IL130789A - Cosmid comprising a dna fragment from klebsiella pneumoniae, contained in a host bacterium - Google Patents

Abstract

A cosmid contained in ATCC 69789 comprising a DNA fragment of about 35 kb isolated from Klebsiella pneumonia wherein (1) the DNA fragment encodes an active glycerol dehydratase enzyme and (2) digestion of the cosmid results in a restriction digest pattern as shown in Fig. 1, columns 1 and 2 of the specification. 881 ט' באייר התשס" ב - April 21, 2002

Description

130789/2 •>Np7ii£j ρτ»η3 Djn wjioiiD n i^pn DNA ujono oon i>ntnp A COSMID COMPRISING A DNA FRAGMENT FROM KLEBSIELLA PNEUMONIAE, CONTAINED IN A HOST BACTERIUM E.I. DU PONT DE NEMOURS & COMPANY GENENCOR INTERNATIONAL INC.

C: 34773 TITLE BIOCONVERSION OF A FERMENTABLE CARBON SOURCE TO 1,3-PROPANEDIOL BY A SINGLE MICROORGANISM FIELD OF INVENTION This invention comprises a process for the bioconversion of a fermentable carbon source to 1 ,3-propanediol by a single microorganism.

BACKGROUND 1 ,3-PropanedioI is a monomer having potential utility in the production of polyester fibers and the manufacture of polyurethanes and cyclic compounds.

A variety of chemical routes to 1,3-propanediol are known. For example ethylene oxide may be converted to 1,3-propanediol over a catalyst in the presence of phosphine, water, carbon monoxide, hydrogen and an acid, by the catalytic solution phase hydration of acrolein followed by reduction, or from hydrocarbons such as glycerol, reacted in the presence of carbon monoxide and hydrogen over catalysts having atoms from group Vm of the periodic table. Although it is possible to generate 1 ,3-propanediol by these methods, they are expensive and generate waste streams containing environmental pollutants.

It has been known for over a century that 1 ,3-propanediol can be produced from the fermentation of glycerol. Bacterial strains able to produce 1,3-propanediol have been found, for example, in the groups Citrobacter, Clostridium, Enterobacter, Ilyobacter, Klebsiella, Lactobacillus, and Pelobacter. In each case studied, glycerol is converted to 1,3-propanediol in a two step, enzyme catalyzed reaction sequence. In the first step, a dehydratase catalyzes the conversion of glycerol to 3-hydroxypropionaldehyde (3-HP) and water, Equation 1. In the second step, 3-HP is reduced to 1,3-propanediol by a NAD+-linked oxidoreductase, Equation 2. The 1,3-propanediol is not metabOlized further and, as a result, Glycerol → 3-HP + H20 (Equation 1) 3-HP + NADH + H+ → 1 ,3-Propanediol + NAD+ (Equation 2) accumulates in high concentration in the media. The overall reaction consumes a reducing equivalent in the form of a cofactor, reduced β-nicotinamide adenine dinucleotide (NADH), which is oxidized to nicotinamide adenine dinucleotide (NAD+).

The production of 1,3-propanediol from glycerol is generally performed under anaerobic conditions using glycerol as the sole carbon source and in the absence of other exogenous reducing equivalent acceptors. Under these conditions, in e.g., strains of Citrobacter, Clostridium, and Klebsiella, a parallel I pathway for glycerol operates which first involves oxidation of glycerol to dihydroxyacetone (DHA) by a NAD+- (or NADP+-) linked glycerol dehydrogenase, Equation 3. The DHA, following phosphorylation to dihydroxyacetone phosphate (DHAP) by a DHA kinase (Equation 4), Glycerol + NAD+ → DHA + NADH + H+ (Equation 3) DHA + ATP → DHAP + ADP (Equation 4) becomes available for biosynthesis and for supporting ATP generation via e.g., glycolysis. In contrast to the 1 ,3-propanediol pathway, this pathway may provide carbon and energy to the cell and produces rather than consumes NADH.

In Klebsiella pneumoniae and Citrobacter freundii, the genes encoding the functionally linked activities of glycerol dehydratase (dnaB), 1 ,3-propanediol oxidoreductase (dhaT), glycerol dehydrogenase (dhaD), and dihydroxyacetone kinase (dnaK) are encompassed by the dha regulon. The dha regulons from Citrobacter and Klebsiella have been expressed in Escherichia coli and have been shown to convert glycerol to 1,3-propanediol.

Biological processes for the preparation of glycerol are known. The overwhelrning majority of glycerol producers are yeasts but some bacteria, other fungi and algae are also known. Both bacteria and yeasts produce glycerol by converting glucose or other carbohydrates through the fructose- 1,6-bisphosphate pathway in glycolysis or the Embden Meyerhof Pamas pathway, whereas, certain algae convert dissolved carbon dioxide or bicarbonate in the chloroplasts into the 3-carbon intermediates of the Calvin cycle. In a series of steps, the 3-carbon intermediate, phosphoglyceric acid, is converted to glyceraldehyde 3-phosphate which can be readily interconverted to its keto isomer dihydroxyacetone phosphate and ultimately to glycerol. Although biological methods of both glycerol and 1,3-propanediol production are known, it has never been demonstrated that the entire process can be accomplished by a single organism.

Neither the chemical nor biological methods described above for the production of 1,3-propanediol is well suited for industrial scale production since the chemical processes are energy intensive and the biological processes require the expensive starting material, glycerol. A method requiring low energy input and an inexpensive starting material is needed. A more desirable process would incorporate a microorganism that would have the ability to convert basic carbon sources such as carbohydrates, or sugars to the desired 1 ,3-propanediol end-product.

Although a single organism conversion of fermentable carbon source other than glycerol or dihydroxyacetone to 1,3-propanediol would be desirable, it has been documented that there are significant difficulties to overcome in such an endeavor. For example, Gottschalk et al. (EP 373 230) teach that the growth of most strains useful for the production of 1,3-propanediol, including Citrobacter freun ii, Clostridium acetobutylicum, Clostridium butyricum, and Klebsiella pneumoniae, is disturbed by the presence of a hydrogen donor such as fructose or glucose. Strains of Lactobacillus brevis and Lactobacillus buchner, which produce 1,3-propanediol in co-fermentations of glycerol and fructose or glucose, do not grow when glycerol is provided as the sole carbon source, and, although it has been shown that resting cells can metabolize glucose or fructose, they do not produce 1,3-propanediol. (Veiga DA Cunha et al., J. Bacteriol. 174, 1013 (1992)). Similarly, it has been shown that a strain of Ilyobacter polytropus, which produces 1,3-propanediol when glycerol and acetate are provided, will not produce 1,3-propanediol from carbon substrates other than glycerol, including fructose and glucose. (Steib et al., Arch. Microbiol. 140, 139 (1984)). Finally Tong et al. (Appl. Biochem. Biotech. 34, 149 (1992)) has taught that recombinant Escherichia coli transformed with the dha regulon encoding glycerol dehydratase does not produce 1,3-propanediol from either glucose or xylose in the absence of exogenous glycerol.

Attempts to improve the yield of 1,3-propanediol from glycerol have been reported where co-substrates capable of providing reducing equivalents, typically fermentable sugars, are included in the process. Improvements in yield have been claimed for resting cells of Citrobacter freundii and Klebsiella pneumoniae DSM 4270 cofermenting glycerol and glucose (Gottschalk et al., supra.; and Tran-Dinh et al., DE 3734 764); but not for growing cells of Klebsiella pneumoniae ATCC 25955 cofermenting glycerol and glucose, which produced no 1,3-propanediol (I-T. Tong, Ph.D. Thesis, University of Wisconsin-Madison (1992)). Increased yields have been reported for the cofermentation of glycerol and glucose or fructose by a recombinant Escherichia coli, however, no 1,3-propanediol is produced in the absence of glycerol (Tong et al., supra.). In these systems, single organisms use the carbohydrate as a source of generating NADH while providing energy and carbon for cell maintenance or growth. These disclosures suggest that sugars do not enter the carbon stream that produces 1,3-propanediol. In no case is 1,3-propanediol produced in the absence of an exogenous source of glycerol. Thus the weight of literature clearly suggests that the production of 1,3-propanediol from a carbohydrate source by a single organism is not possible.

The problem to be solved by the present invention is the biological production of 1,3-propanediol by a single organism from an inexpensive carbon substrate such as glucose or other sugars. The biological production of 1,3-propanediol requires glycerol as a substrate for a two step sequential reaction in which a dehydratase enzyme (typically a coenzyme B ^-dependent dehydratase) converts glycerol to an intermediate, 3-hydroxypropionaldehyde, which is then reduced to 1,3-propanediol by a NADH- (or NADPH) dependent oxidoreductase. The complexity of the cofactor requirements necessitates the use pf a whole cell catalyst for an industrial process which utilizes this reaction sequence for the production of 1,3-propanediol. Furthermore, in order to make the process economically viable, a less expensive feedstock than glycerol or dihydroxyacetone is needed. Glucose and other carbohydrates are suitable substrates, but, as discussed above, are known to interfere with 1,3-propanediol production. As a result no single organism has been shown to convert glucose to 1,3-propanediol.

Applicants have solved the stated problem and the present invention provides for bioconverting a fermentable carbon source directly to 1,3-propanediol using a single organism. Glucose is used as a model substrate and the bioconversion is applicable to any existing microorganism.

Microorganisms harboring the gene for a dehydratase are able to convert glucose and other sugars through the glycerol degradation pathway to 1,3-propanediol with good yields and selectivities. Furthermore, the present invention may be generally applied to include any carbon substrate that is readily converted to glycerol, dihydroxyacetone, or C3 compounds at the oxidation state of glycerol (e.g., glycerol 3-phosphate) or dihydroxyacetone (e.g., dihydroxyacetone phosphate or glyceraldehyde 3-phosphate).

SUMMARY OF THE INVENTION The present invention comprises a process for the bioconversion of a carbon substrate to 1 ,3-propanediol by a single microorganism having at least one gene capable of expressing a dehydratase enzyme by contacting said microorganism with said substrate. The microorganism can be a wild type, or genetically altered, such as a recombinant microorganism or a mutant of a microorganism. Preferably, the dehydratase enzyme is a glycerol dehydratase enzyme or a diol dehydratase enzyme.

The present invention further comprises the product of the above process.

The present invention further comprises a cosmid comprising a DNA fragment of about 35 kb isolated from Klebsiella pneumoniae wherein said fragment encodes an active glycerol dehydratase enzyme having the restriction digest in Figure 1, columns 1 and 2. This cosmid, when transferred into a microorganism permits metabolism of a carbon substrate, in particular glucose, to 1,3-propanediol.

The present invention further comprises a transformed microorganism comprising a host microorganism and the above cosmid or any DNA fragment of said cosmid encoding an active functional protein other than a glycerol dehydratase enzyme.

BRIEF DESCRIPTION OF THE FIGURES Figure 1 shows restriction digests (EcoR 1, BamH 1, EcoR V and Notl) of cosmids pkPl, pKP2 and pKP4 labeled as columns 1, 2 and 4, respectively, and separation on a 0.8% agarose gel electrophoresis. Molecular size markers were loaded on the lanes in the end. Columns labeled as numbers 1 and 2 represent cosmids containing a glycerol dehydratase enzyme.

Figure 2 shows a partial physical map of pKPl and the position of the genes based on DNA sequence. The genes were identified based on comparison of deduced open reading frames with the Genbank data base using the Tfasta program provided by a sequence analysis software of the University of Wisconsin [Genetics Computer Group, Verison 7, April, 1991, 575 Science Drive, Madison, WI 53711].

BRIEF DESCRIPTION OF BIOLOGICAL DEPOSITS The transformed E. coli DH5a containing cosmid pKPl containing a portion of the Klebsiella genome encoding the glycerol dehydratase enzyme was deposited on 18 April 1 95 with the ATCC under the terms of the Budapest Treaty and is identified by the ATCC number ATCC 69789. The transformed E. coli DH5a containing cosmid pKP4 containing a portion of the Klebsiella genome encoding a diol dehydratase enzyme was deposited on 18 April 1995 with the ATCC under the terms of the Budapest Treaty and is identified by the ATCC number ATCC 69790. A recombinant Pseudomonas sp. comprising a DNA fragment encoding dhaBl, dhaB2 and dhaB3 and dhaT was deposited on April 11, 1996 with the ATCC under the terms of the Budapest Treaty and is designated by the ATCC Accession number ATCC 55670. A recombinant Pichia pastoris strain comprising a DNA fragment encoding dhaBl, dhaB2 and dhaB3 and dhaT was deposited on April 11, 1996 with the ATCC under the terms of the Budapest Treaty and is designated by the ATCC Accession number ATCC 74363. "ATCC" refers to the American Type Culture Collection international depository located at 12301 Parklawn Drive, Rockville, MD 20852 U.S.A.

DETAILED DESCRIPTION OF THE INVENTION As used herein the following terms may be used for interpretation of the claims and specification.

The term "genetically altered or genetically altered microorganism" refers to any microorganism, suitable for use in the present invention, which has undergone an alteration of the native genetic machinery of the microorganism. Microorganisms may be genetically altered by undergoing transformation by vectors comprising heterologous nucleic acid fragments, mutagenesis with mutagenizing agents (e.g., UV light, ethanesulfonic acid) or any other method whereby stable alterations of the cell genome occur.

The term "construct" refers to a plasmid, virus, autonomously replicating sequence, genome integrating sequence, phage or nucleotide sequence, linear or circular, of a single- or double-stranded DNA or RNA, derived from any source, in which a number of nucleotide sequences have been joined or recombined into a unique construction which is capable of introducing a promoter fragment and DNA sequence for a selected gene product along with appropriate 3' untranslated sequence into a cell.

The term "transformation" or "transfection" refers to the acquisition of new genes in a cell after the incorporation of nucleic acid. The acquired genes may be integrated into chromosomal DNA or introduced as extrachromosomal replicating sequences. The term "transformant" refers to the product of a transformation. The term "genetically altered" refers to the process of changing hereditary material by transformation or mutation.

The term "expression" refers to the transcription and translation to gene product from a gene coding for the sequence of the gene product.

The term "plasmid" or "vector" or "cosmid" as used herein refers to an extra chromosomal element often carrying genes which are not part of the central metabolism of the cell, and usually in the form of circular double-stranded DNA molecules.

'· The term "dehydratase enzyme" will refer to any enzyme that is capable of isomerizing or converting a glycerol molecule to the product 3-hydroxypropion-aldehyde. For the purposes of the present invention the dehydratase enzymes include a glycerol dehydratase and a diol dehydratase having preferred substrates of glycerol and 1,2-propanediol, respectively.

The term "carbon substrate" or "carbon source" means any carbon source capable of being metabolized by a microorganism wherein the substrate contains at least one carbon atom, provided that the carbon substrate is other than glycerol or dihydroxy acetone.

The term "ATCC" will stand for the American Type Culture Collection, 12301 Parklawn Drive, Rockville, MD 20852, U.S.A.

The present invention provides a method for a biological production of 1,3-propanediol from a fermentable carbon source in a single organism. The method incorporates a microorganism containing a dehydratase enzyme which is contacted with a carbon substrate and 1,3-propanediol is isolated from the growth media. The single organism may be a wild type organism or may be a genetically altered organism harboring a gene encoding a dehydratase enzyme.

, The present method provides a rapid, inexpensive and environmentally responsible source of 1,3-propanediol monomer useful in the production of polyesters and other polymers.

Construction of Recombinant Organisms: Recombinant organisms containing the necessary genes that will encode the enzymatic pathway for the conversion of a carbon substrate to 1,3-propanediol may be constructed using techniques well known in the art. In the present invention genes encoding dehydratase enzyme were isolated from a native host such as Klebsiella and used to transform the E. coli host strains DH5a, ECL707 and AA200.

Methods of obtaining desired genes from a bacterial genome are common and well known in the art of molecular biology. For example, if the sequence of the gene is known, suitable genomic libraries may be created by restriction endonuclease digestion and may be screened with probes complementary to the desired gene sequence. Once the sequence is isolated, the DNA may be amplified using standard primer directed amplification methods such as polymerase chain reaction (U.S. 4,683,202) to obtain amounts of DNA suitable for transformation using appropriate vectors.

Alternatively, cosmid libraries may be created where large segments of genomic DNA (35-45kb) may be packaged into vectors and used to transform appropriate hosts. Cosmid vectors are unique in being able to accommodate large quantities of DNA. Generally cosmid vectors have at least one copy of the cos DNA sequence which is needed for packaging and subsequent circularization of the foreign DNA. In addition to the cos sequence these vectors will also contain an origin of replication such as ColEl and drug resistance markers such as a gene resistant to ampicillin or neomycin. Methods of using cosmid vectors for the transformation of suitable bacterial hosts are well described in Sambrook, J. et al., Molecular Cloning: A Laboratory Manual. Second Edition (1989) Cold Spring Harbor Laboratory Press, herein incorporated by reference.

Typically to clone cosmids, foreign DNA is isolated and ligated, using the appropriate restriction endonucleases, adjacent to the cos region of the cosmid vector. Cosmid vectors containing the linearized foreign DNA is then reacted with a DNA packaging vehicle such as bacteriophage λ. During the packaging process the cos sites are cleaved and the foreign DNA is packaged into the head portion of the bacterial viral particle. These particles are then used to transfect suitable host cells such as E. coli. Once injected into the cell, the foreign DNA circularizes under the influence of the cos sticky ends. In this manner large segments of foreign DNA can be introduced and expressed in recombinant host #cells.

Cosmid vectors and cosrnid transformation methods were used within the context of the present invention to clone large segments of genomic DNA from bacterial genera known to possess genes capable of processing glycerol to 1,3-propanediol. Specifically, genomic DNA from K. pneumoniae was isolated by methods well known in the art and digested with the restriction enzyme Sau3A for insertion into a cosmid vector Supercos 1™ and packaged using GigapackU packaging extracts. Following construction of the vector E. coli XLl-Blue MR cells were transformed with the cosmid DNA. Transformants were screened for the ability to convert glycerol to 1 ,3-propanediol by growing the cells in the presence of glycerol and analyzing the media for 1 ,3-propanediol formation.

Two of the 1,3-propanediol positive transformants were analyzed and the cosmids were named pKPl and pKP2. DNA sequencing revealed extensive homology to the glycerol dehydratase gene from C.freundii, demonstrating that these transformants contained DNA encoding the glycerol dehydratase gene. Other 1,3-propanediol positive transformant were analyzed and the cosmids were named pKP4 and pKP5. DNA sequencing revealed that these cosmids carried DNA encoding a diol dehydratase gene.

Although the instant invention utilizes the isolated genes from within a Klebsiella cosmid, alternate sources of dehydratase genes include, but are not limited to, Citrobacter, Clostridia, and Salmonella.

Other genes that will positively affect the production of 1,3-propanediol may be expressed in suitable hosts. For example it may be highly desirable to over-express certain enzymes in the glycerol degradation pathway and/or other pathways at levels far higher than currently found in wild type cells. This may be accomplished by the selective cloning of the genes encoding those enzymes into multicopy plasmids or placing those genes under a strong inducible or constitutive promoter. Methods for over-expressing desired proteins are common and well known in the art of molecular biology and examples may be found in Sambrook, supra. Furthermore, specific deletion of certain genes by methods known to those skilled in the art will positively affect the production of 1,3-propanediol.

Examples of such methods can be found in Methods in Enzymology, Volume 217, R. Wu editor, Academic Press:San Diego 1993.

Mutants: In addition to the cells exemplified it is contemplated that the present method will be able to make use of cells having single or multiple mutations specifically designed to enhance the production of 1,3-propanediol. Cells that normally divert a carbon feed stock into non-productive pathways, or that exhibit significant catabolite repression could be mutated to avoid these phenotypic deficiencies. For example, many wild type cells are subject to catabolite repression from glucose and by-products in the media and it is contemplated that mutant strains of these wild type organisms, capable of 1,3-propanediol production that are resistant to glucose repression, would be particularly useful in the present invention.

Methods of creating mutants are common and well known in the art. For example, wild type cells may be exposed to a variety of agents such as radiation or chemical mutagens and then screened for the desired phenotype. When creating mutations through radiation either ultraviolet (UV) or ionizing radiation may be used. Suitable short wave UV wavelengths for genetic mutations will fall within the range of 200 nm to 300 nm where 254 nm is preferred. UV radiation in this wavelength principally causes changes within nucleic acid sequence from guanidine and cytosine to adenine and thymidine. Since all cells have DNA repair mechanisms that would repair most UV induced mutations, agents such as caffeine and other inhibitors may be added to interrupt the repair process and maximize the number of effective mutations. Long wave UV mutations using light in the 300 nm to 400 nm range are also possible but are generally not as effective as the short wave UV light unless used in conjunction with various activak>rs such as psoralen dyes that interact with the DNA.

Mutagenesis with chemical agents is also effective for generating mutants and commonly used substances include chemicals that affect nonreplicating DNA such as HN02 and NH2OH, as well as agents that affect replicating DNA such as acridine dyes, notable for causing frameshift mutations. Specific methods for creating mutants using radiation or chemical agents are well documented in the art. See for example Thomas D. Brock in Biotechnology: A Textbook of Industrial Microbiology. Second Edition ( 1989) Sinauer Associates, Inc., Sunderland, MA., or Deshpande, Mukund V., Appl. Biochem. BiotechnoL, 36, 227, (1992), herein incorporated by reference.

After mutagenesis has occurred, mutants having the desired phenotype may be selected by a variety of methods. Random screening is most common where the mutagenized cells are selected for the ability to produce the desired product or intermediate. Alternatively, selective isolation of mutants can be performed by growing a mutagenized population on selective media where only resistant colonies can develop. Methods of mutant selection are highly developed and well known in the art of industrial microbiology. See Brock, Supra,, DeMancilha et al.. Food Chem., 14, 313, (1984).

Mutations and transformations in the 1.3-propanediol production pathway: Representative enzyme pathway. The production of 1, 3-propanediol from glucose can be accomplished by the following series of steps. This series is representative of a number of pathways known to those skilled in the art. Glucose is convened in a series of steps by enzymes of the glycolytic pathway to dihydroxyacetone phosphate (DHAP) and 3-phosphoglyceraldehyde (3-PG). Glycerol is then formed by either hydrolysis of DHAP to dihydroxyacetone (DHA) followed by reduction, or reduction of DHAP to glycerol 3-phosphate (G3P) followed by hydrolysis. The hydrolysis step can be catalyzed by any number of cellular phosphatases which are known to be non-specific with respect to their substrates or the activity can be introduced into the host by recombination. The reduction step can be catalyzed by a NAD+ (or NADP+) linked host enzyme or the activity can be introduced into the host by recombination. It is notable that the dha regulon contains a glycerol dehydrogenase (E.C. 1.1.1.6) which catalyzes the reversible reaction of Equation 3.

Glycerol → 3-HP + H20 (Equation 1) 3-HP + NADH + H+ → 1,3-Propanediol + NAD+ (Equation 2) Glycerol + NAD+ → DHA + NADH + H+ (Equation 3) Glycerol is converted to 1, 3-propanediol via the intermediate 3-hydroxy-propionaldehye (3-HP) as has been described in detail above. The intermediate 3-HP is produced from glycerol, Equation 1, by a dehydratase enzyme which can be encoded by the host or can introduced into the host by recombination. This dehydratase can be glycerol dehydratase (E.C. 4.2.1.30), diol dehydratase (E.C. 4.2.1.28) or any other enzyme able to catalyze this transformation.

Glycerol dehydratase, but not diol dehydratase, is encoded by the dha regulon. 1,3-Propanediol is produced from 3-HP, Equation 2, by a NAD+- (or NADP+) linked host enzyme or the activity can introduced into the host by recombination. This final reaction in the production of 1, 3-propanediol can be catalyzed by 1, 3-propanediol dehydrogenase (E.C. 1.1.1.202) or other alcohol dehydrogenases.

Mutations and transformations that affect carbon channeling. A variety of mutant organisms comprising variations in the 1 ,3 -propanediol production pathway will be useful in the present invention. For example the introduction of a triosephosphate isomerase mutation (tpi-) into the microorganism of the present invention is an example of the use of a mutation to improve the performance by carbon channeling. The mutation can be directed toward a structural gene so as to impair or improve the activity of an enzymatic activity or can be directed toward a regulatory gene so as to modulate the expression level of an enzymatic activity.

Alternatively, transformations and mutations can be combined so as to control particular enzyme activities for the enhancement of 1,3-propanediol production. Thus it is within the scope of the present invention to anticipate modifications of a whole cell catalyst which lead to an increased production of 1,3-propanediol.

Media and Carbon Substrates: Fermentation media in the present invention must contain suitable carbon substrates. Suitable substrates may include but are not limited to monosaccharides such as glucose and fructose, oligosaccharides such as lactose or sucrose, polysaccharides such as starch or cellulose or mixtures thereof and unpurified mixtures from renewable feedstocks such as cheese whey permeate, cornsteep liquor, sugar beet molasses, and barley malt. Additionally the carbon substrate may also be one-carbon substrates such as carbon dioxide, or methanol for which metabolic conversion into key biochemical intermediates has been demonstrated. Glycerol production from single carbon sources (e.g., methanol, formaldehyde or formate) has been reported in methylotrophic yeasts (K. Yamada et al., Agric. Biol Chem., 53(2) 541-543, (1989)) and in bacteria (Hunter et.al., Biochemistry, 24, 4148-4155, (1985)). These organisms can assimilate single carbon compounds, ranging in oxidation state from methane to formate, and produce glycerol. The pathway of carbon assimilation can be through ribulose monophosphate, through serine, or through xylulose-monophosphate (Gottschalk, Bacterial Metabolism. Second Edition, Springer- Verlag: New York (1986)). The ribulose monophosphate pathway involves the condensation of formate with ribulose-5-phosphate to form a 6 carbon sugar that becomes fructose and eventually the three carbon product glyceraldehyde-3-phosphate. Likewise, the serine pathway assimilates the one-carbon compound into the glycolytic pathway via methylenetetrahydrofolate.

In addition to one and two carbon substrates methylotrophic organisms are also known to utilize a number of other carbon containing compounds such as methylamine, glucosamine and a variety of amino acids for metabolic activity. For example, methylotrophic yeast are known to utilize the carbon from methylamine to form trehalose or glycerol (Bellion et al., Microb. Growth CI Compd., [Int. Symp.], 7th (1993), 415-32. Editor(s): Murrell, J. Collin; Kelly, Don P. Publisher: Intercept, Andover, UK). Similarly, various species of Candida will metabolize alanine or oleic acid (Suiter et al., Arch. Microbiol. (1990), 153(5), 485-9). Hence it is contemplated that the source of carbon utilized in the present invention may encompass a wide variety of carbon containing substrates and will only be limited by the choice of organism.

, Although it is contemplated that all of the above mentioned carbon substrates and mixtures thereof are suitable in the present invention, preferred carbon substrates are glucose, fructose, sucrose or methanol.

In addition to an appropriate carbon source, fermentation media must contain suitable minerals, salts, cofactors, buffers and other components, known to those skilled in the art, suitable for the growth of the cultures and promotion of the enzymatic pathway necessary for 1 ,3-propanediol production. Particular attention is given to Co(II) salts and/or vitamin B 12 or precursors thereof.

Culture Conditions: Typically cells are grown at 30°C in appropriate media. Preferred growth media in the present invention are common commercially prepared media such as Luria Bertani (LB) broth, Sabouraud Dextrose (SD) broth or Yeast medium (YM) broth. Other defined or synthetic growth media may also be used and the appropriate medium for growth of the particular microorganism will be known by someone skilled in the art of microbiology or fermentation science. The use of agents known to modulate catabolite repression directly or indirectly, e.g., cyclic adenosine 2':3'-monophosphate, may also be incorporated into the reaction media. Similarly, the use of agents known to modulate enzymatic activities (e.g., methyl viologen) that lead to enhancement of 1,3-propanediol production may be used in conjunction with or as an alternative to genetic manipulations.

'· Suitable pH ranges for the fermentation are between pH 5.0 to pH 9.0 where pH 6.0 to pH 8.0 is preferred as the initial condition.

Reactions may be performed under aerobic or anaerobic conditions where anaerobic or microaerobic conditions are preferred.

Batch and Continuous Fermentations: The present process employs a batch method of fermentation. A classical batch fennentation is a closed system where the composition of the media is set at the beginning of the fermentation and not subject to artificial alterations during the fermentation. Thus, at the beginning of the fermentation the media is inoculated with the desired organism or organisms and fermentation is permitted to occur adding nothing to the system. Typically, however, a "batch" fermentation is batch with respect to the addition of carbon source and attempts are often made at controlling factors such as pH and oxygen concentration. In batch systems the metabolite and biomass compositions of the system change constantly up to the time the fermentation is stopped. Within batch cultures cells moderate through a static lag phase to a high growth log phase and finally to a stationary phase where growth rate is diminished or halted. If untreated, cells in the stationary phase will eventually die. Cells in log phase generally are responsible for the bulk of production of end product or intermediate.

A variation on the standard batch system is the Fed-Batch system.

Fed-Batch fermentation processes are also suitable in the present invention and comprise a typical batch system with the exception that the substrate is added in increments as the fermentation progresses. Fed-Batch systems are useful when catabolite repression is apt to inhibit the metabolism of the cells and where it is desirable to have limited amounts of substrate in the media. Measurement of the actual substrate concentration in Fed-Batch systems is difficult and is therefore estimated on the basis of the changes of measurable factors such as pH, dissolved oxygen and the partial pressure of waste gases such as CO2. Batch and Fed-Batch fermentations are common and well known in the art and examples may be found in Brock, supra.

Although the present invention is performed in batch mode it is contemplated that the method would be adaptable to continuous fermentation methods. Continuous fermentation is an open system where a defined fermentation media is added continuously to a bioreactor and an equal amount of conditioned media is removed simultaneously for processing. Continuous fermentation generally maintains the cultures at a constant high density where cells are primarily in log phase growth.

Continuous fermentation allows for the modulation of one factor or any number of factors that affect cell growth or end product concentration. For example, one method will maintain a limiting nutrient such as the carbon source or nitrogen level at a fixed rate and allow all other parameters to moderate. In other systems a number of factors affecting growth can be altered continuously while the cell concentration, measured by media turbidity, is kept constant.

Continuous systems strive to maintain steady state growth conditions and thus the cell loss due to media being drawn off must be balanced against the cell growth rate in the fermentation. Methods of modulating nutrients and growth factors for continuous fermentation processes as well as techniques for maximizing the rate of product formation are well known in the art of industrial microbiology and a variety of methods are detailed by Brock, supra.

It is contemplated that the present invention may be practiced using either batch, fed-batch or continuous processes and that any known mode of fermentation would be suitable. Additionally, it is contemplated that cells may be immobilized on a substrate as whole cell catalysts and subjected to fermentation conditions for 1,3-propanediol production.

Identification and Purification of 1.3-propanediol: , Methods for the purification of 1,3-propanediol from fermentation media are known in the art. For example propanediols can be obtained from cell media by subjecting the reaction mixture to extraction with an organic solvent, distillation and column chromatography (U.S. 5,356,812). A particularly good organic solvent for this process is cyclohexane (U.S. 5,008,473). 1,3-Propanediol may be identified directly by submitting the media to high pressure liquid chromatography (HPLC) analysis. Preferred in the present invention is a method where fermentation media is analyzed on an analytical ion exchange column using a mobile phase of 0.01N sulfuric acid in an isocratic fashion.

Cells: Cells suitable in the present invention comprise those that harbor a dehydratase enzyme. Typically the enzyme will be either a glycerol dehydratase or a diol dehydratase having a substrate specificity for either glycerol or 1.2- propanediol, respectively. Dehydratase enzymes are capable of converting glycerol to hydroxypropionaldehyde (3-HPA) which is then converted to 1.3- propanediol. Cells containing this pathway may include mutated or recombinant organisms belonging to the genera Citrobacter, Enterobacter, Clostridium, Klebsiella, Samonella, and Lactobacillus. Microorganisms known by persons skilled in the art to produce glycerol by fermentation, e.g., Aspergillus, Saccharomyces, Zygosaccharomyces, Pichia, Kluyveromyces, Candida, Hansenula, Dunaliella, Debarvomyces, Mucor, Torylopsis, and Methylobacteria, may be the hosts for a recombinant dehydratase enzyme. Other cells suitable as hosts in the present invention include Bacillus, Escherichia, Pseudomonas and Streptomyces . While not wishing to be bound by theory, it is believed that organisms, belonging to the above mentioned groups, exist in nature that are suitable for the present invention.

On the basis of applicants' experimental work it is contemplated that a wide variety of cells may be used in the present invention. Applicants have demonstrated for example that cells varying widely in genetic and phenotypic composition are able to bioconvert a suitable carbon substrate to 1,3-propanediol. Cells exemplified include: a K. pneumoniae mutant strain constitutive for the dha genes, recombinant E. coli strains comprising elements of the Klebsiella genome containing genes encoding either glycerol or diol dehydratase, and recombinant E. coli (tpi~) strains also transfected with elements of the Klebsiella genomes and harboring a mutation in the gene encoding the triosephosphate isomerase enzyme.

Although £. coli transformants containing the dha regulon from Klebsiella pneumonia were able to convert glycerol to 1 ,3-propanediol even in the presence pf glucose or xylose (Tong et al., Appl. Biochem. Biotech., 34, 149 (1992)) no 1,3-propanediol was detected by these organisms in the presence of glucose alone. In direct contrast to this disclosure, applicants have discovered that three strains of E. coli, containing either of two independently isolated cosmids comprising the dha regulon from Klebsiella pneumonia, produced 1,3-propanediol from a feed of glucose with no exogenously added glycerol present. E. coli strain ECL707, containing cosmid vectors pKP-1 or pKP-2 comprising the K. pneumoniae dha regulon, showed detectable though modest production of 1,3-propanediol from glucose in the absense of exogenously added glycerol, (Example 3). Recombinant E. coli strains constructed from an alternate host organism, DH5a, also containing cosmid vectors pKP-1 or pKP-2, were found to be more effective than the ECL707 recombinants in producing 1,3-propanediol from glucose under the appropriate conditions, (Example 2). Most effective in producing 1,3-propanediol from glucose under the conditions of Example 3 were the recombinant E. coli strains AA200 containing cosmid vectors pKP-1 orpKP-2, Example 1. E. coli strain AA200 contains a defective triosephosphate isomerase enzyme (tpi~).

A strain of AA200-pKPl, selected for further study from a pool of independent isolates from the transformation reaction, converted glucose to 1,3-propanediol in a two stage reaction. In the first stage, the strain AA200-pKPl-5 was grown to high cell density in the absence of glucose and glycerol. In the second stage, the grown cells, suspended in a medium containing glucose but no glycerol, converted glucose to 1,3-propanediol with high conversion and selectivity, Example 4. Although differing immumochemically, chromatographically, and genetically, the coenzyme B^-dependent enzymes glycerol dehydratase (E.C. 4.2.1.30) and diol dehydratase (E.C. 4.2.1.28) catalyze the conversion of glycerol to 3-hydroxypropionaldehyde. Glycerol dehydratase, but not diol dehydratase, is encompassed by the dha regulon. K. pneumoniae ATCC 8724, containing a diol dehydratase but not a glycerol dehydratase converts glycerol to 1,3-propanediol (Forage et al., J Bacterioi, 149, 413, (1982)). Recombinant E. coli strains ECL707 and AA200, containing cosmid vector pKP4 encoding genes for a diol dehydratase, converted glucose to 1,3-propanediol, Example 1 and Example 3.

K. pneumoniae ECL2106, prepared by mutagenesis from a naturally occurring strain (Ruch et al., /. Bacteriol. 124, 348 (1975)), exibits constitutive expression of the dha regulon (Ruch et al., supra; Johnson et al., J. Bacteriol. 164, 479 (1985)). A strain derived from K. pneumoniae ATCC 25955, displaying the same phenotype, has been similarly prepared (Forage et al., /. Bacteriol. 149, 413 (1982)). Expression of the Klebsiella dha structural genes is, in part, controlled by a repressor (product of dha R) (Sprenger et al., /. Gen Microbiol. 135, 1255 (1989)). Applicants have shown that ECL2106, which is constitutive for the dha structural genes, produced 1,3-propanediol from a feed of glucose in the absence of exogenously added glycerol, Example 5. This is in contrast to wild type K. pneumoniae ATCC 25955 which did not produce detectable levels of 1,3-propanediol under the same conditions, Example 5.

The expression of the dha structural genes in ECL2106 is further controlled by catabolite expression (Sprenger et al., /. Gen Microbiol. 135, 1255 (1989)). Elimination of catabolite repression can be achieved by placing the necessary structural genes under the control of alternate promotors as has been demonstrated for 1,3-propanediol oxidoreductase (dhaT) from C.freundii and diol dehydratase from K. oxytoca ATCC 8724 (Daniel et al., /. Bacteriol. 177, 2151 (1995) and Tobimatsu et al., /. Biol. Chem. 270, 7142 (1995)). By eliminating catabolite repression from ECL2106 in this manner, an improvement in the production of 1,3-propanediol from glucose in the absence of an exogenous source of glycerol is achieved. An even further improvement is obtained by appropriate carbon channelling as is described, by example, with the tpi~ mutation.

As the dha regulons of Citrobacter and Klebsiella sp. are strikingly similar, one of skill in the art will appreciate that teachings that involve the production of 1,3-propanediol from glucose in the absence of an exogenous source of glycerol for Klebsiella sp. applies to Citrobacter sp. as well.

Furthermore, as the metabolism of glycerol by C. butyricum is comparable to that of K. pneumoniae [Zeng et al., Biotechnol. and Bioeng. 44, 902 (1994)], teachings will extend to Clostridia sp. as well.

EXAMPLES GENERAL METHODS Procedures for phosphorylations, ligations and transformations are well known in the art. Techniques suitable for use in the following examples may be found in Sambrook, J. et al., Molecular Cloning: A Laboratory Manual. Second Edition, Cold Spring Harbor Laboratory Press (1989).

Materials and methods suitable for the maintenance and growth of bacterial cultures are well known in the art. Techniques suitable for use in the following examples may be found in Manual of Methods for General Bacteriology (Phillipp Gerhardt, R. G. E. Murray, Ralph N. Costilow, Eugene W. Nester, Willis A. Wood, Noel R. Krieg and G. Briggs Phillips, eds), American Society for Microbiology, Washington, DC. (1994) or Thomas D. Brock in Biotechnology: A Textbook of Industrial Microbiology. Second Edition (1989) Sinauer Associates, Inc., Sunderland, MA. All reagents and materials used for the growth and maintenance of bacterial cells were obtained from Aldrich Chemicals (Milwaukee, WI), DIFCO Laboratories (Detroit, MI), GIBCO/BRL (Gaithersburg, MD), or Sigma Chemical Company (St. Louis, MO) unless otherwise specified.

The meaning of abbreviations is as follows: "h" means hour(s), "min" means minute(s), "sec" means second(s), "d" means day(s), "mL" means milliliters, "L" means liters, 50amp is 50 μ¾ πιί ampicillin, and LB-50amp is Luria-Bertani broth containing 50 μ¾/ηιί ampicillin.

Within the tables the following abreviations are used. "Con." is conversion, "Sel." is selectivity based on carbon, and "nd" is not detected.

Enzyme Assays Glycerol dehydratase activity in cell free extracts was determined using 1,2-propanediol as substrate. The assay, based on the reaction of aldehydes with methylbenzo-2-thiazolone hydrazone, has been described (Forage and Foster, Biochim. Biophys. Acta, 569, 249 (1979)). The activity of 1,3-propanediol oxidoreductase, sometimes referred to as 1,3-propanediol dehydrogenase, was determined in solution or in slab gels using 1,3-propanediol and NAD+ as substrates as has also been described. Johnson and Lin, J. Bacteriol., 169, 2050 (1987).

Isolation and Identification 1.3-propanediol The conversion of glycerol to 1,3-propanediol was monitored by HPLC. Analyses were performed using standard techniques and materials available to one of skill in the art of chromatography. One suitable method utilized a Waters Maxima 820 HPLC system using UV (210 nm) and RI detection. Samples were injected onto a Shodex SH-1011 column (8 mm x 300 mm, purchased from Waters, Milford, MA) equipped with a Shodex SH-1011P precolumn (6 mm x 50 mm), temperature controlled at 50°C, using 0.01 N H2SO4 as mobile phase at a flow rate of 0.5 mL/min. When quantitative analysis was desired, samples were prepared with a known amount of trimethylacetic acid as external standard. Typically, the retention times of glycerol (RI detection), 1,3-propanediol (RI detection), and trimethylacetic acid (UV and RI detection) were 20.67 min, 26.08 min, and 35.03 min, respectively.

Production of 1,3-propanediol was confirmed by GC/MS. Analyses were performed using standard techniques and materials available to one of skill in the an of GC/MS. One suitable method utilized a Hewlett Packard 5890 Series Π gas chromatograph coupled to a Hewlett Packard 5971 Series mass selective detector The chromosomal DNA was partially digested with Sau3A as outlined by Sambrook et al., supra. DNA (2 ug) was digested with 2 units of Sau3A (Promega, Madison, WI) at room temperature in 200 μΐ of total volume. At 0, 5, 10 and 20 min, samples (50 μΐ) were removed and transferred to tubes containing 5 umol of EDTA. These tubes were incubated at 70°C for 10 min. An aliquot (2 ΐ) was withdrawn and analyzed on a 0.5% agarose gel electrophoresis to determine the level of digestion and the rest of the sample (48 μ ) was stored at -20°C. The gel was stained with ethidium bromide and visualized under UV to determine the partial digestion of the chromosomal DNA. A decrease in the size of the chromosomal DNA with increase in time was observed showing that the decrease in the size of the chromosomal DNA is due to the action of Sau3A. DNA was extracted from rest of the sample by standard protocol methods (Sambrook et al., supra).

A cosmid library of partially digested DNA from K. pneumoniae was prepared using Supercos cosmid vector kit and Gigapackll packaging extracts using reagents purchased from Stratagene (La Jolla, CA). The instructions provided by the manufacturer were followed. The packaged K. pneumoniae contained 4 x 10^ to 1.0 x 10^ phage titer as determined by transfecting E. coli XL 1 -Blue MR.

Cosmid DNA was isolated from 6 of the E. coli transformants and found to contain large insert of DNA (25 to 30 kb).

EXAMPLE 1 Cloning and Transformation of E. coli Host Cells with cosmid DNA for the Expression of 1.3-propanediol Media Synthetic S12 medium was used in the screening of bacterial transformants for the ability to make 1,3-propanediol. S 2 medium contains: 10 mM ammonium sulfate, 50 mM potassium phosphate buffer, pH 7.0, 2 mM MgC-2, 0.7 mM CaCl2, 50 μΜ MnCi2, 1 μΜ FeCl3, 1 μΜ ZnCl2, 1.7 μΜ CuS04, 2.5 μΜ CoCl2, 2.4 μΜ Na2Mo04, and 2 μΜ thiamine hydrochloride.

Medium A used for growth and fermentation consisted of: 10 mM ammonium sulfate; 50 mM MOPS/KOH buffer, pH 7.5; 5 mM potassium phosphate buffer, pH 7.5; 2 mM MgCl2; 0.7 mM CaCl2; 50 uM MnCl2; 1 μΜ FeCl3; 1 μΜ ZnCl2; 1.72 μΜ CuS04; 2.53 μΜ CoCl2; 2.42 μΜ Na2Mo04; 2 μΜ thiamine hydrochloride; 0.01% yeast extract; 0.01% casamino acids; 0.8 g/mL vitamin B12; and 50 amp. Medium A was supplemented with either 0.2% glycerol or 0.2% glycerol plus 0.2% D-glucose as required.

Cells: Klebsiella pneumoniae ECL2106 (Ruch et al., J. Bacteriol., 124, 348 (1975)), also known in the literature as K aerogenes or Aerobacter aerogenes, was obtained from E. C. C. Lin (Harvard Medical School, Cambridge, MA) and was maintained as a laboratory culture.

Klebsiella pneumoniae ATCC 25955 was purchased from American Type Culture Collection (Rockville, MD).

E. coli DH5a was purchased from Gibco/BRL and was transformed with the cosmid DNA isolated from Klebsiella pneumoniae ATCC 25955 containing a gene coding for either a glycerol or diol dehydratase enzyme. Cosmids 19 containing t e g ycero e y ratase were ent ed as p an p P and cosmid containing the diol dehydratase enzyme were identified as pKP4.

Transformed DH5 E. coli ECL707 (Sprenger et al., J. Gen. Microbiol., 135, 1255 (1989)) was obtained from E. C. C. Lin (Harvard Medical School, Cambridge, MA) and was similarly transformed with cosmid DNA from Klebsiella pneumoniae. These transformants were identified as ECL707-pKPl and ECL707-pKP2, containing the glycerol dehydratase gene and ECL707-pKP4 containing the diol dehydratase gene.

E. coli AA200 containing a mutation in the tpi gene (Anderson et al., J. Gen Microbiol., 62, 329 (1970)) was purchased from the E. coli Genetic Stock Center, Yale University (New Haven, CT) and was transformed with Klebsiella cosmid DNA to give the recombinant organisms AA200-pKPl and AA200-pKP2, containing the glycerol dehydratase gene, and AA200-pKP4, containing the diol dehydratase gene.

DH5a: Six transformation plates containing approximately 1,000 colonies of E. coli XLl-Blue MR transfected with K. pneumoniae DNA were washed with 5 mJL LB medium and centrifuged. The bacteria were pelleted and resuspended in 5 mL LB medium + glycerol. An aliquot (50 ί) was inoculated into a 15 mL tube containing S12 synthetic medium with 0.2% glycerol + 400 ng per ml of vitamin B i2 + 0.001% yeast extract + 50amp. The tube was filled with the medium to the top and wrapped with parafilm and incubated at 30°C. A slight turbidity was observed after 48 h. Aliquots, analyzed for product distribution as described above at 78 h and 132 h, were positive for 1,3-propanediol, the later time points containing increased amounts of 1,3-propanediol.

The bacteria, testing positive for 1,3-propanediol production, were serially diluted and plated onto LB-50amp plates in order to isolate single colonies. Forty eight single colonies were isolated and checked again for the production of 1,3-propanediol. Cosmid DNA was isolated from 6 independent clones and transformed into E. coli strain DH5a. The transformants were again checked for the production of 1,3-propanediol. Two transformants were characterized further and designated as DH5a-pKPl and DH5a-pKP2.

A 12.1 kb EcoRI-Sall fragment from pKPl, subcloned into pIBI31 (IBI Biosystem, New Haven, CN), was sequenced and termed pHK28-26 (SEQ ID NO: 1 ). Sequencing revealed the loci of the relevant open reading frames of the dha operon encoding glycerol dehydratase and genes necessary for regulation. 130789/2 Referring to SEQ ID NO:l, a fragment of the open reading frame for dhaK encoding dihydroxyacetone kinase is found at bases 1-399; the open reading frame dhaD encoding glycerol dehydrogenase is found at bases 983-2107; the open reading frame dhaR encoding the repressor is found at bases 2209-4134; the open reading frame dhaT encoding 1,3-propanediol oxidoreductase is found at bases 5017-6180; the open reading frame dhaBl encoding the alpha subunit glycerol dehydratase is found at bases 7044-8711; the open reading frame dhaBl encoding the beta subunit glycerol dehydratase is found at bases 8724-9308; the open reading frame dhaB3 encoding the gamma subunit glycerol dehydratase is found at bases 9311-9736; and the open reading frame dhaBX, encoding a protein of unknown function is found at bases 9749-11572.

Single colonies of E. coli XL 1 -Blue MR transfected with packaged cosmid DNA from K. pneumoniae were inoculated into microtiter wells containing 200 uL of S15 medium (ammonium sulfate, 10 mM; potassium phosphate buffer, pH 7.0, 1 mM; MOPS/KOH buffer, pH 7.0, 50 mM; MgCl2, 2 mM; CaCl2, 0.7 mM; MnCl2, 50 μΜ; FeCl3, 1 μΜ; ZnCl2, 1 μΜ; CuSC>4, 1.72 μΜ; CoCl2, 2.53 Μ; Na2MoC>4, 2.42 μΜ; and thiamine hydrochloride, 2 μΜ) + 0.2% glycerol + 400 ng/mL of vitamin Bj2 +■ 0.001% yeast extract + 50 ampicillin. In addition to the microtiter wells, a master plate containing LB-50 amp was also inoculated. After 96 h, 100 μΐ, was withdrawn and centrifuged in a Rainin microfuge tube containing a 0.2 micron nylon membrane filter. Bacteria were retained and the filtrate was processed for HPLC analysis. Positive clones demonstrating 1,3-propanediol production were identified after screening approximately 240 colonies. Three positive clones were identified, two of which had grown on LB-50 amp and one of which had not. A single colony, isolated from one of the two positive clones grown on LB-50 amp and verified for the production of 1,3-propanediol, was designated as pKP4. Cosmid DNA was isolated from E. coli strains containing pKP4 and E. coli strain DH5a was transformed. An independent transformant, designated as DH5a-pKP4, was verified for the production of 1,3-propanediol.

ECL707: E. coli strain ECL707 was transformed with cosmid K. pneumoniae DNA corresponding to pKPl, pKP2, pKP4 and the Supercos vector alone and was named ECL707-pKPl, ECL707-pKP2, ECL707-pKP4, and ECL707-sc, respectively. ECL707 is defective in glpK, gld, and ptsD which encode the ATP-dependent glycerol kinase, NAD+-linked glycerol dehydrogenase, and enzyme II for dihydroxyacetone of the phosphoenolpyruvate dependent phosphotransferase system, respectively.

Twenty single colonies of each cosmid transformation and five of the Supercos vector alone (negative control) transformation, isolated from LB-50amp plates, were transferred to a master LB-50amp plate. These isolates were also tested for their ability to convert glycerol to 1 ,3-propanediol in order to determine if they contained dehydratase activity. The transformants were transferred with a sterile toothpick to microtiter plates containing 200 pL of Medium A supplemented with either 0.2% glycerol or 0.2% glycerol plus 0.2% D-glucose. After incubation for 48 hr at 30°C, the contents of the microtiter plate wells were filtered through an 0.45 μ nylon filter and chromatographed by HPLC. The results of these tests are given in Table 1.

Table 1 Conversion of glycerol to 1,3-propanediol by transformed ECL707: number of positive isolates/number of isolates tested Transformant Glycerol Glycerol plus Glucose ECL707-pKPl 19/20 19/20 ECL707-pKP2 18/20 20/20 ECL707-pKP4 0/20 20/20 ECL707-sc 0/5 0/5 AA200: E. coli strain AA200 was transformed with cosmid K. pneumoniae DNA corresponding to pKPl, pKP2, pKP4 and the Supercos vector alone and was named AA200-pKPl, AA200-pKP2, AA200-pKP4, and AA200-sc, respectively. Strain AA200 is defective in triosephosphate isomerase, (fpi").

Twenty single colonies of each cosmid transformation and five of the empty vector transformation were isolated and tested for their ability to convert glycerol to 1,3-propanediol as described for E. coli strain ECL707. The results of these tests are given in Table 2.

Table 2 Conversion of glycerol to 1,3-propanediol by transformed AA200: Number of positive isolates/number of isolates tested Transformant Glycerol Glycerol plus Glucose AA200-pKPl 17/20 17/20 AA200-pKP2 17/20 17/20 AA200-pKP4 2/20 16/20 EXAMPLE 2 Conversion of D-glucose to 1.3-propanediol bv E. coli strain AA200. transformed with Klebsiellia pneumoniae DNA containing dehydratase activity Glass serum bottles, filled to capacity with media (ca. 14 mL of Medium A supplemented with 10 μg/mL kanamycin and 0.2% D-glucose, plus or minus ,0.5-1.0 mM cyclic adenosine 2':3'-monophosphate (cAMP)), were innoculated with selected single colony isolates of E. coli strain AA200 containing the K. pneumoniae dha regulon cosmids pKPl or pKP2, the K. pneumoniae pdu operon pKP4, or the Supercos vector alone. In order to avoid contact with glycerol, the innoculation was performed from either an agar plate of LB-50amp or from a liquid culture of the same medium. The reactions were incubated for ca. 72 hr at 30°C while shaking at 250 rpm. Growth was determined by the change in absorbance at 600 nm where the initial OD^ was 0.020 AU. The extent of glucose depletion and product distribution were determined by HPLC. Single colony isolates are identified by a numbered suffix "-x", e.g., AA200-pKPl-x. Cumulative results are presented in Table 3 and Table 4.

Table 3 Conversion of 0.2% D-glucose to 1,3 -propanediol bv transformed E. coli strain AA200: without cAMP [1,3 -propane- Transformant OD^n dioll (mM) Con. (%) Sel. (%) AA200-pKPl-3 0.056 0.05 17 1 AA200-pKPl-5 0.115 nd 0 0.007 nd 0 0.076 0.2 14 5 ΑΑ2ΘΟ-ρΚΡ1-20 0.116 nd 27 0 0.205 0.3 17 8 AA200-pKP2-10 0.098 0.2 13 7 AA200-pKP2-14 0.117 0.5 17 14 0.129 0.2 19 5 AA200-pKP2-20 0.094 nd 11 0 AA200-pKP4-4 0.198 0.1 28 2 AA200-pKP4-19 0.197 0.2 34 3 AA200-pKP4-20 0.206 0.1 38 1 AA200-sc-l 0.097 nd 22 0 0.176 nd 46 0 Table 4 Conversion of 0.2% D-glucose to 1 ,3-propanediol by transformed E. coli strain AA200: with cAMP [1,3-propane- Transformant OD^nn dioH (raM) % Con. % Sel.

AA200-pKPl-3 0.102 0.5 19 12 AA200-pKPl-5 0.088 1.5 21 37 0.236 1.4 24 28 0.071 0.8 15 23 AA200-pKPl-20 0.153 nd 40 0 0.185 0.9 27 16 AA200-pKP2-10 0.098 0.2 13 7 AA200-pKP2-14 0.213 2.0 26 27 0.155 0.6 25 12 AA200-pKP2-20 0.198 1.2 40 14 AA200-pKP4-4 0.218 0.1 31 2 AA200-pKP4-19 0.223 0.2 37 3 AA200- KP4-20 0.221 0.2 35 3 AA200-SC-1 0.111 nd 23 0 0.199 nd 49 0 0.122 nd 25 0 aThe identity of 1 ,3-propanediol was verified by GC/MS as described GENERAL METHODS.

EXAMPLE 3 Conversion of D-glucose to 1.3-propanediol by E. coli strain DH5cc. transformed with Klebsiellia pneumoniae DNA containing dehydratase activity E. coli strain DH5cc, containing the K. pneumoniae dha regulon cosmids pKPl or pKP2, were tested for their ability to convert D-glucose to 1,3-propanediol exactly as described in Example 1. The results are presented in Table 5.

Table 5 Conversion of 0.2% D-glucose to 1,3-propanediol by transformed E. coli strain DH5a: plus (+) and minus (-) cAMP [1,3-propane- Trans formant OD^nn dioll (mM) % Con. % Sel.

DH5a-pKPl (-) 0.630 0.5 100 2 DH5a-pKPl (+) 0.774 0.6 100 3 DH5cc-pKP2 (-) 0.584 0.6 100 3 DH5a-pKP2 (+) 0.699 0.7 100 3 EXAMPLE 4 Conversion of D-glucose to 1.3-propanediol by E. coli strain ECL707. transformed with Klebsiellia pneumoniae DNA containing dehydratase activity E. coli strain ECL707, containing the K. pneumoniae dha regulon cosmids pKPl or pKP2, the K. pneumoniae pdu operon pKP4, or the Supercos vector alone, were tested for their ability to convert D-glucose to 1,3-propanediol exactly as described in Example 1. In each case, conversion was quantitative. The results are presented in Table 6.

Table 6 Conversion of D-glucose to 1,3-propanediol by transformed E. coli strain ECL707: with and without cAMP [1,3-propane- [1,3-propane Transformant ODsnn dioll (mM) OD^ dioll (mM) (without cAMP) (with cAMP) ECL7(57-pKPl-l 0.607 0.1 0.475 0.1 ECL707-pKPl-3 0.619 0.1 0.422 0.1 ECL707-pKPl-7 0.582 0.2 0.522 0.2 ECL707-pKPl-10 0.593 0.1 0.408 0.1 ECL707-pKPl-18 0.584 0.1 0.433 0.1 ECL707-pKP2-4 0.588 0.1 0.408 0.1 ECL707-pKP2-5 0.630 0 1 0.516 0.2 ECL707-pKP2-8 0.542 0.1 0.486 0.1 ECL707-pKP2-15 0.589 0.1 0.485 0.1 ECL707-pKP2-19 0.577 0.1 0.504 0.1 ECL707-pKP4-8 0.499 nd 0.361 <0.1 ECL707-pKP4-9 0.544 nd 0.354 nd ECL707-pKP4-10 0.515 nd 0.265 <0.i ECL707-pKP4-14 0.512 nd 0.318 <0.1 ECL707-pKP4-17 0.545 nd 0.388 <0.1 ECL707-SC-1 0.592 nd 0.385 nd EXAMPLE 5 Two stage conversion of D-glucose to 1.3 -propanediol by Escherichia coli AA200-pKPl-5 Baffled flasks (250 mL) containing 50 mL LB -amp medium were inoculated with single colonies of AA200-pKPl-5. The cells were grown, in duplicate, overnight at 30 or 37°C with shaking (250 rpm).

Grown cultures were spun (10 minutes, 10,000 rpm, 4°C) and resuspended in production medium without glucose (10 mM (NH4)2S04; 5 mM potassium phosphate buffer, pH 7.5; 50 mM MOPS, pH 7.5; 0.01% yeast extract; 0.01% casamino acids; 0.8 pg/mL vitamin B 12; and 50 pg/mL ampicillin) containing either trace metals A: (0.08 μΜ CoCl2, 0.06 μΜ CuCl2, 7 pM FeS04, 2 pM H3BO4, 0.2 μΜ MnCl2, 0.1 pM Na2Mo0 , 0.08 pM NiCl2, 0.3 μΜ ZnS04, and 0.03 mM thiamine) or trace metals B: (0.7 mM CaCl2, 2.53 μΜ CoCl2, 1.72 μΜ CuS04, 1.0 μΜ FeCl3, 2 mM MgCl2, 0.05 mM MnCl2, 2.42 μΜ Na2Mo04, 1.0 μΜ ZnCl2, and 0.03 mM thiamine). The cells were spun a second time, resuspended in 50 mL fresh production medium containing D-glucose and dispensed into 60 mL serum bottles which were capped and sealed with butyl rubber septa. The bottles were shaken (250 rpm) and samples withdrawn with a syringe through the septum and filtered through a 0.2 p filter before analysis. Results are shown in Table 7 and Table 8; residual glucose was measured by enzymatic analysis (Biochemistry Analyzer, Yellow Springs Instruments Co., Inc.) and 1,3-propanediol was analyzed by HPLC.

Table 7 Conversion of 0.2% D-glucose to 1,3-propanediol by Escherichia coli AA200-pKPl-5.

Duplicate reactions were performeda Time [Glucose] [1,3 -propaneCon. Sel.

Experiment (days) (mM) diol! (mM) (%) W #1 1 0.1 2.3 99 10 #1 4 0.1 2.3 99 10 #2 1 2.8 2.3 75 14 #2 4 0.1 2.4 99 11 aThe reactions mixtures, containing trace metals A, were incubated at 37°C.

Table 8 Conversion of 1% D-glucose to 1.3 -propanediol by Escherichia coli AA200-kPl-5a time [glucose] [1,3-propane- Con. Sel. (days) (mM) dioll (mM) (%) (%) 0 53 0 0 0 1 39 5.6 26 20 2 35 8.3 34 23 3 33 8.4 38 21b aThe reactions mixtures, containing trace metals B, were incubated at 30°C ^At the end of the reaction, the presence of 1,3-propane-diol was confirmed by GC MS and ^C-NMR (300 MHz).

EXAMPLE 6 Conversion of D-glucose to 1.3-propanediol bv Klebsiella pneumoniae ECL2106 but not by Klebsiella pneumoniae ATCC 25955 Glass serum bottles, filled to capacity (ca 14 mL) with media, were lightly innoculated from a LB agar plate containing K. pneumoniae ECL2106 or K. pneumoniae ATCC 25955. The media contained 50 mM glucose, 3 mM (NH4)2S04, 0.9 mM CaCl2, 4 uM CoCl2, 0.06 μΜ CuCl2, 7 μΜ FeS04, 2 μΜ H3B04, 0.8 mM MgS04, 0.2 uM MnCl2, 0.1 μΜ Na2Mo04, 0.08 uM NiCl2, 0.3 μΜ ZnS04, 0.1 mg/mL DL-cysteine, 10 μΜ emylenediaminetetraacetic acid, 0.8 g/ml vitamin Bi2, potassium phosphate as indicated in Table 9, and either 50 mM HEPES or 50 mM MOPS buffer, pH 7.5. The reactions were incubated for 47 hr at 30°C while shaking at 250 rpm. Otherwise, the reaction was performed as described in Example 1. The results are given in Table 9.

Table 9 Conversion of D-glucose to 1, 3-propanediol by Klebsiella pneumoniae ECL2106 but not by Klebsiella pneumoniae ATCC 25955 Strain Buffer Pi [Glucose] [1,3-Propane- (mM) (mM) dioll (mM) 2106 MOPS 5.0 11.4 0.2 2106 MOPS 2.5 13.9 0.2 2106 MOPS 1.3 14.8 0.1 2106 MOPS 0.6 15.8 0.1 2106 HEPES 5.0 21.1 0.1 2106 HEPES 2.5 23.4 0.1 2106 HEPES 1.3 26.4 0.1 2106 HEPES 0.6 27.5 0.1 25955 MOPS 5.0 4.4 nd 25955 MOPS 2.5 5.4 nd 25955 MOPS 1.3 2.8 nd 25955 MOPS 0.6 7.8 nd 25955 HEPES 5.0 7.0 nd 25955 HEPES 2.5 13.5 nd 25955 HEPES 1.3 10.2 nd 25955 HEPES 0.6 17.8 nd EXAMPLE 7 Conversion of D-glucose to 1.3-propanediol bv recombinant Pichia pastoris Construction of general purpose expression plasmid The 0.9 kb EcoRl/Xbal fragment in pHIL-D4 (Phillips Petroleum, Bartlesville, OK) was replaced by the 0.9 kb EcoRl Xbal fragment from pA0815 (Invitrogen, San Diego, CA) to generate the plasmid pHIL-D4B2 which contains the following elements: 5'AOXl, P. pastoris methanol inducible alcohol oxidase I (AOX1) promoter; AOX1 term, P. pastoris AOX I transcriptional termination region; HIS4, P. pastoris histidinol dehydrogenase-encoding gene for selection in his4 hosts; kan, sequence derived from transposon ΎΏ903 encoding aminoglycoside 3'-phosphotransferase, conferring kanamycin, neomycin and G418 resistance in a wide variety of hosts, and useful as an indicator of cassette copy number; 3'AOXJ , P. pastoris sequence downstream from AOX1 , used in conjunction with 5'AOXl for site-directed vector integration; ori, pBR322 origin of DNA replication allowing plasmid manipulations in E. coli; and amp, β-lactamase gene from pBR322 conferring resistance to ampicillin. An additional feature of pHIL-D4B2 is that multiple expression cassettes {5'AOXl - gene -AOXlterm) can easily be placed onto one plasmid by subcloning cassettes on Bgl2/Xbal fragments into BamHl/Xbal sites.

Construction of plasmid for co-expression of dhaBl and dhaB2 The open reading frames for dhaBl and dhaB2 were amplified from pHK28-26 by PCR using primers (SEQ ID NO:2 with SEQ ID NO:3 and SEQ ID NO:4 with SEQ ID NO:5 for dhaBl and dhaBl, respectively) incorporating EcoRl sites at the 5' ends ( 10 mM Tris pH 8.3, 50 mM KC1, 1.5 mM MgCl2, 0.0001% gelatin, 200 μΜ dATP, 200 μΜ dCTP, 200 μΜ dGTP, 200 μΜ dTTP, 1 μΜ each primer, 1- 10 ng target DNA, 25 units/mL Amplitaq® DNA polymerase Perkin Elmer Cetus, Norwalk CT). PCR parameters were 1 min at 94°C, 1 m n at 55°C, 1 min at 72°C, 35 cycles. The products were subcloned into the EcoRl site of pHIL-D4B2 to generate the expression plasmids pMP19 and pMP20 containing dfiaBl and dhaB2, respectively.

The dhaBl expression cassette on a Bgl2/Xbal fragment from pMP19 was subcloned into the BamHl/Xbal site of pMP20 to generate pMP21. Plasmid pMP21 contains expression cassettes for both dhaB2 and dhaBl, and the HIS4 selectable marker.

'Construction of plasmid for co-expression of dhaB3 and dhaT The open reading frames for dhaT and dhaBS were amplified by PCR from pHK28-26 (SEQ ID NO: l) using primers (SEQ ID NO:6 with SEQ ID NO:7 and SEQ ID NO:8 with SEQ ID NO:9 for dhaT and dhaBS, respectively) incorporating EcoRl sites at the 5' ends. The products were subcloned into the EcoRl site of pHIL-D4B2 to generate the expression plasmids pMP17 and pMP18 containing dhaT and dhaB3, respectively.

The dhaT expression cassette on a Bgl2/Xbal fragment from pMP17 was subcloned into the BamHl/Xbal site of p P18 to generate pMP22 which contains expression cassettes for both dhaT and dhoB3.

The 4.1 kb EcoRl fragment containing SUC2 was deleted from pRK20 (Phillips Petroleum, Bartlesville, OK) to generate pMP2. SUC2 encodes for an invertase which may be used a second selectable marker in Pichia. The 4.0 kb Hind3 fragment containing lacZwas deleted form pMP2 to generate pMP3. The 0.4 kb Hind3 fragment containing AOXl term from pHIL-D4 was subcloned into the Hind3 site of pMP3 to generate pMPlO.

The 2.0 kb Bgl2/Xbal fragment in pMPIO was replaced with the 5.0 kb Bgl2/Xbal fragment containing the dhaB3 and dhaT expression cassettes from pMP22 to generate pMP23. The 5.4 kb Pstl/Bgl2 fragment containing SUC2 from pRK20 was subcloned into the Pstl/Bgl2 sites of pSP73 (Promega, Madison, WI) to generate pMPl la. Plasmid pMPl la was cut with EcoRl, filled with T4 DNA polymerase and religated to generate pMPl lb. The 1.1 kb Pst Bgl2 fragment in pMPIO was replaced with the 5.4 kb Bgl2 Pstl fragment containing SUC2 from pMPl lb to generate pMP12.

The 1.0 kb Scal/Bgl2 fragment in pMP23 was replaced with the 5.2 kb Scal/Bgl2 fragment containing SUC2 from pMP12 to generate pMP24. Plasmid pMP24 contains expression cassettes for both dhaT and dhaB3, and the SUC2 selectable marker.

Transformation of P. pastoris with dhaBl ldhaB2 expression plasmid pMP21 P. pastoris strain GTS1 \5{his4) (Phillips Petroleum, Bartlesville, OK) was transformed with 1-2 ug of Bgl2-linearized plasmid pMP21 using the spheroplast transformation method described by Cregg et al., (Mol. Cell. Biol. 5, 3376, (1985)). Cells were regenerated on plates without histidine for 3-4 days at 30°C. All transformants arise after integration of plasmid DNA into the chromosome. Transformants were patched onto a YPD (1% Bacto yeast extract, 2% peptone, 2% glucose) master plate.

Screening of P. pastoris transformants for dhaBl and dhaB2 Chromosomal DNA was prepared from his+ transformants described above and subjected to PCR analysis with primers specific for dhaBl and dhaB2. High copy number strains were selected from transformants containing both dhaBl and dhaB2 by growth in YPD media supplemented with increasing levels of G418 (Sigma, St. Louis, MO) up to 2000 μg mL. Resistance to a high level of G418 suggests significant duplication of expression cassettes.

Secondary transformation of P. pastoris with dhaB3ldhaT expression plasmid pMP24 Transformants with resistance to a high level of G418 as described above were re-transformed with plasmid pMP24 using the spheroplast transformation method. Cells were first regenerated on non-selective plates for 2 days at 30°C, after which top agar containing the regenerated cells was scraped from the plate and vortexed extensively in 20 mL water. After passing through 4 folds of cheesecloth, the cells were pelleted by centrifugation and resuspended in 10 mL water. Aliquots of 200 uL were plated onto sucrose plates, and incubated for 2 days at 30°C. All transformants arise after integration of plasmid DNA into the chromosome. Transformants appear as large colonies in a background of small colonies, and require isolation. After 24 h growth with shaking at 30°C in Msu media (per L, 13.4 g yeast nitrogen base w/o amino acids, 10 g sucrose, 0.4 g biotin), transformants were streaked onto Msu plates (Msu media plus 15 g L agar) and grown for 2 days at 30°C. Large isolated colonies were patched onto a YPD master plate.

Screening of P. pastoris double transformants for dhaBl. dhaB2. dhaB3. and dhaT and their corresponding enzyme activities Chromosomal DNA was prepared from suc+ double transformants described above and subjected to PCR analysis with primers specific for dhaBl, dhaB2, dhaB3, and dhaT. Thus, the presence of all four open reading frames was confirmed.

The presence of active glycerol dehydratase (dhaB) and 1,3 -propanediol oxido-reductase (dhaT) was demonstrated using in vitro enzyme assays. Additionally, Western blot analysis confirmed protein expression from all four open reading frames. Cell free extracts for these protein characterizations were prepared as follows. Double transformants containing dhaBl, dhaB2, dhaB3, and 130789/2 dhaT were grown aerobically with shaking at 30°C in MGY (Per L, 13.4 g yeast nitrogen base w/o amino acids, 0.4 mg biotin, 10 mL glycerol) for 2 days. The cells were pelleted by centrifugation, resuspended in MM (Per L, 13.4 g yeast nitrogen base w/o amino acids, 0.4 mg biotin, 5 mL methanol) and incubated as above. After approximately 24 h, the cells were harvested, resuspended in buffer (0.1 M tricene/KOH buffer, pH 8.2, 50 mM KC1, and 2% 1,2-propanediol), mechanically disrupted (using a glass rod while vortexing in the presence of glass beads), and centrifuged.

One strain that showed positive for the presence of all four open reading frames (dhoBJ, dhaB2, dhaB3, and dhaT) and their corresponding activities was designated MSP42.81 and was selected for further study.

In vivo production of 1.3 -propanediol using recombinant Pichia pastoris P pastoris MSP42.81 were grown in a BiostatB fermenter (B Braun Biotech, Inc.) in 1.5 L minimal medium containing 8.5 g L KH2PO4, 2.1 g L 10 g L glycerol, 2.3 g L MgS04-7H20, 0.18 g/L CaS04-2H20, and 0.29 ml/L PTMI. PTMI is a stock mineral solution containing 24 mM Q1SO4, 4.8 mM KI, 18 mM MnS04) 0.8 mM Na2Mo04, 0.3 mM H3BO3, 2.1 mM C0CI2, 70 mM ZnS04, 26 mM H2SO4, 234 mM FeS04, and 0.8 mM biotin. The fermenter was controlled at pH 5.0 with addition of 2M N¾OH, 30°C, and 30% dissolved oxygen tension through agitation control. A culture of P pastoris MSP42.81 grown in YM broth at 30°C was used as an inoculum; 20 mL of the culture was used to inoculate the fermenter.

When glycerol was shown to be depleted (24 h after inoculation), induction of the AOX promoters was initiated by the addition of a methanol feed. The feed contained 1 liter of methanol, 5 mL PTMI and 5 mL of a stock biotin solution prepared as 0.2 g/L in water. The methanol solution was added manually to maintain an average concentration of 0.5% as determined by HPLC analysis. Fifty mL of cells OD600 = 20 AU) were removed from the reactor after 15 h of induction.

The 50 mL cell suspension was pelleted and resuspended in 12.5 mL nitrogen sparged base medium (6.7 g/L yeast nitrogen base, 3.0 g L yeast extract, 1.0 g L K2HP04, 1.0 g/L KH2P0 , 3.0 g L (NH,) 2S04, titrated to pH 7.2 and filter sterilized). Coenzyme B12, prepared as a stock solution at 2 mg mL in nitrogen sparged water, was added to the cell suspension to give a final concentration of 20 μg/mL. Three media stock solutions were prepared in base medium containing 1% glucose, 0.5% glucose and 0.5% glycerol (w/v), and 1.0% glycerol (w/v). A stock solution of chloroquine (1.06 g/50 mL, pH 7.2) was also prepared.

Two mL of medium 1, 2, or 3, and 1 mL mixtures of chloroquine and water to give the final concentrations listed in Table 10 were placed in 10 mL crimp sealed serum bottles and sparged with nitrogen before adding 1 mL of cells with coenzyme B 12 mixture. The serum bottles were incubated at 30°C with shaking. Samples taken immediately after the addition of cells and after 24 h incubation were analyzed by HPLC. The results are shown in Table 10.

Table 10 In vivo production of 1.3-propanediol using recombinant Pichia pastoris chloroquine 1 ,3-propanediol reaction medium3 (mM) (mM) 1 glub 0 0.04 2 glu 2.5 0.2 3 glu 5.0 0.1 4 glu 10.0 0.1 5 glub/gly 0 0.2 6 glu/gly 2.5 0.4 7 glu/gly 5.0 0.4 8 glu/gly 10.0 1.2C 9 gly 0 0.2 10 gly 2.5 0.3 11 gly 5.0 0.3 12 giy 10.0 1.4C aLess than 10% of each substrate was used in 24 h unless noted. bNo glucose remained after 24 h.

^The presence of 1 ,3-propanediol was confimed by GC MS as described in GENERAL METHODS.

EXAMPLE 8 Use of P. pastoris double transformant for production of 1.3-propanediol from glucose Double transformants containing dhaBl, dh B2, dhaB3, and dhaT are grown aerobically with shaking at 30°C in MD (Per L, 13.4 g yeast nitrogen base w/o amino acids, 0.4 mg biotin, 20 g glucose). After 2 d, the cells are pelleted by centrifugation, resuspended in MM (Per L, 13.4 g yeast nitrogen base w/o amino acids, 0.4 mg biotin, 5 mL methanol) and incubated as above. After approximately 24 h, glucose and vitamin are added to final concentrations of 10 g/L and 1 mg/L, respectively and incubated anaerobically at 30°C. After 24 h the presence of 1,3 -propanediol is determined by HPLC.

EXAMPLE 9 Plasmid construction for the transformation of Sacch romvces cerevisiae Construction of General Purpose Expression Plasmids Two types of expression plasmids were created, those that could integrate by recombination into chromosomes, and those that could exist as replicating episomal elements. For each type of general expression plasmid a yeast promoter vas present and separated from a yeast transcription terminator by fragments of DNA containing recognition sites for one or more restriction endonucleases. Each type of general expression plasmid also contained the gene for β-lactamase for selection in E. coli on media containing ampicillin, an origin of replication for plasmid maintainence in E. coli, and either a 2 micron origin of replication for episomal elements or sequences homologous to those found in S. cerevisiae chromosomes for recombination and integration of introduced DNA into chromosomes. The selectable nutritional markers used for yeast and present on the expression plasmids were one of the following: HIS 3 gene encoding imidazoleglycerolphosphate dehydratase, URA3 gene encoding orotidine 5'-phosphate decarboxylase, TRPl gene encoding N-(5'-phosphoribosyl)- anthranilate isomerase and LEU2 encoding β-isopropylmalate dehydrogenase. The yeast promoters used were ADHl or GALl, and the transcription terminators ADHl, CYC 1 or AOX1; the latter from Pichia pastoris.

Plasmid pGADGH (Clontech, Palo Alto, CA) was digested with Hindm and the single -strand ends converted to EcoRI ends by ligation with HindUI-Xmnl and EcoRI-XmnI adaptors (New England Biolabs, Beverly, MA).

Selection for plasmids with correct EcoRI ends was achieved by ligation to a kanamycin resistance gene on an EcoRI fragment from plasmid pUC4K (Pharmacia Biotech, Uppsala), transformation into E. coli strain DH5a and selection on LB plates containing 25 μg/mL kanamycin. The resulting plasmid (pGAD/KAN2) was digested with SnaBI and EcoRI and a 1.8 kb fragment with the ADHl promoter was isolated. This fragment was Iigated into the integrative vector pRS406 (Stratagene, La olla, CA) previously digested with HincII and EcoRI. Positive clones were identified by insertional-inactivation of the plasmid-encoded lacZ alpha peptide and the presence of the ADHl promoter fragment. The resulting plasmid (pMCK3) was digested with EcoRI and Smal and Iigated to the 0.2 kb ADHl terminator fragment released from plasmid pGBT9 (Clontech, Palo Alto, CA) by digestion with EcoRI and Nael. The resulting plasmid (pMCK5) contains both ADHl promoter and terminator sequences, URA3 and serves as an integration plasmid.

Plasmid pGADGH was digested with SnaBI and HindHI and a 1.8 kb fragment containing the ADHl promoter isolated. This fragment was ligated into the integrative vector pRS405 (Stratagene, La JoUa, CA) previously digested with Smal and HindHI. Positive clones were identified by insertional-inactivation of the plasmid-encoded lacZ alpha peptide and the presence of the ADHl promoter fragment. The resulting plasmid (pMCK4) contains an ADHl promoter, LEU2 and serves as an integration plasmid.

Plasmid pGBT9 was digested with SnaBI and EcoRI, and the 1.5 kb ADH1/GAL4 fragment replaced by a 1.8 kb ADHl promoter fragment isolated from pGAD/KAN2 by digestion with SnaBI and EcoRI. The resulting vector (pMCKl 1) is a replicating plasmid in yeast with ADHl promoter and terminator and a TRPl marker.

The open reading frames for dhaT, dhaB3, and dhaBl were amplified from pHK28-26 (SEQ ID NO: 1) by PCR using primers (SEQ ID NO:6 with SEQ ID NO:7, SEQ ID NO:8 with SEQ ID NO:9, and SEQ ID NO:2 with SEQ ID NO:3 for dhaT, dhaB3, and dhaBl, respectively) incorporating EcoRI sites at the 5' ends (10 mM Tris pH 8.3, 50 mM KC1, 1.5 mM MgCl2, 0.0001% gelatin, 200 μΜ dATP, 200 μΜ dCTP, 200 μΜ dGTP, 200 uM dTTP, 1 uM each primer, 1-10 ng target DNA, 25 units/mL Amplitaq® DNA polymerase (Perkin Elmer Cetus, Norwalk CT). PCR parameters were 1 min at 94°C, 1 min at 55°C, 1 min at 72°C, 35 cycles. The products were subcloned into the EcoRI site of pHIL-D4 (Phillips Petroleum, Bartlesville, OK) to generate the plasmids pMP13, pMP14, and pMP15 containing dhaT, dhaBS, and dhaBl, respectively.

Construction of Plasmids for Expression of dhaT The integration vector pMCK5 was digested with EcoRI and dephosphorylated. The dhaT gene was excised as an EcoRI fragment from plasmiH pMP13 and ligated to pMCK5. The resulting plasmid (pMCK7) has dhaT correctly orientated for transcription from the ADHl promoter and contains the URA3 marker.

The replicating plasmid pGAD/KAN2 was digested with EcoRI to remove the kanamycin resistance fragment, dephosphorylated, and ligated to the dhaT EcoRI fragment from pMP13. The resulting plasmid (pMCK13) has dhaT correctly oriented for transcription from the ADHl promoter and contains the LEU2 marker.

Construction of Plasmids for Expression of dhaBl The integration vector pMCK5 was digested with EcoRI, and dephosphorylated. The dhaBl gene was excised as an EcoRI fragment from plasmid pMP15 and ligated to pMCK5. The resulting plasmid (pMCK8) has dhaBl correctly orientated for transcription from the ADH1 promoter and contains the URA3 marker.

The replicating plasmid pGADGH was digested with HindlH, dephosphorylated, and ligated to the dhaBl Hindlll fragment from pMP15. The resulting plasmid (pMCKlO) has dhaBl correctly oriented for transcription from the ADH1 promoter and contains the LEU2 marker.

The integration vector pMCK4 was digested with HindUI and dephosphorylated. The dhaBl gene was excised as an HindUI fragment from plasmid pMP15 and ligated to pMCK4. The resulting plasmid (pMCK16) has dhaBl correctly orientated for transcription from the ADH1 promoter and contains the LEU2 marker.

Construction of Plasmids for Expression of dhaB2 The replicating plasmid pMCKl 1 was digested with EcoRI, dephosphorylated, and ligated to the dhoB2 EcoRI fragment from pMP20. The resulting plasmid (pMCK17) has dhaBl correctly oriented for transcription from the ADH1 promoter and contains the TRP1 marker.

Construction of Plasmids for Expression of dhaB3 The replicating plasmid pYES2 was digested with EcoRI, dephosphorylated, and ligated to the dhaB3 EcoRI fragment from pMP14. The resulting plasmid (pMCKl) has dhaB3 correctly oriented for transcription from the GAL1 promoter and contains the URA3 marker.

The replicating plasmid pGAD AN2 was digested with EcoRI, dephosphorylated, and ligated to the dliaB3 EcoRI fragment from pMP14. The resulting plasmid (pMCK15) has dhaB3 correctly oriented for transcription from the ADH1 promoter and contains the LEU2 marker.

Transformation of S. cerevisiae with dha expression plasmids "S. cerevisiae strain YPH499 {ura3-52 lys2-801 adel-101 trpl-del63 his3-del200 leu2-dell) was transformed with 1-2 μg of plasmid DNA using a LiCl/polyethylene glycol protocol published by Stratagene (Catalog #217406). Colonies were grown on Supplemented Minimal Medium (SMM - 0.67% yeast nitrogen base without amino acids, 2% glucose) for 3-4 days at 29°C with one or more of the following additions: adenine sulfate, uracil, L-tryptophan, L-histidine, L-leucine, L-lysine. Colonies were streaked on selective plates and used to inoculate liquid media. Depending on the vector used, colonies arose either after integration of plasmid DNA or from replication of an episome. In addition to transformations with single plasmid types, co-transfonnations with two or more plasmids were carried out.

EXAMPLE 10 Construction of alternate replicating and integration plasmids for the transformation of S. cerevisi e Construction of Plasmids for Expression of dhaT The integration vector pRS404 (Stratagene, La Jolla, CA) is digested with Kpnl and Sacl. The dhaT gene with flanking promoter and terminator is excised as a Kpnl/Sacl fragment from plasmid pMCK7 and ligated to pRS404. The resulting plasmid has dhaT correctly orientated for transcription from the ADHl promoter and contains the TRP1 marker.

Construction of Plasmids for Expression of dhaBl The integration vector pRS403 (Stratagene, La Jolla, CA) is digested with Qal and Aatll. The dhaBl gene with flanking promoter and terminator is excised as a Clal/AatH fragment from plasmid pMCK8 and ligated to pMCK5. The resulting plasmid has dhaBl correctly orientated for transcription from the ADHl promoter and contains the HIS 3 marker.

The replicating plasmid pYES2 (Invitrogen, San Diego, CA) is digested with HindHI and SnaBI, and the GAL1 promoter element is replaced by ligation with a SnaBI and HindHI digested ADHl promoter fragment from pGADGH. A dhaBl HindHI and Xbal fragment from pMP19 is ligated to those sites in the modified, ADHl promoter version of pYES2. The resulting plasmid has dhaBl correctly oriented for transcription from the ADHl promoter and contains the URA3 marker.

Construction of Plasmids for Expression of dhaB2 The integration vectors pRS403, pRS404, pRS405 and pRS406 are digested with Smal. The dhoB2 gene with flanking promoter and terminator is excised as a SnaBI/Nael fragment from plasmid pMCK17 and ligated to each of the integration vectors. The resulting plasmids have dhaBl correcdy orientated for transcription from the ADHl promoter and contain either the LEU2, TRP1, URA3 or HIS 3 markers.

Construction of Plasmids for Expression of dhaB3 The integration vector pRS404 is digested with Pstl and HincII. The dhaB3 gene with flanking promoter and terminator is excised as a Pstl/EcoRV fragment from plasmid pMCK15 and is ligated to pRS404. The resulting plasmid has dhaB3 correctly orientated for transcription from the ADHl promoter and contains the TRP1 marker.

EXAMPLE 11 Screening of S. cerevisiae for dha trans formants and conversion of D-glucose to 1.3-propanediol bv recombinant Screening of S. cerevisiae for dha Genes Chromosomal DNA from Ura+, His+, Trp+ or Leu+ transformants, constructed as described in Examples 8 and 9, is analyzed by PCR using primers specific for each gene, as described for Pichia pastoris (SEQ ID NO:2-9).

Production of 1.3-Propanediol from D-Glucose bv 5. cerevisiae Transformed with dha Genes Transformants containing dhaT, dhaBl, dhaB2 and dhaB3, constructed as described in Examples 8 and 9, are grown aerobically or anaerobically with shaking at 29°C in SMM supplemented with 20 mg/1 adenine sulfate, 30 mg/1 L-lysine, 1 mg/1 vitamin Bj2- Growth continues until stationary phase is reached and the presence of 1 ,3-propanediol is determined by HPLC.

EXAMPLE 12 Production of 1.3 -propanediol from D-glucose bv Clostridium pasteurianum ATCC 6013 under a hydrogen atmosphere General growth conditions for Clostridium pasteurianum Clostridium pasteurianum ATCC 6013 was grown in 60 mL crimp sealed serum bottles containing 10 mL of medium, unless noted. The crimped bottles containing the medium were aseptically sparged with nitrogen prior to innoculation. Basal medium (Medium A), adjusted to pH 7.2, contained the following components in g/L:KH2P04, 1.4; NaH2P04, 0.69; NH4C1, 1.8-2.5; KC1, 0.50; MgS047 H20, 0.50; CaCl2, 0.025; NaCl, 1.0; yeast extract, 2.0; cysteine'HCl, 0.50; sodium bicarbonate, 2.5; p-amino benzoic acid, 0.0080; biotin, 0.000040; sodium citrate 2 H20, 0.10; FeS04'7 H20, 0.050; CoCl2 6 H20, 6.010; MnCl24 H20, 0.0010; ZnCl2, 0.00050; Na2Mo04 2 H20, 0.0025; NiCl2'6 H20, 0.010; and CuS04'5 H20, 0.0050; to which carbon components were added as indicated below. All incubations were performed at 30°C with shaking at 250 rpm.

A 10 mL batch of Medium A supplemented with 5% glucose was inoculated with 1 ml of a frozen stock of Clostridium pasteurianum ATCC 6013 which contained approximately 15% (v/v) glycerol, in duplicate. After 96 h, 0.5 ml of the growing cell suspensions was passed into 10 ml of fresh medium and growth was continued. After 24 h, the atmosphere in the newly innoculated vials was pressurized to 30 psi with hydrogen gas and incubation was continued for a further 96 h. The aqueous phase was sampled at the beginning and end of the final 96 h for analysis by HPLC as described in the GENERAL METHODS. The results are shown in Table 11.

Table 11 Conversion of D-glucose to 1.3-propanediol bv Clostridium pasteurianum ATCC 6013 under a hydrogen atmosphere Time Glucose Glycerol 1,3-Propane- ' (h) Replicate (inM) (mM) diol (mM)a 0 A 114 nd 2.7 96 A 47 nd 3.4 0 B 119 0.1 2.0 96 B 59 0.1 2.5 aThe presence of 1 ,3-propanediol was confirmed by GC MS as described in the GENERAL METHODS.

EXAMPLE 13 Production of 1.3-propanediol from D-glucose bv Clostridium pasteurianum ATCC 6013 in the presence of methyl viologen Experiment 1. All Cells were grown according to the protocol in Example 12. A 10 mL batch of Medium A supplemented with 5% glucose was inoculated with 1 ml of a frozen stock of Clostridium pasteurianum ATCC 6013 which contained approximately 15% (v/v) glycerol, in duplicate. After 96 h, 0.5 ml of the growing cell suspensions was passed into 10 ml of fresh medium and growth was continued. After 24 h, methyl viologen (l,l'-dimethyl-4,4'-bipyridinium dichloride) was added to the newly innoculated vials to a final concentration of 1 mM and incubation was continued for a further 96 h. The aqueous phase was sampled at the beginning and end of the final 96 h for analysis by HPLC as described in the GENERAL METHODS. The results are shown in Table 12.

Table 12 Conversion of D-glucose to 1.3-propanediol bv Clostridium pasteurianum ATCC 6013 in the presence of methyl viologen Time Glucose Glycerol 1,3 -Propane(h) Replicate (mM) (mM) diol (mM)a 0 A 113 0.3 2.4 96 A 28 1.8 3.4 0 B 87 0.3 2.1 96 B 40 3.2 4.4 aThe presence of 1 ,3 -propanediol was confirmed by GC/MS as described GENERAL METHODS.

Experiment 2. Medium A supplemented with 1% (v/v) glycerol + 1% (w/v) glucose was inoculated from a frozen stock of Clostridium pasteurianum ATCC 6013, which contained approximately 15% (v/v) glycerol, at a ratio of 0.2 ml frozen stock per 20 ml medium. After 48 h, 10 mL of the cell suspension was added to 90 ml of fresh medium and growth was continued for 24 h. The 100 ml cell suspension was chilled on ice and the cells collected by centrifugation under anaerobic conditions. The cells were washed 3x in anaerobic buffer (50 mM phosphate buffer, pH 7.2 + 0.5 g/L cysteine*HCl, previously gassed with N2 and autoclaved under N2) and resuspended in anaerobic buffer to a volume of 8 ml. In duplicate experiments, one mL of this cell suspension was inoculated into 10 ml of Medium A supplemented with 1% glucose and 0 mM, 1 mM, 5 mM, or 10 mM methyl viologen and incubated for 240 h. The aqueous phase was sampled at the beginning and end of the final 240 h for analysis by HPLC as described in the GENERAL METHODS. The results are shown in Table 13.

Table 13 Conversion of D-glucose to 1.3-propanediol bv Clostridium pasteurianum ATCC 6013 in the presence of methvi viologen (MV) MV Time Glucose Glycerol 1,3 -Propane(mM) (h) Replicate (mM) (mM) diol (mM)a 0 0 A 38 nd nd 240 Ab 40 nd nd 0 B nd nd nd 240 B" nd nd nd 1 0 A 40 nd nd 240 Ab nd 4 1 0 B 45 nd nd 240 Bb nd 5 2 5 0 A 37 nd nd 240 A nd 4 2 0 B 38 nd nd 240 B nd 3 1 10 0 A 40 nd nd 240 A nd 2 5 0 B 43 nd nd 240 B nd 3 1 aThe presence of 1 ,3 -propanediol was confirmed by GC MS as described in the GENERAL METHODS. bBy 120 h, glucose was depleted and additional glucose, 1% final concentration, was added.

Experiment 3. Clostridium pasturanium ATCC 6013 was initially maintained in thioglycollate medium (Difco®) and transferred to Medium A supplemented with 0.4% glucose for all subsequent studies. After several transfers through the latter medium, an inoculum was prepared by growing a 1 ml aliquot of stock culture in 10 mL the described medium overnight. A series of serum bottles containing methyl viologen at the concentrations indicated in Table 14 in fresh medium bottles were inoculated with 1 ml of the overnight culture and again incubated for the times indicated in Table 14. Bottles were periodically sampled for glucose utilization and analyzed for the presence of 1,3-propanediol and glucose by HPLC as describerd in the GENERAL METHODS. Table 14 summarizes the analytical results.

Table 14 Production of 1.3-propanedioI from plucose bv Clostridium pasturanium ATCC 6013 methyl viologen time glucose 1,3-propanediol bottle (mM) (days) (mM) (mM)a 1 0 0 22.5 0 1 0 5 0 0 2 0.1 0 26.5 0 2 0.1 5 0 0 3 1.0 0 25.3 0 3 1.0 5 10.0 2.4 3 1.0 9 0 2.4 aThe identity of 1,3-propanediol was verified by GC MS as described in the GENERAL METHODS.

EXAMPLE 14 Construction of General Purpose Expression Plasmids for use in transformation of Bacillus. Streptomyces. and Pseudomonas species The expression vector pTacIO The E. coli expression vector, pTacIQ contains the laclq gene (Farabaugh, Nature 274, 673 (1978)) and tac promoter (Amann et al., Gene 25, 167 (1983)) inserted into the EcoRI of pBR322 (Sutcliffe et al., Cold Spring Harb. Symp.

Quant. Biol. 43, 77 (1979)). A multiple cloning site and terminator sequence (SEQ ID NO 10) replaces the pBR322 sequence from EcoRI to Sphl.

Subcloning the glycerol dehydratase genes (dhaBI. 2. 3) The open reading frame for dhaB3 gene was amplified from pHK28-26 by PCR using primers (SEQ ID NO:41 and 42), incorporating an EcoRI site at the 5' end and a Xbal site at the 3 ' end. The product was subcloned into pLitmus29 (New England Biolab, Inc., Beverly, MA) to generate the plasmid pDHAB3 containing dhaB 3.

The region containing the entire coding region for the four genes of the dhaB operon from pHK28-26 was cloned into pBluescriptll KS+ (Stratagene, La Jolla, CA) using the restriction enzymes Kpnl and EcoRI to create the plasmid pM7.

The dhaBX gene was removed by digesting the plasmid pM7, which contains dhaB (1 ,2,3,4), with Apal and Xbal (deleting part of dhaB 3 and all of dhaBX). The resulting 5.9 kb fragment was purified and ligated with the 325-bp Apal-Xbal fragment from plasmid pDHAB3 (restoring the dhaB 3 gene) to create pMl l, which contains dhaB(l,2,3).

The open reading frame for the dhaBl gene was amplified from pHK28- 26 by PCR using primers (SEQ ID NO: 11 and SEQ ID NO: 12) incorporating a HindlTT site and a consensus RBS ribosome binding site at the 5' end and a Xbal site at the 3 ' end. The product was subcloned into pLitmus28 (New England Biolab, Inc.) to generate the plasmid pDTl containing dhaBl.

A Notl-Xbal fragment from pMl 1 containing part of the dhaBl gene, the dhaB 2 gene and the dhaB3 gene was inserted into pDTl to create the dhaB expression plasmid, pDT2. The HindHI-Xbal fragment containing the dha (1,2,3) genes from pDT2 was inserted into pTacIQ to create pDT3.

Subcloning the 1.3 -propanediol dehydrogenase gene (dhaT) The Kpnl-Sacl fragement of pHK28-26, containing the complete 1,3 -propanediol dehydrogenase (dhaT) gene, was subcloned into pBluescriptll KS+ creating plasmid pAHl . The dhaT gene was amplified by PCR from pAHl as template DNA using synthetic primers (SEQ ID NO 13 with SEQ ID NO 14) incorporating an Xbal site at the 5' end and a BamHI site at the 3' end. The product was subcloned into pCR -Script (Stratagene) at the Srfl site to generate the plasmids pAH4 and pAH5 containing dhaT. The plasmid pAH4 contains the dhaT gene in the correct orientation for expression from the lac promoter in pCR -Script and pAH5 contains the dhaT gene in the opposite orientation. The Xbal-BamHI fragment from pAH4 containing the dhaT gene was inserted into pTacIQ to generate plasmid pAH8. The Hindll-BamHI fragment from pAH8 containing the RBS and dhaT gene was inserted into pBluescriptll KS+ to create pAHl 1. The Hindlll-Sall fragment from pAH8 containing the RBS, dhaT gene and terminator was inserted into pBluescriptll S + to create pAH12.

Construction of an expression cassette for dhaB(1.2.3) and dhaT An expression cassette for the dhaB(l,2,3) and dhaT was assembled from the individual dhaB(l ,2,3) and dhaT subclones described above using standard molecular biology methods. The Spel-Kpnl fragment from pAH8 containing the RBS, dhaT gene and terminator was inserted into the Xbal-Kpnl sites of pDT3 to create pAH23. The Smal-EcoRI fragment between the dhaB3 and dhaT gene of pAH23 was removed to create pAH26. The Spel-Notl fragment containing an EcoRI site from pDT2 was used to replace the Spel-Notl fragment of pAH26 to generate pAH27.

Construction of expression cassette for dhaT and dhaB( 1.2.3) An expression cassette for dhaT and dhoB(l ,2,3) was assembled from the individual dhaB(l,2,3) and dhaT subclones described previously using standard molecular biology methods. A Spel-Sacl fragment containing the dhaB(l,2,3) genes from pDT3 was inserted into pAHl 1 at the Spel-Sacl sites to create pAH24.

EXAMPLE 15 Construction of Expression Cassetts and Plasmids for the Expression of dhaB(12.3) and dhaT in S. lividans Subcloning the glucose isomerase promoter Two versions of the glucose isomerase promoter from the vector pCOsl21 (SEQ ID NO: 15) or pC01211ow (SEQ ID NO: 16) were amplified by PCR using primers (SEQ ID NO: 17 and 18) incorporating Spel and EcoRI sites at the 5' end and a Hindin site at the 3' end. The products were subcloned into pLitmus29 (New England Biolabs, Inc., Beverly, MA) to generate the plasmids pDT7 and pDT8.

Construction of an expression cassette for dhaB(1.2.3) and dhaT '•The 4.1 kb expression cassette for dhaB(l,2,3) and dhaT horn pAH27 (Example 14) was inserted into pDT7 or pDT8 using the restriction enzymes Hindm and Sail to create pDTl l and pDT12, respectively.

Construction of a plasmid for co-expression of dhaBd .2.3) and dhaT in Streptomyces The 4.3 kb expression cassette for dhaB( 1 ,2,3) and dhaT was removed from pDTl 1 or pDT12 by digestion with EcoRI and Sail. The vector pIJ488-101 was digested with the restriction enzymes EcoRI and Xbal. The expression cassette and vector were ligated along with a Sal-Xba Linker (SEQ ID NO 19 and 20) to create pDT13 and pDT14, respectively. pD488-101 consists of replication origin from ρΠΙΟΙ from Streptomyces lividans (Kendall and Cohen, /. Bacteriol. 170, 4634 ((1988)) and pUC18 from E. coli (Norrander et al., Gene, 26, 101 ((1983)). The sequences are derived as follows: bases 1-2086 are from pIJlOl (1-2086), and bases 7688-8437 are from pIJlOl (8080-8830). Bases 2087-3368 are from the thiostrepton resistance gene from 5. azureus (Thompson et al., Gene, 20, 51 (1982)). Bases 3369-7687 are from pUC18 containing the erythromycin resistance gene from S. erythreus (Thompson et al., supra) inserted at the Kpnl site.

EXAMPLE 16 Production of 1.3-propanediol from D-elucose using recommbinant Streptomyces lividans Transformation of Streptomyces lividans with dhaB(12.3) and dhaT The plasmids pDT13 or pDT14 are transformed into Streptomyces lividans TK23 using standard protoplast transformation techniques (Hopwood et al., Genetic Manipulation of Streptomyces. The John Innes Foundation (1985)). The transformations are selected on plates containing 50 pg/ml thiostrepton incubated at 30°C. Spores from the transformants are replated to obtain pure cultures.

Detection of glycerol dehydratase and 1.3-propanediol dehydrogenase activity The Streptomyces transformants are grown in 25 ml of TSB (Tryptone- Soy Broth, Difco, Detroit, MI) plus 50 pg/ml thiostrepton at 30°C for 2-4 days. The cells are harvested by centrifugation and resuspended in 1 ml of 100 mM Tris buffer, pH 7.4. The cells are broken using a French Press (20,000 psi) and the cell extract is assayed for glycerol dehydratase and 1, 3-propanediol dehydrogenase as described in GENERAL METHODS.

Production of 1.3-propanediol from D-glucose in recombinant Streptomyces lividans Growth for demonstration of 1,3 -propanediol production by Streptomyces lividans TK23 containing pDT13 or pDT14 proceeds aerobically at 30°C in shake-flask cultures (erlenmeyer flasks, liquid volume 1/10th of total volume).

Cultures in rich media shake-flasks are started by inoculation from two-days old TSA-plates (Trypticase Soy Agar, BBL #11043). Rich media are either TSB (Trypticase Soy Broth; BBL #11768), Liquid Broth ((which contains per liter: 16 g tryptone, 10 g yeast extract, and 5 g NaCl, 2YT , medium B (TSB supplemented with per L: 10.0 g glucose, 2 ml Modified Balch's Trace-Element Solution, in which NTA is replaced by citric acid, 2.0 g Na2C03. 4.0 g K2HP04, 1 mg vitamin B^, final pH = 7.2). or medium C (medium B, at pH 6.4). The composition of Balch's trace-element solution can be found in Methods for General and Molecular Bacteriology (P. Gerhardt et al., eds) p. 158, American Society for Microbiology, Washington, DC (1994). Cultures in nTinimal media shake-flasks are started by inoculation from two-days old liquid TSB cultures, 130789/2 using a 1/30 (v/v) inoculum. Minimal media are either: MM322 (which contains per liter: 12.0 g glucose, 11.3 g 2HPO4, 1.0 g (NH^SC^ 0.2 g Difco yeast extract, 0.1 g NaCl, 2 mg vitamin B12 and 10 ml Modified Batch's Trace-Element Solution modified as above, final pH 6.7 (HCl)); medium D (medium MM322 supplemented with 2 g Na2C03/L, final pH 7.2); or medium E (medium D, final pH 6.4). Media B and C and the minimal media are filter-sterilized, the other media are autoclaved.

The shake-flasks are incubated at 30°C with vigorous shaking for two days, after which they are sampled for HPLC analysis of the supernatant. Glucose is added, the culture is incubated for 1 h under aerobic conditions, after which the culture is transferred to 25 ml volume glass tubes (which are nearly filled to the top). These tubes are subsequently incubated under anaerobic conditions at 30°C. After incubating for 2-5 days, 1,3-propanediol in the supernatant is detected by HPLC as described in GENERAL METHODS.

EXAMPLE 17 Construction of general purpose plasmids and plasmids for the overexpression of dhaBCJ-3) and dhaT in Bacillus Construction of general purpose expression plasmids The replicative high copy number shuttle vector pVS02 is used to co-express dhaB(l-3) and dhaT in Bacillus. pVS02 was contracted by cloning an EcoRI/BamHl fragment carrying an alkaline serine protease from Bacillus lentus fused to the B. subtilis apr promoter into pBS19. pBS19 is a derivative of pBS42 (Band and Henner, DNA 3, 17 (1984)) in which the EcoRI/BamHI fragment has been replaced by the EcoRI/Hindlll polylinker from pUC19 (Boehringer Mannheim). To facilitate sequencing and PCR reactions, a 45 bp synthetic linker (SEQ ID NO 21) was introduced by PCR between the end of the protease gene and the transcriptional terminator.

The replicative low copy number shutle vector pSS15-B is used to co-express dhaB(l-3) and dhaT in Bacillus. Plasmid pSS15-B was constructed by digesting plasmid pHP13 (Haima et al., Mol. Gen. Genet. 209, 335 (1987)) with Hindlll/Sall (sites present in polylinker), filling the ends with T4 DNA Polymerase and religating to generate pSS13. A 2 kb EcoRI/BamHI fragment from pVS02 was inserted into the EcoRI/BamHI site of plasmid pSS13 to create pSS15-B.

Plasmids for the over-expression of a dhaB(l-3) and dhaT cassette In order to create a Bacillus consensus ribosome binding site at the 5' end of dhaT, an EcoRI/Xba linker obtained by annealing synthetic primers (SEQ ID NO:22 with SEQ ID NO:23) was inserted into the EcoRI Xbal site of pAH23 to create pM17. A Hindlll/Bglll linker, using synthetic primers (SEQ ID NO:24 with SEQ ID NO: 25) was added at the HindH/bglH site of plasmid pM17 to introduce a Sail site at the 5' end of dhaBl to create pM20. The 0.3 kb Mlul/Kpnl fragment from plasmid pM20 was replaced with the 0.3 kb Mlul/Kpnl from plasmid pAH4 to introduce a HindHI site to create pM21.

A Sall-Xbal linker (SEQ ID NO 26 & 27) was inserted into pAH5 which was digested with the restriction enzymes, Sall-Xbal to create pDT15. The linker destroys the Xbal site and changes the reading frame so that the dhaT gene will be fused to the open reading frame of protease coding sequence of plasmids pSS15-B and pVS2. The 1 kb Sall-Mlul fragment from pDT15 was then inserted into pAH24, replacing the existing Sall-Mlul fragment to create pDT17.

A Sall-Xbal linker (SEQ ID NO 28 & 29) was inserted into pAH5 which was digested with the restriction enzymes Sall-Xbal, to create pDT16. The linker destroys the Xbal site and changes the reading frame so that the dhaT gene will be fused to the open reading frame of poly-His coding sequence of pUSHl (Schon and Schuman, Gene 147, 91 (1994)). The 1 kb Sall-Mlul fragment from pDT16 was then inserted into pAH24 replacing the existing Sall-Mlul fragment to create pDT18.

Plasmid pDT4 (containing dh B(l-3)) was constructed by introducing the 2.7 kb EcoRI/Xbal fragment from pDT2 into pUC18 (Boehringer Mannheim) digested with EcoRI Xbal.

Plasmids for the over-expression of dha B( 1-3) in Bacillus using a lac-based inducible system A 2.7 kb Bgin/HindHI fragment containing dhaB(l-3) from plasmid pDT4 was cloned into the HmdlTI/BamHI site in the polylinker of pUSHl to create pM26. The dhaBl gene was fused to the open reading frame of poly-His coding' sequence of pUSHl.

EXAMPLE 18 Plasmids for the over-expression of dhaT and dhaB(l-3) cassette in Bacillus.: Constitutive and inducible Plasmids for the over-expression of dhaT and dhaB(l-3) cassette in Bacillus. Constitutive A Sall/Hindlll fragment from plasmid pM 21, containing dhaB(l-3) and dhaT, is ligated with the 5 kb Sal/HindlH pVS02 vector to create pM22. pM22 has dhaB and dhaT under the apr promoter in a high copy number vector.

A Sall/Hindlll fragment from plasmid pM21, is ligated with the 5.8 kb Sal/Hindm fragment from pSS15-B to create pM23. pM23 has dhaB and dhaT under the apr promoter in a low copy number vector. pDT17 is digested with Sacl, ends are filled with T4 DNA polymerase, and DNA is digested with Sail to release the fragment containing dhaT and dhaB. The fragment is then ligated to pSS15-B and pVS02 digested with Hindlll (ends blunted with T4 DNA polymerase) and Sail to create pM25 and pM27, respectively.

Plasmids for the over-expression of dhaT and dha B(l-3) cassette in Bacillus using a lac-based inducible system. pDT18 was open with Sacl, ends were filled with T4 DNA polymerase, and DNA is digested with Sail to release the fragment containing dhaT and dhaB, both genes contain a Bacillus consensus ribosome binding site. The fragment is then ligated to pUSHl (Schon and Schuman, supra) digested with HindlQ (ends blunted with T4 DNA polymerase) and Sail to create pM24.

EXAMPLE 19 Conversion of D- lucose to 1.3 -propanediol bv recombinant Bacillus Transformation of plasmids into Bacillus The plasmids pM22, 23, 24 and 25 are transformed into B. licheniformis BG307 by natural transformation (McCuen and Thome, /. Bacteriol. 107, 636-645 (1971)). The same plasmids are transformed into B. licheniformis strain BG188 using standard protoplast transformation techniques (Pragai et al., Microbiology 140, 305 (1994)). B. subtilis strains BG2864 and BG2232 are transformed with plasmids by natural transformation. Transformants containing plasmids are selected on LA plates contaming 7.5 ug/ml Chloramphenicol at 30°C and 37°C.

Plasmids pM24 and pM26 are also transformed into B. subtilis strain 1E62 (Saito et al., Mol. Gen. Genet. 170, 117 (1979)) and transformants containing plasmids are selected on LA plates containing 10 ug/mL Erythromycin and 20 ug/ml Kanamycin at 32°C.

Detection of glycerol dehydratase and 1.3 -propanediol dehydrogenase activity The Bacillus transformants are grown in 25 mL of LB (Difco) plus antibiotics at 32°C or 37°C for 2-4 days. The cells are harvested by centrifugation and resuspended in I ml of 100 mM Tris buffer, pH 7.4. The cells are broken using a French Press (20,000 psi) and the cell extract is assayed for glycerol dehydratase and 1 ,3-propanediol dehydrogenase activity as described in GENERAL METHODS. 130789 2 Production of 1.3-propanediol in Bacillus Growth conditions for Bacillus Growth for demonstration of 1,3-propanediol production by Bacillus licheniformis and Bacillus subtilis proceeds aerobically at 30°C or 35°C (as indicated) in shake-flask cultures (erlenmeyer flasks) and in 15.5 L (total volume) Biolafitte fermenters (working volume 7-10 liters).

Cultures in LBG (which contains per liter: 16 g tryptone, 10 g glucose, 10 g yeast extract, and 5 g NaCl) shake-flasks are started by inoculation from one-day old TSA-plates (Trypticase Soy Agar, BBL #11043). These shake-flasks are then used to inoculate either fermenters or shake-flasks in which the demonstration proper of D-glucose to 1,3-propanediol conversion is demonstrated.

Batch cultures in shake-flasks Rich media are either TSB (Trypticase Soy Broth; BBL #11768), LBG, medium B (TSB supplemented with per L: 10.0 g glucose, 2 ml Modified Balch's Trace-Element Solution, in which NTA is replaced by citric acid, 2.0 g Na2C03, 4.0 g K2HPO4, 1 mg vitamin B12, final pH = 7.2), or medium C (medium B, at pH 6.4). The composition of Balch's trace-element solution can be found in Methods for General and Molecular Bacteriology (P. Gerhardt et al., eds) p. 158, American Society for Microbiology, Washington, DC (1994). Minimal media are either: MM322 (which contains per liter: 12.0 g glucose, 11.3 g 2HPO4, 1.0 g (NH^SO^ 0.2 g Difco yeast extract, 0.1 g NaCl, 2 mg vitamin B12 and 10 ml Modified Balch's Trace-Element Solution modified as above, final pH 6.7 (HC1)); medium D (medium MM322 supplemented with 2 g Na2C03/L, final pH 7.2); or medium E (medium D, final pH 6.4). Media B and C and the minimal media are filter-sterilized, the other media are autoclaved.

The shake-flasks are incubated at 30°C with vigorous shaking for one day, after which they are sampled for HPLC analysis of the supernatant. Glucose is added, the culture is incubated for 1 hr under aerobic conditions, after which the culture is transferred to 25 ml volume glass tubes (which are nearly filled to the top). These tubes are subsequently incubated under anaerobic conditions at 30°C. After incubation for 1-5 days, 1,3-propanediol in the supernatant is detected by HPLC as described in GENERAL METHODS.

Batch and fedbatch cultures in fermenters A 600-ml total volume culture from a shake-flask (LBG medium) is used to inoculate a fermenter with 6.4 liter of medium, batched in and autoclaved for 30 minutes (minimal media) or 45 minutes (rich media) Typical minimal media in the fermenter is medium D, typical 'rich' media is media D with an additional 130789/2 50 g yeast extract/L. Filter-sterilized additions (vitamin B12 or compensations for auxotrophy) are performed after the fermenter has been autoclaved, using a syringue and a septum-port in the fermenter lid.

Back pressure (BP, 0.1-0.5 bar), aeration (liters of air per minute, 0.4-1 wm), stirring (rpm, 200-600), temperature (T, 30-37 °C), Dissolved Oxygen (%DO), and pH (5.8-7.2, by NH4OH and H2S04 or H3PO4 addition) are monitored and controlled at the desired values, as indicated.

After inoculation, the cells are allowed to grow for at least 4 hours. For anaerobic growth/production, the %DO is allowed to go to 0% by either reduction of rpm and BP, additionally by replacing the air going in by N2, as indicated.

Fermenters and shake-flasks are sampled for OD550 readings (growth) and an enzymatic glucose assay on the supemate; supemate is also prepared for HPLC analysis via our standard procedure, as outlined.

EXAMPLE 20 Transformation of Pseudomonas with dhaB(1.2.3) and dhaT and demonstration of glycerol dehydratase and 1.3-propanediol dehydrogenase activity Construction of plasmids for co-expression of dh B(l.Z3) and dhaT in Pseudomonas The 4.1 kb expression cassette for dhaB(l,2,3) and dhaT hem pAH27 was inserted into the vectors pMMB66EH ( Gene 48, 119 (1986)) and pMMB207 (Morales et al., Gene 97, 39 (1991)) using the restriction enzymes EcoRI and Sail to create pDTIO and pDT9, respectively.

Transformation of P. aeruginosa PAO 2845 with the pDT9 expression plasmid P. aeruginosa PAO 2845 cells were prepared for transformation by overnight growth at 37°C with shaking at 200 rpm in L-broth. A 1:25 inoculation of the culture was made into 25 ml of fresh prewarmed and preaerated L-broth. The fresh culture was incubated 2-3 h to early log phase at 37°C and 200 rpm. The cells, collected by centrifugation, were washed twice in 10 mis of ice cold 0.15 M MgCl2 containing 5% dimethylsulfoxide and resuspended in 2 mL of the dimethylsulfoxide solution. The cell suspension (0.2 mL) was combined with 100-200 ng pDT9 DNA and placed on ice for 60 minutes. The reaction mixture was heat shocked at 37°C for 2 minutes and transferred to ice for 5 minutes. L-broth (0.5 mL) was added and the cells were incubated for 20 minutes at 37°C. Single colonies were obtained from nutrient agar plates supplemented with 37.5 ug/ml chloramphenicol.

Conjugal transfer of pDTIO into P. aeruginosa PAOl The plasmid, pDTIO was mated into PAOl by the method of Figurski and Helinski (1979). pDTIO was transformed into K coli AC80 (Chakrarty et. al., 07 Proc. Natl. Acad. Sci. U. S. A. 75, 3109 (1978)) to create a donor strain. The helper strain was E. coli HB101 containing pRK2013 (Figurski and Helinski, Proc. Natl. Acad. Sci. U. S. A. 76, 1648 (1979)). The recipient strain was Pseudomonas aeruginosa PAOl (Royle et al., J. Bacteriol. 145, 145 (1981)). Cultures (5 mL) of each of the strains were grown overnight in LB at 37°C. The cells were washed in 0.9% NaCl and resuspended in 200 μΐ. The cells were mixed together and spread on a LA plate (Luria Agar, Difco). The plate was incubated at 37°C for 6 hrs. The cells were removed from the plate and transferred to PI A (Difco) plates containing 250 μg/ml carbencillin and grown overnight. Single colonies were isolated on the same media.

Detection of glycerol dehydratase and 1.3-propanediol dehydrogenase activity Pseudomonas aruginosa PAOl/pDTlO was grown in 25ml 2XYT (16 g/L Yeast Extract, 16 g L Tryptone, 5g/L NaCl) plus 250 μg/ml carbenicillin, O.lmM IPTG overnight at 37°C. The cells were harvested by centrifugation and resuspended in 1 ml of lOOmM Tris buffer pH7.4. The cells were broken by French Press at 15,000 psi. The crude extract was then assayed for glycerol dehydratase and 1,3 -propanediol dehydrogenase activity using standard assays. Protein determination was by Bio-Rad method. Specific activity for glycerol dehydratase was 5 U/mg. Specific activity for 1,3-propanediol dehydrogenase was 20 U/mg. Similarly prepared, crude extract from P. aeruginosa PAO 2845 transformed with pDT9 contained 0.05 U/mg glycerol dehydratase activity.

EXAMPLE 21 Production of 1.3-propanediol from glucose using Pseudomonas aeruginosa General growth conditions Pseudomonas aeruginosa strain PA02845 from PGSC (Pseudomonas Genetic Stock Center, East Carolina School of Medicine, Greenville, NC) is grown in basal medium, HEPES0.1, with the following components: NH4CI, 9.52 mM; MgCl26H20, 0.523 mM; K2S04, 0.276 mM; HEPES (N-[2-Hydroxyethyl]piperazine-N-[2-ethanesulfonic acid]), 40 mM; Tricine (N-Tris (hydroxymethyl) methyl glycine), 4 mM; FeS04-7H20, 0.010 mM; K2HP04, 0.132 mM. A trace mineral solution is added to give final concentrations of these components in g/L: sodium citrate-6H20, 0.001; FeS04-7H40, 0.0005; CoCl2-6H20, 0.0001; MnCl2-4H20, 0.00001; ZnCl2, 0.000005; Na2Mo04-2H20, 0.000025; NiCl2-6H20, 0.0001; CuS02-2H20, 0.00005. HEPES0.1 is used in all experiments; supplementations are noted where they occur.

Construction of a glycerol negative mutant of P. aeruginosa PAO 2845 bv gene interruption P. aeruginosa PAO 2845 is grown overnight in Nutrient Broth (Difco, Detroit, MI) at 37°C and 200 rpm shaking. Cells are recovered by centrifugation and DNA extracted from cells using a standard alkaline lysis procedure (Sambrook 1989). The open reading frame for glpR (glycerol catabolism 'regulatory protein gene, Genbank ACCESSION # M60805) is amplified from P. aeruginosa PAO 2845 by PCR using primers JJ-gplR-5'and JJ-glpR-3'(Seq ID Nos. 30 and 31, respectively), incorporating EcoRl sites at the 5' ends. This DNA fragment is then ligated into plasmid pARO180 (Parke,D., 1990) at its unique EcoRl restriction site resulting in plasmid pJJIO. E. coli transformed with DNA from the pJJIO ligation mix are spread on Nutrient Agar (Difco, Detroit, MI) containing 50 pg/ml ampicillin and 0.08 mg/ml Xgal (5-bromo-4- chloro-3-indolyl^-D-galactoside). White colonies, indicating a high probability of glpR insertion, are picked and transferred to LB medium supplemented with 50 pg/ml ampicillin. From cells harvested after overnight growth at 37°C and 200 rpm shaking, pJJIO DNA is recovered.

The kanamycin cassette region from pUC4K (Pharmacia Cat. No. 27-4958-01) is amplified by PCR using primers (SEQ ID NO:32 AND 33) appropriately designed to amplify the region and modify the temiirii of the fragment to be compatible with restriction enzyme Styl (Promega, Madison, WI) resulting in the 4 kb fragment pUC4K-ifyl. The pUC4K-jryl DNA fragment is subcloned into the Sr l site within the glpR gene of plasmid pJJIO, generating plasmid pJJl 1. E. coli transformed with DNA from the pJJl 1 ligation mixture are spread on LB agar supplemented with 25 pg/mL kanamycin and 50 pg/mL ampicillin. DNA from 5-20 isolated colonies is individually collected following overnight growth at 37°C in LB medium supplemented with 25 pg/ml kanamycin and 50 pg/ml ampicillin. The presence of the desired plasmid DNA is confirmed by gel electrophoresis.

P. aeruginosa PAO 2845 is transformed with pJJl 1 DNA following standard protocols. Briefly, P. aeruginosa cells are prepared for transformation by overnight growth at 37°C with shaking at 200 rpm in L-broth. A 1 :25 inoculation of this overnight culture is made into 25 ml of fresh prewarmed and preaerated L-broth. The fresh culture is incubated for 2-3 h (to early log phase) at 37°C and 200 rpm. Cells are centrifuged and supernatant decanted. Collected cells are resuspended in 10 mL of ice cold sterile 0.15 M MgCl2 containing 5% dimethylsulfoxide and held on ice for 5-10 minutes. Cells are centrifuged, separated from the supernatant and resuspended in 10 ml of ice cold sterile 0.15 M MgCl2 containing 5% dimethylsulfoxide and held on ice for 5-10 minutes. After a final centrifugation and separation from the supernatant, the cells are resuspended in 2 ml of ice cold sterile 0.15 M MgCl2 containing 5% dimethylsulfoxide. A 0.2 mL aliquot of the cold cell concentrate is combined with 100-200 ng pJJl 1 DNA in a prechilled 1.5 mL polypropylene centrifuge tube and the mixture is held on ice for 60 minutes. The tube is then rapidly transferred to a 37°C water bath for 2 minutes and immediately returned to ice for 5 minutes. Approximately 0.5 ml of L-broth is added and the cells are incubated for 0.3-1 hour with gende shaking at 37°C. Following the recovery incubation, 10 μΐ and 50 μΐ aliquots of the cell suspension are spread on nutrient agar plates supplemented with 50 pg/mL kanamycin. Colonies developing on the selective medium are screened for growth on agar plates with HEPES0.1 medium supplemented with 1% succinic acid or 1% glycerol. Clones unable to grow on glycerol, but capable of growth on succinate, are preserved for later use by freezing in 15% glycerol.

Transformation of gplRzP. aeruginosa PAO 2845 with pDT9 P. aeruginosa is prepared for transformation by the method described above. A 0.2 ml aliquot of the cold cell concentrate is combined with 100-200 ng pDT9 DNA in a prechilled 1.5 mL polypropylene centrifuge tube and the mixture held on ice for 60 minutes. The tube is then rapidly transferred to a 37°C water bath for 2 minutes and immediately returned to ice for 5 minutes. Approximately 0.5 mL of L-broth and cells are incubated 0.3-1 h with gende shaking at 37°C. Following the recovery incubation, 10 pL and 50 pL aliquots of the cell suspension are spread on nutrient agar plates supplemented with 37.5 pg/mL chloroamphenicol .

Screening of glpR' P. aeruginosa PAO 2845 transformants for the presence ΡΡΤ2· The transformants from above are plated on nutrient agar plates supplemented with 37.5 pg/mL chloramphenicol grown overnight at 37°C. From the colonies appearing on these selective plates, approximately twenty are picked and transferred to 10 mL Nutrient Broth (Difco, Detroit, MI) containing 37.5 pg/mL chloramphenicol and grown overnight at 37°C and 200 rpm shaking. To confirm the presence of the pDT9 plasmid in the selected transformants, plasmid DNA is extracted, purified and cut with EcoRl (Promega, Madison, WI). The molecular weight of the linearized DNA is analyzed by gel electrophoresis. In addition, PCR amplification using primer pairs with sequences common to dhaT, dhaBl, dhaB2, and dhaB3 (SEQ ID NO:34 AND 35, 36 AND 37, 8 AND 9, 4 AND 5, respectively) followed by fragment molecular weight characterization using gel electrophoresis is used to confirm the presence of the desired genes.

Metabolic screening of glpR- P. aeruginosa PAO 2845 transformed with pDT9 One to twenty clones are selected from the positive transformants above for further characterization. Cells are grown aerobically on Nutrient Broth supplemented with 37.5 pg/L chloramphenicol overnight at 30°C with shaking at 250 rpm. Cells are transferred at a 1:8 dilution into the same medium with 1.5 mM IPTG (isopropyl-P-D-thiogalactoside) and grown for 4-6 h. Cells are then harvested by centrifugation and washed once with HEPESO.1 medium supplemented with 10 g/1 glycerol, 0.03 g/L beef extract, 0.05 g/1 peptone, 0.05 g 1 yeast extract (all Difco, Detroit, MI) and 0.2% KNO3. The cells are then resuspended at 1/5 the original volume, with no air space, in a small vial and incubated at 30°C with shaking at 100 rpm for 18-72 h. Cells are removed by centrifugation and the supematants analyzed for the presence of 1 ,3-propanediol by HPLC. In addition, the chemical identity of 1,3-propanediol is confirmed by gas chromatography-mass spectrometry.

Production of 1.3-propanediol from glucose bv glpR- *. aeruginosa PAO 2845 transformed with pDT9 From the screening procedure above, one to five clones which produce the greatest amount of 1,3-propanediol from glycerol are grown aerobically on nutrient broth supplemented with 37.5 pg/L chloramphenicol overnight at 30°C with shaking at 250 rpm. Cells are transferred at a 1:8 dilution into the same medium with 1.5 mM IPTG, allowed to grow for 4-6 h, harvested by centrifugation and washed once with HEPESO.1 medium supplemented with 10 g/L glucose, 0.03 g/L beef extract, 0.05 g/L peptone, 0.05 g/L yeast extract and 0.2% KNO3. The cells are then resuspended at 1/5 the original volume, with no air "space, in a small vial. Cells are incubated at 30°C with shaking at 100 rpm for approximately 36 hours. Cells are removed by centrifugation and the supematants analyzed for the presence of 1 ,3-propanediol by HPLC In addition, the chemical identity of 1,3-propanediol is confirmed by gas chromatography-mass spectrometry.

EXAMPLE 22 Construction of Expression Cassettes for Expresion of dhaB l . dhaBl. dhaBl and dhaT in Aspergillus niger: General Expression Cassette (pAEX): The 1.4 kb. Spel-EcoRV fragment from the plasmid pGPTpyrG (Berka, RM, Garnett, CC (1989) "The development of gene expression systems for filamentous fungi." Biotechnol. Adv. 7: 127-154), containing sufficient portions for proper regulation of the Aspergillus niger gla A promoter and terminator, was ligated into the Spel and EcoRV sites in the polylinker of pLITMUS39 (New England Biolabs, Beverly, MA).

Individual Clone Expression Cassettes for A. niger: The open reading frames (ORF's) for individual Klebsiella pneumoniae dhaB subunits and the dhaT subunit were cloned and ligated into the general expression vector (pAEX) separately, using the same cloning strategy: Primer pairs for PCR amplification of each individual dhaB ORF and the dhaT ORF were designed to match the 5' and 3' ends sequence for each ORF based on known sequence of the entire gene operon {dhaBl, dhaB2, and dhaB 3 and dhaT: SEQ ID NO:38 and 12, 39 and 40, 41 and 42, 43 and 44, respectively). In addition to the matching sequence, the primers for the 5' end of each ORF were designed to include an EcoRl restriction site followed by a Bgl Π restriction site at the 5' most end of the sequence as well as the five base sequence CAGCA upstream of the first ATG of each ORF. Primers designed to match the 3 ' ends of each ORF placed an Xbal restriction site downstream of the translation stop codon, at the 3' most end of the clone.

Individual clone fragments for the dhaB and dhaT ORF's were amplified by PCR from the plasmid pHK26-28, containing the entire K. pneumoniae dha operon, using the primers described above. The individual ORF clone fragments were isolated based on their respective molecular weights (dhaBl= 1540bp; dhaB2=601bp; dhaB3=448bp; dhaT=l 187bp). Using the unique EcoRl and Xbal restriction sites designed in the PCR primers, each individual dhaB and dhaT ORF fragment was ligated into the EcoRl and Xbal restriction sites in the polylinker of pLITMUS29 (New England Biolabs). The dhaBl and dhaBl clones in pLITMUS29 were confirmed to be correct by sequencing. A unique 1363 b.p. Ncol -EcoRV restriction fragment from the coding region of dhaBl clone in pLTTMUS29 was removed and replaced with the corresponding restriction fragment from pHK26-28. A unique 783 b.p. Tthl 11 1-Mlu I restriction fragment from the coding region of dhaT clone in pLITMUS29 was replaced with the corresponding restriction fragment from pHK26-28. The 5 ' and 3' end sequences of the dhaBl and dhaT clones, approximately 250 b.p. which includes some sequence from the substituted fragment, was confirmed to be correct by sequencing.

The unique Bgl Π-Xbal restriction fragments containing the ORF's of dhaBl, dhaBl, dhaBl and dhaT clones in pLITMUS29 were ligated into the Bgl Π-Xbal restriction sites in the general expression vector pAEX separately, placing expression of each clone under the control of the A. niger glaA promoter and terminator. Each resulting vector was named by the respective ORF, i.e.: pAEX.dh Bl, pAEX:dhaB2, pAEX:dfiaB3 and pAEXidhaT.

Dual Expression Cassette Vectors for A. niger: The unique SnaBl-Stul restriction fragment containing the dhaBl expression cassette (consisting of the A. niger glaA promoter, the dhaBl ORF, and terminator) was isolated from the vector pAEX:dhaB 1 and ligated into the unique SnaBl restriction site in the pAEX:dhaB2 vector. The approx. 2.2 kb Scal-Smal restriction fragment from pBH2 (Ward et. al., Exp. Myc. 13, 289 (1989) containing the Aspergillus nidulans pyrG auxotrophy selectable marker, was ligated into the unique Stul restriction site in the vector containing the dhaBl and dhaBl expression cassettes. This vector was named pAEX:Bl+B2.

The unique Spel-Hind ΠΙ restriction fragments containing the entire expression cassettes for dhaB3 and dhaT were isolated from the respective pAEX:dhaB3 and pAEX dhaT vectors. The two expression cassette fragments were simultaneously ligated, in tandem, in the unique Hind ΙΠ restriction site in the vector pUC18. This vector was named pAEX:B3+T.

Transformation . Isolation of Transformants and Confirmation of Integration of Expression Cassettes: Aspergillis niger strain FS1 ipyrG' ) was co-transformed with the two expression vectors pAEX:Bl+B2 and pAEX:B3+T using the method of Campbell (Campbell et. al. (1989), "Improved transformation efficiency of A. niger using homologous niaD gene for nitrate reductase." Curr. Genet. 16:53-56).

Transformants were selected for by their ability to grow on selective media without uridine. Genomic DNA of transformants was digested with Hind EQ and Spel to liberate fragments of predictable molecular weights, demonstrating integration of intact expression cassettes. Detection of each expression cassette was done by Southern analysis, probing with individual genes separately.

EXAMPLE 23 Production of 1.3 -propanediol from maltose using A. niger Confirmation of Expression of each dh B and dhaT gene: Expression of each ORF is tested by growing transformants, that have the pAEX:Bl+B2 and pAEX:B3+T vectors integrated, in 10% CSL media as a seed culture then transferring 1/10 volume of the seed culture to MBM carbon media + 12% Maltose for induction of the glaA promoter. MBM media consisted of the following components: MBM Carbon Medium: NaH2P04 0.70 g/L K2HP04 0.70 g/L ΚΗ2Ρ04 0.70 g/L MgS04-7H20 1.40g/L ( H4)2S04 10.5 g/L CaCl2-2H20 0.70 g L NH4N03 3.50 g L Sodium Citrate 14 g L FeCl2-4H20 l.O mg/L ZnCl2 5.87 mg L CuCl2-2H20 0.42 mg/L MnCl2-4H20 0.21 mg/L Na2B4O7 10H20. 0.07mg/L Folic Acid 0.174 mg/L P ridoxine-HCl 6.12 mg/L Riboflavin 1.83 mg/L Pantothenic Acid 23.60mg/L Nicotinic Acid 26.66 mg/L Biotin 0.49 mg/L Thiamine HCl 1.39 mg/L Maltose 120.0 g/L Carbenicillin 0.035 mg/L Streptomycin 0.035 mg L Tween 80 0.07% (w/v) Mazu Antifoam 0.14% (v/v) 10% CSL Medium consisted of the following components: 10% CSL Medium: Corn Steep Liquor (50% Solids) 10% (w/v) Ν3Η2ΡΟ4·Η20 1.0 g/L MgS040.50 g/L Maltose 100.0 g/L Glucose 10.0 g/L Mazu Antifoam 0..003% (v/v) mRNA is isolated from transformant cultures (Fast Track 2 Kit, Invitrogen Corp.) and Northern analysis performed with chemiluminescence (Genius System, Boehringer-Mannheim) to detect transcribed message from each gene.

Expression of each of the proteins coded by the dhaB and dhaJ genes is demonstrated by Western analysis, using anti-dhaB and anti-dhaT antibodies, against intracellular protein of transformants.

In-vitro activity of the expressed proteins from intracellular protein extracts of transformants is demonstrated by standard dhaB and dhaT activity assays as described in GENERAL METHODS.

Demonstration of Production of 1.3-Propanediol from Maltose as the Carbon Source in Aspergillus niger: Aspergillus niger strain FS1 was grown in 10% CSL media as a seed culture and transferred as a 1:10 dilution to MBM Carbon media + 12% Maltose. Culture supematent was demonstrated to contain 6 g/L glycerol produced by Aspergillus . Analysis of glycerol was done by HPLC.

Production of 1,3-propanediol from intracellular glycerol is demonstrated by growing transformants that have been confirmed to be expressing the dhaB and dhaT genes under the conditions above. The presence of 1,3-propanediol is demonstrated by HPLC and GC analysis.

EXAMPLE 24 Production of 1.3-propanediol from substrates other than glycerol bv Lactobacillus reuteri ATCC 23272 Lactobacillus reuteri ATCC 23272 was maintained on MRS (Difco, Detroit, MI) plates. A colony from a plate was used to inoculate 70 mL Lactobacillus MRS broth (Difco #0881-17) supplemented with 25 raM NaHC03 in a 250 ml Erlenmeyer flask. The flask was incubated in an anaerobic atmosphere (5-7% H2, 2-8% C02, 85-93% N2) at 32°C.

HPLC analysis of Lactobacillus MRS broth showed a component with the retention time of glycerol. .Lactobacillus MRS broth was treated by alkaline boiling and analyzed for glycerol by HPLC and enzymatic assay. At most, 0.25 g L glycerol could be detected in the initial medium; if all of this glycerol was transformed to 1,3-propanediol, 0.21 g/L propanediol could be said to have been produced from glycerol.

After 10 days of incubation, a sample from the Lactobacillus reuteri culture flask was removed, analyzed by HPLC and GC-MS, and compared to an initial medium sample. Correcting for the glycerol present in the medium, 1.35 g/L 1,3-propanediol was produced by Lactobacillus reuteri from substrates other than glycerol.

SEQUENCE LISTING (1) GENERAL INFORMATION: APPLICANT: LAFFEND, NAGARAJAN, VASANTHA NAKAMURA, CHARLES E. (ii) TITLE OF INVENTION: BIOCONVERSION OF A FERMENTABLE CARBON SOURCE TO 1, 3-PROPANE- DIOL BY A SINGLE MIRCRO- ORGANISM (iii) NUMBER OF SEQUENCES: 44 (iv) CORRESPONDENCE ADDRESS: (A) ADDRESSEE: E.I. DU PONT DE NEMOURS AND COMPANY (B) STREET: 1007 MARKET STREET (C) CITY: WILMINGTON (D) STATE: DELAWARE (E) COUNTRY: U.S.A.

(F) ZIP: 19898 (v) COMPUTER READABLE FORM: (A) MEDIUM TYPE: DISKETTE, 3.50 INCH (B) COMPUTER: IBM PC COMPATIBLE (C) OPERATING SYSTEM: MICROSOFT WINDOWS 3.1 (D) SOFTWARE: MICROSOFT WORD 2.0C CURRENT APPLICATION DATA (A) APPLICATION NUMBER: (B) FILING DATE: (C) CLASSIFICATION: (viii) ATTORNEY/AGENT INFORMATION: (A) NAME: FLOYD, LINDA A.

(B) REGISTRATION NUMBER: 33,692 (C) REFERENCE/DOCKET NUMBER: CR-9715-A TELECOMMUNICATION INFORMATION (A) TELEPHONE: 302-892-8112

Claims (4)

130789/2 -76-

1. A cosmid contained in ATCC 69789 comprising a DNA fragment of about 35 kb isolated from Klebsiella pneumoniae wherein 1) the DNA fragment encodes an active glycerol dehydratase enzyme and 2) digestion of the cosmid results in a restriction digest pattern as shown in FIG. 1, columns 1 and 2.

2. A host bacterium transformed with the cosmid of claim 1.

3. The host bacterium of claim 2 which is deposited with the American Type Culture Collection and having accession number ATCC 69789.

4. A host bacterium comprising the cosmid of claim 1, wherein at least one DNA fragment of said cosmid encodes 1,3-propanediol oxidoreductase, and wherein said host converts a carbon source, other than glycerol or dihydroxyacetone, to 1,3-propanediol. SANFORD t. COLB & Co C: 34773