significantly increased FDH-activities compared to the initial strain P. pastoris CBS7435+XRsyn_mut . The sterile control is depicted as Blank.

Fig. 28 shows the xylose reductase activity of the strains P. pastoris CBS7435, P. pastoris CBS7435+XRsyn_mut which was used for the transformation and the resulting -FLD+FDH cotransformants . In addition, the introduced, additional FLD- and FDH-activity was detected and most of the transformants showed significantly increased FLD-activity and equal or slightly elevated FDH-activity compared to the initial strain P. pastoris CBS7435+XRsyn_mut . The sterile control is depicted as Blank.

Fig. 29 shows the xylose reductase activity of the strains P. pastoris D12-B1H (Δdas1das2 ) , P. pastoris D12-B1H (Δdas1das2 ) +XRsyn_mut which was used for the transformation and the resulting -FLD co-transformants . In addition, the introduced, additional FLD-activity was detected and most of the" transformants showed significantly increased FLD-activity compared to the initial strain P. pastoris D12-B1H ( Δdas1das2) +XRsyn_mut . The sterile control is depicted as Blank.

Fig. 30 shows the xylose reductase activity of the strains P. pastoris D12-B1H (Δdas1das2 ) , P. pastoris D12-B1H (Δdas1das2 ) +XRsyn_mut which was used for the transformation and the resulting -FDH co-transformants . In addition, the intro duced, additional FDH-activity was detected and most of the transformants showed significantly increased FDH-activity compared to the initial strain P. pastoris D12-B1H ( Δdas1das2) +XRsyn_mut . The sterile control is depicted as Blank.

Fig. 31 shows the xylose reductase activity of the strains P. pastoris D12-B1H (Δdas1das2 ) , P. pastoris . D12-B1H (Δdas1das2 ) +XRsyn_mut which was used for the transformation and the resulting -FLD+FDH co-transformants . In addition, the introduced, additional FLD- and FDH-activity was detected and most of the transformants showed significantly increased FDH- and FDHactivity compared to the initial strain P. pastoris D12-B1H (Δdas1das2 ) +XRsyn_mut .

Fig. 32 shows KU70-knockout cassette with a KU70-homologous 3'- and 5' -end, two FRT-sequences and in-between these FRTsequences, an FLP recombinase and a Zeocin resistance cassette are located.

Fig. 33 shows glycerol kinase knock out cassette with a 3'- and 5' -end homologous to the glycerol kinase gene and in-between a Zeocin resistance cassette.

EXAMPLES :

Primers and Media used in the examples :

Example 1 :

The sequence of the Hansenula polymorphs {Pichia angusta) dihydroxyacetone synthase (P06834) was used to identify the DAS gene(s) in the P. pastoris genome sequence. 3 sequences were identified with significant homology to the H. polymorpha DHAS sequence. The best hit, named DASIp, showed 68% identity, and the second best hit, named DAS2p, showed 61% identity to the amino acid sequence of the H. polymorpha DHAS. The third hit, with an identity of only 39% was found to be highly similar to the transketolase genes of Pichia stipitis (69% identity), S. cerevisiae (67%) and Kluyveromyces lactis (67%) . Therefore, the latter gene was named TKLl. The occurrence of two DAS genes is special among the methylotrophic yeasts, since only one gene was found in most strains studied so far. Alignments of the P. pastoris DASl, DAS2 and the H. polymorpha DAS protein sequences are shown in Fig. 2. Beside the general transketolase signatures and thiamine pyrophosphate binding domain and pyrimidine binding domain both genes also possess a C-terminal peroxisomal targeting sequence 1 (PTSl, -DKL) similar to the signal sequences of other peroxisomal proteins involved in methanol utilisation in methylotrophic yeasts.

To decrease the fraction of methanol shuttled to the assimilatory pathway both DHAS-encoding genes were deleted employing ADEl and URA3 auxotrophic markers, respectively. While knocking out the DASl gene did not result in a significantly reduced growth rate on methanol as compared to the wild type strain, a strong decrease in the specific growth rate on methanol after knocking out the DAS2 gene was observed. A similar phenotype was observed when both genes were knocked out. Interestingly, both strains (ΔDAS2 and ΔDAS1ΔDAS2) still grew on methanol as sole carbon and energy source (Table 1) .

Table 1. Specific growth rates of P. pastoris reference and DAS knockout strains. Values represents mean vaules ± standard deviations maximum specific growth rate during the exponential rowth hase of 3 inde endent shake flask cultivations.

These results show a clearly reduced growth rate on methanol by knocking out one or two of the DAS genes in P. pastoris .

To evaluate the P. pastoris methanol oxidation system for cofactor regeneration the S. cerevisiae 2, 3-butanediol dehydrogenase gene BDHl ( YALO6OW) , amplified from S. cerevisiae strain BY4741 was introduced. Expression of this gene was controlled by the Δl AOXl promoter variant with increased activity as compared to the wild type AOXl promoter (Hartner FS et al., Nucleic Acid Res., 36 (12) (2008) :e76. The S. cerevisiae 2,3 butanediol dehydrogenase gene was chosen as a model because of its very high turnover number among oxidoreductases of -98,000 min^-1 and sufficiently low K_M values for NADH and the substrate (3.R/3S) -acetoin. While no significant activity on the substrate could be obtained with wild type P. pastoris strains lacking the recombinant BDHl gene, significant reduction of (3JR/3S) -acetoin could be found in recombinant strains expressing the BDHl gene (Table 2) .

Table 2. Time course of conversion of acetoin to 2, 3-butanediol in shake flasks at a cell density of 60 g/L, 25 g/L substrate and 60 g/L methanol. Strain Conversion [%]

After cell growth on minimal glucose medium to the stationary phase and a methanol induction phase for approx. 12 h acetoin was added to a final concentration of 40 g/L (0.45 M) and 6% (v/v) methanol (-1.5 M) to start the bioreduction. The cultures were further incubated for 24 h at 28 °C with vigorous shaking at 320 rpm in deepwell plates. The conversion rate of the best strains of the wild type and the daslΔ das2Δ was 15 and 20%, respectively. Although a high concentration of the substrate and methanol was added to the cultures, all of them were still growing under these conditions but at a lowered rate. Due to the reduced growth rate of the daslΔ das2Δ knockout strains on methanol the start OD and the specific growth rate during bioconversion of these strains was lower than for the wild type strains (Table 2) . An approx. 2-times faster conversion of 3Racetoin to the (R, R) -2, 3-butanediol than of 3S-acetoin to meso2, 3-butanediol could also be observed. (S, S) -2, 3 butandiol was not found. Therefore the whole-cell biocatalyst is assumed to have a high enantioselectivity as found for the isolated enzyme. Time course analysis of the microscale bioconversion showed that a high conversion within the first -4 h of the process could be achieved and no further reaction occurred after this initial conversion. Within the first 4 h a space-time yield of ~11 mmol L^-1 h^'1 and a specific conversion rate of -1.4 mmol (g CDW)^-1 h^-"1 was obtained. Assuming the methanol consumption rate during bioconversion correlates with specific growth rate the methanol consumption rate can be calculated to be -6 mmol g^'1 h^-1 at a specific growth rate of 0.045 h^-1 (based on 17.7 mmol L^-1 h^-1 at μ_maχ = 0.15 h^-1) .

To perform the reaction with higher load of biocatalyst, cells were cultivated in a bioreactor in a low-salt medium with 40 g/L glucose according to Hellwig et al. (Biotechnol Appl Biochem 1999, 30 (Pt3) , 267) followed by a glucose-fed-batch to further increase the biomass to approx. 60 g cell dry weight (CDW) _. L^-1. After biomass growth, cells were induced with methanol for 12 h. The cultures were aliquoted to 3 times 25 mL in 250 mL baffled shake flasks and the bioconversion was started by. adding 2.5 mL of a 30.5% (w/w) acetoin solution in methanol. After 2.5 h of bioconversion a yield of 84.8 ± 0.7 % was obtained with the wild type background strain 742D4 while' the daslΔ das2Δ background strain 764D10 had a yield of 88 ± 3 %. This conversion rate results in space-time yields of 91 and 95 mmol L^-1 h^-1 and a specific conversion rate of 1.30 and 1.42 mmol (g CDW)^-1 h^-1, respectively (Table 3) .

Table 3. Specific conversion rate and space-time yield (STY) of acetoin to 2, 3-butanediol conversion in shake flasks at a cell densit of 60 /L 25 /L substrate and 60 /L methanol.

After 4 hours a conversion of 91 ± 1 % was found for both strains which tend to decrease after this point. Nevertheless, high conversion rate of the substrate acetoin with methanol as cosubstrate and no addition of NADH could be observed.

In summary, it could be shown that the methanol dissimilation pathway of Pichia pastoris could be efficiently used for NADH regeneration during whole-cell bioreductions . Considering the results described above together with the P. pastoris expression system characteristics, methanol could fulfil several main roles: 1) as inducer of the recombinant dehydrogenase and the endogenous NADH regeneration system consisting of the methanol dissimilation enzymes formaldehyde dehydrogenase and formate dehydrogenase, 2) as co-substrate for NADH regeneration, 3) as solvent for the substrate and 4) as substrate for minimal cell growth. This redesigned catalyst platform can be used for wholecell applications where NADH needs to be recycled and due to its simplicity it could further boost the use of recombinant wholecell biocatalysts in chemical and pharmaceutical industry.

Material and Methods: Chemicals and media

Unless otherwise stated, all chemicals were purchased from Carl Roth GmbH (Germany) and Becton, Dickinson and Company (USA) , respectively. Sterile water was purchased from Fresenius Kabi Austria (Austria) . Media were prepared essentially as described earlier (Weis R et al . FEMS Yeast Res 2004, 5, 179). Primers were purchased from either MWG Biotech AG (Germany) or Invitrogen Corp. (USA) .

BMD1% (200 mM potassium phosphate buffer, pH 6, 1.34% yeast nitrogen base, 10g/L D-glucose, 4-10^~5% D-biotin) , BMM (200 mM potassium phosphate buffer, pH 6, 1.34% yeast nitrogen base, 0.5% (v/v) methanol, 4-10^-5% D-biotin), BMM2 (200 mM potassium phosphate buffer, pH 6, 1.34% yeast nitrogen base, 1% (v/v) methanol, 4'10^~5% D-biotin), MD agar (20g/L D-glucose, 1.34% yeast nitrogen base, 15 g/L agar, 4'10^-5% D-biotin) .

Strains

E. coli XL-I blue (Stratagene, USA) was used for all E. coli cloning experiments.

P. pastoris strains CBS7435, JC254 and JC301 were used as hosts for yeast experiments. P. pastoris strain X-33 was used as template for PCR amplification of genomic DNA fragments and strain KM71H was used as reference strain for growth rate determination .

a Centraalbureau voor Schimmelcultures

Molecular biology

500 bp DASl and DAS2 integration sites were amplified using primer pairs Dasldelfw/ DaslAde5'rv, DaslAde3' fw/Dasldelrv, Das2delfw2/Das2Ura5'rv and Das2Ura3 ' fw/Das2delrv employing Phusion™ high fidelity DNA-polymerase (Finnzymes Oy, Finland) according to the manufacturers protocol. ADEl and URA3 marker cassettes were amplified from plasmids pBLADE-IX and pBLURA-IX using primer pairs DaslAde5 ' fw/DaslAde3' rv and

Das2Ura5' fw/Das2Ura3' rv employing Phusion™ high fidelity DNApolymerase according to the manufacturers protocol.

DASl and DAS2 knockout cassettes were synthesized by overlap-extension PCR (oePCR) reactions by mixing 5 ng of the DAS fragments and 10 ng of the ADEl or URA3 cassettes and the outer primer pair (Dasldelfw/ DASldelrv or Das2delfw/ DAS2delrv) employing Phusion™ high fidelity DNA-polymerase according to the manufacurers protocol. The final PCR products were TOPO^® cloned into vector pCR^®-Blunt II-TOPO^® (Invitrogen Corp., San Diego, CA, USA) according to the manufacturers protocol. The resulting vectors pCR2-dDASl-ADE and pCR2-dDAS2-URA were used to amplify the knockout cassettes as templates instead of the single fragments as already described for the oePCR. 5 μg of the PCR products were purified using the Wizard^® SV Gel and PCR Clean-Up System (Promega Corp., USA) . The PCR products were eluted with 20 μl of H₂O to concentrate the DNA and transformed into appropriate P. pastoris strains according to the condensed protocol (LinCereghino J. et al. Biotechniques 2005, 38, 44) . In order to complement auxotrophies when needed the URA3 or HIS4 marker cassettes were amplified using primer pairs Ura_fw/ Ura_rv and His_fw/ His_rv employing Phusion™ high fidelity DNA-polymerase according to the manufacturers protocol. The resulting PCR products were purified using the Wizard^® SV Gel and PCR Clean-Up System and 1-2 μg of DNA were transformed into appropriate P. pastoris strains.

P. pastoris JC301 was transformed with dDASl-ADEl knockout cassette and the HIS4 complementation cassette. After transformation strains were plated on MD agar plates complemented with 50 mg/L uracil. One positive transformant with correct integra tion (D1-D8H) was taken and complemented with the URA3 cassette to generate the auxotrophic DASl knockout strain D1-D8HU. To generate the DAS2 knockout strain, P. pastoris JC254 was transformed with the dDAS2-URA3 knockout cassette. One positive transformant with correct integration (D1-D8H) was selected and named D2-24. To generate the Δdasl Δdas2 knockout strain, D1-D8H was taken to knockout of the DAS2 gene by transformation of the dDAS2-URA3 cassette. Again, one positive transformant with correct integration was selected and named D12-B1H.

Colony PCR of transformants

Correct integration of knockout cassettes on both sides was confirmed by colony PCR. A single colony from a fresh YPD agar plate was resuspended in 100 μl H₂O, heated to 95°C for 5 min and centrifuged for 2 min in a tabletop centrifuge. 3 μL of supernatant was used as template for the PCR. All PCR reactions were performed in 25 μL total volume using 1 U GoTaq^® DNA polymerase (Promega Corp., USA) with 200 μM of each dNTP, 5 pmol of each primer in appropriate buffer conditions in a thermocycler for 40 cycles. Primer pairs AdeSeq3' fw/DAS1__col_fw and AdeSeq5' rev/DASl_col_rv were used to control correct integration on the 5' and the 3' site of DASl-ADEl knockout cassette, respectively. Primer pairs DAS2_col_fw/UraSeq5' rv and

UraSeq3' fw/DAS2_col_rv were used to control correct integration on the 5' and the 3' site of DAS2-URA3 knockout cassette, respectively.

Shake flask cultures and growth rate determination

P. pastoris strains were cultivated in 50 mL of BMD1% or BMM in 250 mL baffled shake flasks at 28°C, 60% humidity and 130 rpm on a rotary shaker. Growth rate was determined during exponential growth phase by determination of optical density at 595 ^' nm

(OD₅₉₅) .

Screening of BDHl transformants

Cultivation of P. pastoris strains was performed in Deepwell-plates. After 60 h growth phase on BMD1% medium strains were induced with BMM2 for approx. 12 h prior to the start of the bioreduction reaction. Bioreduction was started by addition of 50 μl 400 g/L (3.R/3S) -acetoin in methanol and the reaction was performed for 24 h. 300 μl of supernatant were extracted with 400 μl ethyl acetate and substrate conversion was determined by GLC-analysis . GLC analysis was performed essentially as described earlier (Gonzalez E. et al. J Biol Chem 2000, 275, 35876) using a Hewlett Packard 6890 instrument equipped with a FID (275 °C) and a Chirasil-DEX CB column (25 m x 0.32 mm, 0.25 μm film) and H₂ as the carrier gas (2.4 ml/min) . The following temperature program was used: isotherm at 75 °C for 8 min, 2 °C/min ramp to 85 °C, and isotherm at 85 °C.

Example 2: Gene synthesis of optimized ADH3 and ADH4 genes Two new ADH genes were identified in Candida magnoliae. The genome region containing these two genes was isolated and cloned into a pUC19 vector (pUC19-CmADHs) . The wild-type genes were amplified from the pUC19-CmADHs vector using the following primers :

CmADH3:

P05-402 ADH3WTfw AAGAATTCAAA-ATGACGACTACTTCAAATGCGCTCG P05-403 ADH3WTrv TTTGCGGCCGC-CTAAGCAATCAAGCCATTGTCGAC

CmADH4 :

P05-404 ADH4WTfw AAGAATTCAAA-ATGACATCTACACCTAATGCCCTTG P05-405 ADH4WTrv TTTGCGGCCGC-CTACGTCGAGAAACCATTGTCC

The codon usage and the GC content were adjusted, stable secondary structures were avoided, and undesirable restriction cleavage sites were removed to optimally adjust the genes to expression in Pichia pastoris.

Two different codon usages were used: the first is derived from the Kazusa homepage (http: //www. kazusa . or . jp/codon/) and corresponds to the average codon usage of all known Pichia pastoris genes; the second was calculated cumulatively from genes highly expressed in P. pastoris (AOXl, DASl₁ FLDl and HbHNL) .

Subsequently, the ADH3 and ADH4 genes and the corresponding strains, which were generated using the codon usage of the Kazusa homepage, were shortened using ADH3K and ADH4K. The genes having the codon usage of the genes highly expressed in Pichia pastoris were labelled ADH3hM and ADH4hM.

The wild-type genes amplified from the pUC19-CmADHs vector were labelled ADH3WT and ADH4WT.

12 oligonucleotides were generated for each gene from the optimized sequences of the Candida magnoliae ADH3 and ADH4 genes. In addition, an EcoRI cleavage site, followed by a Kozak sequence, was inserted at the 5' terminal of every gene, and a WotI cleavage site was inserted at the 3' terminal of every gene .

The genes were generated in two steps: In the first step, two fragments were amplified using PCR. 30 pmol of external primers and 1.5 pmol of internal oligonucleotides were used for each amplification.

Composition of the PCR mixture of the two fragments:

5 μL of oligonucleotide mixture, 5 μL of dNTP's (2 mM each), 10 μL of Phusion™ buffer (5x) , 0.5 μL of Phusion™ polymerase (2 U/μL) , and 29.5 μL of double-distilled water (ddH₂O) to a total volume of 50 μL.

The following temperature program was used for PCR: 98 °C for 30 seconds, 98°C for 10 seconds, 58°C for 20 seconds, 72°C for 20 seconds, 72 °C for 5 minutes, then store at 4 °C. The PCR was conducted for 30 cycles.

The PCR fragments were purified using Wizard^® SV Gel and PCR Clean-Up System (Promega) . 30 μL of ddH₂O were used for elution. The concentration of the two fragments was 10 ng/μL.

In the second step, an overlap extension PCR (oePCR) was carried out, wherein the two purified fragments were combined to form one gene .

OePCR composition:

3 μL each of initial and terminal primer, 1 μL each of fragment 1 and fragment 2, 5 μL of dNTP's (2 mM each), 10 μL of Phusion™ buffer (5x), 0.5 μL of Phusion™ polymerase (2 U/μL) and 26.5 μL of ddH₂O to a total volume of 50 μL .

The following temperature program was used for oePCR:

98°C for 30 seconds, 98°C for 10 seconds, 58°C for 20 seconds, 72°C for 40 seconds, 72°C for 5 minutes, then store at 4 °C. The PCR was conducted for 30 cycles.

The PCR products of the wild-type ADH genes and the Kazusa ADH genes were purified using Wizard^® SV Gel and PCR Clean-Up System according to the instructions and then cloned into the pCR^®-Blunt II-TOPO^® vector.

The unpurified PCR products of the optimized ADH hM genes were cloned into the pJETl / blunt cloning vector using a GeneJET™ PCR Cloning Kit (but using only half the preparation volume) .

Subsequently, both the TOPO^® and the GeneJET™ cloning products were transformed in chemically competent E.coli DH5α-T1^R cells . 23 transformants each were checked for an insert of correct size using colony PCR.

The colony PCR was carried out using GoTaq^® DNA polymerase. For the PCR, a single colony was suspended in the following PCR mixture after being secured to a fresh agar plate: 5 pmol of each external primer, 2.5 μL of dNTP's (2 mM each), 5 μL of Green GoTaq^® reaktion buffer (5x) , 0.3 μL of GoTaq^® DNA polymerase (5 U/μL) , and 16.2 μL of ddH₂θ to a total volume of 25 μL.

The following temperature program was used for colony PCR:

95°C for 3 minutes, 95°C for 30 seconds, 55°C for 30 seconds, 72°C for 1 minute, 72°C for 10 minutes, then store at 4°C. The PCR was conducted for 35 cycles.

Eight clones each having an insert were streaked on LB agar plates (+ 100 μg/mL of ampicillin or 25 μg/mL of zeocin) for plasmid isolation.

Plasmid isolation was effected using a Wizard® Plus SV Minipreps DNA Purification System according to the instructions of the manufacturer. Eight plasmids per construct were sequenced.

Sequencing brought the following results:

ADH3K, ADH4K and ADH4hM, ADH3hM: no errors

The genes found had the following sequences:

>ADH3WT gaattcaaaatgacgactacttcaaatgcgctcgtcactggaggcagccgcggcattggcgctg cttccgccattaagctggctcaggagggctacagtgttacgctggcctctcgcagtgttgataa actgaatgaagtaaaggcgaaactcccaattgtacaggacgggcagaagcactacatttgggaa ctcgatctggctgatgtggaagctgcttcgtcgttcaagggtgctcctttgcctgctagcagct acgacgtcttcgtttcgaacgcgggcgtcgctgcgttctcgcccacagccgaccacgatgataa ggagtggcagaacttgctcgccgtgaacttgtcgtcgcccattgccctcacgaaggccctcttg aaggacgtctccgaaaggcctgcggacaatccgttgcagattatctacatttcgtcggtggccg gcttgcatggcgccgcgcaggtcgccgtgtacagtgcatctaaggccggtcttgatggttttat gcgctccgtcgcccgtgaggtgggcccgaagggcatccatgtgaactccatcaaccccggatac accaagactgaaatgaccgcgggcattgaagccctgcctgatttgcctatcaaggggtggatcg agcccgaggcaattgctgacgcggttctgtttctggcaaagtccaagaatatcaccggcacaaa cattgtggtcgacaatggcttgattgcttaggcggccgc

>ADH4WT gaattcaaaatgacatctacacctaatgcccttgtcacgggaggcagccgcggcattggcgctt ccgccgccatcaagctggctcaagaagggtacagcgtcacgctggcgtcccgcgaccttgagaa acttaacgaggtcaaggacaagctgccaatcgtgaggggtggacagaaacactacgtttggcaa ctcgatcttgccgatgtattggctgcatcgtctttcaaggcggctcctctgccggccagcagct acgatttgtttgtttcgaacgccggaattgcccagttctcgcccacggcagagtatactaatag tgagtggctgaacattatgaccattaacttagtgtccccgattgccctgacgaaggctcttttg caggccgtttctgggaggtcgagcgagaacccgtttcagatcgtattcatctcgtcggttgcag cactacgtggcgttgcacaaacggccgtctacagtgcgtcgaaggctggtactgatggattcgc acgctcacttgctcgcgaactaggtcctcaaggcgtccatgtgaacgtggtgaaccctggctgg actaagacagacatgacggaaggagtcgaaaccccaaaggacatgcccattaagggctggatcc agcctgaggcaattgctgatgctgtagtattccttgcgaggtcgaaaaacattaccggcgcgaa tattgtagtggacaatggtttctcgacgtaggcggccgc

>ADH3K gaattcaaaatgacgacaacttcaaacgctttggttactggtggatctagaggtattggtgcag cttctgcgattaagttggctcaagaaggttactcggtaactttggcttctagatccgttgacaa gttgaacgaggtaaaggctaagttacccatagtacaagatgggcaaaagcactacatctgggaa ttggatttggctgatgttgaagctgctagttcgtttaagggagcaccattgccagcttcttcat acgatgttttcgtgtctaatgctggtgttgcagctttttctcctactgcagatcatgacgataa ggaatggcaaaacctactcgcagtaaacctctcatctccaattgcccttacaaaggcactactg aaggatgtctctgaaagacctgctgataaccctctccaaattatttacatctcctctgttgcag gacttcatggtgctgctcaagttgctgtatactcagctagtaaagcaggattagacgggtttat gcgtagtgtagcaagagaagttggaccaaagggcattcatgtcaacagtattaacccaggttac accaagacagaaatgacagcaggaattgaagctttgccagatttgcctatcaagggttggatag aacctgaagcaattgcagatgcagtcttgttcctagctaagtccaagaacataaccggaactaa catcgttgtcgataacggtcttatcgcgtaagcggccgc

>ADH4K ^• gaattcaaaatgactagtacacctaacgctctagtaacaggaggtagtagaggaattggtgcatctgetgcaataaagttggcacaagaaggatactctgtcacacttgcatctagagacctagagaagttgaacgaggttaaggacaagttgccaatagttagaggtggtcaaaagcattacgtatggcaactggatttggctgatgtattggcagcttcctcttttaaggcagctcctttacctgcttcatcctacgacttgtttgtctcaaatgctggtattgcccaattctcacctacagctgaatacaccaacagtgaatggctcaacatcatgaccatcaaccttgtgagtccaattgcgttgactaaggcactacttcaagcagtttcaggtagatcgagtgaaaacccctttcaaatcgtgtteatcagctcagtagctgetttgagaggtgttgcacaaactgetgtctactctgcttctaaagctggaactgatggtttcgcaagatccttggctagagaattgggtccacaaggagttcacgttaacgtcgttaacccaggatggactaagaccgatatgacagaaggtgttgaaactccgaaagacatgcccataaagggatggatacaaccagaagctattgcagatg- cagtggttttcttagcacgatcgaagaacattactggagcaaacattgttgtcgataacggattttcgacgtaagcggccgc

>ADH3hM gaattcgaaacgatgactaccacttctaacgcattggttactggtggatctcgtggtattggtgcagcttctgccattaagttggctcaagaaggttactctgtcactttggcttccagatccgtggacaagcttaacgaggttaaggctaagcttcctatcgttcaagacggacaaaagcactacatctgggaattggatcttgctgatgttgaagcagcttcttccttcaaaggtgcacctttaccagcttcctcttacgacgttttcgtttccaatgctggtgttgctgcattctctccaactgcagaccacgatgataaggagtggcagaacttgettgetgttaacctttcttcccctattgcattgaccaaggcattgctgaaggacgtctctgaaagaccagctgacaatcctttgcagattatctacatctcctctgttgcaggtcttcacggtgcagctcaagttgcagtttactctgcatccaaagctggattggacggtttcatgagatccgttgcaagagaggttggtccaaagggtattcacgtcaactccattaacccaggttacaccaagactgagatgactgetggaattgaggcattaccagacttgccaatcaagggttggatcgaacctgaggctattgctgatgctgttctgttcttggctaagtccaagaacatcaccggaactaacatcgtggttgacaacggtttgattgcctaagcggccgc

>ADH4hM gaattcgaaacgatgacttctactccaaacgcattggttactggtggatctagaggtattggagcatctgctgcaatcaagttggctcaagaaggatactctgttacccttgcttctagagaccttgagaagttgaacgaggttaaggacaagttgccaattgtcagaggtggacaaaagcactacgtttggcagttggatttggctgacgttttggctgcatcttccttcaaagctgcacctttgccagcatcttcctacgacttgttcgtctctaacgctggtattgcacagttctctccaactgctgagtacaccaactctgagtggttgaacatcatgaccatcaacttggtctctcctattgccttgactaaggctctgttgcaagctgtttctggaagatcctctgaaaaccctttccaaattgtcttcatctcctctgttgctgcattgagaggtgttgcacaaactgctgtgtactctgcttccaaagctggtactgacggttttgccagatccttggctagagaacttggaccacaaggtgttcacgtcaacgtggttaacccaggttggaccaagactgacatgactgagggtgttgaaactccaaaggacatgcctattaagggttggattcaacctgaggctattgcagatgctgttgtcttcttggctagatccaagaacatcactggtgctaacatcgtcgttgataacggattctctacctaagcggccgc

The protein sequences of ADH3 and ADH4 are: ADH3

MTTTSNALVTGGSRGIGAASAIKLAQEGYSVTLASRSVDKLNEVKAKLPIVQDGQKHYIWELDLADVEAASSFKGAPLPASSYDVFVSNAGVAAFSPTADHDDKEWQNLLAVNLSSPIALTKALLKDVSERPADNPLQIIYISSVAGLHGAAQVAVYSASKAGLDGFMRSVAREVGPKG1HVNSINPGYTKTEMTAGIEALPDLPIKGWIEPEAIADAVLFLAKSKNITGTNIVVDNGLIA ADH4 :

MTSTPNALVTGGSRGIGASAAIKLAQEGYSVTLASRDLEKLNEVKDKLPIVRGGQKHYVWQLD¬

LADVLAASSFKAAPLPASSYDLFVSNAGIAQFSPTAEYTNSEWLNIMTINLVSPIALTKALL¬

QAVSGRSSENPFQIVFISSVAALRGVAQTAVYSASKAGTDGFARSLA¬

RELGPQGVHVNVVNPGWTKTDMTEGVETPKDMPIKGWIQPEAIADAVVFLARSKNITGA¬

NIVVDNGFST

Example 3: Cloning the genes into the pPICZ BΔl vector

All three ADH3 and all three ADH4 genes were cloned into the pPICZ BΔl vector using EcoRI and NotI cleavage sites. Ligation was effected using about 50 ng of cut vector, about 20 ng of cut insert (about 1.5 times molar excess of insert) and 2 units of T4 DNA ligase in T4 ligase buffer at 16°C over night.

All ligation reactions were transformed in chemically competent E.coli DH5α-Tl^R cells and on LB-zeocin (25 μg/mL) agar plates .

Seven transformants per gene were checked for the presence of the gene using colony PCR. Four transformants each were subjected to plasmid isolation using the Wizard® Plus SV Minipreps DNA Purification System, and the isolated plasmid DNA was control-cut using EcoRI and Notl. Two positive plasmids each were sequenced for a final check of correct cloning. None of the coding sequences had an error. One strain each was deposited and used for all further preparations.

Example 4: Pichia pastoris transformation and microscale cultivation

For Pichia pastoris transformation, the plasmids were linearized over night using Bg1II. Pichia pastoris transformation was carried out according to common methods. 2.5 μg of DNA were used for transformation.

After selection on YPD-Zeo (100 μg/mL) plates at 3O°C, 44 colonies each were cultivated in 96 deep well plates in 300 μL of 1% glucose minimal medium (BMD 1%) for screening using whole cell conversion known in the art.

Example 5: Whole cell conversions of alcohol dehydrogenases from Candida magnoliae

Strains expressing ADH3 and ADH4 in the wild type and in a daslΔ das2Δ double knock-out background were used for whole cell conversions with acetophenone, with methanol as a cosubstrate. The whole cell conversions were carried out in 96 deep well plates and in shake flasks.

After a growth phase in BMD1% for 60 h, followed by methanol induction for about 12 h, whole cell conversion was started by adding 50 μL of acetophenone solution (80 mM dissolved in MeOH) . The cells were harvested after 8 hours (5 minutes at 4000 rpm) . 100 μL of the clear supernatant were incubated with 100 μL of methanol for 10 minutes at 60°C to precipitate proteins. This was followed by 10 minutes of centrifugation at 4000 rpm at 4 °C, and 100 μL of the supernatant, again clear, were used for HPLC measurement .

The rates of conversion of acetophenone to 1-phenyl ethanol was determined using reversed phase HPLC.

Column: Purosphere RP-18 250 mm x 4.6 mm, 5 μm

Mobile phase: acetonitril/water 80/20

Detector: UV/VIS detector, 210 nm wave length

Flow rate: 1 mL/min

For conversion in 2 L baffled shake flasks, a ADH3hM and a ADH4hM clone, and the wild-type CBS 7435 were cultivated in 150 mL BMD1% for 60 hours at 110 rpm and 28 °C. After 60 hours, the shaked cultures were induced using 15 mL of BMMlO, and using 750 μL of methanol 8 hours later. After a 24 hour induction phase, acetophenone solution (1.7 M) was added to an final concentration of 1 g/L in the course of the following induction.

Samples were taken at different times to get a clearer impression of the exact time of conversion. The samples were measures using reversed phase HPLC.

The 1-phenyl ethanol peak appeared after a retention time of 3.2 minutes, and the acetophenone peak appeared at 3.7 minutes. Conversion was calculated from the integrals. of the peak areas after calibration using standard solutions of the substrate and the product. It turned out that 1 g/L of substrate was already converted completely after two hours. For comparison, the conversion of the CBS 7435 strain showed after 24 hours, a conversion rate which was not more than 10%.

Example 6: Determination of lysate activity For the photometric assay, the strains expressing ADH3hM and ADH4hM in the wild-type background, and the wild type (CBS 7435) itself were cultivated in a 50 mL shaking culture.

After 60 hours, the shaking cultures were induced using 5 mL of BMMlO, and using 250 μL of methanol 8 hours later. After a 24 hour induction phase, the cells were harvested by centrifugation (10 minutes, 4000 rpm) . The pellet was frozen and subsequentlydigested after resuspension in 20 mL of buffer A (10% glycerol, 1 mM DTT, 0.1 mM EDTA, 50 mM potassium phosphate buffer pH=7,5) using a homogenizer.

The lysate was used for the photometric assay:

20 μL 10 mM cosubstrate (NADH or NADPH)

10 μL cell lysate

960 μL 50 mM potassium phosphate buffer + 1 mM MgCl₂, pH = 6.5

10 μL IM substrate

The reaction was monitored using a photometer (DU 800 spectrophotometer/ Beckman Coulter Inc, Fullerton, CA, USA) at 340 nm for 5 minutes. The reaction was started by adding 10 μL of IM substrate solution.

Absorption was found to decrease at 340 nm due to the conversion of NAD(P)H to NAD(P)⁺.

The protein concentration was determined using the BCA Protein Assay (Pierce Biotechnology Inc.) according to the manufacturer's protocol with bovine serum albumin (BSA) as standard. Resulting protein concentrations were used for the calculation of specific activities according to the formula 2.

The specific activities were calculated using the following formulae :

ΔE * 1000 activity = d * ε

Formula 1: Calculation of activity, ΔE: Difference of absorption per minute, ε: constant of proportionality ε at 340 nm is 6620 mL mmol^'1 cm^-1, layer thickness d = 1 cm. Formula 2: Calculation of specific activity

Conversion was only measurable when NADPH was added. When NADH was used as a cofactor, no activity was detectable. This result clearly shows that both ADH enzymes prefer NADPH as a cofactor.

Example 7: Whole cell conversions in daslΔ das2Δ double knock-out background

Although it is an established fact that both ADH genes are NADPH dependent, they were transformed in the daslΔ das2Δ double knock-out background. Again, acetophenone with methanol as cosubstrate was used for the whole cell conversions. The whole cell conversions were carried out in 96 deep well plates as in the previous examples.

Screening was carried out as described in example 4.

The screeing results were surprising. Although both ADH genes are NADPH dependent, there were clones that converted more than 75% of acetophenone within 8 hours. This shows that NADPH may also be regenerated in the daslΔ das2Δ double knock-out background.

Example 8: Construction of an E. coli/P. pastoris shuttle vector for the over-expression of P. pastoris FLD in P. pastoris

The NADH-regenerating enzyme Formaldehyde Dehydrogenase (FLD) which is an enzyme of the P. pastoris methanol dissimilation pathway was overexpressed in P. pastoris in order to increase the cofactor regeneration capability.

First, the complete P. pastoris FLD gene was amplified from genomic DNA of Pichia pastoris CBS7435 using the primer pair FLD-BgIF/ FLDTTXhoR and employing the Phusion™ High-Fidelity DNA-polymerase (Finnzymes Oy, Finland) according to the manufacturer's protocol. The resulting PCR-fragment was ligated into the pJETl/blunt vector provided within the GeneJET™ PCR Cloning Kit from Fermentas and used for E. coli transformation. All steps were performed according to the manufacturer's protocol. The sequence of the PCR product and the resulting pJETl-FLD vector were controlled by sequencing.

This vector served as template for amplifying the FLD gene again with Phusion™ High-Fidelity DNA-polymerase using the primer pair PpFLD_spe_fw/PpFLD_not_rv. The PCR-fragment was purified according to the QIAquick PCR Purification Kit (Qiagen) , digested with Spel and NotI (from Fermentas) , again purified using the QIAqick Gel Extraction Kit (Qiagen) and used for the ligation into the Spel/NotI cut E. coli/P. pastoris shuttle vector 'Kan3,2-8' (see Fig. 3) . The nucleotide sequence of shuttle vector 'Kan3,2-8' is shown in .Fig. 4.

The resulting Kan3,2-8-FLD vector was controlled by sequencing, linearised by Sad digestion and used for P. pastoris transformation. First, the strains P. pastoris CBS7435, P. pastoris 742D4 [BDH1) and P. pastoris 764D10 {das1 das2 BDHl) were used. P. pastoris transformations were performed as known in the art. Transformants were selected on YPD/Geneticin-agar plates (300 μg/mL Geneticin - final concentration) . Then, transformants were cultivated in 96-well deep well plates and assayed for increased FLD activity. In addition, BDH-activity was investigated and finally, only those strains were chosen for further investigation which showed BDH-activity and increased FLD-activity .

Cultivation in 96-well deep well plates and cell disruption:

After an initial incubation phase on BMD1% medium at 320 rpm, 28°C and 80% air humidity for -60 h, the cultures were induced with BMM2 and with BMMlO after ~70h, -84 h and -108 h. After -132 h the cultures were harvested. Therefore, 100 μL of each culture were transferred to V-bottom microtiter plates from Greiner Bio-One (#651101) and the cells were harvested by centrifugation (Eppendorf Centrifuge 5810R: 4000 rpm, 4°C, 10 min) . The supernatant was discarded, 50 μL Y-Per^® Plus (Pierce) were added and the mixture was incubated for 30 min at RT and 1400 rpm using a TITRAMAX 1000 shaker from Heidolph. Then, cell debris were removed by centrifugation (Eppendorf Centrifuge 5810R: 4000 rpm, 4 °C, 10 min) and 10 μL of the supernatant were employed for the photometric FLD-activity assay.

FLD-activity assay:

The FLD-activity was determined photometrically tracking the increase in absorption of NADH (at 340 nm) for 2-5 min. The reaction mixture (200 μL) consisted of 150 mM potassium phosphate buffer (pH 7.5), 9 mM glutathione (reduced), 3 mM formaldehyde, 200 μM NAD⁺ and 10 μL crude cell lysate. The reaction was started by the addition of the NAD⁺ stock solution and the absorption of NADH was tracked at 340 nm.

BDH-activity assay:

After a 60 h incubation phase on BMD1% medium, the cultures were induced with BMM2 and again incubated at 320 rpm, 28 °C and 80% air humidity for approx. 12 h. Then, the bioreduction reaction was started by the addition of 50 μL of an 8 mM rac-acetoin solution in BMMlO. The reaction was performed for 24 h. Then, 300 μL of the supernatant were extracted with 800 μL of ethyl acetate and the substrate conversion was determined by GCanalysis. A Hewlett Packard 6890 instrument was equipped with an FID detector (275°C) and a Chirasil-DEX CB column (25 m x 0.32 mm, 0.25 μm film), and H₂ was used as carrier gas (2.4 mL/min) . The following temperature program was used: 65°C - 6.5 min; 50°C/min to 80°C; 80°C - 0.7 min; 2°C/min to 85°C; 85°C - 3 min. Retention times: rac-acetoin: 3.88 min and 4.49 min, Dbutanediol: 11.59 min, L-butanediol : 11.13 min, meso-butanediol : 12.56 min.

In Fig. 5, landscapes of resulting transformants are depicted. Some transformants show significantly increased FLDactivity compared to the initial strains used for the transformation. The potential of thus engineered P. pastoris strains coexpressing an interesting oxidoreductase such as the S. cerevisiae Butanediol Dehydrogenase (S. c. BDH) and additional FLD genes was tested in biotransformations (Example 12) .

Example 9: Construction of an E. coli/P. pastoris shuttle vector for the over-expression of P. pastoris FDH in P. pastoris

The P. pastoris FDH gene was amplified from genomic DNA of Pichia pastoris CBS7435 using the primer pair PpFDH_fw / PpFDH__rv and employing the Phusion™ High-Fidelity DNA-polymerase (Finnzymes) according to the manufacturer's protocol. The PCRfragment was purified according to the QIAquick PCR Purification Kit (Qiagen) , digested with EcoRI and Notl (from Fermentas) , again purified using the QIAqick Gel Extraction Kit (Qiagen) and used for the ligation into the EcoRI/Notl cut E. coli/P. pastoris shuttle vector 'Kan3,2-8' (see Figs. 1 and 2) . The resulting vector was named Kan3, 2-8-FDH. Its sequence and also the sequence of the FDH-PCR product were controlled by sequencing. Compared to the published P. pastoris FDH gene (US 7,087,418) , two nucleotide exchanges were found (see Fig. 6) .

The vector Kan3, 2-8-FDH was linearised by BgIII digestion. First, the strains P. pastoris CBS7435, P. pastoris 742D4 (BDHl) and P. pastoris 764D10 (das1 das2 BDHl) were used. Transformants were selected on YPD/Geneticin (300 μg/mL)-agar plates. Then, transformants were cultivated in 96-well deep well plates and assayed for increased FDH activity. In addition, BDH-activity was investigated and finally, only those strains were chosen for further investigation which showed BDH-activity and increased FDH-activity.

Cultivation in 96-well deep well plates and cell disruption:

The cultivation of P. pastoris transformants in 96-well deep well plates and the subsequent cell lysis with Y-Per^® were performed as described in Example 8.

FDH-activity assay:

The FDH-activity was determined photometrically, tracking the increase in absorption of NADH for 2-5 min. The reaction mixture (200 μL) consisted of a 50 mM potassium phosphate buffer (pH 7.5), 300 mM sodium formate, 200 μM NAD⁺ and 10 μL of crude cell lysate. The reaction was started by the addition of the NAD⁺ stock solution and the absorption of NADH was tracked at 340 nm.

BDH-activity assay:

The BDH-activity was assayed according to Example 8.

In Fig. 7, landscapes of resulting transformants are depicted. Some transformants show significantly increased FDHactivity compared to the initial strains used for the transformation. The potential of thus engineered P. pastoris strains coexpressing an interesting oxidoreductase such as the S. cerevisiae Butanediol Dehydrogenase (S. c. BDH) and additional FDH genes will be evaluated in medium scale-biotransformations .

Example 10: Construction of E. coli/P. pastoris shuttle vector for the over-expression of P. pastoris FLD and P. pastoris FDH in P. pastoris The E. coli/P. pastoris shuttle vectors 'Kan3,2-8', 'Kan3,28-FLD' and 'Kan3, 2-8-FDH' served as template for the following PCR reactions: The primer pair FDH_spe_fw/OeFDHtt_rv was used to amplify the FDH gene from plasmid Kan3, 2-8-FDH. The primer pairs OeFDHtt_fw/ 0ettA0X_rv and OettAOX_fw/OeAOXFLD_rv were used to amplify the AOXTT-sequence and the P(AOXl_syn) promoter sequence from plasmid Kan3,2-8, respectively. The primer pairs OeAOXFLD_fw/ MutFLD_rv and MutFLD_fw/PpFLD_not_rv were used to amplify the FLD gene from plasmid Kan3,2-8-FLD and removing the internal BgIII restriction recognition site. All PCRs were performed employing the Phusion™ High-Fidelity DNA-polymerase (Finnzymes Oy, Finland) according to the manufacturer's protocol .

Step-by-step, the fragments were ligated by overlap extension PCR (oePCR) . First the AOXTT-sequence was fused with the P(AOXl_syn) promoter sequence. Then the two FLD-fragments were linked and finally, the FDH gene was connected to the AOXTTP(AOXl_syn) fragment and the FLD gene. All oe-PCR reactions were performed in two steps. First, 10 ng of the largest DNA-fragment and equimolar amounts of all additional fragments were mixed with 10 μL of the 5x Phusion HF Buffer, 1 μL of a 10 mM dNTP-mix and 0.3 μL of the Phusion High-Fidelity DNA Polymerase (2 U/μL) . ddH₂θ was added to achieve a final volume of 50 μL. The following temperature programme was used: 30 sec 98°C, 20 cycles (10 sec 98°C, 20 sec 60°C, 105 sec 72°C) and a final extension at 72°C for 7 min. Second, 4 μL of each outer primer stock solution (10 pmol/μL) , 10 μL of the 5x Phusion HF Buffer, 1 μL of a 10 mM dNTP-mix and 0.3 μL of the Phusion High-Fidelity DNA Polymerase (2 U/μL) were added. Again ddH₂O was used to achieve a final volume of 100 μL. For the second oePCR-step, the same temperature programme was used.

The final oePCR-product (see Fig. 8) was ligated into the pJET1.2/blunt vector provided within the CloneJET™ PCR Cloning Kit from Fermentas and used for E. coli transformation. All steps were performed according to the manufacturer's protocol. The sequence of the resulting pJETl .2-FDH+FLD vector was controlled by sequencing.

The FDH-FLD-double cassette was amplified by PCR using the pJETl .2-FDH+FLD vector as template, the primer pair FDH_spe_fw/PpFLD_not_rv and the Phusion™ High-Fidelity DNA polymerase (Finnzymes Oy, Finland) according to the manufacturer's protocol. After purification with the QIAquick PCR Purification Kit (Qiagen) , the fragment was digested with Spel and Notl (from Fermentas) , again purified with the QIAqick Gel Extraction Kit (Qiagen) and used for the ligation into the Spel/NotI cut E. coli/P. pastoris shuttle vector 'Kan3,2-8' (see Figs. 3 and 4 for plasmid map and sequence) . The resulting E. coli/P. pastoris shuttle vector was named 'Kan3, 2-8-FDH+FLD' and controlled by sequencing.

Then, the vector was linearised with BgIII (from Fermentas) and used for P. pastoris transformations. Again, the strains P. pastoris CBS7435, P. pastoris 742D4 (BDHl) and P. pastoris 764D10 {dasl das2 BDHl) were used. Transformants were selected on YPD/Geneticin (300 μg/mL) -agar plates. Then, transformants were cultivated in 9β-well deep well plates and assayed for increased FDH, FLD and remaining BDH activity.

Cultivation in 96-well deep well plates and cell disruption:

The cultivation of P. pastoris transformants in 96-well deep well plates and the subsequent cell lysis with Y-Per^® were performed as described in Example 8.

FLD-activity assay:

The FLD-activity assay was performed according to Example 8.

FDH-activity assay:

The FDH-activity assay was performed according to Example 9.

BDH-activity assay:

The BDH-activity was assayed according to Example 8.

In Figs. 9 and 10, landscapes of resulting transformants are depicted. Some transformants show significantly increased FDHand FLD-activity compared to the initial strains used for the transformation. The thus engineered P. pastoris strains provide a more efficient cofactor regeneration system and consequently the basis for improved whole cell biotransformations which require the regeneration of NAD (P) H-cofactors .

Example 11: Biotransformation processes with P. pastorisS. c.BDH-strains

Optimisation of substrate and cosubstrate concentrations

The effects of different substrate concentrations on the catalytic activities of P. Pastoris -S. c. BDH-strains were investigated in deep well plates. P. pastoris cells were first cultivated in 25 mL YPD medium in 250 mL baffled shake flasks at 28°C and 120 rpm (Certomat BS-I, B. Braun Biotech, Germany) . After a cultivation period of 60 hours, 25 mL BM2M containing methanol were added to induce protein expression. After 12 hours of induction at 28 °C and 120 rpm, 450 μL of the cell suspension were transferred to each well of a deep well plate. 50 μL substrate solution containing acetoin and methanol in 200 mmol/L KPi buffer, pH 6.0, were added.

For investigation of the substrate acetoine the methanol concentration was 1% and acetoine was varied in the range of 5 to 100 mmol/L. Biotransformation processes were carried out at 28 °C, 320 rpm and 80% humidity (HT Infors Multitron 2 shaker) . After 2 and 5 hours 400 μL biotransformation solution were mixed with 100 μL 50 mmol/L n-butanol solution, which solved as internal standard for gas chromatography. The resulting solution was extracted with 500 μL ethyl acetate. The organic phase was analyzed via gas chromatography on a Varian CP7503 gas chromatograph with a Chirasil DEX-CD column (25 m x 0,32 mm i. d.) with a flame ionization detector and hydrogen as carrier gas. Compounds were detected using the following temperature program: 6.5 min - 65°C, 50°C/min to 80°C, 0.7 min - 80°C, 2°C/min to 85°C, 3.0 min - 85°C. Typical retention times were: n-butanol : 3.3 min, (S)-acetoin: 3.8 min, (R) -acetoin: 4.4 min, (2S,3S)butanediol : 11.1 min, (2.R, 3R) -butanediol : 11.6 min, mesobutanediol: 12.5 min.

No substrate inhibition by acetoin was detected in the investigated concentration range (Fig. 11) .

For investigation of the cosubstrate methanol the acetoin concentration was 100 mmol/L and methanol was varied in the range of 1 to 10% (v/v) . Biotransformation processes were performed in the same manner as described before.

No inhibition of the catalytic activity was detected up to 10% (v/v) methanol (Fig 12) .

Stability studies (repetitive batches)

The stability of P. pastoris -S.c.-BDH whole cell biocatalysts and the possibility to use this catalyst in continuously operated biotransformation processes was evaluated by performing repetitive batch studies.

P. pastoris cells were first cultivated in 25 mL YPD medium and induced by adding 25 mL BM2M as described previously. Biomass was then removed from the cultivation medium by centrifuga tion (5 min, 4000 rpm, 4°C, Centrifuge 5810R, Eppendorf, Hamburg, Germany) . The supernatant was removed and the cell pellet resuspended in 50 mL biotransformation substrate solution containing acetoin and methanol. Two different buffers with different pH values were applied and compared.

The biotransformation substrate solution including P. pastoris cells were then transferred back to a 250 mL baffled shake flask. The biotransformation conditions were 28 °C and 120 rpm. Samples were taken periodically and analyzed via gas chromatography as previously described.

After 4 hours, cells were again removed from the biotransformation solution by centrifugation. The supernatant was removed and the cell pellet again resuspended in 50 mmol/L fresh biotransformation solution and the biotransformation process was repeated as described above.

Fig. 13 shows the increasing concentration of the product 2, 3-butanediol . The catalytic activity in both cycles was almost similar, indicating that P. pastoris should be a stable whole cell biocatalyst in continuously operated applications. The catalytic activity at pH 6.0 is slightly higher compared to the activity at pH 7.5.

Table. Conversion of acetoin to 2, 3-butanediol in deepwell plates: conversion, optical density at the start of bioconversion and the specific growth rate (μ) during the bioconversion are shown.

Example 12: Improving P. pastoris-S . c . BDH-strains for wholecell biotransformations Kinetic studies 1.) Construction of the strain P. pastoris-S . c. -BDH+AOX1 The first enzyme of the methanol dissimilation pathway, the alcohol oxidase (AOXl) was overexpressed in P. pastoris in order to study the effects of its overexpression on cofactor regeneration via the methanol dissimilation pathway.

First, the complete P. pastoris AOXl gene (GenBank No. U96967) was amplified from genomic DNA of Pichia pastoris CBS7435 using the primer pair Eco_AOXl_fw (accgaattcatggctatccccgaagag) / Not_A0Xl_rv (accgcggccgcttagaatctagcaagacc) and employing the Phusion™ High-Fidelity DNA-polymerase (Finnzymes Oy, Finland) according to the manufacturer's protocol. The PCR-fragment was purified according to the QIAquick PCR Purification Kit (Qiagen) , digested with EcoRI and Notl (from Fermentas) , again purified using the QIAqick Gel Extraction Kit (Qiagen) and used for the ligation into the EcoRI/Notl cut E. coli/P. pastoris shuttle vector 'Kan3,2-8' (see Fig. 3) . The nucleotide sequence of the shuttle vector 'Kan3,2-8' is shown in Fig. 4.

The resulting Kan3, 2-8-AOX1 vector was controlled by sequencing, linearised by SacI digestion and used for P. pastoris transformation. First, the strains P. pastoris CBS7435, P. pastoris 742D4 (BDHl) and P. pastoris 764D10 (dasl das2 BDHl) were used. P. pastoris transformations were performed according to common methods. Transformants were selected on YPD/Geneticinagar plates (300 μg/mL Geneticin - final concentration) . Then, transformants were selected for the determination of increased AOXl-copy number by RT-PCR (real time polymerase chain reaction) . In addition, BDH-activity was investigated and finally, only those strains were chosen for further investigation which showed BDH-activity and increased AOXl copy number.

RT-PCR:

Copy numbers of the expression cassette were determined by quantitative real-time PCR using the P. pastoris ARG4 gene as reference gene. Quantitative real-time PCR was performed using the Power SYBR^® Green PCR Master Mix (Applied Biosystems) in an ABI PRISM 7300 Real Time PCR System (Applied Biosystems) . AOXl_fw_RT (GAAGCTGCCCTGTCTTAAACCTT) /A0Xl_rv_RT

(CAAAAGCTTGTCAATTGGAACCA) and ARG4-RTfw (TCCTCCGGTGGCAGTTCTT) /ARG4-RTrv (TCCATTGACTCCCGTTTTGAG) were used as primers at concentrations of 200 nM with 1 ng of genomic DNA as template. Tern perature conditions were 10 min at 95°C, 40 cycles of 15 s at 95°C and 60 s at 60°C followed by a dissociation step (15 s at 95°C, 30 s at 60 °C, 15 s at 95°C) at the end of the last cycle.

2.) Fermentation:

250 mL baffled shake flasks containing 50 mL BMGY medium (Table 5) were inoculated with a single colony of Pichia pastoris CBS7435, Pichia pastoris CBS7435+BDH, Pichia pastoris CBS7435+BDH+AOX1, Pichia pastoris CBS7435+BDH+FLD and Pichia pastoris CBS7435+BDH+FDH, respectively and incubated at 28 °C and 120 rpm (Certomat BS-I, B. Braun Biotech, Germany) for approximately 24 hours.

As second pre-culture, 1 L baffled shake flasks containing 200 mL YPD medium were inoculated with 500 μl of the first preculture and incubated at 28 °C and 120 rpm until an optical density of 10 was reached. 400 mL of the second pre-culture were transferred to the fermenter (Biostat CT5-2, B. Braun Biotech, Germany) containing 3.5 L modified basal salt medium.

Fermentation started with a batch phase where glycerol from modified basal salt medium was used as sole carbon source. The pH value was set to 5.0 and adjusted by adding a 25% ammonia solution (technical quality) . For the cultivation the following parameters were chosen: temperature: 28°C, aeration: 2.5 - 1OL air/min, agitation control: 500 - 1500 rpm and oxygen partial pressure: >30% of the saturation concentration. After -16 hours the whole amount of glycerol was metabolized and the first fedbatch phase was started by feeding a 700 g/L glycerol solution containing 4,35 mL/L PTMl solution. The glycerol solution was fed over a time period of 25 hours and was increased stepwise. Then the second fed-batch phase was started by feeding a methanol solution containing 4,35 mL/L PTMl solution. Methanol was fed over a time period of 72 hours and feeding was increased stepwise .

Biomass was harvested by centrifugation (20 min, 4000 rpm, 4°C, Avanti J-20 XP, Beckman Coulter, Germany, JA-10), washed with a 50 mmol/L KPi buffer, pH 7.5, and the cell pellets were stored at -20°C.

3.) cell disruption

Cell pellets were suspended in 50 mmol/L KPi buffer, pH 7.5, and disrupted by sonification (10 min, Sonifier 250, Branson, USA) . Cells and cell debris were removed by ultracentrifugation (rotor TI70, 40 min, 40000 rpm, 4°C, Optima LE-80K Ultrazentrifuge, Beckman Coulter, Germany) . The supernatant was investigated with respect to FLD, FDH and AOXl-activity .

4.) kinetic investigation

The kinetic properties of the investigated enzymes FLD and FDH were analyzed photometrically by monitoring the NAD consumption at 340 nm (Spectramax plus, Molecular Devices, USA) over a time period of 2 minutes. Protein concentration was determined by a BCA protein assay (Thermo Scientific) according to the manufacturer's protocol.

Kinetic properties of FLD:

The kinetic properties of formaldehyde dehydrogenase (FLD) were determined applying varying concentrations of formaldehyde in the range of 0.0025 to 10 mmol/L and varying concentrations of the cofactor NAD in the range of 0.02 to 7 mmol/L. The assay conditions for varying concentrations of formaldehyde were: 28°C, 5 mmol/L NAD, 10 mmol/L glutathione, 0.25 mg/mL protein in 50 mmol/L potassium phosphate buffer, pH 7.5. The assay conditions for varying concentrations of NAD were: 28 °C, 2.5 mmol/L formaldehyde, 10 mmol/L glutathione, 0.25 mg/mL protein in 50 mmol/L potassium phosphate buffer, pH 7.5.

Kinetic properties of FDH:

The kinetic properties of formate dehydrogenase (FDH) were determined applying varying concentrations of formate in the range of 0.05 to 100 mmol/L and varying concentrations of the cofactor NAD in the range of 0,02 to 5 mmol/L. The assay conditions for varying concentrations of formate were: 28 °C, 1 mmol/L NAD, 0.25 mg/mL protein in 50 mmol/L potassium phosphate buffer, pH 7.5. The assay conditions for varying concentrations of NAD were: 28 °C, 200 mmol/L formate, 0.25 mg/mL protein in 50 mmol/L potassium phosphate buffer, pH 7.5.

Kinetic properties of AOX:

The AOX assay was performed as described by C. Jungo et al. (Quantitative characterization of the regulation of the synthesis of alcohol oxidase and of the expression of recombinant avidin in a Pichia pastoris Mut⁺ strain. Enzyme and Microbial Technology 39: 936-944, 2006) . The only difference was that the assay was performed at 28 °C and in a 50 mM KPi buffer pH 7.5. Calibration was made with an AOX solution from P. pastoris pur- chased from Sigma Aldrich (A2404-250U) . Varying concentrations of the substrate methanol were applied in the range of 0.001% to 10% (v/v) .

The kinetic properties of the enzymes AOX, FLD and FDH were derived from the following equations:

The following kinetic parameters were determined for the following five strains: Pichia pastoris CBS7435, Pichia pastoris CBS7435+BDH, Pichia pastoris CBS7435+BDH+AOX1, Pichia pastoris CBS7435+BDH+FLD and Pichia pastoris CBS7435+BDH+FDH.

Pichia pastoris CBS7435:

Pichia pastoris CBS7435+BDH+FDH:

A model was built based on the kinetic properties of the enzymes FLD, FDH and AOX. This model helped to determine the NADH regeneration rates for the methanol metabolism of the five engineered strains. Regarding the kinetic parameters, it could be proposed that the overexpression of the formaldehyde dehydrogenase should be the most effective step in order to increase the internal production of NADH. In order to confirm this prediction, whole cell biotransformation processes applying acetoin as subtrate were carried out. A comparison of all investigated strains pointed out that the highest catalytic activity for the reduction of acetoin and production of the corresponding 2,3butanediol was reached with the engineered Pichia pastoris CBS7435+BDH+FLD strain which showed increased expression of the intracellular enzyme formaldehyde dehydrogenase.

Biotransformations with engineered P. pastoris-S . c. BDHstrains

Biotransformation processes were carried out applying either the wild type of Pichia pastoris with recombinant BDH or an engineered strain with overexpressed FLD, FDH and AOXl, respectively and recombinant BDH. In addition, the Pichia pastoris CBS7435 wild type strain without BDH was tested for comparison. P. pastoris cells were cultivated in 25 mL YPD medium in 250 mL baffled shake flasks at 28°C and 120 rpm (Certomat BS-I, B. Braun Biotech, Germany) . After a cultivation period of 60 hours, 25 mL BM2M containing methanol were added to induce protein expression .

After 12 hours of induction at 28°C and 120 rpm, 10 mL biotransformation substrate solution were added to the Pichia pastoris culture and biotransformation was started. After addition of acetoin and methanol, samples were taken periodically and analyzed via gas chromatography. 400 μL biotransformation medium were mixed with 100 μL of a 50 mmol/L n-butanol solution which served as internal standard for gas chromatography. The resulting solution was extracted with 500 μL ethyl acetate. The organic phase was analyzed via gas chromatography on a Varian CP7503 gas chromatograph with a Chirasil DEXCD column (25 m x 0,32 mm i. d.) with a flame ionization detector and hydrogen as carrier gas. Compounds were detected using the following temperature program: 6.5 min 65°C, 50°C/min, 0.7 min 80°C, 2°C/min, 3.0 min 85°C. Typical retention times were: n-butanol: 3.3 min, (S)-acetoin: 3.8 min, (R)-acetoin: 4.4 min, (2S, 3S) -butanediol: 11.1 min, (2R, 3R) -butanediol : 11.6 min, meso-butanediol : 12.5 min.

The P. pastoris strain with overexpressed FLD showed a slightly higher catalytic activity (Fig. 14) . The initial reaction velocities of both strains were 50.1 U/gCww for P. pastoris BDH 742 D4 and 67.9 U/gCww for P. pastoris BDH 742 D4 FLD.

These first results were confirmed by comparing batch biotransformation processes employing the P. pastoris strains: Pichia pastoris CBS7435 (WT) , Pichia pastoris CBS7435+BDH (WTBDH), Pichia pastoris CBS7435+BDH+AOX1 (AOX-BDH), Pichia pastoris CBS7435+BDH+FLD (FLD-BDH) and Pichia pastoris CBS7435+BDH+FDH (FDH-BDH) (see Fig. 14b) .

Since the presence of evaluated diacetyl concentrations (>20 mmol/L) inhibits the reduction of acetoin to 2, 3-butanediol, the two-step reduction of diacetyl to 2, 3-butanediol is only feasible with some experimental restrictions. Fig. 14c shows the comparison of biocatalytic reduction of diacetyl in batch mode and fed-batch mode. When 100 mmol/L diacetyl are applied in a batch process only marginal formation of the product 2, 3-butanediol was observed. When diacetyl is applied stepwise to the biotransformation process, higher amounts of 2, 3-butanediol can be produced.

Both biotranformation processes were carried out in 300 mL shake flasks with baffles in 50 mmol/L KPi buffer, pH 6.0 at 28°C and 120 rpm. The reaction volume was 20 mL and the biocatalyst concentration 15 gc_Dw/L. For the biotransformation in batch mode, 100 mmol/L diacetyl were applied. The biotransformation in fedbatch mode was started applying 20 mmol/L diacetyl and 10% methanol. After 2, 4, 6 and 8 hours further 52,44 μmol/L of diacetyl were added so that finally the same amount of diacetyl was applied to both biotransformation processes.

Example 13: Expression of C. tenuis Xylose reductase (CtXR) in P. pastoris and improving P. pastoris-CtXR-strains for wholecell biotransformations

Expression of the wild type gene of C. tenuis Xylose reductase in P. pastoris .

E. coli plasmids with the C. tenuis Xylose reductase wild type gene (CtXRwt; see Genbank accession number AF074484) and a double mutant thereof (CtXR-K274R-N276D) encoding for a xylose reductase with increased NADH-activity (B. Petschacher, et al. Biochemical Journal 2005, 385, 75) were provided.

The primer pairs XR-Eco_f/Bgl_remo_r and Bgl_remo_f/XR-Not-r were used to amplify fragments of CtXRwt and CtXR-K274R-N276D, respectively. At the same time, the internal BgIII restriction recognition site was removed. Fragment 1 and 2 were linked by oePCR according to the protocol depicted in Example 10, using the outer primers XR-Eco_f/XR-Not-r for a final amplification step. The oePCR products were purified with the QIAqick PCR Purification Kit (Qiagen) , cut with EcoRI and NotI (from Fermentas) , again purified with the QIAqick Gel Extraction Kit (Qiagen) and used for the ligation into the EcoRI/Notl cut E. coli/ P. pastoris shuttle vector 'T2' (see Figs. 15 and 16 for plasmid map and sequence) . The resulting E. coli/P. pastoris shuttle vectors were named 'T2-CtXRwt' and 'T2-CtXRmut' , respectively and controlled by sequencing.

Then, the vector was linearised with Bg1II (from Fermentas) and used for P. pastoris transformations. The strains P. pastoris CBS7435 and P. pastoris D12-B1H {dasl das2) and strains from Example 8, Example 9 and Example 10, respectively, namely P. pastoris CBS7435-FDH, P. pastoris CBS7435-FLD and P. pastoris CBS7435-FDH+FLD were also used. The transformants were selected on YPD/Zeocin (50 μg/mL) -agar plates. Then, transformants were cultivated in 96-well deep well plates and assayed for increased FDH and FLD activity.

Cultivation in 96-well deep well plates and cell disruption:

The cultivation of P. pastoris transformants in 96-well deep well plates and the subsequent cell lysis with Y-Per^® were performed as described in Example 8. Xylose reductase-activity assay:

The xylose reductase activity was determined photometrically, tracking the decrease in absorption of NADPH (CtXRwt) and NADH (CtXR-K274R-N276D = CtXRmut), respectively. The reaction mixture (200 μL) consisted of a 50 mM potassium phosphate buffer (pH 7.0), 850 mM D (+) -xylose, and 300 μM NADPH (CtXRwt) and 400 μM NADH ( CtXR-K274R-N276D = CtXRmut), respectively. In addition, 10 μL of crude cell lysate were added per well. The reaction was started by the addition of the NAD(P)H stock solution and the decrease in absorption of NAD(P)H was tracked at 340 nm for 2-5 min.

The best P. pastoris D12-B1H {dasl das2) -CtXRwt and -CtXRK274R-N276D strain, respectively, was used for a P. pastoris cotransformation. Linearised E. coli/P. pastoris shuttle vectors of Example 8, 9 and 10, namely 'Kan3, 2-8-FDH+FLD' , 'Kan3,2-8FDH' and 'Kan3, 2-8-FLD' were employed. The transformants were selected on YPD/Geneticin (300 μg/mL) -agar plates. Then, transformants were cultivated in 96-well deep well plates and assayed for increased FDH and FLD activity. The remaining Xylose reductase activity was also determined.

FLD-activity assay:

The FLD-activity assay was performed according to Example 8.

FDH-activity assay:

The FDH-activity assay was performed according to Example 9.

Fig. 17 shows the successful expression of CtXRwt and CtXRmut (CtXR-K274R-N276D) in P. pastoris. Most of the transformants show significantly increased Xylose reductase-activity compared to the initial P. pastoris strains. Figs. 18 - 23 depict the results for the coexpression of NADH-regeneration enzymes (FLD, FDH and FLD+FDH, respectively) and genes encoding for Xylose reductase. In all cases, significant Xylose reductase-activity was detected. The activity for FLD and FDH remained comparable to the initial P. pastoris strains used for the co-transformation (P. pastoris CBS7435-FDH, P. pastoris CBS7435-FXD and P. pastoris CBS7435-FDH+FLD) .

Expression of a synthetic gene of C. tenuis Xylose reductase in P. pastoris

A synthetic CtXR gene (Fig. 24) was designed in order to improve the expression of CtXR in the recombinant host P. pastoris. The CtXR wild type gene was codon optimized according to the high-methanol-codon usage of Pichia pastoris which was calculated based on highly expressed genes in P. pastoris such as the genes encoding for AOXl, DASl, FLDl and HbHNL. In addition, mRNA 3' -end-processing signals and restriction enzyme recognition sequences were eliminated. This was done with the program GeneDesigner from DNA2.0. The resulting sequence was again checked with the program Letol.O from Entelechon and afterwards, the synthetic gene was ordered at Genscript.

This synthetic gene was then amplified with Phusion™ HighFidelity DNA-polymerase (Finnzymes) according to the manufacturer' s protocol and using the primer pair ctxr syn E fwd/CtXR syn N rev. For the generation of the mutant CtXR-K274R-N276D, first, two fragments were amplified with the following primer pairs 'ctxr syn E fwd/ K274R N276D syn rev' and 'K274R, N276D syn n fwd/ ctxr syn N rev' , respectively. These two fragments were fused by oePCR according to Example 10, using the outer primers 'ctxr syn E fwd/ctxr syn N rev' for a final amplification step. Then, the amplified CtXRsyn_wt and CtXRsyn_mut gene were purified with the QIAqick PCR Purification Kit (Qiagen) , cut with EcoRI and Notl (from Fermentas), again purified with the QIAqick Gel Extraction Kit (Qiagen) and used for the ligation into the EcoRI/Notl cut E. coli/P. pastoris shuttle vector 'T2' (see Figs. 15 and 16 for plasmid map and sequence). The resulting E. coli/P. pastoris shuttle vectors were named 'T2CtXRsyn_wt' and 'T2-CtXRsyn_mut' , respectively. The sequence was controlled by sequencing.

Then, the vector was linearised with BgIII (from Fermentas) and used for P. pastoris transformations. The strains P. pastoris CBS7435 and P. pastoris D12-B1H (das1 das2) , respectively, were used. The transformants were selected on YPD/Zeocin (50 μg/mL) -agar plates. Then, transformants were cultivated in 96well deep well plates and assayed for increased FDH and FLD activity.

Cultivation in 96-well deep well plates and cell disruption:

The cultivation of P. pastoris transformants in 96-well deep well plates and the subsequent cell lysis with Y-Per^® were performed as described in Example 8.

Xylose reductase-activity assay:

The Xylose reductase activity assay was performed according to Example f - 'Expression of the wild type gene of C. tenuis Xylose reductase in P. pastoris' .

The best P. pastoris D12-B1H {dasl das2) -CtXRsyn_wt and CtXRsyn_mut strain and the best P. pastoris CBS7435-CtXRsyn_wt and -CtXRsyn_mut strain, respectively, was used for a P. pastoris co-transformation. Linearised E. coli/P. pastoris shuttle vectors of Example a, b and c, namely 'Kan3, 2-8-FDH+FLD' , 'Kan3,2-8-FDH' and 'Kan3, 2-8-FLD' were employed and the transformations were performed. The transformants were selected on YPD/Geneticin (300 μg/mL) /Zeocin (50 μg/m) -agar plates. Then, transformants were cultivated in 9β-well deep well plates and assayed for increased FDH and FLD activity. The remaining Xylose reductase activity was also determined.

FLD-activity assay:

The FLD-activity assay was performed according to Example 8.

FDH-activity assay:

The FDH-activity assay was performed according to Example 9.

In Fig. 25, landscapes of P. pastoris-CtXRsyn_wt and - CtXRsyn_mut {CtXR-K214R-N216D) transformants are depicted. Most of the transformants show a significantly increased Xylose reductase activity compared to the initial P. pastoris wild type strains used for the transformation.

The Figs. 26 - 31 depict the results for the co-expression of genes encoding for NADH-regeneration enzymes (FLD, FDH and FLD+FDH, respectively) and synthetic genes encoding for Xylose reductase. For most of the transformants, the Xylose reductaseactivity was comparable to the Xylose reductase-activity of the strains used for the co-transformation (P. pastoris CBS7435CtXRsyn_wt and -CtXRsyn__mut) . The activity for FLD and FDH was in most of the cases significantly or slightly improved compared to the initial P. pastoris strains used for the cotransformation (P. pastoris CBS7435-FDH, P. pastoris CBS7435-FLD and P. pastoris CBS7435-FDH+FLD) .

Example 14: Engineered P. pastoris strain for efficient conversion of glycerol

Generation of the strain P. pastoris ΔKU70 for the efficient gene-knockout in P. pastoris:

First, a KU70-knockout cassette was constructed which consisted of a 3'- and 5' -flanking arm, homologous to the KUlO gene, two adjacent FRT-sequences and in-between, an expression cassette for the FLP recombinase and the Zeocin resistance cas sette ( see Fig . 32 ) .

This knockout cassette was ligated into the pJETl .2/blunt vector provided within the CloneJET™ PCR Cloning Kit from Fermentas and used for E. coli transformation. All steps were performed according to the manufacturer's protocol. The resulting vector was named pJET-48-4 and its sequence surrounding the FRTsequences were controlled by sequencing.

The KU70-knockout-cassette was then amplified by PCR using the pJET-48-4 vector as template, the primer pair UpStrm2_fwd/ DnStrm_rev and and the Phusion™ High-Fidelity DNA-polymerase (Finnzymes Oy, Espoo, Finland) according to the manufacturer's protocol. After purification with the QIAqick Gel Extraction Kit (Qiagen) , the linear fragment was used for P. pastoris CBS7435 transformation. Transformants were selected on YPD/Zeocin (50 μg/mL)-agar plates. Then, the existance of a single copy and the correct integration of the cassette were controlled by PCR. Positive transformants were then repeatedly streaked on minimal methanol-agar medium and incubated at 28 °C. Thus, the FLP recombinase which was under the control of the P_A0X1 promoter was expressed and catalyzed the removal of the sequence located inbetween the two FRT-sequences leaving one FRT-sequence behind. The thus generated new strains were again cultivated on minimal dextrose and YPD/Zeocin-agar plates. Finally, the strain P. pastoris CBS7435 48-4-3 / 2,4 (P. pastoris CBS7435 ΔKU70) which lost its Zeocin resistance but was KU70 negative was chosen for further investigation.

Generation of a P. pastoris glycerol kinase (glykin) knockout cassette :

The P. pastoris glycerol kinase gene was amplified from genomic DNA from P. pastoris CBS7435 using the primer pair glykin_fwd/glykinl_rev and the Phusion™ High-Fidelity DNApolymerase (Finnzymes Oy, Finland) according to the manufacturer's protocol.

Then a simple glykin knockout cassette was constructed linking a 3'- and a 5' -flanking arm which were homologous to the 3'and 5' -end of the glykin gene with a zeocin resistance cassette (see Fig. 33) . Therefore, the primer pair AD-zeo-1-f/zeo-AD-2-r was used to amplify the Zeocin resistance cassette from the Invitrogen standard plasmid pPICZ . The homologous flanking regions were amplified using the primer pairs glykinl_fwd/ GK-zeo-1-r and GK-zeo-2-f/glykin_rev. Again the Phusion™ High-Fidelity DNApblymerase was used according to the manufacturer's protocol. After purification of the PCR-products with the QIAqick Gel Extraction Kit (Qiagen) the fragments were linked by oePCR according to Example c, using the outer primers glykinl_fwd/glykin_rev for a final amplification step.

The oePCR product was again purified using the QIAqick Gel Extraction Kit (Qiagen) and the resulting linear fragment was directly employed for P. pastoris CBS7435 ΔKU70 transformation. Transformants were selected on YPD/Zeocin (50 μg/mL)-agar plates. Then, transformants were cultivated in 250 μL YPD-medium in 96-well deep well plates (320 rpm, 28 °C, 80% air humidity, 12 h) . These cultures were replicated on YPD/Zeocin (50 μg/mL) -agar plates and minimal glycerol medium agar plates (MG1%) using a 96-pin stemp and again incubated at 28 °C. After 60 h, most of the transformants showed normal growth on YPD/Zeocin but significantly reduced growth on MG1% medium.

Thus, the newly generated P. pastoris strains provide the basis for an efficient yeast whole cell biocatalyst for a multistep conversion such as the conversion of glycerol to e.g. 1, 3-propanediol as the assimilatory pathway of glycerol was interrupted by introducing the glycerol kinase knockout. In addition, a new auxotrophic P. pastoris platform strain for molecular biology was generated. A first experiment proved the possibility to complement the auxotrophy and to select corresponding transformants on MG 1% medium.

Claims

Claims :

1. Recombinant Pichia pastoris cell comprising at least one mutation in the dihydroxyacetone synthase gene 2 (DAS2) resulting in a reduced dihydroxyacetone synthase activity compared to a wild-type Pichia pastoris cell, wherein the DAS2 gene comprises a nucleotide sequence having at least 80% identity with SEQ ID No. 1.

2. Cell according to claim 1, characterised in that said cell comprises further a mutation of the dihydroxyacetone synthase gene 1 (DASl) .

3. Cell according to claim 1 or 2, characterised in that the mutation is a deletion-, an insertion or a substitution.

4. Cell according to any one of claims 1 to 3, characterised in that the cell comprises an increased formaldehyde dehydrogenase

(FLD) enzyme activity and/or formate dehydrogenase (FDH) enzyme activity compared to a wild-type Pichia pastoris cell.

5. Cell according to claim 4, characterised in that the increased enzyme activity is that of an enzyme, endogenous to said cell, encoded by a nucleic acid coding sequence operably linked to at least one regulatory sequence not natively associated with said nucleic acid coding sequence, whose expression is increased as compared to the expression of the enzyme activity when said nucleic acid coding sequence is associated with its native regulatory sequence or of an enzyme, exogenous to said cell, encoded by a nucleic acid coding sequence, operably linked to at least one regulatory sequence.

6. Cell according to any one of claims 1 to 5, characterised in that said cell comprises at least, one exogenous nucleic acid molecule encoding at least one NADH or NADPH dependent enzyme operably linked to at least one regulatory sequence.

7. Cell according to claim 5 or 6, characterised in that the at least one regulatory sequence is a promoter, preferably an alcohol oxygenase promoter, more preferably an alcohol oxygenase promoter from Pichia pastoris, even more preferably AOXl or a variant thereof.

8. Cell according to claim 7, characterised in that the at least one NADH or NADPH dependent enzyme is selected from the group consisting of butanediol dehydrogenase (EC 1.1.1.4), xylose reductase (EC 1.1.1.21), alcohol dehydrogenase (EC 1.1.1.1) enoate reductase (EC 1.3.1.31) and NADPH-kenoprotein reductase (EC 1.6.2.4) .

9. Cell according to claim 8, characterised in that the butanediol dehydrogenase is of Saccharomyces cerevisiae (Ace. No. NP009341) , xylose reductase of Candida tenuis (Ace. No. AAC25601), alcohol dehydrogenase of Candida magnoliae.

10. Cell according to claim 9, characterised in that the xylose reductase of Candida tenuis comprises a K274R and/or N276D mutation.

11. Cell according to any of claims 1 to 10, characterised in that the glycerol kinase gene of said cell comprises at least one mutation resulting in a reduced glycerol kinase activity compared to a wild-type Pichia pastoris cell.

12. Use of a cell according to any one of claims 1 to 11 in whole cell biocatalysis .

13. A method for catalysing an educt to a product comprising a) cultivating a cell according to any one of claims 1 to 11, b) incubating an educt with the culture of a) , and c) isolating the synthesised product.