Organisms for the production of 1,3-butanediol

Abstract

A non-naturally occurring microorganism comprising a microorganism having a 1,3-butanediol (1,3-BDO) pathway having at least one exogenous nucleic acid encoding a 1,3-BDO pathway enzyme expressed in sufficient amount to produce 1,3-BDO. A method for producing 1,3-BDO comprising cultivating such a microorganism under a plurality of conditions for a sufficient period of time to produce 1,3-BDO.

Description

[0001] This application is a divisional application of the invention application filed on April 30, 2010, with application number 201080029715.X and entitled "Biotechnology for the Production of 1,3-Butanediol".

[0002] Application instructions

[0003] This application claims priority to U.S. Provisional Application No. 61 / 174,473, filed April 30, 2009, the entire contents of which are incorporated herein by reference. Background of the Invention

[0005] Generally, this invention relates to biosynthetic processes and organisms capable of producing organic compounds. More specifically, this invention relates to non-naturally occurring organisms capable of producing the chemical commodity 1,3-butanediol.

[0006] 1,3-Butanediol (1,3-BDO) is a tetracarbon diol traditionally produced from acetylene through the hydration of acetylene. The resulting acetaldehyde is then converted to 3-hydroxybutyraldehyde, which is subsequently reduced to form 1,3-BDO. In recent years, acetylene has been replaced by the cheaper ethylene as the source of acetaldehyde. 1,3-BDO is commonly used as an organic solvent in food flavorings. It is also used as a comonomer for polyurethanes and polyester resins and is widely used in hypoglycemic drugs. Optically active 1,3-BDO is a useful starting material for the synthesis of bioactive compounds and liquid crystals. The primary commercial use of 1,3-butanediol is subsequent dehydration to provide 1,3-butadiene (Ichikawa et al., J. of Molecular Catalysis A-Chemical, 256:106-112 (2006); Ichikawa et al., J. of Molecular Catalysis A-Chemical, 231:181-189 (2005)), a petrochemical product used in the manufacture of synthetic rubbers (such as tires), latexes, and resins, with a production rate of 25 billion lbs per year. This reliance on petroleum-based feedstocks for acetylene or ethylene necessitates the development of pathways for the production of 1,3-butanediol and butadiene based on renewable feedstocks.

[0007] Therefore, there is a need to develop microorganisms and methods for producing 1,3-BDO using them. This invention meets this need and also provides related advantages. Invention Overview

[0009] In some embodiments, the present invention targets a non-naturally occurring microorganism, including microorganisms possessing a 1,3-butanediol (1,3-BDO) pathway having at least one exogenous nucleic acid encoding a 1,3-BDO pathway enzyme expressed in sufficient quantities to produce 1,3-BDO. The 1,3-BDO pathway includes enzymes selected from: 2-amino-4-ketovalerate (AKP) thiolytic enzyme, AKP dehydrogenase, 2-amino-4-hydroxyvalerate transaminase, 2-amino-4-hydroxyvalerate oxidoreductase (deamination), 2-oxo-4-hydroxyvalerate decarboxylase, 3-hydroxybutyraldehyde reductase, AKP transaminase, AKP oxidoreductase (deamination), 2,4-dioxovalerate decarboxylase, 3-oxobutyraldehyde reductase (ketone reduction), 3-oxobutyraldehyde reductase (aldehyde reduction), 4-hydroxy-2-butanone reductase, AKP decarboxylase, 4-amino... Butyl-2-ketotransaminase, 4-aminobutyric-2-keto oxidoreductase (deamination), 4-aminobutyric-2-keto ammonia-lyase, butenone hydratase, AKP ammonia-lyase, acetoacrylate decarboxylase, acetoacetyl-CoA reductase (CoA-dependent, aldehyde formation), acetoacetyl-CoA reductase (CoA-dependent, alcohol formation), acetoacetyl-CoA reductase (keto reduction), 3-hydroxybutyryl-CoA reductase (aldehyde formation), 3-hydroxybutyryl-CoA reductase (alcohol formation), 4-hydroxybutyryl-CoA dehydratase, and crotonic acid oxidase.

[0010] In some embodiments, the present invention relates to a method for producing 1,3-BDO, the method comprising culturing such non-naturally occurring microorganisms under various conditions for a sufficient period of time to produce 1,3-BDO. Attached Figure Description

[0011] Figure 1 The pathway from alanine to 1,3-BDO is shown. The enzymes are as follows: A) AKP thiolyase, B) AKP transaminase or AKP oxidoreductase (deamination), C) 2,4-dioxovalerate decarboxylase, D) 3-oxobutyraldehyde reductase (aldehyde reduction), E) AKP decarboxylase, F) 4-aminobutyric acid-2-ketoamine-lyase, G) butenone hydratase, H) 4-hydroxy,2-butanone reductase, I) AKP ammonia-lyase, J) acetylacrylate decarboxylase, K) 4-aminobutyric acid-2-ketotransaminase or 4-aminobutyric acid-2-keto oxidoreductase (deamination), L) AKP dehydrogenase, M) 2-amino-4-hydroxyvalerate transaminase or 2-amino-4-hydroxyvalerate oxidoreductase (deamination), N) 2-oxo-4-hydroxyvalerate decarboxylase, O) 3-oxobutyraldehyde reductase (keto reduction), and P) 3-hydroxybutyraldehyde reductase.

[0012] Figure 2The pathway from acetoacetyl-CoA to 1,3-butanediol is illustrated. The enzymes involved are as follows: A) acetoacetyl-CoA reductase (CoA-dependent, aldehyde formation), B) 3-oxobutyraldehyde reductase (ketone reduction), C) 3-hydroxybutyraldehyde reductase, D) acetoacetyl-CoA reductase (CoA-dependent, alcohol formation), E) 3-oxobutyraldehyde reductase (aldehyde reduction), F) 4-hydroxy,2-butanone reductase, G) acetoacetyl-CoA reductase (ketone reduction), H) 3-hydroxybutyryl-CoA reductase (aldehyde formation), and I) 3-hydroxybutyryl-CoA reductase (alcohol formation).

[0013] Figure 3 The pathway from 4-hydroxybutyryl-CoA to 1,3-butanediol is shown. The enzymes are as follows: A) 4-hydroxybutyryl-CoA dehydratase, B) crotonylase, C) 3-hydroxybutyryl-CoA reductase (aldehyde formation), D) 3-hydroxybutyraldehyde reductase, and E) 3-hydroxybutyryl-CoA reductase (alcohol formation).

[0014] Figure 4 This shows an aldehyde dehydrogenase that exhibits significant activity against 3-hydroxybutyl-CoA.

[0015] Figure 5 The specific activity of bld from Clostridium saccharoperbutylacetonicum to 3-hydroxybutyryl-CoA is shown before and after dialysis.

[0016] Figure 6 The concentrations of 1,3-BDO are shown when 3-hydroxybutyraldehyde is added as a substrate and in the control sample without a substrate. The GI number of the alcohol dehydrogenase is shown.

[0017] Figure 7 The concentrations of 1,3-BDO are shown when 3-hydroxybutyryl-CoA is added as a substrate and in the control sample without a substrate. The GI number of the alcohol dehydrogenase is shown. The GI number of the aldehyde dehydrogenase tested simultaneously is 163762382. Invention Details

[0019] This invention targets, in part, non-naturally occurring microorganisms that express genes encoding enzymes that catalyze the production of 1,3-butanediol (1,3-BDO). The pathway for 1,3-butanediol production disclosed herein is based on three precursors: (i) D-alanine, (ii) acetoacetyl-CoA, and (iii) 4-hydroxybutyryl-CoA. Successfully designing these pathways requires identifying a suitable set of enzymes with sufficient activity and specificity, cloning their corresponding genes into a production host, optimizing fermentation conditions, and determining the product formation after fermentation.

[0020] The conversion of alanine to 1,3-BDO can be achieved through multiple pathways involving approximately five enzymatic catalytic steps, such as... Figure 1 As shown. In the first step of all pathways (step A), alanine and acetyl-CoA are bound by 2-amino-4-ketovalerate thiolysolase (a highly selective enzyme). The product of this reaction, 2-amino-4-oxovalerate (AKP), can then be transaminated, reduced, decarboxylated, or deaminated, as... Figure 1 As shown. Further synthetic steps for the production of 1,3-BDO are discussed in detail below. The theoretical yield of 1,3-BDO from each of these pathways is calculated to be approximately 1.09 mol / mol of glucose consumed.

[0021] Figure 2 Several pathways for the production of 1,3-BDO from acetoacetyl-CoA are outlined. Each of these pathways utilizes three reducing equivalents and provides a theoretical yield of 1 mole of 1,3-BDO per mole of glucose consumed. Other carbon sources, such as syngas, can also be used to produce acetoacetyl-CoA. Glucose vaporization to form syngas will result in the highest theoretical yield of 1.09 moles of 1,3-BDO per mole of glucose consumed (assuming 6 moles of CO and 6 moles of H2 are obtained from glucose).

[0022] 6CO + 6H₂ → 1.091C₄H 10 O2 + 1.636CO2 + 0.545H2

[0023] 4-Hydroxybutyryl-CoA is an important starting metabolite from which many industrially useful compounds, including 1,3-BDO, can be produced, such as... Figure 3 As shown. Although 4-hydroxybutyryl-CoA is not a very common central metabolite, the method for designing strains to synthesize 4-hydroxybutyryl-CoA has been previously described by the applicant in U.S. Patent Application No. 2009 / 0075351. Assuming glucose as the carbohydrate starting material, the pathway from 4-hydroxybutyryl-CoA to 1,3-butanediol has a theoretical yield of 1.09 mol / mol of product.

[0024] This invention also relates in part to methods for producing 1,3-BDO by culturing these non-naturally occurring microorganisms. Dehydration of the 1,3-BDO produced by the organisms and methods described herein provides an opportunity to produce renewable butadiene in small-scale end-use facilities, avoiding the need to transport this flammable and reactive chemical.

[0025] As used herein, the term "non-naturally occurring" when used with respect to the microorganisms of the present invention means that the microorganism has at least one genetic alteration that would not normally be found in naturally occurring strains of the relevant species (including wild-type strains of the relevant species). Genetic alterations include, for example, modifications that introduce expressible nucleic acids encoding metabolic polypeptides, additions of other nucleic acids, deletions of nucleic acids, and / or other functional disruptions of the microbial genetic material. These modifications include, for example, coding regions and functional segments of heterologous polypeptides, homologous polypeptides, or heterologous and homologous polypeptides of the relevant species. Other modifications include, for example, non-coding regulatory regions in which the modification alters the expression of a gene or operon. Exemplary metabolic polypeptides include enzymes or proteins within the biosynthetic pathway of 1,3-butanediol.

[0026] Metabolic modification refers to biochemical reactions that alter a organism's naturally occurring state. Therefore, non-naturally occurring microorganisms can possess genetic modifications to nucleic acids encoding metabolic peptides or functional fragments thereof. This article discloses exemplary metabolic modifications.

[0027] As used herein, the term "isolated" when referring to microorganisms means an organism that, when found in nature, is substantially free of at least one component. This term includes microorganisms that have separated from some or all of their components when found in their natural environment. It also includes microorganisms that have separated from some or all of their components when found in a non-natural environment. Therefore, when an isolated microorganism is found in nature or when it is cultured, stored, or survives in a non-natural environment, it is partially or wholly isolated from other substances. Specific examples of isolated microorganisms include partially pure microorganisms, substantially pure microorganisms, and microorganisms cultured in non-natural media.

[0028] As used herein, the term "microorganism" is intended to mean any organism that exists as a microscopic cell contained within the domains Archaea, Bacteria, or Eukaryotes. Therefore, the term includes prokaryotic or eukaryotic cells or organisms of microscopic size and includes all kinds of bacteria, archaea, and eubacteria, as well as eukaryotic microorganisms such as yeast and fungi. The term also includes any kind of cell culture that can be cultured for the production of biochemicals.

[0029] As used herein, the term “CoA” or “coenzyme A” refers to an organic cofactor or prosthetic group (the non-protein portion of an enzyme) whose presence is required for the activity of many enzymes (apoenzyme proteins) to form an active enzyme system. Coenzyme A plays a role in certain condensing enzymes, in the transfer of acetyl or other acyl groups, in fatty acid synthesis and oxidation, pyruvate oxidation, and in other acetylation processes.

[0030] As used herein, the term "substantially anaerobic" when used with respect to culture or incubation conditions means that the oxygen content is less than about 10% of the saturation level of dissolved oxygen in a liquid culture medium. The term is also intended to include sealed chambers containing liquid or solid culture media maintained in an atmosphere of less than about 1% oxygen.

[0031] As used herein, “exogenous” means that the relevant molecule or activity is introduced into the host microorganism. The molecule may be introduced, for example, by introducing a coding nucleic acid into the host genetic material, such as by integration into the host chromosome or as non-chromosomal genetic material (e.g., plasmids). Therefore, when used in relation to the expression of a coding nucleic acid, the term refers to the introduction of the coding nucleic acid into the microorganism in an expressible form. When used in relation to biosynthetic activity, the term refers to the activity of the relevant organism introduced into the host. The source may be, for example, a homologous or heterologous coding nucleic acid expressing the activity after its introduction into the host microorganism. Therefore, the term “endogenous” refers to the relevant molecule or activity present within the host. Similarly, when used in relation to the expression of a coding nucleic acid, the term refers to the expression of the coding nucleic acid contained within the microorganism. The term “heterogeneous” refers to a molecule or activity from a different source than the relevant species, while “homogeneous” refers to a molecule or activity from the host microorganism. Therefore, the exogenous expression of the coding nucleic acid of the present invention can utilize one or both of heterologous and homologous coding nucleic acids.

[0032] The non-naturally occurring microorganism of the present invention may contain stable genetic alterations, meaning that the microorganism can be cultured for more than five generations without loss of alteration. Generally, stable genetic alterations include modifications that persist for more than 10 generations, specifically, stable modifications that persist for more than about 25 generations, and more specifically, stable genetic modifications that persist for more than 50 generations, including permanent alterations.

[0033] Those skilled in the art will understand that the genetic alterations described herein (including the metabolic modifications exemplified herein) are appropriate source biological descriptions of the genetic material (e.g., genes) required for a suitable host organism (e.g., *E. coli*) and its corresponding metabolic responses or desired metabolic pathways. However, given the availability of complete genome sequencing of a wide variety of organisms and a high level of skill in the field of genomics, those skilled in the art will be able to readily apply the teachings and guidance provided herein to substantially all other organisms. For example, the exemplary *E. coli* metabolic alterations described herein can be readily applied to other species by incorporating the same or similar coding nucleic acids from species different from those concerned. Generally, such genetic alterations include, for example, genetic alterations of species homologs, and specifically, substitutions of orthologous, paralogous, or non-paralogous genes.

[0034] Orthologous genes are related through vertical descent and result in one or more genes having substantially the same or similar functions in different organisms. For example, mouse epoxide hydrolase and human epoxide hydrolase can be considered orthologous due to their biological function of epoxide hydrolysis. Genes are considered orthologous when they share a sufficient amount of sequence similarity to indicate they are homologous or related through evolution from a common ancestor. Genes are also considered orthologous if they share a sufficient amount of three-dimensional structure but not necessarily a sufficient amount of sequence similarity to indicate they evolved from a common ancestor to the point that basic sequence similarity is unrecognizable. Orthologous genes can encode proteins with sequence similarity of approximately 25% to 100% amino acid sequence identity. Genes encoding proteins with less than 25% shared amino acid similarity also show similarity in their three-dimensional structures, and these genes can also be considered to have arisen through vertical descent. Members of the serine protease family of enzymes (including tissue-type plasminogen activator and elastase) are considered to have arisen through vertical descent from a common ancestor.

[0035] Orthologous genes include genes or their encoded gene products that diverge in structure or overall activity, for example, through evolution. For example, three genes and their corresponding products are considered orthologous when one species encodes gene products exhibiting two functions, and when such functions have been differentiated into different genes within a second species. For the production of biochemical products, those skilled in the art will understand that orthologous genes containing metabolic activities to be introduced or disrupted will be selected for the construction of said non-naturally occurring microorganisms. Examples of orthologous genes exhibiting their respective activities are those where different activities between two or more species or within a single species have been differentiated into different gene products. Specific examples are the separation of elastase proteolysis and plasminogen proteolysis (two types of serine protease activities) into different molecules as plasminogen activator and elastase. A second example is the separation of mycoplasma 5'-3' exonuclease and Drosophila DNA polymerase III activity. A DNA polymerase from the first species can be considered orthologous to one or both of the exonuclease and polymerase from the second species, and vice versa.

[0036] Conversely, paralogous proteins are homologous species that are related through, for example, replication and then evolutionary divergence and have similar or common, but not identical, functions. Paralogous proteins can originate, for example, from the same species or from different species. For example, microsomal epoxide hydrolases (epoxide hydrolase I) and soluble epoxide hydrolases (epoxide hydrolase II) can be considered paralogous because they represent two different classes of enzymes that co-evolved from a common ancestor, catalyze different reactions, and have different functions within the same species. Paralogous proteins are proteins from the same species that have significant sequence similarity to each other, indicating that they are homologous or related through co-evolution from a common ancestor. Groups of paralogous protein families include HipA homologs, luciferase genes, peptidases, and others.

[0037] Non-orthologous gene substitution is the substitution of a non-orthologous gene from one species for the function of a related gene in a different species. Substitution includes, for example, the ability to perform a function substantially the same or similar in the species of origin to the related function in the different species. While generally, non-orthologous gene substitution can be considered as relating to a known gene structure encoding the related function, genes with poor structural relevance but similar function, and their corresponding gene products, will still fall within the meaning of the term as used herein. Functional similarity requires, for example, at least some structural similarity at the active site or binding region of the non-orthologous gene product compared to the gene encoding the function being substituted. Therefore, orthologous genes include, for example, paralogous or unrelated genes.

[0038] Therefore, when identifying and constructing the non-naturally occurring microorganisms of the present invention with the ability to biosynthesize 1,3-BDO, those skilled in the art will apply the teachings and guidance provided herein to specific species and will understand that the identification of metabolic modifications may include the identification and inclusion or inactivation of orthologous genes. Those skilled in the art may also utilize these evolutionarily related genes if paralogous and / or non-orthologous gene substitutions encoding enzymes that catalyze similar or substantially similar metabolic reactions are present in the reference microorganism.

[0039] Orthologous, paralogous, and non-orthologous gene substitutions can be determined using methods well known to those skilled in the art. For example, examining the nucleic acid or amino acid sequences of two polypeptides will reveal sequence identity and similarity between the compared sequences. Based on this similarity, those skilled in the art can determine whether the similarity is high enough to indicate that the proteins are related through evolution from a common ancestor. Algorithms well known to those skilled in the art (such as Align, BLAST, Clustal W, and others) compare and determine the similarity or identity of the original sequences, and also determine the presence or size of gaps within the sequences, which can be assigned weights or scores. Such algorithms are also known in the art and are equally applicable to determining the similarity or identity of nucleotide sequences. The calculation of parameters for determining sufficient similarity to determine relevance is based on well-known methods for calculating statistical similarity or the probability of finding similar matches in random polypeptides and the degree of said match. If desired, computer comparisons of two or more sequences can also be visually optimized by those skilled in the art. Related gene products or proteins should have a high degree of similarity, for example, 25% to 100% sequence identity. Unrelated proteins can exhibit homology, which is essentially the same probability as expected if a database of a considerable size is scanned (approximately 5%). Sequences between 5% and 24% may or may not represent sufficient homology to conclude that the compared sequences are related. Given the size of the dataset, additional statistical analyses can be performed to determine the degree of this match and thus the relevance of these sequences.

[0040] Exemplary parameters for using, for example, the BLAST algorithm to determine the correlation of two or more sequences can be described below. In short, amino acid sequence alignment can be performed using BLASTP version 2.0.8 (January 5, 1999) with the following parameters: matrix: 0BLOSUM 62; vacancy open: 11; vacancy extension: 1; x_dropoff: 50; expectation: 10.0; sequence length: 3; filter: on. Nucleic acid sequence alignment can be performed using BLASTN version 2.0.6 (September 16, 1998) with the following parameters: match: 1; mismatch: -2; vacancy open: 5; vacancy extension: 2; x_dropoff: 50; expectation: 10.0; sequence length: 11; filter: off. Those skilled in the art will recognize that modifications to the above parameters will, for example, increase or decrease the stringency of the comparison and the determination of the correlation of two or more sequences.

[0041] In some embodiments, the present invention provides a non-naturally occurring microorganism comprising a microorganism having a 1,3-butanediol (1,3-BDO) pathway, the 1,3-butanediol (1,3-BDO) pathway having at least one exogenous nucleic acid encoding a 1,3-BDO pathway enzyme expressed in sufficient quantities to produce 1,3-BDO. The 1,3-BDO pathway includes enzymes selected from: 2-amino-4-ketovalerate (AKP) thiolytic enzyme, AKP dehydrogenase, 2-amino-4-hydroxyvalerate transaminase, 2-amino-4-hydroxyvalerate oxidoreductase (deamination), 2-oxo-4-hydroxyvalerate decarboxylase, 3-hydroxybutyraldehyde reductase, AKP transaminase, AKP oxidoreductase (deamination), 2,4-dioxovalerate decarboxylase, 3-oxobutyraldehyde reductase (ketone reduction), 3-oxobutyraldehyde reductase (aldehyde reduction), 4-hydroxy-2-butanone reductase, AKP decarboxylase, 4-amino... Butyl-2-ketotransaminase, 4-aminobutyric-2-keto oxidoreductase (deamination), 4-aminobutyric-2-keto ammonia-lyase, butenone hydratase, AKP ammonia-lyase, acetoacrylate decarboxylase, acetoacetyl-CoA reductase (CoA-dependent, aldehyde formation), acetoacetyl-CoA reductase (CoA-dependent, alcohol formation), acetoacetyl-CoA reductase (keto reduction), 3-hydroxybutyryl-CoA reductase (aldehyde formation), 3-hydroxybutyryl-CoA reductase (alcohol formation), 4-hydroxybutyryl-CoA dehydratase, and crotonic acid oxidase.

[0042] Any combination and any number of the enzymes described above can be introduced into a host microorganism to complete the 1,3-BDO pathway, such as... Figure 1-3 As shown. For example, the non-naturally occurring microorganism may include one, two, three, four, five, or all of the nucleic acids in the 1,3-BDO pathway, each nucleic acid encoding a 1,3-BDO pathway enzyme. These nucleic acids may include heterologous nucleic acids, additional copies of existing genes, and gene regulatory elements, as further explained below. The pathways of the non-naturally occurring microorganisms described in this invention are also suitably designed for cultivation in substantially anaerobic media.

[0043] In some embodiments, the non-naturally occurring microorganism possessing the 1,3-BDO pathway comprises a set of 1,3-BDO pathway enzymes. A set of 1,3-BDO pathway enzymes represents a group of enzymes capable of converting alanine, acetoacetyl-CoA, or 4-hydroxybutyryl-CoA into 1,3-BDO (e.g., ...). Figure 1-3 The enzyme (shown) exemplifies the conversion of alanine to 1,3-BDO (according to...). Figure 1A group of 1,3-BDO pathway enzymes includes (a) (1) 2-amino-4-ketovalerate (AKP) thiolylase; (2) AKP dehydrogenase; (3) 2-amino-4-hydroxyvalerate transaminase or oxidoreductase (deamination); (4) 2-oxo-4-hydroxyvalerate decarboxylase; and (5) 3-hydroxybutyraldehyde reductase; (b) (1) 2-amino-4-ketovalerate (AKP) thiolylase; (2) AKP transaminase or oxidoreductase (deamination); (3) 2, (c) (1) 2-amino-4-ketovalerate (AKP) thiolysis enzyme; (2) AKP transaminase or oxidoreductase (deamination); (3) 2,4-dioxovalerate decarboxylase; (4) 3-oxovalerate reductase (aldehyde reduction); and (5) 4-hydroxy-2-butanone reductase; (d) (1) 2-amino-4-ketovalerate (AKP) thiolysis enzyme. (1) 2-amino-4-ketovalerate (AKP) decarboxylase; (2) AKP decarboxylase; (3) 4-aminobutyraldehyde transaminase or oxidoreductase (deamination); (4) 3-oxobutyraldehyde reductase (ketone reduction); and (5) 3-hydroxybutyraldehyde reductase; (e)(1) 2-amino-4-ketovalerate (AKP) thiolyte; (2) AKP decarboxylase; (3) 4-aminobutyraldehyde transaminase or oxidoreductase (deamination); (4) 3-oxobutyraldehyde reductase (aldehyde reduction); and (5) 4-hydroxy-2-butyraldehyde reductase (ketone reduction); Ketone reductase; (f)(1) 2-amino-4-ketovalerate (AKP) thiolyase; (2) AKP decarboxylase; (3) 4-aminobutyric acid-2-ketoamine-lyase; (4) butenone hydratase; and (5) 4-hydroxy-2-butanone reductase; and (g)(1) 2-amino-4-ketovalerate (AKP) thiolyase; (2) AKP ammonia-lyase; (3) acetylacrylate decarboxylase; (4) butenone hydratase; and (5) 4-hydroxy-2-butanone reductase;

[0044] An example is the conversion of acetylacetyl-CoA to 1,3-BDO (according to...) Figure 2A group of 1,3-BDO pathway enzymes include (h) (1) acetyl-CoA reductase (CoA-dependent, aldehyde formation); (2) 3-oxobutyraldehyde reductase (ketone reduction); and (3) 3-hydroxybutyraldehyde reductase; (i) (1) acetyl-CoA reductase (CoA-dependent, alcohol formation) and (2) 4-hydroxy-2-butanone reductase; (J) (1) acetyl-CoA reductase (CoA-dependent, aldehyde formation); (2) 3-oxobutyraldehyde reductase (aldehyde reduction); and (3) 4-hydroxy-2-butanone reductase; (k) (1) acetyl-CoA reductase (ketone reduction) and (2) 3-hydroxybutyryl-CoA reductase (alcohol formation); and (l) (1) acetyl-CoA reductase (ketone reduction); (2) 3-hydroxybutyryl-CoA reductase (aldehyde formation); and (3) 3-hydroxybutyraldehyde reductase;

[0045] An example of converting 4-hydroxybutyryl-CoA to 1,3-BDO (according to...) Figure 3 A group of 1,3-BDO pathway enzymes include (m)(1) 4-hydroxybutyryl-CoA dehydratase; (2) crotonylase; and (3) 3-hydroxybutyryl-CoA reductase (alcohol formation); and (n)(1) 4-hydroxybutyryl-CoA dehydratase; (2) crotonylase; (3) 3-hydroxybutyryl-CoA reductase (aldehyde formation); and (4) 3-hydroxybutyraldehyde reductase.

[0046] The conversion of alanine to 1,3-BDO can be achieved through multiple pathways in approximately five enzymatic steps, such as... Figure 1 As shown. In the first step of all pathways (step A), alanine and acetyl-CoA are bound by 2-amino-4-ketovalerate thiolysolase (a highly selective enzyme). The product of this reaction, 2-amino-4-oxovalerate (AKP), can then be transaminated, reduced, decarboxylated, or deaminated, as... Figure 1 As shown.

[0047] In one pathway, AKP is converted to 2,4-dioxolaric acid, a 2-keto acid structurally similar to α-ketoglutarate, by transaminase or deaminase (step B). 2,4-Dioxolaric acid is then converted to 3-oxobutyraldehyde by 2-keto acid decarboxylase (step C). The ketone and acetaldehyde groups are reduced to their respective alcohols to obtain 1,3-butanediol. These reduction reactions can be carried out in either order to form the intermediate 3-hydroxybutyraldehyde (steps O and P) or 4-hydroxy,2-butanone (steps D and H).

[0048] In another pathway, the 4-oxo group of AKP is first reduced to a secondary alcohol by AKP dehydrogenase (step L). The product, 2-amino-4-hydroxyvalerate, is then converted to 2-oxo-4-hydroxyvalerate (step M). The resulting 2-keto acid is decarboxylated to 3-hydroxybutyraldehyde (step N). In the final step of this pathway, the aldehyde of 3-hydroxybutyraldehyde is reduced to a primary alcohol by 3-hydroxybutyraldehyde to form 1,3-butanediol (step P).

[0049] Another pathway involves decarboxylation of AKP via an amino acid decarboxylase (step E). The decarboxylation product 4-aminobutyr-2-one can be transaminated or oxidatively deaminated to 3-oxobutyraldehyde (step K) or deaminated to butenone (step F). When 3-oxobutyraldehyde is formed, two alcohol-forming reduction steps are used to form 1,3-butanediol, as described above (steps O and P or steps D and H). The deamination product butenone is then hydrolyzed to 4-hydroxy,2-butanone (step G), and 4-hydroxy,2-butanone is reduced to 1,3-butanediol by 4-hydroxy-2-butanone reductase (step H).

[0050] Another route still involves deamination of AKP to acetoacrylate (step I). Acetate is decarboxylated to butenone (step J), which is then converted to 1,3-butanediol by butenone hydratase (step G) and 4-hydroxy,2-butanone reductase (step H).

[0051] Based on the pathway for producing 1,3-BDO from alanine described above, in some embodiments, the non-naturally occurring microorganism possesses a set of 1,3-BDO pathway enzymes, including (1) 2-amino-4-ketovalerate (AKP) thiolyte; (2) AKP dehydrogenase; (3) 2-amino-4-hydroxyvalerate transaminase or oxidoreductase (deamination); (4) 2-oxo-4-hydroxyvalerate decarboxylase; and (5) 3-hydroxybutyraldehyde reductase. Any number of nucleic acids encoding these enzymes can be introduced into the host microorganism, including one, two, three, four, or all five nucleic acids encoding these enzymes. When one, two, three, or four exogenous nucleic acids are introduced, these nucleic acids can be any arrangement of the five nucleic acids.

[0052] In other embodiments, the non-naturally occurring microorganism possesses a set of 1,3-BDO pathway enzymes, including (1) 2-amino-4-ketovalerate (AKP) thiolyte; (2) AKP transaminase or oxidoreductase (deamination); (3) 2,4-dioxovalerate decarboxylase; (4) 3-oxobutyraldehyde reductase (ketone reduction); and (5) 3-hydroxybutyraldehyde reductase. Any number of nucleic acids encoding these enzymes can be introduced into the host microorganism, including one, two, three, four, or all five nucleic acids encoding these enzymes. When one, two, three, or four exogenous nucleic acids are introduced, these nucleic acids can be any arrangement of the five nucleic acids.

[0053] In other embodiments, the non-naturally occurring microorganism possesses a set of 1,3-BDO pathway enzymes, including (1) 2-amino-4-ketovalerate (AKP) thiolyte; (2) KP transaminase or oxidoreductase (deamination); (3) 2,4-dioxovalerate decarboxylase; (4) 3-oxobutyraldehyde reductase (aldehyde reduction); and (5) 4-hydroxy-2-butanone reductase. Any number of nucleic acids encoding these enzymes may be introduced into the host microorganism, including one, two, three, four, or all five nucleic acids encoding these enzymes. When one, two, three, or four exogenous nucleic acids are introduced, these nucleic acids can be any arrangement of the five nucleic acids.

[0054] In a further embodiment, the non-naturally occurring microorganism possesses a set of 1,3-BDO pathway enzymes, including (1) 2-amino-4-ketovalerate (AKP) thiolytic enzyme; (2) AKP decarboxylase; (3) 4-aminobutyrone-2-ketotransaminase or oxidoreductase (deamination); (4) 3-oxobutyraldehyde reductase (ketoreduction); and (5) 3-hydroxybutyraldehyde reductase. Any number of nucleic acids encoding these enzymes may be introduced into the host microorganism, including one, two, three, four, or all five nucleic acids encoding these enzymes. When one, two, three, or four exogenous nucleic acids are introduced, these nucleic acids may be any arrangement of the five nucleic acids.

[0055] In a further embodiment, the non-naturally occurring microorganism possesses a set of 1,3-BDO pathway enzymes, including (1) 2-amino-4-ketovalerate (AKP) thiolytic enzyme; (2) AKP decarboxylase; (3) 4-aminobutyrone-2-ketotransaminase or oxidoreductase (deamination); (4) 3-oxobutyraldehyde reductase (aldehyde reduction); and (5) 4-hydroxy-2-butanone reductase. Any number of nucleic acids encoding these enzymes may be introduced into the host microorganism, including one, two, three, four, or all five nucleic acids encoding these enzymes. When one, two, three, or four exogenous nucleic acids are introduced, these nucleic acids may be any arrangement of the five nucleic acids.

[0056] In a further embodiment, the non-naturally occurring microorganism possesses a set of 1,3-BDO pathway enzymes, including (1) 2-amino-4-ketovalerate (AKP) thiolyase; (2) AKP decarboxylase; (3) 4-aminobutyric acid-2-ketoamine lyase; (4) butenone hydratase; and (5) 4-hydroxy-2-butanone reductase. Any number of nucleic acids encoding these enzymes may be introduced into the host microorganism, including one, two, three, four, or all five nucleic acids encoding these enzymes. When one, two, three, or four exogenous nucleic acids are introduced, these nucleic acids may be any arrangement of the five nucleic acids.

[0057] In a further embodiment, the non-naturally occurring microorganism possesses a set of 1,3-BDO pathway enzymes, including (1) 2-amino-4-ketovalerate (AKP) thiolyase; (2) AKP ammonia-lyase; (3) acetylacrylate decarboxylase; (4) butenone hydratase; and (5) 4-hydroxy-2-butanone reductase. Any number of nucleic acids encoding these enzymes may be introduced into the host microorganism, including one, two, three, four, or all five nucleic acids encoding these enzymes. When one, two, three, or four exogenous nucleic acids are introduced, these nucleic acids may be any arrangement of the five nucleic acids.

[0058] Figure 2 Several pathways for the production of 1,3-butanediol from acetoacetyl-CoA are outlined. One pathway via steps A, B, and C utilizes (i) a CoA-dependent, aldehyde-forming acetoacetyl-CoA reductase to convert acetoacetyl-CoA to 3-oxobutyraldehyde (O2-CoA). Figure 2 Step A), (ii) 3-O-butyraldehyde reductase, to reduce 3-oxobutyraldehyde to 3-hydroxybutyraldehyde ( Figure 2 Steps B) and (iii) finally involve 3-hydroxybutyraldehyde reductase to form 1,3-butanediol ( Figure 2 Step C).

[0059] Alternatively, acetoacetyl-CoA can be reduced by acetoacetyl-CoA reductase, which forms the aldehyde, to form 4-hydroxy,2-butanone. Figure 2 Step D). The formation of 4-hydroxy,2-butanone can also be achieved by reducing 3-oxobutanol ( ) with the aldehyde-reducing 3-oxobutanol reductase. Figure 2 Step E). Ultimately, 4-hydroxy,2-butanone can be reduced by 4-hydroxy-2-butanone reductase to form 1,3-BDO (… Figure 2 Step F).

[0060] Another 1,3-BDO formation pathway still relies on the reduction of acetyl-CoA to 3-hydroxybutyryl-CoA via the ketone-reducing acetyl-CoA reductase. Figure 2 Step G). This enzyme reduces the ketone function in acetyl-CoA to a hydroxyl group. 3-Hydroxybutyryl-CoA can be reduced by a bifunctional alcohol-forming 3-hydroxybutyryl-CoA reductase to form 1,3-butanediol (…). Figure 2 (Step I). Alternatively, it can first be reduced to 3-hydroxybutyraldehyde by the 3-hydroxybutyryl-CoA reductase formed by the aldehyde (Step H), and 3-hydroxybutyraldehyde can then be reduced as shown in Step C.

[0061] In some embodiments, based on the pathway described above for the production of 1,3-BDO from acetoacetyl-CoA, the non-naturally occurring microorganism possesses a set of 1,3-BDO pathway enzymes, including (1) acetoacetyl-CoA reductase (CoA-dependent, aldehyde-forming); (2) 3-oxobutyraldehyde reductase (ketone reduction); and (3) 3-hydroxybutyraldehyde reductase. Any number of nucleic acids encoding these enzymes may be introduced into the host microorganism, including one, two, or all three nucleic acids encoding these enzymes. When one or two exogenous nucleic acids are introduced, these nucleic acids can be any arrangement of the three nucleic acids.

[0062] In other embodiments, the non-naturally occurring microorganism possesses a set of 1,3-BDO pathway enzymes, including (1) acetyl-CoA reductase (CoA-dependent, alcohol-forming) and (2) 4-hydroxy-2-butanone reductase. Any number of nucleic acids encoding these enzymes may be introduced into the host microorganism, including one or both of the nucleic acids encoding these enzymes. When a foreign nucleic acid is introduced, such a nucleic acid may be either of the two nucleic acids.

[0063] In a further embodiment, the non-naturally occurring microorganism possesses a set of 1,3-BDO pathway enzymes, including (1) acetylacetyl-CoA reductase (CoA-dependent, aldehyde-forming); (2) 3-oxobutyraldehyde reductase (aldehyde reduction); and (3) 4-hydroxy-2-butanone reductase. Any number of nucleic acids encoding these enzymes may be introduced into the host microorganism, including one, two, or all three nucleic acids encoding these enzymes. When one or two exogenous nucleic acids are introduced, these nucleic acids may be any arrangement of the three nucleic acids.

[0064] In a further embodiment, the non-naturally occurring microorganism possesses a set of 1,3-BDO pathway enzymes, including (1) acetyl-CoA reductase (ketone reduction) and (2) 3-hydroxybutyryl-CoA reductase (alcohol formation). Any number of nucleic acids encoding these enzymes may be introduced into the host microorganism, including one or both of the nucleic acids encoding these enzymes. When a foreign nucleic acid is introduced, such a nucleic acid may be either of the two nucleic acids.

[0065] In a further embodiment, the non-naturally occurring microorganism possesses a set of 1,3-BDO pathway enzymes, including (1) acetyl-CoA reductase (ketone reduction); (2) 3-hydroxybutyryl-CoA reductase (aldehyde formation); and (3) 3-hydroxybutyraldehyde reductase. Any number of nucleic acids encoding these enzymes may be introduced into the host microorganism, including one, two, or all three nucleic acids encoding these enzymes. When one or two exogenous nucleic acids are introduced, these nucleic acids may be any arrangement of the three nucleic acids.

[0066] 4-Hydroxybutyryl-CoA is an important starting metabolite from which many industrially useful compounds can be produced. Although 4-hydroxybutyryl-CoA is not a very common central metabolite, methods for designing strains to synthesize 4-hydroxybutyryl-CoA have been described in Burk et al. (US 20090075351). An exemplary method involves the synthesis of 4-hydroxybutyryl-CoA from succinyl-CoA by employing genes encoding succinyl butyrate dehydrogenase (CoA-dependent), 4-hydroxybutyrate dehydrogenase, 4-hydroxybutyrate kinase, and phosphobutyryltransferase activities.

[0067] The first step in the pathway involves the dehydration of 4-hydroxybutyryl-CoA. Figure 3Step A), followed by hydration of crotonyl-CoA to form 3-hydroxybutyryl-CoA (Step B). 3-hydroxybutyryl-CoA then undergoes two reduction steps performed by either of the two enzymes (Steps C and D) or a single bifunctional enzyme (Step E) to form 1,3-butanediol.

[0068] Therefore, in some embodiments, the non-naturally occurring microorganism possesses a set of 1,3-BDO pathway enzymes, including (1) 4-hydroxybutyryl-CoA dehydratase; (2) crotonylase; and (3) 3-hydroxybutyryl-CoA reductase (alcohol formation). Any number of nucleic acids encoding these enzymes can be introduced into the host microorganism, including one, two, or all three nucleic acids encoding these enzymes. When one or two exogenous nucleic acids are introduced, these nucleic acids can be any arrangement of the three nucleic acids.

[0069] In other embodiments, the non-naturally occurring microorganism possesses a set of 1,3-BDO pathway enzymes, including (1) 4-hydroxybutyryl-CoA dehydratase; (2) crotonylase; (3) 3-hydroxybutyryl-CoA reductase (aldehyde formation); and (4) 3-hydroxybutyral reductase. Any number of nucleic acids encoding these enzymes may be introduced into the host microorganism, including one, two, three, or all four nucleic acids encoding these enzymes. When one, two, or three exogenous nucleic acids are introduced, these nucleic acids may be any arrangement of the four nucleic acids.

[0070] In another embodiment, the present invention provides a non-naturally occurring microorganism having a 1,3-BDO pathway, wherein the non-naturally occurring microorganism contains at least one exogenous nucleic acid encoding an enzyme or protein that converts a substrate into a product selected from: alanine to 2-amino-4-oxovalerate, 2-amino-4-oxovalerate to 2-amino-4-hydroxyvalerate, 2-amino-4-hydroxyvalerate to 2-oxo-4-hydroxyvalerate, 2-oxo-4-hydroxyvalerate to 3-hydroxybutyraldehyde, and 3-hydroxybutyraldehyde to 1,3-BDO.

[0071] In another embodiment, the present invention provides a non-naturally occurring microorganism having a 1,3-BDO pathway, wherein the non-naturally occurring microorganism contains at least one exogenous nucleic acid encoding an enzyme or protein that converts a substrate into a product selected from: alanine to 2-amino-4-oxovalerate, 2-amino-4-oxovalerate to 2,4-dioxovalerate, 2,4-dioxovalerate to 3-oxobutanal, 3-oxobutanal to 4-hydroxy-2-butanone, and 4-hydroxy-2-butanone to 1,3-BDO.

[0072] In another embodiment, the present invention provides a non-naturally occurring microorganism having a 1,3-BDO pathway, wherein the non-naturally occurring microorganism contains at least one exogenous nucleic acid encoding an enzyme or protein that converts a substrate into a product selected from: alanine to 2-amino-4-oxovalerate, 2-amino-4-oxovalerate to 4-aminobutyr-2-one, 4-aminobutyr-2-one to 3-oxobutyraldehyde, 3-oxobutyraldehyde to 3-hydroxybutyraldehyde, and 3-hydroxybutyraldehyde to 1,3-BDO.

[0073] In another embodiment, the present invention provides a non-naturally occurring microorganism having a 1,3-BDO pathway, wherein the non-naturally occurring microorganism contains at least one exogenous nucleic acid encoding an enzyme or protein that converts a substrate into a product selected from: alanine to 2-amino-4-oxovalerate, 2-amino-4-oxovalerate to 4-aminobut-2-one, 4-aminobut-2-one to 3-oxobutanal, 3-oxobutanal to 4-hydroxy-2-butanone, and 4-hydroxy-2-butanone to 1,3-BDO.

[0074] In another embodiment, the present invention provides a non-naturally occurring microorganism having a 1,3-BDO pathway, wherein the non-naturally occurring microorganism contains at least one exogenous nucleic acid encoding an enzyme or protein that converts a substrate into a product selected from: alanine to 2-amino-4-oxovalerate, 2-amino-4-oxovalerate to 4-aminobut-2-one, 4-aminobut-2-one to butenone, butenone to 4-hydroxy-2-butanone, and 4-hydroxy-2-butanone to 1,3-BDO.

[0075] In another embodiment, the present invention provides a non-naturally occurring microorganism having a 1,3-BDO pathway, wherein the non-naturally occurring microorganism comprises at least one exogenous nucleic acid encoding an enzyme or protein that converts a substrate into a product selected from: alanine to 2-amino-4-oxovalerate, 2-amino-4-oxovalerate to acetoacrylate, acetoacrylate to butenone, butenone to 4-hydroxy-2-butanone, and 4-hydroxy-2-butanone to 1,3-BDO.

[0076] Therefore, the present invention provides a non-naturally occurring microorganism comprising at least one exogenous nucleic acid encoding an enzyme or protein, wherein the enzyme or protein transforms the substrate and the product of the 1,3-BDO pathway that converts alanine to 1,3-BDO, by means of... Figure 1 The pathway shown is an example.

[0077] In another embodiment, the present invention provides a non-naturally occurring microorganism having a 1,3-BDO pathway, wherein the non-naturally occurring microorganism contains at least one exogenous nucleic acid encoding an enzyme or protein that converts a substrate into a product selected from: acetoacetyl-CoA to 4-hydroxy-2-butanone and 4-hydroxy-2-butanone to 1,3-BDO.

[0078] In another embodiment, the present invention provides a non-naturally occurring microorganism having a 1,3-BDO pathway, wherein the non-naturally occurring microorganism contains at least one exogenous nucleic acid encoding an enzyme or protein that converts a substrate into a product selected from: acetoacetyl-CoA to 3-oxobutyraldehyde, 3-oxobutyraldehyde to 4-hydroxy-2-butanone, and 4-hydroxy-2-butanone to 1,3-BDO.

[0079] In another embodiment, the present invention provides a non-naturally occurring microorganism having a 1,3-BDO pathway, wherein the non-naturally occurring microorganism contains at least one exogenous nucleic acid encoding an enzyme or protein that converts a substrate into a product selected from: acetoacetyl-CoA to 3-oxobutyraldehyde, 3-oxobutyraldehyde to 3-hydroxybutyraldehyde, and 3-hydroxybutyraldehyde to 1,3-BDO.

[0080] In another embodiment, the present invention provides a non-naturally occurring microorganism having a 1,3-BDO pathway, wherein the non-naturally occurring microorganism contains at least one exogenous nucleic acid encoding an enzyme or protein that converts a substrate into a product selected from: acetoacetyl-CoA to 3-hydroxybutyryl-CoA, 3-hydroxybutyryl-CoA to 3-hydroxybutyraldehyde, and 3-hydroxybutyraldehyde to 1,3-BDO.

[0081] In another embodiment, the present invention provides a non-naturally occurring microorganism having a 1,3-BDO pathway, wherein the non-naturally occurring microorganism contains at least one exogenous nucleic acid encoding an enzyme or protein that converts a substrate into a product selected from: acetoacetyl-CoA to 3-hydroxybutyryl-CoA and 3-hydroxybutyryl-CoA to 1,3-BDO.

[0082] Therefore, the present invention provides a non-naturally occurring microorganism comprising at least one exogenous nucleic acid encoding an enzyme or protein, wherein the enzyme or protein transforms the substrate and the product of the 1,3-BDO pathway that converts acetyl-CoA to 1,3-BDO, by means of... Figure 2 The pathway shown is an example.

[0083] In another embodiment, the present invention provides a non-naturally occurring microorganism having a 1,3-BDO pathway, wherein the non-naturally occurring microorganism comprises at least one exogenous nucleic acid encoding an enzyme or protein that converts a substrate into a product selected from: 4-hydroxybutyryl-CoA to crotonyl-CoA, crotonyl-CoA to 3-hydroxybutyryl-CoA, 3-hydroxybutyryl-CoA to 3-hydroxybutyraldehyde, and 3-hydroxybutyraldehyde to 1,3-BDO.

[0084] In another embodiment, the present invention provides a non-naturally occurring microorganism having a 1,3-BDO pathway, wherein the non-naturally occurring microorganism contains at least one exogenous nucleic acid encoding an enzyme or protein that converts a substrate into a product selected from: 4-hydroxybutyryl-CoA to crotonyl-CoA, crotonyl-CoA to 3-hydroxybutyryl-CoA, and 3-hydroxybutyryl-CoA to 1,3-BDO.

[0085] Therefore, the present invention provides a non-naturally occurring microorganism comprising at least one exogenous nucleic acid encoding an enzyme or protein, wherein the enzyme or protein transforms the substrate and the product of the 1,3-BDO pathway that converts 4-hydroxybutyryl-CoA to 1,3-BDO, by means of... Figure 3 The pathway shown is an example.

[0086] Successfully designing any of these pathways requires identifying a suitable set of enzymes with sufficient activity and specificity, cloning their corresponding genes into the production host, optimizing fermentation conditions, and determining the product formation after fermentation.

[0087] To design a production host for producing any of the aforementioned products, one or more exogenous DNA sequences can be expressed in a microorganism. Furthermore, the microorganism can cause the loss of function of endogenous genes. These modifications will enable the production of 1,3-BDO using renewable raw materials.

[0088] Below, we describe many genes that encode the biochemical properties of enzymes, which in turn affect... Figure 1 , 2 Each of the steps shown in 3 is catalyzed. While we have described this method for E. coli, those skilled in the art can apply these teachings to virtually any other organism. Specifically, genes that are natural to E. coli are listed, in addition to genes in other organisms that, when properly cloned and expressed, can be applied to catalyze appropriate transformations.

[0089] This invention is generally described with reference to metabolic reactions, reactants, or products thereof, or specifically with reference to one or more nucleic acids or genes encoding enzymes (which are involved in or catalyze the referred metabolic reaction, reactant, or product) or proteins (which are involved in the referred metabolic reaction, reactant, or product). Unless expressly stated herein, those skilled in the art will understand that reference to a reaction also constitutes a reference to the reactant and the product of said reaction. Similarly, unless expressly stated herein, reference to a reactant or product is also a reference to the reaction, and reference to any of these metabolic components is also a reference to the gene encoding the catalyzing enzyme or the protein involved in the referred reaction. Likewise, taking into account the well-known fields of metabolic biochemistry, enzymology, and genomics, references herein to genes or encoding nucleic acids also constitute a reference to the corresponding encoded enzyme and the reaction it catalyzes or the protein involved in the reaction, as well as the reactant and the product of the reaction.

[0090] Figure 1-3 All the transformations described are categorized into eight general transformation classes as shown in Table 1. The following describes several biochemical characteristics of genes within each class. Specifically, a list is provided of genes that, when properly cloned and expressed, can be used for catalysis. Figure 1-3 Appropriately transformed genes. Further information for use is provided in Tables 35-37 below. Figure 1-3 An exemplary gene for each of the steps.

[0091] Table 1 shows the types of enzymes available for converting common, ubiquitous central metabolic intermediates into 1,3-butanediol. The first three digits of each label correspond to the first three digits of the enzyme committee number, which indicates the general conversion type independent of substrate specificity.

[0092] Table 1

[0093] mark Function 1.1.1.a Oxidoreductases (from ketones to hydroxyl groups or from aldehydes to alcohols) 1.1.1.c Oxidoreductase (2 steps, acyl-CoA to alcohol) 1.2.1.b Oxidoreductase (acyl-CoA to aldehyde) 1.4.1.a Oxidoreductase (deamination) 2.3.1.b transacylase 2.6.1.a transaminase 4.1.1.a Carboxyl lyase 4.2.1.a hydrolytic enzymes 4.3.1.a ammonia-lyase

[0094] Figure 1 , 2 Many of the transformations in section 3 belong to the class of oxidoreductases, which reduce aldehydes to alcohols. For example, Figure 1 Steps D and P, catalyzed by 3-oxobutyraldehyde reductase (aldehyde reduction) and 3-hydroxybutyraldehyde reductase respectively, belong to this class. Similarly, Figure 2 Steps C and E, catalyzed by 3-hydroxybutyraldehyde reductase and 3-oxobutyraldehyde reductase (aldehyde reduction), are also oxidoreductases that convert aldehydes into alcohols.

[0095] Figure 3 The pathways involved involve oxidoreductases, such as 3-hydroxybutyraldehyde reductase in step D.

[0096] Exemplary genes encoding enzymes that catalyze the conversion of aldehydes to alcohols (i.e., alcohol dehydrogenases or equivalence aldehyde reductases) include alrA (Tani et al., Appl. Environ. Microbiol., 66:5231-5235 (2000)) encoding medium-chain alcohol dehydrogenases (C2-C14), ADH2 from Saccharomyces cerevisiae (Atsumi et al., Nature, 451:86-89 (2008)), yqhD from Escherichia coli which has a preference for molecules longer than C3 (Sulzenbacher et al., J. of Molecular Biology, 342:489-502 (2004)), and bdh I and bdh II from Clostridium acetone-butanol which convert butyraldehyde to butanol (Walter et al., J. of Bacteriology, 174:7149-7158 (1992)). The yqhD gene product utilizes NADPH as a cofactor to catalyze the reduction of acetaldehyde, malondialdehyde, propionaldehyde, butyraldehyde, and acrolein (Perez et al., J. Biol. Chem., 283:7346-7353 (2008)). The adhA gene product from *C. motile fermentum* has been shown to be active against many aldehydes, including formaldehyde, acetaldehyde, propionaldehyde, butyraldehyde, and acrolein (Kinoshita et al., Appl. Microbiol. Biotechnol, 22:249-254 (1985)). Other candidate aldehyde reductases are encoded by bdh from *C. saccharoperbutylacetonicum* and Cbei_1722, Cbei_2181, and Cbei_2421 from *C. beijerinckii*.

[0097] Data associated with the sequence of each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 2 below.

[0098] Table 2

[0099] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> alrA BAB12273.1 9967138 Acinetobacter strain M-1 ADH2 NP_014032.1 6323961 brewing yeast yqhD NP_41748401 16130909 E. coli bdh I NP_349892.1 15896543 Clostridium acetonebutanol bdh II NP_349891.1 15896542 Clostridium acetonebutanol adhA YP_162971.1 56552132 Motile fermentation monoclonal bacteria bdh BAF45463.1 124221917 Clostridium glycoacetate polybutanol Cbei_1722 YP_001308850 150016596 Clostridium beyerridis Cbei_2181 YP_001309304 150017050 Clostridium beyerridis Cbei_2421 YP_001309535 150017281 Clostridium beyerridis

[0100] The enzyme exhibiting 3-hydroxybutyraldehyde reductase activity (EC 1.1.1.61) also belongs to this category. This type of enzyme has been characterized in *Ralstonia eutropha* (Bravo et al., J. Forensic Sci., 49:379-387 (2004)), *Clostridium kluferum* (Wolff et al., Protein Expr. Purif., 6:206-212 (1995)), and *Arabidopsis thaliana* (Breitkreuz et al., J. Biol. Chem., 278:41552-41556 (2003)). Another gene is the alcohol dehydrogenase adhI from *Bacillus thermoglucosidase* (Jeon et al., J. Biotechnol., 135:127-133 (2008)). Sequence data associated with each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 3 below.

[0101] Table 3

[0102] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> 4hbd YP_726053.1 113867564 Optimal Rawlstonella H16 4hbd L21902.1 146348486 Clostridium clavatum DSM 555 4hbd Q94B07 75249805 Arabidopsis adhI AAR91477.1 40795502 Thermoglucosidase Bacillus thuringiensis M10EXG

[0103] Another exemplary enzyme is 3-hydroxyisobutyrate dehydrogenase, which catalyzes the reversible oxidation of 3-hydroxyisobutyrate to methylmalonyl hemialdehyde. This enzyme is involved in the degradation of valine, leucine, and isoleucine and has been found in bacteria, eukaryotes, and mammals. The enzyme encoded by P84067 from *Thermophilus HB8* has been structurally characterized (Lokanath et al., J. Mol. Biol., 352:905-917 (2005)). The reversibility of human 3-hydroxyisobutyrate dehydrogenase has been demonstrated using isotopically labeled substrates (Manning et al., Biochem J., 231:481-484 (1985)). Other genes encoding this enzyme include 3hidh in Homo sapiens (Hawes et al., Methods Enzymol, 324:218-228 (2000)) and rabbits (Hawes et al., ibid.; Chowdhury et al., Biosci. Biotechnol Biochem., 60:2043-2047 (1996)), mmsB in Pseudomonas aeruginosa and Pseudomonas putridae (Liao et al., US Patent 20050221466), and dhat in Pseudomonas putridae (Aberhart et al., J. Chem. Soc., 6:1404-1406 (1979); Chowdhury et al., ibid.; Chowdhury et al., Biotechnol Biochem., 67:438-441 (2003)). Data associated with the sequence of each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 4 below.

[0104] Table 4

[0105] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> P84067 P84067 75345323 Thermophilic bacteria 3hidh P31937.2 12643395 Homo sapiens 3hidh P32185.1 416872 rabbit mmsB P28811.1 127211 Pseudomonas aeruginosa mmsB NP_746775.1 26991350 Pseudomonas putida dhat Q59477.1 2842618 Pseudomonas putida

[0106] The conversion of ketones to their corresponding hydroxyl groups by oxidoreductases is also a synthetic step in the disclosed pathway. Clearly, Figure 1 The reactions L, O, and H catalyzed by AKP dehydrogenase, 3-oxobutyraldehyde reductase (ketone reduction), and 4-hydroxy-2-butanone reductase, respectively, represent this category of transformations. The latter two transformations occur in... Figure 2 This also occurs in steps B and F respectively. On similar routes, in... Figure 2 In step G, acetoacetyl-CoA reductase reduces acetoacetyl-CoA to 3-hydroxybutyryl-CoA.

[0107] 2-Amino-4-hydroxyvalerate is obtained by dehydrogenase reduction of the 4-oxo group of 2-amino-4-oxovalerate (AKP). Figure 1 Step L). This reaction is very similar to the NAD(P)H-dependent reduction from aspartic semialdehyde to homoserine catalyzed by homoserine dehydrogenase (EC 1.1.13). In many organisms (including E. coli), homoserine dehydrogenase is a bifunctional enzyme that also catalyzes the ATP-dependent conversion of aspartic acid to aspartic acyl-4-phosphate (Starnes et al., Biochemistry, 11:677-687 (1973)). The functional domains are independently catalyzed and connected by a linker region (Sibilli et al., J. Biol. Chem., 256:10228-10230 (1981)) and both domains are allosterically repressed by threonine. The homoserine dehydrogenase domain of the *E. coli* enzyme encoded by thrA was isolated from, characterized, and found to exhibit high catalytic activity and reduced inhibitory activity due to threonine (James et al., *Biochemistry*, 41:3720-3725 (2002)). This can be applied to other bifunctional threonine kinases, including, for example, *Lactobacillus plantarum* (Cahyanto et al., *Microbiology*, 152:205-112 (2006)) and *Arabidopsis thaliana* hom1. Monofunctional homoserine dehydrogenases encoded by *Saccharomyces cerevisiae* hom6 (Jacques et al., *Biochem. Biophys. Acta*, 1544:28-41 (2001)) and *Lactobacillus plantarum* hom2 (Cahyanto et al., ibid.) have been functionally expressed and characterized in *E. coli*. Sequence-associated data for each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 5 below.

[0108] Table 5

[0109] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> thrA AAC73113.1 1786183 Escherichia coli K12 akthr2 O81852 75100442 Arabidopsis hom6 CAA89671 1015880 brewing yeast hom1 CAD64819 28271914 Lactobacillus plantarum hom2 CAD63186 28270285 Lactobacillus plantarum

[0110] Acetylacetyl-CoA reductase (AAC) catalyzes the reduction of acetylacetyl-CoA to 3-hydroxybutyryl-CoA. Figure 2 Step G) is involved in the pathway of acetyl-CoA fermentation to butyrate in multiple Clostridium species and has been studied in detail (Jones et al., Microbiol. Rev., 50:484-524 (1986)). An enzyme encoded by hbd from Clostridium acetobutyrate has been cloned and functionally expressed in Escherichia coli (Youngleson et al., J. Bacteriol., 171:6800-6807 (1989)). In addition, subunits of two fatty acid oxidation complexes encoded by fadB and fadJ in Escherichia coli function as 3-hydroxyacyl-CoA dehydrogenases (Binstock et al., Methods Enzymol, 71C:403-411 (1981)). Other genes shown to reduce acetoacetyl-CoA to 3-hydroxybutyryl-CoA include phbB from *Azologia procumbens* (Ploux et al., Eur. J. Biochem., 174:177-182 (1988)) and phaB from *Rhodotorula gracilis* (Alber et al., Mol. Microbiol., 61:297-309 (2006)). The former gene is NADPH-dependent, its nucleotide sequence has been determined (Peoples et al., Mol. Microbiol. 3:349-357 (1989)), and it has been expressed in *Escherichia coli*. Substrate specificity studies of this gene have concluded that, in addition to acetoacetyl-CoA, it can accept 3-oxopropionyl-CoA as a substrate (Ploux et al., ibid.). Other genes include Hbd1 (C-terminal domain) and Hbd2 (N-terminal domain) in Clostridium kluferum (Hillmer and Gottschalk, Biochim. Biophys. Acta 3334:12-23 (1974)) and HSD17B10 in bovine taurus (Wakil et al., J. Biol. Chem., 207:631-638 (1954)). Sequence data associated with each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 6 below.

[0111] Table 6

[0112] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> fadB P21177.2 119811 E. coli fadJ P77399.1 3334437 E. coli Hbd2 EDK34807.1 146348271 Clostridium clavatum Hbd1 EDK32512.1 146345976 Clostridium clavatum hbd P52041.2 Clostridium acetonebutanol HSD17B10 O02691.3 3183024 ox phbB P23238.1 130017 *Agromycetes pubescens* phaB YP_353825.1 77464321 Rhodopsyllosis

[0113] Many similar enzymes have been found in other species of Clostridium and Agrobacterium tumefaciens (Berg et al., Archaea. Science, 318:1782-1786 (2007)), as shown in Table 7.

[0114] Table 7

[0115] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> Hbd NP_349314.1 NP_349314.1 Clostridium acetonebutanol Hbd AAM14586.1 AAM14586.1 Clostridium beyerridis Msed_1423 YP_001191505 YP_001191505 Diligent golden cocci Msed_0399 YP_001190500 YP_001190500 Diligent golden cocci Msed_0389 YP_001190490 YP_001190490 Diligent golden cocci Msed_1993 YP_001192057 YP_001192057 Diligent golden cocci

[0116] Exemplary alcohol dehydrogenases that convert ketones to hydroxyl groups are secondary alcohol dehydrogenases that have been shown to convert acetone to isopropanol in *Clostridium beyerei* (Ismaiel et al., J. Bacteriol, 175:5097-5105 (1993)) and *Thermophyton floccosum* (Lamed et al., Biochem. J., 195:183-190 (1981)). The adhA gene product from *Thermophyton floccosum*, which exhibits maximal activity against 2-pentanol and acetone aldehyde, shows very broad specificity, including isopropanol and acetone (Vander et al., Eur. J. Biochem., 268:3062-3068 (2001)). Another secondary alcohol dehydrogenase active to isopropanol and acetone is encoded by the product of the adh-A gene from Rhodococcus aureus (Edegger et al., Chem. Commun. (Camb), 2402-2404 (2006); Kosjek et al., Biotechnol. Bioeng., 86:55-62 (2004)). These genes, along with other genes, are listed in Table 8 below.

[0117] Table 8

[0118] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> adh AAA23199.2 60592974 Clostridium beyerii NRRL B593 adh P14941.1 113443 Brucella thermophila HTD4 adhA AAC25556 3288810 Thermococcus adh-A CAD36475 21615553 Rhodococcus

[0119] Alternatively, several exemplary alcohol dehydrogenases exist that convert ketones to hydroxyl functional groups. Two such enzymes from *Escherichia coli* are encoded by malate dehydrogenase (mdh) and lactate dehydrogenase (ldhA). Furthermore, lactate dehydrogenases from *R. eutrophus* have been shown to exhibit high activity against substrates of varying chain lengths, such as lactate, 2-oxobutyric acid, 2-oxovalerate, and 2-oxoglutarate (Steinbuchel et al., *Eur. J. Biochem.*, 130:329-334 (1983)). The conversion of oxofunctional to hydroxyl groups can also be catalyzed by 2-acetyl-3-butanediol reductase, an enzyme reportedly found in rats and placenta (Suda et al., Arch. Biochem. Biophys., 176:610-620 (1976); Suda et al., Biochem. Biophys. Res. Commun., 77:586-591 (1977)). All these enzymes can provide 3-oxobutyraldehyde reductase and 4-hydroxy-2-butanone reductase. Other enzymes used in these steps are the cloned and characterized mitochondrial 3-hydroxybutyrate dehydrogenase (bdh) from the human heart (Marks et al., J. Biol. Chem. 267:15459-15463 (1992)). This enzyme is a dehydrogenase that acts on 3-hydroxy acids. Sequence-related data for each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 9 below.

[0120] Table 9

[0121] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> mdh AAC76268.1 1789632 E. coli ldhA NP_415898.1 16129341 E. coli ldh YP_725182.1 113866693 Eutrophic Rollstonella bdh AAA58352.1 177198 Homo sapiens

[0122] Many organisms can catalyze the reduction of 4-hydroxy-2-butanone to 1,3-butanediol, including those belonging to the genera Bacillus, Brevibacterium, Candida, and Klebsiella, as described by Matsuyama et al. (1995).

[0123] Figure 2 and 3 Several transformations in this process rely on a two-step reduction from acyl-CoA to the corresponding alcohol. For example, Figure 2 Steps D and I (involving acetyl-CoA reductase (CoA-dependent, alcohol formation) and 3-hydroxybutyryl-CoA reductase (alcohol formation)) and Figure 3 Step E (involving 3-hydroxybutyryl-CoA reductase (alcohol formation)) illustrates this type of transformation.

[0124] Exemplary two-step oxidoreductases that convert acyl-CoA to alcohols include those that convert substrates such as acetyl-CoA to ethanol (e.g., adhE from *E. coli* (Kessler et al., FEBS. Lett., 281:59-63 (1991)) and butyryl-CoA to butanol (e.g., adhE2 from *Clostridium acetobutanol* (Fontaine et al., J. Bacteriol., 184:821-830 (2002)). Besides converting acetyl-CoA to alcohols... CoA is reduced to ethanol, and an enzyme encoded by adhE in Leuconostoc mesenteroides has been shown to oxidize the branched compound isobutyraldehyde to isobutyryl-CoA (Kazahaya et al., J. Gen. Appl. Microbiol., 18:43-55 (1972); Koo et al., Biotechnol. Lett., 27:505-510 (2005)). Sequence-related data for each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 10 below.

[0125] Table 10

[0126] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> adhE NP_415757.1 16129202 E. coli adhE2 AAK09379.1 12958626 Clostridium acetonebutanol adhE AAV66076.1 55818563 Leuconostoc mesentery

[0127] Another exemplary enzyme can convert malonyl-CoA to 3-HP. An NADPH-dependent enzyme with this activity was characterized in *Flexobacter orangeii*, in which it participates in the 3-hydroxypropionic acid cycle (Hugler et al., J. Bacteriol., 184:2404-2410 (2002); Strauss et al., Eur. J. Biochem., 215:633-643 (1993)). This enzyme (in a quantity of 300 kDa) exhibits high substrate specificity and shows little sequence similarity to other known redox enzymes (Hugler et al., ibid.). No enzyme has been shown to catalyze this specific reaction in other organisms; however, there is biological evidence that other organisms may possess similar pathways (Klatt et al., Environ. Microbiol., 9:2067-2078 (2007)). Enzymes in other organisms (including Roseiflexus castenholzii, erythrocyte NAP1, and marine gamma-proteobacterium HTCC2080) can be inferred from sequence similarity.

[0128] Table 11

[0129] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> mcr AAS20429.1 42561982 Orange-green flexorum Rcas_2929 YP_001433009.1 156742880 Roseiflexus castenholzii NAP1_02720 ZP_01039179.1 85708113 Red bacterium NAP1 MGP2080_00535 ZP_01626393.1 119504313 Marine gamma-proteobacterium HTCC2080

[0130] Longer-chain acyl-CoA molecules can be reduced by enzymes, such as jojoba (Symonds, California) FAR, which encodes alcohol-forming fatty acyl-CoA reductase. Overexpression of FAR in E. coli leads to increased FAR activity and accumulation of fatty alcohols (Metz et al., Plant Physiology, 122:635-644 (2000)) (FAR, AAD38039.1, 5020215, Symonds, California).

[0131] The pathway disclosed in this article involves many oxidoreductase conversions that transform acyl-CoA into aldehydes. Specifically, Figure 2 Steps A and H, catalyzed by acetyl-CoA reductase (aldehyde formation) and 3-hydroxybutyryl-CoA reductase (aldehyde formation), and from... Figure 3 The diagram shows step C of the conversion catalyzed by 3-hydroxybutyryl-CoA reductase.

[0132] Several acyl-CoA dehydrogenases are capable of reducing acyl-CoA to its corresponding aldehyde. Exemplary genes encoding such enzymes include Acinetobacter calciacetate acr1 (Reiser et al., J. of Bacteriology, 179:2969-2975 (1997)) encoding fatty acyl-CoA reductase, Acinetobacter spermatophore M-1 (Ishige et al., Appl. Environ. Microbiol., 68:1192-1195 (2002)) in Clostridium kluferum, and the sucD gene in Clostridium kluferum encoding CoA- and NADP-dependent succinate semialdehyde dehydrogenase (Sohling et al., J. Bacteriol., 178:871-880 (1996)). The sucD enzyme of *Porphyromonas gingivalis* is another succinate hemialdehyde dehydrogenase (Takahashi et al., J. Bacteriol., 182:4704-4710 (2000)). The acylated acetaldehyde dehydrogenase encoded by bphG in *Pseudomonas spinosa* remains another enzyme shown to oxidize and acylate acetaldehyde, propionaldehyde, butyraldehyde, isobutyraldehyde, and formaldehyde (Powlowski et al., J. Bacteriol., 175:377-385 (1993)). In addition to reducing acetyl-CoA to ethanol, the enzyme encoded by adhE in *Leuconostoc mesenteroides* has been shown to oxidize the branched compound isobutyraldehyde to isobutyryl-CoA (Kazahaya et al., ibid.; Koo et al., ibid.). Butyraldehyde dehydrogenase catalyzes a similar reaction in solvent-genic organisms (such as Clostridium saccharoacetate-polybutanol) for the conversion of butyryl-CoA to butyraldehyde (Kosaka et al., Biosci. Biotechnol. Biochem., 71:58-61 (2007)). Other candidate aldehyde dehydrogenases have been found in olefin-desulfurizing bacteria (Desulfatibacillum alkenivorans), Citrobacter krusei, Salmonella enterica, Lactobacillus brevis, and Bacillus selenitireducens. Sequence-related data for each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 12 below.

[0133] Table 12

[0134]

[0135] Other enzymes that convert acyl-CoA to their corresponding aldehydes include malonyl-CoA reductases, which convert malonyl-CoA to malonyl hemialdehyde. Malonyl-CoA reductases are key enzymes in autotrophic carbon fixation via the 3-hydroxypropionic acid cycle in thermophilic archaea (Berg et al., ibid.; Thauer, RK, Science, 318: 1732-1733 (2007)). These enzymes utilize NADPH as a cofactor and have been characterized in *Aureococcus tarda* and *Sulphurella* species (Alber et al., J. Bacteriol., 188: 8551-8559 (2006); Hugler et al., ibid.). These enzymes are encoded by Msed_0709 in *Aureococcus tarda* (Alber et al., ibid.; Berg et al., ibid.). A gene encoding malonyl-CoA reductase from the sulfur-bearing bacterium *Tokodaii* was cloned and heterologously expressed in *Escherichia coli* (Alber et al., ibid.). This enzyme has also been shown to catalyze the conversion of methylmalonyl-CoA to its corresponding aldehyde (2007). Although the aldehyde dehydrogenases of these enzymes are functionally similar to bifunctional dehydrogenases from *Flexobacterium orangeense*, there is little sequence similarity. The two malonyl-CoA reductases show high sequence similarity relative to aspartate hemialdehyde dehydrogenase (a catalyzed reduction of aspartic-4-phosphate to aspartic hemialdehyde and the simultaneous dephosphorylation). Other genes can be identified by protein sequence homology in other organisms, including *Tokodaii* sulfidea and *Tokodaii* sulfidea, and are listed below. Another enzyme targeting CoA-acylated aldehyde dehydrogenase is the *ald* gene from *Clostridium beyerridis* (Toth et al., *Appl. Environ. Microbiol.*, 65: 4973-4980 (1999)). This enzyme has been reported to reduce acetyl-CoA and butyryl-CoA to their corresponding aldehydes. This gene is very similar to the acetaldehyde dehydrogenase *eutE* encoding *Salmonella typhimurium* and *Escherichia coli* (Toth et al., ibid.). Sequence data associated with each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 13 below.

[0136] Table 13

[0137] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> MSED_0909 YP_001190808.1 146303492 Diligent golden cocci mcr NP_378167.1 15922498 sulfur leaf fungus tokodaii asd-2 NP_343563.1 15898958 Sulfur mineral sulfide leaf fungus Saci_2370 YP_256941.1 70608071 Acidophilic thermosulfur leaf fungus Ald AAT66436 9473535 Clostridium beyerridis eutE AAA80209 687645 Salmonella typhimurium eutE P77445 2498347 E. coli

[0138] The oxidative deamination of amino groups to their corresponding bridging groups is catalyzed by a deamination oxidoreductase of EC grade 1.4.1. This enzyme utilizes NAD+. + NADP + or FAD + As an acceptor, this level of enzyme can perform the following conversion: 2-amino-4-oxovalerate to 2,4-dioxovalerate ( Figure 1Step B), from 2-amino-4-hydroxyvalerate to 2-oxo-4-hydroxyvalerate ( Figure 1 Step M) and 4-aminobutyr-2-one to 3-oxobutyraldehyde ( Figure 1Step K). Exemplary oxidoreductases acting on similar substrates include gdhA-encoded glutamate dehydrogenase (deamination), ldh-encoded leucine dehydrogenase (deamination), and nadX-encoded aspartate dehydrogenase (deamination). The gdhA gene product from *Escherichia coli* (McPherson et al., *Nucleic. Acids Res.* 11:5257-5266 (1983); Korber et al., *J. Mol. Biol.* 234:1270-1273 (1993)), and gdh from *Thermophyton floccosum* (Kort et al., *Extremophiles*...) are examples of this type of oxidoreductase. 1:52-60 (1997); Lebbink et al., J.Mol.Biol.280:287-296 (1998); Lebbink et al., J.Mol.Biol.289:357-369 (1999)) and gdhA1 from halophilic bacteria (Ingoldsby et al., Gene.349:237-244 (2005)) catalyze the reversible interconversion of glutamate to 2-oxoglutarate and ammonia, with NADP(H), NAD(H) or both preferred, respectively. Other candidate glutamate dehydrogenase genes have been found in Bacillus subtilis (Khan et al., Biosci. Biotechnol Biochem. 69:1861-1870 (2005)), tobacco (Purnell et al., Planta 222:167-180 (2005)), rice (Abiko et al., Plant Cell Physiol 46:1724-1734 (2005)), Mediterranean halophilic bacteria (Diaz et al., Extremophiles. 10:105-115 (2006)), halophilic bacteria (Hayden et al., FEMS Microbiol Lett. 211:37-41 (2002)) and yeast (Roca et al., ApplEnviron. Microbiol 69:4732-4736 (2003)). Tobacco enzymes are composed of α and β subunits encoded by gdh1 and gdh2 (Purnell et al., Planta 222:167-180 (2005)). The ldh gene of Bacillus cereus encodes a LeuDH protein that accepts a wide variety of substrates, including leucine, isoleucine, valine, and 2-aminobutyric acid (Stoyan et al., J. Biotechnol 54:77-80 (1997); Ansorge et al., Biotechnol Bioeng. 68:557-562 (2000)).The biosynthesis of NAD involves the nadX gene, which encodes an aspartate dehydrogenase, from *Thermophyton floccosum* (Yang et al., J. Biol. Chem. 278:8804-8808 (2003)). Sequence data associated with each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 14 below.

[0139] Table 14

[0140] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> gdhA P00370 118547 E. coli gdh P96110.4 6226595 Thermophyton rubrum gdhA1 NP_279651.1 15789827 halophilic bacteria rocG NP_391659.1 16080831 Bacillus subtilis gdh1 AAR11534.1 38146335 tobacco gdh2 AAR11535.1 38146337 tobacco GDH Q852M0 75243660 rice GDH Q977U6 74499858 Mediterranean halophilic bacteria GDH P29051 118549 halophilic bacteria GDH2 NP_010066.1 6319986 brewing yeast ldh P0A393 61222614 Bacillus cereus nadX NP_229443.1 15644391 Thermophyton rubrum

[0141] An enzyme with 4-aminobutyronitrile oxidoreductase (deamination) activity is required to convert 4-aminobutyronitrile into its corresponding aldehyde. Figure 1 (Step K). Exemplary candidate enzymes include 3,5-diaminohexanoate dehydrogenase (EC 1.4.1.11) and lysine 6-dehydrogenase (EC 1.4.1.18). 3,5-diaminohexanoate dehydrogenase interconverts 3-amino acids and 3-oxoacids and has been characterized in organisms that ferment lysine. Recently, the gene kdd encoding 3,5-diaminohexanoate dehydrogenase was discovered in *Fusobacterium nucleatum* (Kreimeyer et al., J Biol. Chem. 282: 7191-7197 (2007)). The enzyme has been purified and characterized in other organisms (Baker et al., J Biol. Chem. 247: 7724-7734 (1972); Baker et al., Biochemistry 13: 292-299 (1974)), but the genes associated with these enzymes are unknown. Candidate enzymes in other organisms can be inferred by sequence homology. The lysDH gene encodes a lysine 6-dehydrogenase that catalyzes the conversion of primary amines to their corresponding aldehydes. This enzyme naturally catalyzes the reversible oxidative deamination of the 6-amino group of L-lysine to form 2-aminoadipic acid-6-hemialdehyde (Misono et al., J Bacteriol. 150: 398-401 (1982)). Exemplary enzymes have been found in *Bacillus stearothermophilus* (Heydari et al., Appl Environ. Microbiol 70: 937-942 (2004)), *Agrobacterium tumefaciens* (Hashimoto et al., J Biochem. 106: 76-80 (1989); Misono and Nagasaki, J Bacteriol. 150: 398-401 (1982)), and *Achromobacterium denitrifyingans* (Ruldeekulthamrong et al., BMB. Rep. 41: 790-795 (2008)). Data associated with the sequence of each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 15 below.

[0142] Table 15

[0143] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> kdd AAL93966.1 19713113 Fusobacterium nucleatum lysDH BAB39707 13429872 Thermophilic lipophilic Bacillus lysDH NP_353966 15888285 Agrobacterium tumefaciens lysDH AAZ94428 74026644 Denitrifying colorless bacteria

[0144] 2-Amino-4-oxovalerate (AKP) thiohydrolase or AKP thiohydrolase (AKPT) Figure 1 Step 1) is a pyridoxal phosphate-dependent enzyme involved in the degradation of ornithine in Clostridium stearothermia (A. Biochemistry, 13: 2898-2903 (1974); Kenklies et al., Microbiology, 145: 819-826 (1999)). Gene clusters encoding the α and β subunits (or-2 (ortA) and or-3 (ortB)) of AKPT have been identified, and the biochemical properties of the enzyme have been characterized (Fonknechten et al., J. Bacteriol., In Press (2009)). The enzyme functions in both directions and reacts with the D-isomer of alanine. Enzyme engineering can be performed to optimize the function of L-alanine as a substrate. AKPT from Clostridium stearothermia has been characterized, but its protein sequence has not yet been published. Enzymes with high sequence homology were found in *Clostridium difficile*, *Alkaliphilus metalliredigenes* QYF, *Thermophilic Anaerobes* X514, and *Thermophilic Anaerobes tengchong* MB4 (Fonknechten et al., ibid.). Sequence-related data for each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 16 below.

[0145] Table 16

[0146] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology <!-- 20 -->]]> ortA(A) YP_001086914.1 126698017 Clostridium difficile 630 ortB(β) YP_001086915.1 126698018 Clostridium difficile 630 Amet_2368(α) YP_001320181.1 150390132 Alkaliphilus metalliredigenes QYF Amet_2369(β) YP_001320182.1 150390133 Alkaliphilus metalliredigenes QYF Teth514_1478(α) YP_001663101.1 167040116 Thermophilic anaerobic bacteria X514 Teth514_1479(β) YP_001663102.1 167040117 Thermophilic anaerobic bacteria X514 TTE1235(α) NP_622858.1 20807687 Tengchong thermophilic anaerobic bacillus MB4 thrC(β) NP_622859.1 20807688 Tengchong thermophilic anaerobic bacillus MB4

[0147] The conversion of 2-amino-4-oxovalerate (AKP) to 2,4-dioxovalerate (step B) Figure 1 This conversion is accomplished via 2-amino-4-oxovalerate transaminase or oxidoreductase (deamination). The choice of appropriate enzyme for this conversion depends on the stereochemistry of the substrate. For example, if the substrate is D-configured, a D-amino acid transaminase (EC 2.6.1.21) can be used, while an L-stereoisomer can be utilized by an L-transaminase, such as aspartate transaminase (EC 2.6.1.1).

[0148] Aspartate aminotransferase transfers the amino group from aspartic acid to α-ketoglutarate, forming glutamic acid and oxaloacetic acid. Aspartic acid is structurally similar to 2-amino-4-oxovalerate. The transformation is catalyzed by, for example, the aspC gene product from *Escherichia coli* (Yagi et al., FEBS Lett., 100: 81-84 (1979); Yagi et al., Methods Enzymol., 133: 83-89 (1985)), AAT2 from *Saccharomyces cerevisiae* (Yagi et al., J. Biochem., 92: 35-43 (1982)) and ASP5 from *Arabidopsis thaliana* (Kwok et al., J. Exp. Bot., 55: 595-604 (2004); De la et al., Plant J., 46: 414-425 (2006); Wilkie et al., Protein Expr. Purif., 12: 381-389 (1998)). It has been shown that enzymes from brown rats induce the transamination of alternative substrates such as 2-aminoadipic acid and 2,4-diaminobutyric acid (Recasens et al., Biochemistry, 19:4583-4589 (1980)). Transaminases acting on other amino acid substrates can also catalyze this conversion. Valine transaminases catalyze the conversion of valine and pyruvate to 2-ketoisovaleric acid and alanine. The *E. coli* gene avtA encodes such an enzyme (Whalen et al., J. Bacteriol., 150:739-746 (1982)). The product of this gene also catalyzes the amination of α-ketobutyric acid to produce α-aminobutyric acid, although the amine donor in this reaction has not yet been identified (Whalen et al., J. Bacteriol., 158:571-574 (1984)). Another candidate enzyme is α-aminoadipic acid transaminase (EC2.6.1.39), an enzyme involved in the biosynthesis and degradation of lysine in some organisms. The lysN-encoded enzyme from *Thermophilus* is activated by several alternative substrates, including oxaloacetate, 2-oxoisohexanoic acid, 2-oxoisovaleric acid, and 2-oxo-3-methylvaleric acid (Miyazaki et al., *Microbiol.* 150: 2327-2334 (2004)). Similar enzymes from *Homo sapiens* have been characterized (Okuno et al., *Enz. Prot.* 47: 136-148 (1993)). Sequence-related data for each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 17 below.

[0149] Table 17

[0150] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> aspC NP_415448.1 16128895 E. coli AAT2 P23542.3 1703040 brewing yeast ASP5 P46248.2 20532373 Arabidopsis got2 P00507 112987 brown rat avtA YP_026231.1 49176374 E. coli lysN BAC76939.1 31096548 Thermophilic bacteria AadAT-II Q8N5Z0.2 46395904 Homo sapiens

[0151] When the substrate is present as a D-stereoisomer, transamination can be catalyzed by D-transaminases (EC2.6.1.21) (also known as D-amino acid transaminases and D-alanine transaminases (DAAT)). This class of enzymes is known for its broad substrate specificity (which is species-specific). D-transaminases encoded by dat from Bacillus YM-1 have been cloned, sequenced (Tanizawa et al., J. Biol. Chem., 264: 2450-2454 (1989)), and their crystal structures have been resolved (Peisach et al., Biochemistry, 37: 4958-4967 (1998)). This enzyme has also been a subject of protein engineering research to alter substrate specificity (Gutierrez et al., Eur. J. Biochem, 267: 7218-7223 (2000); Gutierrez et al., Protein Eng., 11: 53-58 (1998)). Other genes have been found in Bacillus licheniformis ATCC10716 (Taylor et al., Biochim. Biophys. Acta., 1350: 38-40 (1997)), Staphylococcus aureus (Pucci et al., J. Bacterial., 177: 336-342 (1995)), and Bacillus subtilis (Martinez-Carrion et al., J. Biol. Chem., 240: 3538-3546 (1965)). Sequence data associated with each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 18 below.

[0152] Table 18

[0153] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> dat P19938 118222 Bacillus YM-1 dat P54692 1706292 Bacillus licheniformis ATCC 10716 dat P54694 1706294 hemolytic staphylococci dat O07597.1 3121979 Bacillus subtilis

[0154] exist Figure 1 In reaction K, 4-aminobutyric acid-2-one is transaminated to form 3-oxobutyraldehyde. This conversion can be catalyzed by transaminases that interconvert terminal amines and aldehydes. Exemplary candidate enzymes are β-alanine / α-ketoglutarate transaminase, GABA transaminase, 3-amino-2-methylpropionic acid transaminase, lysine-6-transaminase, 2,4-diaminobutyric acid transaminase, putrescine transaminase, and diamine transaminase.

[0155] Cargill has developed and patented a β-alanine / α-ketoglutarate transaminase for the production of 3-HP from β-alanine via malonyl-semialdehyde (Chandra et al., ARch. Microbiol., 176: 443-451 (2001)). The SkPYD4 gene product in *Saccharomyces cerevisiae* also shows a preference for β-alanine as an amino donor (Aberhart et al., J. Chem. Soc. 6: 1404-1406 (1979)). SkUGA1 encodes the homologue of *Saccharomyces cerevisiae* GABA transaminase UGA1 (Ichikawa et al., J. MoL Catalysis A-Chem., 256: 106-112 (2006)), while SkPYD4 encodes the enzyme involved in the transamination of β-alanine and GABA (Aberthart et al., ibid.). 3-Amino-2-methylpropionic acid transaminase catalyzes the conversion from methylmalonyl hemialdehyde to 3-amino-2-methylpropionic acid. This enzyme has been characterized in brown rats and wild boars and is encoded by Abat (Chopra et al., Protein Expr. Purif., 25: 533-540 (2002); Kuznetsova et al., FEMS Microbiol. Rev., 29: 263-279 (2005)). Candidate enzymes with high sequence homology to 3-amino-2-methylpropionic acid transaminase in other organisms include Gta-1 in nematodes and gabT in Bacillus subtilis. Furthermore, one of the native GABA transaminases encoded by the gabT gene in *E. coli* has been shown to have broad substrate specificity (Fontaine et al., J. Bacteriol., 184: 821-830 (2002); Kanamasa et al., Appl. Microbiol Biotechnol., 80: 223-229 (2008)). The puuE gene encodes another 4-aminobutyric acid transaminase in *E. coli* (Drummond et al., J. Biol. Chem., 235: 318-325 (1960)).

[0156] Lysine-6-transaminase converts lysine to α-aminohexanedioic acid hemialdehyde. Candidate enzymes have been characterized in *Candida utilis* (Hammer et al., *J Basic Microbiol* 32: 21-27 (1992)), *Flavobacterium spp.* (Fujii et al., *J Biochem.* 128: 391-397 (2000)), and *Streptomyces pubescens* (Romero et al., *J Ind. Microbiol Biotechnol* 18: 241-246 (1997)). Recombinant lysine-6-transaminase from *Streptomyces pubescens* is functionally expressed in *Escherichia coli* (Tobin et al., *J Bacteriol.* 173: 6223-6229 (1991)). *Flavobacterium spp.* enzymes are specific for α-ketoglutarate, which is the amino acceptor (Soda et al., *Biochemistry* 7: 4110-4119 (1968)). Enzymes with diaminobutyric acid (DABA) transaminase activity are encoded by the product of the dat gene in *Acinetobacter baumannii* (Ikai et al., J Bacteriol. 179: 5118-5125 (1997)). In addition to its native substrate 2,4-diaminobutyric acid (2,4-diaminobutyric acid), DAT transaminases the terminal amines of lysine, 4-aminobutyric acid, and ornithine. Candidate putrescine transaminases are encoded by ygjG in *Escherichia coli* and spuC in *Pseudomonas aeruginosa* (Lu et al., J Bacteriol. 184: 3765-3773 (2002)). The ygiG gene product reacts with alternative substrates cadaverine, spermidine, and 1,7-diaminoheptanoic acid (Samsonova et al., BMC. Microbiol 3: 2 (2003); Kim, J Biol. Chem. 239: 783-786 (1964)).

[0157] Data associated with the sequence of each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 19 below.

[0158] Table 19

[0159] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> SkyPYD4 ABF58893.1 98626772 Kluyveri yeast SkUGA1 ABF58894.1 98626792 Kluyveri yeast UGA1 NP_011533.1 6321456 brewing yeast Abat P50554.3 122065191 brown rat Abat P80147.2 120968 wild boar GTA-1 Q21217.1 6016091 nematodes gabT P94427.1 6016090 Bacillus subtilis gabT P22256.1 16130576 Escherichia coli K12 puuE NP_415818.1 16129263 Escherichia coli K12 lat BAB13756.1 10336502 Flavobacterium sarcodactylis lat AAA26777.1 153343 Streptomyces with small sticks dat P56744.1 6685373 Acinetobacter baumannii ygjG NP_417544 145698310 E. coli spuC AAG03688 9946143 Pseudomonas aeruginosa

[0160] Figure 1In step C, 2,4-dioxolanoic acid is decarboxylated by 2,4-dioxolanoic acid decarboxylase to form 3-oxobutyraldehyde. 2,4-dioxolanoic acid is similar to the natural substrates of pyruvate decarboxylase (EC 4.1.1.1) and benzoylcarboxylate decarboxylase (EC 4.1.1.7). Pyruvate decarboxylase (PDC) (also known as keto acid decarboxylase) is a key enzyme catalyzing the decarboxylation of pyruvate to aldehydes in alcoholic fermentation. Enzymes from *Saccharomyces cerevisiae* have a broad substrate range for aliphatic 2-keto acids, including 2-ketobutyric acid, 2-ketovalerate, 3-hydroxypyruvate, and 2-phenylpyruvate (Li et al., Biochemistry, 38: 10004-10012 (1999)). Extensive studies have been conducted on this enzyme to modify its activity, its design, and its functional expression in *E. coli* (Killenberg-Jabs et al., *Eur. J. Biochem.*, 268: 1698-1704 (2001); Li et al., ibid.; Schure et al., *Appl. Environ. Microbiol.*, 64: 1303-1307 (1998)). PDCs encoded by pdcs from *M. motile fermentum* also possess a broad substrate range and have been the subject of targeted engineering studies to alter their affinity for different substrates (Siegert et al., *Protein Eng. Des. Sel.*, 18: 345-357 (2005)). The crystal structure of this enzyme is available (Killenberg-Jabs, ibid.). Other well-characterized PDC enzymes include those from *Acetobacter pastoris* (Chandra et al., Arch. Microbiol. 176: 443-451 (2001)) and *Kluyveromyces lactis* (Krieger et al., Eur. J. Biochem., 269: 3256-3263 (2002)). Sequence-associated data for each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 20 below.

[0161] Table 20

[0162] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> pdc P06672.1 118391 Motile fermentation monoclonal bacteria pdc1 P06169 30923172 brewing yeast pdc Q8L388 20385191 Acetobacter pastoris pdc1 Q12629 52788279 Lactobacillus kurrovis

[0163] Similar to PDC, benzoylformate decarboxylase (EC 4.1.1.7) has a broad substrate range and has been a target of enzyme engineering research. Enzymes from *Pseudomonas putida* have been extensively studied, and their crystal structures have been obtained (Polovnikova et al., *Biochemistry* 42: 1820-1830 (2003); Hasson et al., *Biochemistry* 37: 9918-9930 (1998)). Site-directed mutagenesis of two residues in the active site of this *Pseudomonas putida* enzyme altered the affinity (Km) for both naturally occurring and non-naturally occurring substrates (Siegert et al., ibid.). The properties of this enzyme have been further modified through directed engineering (Lingen et al., *Chembiochem* 4: 721-726 (2003); Lingen et al., *Protein Eng.* 15: 585-593 (2002)). The enzyme encoded by mdlC from *Pseudomonas aeruginosa* has also been experimentally characterized (Barrowman et al., FEMS Microbiology Letters, 34: 57-60 (1986)). Other genes from *Pseudomonas schlegelii*, *Pseudomonas fluorescens*, and other organisms can be identified by sequence homology inference or by using a growth selection system cultured in *Pseudomonas putida* (Henning et al., Appl. Environ. Microbiol., 72: 7510-7517 (2006)). Sequence-associated data for each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 21 below.

[0164] Table 21

[0165] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> mdlC P20906.2 3915757 Pseudomonas putida mdlC Q9HUR2.1 81539678 Pseudomonas aeruginosa dpgB ABN80423.1 126202187 Pseudomonas schrenckii ilvB-1 YP_260581.1 70730840 Pseudomonas fluorescens

[0166] The third enzyme capable of decarboxylating 2-oxoacids is α-ketoglutarate decarboxylase (KGD). The substrate range of this class of enzymes has not yet been investigated. Genomatica has cloned KDC from Mycobacterium tuberculosis (Tian et al., Proc. Natl. Acad. ScL USA, 102: 10670-10675 (2005)) and has functionally expressed it in Escherichia coli. KDC enzyme activity has been detected in several species of rhizobia, including *Rhizobium spp.* soybean and *Rhizobium spp.* lobelea* (Green et al., J. Bacteriol., 182: 2838-2844 (2000)). Although the KDC-encoding genes have not yet been isolated from these organisms, genomic sequences are available, and several genes in each genome have been annotated as putative KDCs. KDCs from Euglena have also been characterized, but the gene associated with this activity has not yet been identified (Shigeoka et al., Arch. Biochem. Biophys., 288: 22-28 (1991)). The sequence of the first twenty amino acids from the N-terminus is MTYKAPVKDVKFLLDKVFKV (Shigeoka et al., ibid.). The gene can be identified by KDC activity sequencing of the gene containing this N-terminal sequence. Sequence data associated with each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 22 below.

[0167] Table 22

[0168] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> kgd O50463.4 160395583 Mycobacterium tuberculosis kgd NP_767092.1 27375563 Bradyrhizobium soybean USDA110 kgd NP_105204.1 13473636 Slow-growing rhizobia in the roots of Pistacia chinensis

[0169] The fourth enzyme used to catalyze this step is branched-chain α-keto acid decarboxylase (BCKA). This class of enzymes has been shown to function on a wide variety of compounds with chain lengths ranging from 3 to 6 carbons (Oku et al., J. Biol. Chem., 263: 18386-18396 (1988); Smit et al., Appl. Environ. Microbiol., 71: 303-311 (2005)). Enzymes from *Lactococcus lactis* have been characterized on a variety of branched and straight-chain substrates (including 2-oxobutyric acid, 2-oxohexanoic acid, 2-oxovalerate, 3-methyl-2-oxobutyric acid, 4-methyl-2-oxobutyric acid, and isohexanoic acid) (Smit et al., ibid.). The structure of this enzyme has been characterized (Berthold et al., D. Biol. Crystallogr., 63: 1217-1224 (2007)). Sequence alignment between *Lactococcus lactis* enzyme and *Fermentomonas motilityis* pyruvate decarboxylase showed that the catalytic residues and substrate recognition residues were almost identical (Siegert et al., ibid.), making this enzyme highly suitable for directed engineering. Other BCKA genes were identified by homology relative to *Lactococcus lactis* protein sequences (kdcA, AAS49166.1, 44921617, *Lactococcus lactis*). Many high-resolution BLASTp hits of this enzyme were annotated as indolepyruvate decarboxylase (EC 4.1.1.74). Indolepyruvate decarboxylase (IPDA) is an enzyme that catalyzes the decarboxylation of indolepyruvate to indolealdehyde in plants and plant bacteria.

[0170] exist Figure 1In step E, 2-amino-4-ketovalerate is decarboxylated by AKP decarboxylase to form 4-aminobut-2-one. This conversion can be catalyzed by amino acid decarboxylases. The choice of appropriate decarboxylase depends on the stereochemical configuration of 4-amino-4-oxovalerate. When this compound is in the D-configuration, D-amino acid decarboxylases can be utilized. One such D-amino acid decarboxylase is diaminopimelic acid decarboxylase (DDC, EC 4.1.1.20). This enzyme decarboxylates the D-stereocenter of racemic diaminopimelic acid, catalyzing the final step in the biosynthesis of lysine. Many organisms have been studied, including *Escherichia coli* (Momany et al., D. Biol. Crystallogr., 58: 549-552 (2002)), *Mycobacterium tuberculosis* (Kefala et al., Acta. Crystallogr. Sect. F. Struct. Biol. Cryst. Commun., 61: 782-784 (2005); Gokulan et al., J. Biol. Chem., 278: 185). DDCs have been studied in *Hypericum methylophilus* (Tsujimoto et al., J. Biotechnol, 124: 327-337 (2006)) and *Helicobacter pylori* (Hu et al., J. Biol. Chem., 283: 21284-21293 (2008)). Ornithine decarboxylase (EC) derived from *Homo sapiens* has been used as an alternative. 4.1.1.17) Ornithine, the D-isomer, has low activity (Qu et al., Biochem.J., 375:465-470 (2003); Fitzgerald et al., DNA, 8:623-634 (1989)) and can be used for decarboxylation in step E. Sequence-related data for each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 23 below.

[0171] Table 23

[0172] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> lysA NP_417315.1 16130742 E. coli lysA AAA25361.1 149964 Mycobacterium tuberculosis lysA BAC92756.1 37196770 Methylophilic bacteria lysA ABW70801.1 158523325 Helicobacter pylori odcl AA59969.1 386989 Homo sapiens

[0173] When 2-amino-4-ketovalerate exhibits L-stereochemistry, amino acid decarboxylases such as aspartate decarboxylase (EC 4.1.1.11), ornithine decarboxylase (EC 4.1.1.17), or lysine decarboxylase (EC 4.1.1.18) can be utilized. An exemplary enzyme is aspartate decarboxylase (EC 4.1.1.11). 2-amino-4-ketovalerate shares structural similarity with aspartic acid (the natural substrate of this enzyme). Aspartate decarboxylase is involved in pantothenic acid biosynthesis and is encoded by panD in *Escherichia coli* (Dusch et al., *Appl. Environ. Microbiol.*, 65: 1530-1539 (1999); Ramjee et al., *Biochem. J.*, 323: 661-669 (1997); Merkel et al., *FEMS Microbiol. Lett.*, 143: 247-252 (1996); Schmitzberger et al., *EMBO J.*, 22: 6193-6204 (2003)). Enzymes from *Mycobacterium tuberculosis* (Chopra et al., *Protein Expr. Purif.*, 25: 533-540 (2002)) and *Corynebacterium glutamicum* (Dusch et al., ibid.) have been expressed and characterized in *Escherichia coli*. Lysine decarboxylase is encoded in the *Escherichia coli* genome by the genes cadA and ldcC. Recently, a CadA-like lysine decarboxylase was discovered in Vibrio parahaemolyticus (Tanaka et al., J. Appl. Microbiol. 104: 1283-1293 (2008)). A lysine decarboxylase encoded by ldc from the ruminant Lunatomia has sequence similarity to ornithine decarboxylases in eukaryotes and accepts both L-lysine and L-ornithine as substrates (Takatsuka et al., Biosci. Biotechnol Biochem. 63: 1843-1846 (1999)). Ornithine decarboxylase candidate enzymes have been found in *Nicotiana spp.* (Lee et al., *Biochem. J. 360: 657-665 (2001)), *Lactobacillus* spp. 30a (Guirard et al., *J Biol. Chem. 255: 5960-5964 (1980)), and *Vibrio vulnificus* (Lee et al., *J Biol. Chem. 282: 27115-27125 (2007)). The residues involved in the substrate specificity of *Vibrio vulnificus* have been elucidated (Lee et al., ibid.).

[0174] Data associated with the sequence of each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 24 below.

[0175] Table 24

[0176] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> panD P0A790 67470411 E. coli panD Q9X4N0 18203593 Corynebacterium glutamicum panD P65660.1 54041701 Mycobacterium tuberculosis cadA AAA23536. 145458 E. coli ldcC AAC73297.1 1786384 E. coli ldc O50657.1 13124043 Lunaemonae of Ruminants cadA AB124819.1 44886078 Vibrio parahaemolyticus AF323910.1: 1..1299 AAG45222.1 12007488 Sticky tobacco odc1 P43099.2 1169251 Lactobacillus 30a VV2_1235 NP_763142.1 27367615 Vibrio vulnificus

[0177] In reaction J( Figure 1 In acetoacrylate, acetoacrylate is decarboxylated to 2-oxobutene by acetoacrylate decarboxylase. The enzyme that catalyzes this conversion has not yet been identified, but similar reactions are catalyzed by enzymes aconitate decarboxylase, 4-oxaloylcrotonate decarboxylase, and cinnamic acid decarboxylase.

[0178] Aconitine decarboxylase catalyzes the final step in itaconic acid biosynthesis in Candida species and the filamentous fungus Aspergillus terreus (Bonnarme et al., J. Bacteriol., 177: 3573-3578 (1995); Willke et al., Appl. Microbiol. Biotechnol., 56: 289-295 (2001)). Cis-aconitine decarboxylase (CAD) encoded by ATEG_09971 (EC 4.1.16) has been found in Aspergillus terreus and other related fungi and has been extensively studied. Recently, this gene has been cloned and functionally characterized (Kanamasa et al., Appl. Microbiol. Biotechnol., 80: 223-229 (2008)) and (WO / 2009 / 014437).

[0179] 4-Oxaloylcrotonate decarboxylase has been isolated from many organisms and characterized. Genes encoding this enzyme include dmpH and dmpE from *Pseudomonas spinosa* (strain 600) (Shingler et al., J. Bacteriol., 174: 711-724 (1992)), xylII and xylIII from *Pseudomonas putida* (Kato et al., Arch. Microbiol., 168: 457-463 (1997); Stanley et al., Biochemistry, 39: 3514 (2000); Lian et al., J. Am. Chem. Soc., 116: 10403-10411 (1994)), and Reut_B5691 and Reut_B5692 from *Ralstonia macrophylla* JMP134 (Hughes et al., J. Bacteriol., 158: 79-83 (1984)). The gene encoding this enzyme from *Pseudomonas spinosa* (strain 600) has been cloned and characterized in *Escherichia coli* (Shingler et al., ibid.). Data associated with the sequence of each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 25 below.

[0180] Table 25

[0181] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> dmpH CAA43228.1 45685 Pseudomonas polysaccharidus CF600 dmpE CAA43225.1 45682 Pseudomonas polysaccharidus CF600 xylII YP_709328.1 111116444 Pseudomonas putida xvlIII YP_709353.1 111116469 Pseudomonas putida Reut_B5691 YP_299880.1 73539513 Optimal Rawlstonella JMP134 Reut_B5692 YP_299881.1 73539514 Optimal Rawlstonella JMP134 ATEG_09971 EAU29420.1 114187720 Aspergillus terreus

[0182] Other grades of decarboxylases catalyzing the conversion of cinnamic acid (phenylacrylic acid) and substituted cinnamic acid derivatives to the corresponding styrene derivatives have been characterized. These enzymes are common in various organisms, and the specific genes encoding these enzymes that have been cloned and characterized in *E. coli* are as follows: pad 1 from *Saccharomyces cerevisiae* (Clausen et al., *Gene*, 142: 107-112 (1994)), pdc from *Lactobacillus plantarum* (Barthelmebs et al., *Appl. Environ. Microbiol.*, 67: 1063-1069 (2001); Rodriguez et al., *J. Agric. Food*). Chem., 56:3068-3072 (2008); Qi et al., Biochem.J., 375:465-470 (2007)), pofK (pad) from Klebsiella acidogenetic bacteria (Uchiyama et al., Biosci. Biotechnol. Biochem., 72:116-123 (2008); Hashidoko et al., Biosci. Biotech. Biochem., 58:217-218 (1994)), pofK (pad) from Pediococcus pentosaceus (Barthelmebs et al., ibid.), and padC from Bacillus subtilis and Bacillus pumilus (Cavin et al., Appl. Environ. Microbiol., 64:1466-1471 (1998)). Ferulic acid decarboxylases from *Pseudomonas fluorescens* have also been purified and characterized (Huang et al., J. Bacteriol., 176: 5912-5918 (1994)). Importantly, this class of enzymes has been shown to be stable and does not require exogenous or endogenous cofactors, thus making these enzymes ideally suited for biotransformation (Sariaslani, FS, Annu. Rev. Microbiol., 61: 51-69 (2007)). Sequence-associated data for each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 26 below.

[0183] Table 26

[0184] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> pad1 AAB64980.1 1165293 brewing yeast pdc AAC45282.1 1762616 Lactobacillus plantarum pad BAF65031.1 149941608 Klebsiella pneumoniae padC NP_391320.1 16080493 Bacillus subtilis pad YP_804027.1 116492292 Pediococcus pentosaceus pad CAC18719.1 11691810 Bacillus pumilus

[0185] Other enzymes used for decarboxylation include acetoacetic acid decarboxylase (EC 4.1.1.4), an enzyme that decarboxylates acetoacetic acid to acetone, and has therefore been studied for its role in bacterial biosolubility. Exemplary bacterial enzymes have been characterized from *Clostridium acetobutyricum* (Benner et al., J. Am. Chem. So. 103: 993-994 (1981); Highbarger et al., Biochemistry 35: 41-46 (1996); Petersen et al., Appl. Environ. Microbiol. 56: 3491-3498 (1990); Rozzel et al., J. Am. Chem. Soc. 106: 4937-4941 (1984)), *Clostridium glycoacetobutyricum* (Kosaka et al., Biosci. Biotechnol Biochem. 71: 58-68 (2007)), and *Clostridium beyerii* (Ravagnani et al., Mol. Microbiol. 37: 1172-1185 (2000)). Acetoacetate decarboxylase activity has also been demonstrated in *Pseudomonas putida* and *Bacillus polymyxa*, but genes associated with this activity have not yet been identified (Matiasek et al., *Curr. Microbiol.* 42: 276-281 (2001)). Bacterial genes in other organisms (such as *Clostridium botulinum* and *Bacillus amyloliquefaciens*) can be identified through sequence homology. In humans and other mammals, acetoacetate decarboxylase catalyzes the final step of the ketone body pathway (Kalapos, *Biochim. Biophys. Acta 1621: 122-139 (2003)), but genes associated with this activity have not yet been identified. Sequence-associated data for each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 27 below.

[0186] Table 27

[0187] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> adc NP_149328.1 15004868 Clostridium acetonebutanol adc AAP42566.1 31075386 Clostridium glycoacetate polybutanol cbei_3835 YP_001310906.1 150018652 Clostridium beyerridis CLL_A2135 YP_001886324.1 187933144 Clostridium botulinum RBAM_030030 YP_001422565.1 154687404 Bacillus amyloliquefaciens

[0188] All of the aforementioned candidate genes can also be used for catalysis. Figure 1 In step N, the decarboxylation of 2-oxo-4-hydroxyvalerate to 3-hydroxybutyraldehyde is performed.

[0189] Butenone hydratase ( Figure 1 Step G), 4-hydroxybutyryl-CoA dehydratase ( Figure 3 Step A) and crotonase ( Figure 3 Step A) is the conversion of the hydrolyzing enzyme type. Specifically, it involves the hydration of butenone to 4-hydroxy-2-butenone. Figure 1Step G) can be accomplished by an enzyme from the hydratase family. Enzymes that can perform this conversion include fumarate hydratase (EC 4.2.1.2), 2-(hydroxymethyl)glutarate dehydratase (EC 4.2.1.-), dimethylmaleate hydratase (EC 4.2.1.85), and citrate hydrolysase (EC 4.2.1.34).

[0190] Fumarate hydratase naturally catalyzes the reversible hydration of fumarate to malate. Although the ability of fumarate hydratase to react with butenone as a substrate is not described in the literature, a wealth of structural information about this enzyme is available, and other researchers have successfully engineered it to alter its activity, inhibition, and localization (Weaver, T., B. Biol. Crystallogr., 61: 1395-1401 (2005)). *Escherichia coli* possesses three fumarate enzymes: FumA, FumB, and FumC, which are modulated by various culture conditions. FumB is oxygen-sensitive and active only under anaerobic conditions, FumA is active under microanaerobic conditions, and FumC is the only enzyme active under aerobic conditions (Tseng et al., J. Bacteriol., 183: 461-467 (2001); Woods et al., Biochem. Biophys. Acta., 954: 14-26 (1988); Guest et al., J. Gen. Microbiol., 131: 2971-2984 (1985)). Other enzymes have been found in Campylobacter jejuni (Smith et al., Int. J. Biochem. CellBiol., 31: 961-975 (1999)), Thermophilus thermophilus (Mizobata et al., Arch. Biochem. Biophys., 355: 49-55 (1998)), and Rhizopus rubrum (Kobayashi et al., J. Biochem., 89: 1923-1931 (1981)). Similar enzymes with high sequence homology include fum1 from Arabidopsis thaliana and fumC from Corynebacterium glutamicum. MmcBC fumarate from Pelotomaculum thermopropionicum is another class of fumarate with two subunits (Shimoyama et al., FEMS Microbiol. Lett., 270: 207-213 (2007)). Data associated with the sequence of each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 28 below.

[0191] Table 28

[0192] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology <!-- 28 -->]]> fumA NP_416129.1 16129570 E. coli fumB NP_418546.1 16131948 E. coli fumC NP_416128.1 16129569 E. coli fumC O69294 9789756 Campylobacter jejuni fumC P84127 75427690 Thermophilic bacteria fumH P14408 120605 brown rat fum1 P93033 39931311 Arabidopsis fumC Q8NRN8 39931596 Corynebacterium glutamicum MmcB YP_001211906 147677691 Pelotomaculum thermopropionicum MmcC YP_001211907 147677692 Pelotomaculum thermopropionicum

[0193] Two other hydrating enzymes are 2-(hydroxymethyl)glutarate dehydratase and dimethylmaleate hydratase, whose roles in the nicotinic acid catabolism of *Backacium* (formerly *Clostridium*) are studied (Alhapel et al., Proc. Natl. Acad. Sci. USA, 103: 12341-12346 (2006)). 2-(hydroxymethyl)glutarate dehydratase is a [4Fe-4S]-containing enzyme that dehydrates 2-(hydroxymethyl)glutarate to 2-methylene-glutarate. This enzyme is encoded by hmd in *Backacium* (Alhapel et al., ibid.). Similar enzymes with high sequence homology have been found in *Bacteroides polychaete*, *Anaerotruncus colihominis*, and the anaerobic halophilic-alkaliophilic thermophilic bacterium *Natranaerobius thermophilic*. These enzymes are homologous to the α and β subunits of [4Fe-4S]-containing bacterial serine dehydratases, such as E. coli enzymes encoded by tdcG, sdhB, and sdaA. Dimethylmaleic acid hydratase (EC 4.2.1.85) is a reversible Fe2+-dependent and oxygen-sensitive enzyme in the cis-aconitase family that hydrates dimethylmaleic acid to form (2R,3S)-2,3-dimethylmalate. This enzyme is encoded by dmdAB in *Backacium* (Alhapel et al., ibid.; Kollmann-Koch et al., Physiol. Chem., 365:847-857 (1984)). Sequence-associated data for each of these exemplary gene products can be found using the GenBank accession numbers shown in Table 29 below.

[0194] Table 29

[0195]

[0196] The other enzyme is 2-methylmalate dehydratase (also known as citrate hydrolysase), a reversible hydrolysase that catalyzes the α,β elimination of water from citrate malate to form mesoconate. This enzyme has been purified and characterized in Clostridium tetani (Wang et al., J. Biol. Chem., 244:2516-2526 (1969)). The activity of this enzyme has also been detected in various bacteria of the genera *Citrobacter* and *Morganella* in the context of the glutamate degradation VI pathway (Kato et al., ibid.). To date, no gene encoding this enzyme has been found in any organism.

[0197] Crotonylase (EC 4.2.1.55) catalyzes the hydration of crotonyl-CoA to form 3-hydroxybutyryl-CoA (CLOSE). Figure 3Step B). These enzymes are required for the formation of n-butanol in some organisms (specifically, Clostridium spp.) and include a step in the 3-hydroxypropionic acid / 4-hydroxybutyric acid cycle in thermophilic archaea such as Sulfophyllum spp., Acidobacteria spp., and Agrobacterium spp. Exemplary genes encoding crotonases can be found in Clostridium acetobutanol (Boynton et al., J. Bacteriol., 178:3015-3024 (1996)), Clostridium klufernum (Hillmer et al., FEBS Lett., 21:351-354 (1972)), and Agrobacterium spp. (Berg et al., ibid.). Enyl-CoA hydratases (which are involved in fatty acid β-oxidation and / or the metabolism of various amino acids) can also catalyze the hydration of crotonyl-CoA to form 3-hydroxybutyryl-CoA (Roberts et al., Arch. Microbiol., 117:99-108 (1978); Agnihotri et al., Bioorg. Med. Chem., 11:9-20 (2003); Conrad et al., J. Bacteriol., 118:103-111 (1974)). Exemplary enyl-CoA hydratases are products of the ech gene from *Pseudomonas putida* (Roberts et al., ibid.). *Pseudomonas putida* enyl-CoA hydratases phaA and phaB have been shown to perform double bond hydroxylation during phenylacetic acid catabolism (Olivera et al., Proc. Natl. Acad. Sci. USA, 95:6419-6424 (1998)). paaA and paaB from *Pseudomonas fluorescens* catalyze similar transformations (Olivera et al., ibid.). Finally, many *E. coli* genes have been shown to exhibit enoyl-CoA hydratase functionality, including maoC (Park et al., *J. Bacteriol.*, 185: 5391-5397 (2003)), paaF (Ismail et al., *Eur. J. Biochem.*, 270: 3047-3054 (2003); Park et al., *Appl. Biochem. Biotechnol.*, 113-116: 335-346 (2004); Park et al., *Biotechnol Bioeng.*, 86: 681-686 (2004)) and paaG (Ismail et al., ibid.; Park et al., ibid.; Park et al., ibid.). Data associated with the sequence of each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 30 below.

[0198] Table 30

[0199] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> crt NP_349318.1 15895969 Clostridium acetonebutanol crt1 YP_001393856 153953091 Clostridium clavatum DSM 555 ech NP_745498.1 26990073 Pseudomonas putida phaA ABF82233.1 26990002 Pseudomonas putida phaB ABF82234.1 26990001 Pseudomonas putida paaA NP_745427.1 106636093 Pseudomonas fluorescens paaB NP_745426.1 106636094 Pseudomonas fluorescens maoC NP_415905.1 16129348 E. coli paaF NP_415911.1 16129354 E. coli paaG NP_415912.1 16129355 E. coli

[0200] Alternatively, the *E. coli* gene products fadA and fadB encode multienzyme complexes involved in fatty acid oxidation, exhibiting enoyl-CoA hydratase activity (Haller et al., *Biochemistry* 39: 4622-4629 (2000); Martinez-Carrion et al., *J. Biol. Chem.* 240: 3538-3546 (1965); Matthies et al., *Appl. Environ. Micriobiol.* 58: 1435-1439 (1992)). The fadB gene product can be activated by knocking out the negative regulator encoded by fadR (Jeng et al., *A. Biochemistry* 13: 2898-2903 (1974)). The fadl and fadJ genes encode similar functions and are naturally expressed under anaerobic conditions (Atsumi et al., *Nature* 451: 86-89 (2008)). Data associated with the sequence of each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 31 below.

[0201] Table 31

[0202] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> fadA YP_026272.1 49176430 E. coli fadB NP_418288.1 16131692 E. coli fadI NP_416844.1 16130275 E. coli fadJ NP_416843.1 16130274 E. coli fadR NP_415705.1 16129150 E. coli

[0203] Reversible condensation of 4-hydroxybutyryl-CoA to crotonyl-CoA ( Figure 3 Step A) is catalyzed by the bifunctional enzyme 4-hydroxybutyryl-CoA dehydratase / ethyleneacetyl-CoA Δ-isomerase. This enzyme first dehydrates 4-hydroxybutyryl-CoA to ethyleneacetyl-CoA, which is then rearranged to form crotonyl-CoA. Enzymes from *Clostridium kluferum* and *Clostridium gammabeta* have been purified, characterized, and sequenced to their N-terminal domains (Scherf et al., Eur. J. Biochem., 215: 421-429 (1993); Scherf et al., Arch. Microbiol., 161: 239-245 (1994)). The abfD gene from *Clostridium gammabeta* and *Clostridium kluferum* perfectly matches these N-terminal amino acid sequences and has been identified as encoding 4-hydroxybutyryl-CoA dehydratase / ethyleneacetyl-CoA Δ-isomerase activity. Similar genes were identified through genome engineering via homology, including abfD from *Porphyromonas gingivalis* and Msed_1220 from *Aureococcus lucida*. Data associated with the sequences of each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 32 below.

[0204] Table 32

[0205] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> abfD YP_001396399.1 153955634 Clostridium klufernum abfD P55792 84028213 Clostridium gammabutyricum abfD YP_001928843 188994591 Porphyromonas gingivalis Msed_1220 YP_001191305.1 146303989 Diligent golden cocci

[0206] Deamination reaction of 2-amino-4-ketovalerate ( Figure 1 Reaction I) and the deamination reaction of 4-aminobutyric-2-one Figure 1 Step F) can be performed by AKP ammonia-lyase and 4-aminobutyric acid-2-ketoammonia-lyase, respectively. These deamination reactions are very similar to the deamination of aspartic acid to fumarate by aspartate enzymes. This enzyme has been extensively studied, and various crystal structures have been obtained. Escherichia coli enzymes have been shown to react with alternative substrates such as aspartic acid phenyl methyl ester, asparagine, benzyl-aspartic acid, and malic acid (Ma et al., Ann. NYAcad. Sci., 672: 60-65 (1992). In an independent study, directed evolution was performed on this enzyme to alter substrate specificity (Asano et al., Biomol. Eng., 22: 95-101 (2005)). Enzymes with aspartate enzyme functionality have also been characterized in Haemophilus influenzae (Sjostrom et al., Bioc. *Hem. Biophys. Aeta., 1324:182-190 (1997)*, *Pseudomonas fluorescens* (Takagi et al., J. Biochem., 96:545-552 (1984)), *Bacillus subtilis* (Sjostrom et al., ibid.), and *Serratia marcescens* (Takagi et al., J. Bacteriol., 161:1-6 (1985)). Sequence-associated data for each of these exemplary gene products can be found using the GenBank registry numbers shown in Table 33 below.

[0207] Table 33

[0208] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> aspA NP_418562 90111690 E. coli aspA P44324.1 1168534 Haemophilus influenzae aspA P07346.1 114273 Pseudomonas fluorescens ansB P26899.1 251757243 Bacillus subtilis aspA P33109.1 416661 Serratia marcescens

[0209] Similar ammonia-lysinic reactions are catalyzed by methylaspartase (EC 4.3.1.2), an enzyme involved in the glutamate fermentation pathway via medocanthalate (Kato et al., ibid.). This enzyme (also known as β-methylaspartase and 3-methylaspartate ammonia-lysase) naturally catalyzes the deamination of threo-3-methylaspartic acid to medocanthalate. 3-methylaspartase from Clostridium tetani has been cloned, functionally expressed in Escherichia coli, and crystallized (Asuncion et al., 57:731-733 (2001); Asuncion et al., J Biol Chem. 277:8306-8311 (2002); Botting et al., 27:2953-2955 (1988); Goda et al., 31:10747-10756 (1992)). In *Citrobacter malonic acid-free*, this enzyme is encoded by BAA28709 (Kato et al., Arch. Microbiol 168: 457-463 (1997)). 3-Methylaspartate enzyme has also been crystallized from *Escherichia coli* YG1002 (Asano et al., FEMS Microbiol Lett. 118: 255-258 (1994)), although the protein sequence is not listed in public databases (such as GenBank). Data associated with the sequences of each of these exemplary gene products can be found using the GenBank accession numbers shown in Table 34 below.

[0210] Table 34

[0211] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> mal AAB24070.1 259429 Clostridium tetani BAA28709 BAA28709.1 3184397 Citrobacter malonic acid-free

[0212] In some embodiments, the 2-amino-4-ketovalerate (AKP) thiolytic enzyme is encoded by one or more genes selected from the following: ortA (α), ortB (β), Amet_236B (α), Amet_2369 (β), Teth514_1478 (α), Teth514_1479 (β), TTE1235 (α), and thrC (β).

[0213] In some embodiments, the AKP dehydrogenase is encoded by one or more genes selected from the following: thrA, akthr2, hom6, hom1, hom2, fadB, fadJ, Hbd2, Hbd1, hbd, HSD17B10, phbB, phaB, Msed_1423, Msed_0399, Msed_0389, Msed_1993, adh, adhA, adh-A, mdh, ldhA, ldh, and bdh.

[0214] In some embodiments, the 2-amino-4-hydroxyvalerate transaminase is encoded by one or more genes selected from the following: aspC, AAT2, ASP5, got2, avtA, lysN, AadAT-II, dat, lat, ygjG, spuC, SkyPYD4, SkUGA1, UGA1, Abat, Abat, Gta-1, gabT, and puuE.

[0215] In some embodiments, the 2-amino-4-hydroxyvalerate oxidoreductase (deamination) is encoded by one or more genes selected from the following: gdhA, gdh, gdhA1, rocG, gdh1, gdh2, GDH, GDH2, ldh, and nadX.

[0216] In some embodiments, the 2-oxo-4-hydroxyvalerate decarboxylase is encoded by one or more genes selected from the following: pdc, pdc1, mdlC, dpgB, ilvB-1, kgd, kdcA, lysA, panD, cadA, ldc, ldcC, AF323910.1:1...1299, odc1, VV2_1235, dmpH, dmpE, xylII, xylIII, Reut_B5691, Reut_B5692, CAD, pad1, pofK(pad), padC, pad, adc, cbei_3835, CLL_A2135, RBAM_030030.

[0217] In some embodiments, the 3-hydroxybutyraldehyde reductase is encoded by one or more genes selected from the following: alarA, ADH2, yqhD, bdh I, bdh II, adhA, 4hbd, adhI, P84067, mmsb, dhat, and 3hidh.

[0218] In some embodiments, the AKP transaminase is encoded by one or more genes selected from the following: aspC, AAT2, ASP5, got2, avtA, lysN, AadAT-II, dat, lat, ygjG, spuC, SkyPYD4, SkUGA1, UGA1, Abat, Gta-1, gabT, and puuE.

[0219] In some embodiments, the AKP oxidoreductase (deamination) is encoded by one or more genes selected from the following: gdhA, gdh, gdhA1, rocG, gdh1, gdh2, GDH, GDH2, ldh, and nadX. In some embodiments, the 2,4-dioxovalerate decarboxylase is encoded by one or more genes selected from the following: pdc, pdc1, mdlC, dpgB, ilvB-1, kgd, kdcA, lysA, panD, cadA, ldc, ldcC, AF323910.1:1...1299, odc1, VV2_1235, dmpH, dmpE, xylII, xylIII, Reut_B5691, Reut_B5692, CAD, pad1, padC, and pad, adc, cbei_3835, CLL_A2135, RBAM_030030.

[0220] In some embodiments, the 3-oxobutyraldehyde reductase (ketone reducer) is encoded by one or more genes selected from the following: thrA, akthr2, hom6, hom1, hom2, fadB, fadJ, Hbd2, Hbd1, hbd, HSD17B10, phbB, phaB, Msed_1423, Msed_0399, Msed_0389, Msed_1993, adh, adhA, adh-A, mdh, ldhA, ldh, and bdh.

[0221] In some embodiments, the 3-oxobutyraldehyde reductase (aldehyde reducer) is encoded by one or more genes selected from the following: alrA, ADH2, yqhD, bdh I, bdh II, adhA, 4hbd, adhI, P84067, mmsb, dhat, and 3hidh.

[0222] In some embodiments, the 4-hydroxy-2-butanone reductase is encoded by one or more genes selected from the following: thrA, akthr2, hom6, hom1, hom2, fadB, fadJ, Hbd2, Hbd1, hbd, HSD17B10, phbB, phaB, Msed_1423, Msed_0399, Msed_0389, Msed_1993, adh, adhA, adh-A, mdh, ldhA, ldh, and bdh.

[0223] In some embodiments, the AKP decarboxylase is encoded by one or more genes selected from the following: pdc, pdc1, mdlC, dpgB, ilvB-1, kgd, kdcA, lysA, panD, cadA, ldc, ldcC, AF323910.1:1...1299, odc1, VV2_1235, dmpH, dmpE, xylII, xylIII, Reut_B5691, Reut_B5692, CAD, pad1, pofK (pad), padC, pad.

[0224] In some embodiments, the 4-aminobutyric-2-one transaminase is encoded by one or more genes selected from the following: aspC, AAT2, ASP5, got2, avtA, lysN, AadAT-II, dat, lat, ygjG, spuC, SkyPYD4, SkUGA1, UGA1, Abat, Gta-1, gabT, and puuE.

[0225] In some embodiments, the 4-aminobutyric-2-one oxidoreductase (deamination) is encoded by one or more genes selected from the following: gdhA, gdh, gdhA1, rocG, gdh1, gdh2, GDH, GDH2, ldh, nadX, kdd, and lysDH.

[0226] In some embodiments, the 4-aminobutyric-2-ketoamine lyase is encoded by one or more genes selected from the following: aspA, ansB, mal, and BAA28709.

[0227] In some embodiments, the butenone hydratase is encoded by one or more genes selected from the following: fumA, fumB, fumC, fumH, fum1, MmcB, MmcC, hmd, BACCAP_02294, ANACOL_02527, NtherDRAFT_2368, dmdA, dmdB, crt, crt1, echpaaA, paaB, phaA, phaB, maoC, paaF, paaG, abfD, Msed_1220, fadA, fadB, fadI, fadJ, and fadR.

[0228] In some embodiments, the AKP ammonia-lyase is encoded by one or more genes selected from the following: aspA, ansB, mal, and BAA28709.

[0229] In some embodiments, the acetylpropionic acid decarboxylase is encoded by one or more genes selected from the following: pdc, pdc1, mdlC, dpgB, ilvB-1, kgd, kdcA, lysA, panD, cadA, ldc, ldcC, AF323910.1:1...1299, odc1, VV2_1235, dmpH, dmpE, xylII, xylIII, Reut_B5691, Reut_B5692, CAD, pad1, pofK(pad), padC, pad, adc, cbei_3835, CLL_A2135, RBAM_30030.

[0230] In some embodiments, the acetyl-CoA reductase (CoA-dependent, aldehyde-forming) is encoded by one or more genes selected from the following: acr1, sucD, bphG, bld, adhE, Msed_0709, mcr, asd-2, Saci_2370, Ald, and eutE.

[0231] In some embodiments, the acetyl-CoA reductase (CoA-dependent, alcohol-forming) is encoded by one or more genes selected from the following: adhE, adhE2, mcr, Rcas_2929, NAP1_02720, MGP2080_00535, and FAR.

[0232] In some embodiments, the acetyl-CoA reductase (ketone reducer) is encoded by one or more genes selected from the following: thrA, akthr2, hom6, hom1, hom2, fadB, fadJ, Hbd2, Hbd1, hbd, HSD17B10, phbB, phaB, Msed_1423, Msed_0399, Msed_0389, Msed_1993, adh, adhA, adh-A, mdh, ldhA, ldh, and bdh.

[0233] In some embodiments, the 3-hydroxybutyryl-CoA reductase (aldehyde formation) is encoded by one or more genes selected from the following: acr1, sucD, bphG, bld, adhE, Msed_0709, mcr, asd-2, Saci_2370, Ald, and eutE.

[0234] In some embodiments, the 3-hydroxybutyryl-CoA reductase (alcohol formation) is encoded by one or more genes selected from the following: adhE, adhE2, mcr, Rcas_2929, NAP1_02720, MGP2080_00535, and FAR.

[0235] In some embodiments, the 4-hydroxybutyryl-CoA dehydratase is encoded by one or more genes selected from the following: fumA, fumB, fumC, fumH, fum1, MmcB, MmcC, hmd, BACCAP_02294, ANACOL_02527, NtherDRAFT_2368, dmdA, dmdB, crt, crt1, ech, paaA, paaB, phaA, phaB, maoC, paaF, paaG, abfD, Msed_1220, fadA, fadB, fadI, fadJ, and fadR.

[0236] In some embodiments, the crotonase is encoded by one or more genes selected from the following: fumA, fumB, fumC, fumH, fum1, MmcB, MmcC, hmd, BACCAP_02294, ANACOL_02527, NtherDRAFT_2368, dmdA, dmdB, crt, crt1, ech paaA, paaB, phaA, phaB, maoC, paaF, paaG, abfD, Msed_1220, fadA, fadB, fadI, fadJ, and fadR.

[0237] The non-naturally occurring microorganisms of this invention can produce 1,3-butanediol by introducing expressible nucleic acids encoding one or more enzymes or proteins involved in one or more 1,3-butanediol biosynthetic pathways. Depending on the host microorganism selected for biosynthesis, some or all of the nucleic acids for a specific 1,3-butanediol biosynthetic pathway can be expressed. For example, if the selected host is deficient in one or more enzymes or proteins for the desired biosynthetic pathway, expressible nucleic acids for the deficient enzymes or proteins can be introduced into the host for subsequent exogenous expression. Alternatively, if the selected host exhibits endogenous expression of some pathway genes but is deficient in others, nucleic acids encoding the deficient enzymes or proteins are required to complete the biosynthesis of 1,3-butanediol. Therefore, the non-naturally occurring microorganisms of this invention can produce 1,3-butanediol by introducing exogenous enzyme or protein activity to obtain the desired biosynthetic pathway; or by introducing one or more exogenous enzyme or protein activities to obtain the desired biosynthetic pathway, which, together with one or more endogenous enzymes or proteins, produces the desired product, such as 1,3-butanediol.

[0238] Depending on the composition of the 1,3-butanediol biosynthesis pathway of the selected host microorganism, the non-naturally occurring microorganisms of the present invention include at least one exogenously expressed 1,3-butanediol pathway-encoding nucleic acid, and even all nucleic acids encoding one or more 1,3-butanediol biosynthesis pathways. For example, 1,3-butanediol biosynthesis can be established in hosts with deficiencies in pathway enzymes or proteins through exogenous expression of the corresponding encoding nucleic acid. In hosts with deficiencies in all enzymes or proteins of the 1,3-butanediol pathway, exogenous expression of all enzymes or proteins in that pathway may be included, although it is understood that all enzymes or proteins of a pathway may be expressed, even if the host contains at least one of the enzymes or proteins of that pathway. For example, exogenous expression of all enzymes or proteins in the pathway for the production of 1,3-butanediol may be included.

[0239] In light of the teachings and guidance provided herein, those skilled in the art will understand that the number of encoding nucleic acids introduced in an expressible formation will be at least equal to the amount of 1,3-butanediol pathway deficiency in the selected host microorganism. Therefore, the non-naturally occurring microorganisms of the present invention may have one, two, three, four, five, or even all nucleic acids encoding enzymes or proteins constituting the 1,3-butanediol biosynthetic pathway disclosed herein. In some embodiments, the non-naturally occurring microorganisms may also include other gene modifications that facilitate or optimize 1,3-butanediol biosynthesis or confer other useful functions to the host microorganism. One such additional functionality may include, for example, enhancement of the synthesis of one or more 1,3-butanediol pathway precursors (e.g., acetyl-CoA).

[0240] Generally, host microorganisms should be selected to produce precursors of the 1,3-butanediol pathway, either as naturally occurring molecules or as engineered products that provide for the regeneration of the desired precursor or for the increased production of precursors naturally produced by the host microorganism. For example, acetyl-CoA is naturally produced in host organisms such as *Escherichia coli*. Host organisms can be engineered to increase precursor production, as disclosed herein. Furthermore, microorganisms engineered to produce desired precursors can be used as host organisms and further engineered to express enzymes or proteins of the 1,3-butanediol pathway.

[0241] In some embodiments, the non-naturally occurring microorganisms of the present invention are derived from hosts containing enzymatic capabilities for the synthesis of 1,3-butanediol. In this particular embodiment, they can be used to increase the synthesis or accumulation of 1,3-butanediol pathway products, for example, to induce 1,3-butanediol pathway reactions to produce 1,3-butanediol. The increase in synthesis or accumulation can be accomplished, for example, by overexpression of nucleic acids encoding one or more of the aforementioned 1,3-butanediol pathway enzymes or proteins. This overexpression of one or more enzymes and / or proteins of the 1,3-butanediol pathway can occur, for example, through exogenous expression of one or more endogenous genes or through exogenous expression of one or more heterologous genes. Therefore, naturally occurring organisms can readily generate the non-naturally occurring microorganisms of the present invention, for example, by overexpressing one, two, three, four, or five, i.e., all, of the nucleic acids encoding 1,3-butanediol biosynthetic pathway enzymes or proteins to produce 1,3-butanediol. Furthermore, non-naturally occurring organisms can be generated by mutagenesis of endogenous genes that cause an increase in enzyme activity in the 1,3-butanediol biosynthetic pathway.

[0242] In particularly useful embodiments, exogenous expression encoding nucleic acids is employed. Exogenous expression confers the ability to tailor expression and / or regulatory elements to the host and to achieve desired expression levels controlled by the user. However, other embodiments may also utilize endogenous expression, for example, by removing negative regulatory effectors or inducing gene promoters when linked to inducible promoters or other regulatory elements. Thus, endogenous genes with naturally occurring inducible promoters can be upregulated by providing appropriate inducers; or the regulatory region of an endogenous gene can be engineered to incorporate inducible regulatory elements, thereby allowing upregulation of the endogenous gene expression when needed. Similarly, inducible promoters may be included as regulatory elements for exogenous genes introduced from non-naturally occurring microorganisms.

[0243] It should be understood that, in the method of the present invention, any nucleic acid of one or more exogenous nucleic acids can be introduced into a microorganism to produce the non-naturally occurring microorganism of the present invention. Nucleic acids can be introduced to confer, for example, a 1,3-butanediol biosynthetic pathway upon the microorganism. Alternatively, nucleic acids encoding an intermediate microorganism can be introduced to produce a biosynthetic intermediate microorganism capable of catalyzing some of the desired reactions, thereby conferring 1,3-butanediol biosynthetic capability. For example, a non-naturally occurring microorganism possessing a 1,3-butanediol biosynthetic pathway may comprise at least two exogenous nucleic acids encoding the desired enzyme or protein. Therefore, it should be understood that any combination of two or more enzymes or proteins of a biosynthetic pathway can be included in the non-naturally occurring microorganism of the present invention. Similarly, it should be understood that, upon request, any combination of three or more enzymes or proteins of a biosynthetic pathway can be included in the non-naturally occurring microorganism of the present invention, etc., provided that the combination of enzymes and / or proteins of the desired biosynthetic pathway results in the production of the corresponding desired product. Similarly, as required, any combination of four or more enzymes or proteins of the biosynthetic pathway can be included in the non-naturally occurring microorganisms of the present invention, provided that the combination of enzymes and / or proteins of the desired biosynthetic pathway results in the production of the corresponding desired product.

[0244] In addition to the biosynthesis of 1,3-butanediol as described herein, the non-naturally occurring microorganisms and methods described in this invention can also be applied in various combinations with each other and with other microorganisms and methods well known in the art to obtain the product biosynthesized via other pathways. For example, besides using a 1,3-butanediol producer, an alternative method for producing 1,3-butanediol is by adding another microorganism capable of converting a 1,3-butanediol pathway intermediate to 1,3-butanediol. One such step includes, for example, fermentation of a microorganism that produces a 1,3-butanediol pathway intermediate. This 1,3-butanediol pathway intermediate can then be used as a substrate for a second microorganism that converts the 1,3-butanediol pathway intermediate to 1,3-butanediol. The 1,3-butanediol pathway intermediate can be directly added to another culture of the second organism, or the original culture of the 1,3-butanediol pathway intermediate producer can be depleted, for example by cell separation, and then the second organism can be subsequently added to the fermentation broth to produce the final product without intermediate purification.

[0245] In other embodiments, the non-naturally occurring microorganisms and methods of the present invention can be combined in a variety of sub-pathways to obtain, for example, the biosynthesis of 1,3-butanediol. In these embodiments, the biosynthetic pathways for the desired product of the present invention can be isolated into different microorganisms, and these different microorganisms can be co-cultured to produce the final product. In such a biosynthetic scheme, the product of one microorganism is used as a substrate for a second microorganism until the final product is synthesized. For example, the biosynthesis of 1,3-butanediol can be accomplished by constructing microorganisms comprising a biosynthetic pathway for converting an intermediate of one pathway into an intermediate or product of another pathway. Alternatively, 1,3-butanediol can also be produced from microbial biosynthesis by co-culturing or co-fermenting two organisms in the same vessel, wherein the first microorganism produces a 1,3-butanediol intermediate and the second microorganism converts the intermediate into 1,3-butanediol.

[0246] In light of the teachings and guidance provided herein, those skilled in the art will understand that there are various combinations and arrangements of the non-naturally occurring microorganisms and methods described herein, together with other microorganisms, co-cultures of other non-naturally occurring microorganisms with sub-pathways, and other chemical and / or biochemical procedures well known in the art for the production of 1,3-butanediol.

[0247] Sources of nucleic acids encoding enzymes or proteins for the 1,3-butanediol pathway can include, for example, any species in which the encoded gene product is capable of catalyzing a reference reaction. Such species include both prokaryotes and eukaryotes, including but not limited to bacteria (including archaea and eubacteria) and eukaryotes (including yeast), plants, insects, animals, and mammals (including humans). Exemplary species for such sources include, for example, *Escherichia coli* and other exemplary species disclosed herein or available as source organisms for the corresponding genes. However, with complete genome sequences of over 550 species available today (more than half of which are available in public databases such as NCBI), including 395 microbial genomes and various yeast, fungal, plant, and mammalian genomes, the identification of genes encoding 1,3-butanediol biosynthetic activity against one or more genes in closely or distantly related species (including, for example, syndrome, ortholog, paralog, and non-syndrome gene substitutions of known genes) and the exchange of gene alterations between organisms are common and well-known in the art. Therefore, the metabolic alterations described herein for the biosynthesis of 1,3-butanediol in specific organisms (such as *Escherichia coli*) can be readily applied in the same manner to other microorganisms, including prokaryotes and eukaryotes. In light of the teachings and guidance provided herein, those skilled in the art will recognize that the metabolic alterations demonstrated in one organism can be applied in other organisms in the same way.

[0248] In certain circumstances, such as when alternative 1,3-butanediol biosynthetic pathways exist in unrelated species, 1,3-butanediol biosynthesis in a host species can be conferred by exogenous expression, for example, from a paralog or multiple paralogs of an unrelated species that replace the reference reaction with a catalytically similar but not identical metabolic reaction. Because of certain differences in metabolic networks between different organisms, those skilled in the art will understand that actual gene usage can vary between different organisms. However, in light of the teachings and guidance provided herein, those skilled in the art will also understand that the teachings and methods of this invention can be applied to all microorganisms by utilizing homologous metabolic alterations made to those exemplary microorganisms herein to construct 1,3-butanediol-synthesizing microorganisms in the relevant species.

[0249] The host microorganism can be selected from, and the non-naturally occurring microorganism can be generated from, for example, bacteria, yeast, fungi, or various other microorganisms suitable for fermentation processes. Exemplary bacteria include species selected from: *Escherichia coli*, *Klebsiella pneumoniae*, *Succinate-producing anaerobic spirochetes*, *Succinate-producing Actinobacillus*, *Succinate-producing Mansonia*, *Rhizobium praecoxibrio*, *Bacillus subtilis*, *Corynebacterium glutamicum*, *Glucosamine oxidase*, *Fermentomonas motilityis*, *Lactococcus lactis*, *Lactobacillus plantarum*, *Streptomyces cerevisiae*, *Clostridium acetobutyrolactone*, *Pseudomonas fluorescens*, and *Pseudomonas putida*. Exemplary yeasts or fungi include species selected from: *Saccharomyces cerevisiae*, *Human yeast probe*, *Kluyveromyces lactis*, *Kluyveromyces marxi*, *Aspergillus terreus*, *Aspergillus niger*, and *Pichia pastoris*. *Escherichia coli* is a particularly useful host organism because it is a well-characterized microorganism suitable for genetic engineering. Other particularly useful host organisms include yeasts, such as *Saccharomyces cerevisiae*.

[0250] Methods for constructing and testing the expression levels of hosts that produce non-naturally occurring 1,3-butanediol can be operated, for example, using recombination and detection methods well known in the art. Descriptions of such methods can be found, for example, Sambrook et al., Molecular Cloning: A Laboratory Manual, 3rd Edition, Cold Spring Harbor Laboratory, New York (2001) and Ausubel et al., Current Protocols in Molecular Biology, John Wiley and Sons, Baltimore, MD (1999).

[0251] Exogenous nucleic acid sequences involved in the pathway for the production of 1,3-butanediol can be stably or transiently introduced into host cells using techniques well-known in the art (including but not limited to conjugation, electroporation, chemical transformation, transduction, transfection, and sonication). For exogenous expression in *E. coli* or other prokaryotic cells, some nucleic acid sequences in the gene or cDNA of the eukaryotic nucleic acid may encode a targeting signal, such as an N-terminal mitochondrial targeting signal or other targeting signals, which can be removed before conversion into the prokaryotic host cell if desired. For example, removal of the mitochondrial leader sequence leads to upregulation of expression in *E. coli* (Hoffmeister et al., J. Biol. Chem. 280:4329-4338 (2005)). For exogenous expression in yeast or other eukaryotic cells, the gene can be expressed in the cytosol without the addition of a leader sequence; or the gene can be directed to mitochondria or other organelles by adding a suitable targeting sequence appropriate for the host cell (such as a mitochondrial targeting or secretion signal). Therefore, it is understood that appropriate modifications to nucleic acid sequences, such as the removal or inclusion of guide sequences, can be incorporated into exogenous nucleic acid sequences to provide the desired properties. Furthermore, techniques well-known in the art can enable codon optimization of genes to achieve optimized protein expression.

[0252] Expression vectors can be constructed to include one or more 1,3-butanediol biosynthetic pathway-encoding nucleic acids operably linked to a functional expression regulatory sequence in a host organism, as demonstrated herein. Expression vectors suitable for host microorganisms of the present invention include, for example, plasmids, phage vectors, viral vectors, episomes, and artificial chromosomes (including vectors operably used for stable integration into the host chromosome and selectable sequences or markers). Furthermore, the expression vector may include one or more selectable marker genes and appropriate expression regulatory sequences. It may also include, for example, selectable marker genes that provide antibiotic or toxin resistance, supplement nutrient deficiencies, or provide key nutrients not present in the culture medium. Expression regulatory sequences may include constitutive and inducible promoters, transcription enhancers, transcription terminators, and other expression regulatory sequences well known in the art. When two or more exogenous encoding nucleic acids are co-expressed, both nucleic acids can be inserted, for example, in a single expression vector or independently. For single-vector expression, the encoding nucleic acid can be operably linked to a common expression regulatory sequence or to different expression regulatory sequences, such as an inducible promoter and a constitutive promoter. The transformation of exogenous nucleic acid sequences involved in metabolic or synthetic pathways can be determined using methods well-known in the art. Such methods include, for example, nucleic acid analysis (such as Northern blotting or mRNA polymerase chain reaction (PCR) amplification or gene product expression immunoblotting or other suitable assays) to analyze the expression of the introduced nucleic acid sequence or its corresponding gene product. Those skilled in the art understand that sufficient amounts of exogenous nucleic acid can be expressed to produce the desired product, and further understand that sufficient expression levels can be optimized using methods well-known in the art and disclosed herein.

[0253] The present invention provides a method for producing 1,3-BDO, the method comprising culturing the non-naturally occurring microorganisms disclosed herein under various conditions for a sufficient time to produce 1,3-BDO, the non-naturally occurring microorganisms including microorganisms incorporating one, two, three, four, five to all exogenous nucleic acids encoding enzymes that complete the 1,3-BDO pathway. The 1,3-BDO pathway comprises a set of 1,3-BDO pathway enzymes, identified as follows: (a)(1) 2-amino-4-ketovalerate (AKP) thiolylase; (2) AKP dehydrogenase; (3) 2-amino-4-hydroxyvalerate transaminase or oxidoreductase (deamination); (4) 2-oxo-4-hydroxyvalerate decarboxylase; and (5) 3-hydroxybutyraldehyde reductase; (b)(1) 2-amino-4-ketovalerate (AKP) thiolylase; (2) AKP transaminase or oxidoreductase (deamination); (3) 2,4-dioxovalerate decarboxylase; and (4) 3-oxobutyraldehyde. (c) (1) 2-amino-4-ketovalerate (AKP) thiolyase; (2) AKP transaminase or oxidoreductase (deamination); (3) 2,4-dioxovalerate decarboxylase; (4) 3-oxobutyraldehyde reductase (aldehyde reduction); and (5) 4-hydroxy-2-butanone reductase; (d) (1) 2-amino-4-ketovalerate (AKP) thiolyase; (2) AKP decarboxylase; (3) 4-aminobutyraldehyde transaminase or oxidoreductase (deamination); (4) 3-oxobutyraldehyde reductase (keto reduction); and (5) 3-hydroxybutyraldehyde reductase; (e)(1) 2-amino-4-ketovalerate (AKP) thiolyase; (2) AKP decarboxylase; (3) 4-aminobutyric acid-2-ketotransaminase or oxidoreductase (deamination); (4) 3-oxobutyraldehyde reductase (aldehyde reduction); and (5) 4-hydroxy-2-butanone reductase; (f)(1) 2-amino-4-ketovalerate (AKP) thiolyase; (2) AKP decarboxylase; (3) 4-aminobutyric acid-2-ketoamine-lyase; (4) butanone hydratase; and (5) 4-hydroxy-2-butanone reductase; (g)(1) 2-amino-4-ketovalerate (AKP) thiolyase; (2) AKP ammonia-lyase; (3) 4-aminobutyric acid-2-ketoamine-lyase; (4) butanone hydratase; and (5) 4-hydroxy-2-butanone reductase; (g)(1) 2-amino-4-ketovalerate (AKP) thiolyase; (2) AKP ammonia-lyase; (3) 4-aminobutyric acid-2-ketoamine-lyase; (4) 3-oxobutyraldehyde reductase; and (5) 4-hydroxy-2-butanone reductase; (4) Acetylacetyl decarboxylase; and (5) 4-hydroxy-2-butanone reductase; (h) (1) Acetylacetyl-CoA reductase (CoA-dependent, aldehyde formation); (2) 3-oxobutanone reductase (ketone reduction); and (3) 3-hydroxybutanone reductase; (i) (1) Acetylacetyl-CoA reductase (CoA-dependent, alcohol formation) and (2) 4-hydroxy-2-butanone reductase; (j) (1) Acetylacetyl-CoA reductase (CoA-dependent, aldehyde formation); (2) 3-oxobutanone reductase (aldehyde reduction); and (3) 4-hydroxy-2-butanone reductase;(k)(1) acetyl-CoA reductase (ketone reduction) and (2) 3-hydroxybutyryl-CoA reductase (alcohol formation); (l)(1) acetyl-CoA reductase (ketone reduction); (2) 3-hydroxybutyryl-CoA reductase (aldehyde formation); and (3) 3-hydroxybutyraldehyde reductase; (m)(1) 4-hydroxybutyryl-CoA dehydratase; (2) crotonylase; and (3) 3-hydroxybutyryl-CoA reductase (alcohol formation); and (n)(1) 4-hydroxybutyryl-CoA dehydratase; (2) crotonylase; (3) 3-hydroxybutyryl-CoA reductase (aldehyde formation); and (4) 3-hydroxybutyraldehyde reductase.

[0254] Purification and / or determination for 1,3-butanediol production assays can be performed using well-known methods. Suitable replicas, such as triplicate cultures, can be cultivated for each designed test strain. For example, the formation of products and byproducts in the designed production host can be monitored. The final product, intermediates, and other organic compounds can be analyzed using various methods, such as HPLC (High Performance Liquid Chromatography), GC-MS (Gas Chromatography-Mass Spectrometry), and LC-MS (Liquid Chromatography-Mass Spectrometry), or other suitable analytical methods utilizing routine procedures well known in the art. The release of products from the fermentation broth can also be tested using the culture supernatant. Byproducts and residual glucose can be quantified using HPLC or other suitable assays and detection methods well known in the art, such as refractive index detectors for glucose and alcohols and ultraviolet detectors for organic acids (Lin et al., Biotechnol. Bioeng. 90:775-779 (2005)). Individual enzyme or protein activities derived from exogenous DNA sequences can also be determined using methods well known in the art (see, for example, WO / 2008 / 115840 and Hanai et al., Appl Environ. Microbiol. 73:7814-7818 (2007)).

[0255] 1,3-Butanediol can be separated from other components in a culture using various methods well known in the art. Such separation methods include, for example, extraction procedures and methods including: continuous liquid-liquid extraction, total evaporation, membrane filtration, membrane separation, reverse osmosis, electrodialysis, distillation, crystallization, centrifugation, extractive filtration, ion exchange chromatography, size exclusion chromatography, adsorption chromatography, and ultrafiltration. All of the above methods are well known in the art.

[0256] Any non-naturally occurring microorganism described herein can be cultured to produce and / or secrete the biosynthetic products of this invention. For example, 1,3-butanediol producers can be cultured for the biosynthetic production of 1,3-butanediol.

[0257] To produce 1,3-butanediol, a recombinant bacterial strain is cultured in a medium containing a carbon source and other essential nutrients. Maintaining anaerobic conditions in the fermenter to reduce consumption throughout the process is highly desirable. Such conditions can be achieved, for example, by first spraying the medium with nitrogen and then sealing the flask with a gasket and clamp cap. For strains that do not follow anaerobic conditions, microaerobic conditions can be employed by making a small hole in the gasket for limited aeration. Exemplary anaerobic conditions have been previously described and are well known in the art. Descriptions of exemplary aerobic and anaerobic conditions can be found, for example, in U.S. Publication No. US-2009-0047719, filed August 10, 2007. As disclosed herein, fermentation can be carried out in batch, fed-batch, or continuous manner.

[0258] If necessary, the pH of the culture medium can be maintained at the desired pH, particularly by adding alkali (such as NaOH or other bases) or acid to maintain a neutral pH (such as a pH of about 7) as needed. The growth rate can be determined by measuring optical density using a spectrophotometer (600 nm), and the glucose uptake rate can be determined by monitoring the consumption of carbon sources over time.

[0259] In addition to the exemplary renewable feedstocks described above, the 1,3-butanediol microorganisms of this invention can also be modified to be cultivated using syngas as their carbon source. In this particular embodiment, one or more proteins or enzymes are expressed in the 1,3-butanediol-producing organism to provide a metabolic pathway for utilizing syngas or other gaseous carbon sources.

[0260] The organisms used in this invention can be utilized, for example, any carbohydrate source that can provide carbon to non-naturally occurring microorganisms. Such sources include, for example, sugars (such as glucose, xylose, arabinose, galactose, mannose, fructose, and starch). Other carbohydrate sources include, for example, renewable feedstocks and biomass. Exemplary types of biomass that can be used as feedstocks in the methods described in this invention include cellulosic biomass, hemicellulose biomass, and lignin feedstocks or lignin fractions of feedstocks. Such biomass feedstocks contain, for example, carbohydrate substrates used as carbon sources (such as glucose, xylose, arabinose, galactose, mannose, fructose, and starch). In light of the teachings and guidance provided herein, those skilled in the art will understand that, in addition to the exemplary feedstocks and biomass described above, renewable feedstocks and biomass can also be used to cultivate the microorganisms of this invention for the production of 1,3-butanediol.

[0261] Therefore, in light of the teachings and guidance provided herein, those skilled in the art will understand that non-naturally occurring microorganisms can be produced that, when cultured on a carbon source (such as syngas, CO, and / or CO2), secrete the biosynthesized compounds of the present invention. Such compounds include, for example, 1,3-butanediol and any intermediate metabolites in the 1,3-butanediol pathway. All that is required is to design one or more desired enzyme or protein activities to achieve the biosynthesis of the desired compound or intermediate (including, for example, some or all inclusions of the 1,3-butanediol biosynthetic pathway). Therefore, the present invention provides a non-naturally occurring microorganism that, when cultured on carbohydrates or other carbon sources, produces and / or secretes 1,3-butanediol and, when cultured on carbohydrates or other carbon sources, produces and / or secretes any intermediate metabolites shown in the 1,3-butanediol pathway. The 1,3-butanediol-producing microorganism of the present invention can facilitate the synthesis of intermediates (e.g., acetyl-CoA).

[0262] The non-naturally occurring microorganisms of the present invention are constructed using methods well known in the art, as demonstrated herein, to exogenously express at least one nucleic acid encoding a 1,3-butanediol pathway enzyme or protein in sufficient quantities to produce 1,3-butanediol. It is understood that the microorganisms of the present invention are cultured under conditions sufficient for the production of 1,3-butanediol. Following the teachings and guidance provided herein, the non-naturally occurring microorganisms of the present invention can achieve the biosynthesis of 1,3-butanediol, producing intracellular concentrations of about 0.1-2000 mM or higher. Generally, intracellular concentrations of 1,3-butanediol are between about 3-1800 mM, particularly between about 5-1700 mM and more particularly between about 8-1600 mM, including about 100 mM, 200 mM, 500 mM, 800 mM or higher. Intracellular concentrations between each of these exemplary ranges and higher can also be obtained from the non-naturally occurring microorganisms of the present invention.

[0263] In some embodiments, the culture conditions include anaerobic or substantially anaerobic culture or maintenance conditions. Exemplary anaerobic conditions have been previously described and are well known in the art. Descriptions of exemplary anaerobic conditions for fermentation processes are found herein and, for example, in U.S. Patent Application No. 2009 / 0047719, filed August 10, 2007. Any of these conditions can be used with the aforementioned non-naturally occurring microorganisms and other anaerobic conditions well known in the art. Under such anaerobic conditions, the 1,3-butanediol producer can synthesize 1,3-butanediol at intracellular concentrations of 5-10 mM or higher, as well as all other concentrations exemplified herein. It is understood that, although the above description refers to intracellular concentrations, the 1,3-butanediol-producing microorganisms can produce 1,3-butanediol intracellularly and / or secrete products into the culture medium.

[0264] Culture conditions may include, for example, liquid culture procedures as well as fermentation and other large-scale culture procedures. As described herein, particularly useful yields of the biosynthetic products of the present invention can be obtained under anaerobic or substantially anaerobic culture conditions.

[0265] As described herein, exemplary culture conditions for achieving 1,3-butanediol biosynthesis include anaerobic culture or fermentation conditions. In some embodiments, the non-naturally occurring microorganisms of the present invention can be maintained, cultured, or fermented under anaerobic or substantially anaerobic conditions. Briefly, anaerobic conditions refer to an oxygen-deficient environment. Substantially anaerobic conditions include, for example, batch or continuous fermentation of the culture medium so that the dissolved oxygen concentration in the culture medium is maintained between 0 and 10% saturation. Substantially anaerobic conditions also include culturing or statically incorporating cells in a liquid culture medium or on solid agar within a sealed chamber maintained in an atmosphere of less than 1% oxygen. The percentage of oxygen can be maintained, for example, by spraying the culture medium with an N2 / CO2 mixture or one or more other suitable non-oxygen gases.

[0266] The culture conditions described herein can be scaled up and continuously increased for the production of 1,3-butanediol. Exemplary culture procedures include, for example, fed-batch fermentation and batch separation; fed-batch fermentation and continuous separation; or continuous fermentation and continuous separation. All of these processes are well known in the art. Fermentation procedures are particularly useful for the industrial-scale biosynthetic production of 1,3-butanediol. Generally, and as with discontinuous culture procedures, continuous and / or near-continuous production of 1,3-butanediol will involve culturing the present invention’s organism for the production of non-naturally occurring 1,3-butanediol in sufficient nutrients and culture medium to maintain and / or near-maintain exponential growth. Continuous culture under such conditions may include, for example, 1 day, 2, 3, 4, 5, 6, or 7 days or more. Furthermore, continuous culture may include 1 week, 2, 3, 4, or 5 or more weeks or even months. Alternatively, the present invention’s organism may be cultured for several hours, depending on the specific application. It is understood that continuous and / or near-continuous culture conditions may also include all time intervals between these exemplary cycles. It is further understood that the cultivation time of the microorganisms described in this invention is a time period sufficient to produce a sufficient quantity of the product for the desired purpose.

[0267] Fermentation procedures are well known in the art. In short, fermentation for the biosynthetic production of 1,3-butanediol can be performed, for example, fed-batch fermentation and batch separation; fed-batch fermentation and continuous separation; or continuous fermentation and continuous separation. Examples of batch and continuous fermentation procedures are well known in the art.

[0268] In addition to the fermentation process described above for the large-scale continuous production of 1,3-butanediol by the 1,3-butanediol producer of the present invention, the 1,3-butanediol producer may also, for example, undergo a simultaneous chemical synthesis process to convert the product into other compounds or the product may be isolated from the fermentation culture and subsequently undergo chemical transformation to convert the product into other compounds (if necessary).

[0269] In some implementations, syngas can be used as a carbon feedstock. Key considerations for syngas fermentation processes include high biomass concentration and good gas-liquid mass transfer (Bredwell et al., Biotechnol Prog., 15:834-844 (1999)). CO has a slightly lower solubility in water than oxygen. Continuous gas jet fermentation can be carried out in a controlled fermenter, with constant flue gas analysis by mass spectrometry and periodic liquid sampling and analysis by GC and HPLC. The liquid phase can operate in batch mode. For example, using… A series of HPLC columns (e.g., HPX-87 series) (BioRad, Hercules CA) were used, and fermentation products (such as alcohols, organic acids, and residual glucose along with residual ethanol) were quantified by HPLC (Shimadzu, Columbia MD) using refractive index detectors for glucose and alcohols and UV detectors for organic acids. The growth rate could be determined by measuring optical density using a spectrophotometer (600 nm). All tubing in these systems was made of glass or metal to maintain anaerobic conditions. Glass powder was used for gas injection to reduce bubble size and improve mass transfer. Various injection rates were tested, ranging from approximately 0.1 to 1 vvm (vapor volume per minute). To obtain accurate measurements of the gas uptake rate, periodic queries were performed, in which the gas flow was temporarily stopped, and the gas phase composition was monitored as a function of time.

[0270] To achieve the overall target yield, cell retention or recycling methods are employed. One method to increase microbial concentration is cell recycling via a tangential flow membrane relative to the side flow. Repeated batch cultures can also be utilized, as previously described for acetic acid production via *M. molluscum* (Sakai et al., J Biosci. Bioeng, 99:252-258 (2005)). Various other methods can also be used (Bredwell et al., Biotechnol Prog., 15:834-844 (1999); Datar et al., Biotechnol Bioeng, 86:587-594 (2004)). Other optimization tests can be performed, for example, at a high pressure of 1.5 atm to improve mass transfer (Najafpour et al., Enzyme and Microbial Technology, 38[l-2], 223-228 (2006)).

[0271] Once satisfactory efficiency is achieved using pure H2 / CO as feedstock, the resulting syngas mixture contains inhibitors that may be present in commercial syngas. For example, a typical impurity profile is 4.5% CH4, 0.1% C2H2, 0.35% C2H6, 1.4% C2H4, and 150 ppm nitric oxide (Datar et al., Biotechnol Bioeng, 86:587-594 (2004)). Tar (representative compounds such as benzene, toluene, ethylbenzene, p-xylene, o-xylene, and naphthalene) is added at ppm levels to test any effects on production. For example, 40 pp NO has been shown to be inhibitory on C. carboxidivorans (Ahmed et al., Biotechnol Bioeng, 97:1080-1086 (2007)). Cultures are tested in shake flasks before being transferred to fermenters. Additionally, different levels of these potential inhibitory compounds are tested to quantify their effects on cell growth. This knowledge is used to establish standards for syngas purity and apply them to large-scale research and production. If any specific component is found to be difficult to reduce or remove from syngas, adaptive evolutionary processes are employed to gradually make the cells tolerate one or more impurities.

[0272] Advances in protein engineering have made it feasible to modify any of the enzymes disclosed herein to act effectively on substrates that are not yet natural to the enzyme. Examples of enzymes with broad specificity from different relevant levels and methods used to evolve such enzymes to act on unnatural substrates are given below.

[0273] One level of enzyme in the pathway disclosed in this paper is the oxidoreductase (1.1.1) that performs the interconversion of ketones or aldehydes with alcohols. This level of enzyme can function on a wide variety of substrates. It was shown that the alcohol dehydrogenase (1.1.1.1) purified from the soil bacterium *Brevibacterium breve* KU 1309 (Hirano et al., J. Biosci. Bioeng. 100:318-322 (2005)) functions against excess fats and highly reactive aromatic alcohols. Table 33 shows the activity of this enzyme and its K+ for different alcohols. m This enzyme is reversible and exhibits very high activity against a variety of aldehydes, as shown in Table 34.

[0274] Table 33

[0275] Substrate Relative activity (%) <![CDATA[K M (MM)]]> 2-Phenylenol 100 0.025 (S)-2-Phenylacetol 156 0.157 (R)-2-Phenylacetol 63 0.020 benzyl alcohol 199 0.012 3-Phenylacetol 135 0.033 ethanol 76 1-Butanol 111 1-Octanol 101 1-Dodecanool 68 1-Phenylephethanol 46 2-Propanol 54

[0276] In this table, the activity of 2-phenylethanol (corresponding to 19.2 U / mg) is considered to be 100%.

[0277] Table 34

[0278]

[0279] Lactate dehydrogenase (1.1.1.27) from *Ralstonia eutropha* is another enzyme that has been shown to have high activity against a variety of 2-oxoacids, such as 2-oxobutyric acid, 2-oxovalerate, and 2-oxoglutarate (a C5 compound similar to 2-oxoadipic acid) (Steinbuchel et al., ibid.). Column 2 of Table 35 shows the activity of ldhA from *Ralstonia eutropha* (formerly *Alcaligenes*) against different substrates (Steinbuchel et al., ibid.).

[0280] Table 35

[0281]

[0282] It has been shown that oxidoreductases that can convert 2-oxoacids to their acyl-CoA counterparts (1.2.1) also accept multiple substrates. For example, the branched-chain 2-keto-acid dehydrogenase complex (BCKAD) (also known as 2-oxoisovalerate dehydrogenase (1.2.1.25)) participates in the branched-chain amino acid degradation pathway, converting 2-keto acid derivatives of valine, leucine, and isoleucine to their acyl-CoA derivatives and CO2. In some organisms (including brown rats (Paxton et al., Biochem. J. 234:295-303 (1986)) and Saccharomyces cerevisiae (Sinclair et al., Biochem. Mol. Biol. Int. 31:911-922 (1993)), this complex has been shown to have a broad substrate range, including straight-chain oxoacids such as 2-oxobutyric acid and α-ketoglutarate, in addition to branched-chain amino acid precursors.

[0283] It has been reported that another class of enzymes, namely transaminases (2.6.1), act on multiple substrates. Aspartate transaminase (aspAT) from Thermococcus fibrillans has been discovered, characterized in Escherichia coli, and recombinant protein characterization has demonstrated that this enzyme has the highest activity for aspartate and α-ketoglutarate, but lower but significant activity for alanine, glutamate, and aromatic amino acids (Ward et al., Archaea. 1:133-141 (2002)). In another instance, a transaminase found in Leishmania from Mexico and characterized in Escherichia coli (Vernal et al., FEMS Microbiol. Lett. 229:217-222 (2003)) has been reported to have broad substrate specificity, specifically for tyrosine (100% activity is considered to be for tyrosine), phenylalanine (90%), tryptophan (85%), aspartic acid (30%), leucine (25%), and methionine (25%) (Vernal et al., Mol. Biochem. Parasitol. 96:83-92 (1998)). It has been reported that, although the sequence homology between these two enzymes is only 6%, tyrosine transaminases from Trypanosoma cruzi exhibit similar broad specificity. Note that the latter enzyme can accept leucine, methionine, as well as tyrosine, phenylalanine, tryptophan, and alanine as highly efficient amino donors (Nowicki et al., Biochim. Biophys. Acta 1546:268-281 (2001)).

[0284] In contrast to the enzymes described in these examples, which naturally possess broad substrate specificity, many enzymes have been modified using directed evolution to broaden their specificity for their unnatural substrates. Alternatively, the substrate preference of enzymes has also been altered using directed evolution. For example, the enantioselectivity of a lipase from *Pseudomonas aeruginosa* has been reported to be significantly improved. This enzyme hydrolyzes p-nitrophenyl-2-methyldecanoic acid with only a 2% enantiomeric excess (ee) favoring the (S)-acid. However, after four consecutive rounds of error-prone mutagenesis and screening, a variant was produced that catalyzes the essential reaction with 81% ee (Reetz et al., Angew. Chem. Int. Ed Engl. 36:2830-2832 (1997)).

[0285] Directed evolution has made it possible to modify enzymes to act on a wide range of unnatural substrates. The substrate specificity of lipases in *Pseudomonas aeruginosa* was amplified by randomizing amino acid residues near the active site. This allowed the enzyme to accept α-substituted hydroxy esters (Reetz et al., Angew. Chem. Int. Ed Engl. 44:4192-4196 (2005)). In another successful attempt, DNA shuffling was used to produce *E. coli* transaminases that accepted β-branched substrates that were poorly accepted by wild-type enzymes (Yano et al., Proc. Natl. Acad. Sci. USA 95:5511-5515 (1998)). Specifically, at the end of four rounds of shuffling, the activity of aspartate transaminase for valine and 2-oxovaline increased by five orders of magnitude, while the activity for the natural substrate aspartate decreased by more than 30-fold. Recently, an algorithm has been used to design retro-aldolases that catalyze the breaking of carbon-carbon bonds in the unnatural and abiotic substrate 4-hydroxy-4-(6-methoxy-2-naphthyl)-2-butanone. These algorithms utilize different combinations of four distinct catalytic motifs to design novel enzymes, and 20 selected designs for experimental characterization showed a fourfold improvement compared to the uncatalyzed reaction (Jiang et al., Science 319:1387-1391 (2008)). Therefore, these engineering methods not only expand the array of substrates on which enzymes can function but also allow for the design and construction of highly efficient enzymes. For example, a DNA shuffling method (temporary template random chimeric growth or RACHITT) has been reported to produce a designed monooxygenase with improved desulfurization of complex substrates and a 20-fold improvement in the conversion rate of unnatural substrates (Coco et al., Nat. Biotechnol. 19:354-359 (2001)). Similarly, the specific activity of the slow-mutated triose phosphate isomerase was improved by 1.3 to 19 times (Hermes et al., Proc. Natl. Acad. Sci. USA 87:696-700 (1990)). The increase in specific activity was achieved by using random mutagenesis over the entire length of the protein, and this improvement can be traced back to a mutation in six amino acid residues.

[0286] The effectiveness of protein engineering methods in altering the substrate specificity of enzymes for desired substrates has also been demonstrated. Isopropyl malate dehydrogenase from *Thermophilus* was modified by altering residues near its active site so that it can now act on malate and D-lactic acid as substrates (Fujita et al., *Biosci. Biotechnol Biochem.* 65:2695-2700 (2001)). This study, along with others, has shown that one or more residues can be modified to alter substrate specificity. A relevant example is dihydroflavonol 4-reductase, whose single amino acid was altered in the putative substrate-binding region, and which preferentially reduces dihydrokaempferol (Johnson et al., *Plant J.* 25:325-333 (2001)). The substrate specificity of a highly specific isocitrate dehydrogenase from *E. coli* is achieved by altering a single residue at its active site from isocitrate to isopropyl malate (Doyle et al., *Biochemistry* 40:4234-4241 (2001)). Similarly, NAD... + The cofactor specificity of 1,5-hydroxyprostaglandin dehydrogenase is altered to NADP by changing several residues near the N-terminus. + (Cho et al., Arch. Biochem. Biophys. 419:139-146 (2003)). Sequence analysis and molecular modeling analysis were used to identify key residues for modification, which were further investigated through site-directed mutagenesis.

[0287] Fucosidase was evolved from galactosidase in *E. coli* through DNA shuffling and screening (Zhang et al., *Proc Natl Acad Sci US.A 94:4504-4509* (1997)). Similarly, aspartate transaminase from *E. coli* was converted to tyrosine transaminase using homology modeling and site-directed mutagenesis (Onuffer et al., *Protein Sci. 4:1750-1757* (1995)). It has been reported that site-directed mutagenesis of two residues in the active site of benzoylformate decarboxylase from *Pseudomonas putida* altered its affinity for both natural and unnatural substrates (K). m(Siegert et al., Protein EngDes Sel 18:345-357 (2005)). Cytochrome C peroxidase (CCP) from Saccharomyces cerevisiae underwent directed molecular evolution to generate mutants with increased activity against the traditional peroxidase substrate guaiacol, thereby changing the substrate specificity of CCP from protein cytochrome c to small organic molecules. After three rounds of DNA shuffling and screening, mutants with 300-fold increased activity against guaiacol and up to 1000-fold increased specificity against this substrate compared to the natural substrate were isolated (Iffland et al., Biochemistry 39:10790-10798 (2000)).

[0288] In some cases, enzymes with substrate preference different from their amphiphilic counterparts have been obtained. For example, the degradation of biphenyl-dioxygenase-mediated polychlorinated biphenyls (PCBs) has been improved by reprogramming genes from two bacterial species, *Pseudomonas alkaligenes* and *Burkholderia cepacia* (Kumamaru et al., *Nat. Biotechnol* 16, 663-666 (1998)). The resulting chimeric biphenyl oxygenases exhibited substrate preference different from their amphiphilic counterparts and improved activity for the degradation of associated benzene compounds and monocyclic aromatic hydrocarbons (PAHs), which were initially undesirable substrates for the enzyme.

[0289] This could not only alter enzyme specificity but also potentially enhance the activity of enzymes with low activity for substrates in their natural state. One study demonstrated that random mutagenesis could significantly improve the activity of racemic amino acid enzymes from *Pseudomonas putida*, which have broad substrate specificity (especially for lysine, arginine, alanine, serine, methionine, cysteine, leucine, and histidine) but low activity for tryptophan (Kino et al., *Appl. Microbiol. Biotechnol.* 73:1299-1305 (2007)). Similarly, the active site of bovine BCKAD was engineered to support the alternative substrate acetyl-CoA (Meng et al., *Biochemistry* 33:12879-12885 (1994)). An interesting aspect of these methods is that even when these mutant enzymes with potent activity have been generated using random methods, the exact mutation or structural change conferring the activity improvement can be identified. For example, in the aforementioned study, the mutation promoting improved activity for tryptophan could be traced back to two distinct locations.

[0290] Directed evolution has also been used to express proteins that are difficult to express. For example, by subjecting horseradish peroxidase to random mutagenesis and gene recombination, mutants with an activity more than 14 times that of the wild type can be extracted (Lin et al., Biotechnol. Prog. 15:467-471 (1999)).

[0291] The final example of directed evolution is extensive modification to obtain a range of desired functions. An enzyme, lactate dehydrogenase from *Bacillus stearothermophilus*, was subjected to site-directed mutagenesis and three amino acid substitutions at sites identified to determine specificity for different hydroxy acids (Clarke et al., *Biochem. Biophys. Res. Commun.* 148:15-23 (1987)). Following these mutations, the specificity for oxaloacetate compared to pyruvate increased to 500, compared to the wild-type enzyme's catalytic specificity of 1000 for pyruvate compared to oxaloacetate. The enzyme was further engineered using site-directed mutagenesis to possess activity against branched-chain substituted pyruvate (Wilks et al., *Biochemistry 29:8587-8591 (1990)). Specifically, the enzyme showed activity against α-ketoisocaprate K+. cat The improvement was 55-fold. Three structural modifications were made to the same enzyme to change its substrate specificity from lactate to malate. This enzyme exhibited high activity and specificity for malate (Wilks et al., Science 242:1541-1544 (1988)). Subsequently, the same enzyme from *Bacillus stearothermophilus* was designed to exhibit high catalytic activity for α-keto acids with positively charged side chains (such as those containing ammonium groups) (Hogan et al., Biochemistry 34:4225-4230 (1995)). A mutant containing an acidic amino acid was introduced at position 102 of the enzyme to support the binding of this side-chain ammonium group. The results showed that this mutant exhibited kJ / kJ / kJ for ω-amino-α-keto acid substrates. cat / K m The value was improved by up to 25-fold. This enzyme was also structurally modified to function as a phenyllactic acid dehydrogenase rather than a lactate dehydrogenase (Wilks et al., Biochemistry 31:7802-7806 (1992)). A restriction site was introduced into the gene targeting this enzyme, allowing for the excision of a gene region. This region encodes a mobile surface loop (residues 98-110) of the polypeptide, which typically seals the active site vacuoles from a large amount of solvent and is a major determinant of substrate specificity. Variable length and sequence loops were inserted into the excision gene and used to synthesize hydroxy acid dehydrogenases with altered substrate specificity. In the case of constructing a longer loop, the activity against pyruvate was reduced by a million-fold, but the activity against phenyllactic acid remained largely unchanged. A specificity of 390,000-fold (K0.05) was achieved. cat / K m The enzyme's 1700:1 selectivity for phenylpyruvate versus pyruvate is exactly what phenyllactate dehydrogenase requires.

[0292] As noted above, directed evolution is a powerful method involving the introduction of mutations into target-specific genes to improve and / or alter the properties of enzymes. It can be achieved by developing and implementing mutations that allow for numerous enzyme variants (e.g., >10). 4 High-throughput screening assays with high sensitivity are used to identify improved and / or altered enzymes. Iterative rounds of mutagenesis and screening are typically performed to optimize enzyme properties. Computational algorithms that can help identify gene regions to be mutagenized have also been developed and can significantly reduce the number of enzyme variants that need to be generated and screened.

[0293] Numerous directed evolution techniques have been developed (see Hibbert, EG, F. Baganz, HC Hailes, JMWard, GJLye, JMWoodley, and PADalby, 2005, Directed evolution of biocatalytic processes. Biomol. Eng 22:11-19; Huisman, GW and JJLalonde, 2007, Enzyme evolution for chemical process applications, pp.717-742. In RNPatel (ed.), Biocatalysis in the pharmaceutical and biotechnology industries. CRC Press; Otten, LG and WJQuax, 2005. Directed evolution: selecting today's biocatalysts. Biomol. Eng 22:1-9; and Sen, S., D. Venkata, V, and B. Mandal, 2007, Developments in directed evolution for improving enzyme functions. Appl. Biochem. Biotechnol). (143:212-223) to effectively create diverse variant libraries and these methods have been successfully applied to improve various properties across many enzyme grades.

[0294] Enzyme properties improved and / or altered through directed evolution include, for example, selectivity / specificity – for the transformation of unnatural substrates; temperature stability – for robust high-temperature treatments; pH stability – for bioprocesses under low or high pH conditions; substrate or product tolerance – to enable high product titers; and binding (K+) properties. m - Expanded substrate binding to include unnatural substrates; inhibitory (K) i – To remove inhibition via product, substrate, or key intermediate; activity (k cat - Increase the rate of enzyme-catalyzed reactions to achieve the desired throughput; expression level - increase protein yield and overall pathway throughput; oxygen stability - for operation of air-sensitive enzymes under aerobic conditions; and anaerobic activity - for operation of aerobic enzymes under hypoxic conditions.

[0295] The exemplary methods described below have been developed for the mutagenesis and diversification of genes to target desired properties of specific enzymes. Any of these methods can be used to alter / optimize the activity of decarboxylases.

[0296] EpPCR (Pritchard, L., D. Corne, D. Keil, J. Rowland and M. Winson, 2005, A general model of error-prone PCR. J Theor. Biol 234:497-509.) is achieved by adding Mn 2+ Ions reduce the fidelity of DNA polymerase in PCR reactions, and random point mutations are introduced by biasing dNTP concentrations or through other conditional alterations. A five-step cloning process that confines mutagenesis to the relevant target gene involves: 1) error-prone PCR amplification of the relevant gene; 2) restriction enzyme digestion; 3) gel purification of the desired DNA fragment; 4) ligation into a vector; 5) transformation of the gene variant into a suitable host; and library screening to improve efficiency. This method can simultaneously generate multiple potentially useful mutations in a single gene. EpPCR can generate a large number of mutants, therefore high-throughput screening assays or selection methods (especially robotic ones) are useful for identifying mutants with the desired characteristics.

[0297] Error-prone rolling circle amplification (epRCA) (Fujii, R., M. Kitaoka and K. Hayashi, 2004, One-step random mutagenesis by error-prone rolling circle amplification. Nucleic AcidsRes 32:1145; and Fujii, R., M. Kitaoka and K. Hayashi, 2006, Error-prone rolling circle amplification: the simplest random mutagenesis protocol. Nat. Protoc. 1:2493-2497) shares many elements with epPCR, except that the entire circular plasmid is used as a template and random 6-primers with exonuclease-resistant phosphate-thioester bonds on the last two nucleotides are used to amplify the plasmid, which is then transformed into cells where it is repeatedly recirculated in tandem. Adjusting Mn 2+ Concentration can slightly alter the mutation rate. This technique utilizes a simple, error-prone, one-step method to create complete copies of plasmids at 3-4 mutations / kbp. No restriction enzyme digestion or specific primers are required. Furthermore, this method is typically available in kit form.

[0298] DNA or family shuffling (Stemmer, WP1994, DNA shuffling by random fragmentation and reassembly: in vitro recombination for molecular evolution. Proc Natl Acad Sci U SA 91:10747-10751; and Stemmer, WP1994. Rapid evolution of a protein invitro by DNA shuffling. Nature 370:389-391) typically involves digesting two or more variant genes with nucleases (such as deoxyribonuclease I or endonuclease V) to generate a pool of random fragments, which are then reassembled in the presence of DNA polymerase through annealing and expansion cycles to create a chimeric gene library. Recombination occurs when fragments act as primers for each other and when one copy acts as a primer for another (template switching). This method can be used with DNA sequences >1kbp. In addition to the mutant recombination created by fragment reassembly, this method introduces point mutations at a rate similar to error-prone PCR in the expansion step. This method can be used to remove harmful random neutral mutations that may contribute to antigenicity.

[0299] Staggered extension (StEP) (Zhao, H., L. Giver, Z. Shao, J.A.Affholter, and F.H.A. R., 1998, Molecular evolution by staggered extension process (StEP) in vitro recombination. Nat. Biotechnol 16:258-261.) always involves a template primer followed by repeated cycles of two-step PCR with denaturation and very short (approximately 5 seconds) annealing / extension. The growing fragment anneals to different templates and extends further; this is repeated until a full-length sequence is formed. Template switching means that most resulting fragments have multiple parents. The combination of low-fidelity polymerases (Taq and Mutazyme) reduces error-prone bias due to their contrasting mutation profiles.

[0300] In random primer recombination (RPR), random sequence primers are used to generate numerous short DNA fragments complementary to different template segments (Shao, Z., H. Zhao, L. Given and F. H. Arnold, 1998, Random-priming in vitro recombination: an effective tool for directed evolution. Nucleic Acids Res 26:681-683.). Point mutations are introduced by base misincorporation and mispriming in epPCR. Based on homology, the short DNA fragments act as primers for each other and are recombinated and reassembled into the full length through repeated thermal cycling. Template removal prior to this step ensures low parental recombination. Similar to most other methods, this method involves multiple iterations to progressively obtain different properties. This technique avoids sequence bias, is independent of gene length, and requires very little parental DNA.

[0301] In heteroduplex recombination, linear plasmid DNA is used to form heteroduplexes that are repaired through mismatch repair (Volkov, AA, Z. Shao and F. H. Arnold. 1999. Recombination and chimeragenesis by invitro heteroduplex formation and in vivo repair. Nucleic Acids Res 27:18; and Volkov, AA, Z. Shao and F. H. Arnold. 2000. Random chimeragenesis by heteroduplex recombination. Methods Enzymol. 328:456-463). The mismatch repair step is at least somewhat mutagenic. Heteroduplex transformation is more efficient than linear homoduplex transformation. This method is suitable for large genes and whole operons.

[0302] Temporary template random chimeric growth (RACHITT) (Coco, WM, WE, Levinson, MJ, Crist, HJ, Hektor, A. Darzins, PTPienkos, CHS, quires, and DJ, Monticello, 2001, DNA shuffling method for generating highly recombined genes and evolved enzymes. Nat. Biotechnol 19:354-359) employs deoxyribonuclease I fragmentation and size fractionation of single-stranded DNA. Homologous fragments are hybridized to a complementary single-stranded DNA scaffold under polymerase-free conditions. Any overlapping unhybridized fragment ends are trimmed by an exonuclease. Gaps between fragments are filled and then ligated to provide a pool of diverse strands of full length that hybridize to the scaffold (which contains U to prevent amplification). The scaffold is then disrupted and replaced by a new strand complementary to the diverse strands by PCR amplification. This method involves only one strand (scaffold) from one parent while primer fragments are from other genes; the parental scaffold is relatively selected. Therefore, no re-annealing of parental fragments occurs. Overlapping fragments are repaired using exonucleases. In other respects, this is similar to DNA shuffling and StEP. Therefore, there should be no siblings, almost no inactivated genes, and no unshuffled parents. The advantage of this technique is that it produces little to no parental genes and generates far more crossovers compared to standard DNA shuffling.

[0303] Truncated template recombination extension (RETT) necessarily involves template switching of the self-growing strand from primers in the presence of a pool of single-stranded DNA fragments (Lee, SH, EJRyu, MJKang, E.-S.Wang, ZCYPiao, KJJJung and Y.Shin, 2003, A new approach to directed gene evolution by recombined extension on truncated templates (RETT). J.Mo lee. Catalysis 26:119-129.). It does not use DNA endonucleases. Single-stranded DNA is generated via DNA polymerase with random primers or through serial deletion using exonucleases. The single-stranded DNA serves only as a template, not a primer. Random primers and exonucleases do not introduce sequence biases such as enzymatically induced breaks like those in DNA shuffling / RACHITT. RETT can be optimized more easily than StEP because it utilizes normal PCR conditions rather than very short extensions. Recombination occurs as part of the PCR step—there is no direct shuffling. This method, because it is uninterrupted, can also be more random than StEP.

[0304] In degenerate oligonucleotide gene shuffling (DOGS), degenerate primers are used to control recombination between molecules (Bergquist, PL and MD Gibbs, 2007, Degenerate oligonucleotide gene shuffling. Methods Mol. Biol 352:191-204; Bergquist, PL, RAR Reeves and MD Gibbs, 2005, Degenerate oligonucleotide gene shuffling (DOGS) and random drift mutagenesis (RNDM): two complementary techniques for enzyme evolution. Biomol. Eng 22:63-72; Gibbs, MD, KMNevalainen and PL Bergquist, 2001, Degenerate oligonucleotide gene shuffling (DOGS): a method for enhancing the frequency of recombination with family shuffling. Gene 271:13-20). This can be used to control the tendency of other methods (such as DNA shuffling) to re-degenerate parental genes. This method can be combined with selective epPCR (epPCR) of gene segments. This would be a good method to inhibit parental sequence reformation. No endonucleases are required. By adjusting the input concentration of the resulting segments, the desired backbone can be selected. This method allows for DNA shuffling from unrelated parents without restriction enzyme digestion and allows for selective epPCR.

[0305] Progressive cutting generates a hybrid enzyme (ITCHY) to create a library of combinations of related genes or gene fragments with 1-base-pair deletions. (Ostermeier et al., Proc Natl Acad Sci US.A. 96:3562-3567 (1999); Ostermeier et al., 1999 Nat. Biotechnol. 17:1205-1209 (1999)) The cuts are introduced into two different gene fragments in opposite directions. These are then ligated together and cloned as a fusion. This technique does not require homology between the two parental genes. When ITCHY is combined with DNA shuffling, the system is called SCRATCHY (see below). The main advantage of both is that homology between parental genes is not required; for example, functional fusions between E. coli and human genes can be created via ITCHY. When an ITCHY library is formed, all possible exchanges are captured.

[0306] Except for the use of phosphate-thiophosphate dNTPs to generate cleavage products, the thioprogressive cleavage to produce a hybrid enzyme (THIO-ITCHY) is almost identical to ITCHY. (Lutz, S., M. Ostermeier and SJ Benkovic, 2001, Rapid generation of incremental truncation libraries for protein engineering using alpha-phosphothioate nucleotides. Nucleic Acids Res 29:E16.) Compared to ITCHY, THIO-ITCHY is easier to optimize and offers greater reproducibility and tunability.

[0307] SCRATCHY-ITCHY, a combination of DNA shuffling and ITCHY, allows for multiple crossovers (Lutz et al., Proc Natl Acad Sci U SA98:11248-11253 (2001)). SCRATCHY combines the best properties of both ITCHY and DNA shuffling. Computational prediction methods can be used for optimization. When sequence identity is below 80%, SCRATCHY is more efficient than DNA shuffling.

[0308] In random drift mutagenesis (RNDM), mutations are performed via epPCR, followed by screening / selection of those that retain useful activity (Bergquist et al., Biomol. Eng. 22:63-72 (2005)). These are then used in DOGS to generate recombinants with multiple active mutants or fusions between active mutants and some other desired parents. Designed to facilitate the isolation of neutral mutations; its purpose is to screen for retained catalytic activity, indicating whether this activity is higher or lower than that in the original gene. RNDM is useful in high-throughput assays when screening can detect activity in the background. RNDM has been used as a front-end for DOGS in terms of generating diversity. This technique sets activity requirements before shuffling or other subsequent steps; neutral drift libraries have been shown to produce higher / faster activity improvements compared to smaller libraries. While published using epPCR, this can be applied to other large-scale mutagenesis methods.

[0309] Sequence saturation mutagenesis (SeSaM) is a random mutagenesis method that: 1) generates a pool of random-length fragments by random incorporation and breakage of phosphate-thionucleotides; this pool is used as a template to 2) extend in the presence of a "universal" base (such as inosine); 3) replication of complement containing inosine provides random base incorporation and thus provides mutagenesis (Wong et al., Biotechnol J. 3:74-82 (2008); Wong Nucleic Acids Res 32:e26; Wong et al., Anal. Biochem. 341:187-189 (2005)). Using this technique, it is possible to generate large libraries of mutants within 2-3 days using a simple method. This is highly non-directional compared to the mutational bias of DNA polymerase. This difference makes this technique a complement (or alternative) to epPCR.

[0310] In synthetic shuffling, overlapping oligonucleotides are engineered to encode “all the genetic diversity in the target” and allow for very high diversity in the shuffling progeny (Ness et al., Nat. Biotechnol 20:1251-1255 (2002)). In this technique, fragments can be engineered to be shuffled. This contributes to increased diversity in the resulting progeny. Sequence / codon biases can be engineered so that more distantly related sequences recombine at rates approaching more closely related sequences, without requiring physical possession of the template gene.

[0311] Nucleotide exchange and excision (NexT) technology utilizes dUTP incorporation followed by treatment with uracil DNA glycosylase and then piperidine to perform endpoint DNA breaks (Muller et al., Nucleic Acids Res 33:e117 (2005)). Internal PCR primer expansion is used to correct the polymerase reassembly of the gene. The size of the shuffle is directly controllable using a varying dUPT::dTTP ratio. This is a simple endpoint reaction utilizing uracil incorporation and breakage. Other nucleotide analogues, such as 8-oxo-guanine, can be used for this method. Furthermore, this technique works well for very short fragments (86 bp) with a low error rate. The chemical breakage of DNA means that almost no unshuffled clones are produced.

[0312] In sequence-homology-independent protein recombination (SHIPREC), a linker is used to facilitate fusion between two distantly related / unrelated genes; nuclease treatment is used to generate a series of chimeras between the two. The result is a single-crossover library of these fusions (Sieber, V., CA. Martinez and F. H. Arnold. 2001. Libraries of hybrid proteins from distantly related sequences. Nat. Biotechnol 19:456-460). This produces a limited number of shuffles; mutagenesis is a separate process. This technique can create chimeric libraries with fractional changes to each of two unrelated parental genes. Homology is not required. SHIPREC was tested by fusing the heme-binding region of bacterial CP450 to the N-terminal region of mammalian CP450; this produced mammalian activity in the more soluble enzyme.

[0313] In site-specific saturation mutagenesis (GSSM), the starting material is a supercoiled double-stranded DNA plasmid with insertions at the desired mutation site and two primers that are degenerate (Kretz, KA, THRichardson, KAGray, DERobertson, X. Tan, and JMShort, 2004, Gene site saturation mutagenesis: a comprehensive mutagenesis approach. Methods Enzymol. 388:3-11). Primers introduce relevant mutations and annealing at the same sequence on opposite strands of the DNA; mutations in the middle of the primer and ~20 nucleotides flanking the correct sequence on each side. The sequences in the primers are NNN or NNK (coding) and MNN (uncoding) (N = all 4, K = G, T, M = A, C). After extension, DpnI is used to digest dam-methylated DNA to eliminate the wild-type template. This technique explores all possible amino acid substitutions at a given site (i.e., a codon). This technology facilitates the generation of all possible substitutions at a site free of meaningless codons, along with identical or nearly identical expression of the most probable alleles. No existing knowledge of the target enzyme's structure, mechanism of action, or region is required. If subsequent shuffling or gene reassembly occurs, this technology creates diverse libraries of recombinants containing all possible combinations of point-up mutations. The practicality of this combined technology has been demonstrated for the successful evolution of over 50 different enzymes and for more than one trait within a given enzyme.

[0314] Combinatorial cassette mutagenesis (CCM) involves the use of short oligonucleotide cassettes to replace limited regions with a large number of possible amino acid sequence changes (Reidhaar-Olson, JF, JUBowie, RMBreyer, JCHu, KLKnight, WALim, MCMossing, D.A. Parsell, KR. Shoemaker, and RTSauer, 1991, Random mutagenesis of protein sequences using oligonucleotide cassettes. Methods Enzymol. 208:564-586; and Reidhaar-Olson, JF, and RTSauer, 1988, Combinatorial cassette mutagenesis as a probe of the informational content of protein sequences. Science 241:53-57). Simultaneous substitution at two or three sites is possible using this technique. Furthermore, this method tests a large number of possible sequence changes at a limited range of sites. It has been used to explore the informational content of λ-repressor DNA-binding regions.

[0315] Combinatorial multiple-cassette mutagenesis (CMCM) is essentially similar to CCM, except that it is employed as part of a larger protocol: 1) using epPCR at a high mutation rate, 2) identifying hotspots and hot regions, and then 3) extending the CMCM to cover the spatially defined regions of the protein sequence (Reetz, MT, S. Wilensek, D. Zha, and KE Jaeger, 2001, Directed Evolution of an enantioselective Enzyme through Combinatorial Multiple-Cassette Mutagenesis. Angew. Chem. Int. Ed Engl. 40:3589-3591). Like CCM, this technique can virtually test all possible alterations on the target region. When used in conjunction with methods for generating random mutations and shuffled genes, it provides an excellent means of generating diverse, shuffled proteins. This method successfully increased the enantioselectivity of enzymes by 51-fold.

[0316] In mutagenic strain technology, conditionally ts-mutant plasmids allow for a 20- to 4000-X increase in the frequency of random and natural mutations during selection and, when selection is not required, allow for the inhibition of the accumulation of harmful mutations (Selifonova, O., F. Valle and V. Schellenberger, 2001, Rapid evolution of novel traits in microorganisms. Appl Environ Microbiol 67:3645-3649). This technology is based on the plasmid-derived mutD5 gene, which encodes a mutant subunit of DNA polymerase III. This subunit binds to endogenous DNA polymerase III and impairs the proofreading ability of polymerase III in any strain containing this plasmid. A wide range of base substitution and frameshift mutations occur. For efficient use, the mutagenic plasmid should be removed once the desired phenotypic effect is achieved; this is accomplished via a temperature-sensitive initiation site of replication, which allows for plasmid elimination at 41°C. It should be noted that mutagenic strains have been investigated for a long time (e.g., see Winter et al., 1996, J. Mol Biol. 260, 359-3680.). A very high rate of spontaneous mutation has been observed in this technique. The conditional nature minimizes unwanted background mutations. This technique can be combined with adaptive evolution to increase the mutagenesis rate and more rapidly obtain the desired phenotype.

[0317] "Precise mutagenesis (LTM) is a multidimensional mutagenesis method for evaluating and optimizing the selection of combinatorial mutations of amino acids" (Rajpal, A., N. Beyaz, L. Haber, G. Cappuccilli, H. Yee, RRBhatt, T. Takeuchi, RALerner, and R. Crea, 2005, A general method for greatly improving the affinity of antibodies by using combinatorial libraries. Proc Natl Acad Sci U SA 102:8466-8471). Instead of saturating each site with all possible amino acid changes, it selects a set of nine to cover the range of amino acid R-group chemicals. Fewer changes at each site allow multiple sites to undergo this type of mutagenesis. >800-fold increases in binding affinity against antibodies, from low nanomolar to picomolar levels, have been achieved using this method. This is a reasonable method for minimizing the number of random combinations and this will improve the ability to discover improved traits by greatly reducing the number of clones to be screened. This has been used in antibody engineering, particularly to improve binding affinity and / or reduce dissociation. This technique can be combined with screening or selection.

[0318] Gene reassembly is a DNA shuffling method that can be applied to multiple genes at once or to create large chimeric (multiplexed) libraries of single genes (see website). www.verenium.com / Pages / Technology / EnzymeTech / TechEnzyTGR.html This technique is typically used in conjunction with ultra-high-throughput screening to challenge the represented sequence space for desired improvement. It allows for multiple gene recombinations independent of homology. The exact number and location of crossover events can be pre-determined using fragments designed through bioinformatics analysis. This technique results in very high levels of diversity with virtually no parental gene recombination and low levels of gene inactivation. Combined with GSSM, it allows for improved activity testing of a wide range of mutations. This method allows for the "mixing" and "fine-tuning" of DNA shuffling, such as optimizing codon usage.

[0319] Computer-simulated protein design automation (PDA) is an optimization algorithm that fixes a structurally defined protein backbone with specific folds and searches for sequence spaces for amino acid substitutions that can stabilize the folds and the overall protein energy (Hayes, RJ, J. Bentzien, MLAry, MYHwang, JMJacinto, J. Vielmetter, A. Kundu, and BIDahiyat, 2002, Combining computational and experimental screening for rapid optimization of protein properties. Proc Natl Acad Sci U SA 99:15926-15931). This technique allows for structure-based entropy prediction based on computer simulations to seek structural tolerance to changes in protein amino acids. Statistical mechanics is used to calculate coupling interactions at each position—structural tolerance to amino acid substitutions is a measure of coupling. Finally, this technique is designed to achieve desired modifications to protein properties while maintaining the integrity of structural properties. The method computationally evaluates and allows for a very large number of possible sequence variants (10). 50 The selection of target sequence variants is related to the best-fit thermodynamic predictions, and clearly, only stability or stability-related properties can be effectively addressed using this technique. This method has been successfully applied to some medical proteins, particularly engineered immunoglobulins. Computer simulation predictions avoid testing an unusually large number of potential variants. Predictions based on existing three-dimensional structures are more likely to succeed than those based on pseudostructures. This technique can easily predict and allow targeted screening for multiple simultaneous mutations, which is impossible with simple experimental techniques due to the exponential increase in the number.

[0320] Iterative saturation mutagenesis (ISM) involves: 1) selecting an appropriate site for enzyme improvement using knowledge of structure / function; 2) saturating mutagenesis at the selected site using Stratagenene QuikChange (or other suitable methods); 3) screening for / selecting the desired properties; and 4) restarting at another site under improved cloning conditions and repeating continuously (Reetz, MT, and JD Carballeira, 2007, Iterative saturation mutagenesis (ISM) for rapid directed evolution of functional enzymes. Nat. Protoc. 2:891-903; and Reetz, MT, JD Carballeira, and A. Vogel, 2006, Iterative saturation mutagenesis on the basis of B factors as a strategy for increasing protein thermostability. Angew. Chem. Int. Ed Engl. 45:7745-7751). This is a proven method that ensures all possible substitutions at a given site contribute to screening / selection.

[0321] Any of the aforementioned methods for mutagenesis can be used alone or in any combination. Furthermore, any or any combination of the directed evolution methods can be used in conjunction with adaptive evolution techniques.

[0322] To generate better producers, metabolic modeling can be used to optimize culture conditions. Modeling can also be used to design gene knockouts for additional optimized pathway applications (see, for example, U.S. Patent Publications US 2002 / 0012939, US 2003 / 0224363, US 2004 / 0029149, US 2004 / 0072723, US 2003 / 0059792, US 2002 / 0168654, and US 2004 / 0009466, and U.S. Patent No. 7,127,379). Modeling analysis allows for reliable prediction of the effects of shifted metabolism on cell growth for more efficient production of 1,3-butanediol.

[0323] A computational framework, OptKnock, is proposed for identifying and designing metabolic alterations to support the biosynthesis of desired products (Burgard et al., Biotechnol. Bioeng. 84:647-657 (2003)). OptKnock is a metabolic modeling and simulation program that proposes gene deletion strategies to produce genetically stable microorganisms that produce target products beyond quotas. Specifically, the framework examines the complete metabolic and / or biochemical network of a microorganism to propose genetic manipulations that force the desired biochemical product to become a specific byproduct of cell growth. By strategically placing gene deletions or other functional gene disruptions to couple biochemical production with cell growth, the growth selection pressure applied to the designed strain after a long period in a bioreactor results in performance improvement due to the combination of forced growth and biochemical production. Finally, when gene deletions are constructed, the likelihood of the designed strain reverting to its wild-type state is negligible because the genes selected by OptKnock are completely removed from the genome. Therefore, this computational method can be used to identify alternative pathways leading to the biosynthesis of a desired product or to combine with the non-naturally occurring microorganisms for further optimization of the biosynthesis of the desired product.

[0324] In short, the term OptKnock is used in this paper to refer to computational methods and systems for modeling cellular metabolism. The OptKnock scheme involves a framework of models and methods that incorporate specific constraints into flux balance analysis (FBA) models. These constraints include, for example, qualitative dynamic information, qualitative regulatory information, and / or DNA microarray experimental data. OptKnock also computes solutions to various metabolic problems by, for example, tightening flux boundaries derived from flux balance models and subsequently probing the efficiency limits of metabolic networks in the presence of gene additions or deletions. The OptKnock computational framework allows for the construction of feasible model representations that enable efficient questioning of the efficiency limits of metabolic networks and provides methods for solving induced mixed-integer linear programming problems. Descriptions of metabolic modeling and simulation methods referred to in this paper as OptKnock can be found, for example, in US2002 / 0168654, WO 2002 / 055995, and US2009 / 0047719.

[0325] Another computational approach for identifying and designing metabolic alterations that support the biosynthetic production of products is a metabolic modeling and simulation system called... The description of the calculation method and system can be found, for example, in US2003 / 0233218 filed on June 14, 2002, and in WO / 2003 / 106998. This is a computational system that can be used to generate computer-simulated network models and simulate fluxes of mass, energy, or charge through the chemical reactions of a biological system. This limits the solution space to any and all possible functions of the chemical reactions in the system, thereby determining a range of permissible activities of the biological system. This approach is called constraint-based modeling because the solution space is limited by constraints such as the known stoichiometry of the included reactions and the reaction thermodynamics and capacity constraints associated with the maximum flux. The space defined by these constraints can be queried to determine the dominant capacity and behavior of the biological system or its biochemical components.

[0326] These computational methods align with biological reality because biological systems are flexible and can achieve the same results in many different ways. Biological systems have been designed through evolutionary mechanisms constrained by fundamental constraints that all biological systems must confront. Therefore, constraint-based modeling strategies encompass these general realities. Furthermore, the ability to continuously impose further constraints on network models through tightening of constraints leads to a smaller solution space, thereby improving physiological efficacy or the accuracy of phenotypic predictability.

[0327] With regard to the teachings and guidance provided herein, those skilled in the art will be able to apply various computational frameworks for metabolic modeling and simulation to design and implement the biosynthesis of desired compounds in host microorganisms. Such metabolic modeling and simulation methods include, for example, those demonstrated above. The OptKnock computational system. To illustrate the invention, this document describes some methods relating to the OptKnock computational framework for modeling and simulation. Those skilled in the art will understand how to utilize OptKnock to apply the identification, design, and implementation of metabolic alterations to any other such metabolic modeling and simulation computational frameworks and methods well known in the art.

[0328] The method described above will provide a set of metabolic reactions to be interrupted. Elimination of each reaction within this set or metabolic modification can produce the desired product as a specific product during the incubation period of the organism. Since the reactions are known, the solution to the bilayer OptKnock problem will also provide the relevant genes encoding one or more enzymes that catalyze each reaction within this set. The identification of a set of reactions and their corresponding genes encoding the enzymes involved in each reaction is typically an automated process, accomplished by associating the reactions with a database of reactions that have relationships between enzymes and their encoding genes.

[0329] Once determined, the set of reactions is implemented in a target cell or organism by disrupting the function of at least one gene encoding each metabolic reaction within the set of reactions in order to produce the desired product. A particularly useful means of disrupting the function of this set of reactions is by deleting each coding gene. However, in some cases, it may be advantageous to disrupt the reaction by other genetic aberrations, including, for example, mutations, deletions, or truncations of coding sequences at arbitrary locations in regulatory regions such as promoters or cis-binding sites for regulatory factors. These latter aberrations, which result in incomplete deletion of the genome, may be useful, for example, when rapid assessment of product coupling is required or when genetic reversion variations are unlikely.

[0330] To identify other productive solutions to the bilayer OptKnock problem described above, which leads to further interruption of reaction sets or to metabolic modifications that may result in biosynthesis (including growth-coupled biosynthesis of the desired product), an optimization method called integer cutting can be implemented. This method proceeds by iteratively solving the exemplary OptKnock problem described above by incorporating another constraint, known as integer cutting, at each iteration. The integer cutting constraint effectively prevents the solution procedure from selecting exactly the same reaction set determined in any previous iteration that forces the coupling of product biosynthesis to growth. For example, if previously determined metabolic modifications coupled to growth specify reactions 1, 2, and 3 for interruption, the subsequent constraint prevents the same reactions from being considered simultaneously in subsequent solutions. This integer cutting method is well known in the art and can be found described, for example, in Burgard et al., Biotechnol. Prog. 17:791-797 (2001). Similar to all the methods described in this paper regarding their use in conjunction with the OptKnock computational framework for metabolic models and simulations, the integer slicing method for reducing redundancy in iterative computational analysis can also be used with other computational frameworks well known in the art (including, for example, Use them together.

[0331] The exemplary methods described herein allow for the biosynthetic production of desired products from cells and organisms, including the forced coupling of the production of targeted biochemical products with the growth of cells or organisms designed to include identified genetic alterations. Therefore, the computational methods described herein allow for the use of OptKnock or... The identification and implementation of metabolic modifications identified by computer simulation methods. Metabolic modifications may include, for example, the addition of one or more enzymes in a biosynthetic pathway and / or the disruption of function of one or more metabolic reactions (including, for example, disruption via gene deletion).

[0332] As discussed above, the OptKnock method is based on the premise that mutant microbial networks, when subjected to long-term growth selection, can evolve toward their computationally predicted maximum growth phenotypic expression. In other words, this method leverages the self-optimizing ability of organisms under selective pressure. The OptKnock framework allows for exhaustive enumeration of gene deletion combinations that couple biochemical production and cell growth based on network stoichiometry. Identification of the optimal gene / response knockout requires a solution to a bilayer optimization problem that selects active response groups so that the resulting optimal growth solution of the network exceeds the quota for producing the relevant biochemicals (Burgard et al., Biotechnol. Bioeng. 84:647-657 (2003)).

[0333] Computer-simulated stoichiometric models of E. coli metabolism can be used to identify genes essential to metabolic pathways, as previously demonstrated and described in, for example, U.S. Patent Publications US 2002 / 0012939, US 2003 / 0224363, US 2004 / 0029149, US 2004 / 0072723, US 2003 / 0059792, US 2002 / 0168654, and US 2004 / 0009466, and U.S. Patent No. 7,127,379. As disclosed herein, the OptKnock mathematical framework can be used to precisely locate gene deletions that lead to the coupled growth of the desired product. Furthermore, solutions to the bilayer OptKnock problem provide only one set of deletions. To enumerate all meaningful solutions, i.e., all knockout sets leading to the formation of coupled growth, an optimization technique called integer cutting can be implemented. This necessitates iteratively solving the OptKnock problem by incorporating another constraint known as integer cutting at each iteration.

[0334] It should be understood that modifications that do not substantially affect the performance of various embodiments of the present invention are also included in the definition of the invention provided herein. Therefore, the embodiments described below are intended to illustrate rather than limit the invention.

[0335] Example I

[0336] Synthesis of 1,3-butanediol from alanine

[0337] This embodiment describes the ability to utilize Figure 1 The alanine pathway is used by microorganisms to produce 1,3-butanediol via steps A, B, C, D, and H.

[0338] Escherichia coli was used as a target organism to design 1,3-butanediol production pathways, such as... Figure 1As shown, *Escherichia coli* provides a good host for generating non-naturally occurring microorganisms capable of producing 1,3-butanediol. *Escherichia coli* is well-suited for genetic manipulation and is known to efficiently produce a variety of products, such as ethanol, acetic acid, formic acid, lactic acid, and succinic acid, under anaerobic or microaerobic conditions.

[0339] To generate Escherichia coli designed to produce 1,3-butanediol, well-known molecular biotechnologies were used to express nucleic acids in E. coli encoding enzymes used in the alanine pathway as previously described (see, for example, Sambrook, ibid., 2001; Ausubel, ibid., 1999; Roberts et al., ibid., 1989).

[0340] Specifically, the genes ortA (YP_001086914.1), ortB (YP_001086915.1), dat (P19938), and pdc (P06672), encoding AKP thiolase, AKP transaminase, and 2,4-dioxovalerate decarboxylase activities, respectively, were cloned into the pZE13 vector (Expressys, Ruelzheim, Germany) under the PA1 / lacO promoter. Furthermore, the genes yqhD (NP_417484.1) and adh (AAA23199.2), encoding 3-oxobutyraldehyde reductase (aldehyde reduction) and A-hydroxy,2-butanone reductase, respectively, were cloned into the pZA33 vector (Expressys, Ruelzheim, Germany) under the PA1 / lacO promoter. These two sets of plasmids were transformed into *E. coli* strain MG1655 to express the proteins and enzymes required for 1,3-butanediol synthesis via the alanine pathway. Note that Escherichia coli has the ability to form D-alanine.

[0341] Genetically engineered organisms were cultured in a glucose-containing medium following procedures well known in the art (see, for example, Sambrook et al., ibid., 2001). The expression of alanine pathway genes was consolidated using methods well known in the art for determining peptide expression or enzyme activity (including, for example, Northern blotting, PCR amplification of mRNA, and Western blotting). The enzyme activity of the expressed enzyme was determined using assays specific to individual activity. The ability of the engineered *E. coli* strain to produce 1,3-butanediol was determined using HPLC, gas chromatography-mass spectrometry (GC-MS), or liquid chromatography-mass spectrometry (LC-MS).

[0342] The efficiency of the functional 1,3-butanediol biosynthesis pathway was further enhanced by optimizing its utilization. In short, the engineered strain was evaluated to determine whether any exogenous genes were expressed at a rate-limited level. For those expressed at low levels, expression could be increased by, for example, by introducing additional gene copy number-limiting enzymes that restrict the flux through the pathway.

[0343] To generate better producers, metabolic modeling can be used to optimize culture conditions. Modeling has also been used to design gene knockouts for additional optimized pathway applications (see, for example, U.S. Patent Publications US 2002 / 0012939, US 2003 / 0224363, US 2004 / 0029149, US 2004 / 0072723, US 2003 / 0059792, US 2002 / 0168654, and US 2004 / 0009466, and see U.S. Patent No. 7,127,379). Modeling analysis allows for reliable prediction of the effects of shifted metabolism on cell growth for more efficient production of 1,3-butanediol. One modeling approach is the two-layer optimization method OptKnock (Burgard et al., Biotechnol. Bioengineer. 84:647-657 (2003)), which is used to select for gene knockouts that collectively lead to better production of 1,3-butanediol. Adaptive evolution can also be used to generate, for example, better producers of alanine or 2-amino-4-oxovalerate intermediates or 1,3-butanediol products. Adaptive evolution is performed to improve growth and production characteristics (Fong and Palsson, Nat. Genet. 36:1056-1058 (2004); Alper et al., Science 314:1565-1568 (2006)). Based on the results, subsequent rounds of modeling, genetic engineering, and adaptive evolution can be applied to 1,3-butanediol producers to further increase production.

[0344] To achieve large-scale production of 1,3-butanediol, the alanine pathway-containing organism described above was cultured in a fermenter using a medium known in the art to support anaerobic growth. Fermentation was carried out in batch, fed-batch, or continuous manner. Anaerobic conditions were maintained by first spraying the medium with nitrogen and then sealing the culture vessels (e.g., flasks sealed with gaskets and clamp caps). Microaerobic conditions could also be employed by providing limited aeration through small holes. The pH of the medium was maintained at pH 7 by adding an acid (e.g., H₂SO₄). The growth rate was determined by measuring the optical density using a spectrophotometer (600 nm), and the glucose uptake rate was determined by monitoring the consumption of the carbon source over time. Byproducts (such as unwanted alcohols, organic acids, and residual glucose) could be quantified using HPLC (Shimadzu) with an HPX-087 column (BioRad) using refractive index detectors for glucose and alcohols and UV detectors for organic acids (Lin et al., Biotechnol. Bioeng., 775-779 (2005)).

[0345] Example II

[0346] Synthesis of 1,3-BDO using acetoacetyl-COA as an intermediate

[0347] This embodiment describes the production of 1,3-butanediol using acetyl-CoA as a precursor. Figure 2 The generation of microorganisms in steps G, H and I).

[0348] Escherichia coli was used as a target organism for design Figure 2 The process involves three pathways: G (conversion from acetoacetyl-CoA to 3-hydroxybutyryl-CoA), H (conversion from 3-hydroxybutyryl-CoA to 3-hydroxybutyraldehyde), and I (conversion from 3-hydroxybutyraldehyde to 1,3-butanediol). *E. coli* provides a suitable host for generating non-naturally occurring microorganisms capable of producing 1,3-butanediol. *E. coli* is well-suited for genetic manipulation and is known to efficiently produce a variety of products, such as ethanol, acetic acid, formic acid, lactic acid, and succinic acid, under anaerobic or microaerobic conditions.

[0349] To generate *E. coli* designed for the production of 1,3-butanediol, well-known molecular biotechnologies were used to express nucleic acids encoding the enzymes used in the disclosed pathways (steps G, H, and I) in *E. coli* (see, for example, Sambrook, ibid., 2001; Ausubel, ibid., 1999; Roberts et al., ibid., 1989). Note that *E. coli* possesses a native thiolase encoding two molecules of condensed acetyl-CoA encoded by atoB (accession number: NP_416728.1) to form acetoacetyl-CoA.

[0350] Furthermore, hbd (NP_349314.1) encoding acetyl-CoA reductase (ketone reduction) was cloned into the pZE13 vector (Expressys, Ruelzheim, Germany) under the PA1 / lacO promoter. This plasmid was transformed into E. coli strain MG1655 to express the enzyme required for the formation of 3-hydroxybutyryl-CoA via acetyl-CoA. Aldehyde dehydrogenases (selected from Table A below) that convert 3-hydroxybutyryl-CoA to 3-hydroxybutyraldehyde and alcohol dehydrogenases (selected from Table B below) that further reduce 3-hydroxybutyraldehyde to 1,3-BDO were also cloned into the pZE13 vector under the PA1 / lacO promoter.

[0351] The genetically engineered organism was cultured in a glucose-containing medium following procedures well known in the art (see, for example, Sambrook et al., ibid., 2001). Expression of the pathway gene was consolidated using methods well known in the art for determining peptide expression or enzyme activity (including, for example, Northern blotting, PCR amplification of mRNA, and Western blotting). The enzyme activity of the expressed enzyme was determined using assays specific to individual activity. The ability of the engineered *E. coli* strain to produce 1,3-butanediol was determined using HPLC, gas chromatography-mass spectrometry (GC-MS), or liquid chromatography-mass spectrometry (LC-MS).

[0352] The efficiency of the functional 1,3-butanediol biosynthesis pathway was further enhanced by optimizing its utilization. In short, the engineered strain was evaluated to determine whether any exogenous genes were expressed at a rate-limited level. For those expressed at low levels, expression could be increased by, for example, by introducing additional gene copy number-limiting enzymes that restrict the flux through the pathway.

[0353] To generate better producers, metabolic modeling can be used to optimize culture conditions. Modeling has also been used to design gene knockouts for additional optimized pathway applications (see, for example, U.S. Patent Publications US 2002 / 0012939, US 2003 / 0224363, US 2004 / 0029149, US 2004 / 0072723, US 2003 / 0059792, US 2002 / 0168654, and US 2004 / 0009466, and see U.S. Patent No. 7,127,379). Modeling analysis allows for reliable prediction of the effects of shifted metabolism on cell growth for more efficient production of 1,3-butanediol. One modeling approach is the two-layer optimization method OptKnock (Burgard et al., Biotechnol. Bioengineer. 84:647-657 (2003)), which is used to select for gene knockouts that collectively lead to better production of 1,3-butanediol. Adaptive evolution can also be used to generate, for example, better producers of acetyl-CoA intermediates or 1,3-butanediol products. Adaptive evolution is performed to improve growth and production characteristics (Fong and Palsson, Nat. Genet. 36:1056-1058 (2004); Alper et al., Science 314:1565-1568 (2006)). Based on the results, subsequent rounds of modeling, genetic engineering, and adaptive evolution can be applied to 1,3-butanediol producers to further increase production.

[0354] To achieve large-scale production of 1,3-butanediol, recombinant organisms were cultured in fermenters using media known in the art to support anaerobic growth. Fermentation was carried out in batch, fed-batch, or continuous manner. Anaerobic conditions were maintained by first spraying the medium with nitrogen and then sealing the culture vessels (e.g., flasks sealed with gaskets and clamp caps). Microaerobic conditions could also be employed by providing limited aeration through small holes. The pH of the medium was maintained at pH 7 by adding an acid (e.g., H₂SO₄). The growth rate was determined by measuring the optical density using a spectrophotometer (600 nm), and the glucose uptake rate was determined by monitoring the consumption of the carbon source over time. Byproducts (such as unwanted alcohols, organic acids, and residual glucose) could be quantified using HPLC (Shimadzu) with an HPX-087 column (BioRad) using refractive index detectors for glucose and alcohols and UV detectors for organic acids (Lin et al., Biotechnol. Bioeng., 90:775-779 (2005)).

[0355] The activities of several aldehyde dehydrogenases for 3-hydroxybutyryl-CoA were tested. The activity of each strain carrying one of the six genes encoding aldehyde dehydrogenases listed in Table A below was tested for 3-hydroxybutyryl-CoA by measuring the release of the CoA moiety in crude bacterial lysate. The genes tested and those found to have significant activity for 3-HBCoA encode proteins with the following registration and GI numbers:

[0356] Table A

[0357]

[0358] To correct for the background activity in the lysis buffer, the measured activity was compared with a negative control containing only the vector "Vo" and no ALD gene. Figure 4 The specific activity of each p-3-hydroxybutyryl-CoA gene tested is shown. Gene IDs are shown on the x-axis.

[0359] Furthermore, the activity of bld (GenBank ID: AAP42563.1, GI: 31075383) for 3-HBCoA was also tested. Next... Figure 5 The activity of this gene for 3-hydroxybutyryl-CoA was shown before and after dialysis.

[0360] The following lists alcohol dehydrogenases that have undergone testing for activity against 3-hydroxybutyraldehyde and demonstrated significant activity.

[0361] Table B

[0362] <![CDATA[ protein ]]> <![CDATA[ GenBank ID number ]]> <![CDATA[ GI number ]]> <![CDATA[ biology ]]> Bdh(Cbei_2181) YP_001309304 150017050 Clostridium beyerridis Bdh(Cbei_1722) YP_001309535.1 150016596 Clostridium beyerridis Bdh(Cbei_2421) YP_001309535.1 150017281 Clostridium beyerridis

[0363] The following scheme was used to demonstrate alcohol dehydrogenase activity (converting 3-hydroxybutyraldehyde to 1,3-BDO) and combined aldehyde and alcohol dehydrogenase activity (converting 3-hydroxybutyryl-CoA to 1,3-BDO).

[0364] Transform chemocompetent cells with plasmids containing either aldehyde dehydrogenase or alcohol dehydrogenase (listed in Tables A and B above). Select colonies from the plate and incubate overnight in LB broth with 100 μg / ml carbenicillin. Inoculate 0 mL of 0 mL culture for each alcohol dehydrogenase, or 1.5 mL of 500 mL culture for each aldehyde dehydrogenase. Incubate cells at 37°C to ~0.7 OD and induce with IPTG. During the 4-hour protein expression period, incubate the culture at 30°C. Divide the cell culture into 30 mL aliquots, centrifuge, and store cell pellets at -80°C. Estimate the final cell density using a sample of the cell culture.

[0365] Combinations of alcohol dehydrogenases and aldehyde dehydrogenases were screened in 96-well plates using 3-hydroxybutyryl-CoA as a substrate and a control (without substrate). Alternatively, to test alcohol dehydrogenase activity, only alcohol dehydrogenase was added with and without the substrate 3-hydroxybutyraldehyde. Cell lysis buffers were prepared on ice in a freezer (4°C). The amount of Bug Buster cell lysis reagent per cell cluster was calculated using the final cell density. Lysozyme (10 μL) and benzonase (10 μL) were added to 35 mL of Bug Buster and gently inverted to mix. First, 50 μM of dithiothreitol (100 mM starting material) was added to the cluster, followed by 1.0 (600 nm) of the Bug Buster enzyme mixture per 0.5 mL OD and gently mixed to resuspend the cell cluster.

[0366] Add 50 μl and 25 μl of a cofactor mixture (4 mM NADH and 4 mM NADPH) at 1 M MOPS (pH = 7.5), 100 μl of aldehyde dehydrogenase cell lysis buffer, and 150 μl of alcohol dehydrogenase cell lysis buffer (either alone or just 150 μl of alcohol dehydrogenase cell lysis buffer) to each well and mix gently. Then, add the relevant substrate to the well. Resuspend 25 mg of 3-hydroxybutyryl-CoA in 250 μl of water and add 5 μl to each well testing alcohol and aldehyde dehydrogenase activity to obtain a final concentration of 1.8 mM. To test alcohol dehydrogenase activity only, add 50 μl of 3-hydroxybutyraldehyde (prepared by mixing 0.6 ml of aldehyde in 5 ml of water with a catalytic base (one NaOH pellet)) to each well. Guthrie, JP (with reference documents) The final concentration of 3-hydroxybutyraldehyde in each well was approximately 50 mM. The 96-well plate was sealed with a plastic PCR seal and incubated overnight (18 hours total) at 30°C with shaking. As protein and cell debris precipitate during incubation, the plate was centrifuged at 4500 x g for 10 minutes, and the supernatant was filtered through a Whatman 96-well filter plate (0.45 μm) before LC-MS analysis. Samples were analyzed for 1,3-butanediol formation.

[0367] Figure 6 The concentrations of 1,3-BDO are shown when 3-hydroxybutyraldehyde is added as a substrate and in the control sample without a substrate. The GI number of the alcohol dehydrogenase is shown.

[0368] Figure 7 The concentrations of 1,3-BDO are shown when 3-hydroxybutyryl-CoA is added as a substrate and in the control sample without a substrate. The GI number of the alcohol dehydrogenase is shown. The GI number of the acetaldehyde dehydrogenase tested simultaneously is 163762382.

[0369] Example 3

[0370] Synthesis of 1,3-BDO using 4-hydroxybutyryl-COA as an intermediate

[0371] This example describes the production of 1,3-butanediol using 4-hydroxybutyryl-CoA as a precursor. Figure 3 The generation of microorganisms in steps A, B, and E).

[0372] Escherichia coli was used as a target organism for design Figure 3 The pathway involves steps A, B, and E. *E. coli* provides a suitable host for generating non-naturally occurring microorganisms capable of producing 1,3-butanediol. *E. coli* is well-suited for genetic manipulation and is known to efficiently produce a variety of products, such as ethanol, acetic acid, formic acid, lactic acid, and succinic acid, under anaerobic or microaerobic conditions.

[0373] To generate *E. coli* strains designed to produce 1,3-butanediol, well-known molecular biotechnologies were used to express nucleic acids encoding the enzymes used in the disclosed pathways (steps A, B, and E) in *E. coli* (see, for example, Sambrook, ibid., 2001; Ausubel, ibid., 1999; Roberts et al., ibid., 1989). Recombinant strains designed to produce large quantities of 4-hydroxybutyryl-CoA have been previously described by the applicant (Burk et al., (US 20090075351)) and will be used to insert the proposed pathway into 1,3-butanediol.

[0374] Furthermore, the genes abfD (YP_3001396399.1), crt (NP_349318.1), and adhE2 (AAK09379.1), respectively encoding the activities of 4-hydroxybutyryl-CoA dehydratase, crotonylase, and 3-hydroxybutyryl-CoA reductase (alcohol formation), were cloned into the pZE13 vector (Expressys, Ruelzheim, Germany) under the PA1 / lacO promoter. This plasmid was transformed into a recombinant Escherichia coli strain that produces 4-hydroxybutyryl-CoA to express the proteins and enzymes required for the synthesis of 1,3-butanediol, which originates from this metabolite.

[0375] The genetically engineered organism was cultured in a glucose-containing medium following procedures well known in the art (see, for example, Sambrook et al., ibid., 2001). Expression of the pathway gene was consolidated using methods well known in the art for determining peptide expression or enzyme activity (including, for example, Northern blotting, PCR amplification of mRNA, and Western blotting). The enzyme activity of the expressed enzyme was determined using assays specific to individual activity. The ability of the engineered *E. coli* strain to produce 1,3-butanediol was determined using HPLC, gas chromatography-mass spectrometry (GC-MS), or liquid chromatography-mass spectrometry (LC-MS).

[0376] The efficiency of the functional 1,3-butanediol biosynthesis pathway was further enhanced by optimizing its utilization. In short, the engineered strain was evaluated to determine whether any exogenous genes were expressed at a rate-limited level. For those expressed at low levels, expression could be increased by, for example, by introducing additional gene copy number-limiting enzymes that restrict the flux through the pathway.

[0377] To generate better producers, metabolic modeling can be used to optimize culture conditions. Modeling has also been used to design gene knockouts for additional optimized pathway applications (see, for example, U.S. Patent Publications US 2002 / 0012939, US 2003 / 0224363, US 2004 / 0029149, US 2004 / 0072723, US 2003 / 0059792, US 2002 / 0168654, and US 2004 / 0009466, and see U.S. Patent No. 7,127,379). Modeling analysis allows for reliable prediction of the effects of shifted metabolism on cell growth for more efficient production of 1,3-butanediol. One modeling approach is the two-layer optimization method OptKnock (Burgard et al., Biotechnol. Bioengineer. 84:647-657 (2003)), which is used to select for gene knockouts that collectively lead to better production of 1,3-butanediol. Adaptive evolution can also be used to generate, for example, better producers of acetyl-CoA intermediates or 1,3-butanediol products. Adaptive evolution is performed to improve growth and production characteristics (Fong and Palsson, Nat. Genet. 36:1056-1058 (2004); Alper et al., Science 314:1565-1568 (2006)). Based on the results, subsequent rounds of modeling, genetic engineering, and adaptive evolution can be applied to 1,3-butanediol producers to further increase production.

[0378] To achieve large-scale production of 1,3-butanediol, recombinant organisms were cultured in fermenters using media known in the art to support anaerobic growth. Fermentation was carried out in batch, fed-batch, or continuous manner. Anaerobic conditions were maintained by first spraying the medium with nitrogen and then sealing the culture vessels (e.g., flasks sealed with gaskets and clamp caps). Microaerobic conditions could also be employed by providing limited aeration through small holes. The pH of the medium was maintained at pH 7 by adding an acid (e.g., H₂SO₄). The growth rate was determined by measuring the optical density using a spectrophotometer (600 nm), and the glucose uptake rate was determined by monitoring the consumption of the carbon source over time. Byproducts (such as unwanted alcohols, organic acids, and residual glucose) could be quantified using HPLC (Shimadzu) with an HPX-087 column (BioRad) using refractive index detectors for glucose and alcohols and UV detectors for organic acids (Lin et al., Biotechnol. Bioeng., 90:775-779 (2005)).

[0379]

[0380]

[0381]

[0382]

[0383]

[0384]

[0385]

[0386]

[0387]

[0388]

[0389]

[0390]

[0391]

[0392]

[0393]

[0394]

[0395]

[0396]

[0397] Although the invention has been described with reference to the disclosed embodiments, those skilled in the art will readily understand that the specific examples and studies detailed above are merely illustrative. It should be understood that various modifications can be made without departing from the spirit of the invention. Therefore, the invention is limited only by the following claims.

Claims

1. A non-naturally occurring bacterium possessing the 1,3-butanediol (1,3-BDO) pathway, wherein the bacterium comprises: (a) At least three exogenous nucleic acids encoding 1,3-BDO pathway enzymes, said 1,3-BDO pathway enzymes comprising (1) a 4-hydroxybutyryl-CoA dehydratase that converts 4-hydroxybutyryl-CoA to crotonyl-CoA; (2) a crotonylase that converts crotonyl-CoA to 3-hydroxybutyryl-CoA; and (3) an alcohol-forming 3-hydroxybutyryl-CoA reductase that converts 3-hydroxybutyryl-CoA to 1,3-butanediol; or (b) At least four exogenous nucleic acids encoding 1,3-BDO pathway enzymes, said 1,3-BDO pathway enzymes comprising (1) a 4-hydroxybutyryl-CoA dehydratase that converts 4-hydroxybutyryl-CoA to crotonyl-CoA; (2) a crotonylase that converts crotonyl-CoA to 3-hydroxybutyryl-CoA; (3) an aldehyde-forming 3-hydroxybutyryl-CoA reductase that converts 3-hydroxybutyryl-CoA to 3-hydroxybutyraldehyde; and (4) a 3-hydroxybutyraldehyde reductase that converts 3-hydroxybutyraldehyde to 1,3-butanediol. The 1,3-BDO pathway enzymes are expressed in sufficient quantities to produce 1,3-BDO.

2. The non-naturally occurring bacteria according to claim 1, wherein at least one of the exogenous nucleic acids is a heterologous nucleic acid.

3. The non-naturally occurring bacteria according to claim 1, wherein the non-naturally occurring bacteria are in a culture medium containing less than 10% dissolved oxygen saturation.

4. The non-naturally occurring bacteria according to claim 1, wherein the alcohol-forming 3-hydroxybutyryl-CoA reductase is encoded by one or more genes selected from the group consisting of: adhE, adhE2, mcr, Rcas_2929, NAP1_02720, MGP2080_ 00535 and FAR .

5. The non-naturally occurring bacteria according to claim 1, wherein the 3-hydroxybutyraldehyde reductase is encoded by one or more genes selected from the group consisting of: alrA , ADH2 , yqhD , bdh I , bdh II , adhA , 4hbd , adhI , P84067 , mmsb , dhat and 3hidh .

6. The non-naturally occurring bacteria according to claim 1, wherein the aldehyde-forming 3-hydroxybutyryl-CoA reductase is encoded by one or more genes selected from the group consisting of: acr1 , sucD , bphG , bld , adhE , Msed_0709 , mcr , asd-2 , Saci_2370 , Ald and eutE .

7. The non-naturally occurring bacteria according to claim 1, wherein the 4-hydroxybutyryl-CoA dehydrase is encoded by one or more genes selected from the group consisting of: fumA , fumB , fumC , fumH , fum1 , MmcB , MmcC , hmd , BACCAP 02294 , ANACOL_02527 , NtherDRAFT_2368 , dmdA , dmdB , crt , crt1 , echpaaA , paaB , phaA , phaB , maoC , paaF , paaG , abfD , Msed_1220 , fadA , fadB , fadI , fadJ and fadR .

8. The non-naturally occurring bacteria according to claim 1, wherein the crotonase is encoded by one or more genes selected from the group consisting of: fumA , fumB , fumC , fumH , fum1 , MmcB , MmcC , hmd , BACCAP _02294 , ANACOL_ 02527 , NtherDRAFT_2368 , dmdA , dmdB , crt , crt1 , echpaaA , paaB , phaA , phaB , maoC , paaF , paaG , abfD , Msed_1220 , fadA , fadB , fadI , fadJ and fadR .

9. The non-naturally occurring bacteria according to claim 8, wherein the bacteria is Escherichia coli.

10. A culture medium containing the non-naturally occurring bacteria as described in claim 1.

11. A composition comprising the culture medium according to claim 10 and biosynthesized 1,3-BDO.

12. A method for producing 1,3-BDO, said method comprising culturing non-naturally occurring bacteria possessing the 1,3-butanediol (1,3-BDO) pathway under various conditions for a sufficient duration to produce 1,3-BDO. The bacteria mentioned above include: (a) At least three exogenous nucleic acids encoding 1,3-BDO pathway enzymes, said 1,3-BDO pathway enzymes comprising (1) a 4-hydroxybutyryl-CoA dehydratase that converts 4-hydroxybutyryl-CoA to crotonyl-CoA; (2) a crotonylase that converts crotonyl-CoA to 3-hydroxybutyryl-CoA; and (3) an alcohol-forming 3-hydroxybutyryl-CoA reductase that converts 3-hydroxybutyryl-CoA to 1,3-butanediol; or (b) At least four exogenous nucleic acids encoding 1,3-BDO pathway enzymes, said 1,3-BDO pathway enzymes comprising (1) a 4-hydroxybutyryl-CoA dehydratase that converts 4-hydroxybutyryl-CoA to crotonyl-CoA; (2) a crotonylase that converts crotonyl-CoA to 3-hydroxybutyryl-CoA; (3) an aldehyde-forming 3-hydroxybutyryl-CoA reductase that converts 3-hydroxybutyryl-CoA to 3-hydroxybutyraldehyde; and (4) a 3-hydroxybutyraldehyde reductase that converts 3-hydroxybutyraldehyde to 1,3-butanediol. The 1,3-BDO pathway enzymes are expressed in sufficient quantities to produce 1,3-BDO.

13. The method of claim 12, wherein the non-naturally occurring bacteria are in a culture medium containing less than 10% dissolved oxygen saturation.

14. The method of claim 12, wherein the alcohol-forming 3-hydroxybutyryl-CoA reductase is encoded by one or more genes selected from the group consisting of: adhE, adhE2, mcr, Rcas_2929, NAP1_02720, MGP2080_00535 and FAR .

15. The method of claim 12, wherein the 3-hydroxybutyraldehyde reductase is encoded by one or more genes selected from the group consisting of: alrA , ADH2 , yqhD , bdh I, bdh II. adhA , 4hbd , adhI , P84067 , mmsb , dhat and 3hidh.

16. The method of claim 12, wherein the aldehyde-forming 3-hydroxybutyryl-CoA reductase is encoded by one or more genes selected from the group consisting of: acr1 , sucD, bphG, bld , adhE , Msed_0709 , mcr , asd-2 , Saci_2370 , Ald ,and eutE .

17. The method of claim 12, wherein the 4-hydroxybutyryl-CoA dehydratase is encoded by one or more genes selected from the group consisting of: fumA , fumB , fumC , fumH , fum1 , MmcB , MmcC , hmd , BACCAP_02294 , ANACOL_ 02527 , NtherDRAFT_2368 , dmdA , dmdB , crt, crt1, echpaaA , paaB , phaA , phaB , maoC , paaF , paaG , abfD , Msed_1220、fadA , fadB , fadI , fadJ and fadR .

18. The method of claim 12, wherein the crotonase is encoded by one or more genes selected from the group consisting of: fumA , fumB , fumC , fumH , fum1 , MmcB , MmcC , hmd , BACCAP_02294 , ANACOL_02527 , NtherDRAFT_2368 , dmdA , dmdB , crt, crt1, echpaaA , paaB , phaA , phaB , maoC , paaF , paaG , abfD , Msed_1220 , fadA , fadB , fadI , fadJ and fadR.

19. The method of claim 12, further comprising isolating 1,3-BDO from the culture.

20. The method according to claim 19, wherein the separation comprises extraction, total evaporation, membrane separation, reverse osmosis, electrodialysis, distillation, crystallization, centrifugation, extraction filtration, ion exchange chromatography, adsorption chromatography, and ultrafiltration.

21. The method of claim 20, wherein the separation comprises distillation or membrane filtration.

22. The method of claim 12, further comprising converting the 1,3-BDO into a compound, polymer, or other product; wherein the compound, polymer, or other product is butadiene, rubber, tire, latex, or resin.

Images