WO1994010325A1 - Recombinant method and host for manufacture of xylitol - Google Patents

Abstract

Novel methods for the synthesis of xylitol are described.

Description

Recombinant method and host for manufacture of xylitol.

Cross Reference to Related Application

This application is a continuation-in-part application of U. S . Appl . No. 07/973 , 325 , filed November 5 , 1992.

Background of the Invention

Field of the Invention

The present invention relates generally to methods of using genetically modified microorganisms for the manu- facture of useful chemical compounds (metabolic enginee¬ ring) and more specifically to constructing microbial strains by genetic manipulation that are capable of conver¬ ting readily available carbon sources, such as D-glucose, into a more valuable product, for example, xylitol.

2. Related Art

Xylitol is a chemical compound of a considerable value as a special sweetener. It is approximately as sweet as sucrose, non-toxic, and non-cariogenic.

Currently, xylitol is produced by chemical hydro- genation of D-xylose. D-xylose is obtained from hydro- lysates of various plant materials where it is always pre¬ sent in a mixture with other pentoses and hexoses. Purification of xylose and also xylitol presents therefore a significant problem. A number of processes of this type are known. U.S. patents 3,784,408, 4,066,711, 4,075,406, and 4,008,285 can be mentioned as examples.

The reduction of D-xylose into xylitol can also be achieved in a microbiological process using either strains isolated from nature (Barbosa, .F.S. et al . , J. Industrial Microbiol . 3:241-251 (1988)) or genetically engineered strains (Hallborn, J. et al . , Biotechnology 9:1090-1095 (1991)). However, obtaining the substrate, D-xylose, in a form suitable for yeast fermentation is also a considerable problem because inexpensive xylose sources such as sulphite liquor from pulp and paper processes contain impurities which inhibit yeast growth.

An attractive alternative method for the manufacture of xylitol would be obtaining it by fermentation of a cheap and readily available substrate, such as D-glucose. However, no microorganisms are known that produce xylitol in significant amounts during one-step fermentation of any common carbon sources other than D-xylose and D-xylulose, both of which are structurally very closely related to xylitol.

On the other hand, many microorganisms, especially osmophilic yeasts, for example Zygosaccharomyces rouxii,

Candida polymorpha , and Torulopsis Candida , produce significant amounts of a closely related pentitol, D- arabitol, from D-glucose (Lewis D.H. & Smith D.C., New

Phytol . 66:143-184 (1967)). Using this property of osmophilic yeasts, H. Onishi and T. Suzuki developed a method for converting D-glucose into xylitol by three consecutive fermentations (Appl . Microbiol . 18:1031-1035

(1969)). In this process, D-glucose was first converted into D-arabitol by fermentation with an osmophilic yeast strain. Second, the D-arabitol was oxidized into D- xylulose in a fermentation with Acetobacter suboxydans. Finally, the D-xylulose was reduced to xylitol in the third fermentation using one of many yeast strains capable of reducing D-xylulose into xylitol.

An obvious disadvantage of this method is that it involves three different fermentation steps, each taking from 2 to 5 days; further additional steps like sterilization and cell removal are also needed, thus increasing processing costs. The yield of the step fermentation process is low and the amount of by-products is high. Thus, a need still exists for methods for the economical production of xylitol in microbial systems from readily available substrates.

Summary of the Invention

The present invention provides methods for constructing recombinant hosts, and the recombinant hosts constructed thereby, such hosts being capable of producing xylitol when grown on carbon sources other than D-xylulose or D-xylose, and other than polymers or oligomers or mixtures thereof. The carbon sources used by the hosts of the invention are inexpensive and readily available. The microorganisms of the invention are also capable of secreting the synthesized xylitol into the culture medium. This goal is achieved through modification of the metabolism of a desired microorganism, preferably a naturally occurring yeast microorganism, by introducing and expressing desired heterologous genes. This goal is also achieved by further modification of the metabolism of such desired microorganism, so as to overexpress and/or inactivate the activity or expression of certain genes homologous to such microorganism in its native state.

Therefore, it is an object of the invention to provide a method for the production of xylitol, such method utilizing a new and novel microbe strain, a recombinant host, also herein termed a genetically engineered microorganism, as the producer of such xylitol, such genetically engineered microorganism producing such xylitol either de novo or in enhanced amounts when compared the native unengineered microorganism. It is a further object of the invention to provide a method for the production of xylitol, such method utilizing a novel metabolic pathway that has been engineered into a microorganism and which results in the de novo or enhanced production of xylitol by such microorganism.

It is a further object of the invention to provide a method for the production of xylitol, such method utilizing a novel metabolic pathway as above, and such pathway modifying the pathway of D-arabitol biosynthesis and/or metabolism, such pathway being modified so that the microorganism now produces xylitol from fermentation of carbon sources that the unmodified host utilizes for D- arabitol biosynthesis. It is a further object of the invention to provide a method for the production of xylitol, such method utilizing the altered D-arabitol pathway above, and such pathway being altered either by the extension of the preexisting pathway for D-arabitol biosynthesis (with additional steps for D-arabitol utilization) or by the substitution of one or more steps of the D-arabitol pathway with similar steps leading to the formation of xylitol.

It is a further object of the invention to provide a method for the production of xylitol, such method utilizing the altered D-arabitol biosynthesis pathway above, and such pathway being altered by extending the pre¬ existing D-arabitol pathway by the introduction and overexpression of the genes coding for D-xylulose-forming D-arabitol dehydrogenase (EC 1.1.1.11) and xylitol dehydrogenase (EC 1.1.1.9) into an D-arabitol-producing microorganism.

It is a further object of the invention to provide a method for the production of xylitol using a novel microorganism as above, such method utilizing the altered D-arabitol biosynthesis pathway above, and such pathway being altered further, by inactivating, using chemically induced mutagenesis or gene disruption, the gene coding for transketolase (EC 2.2.1.1) or the gene coding for D- xylulokinase (EC 2.7.1.17) in such microorganism. It is a further object of the invention to provide a method for the production of xylitol using a novel microorganism as above, such method utilizing a genetically-engineered altered overexpression of the genes coding for the enzymes of the oxidative branch of the pentose-phosphate pathway, and specifically D-glucose-6- phosphate dehydrogenase (EC 1.1.1.49) and/or 6-phospho-D- gluconate dehydrogenase (EC 1.1.1.44) in such microorganism.

It is a further object of the invention to provide a method for the production of xylitol using a novel microrganism as above, such method utilizing a genetically- engineered altered overexpression of the genes coding for the enzymes of the oxidative branch of the pentose- phosphate pathway, as well as the D-ribulose-5-phosphate epimerase gene (EC 5.1.3.1).

Brief Description of the Drawings

Figure 1 is a restriction map of the insert in plasmid pARL2. This insert is that of the Klebsiella terrigena Phpl chromosomal locus and contains the K . terrigena D-arabitol dehydrogenase gene. The open box represents K. terrigena chromosomal DNA. The arrow shows the location and direction of the D-arabitol dehydrogenase (EC l.l.l.ll) gene in this DNA.

Figure 2 shows the construction of pYARD from pADH and pAAH5. On the plasmid diagrams, the single line (-) indicates bacterial sequences; the wavy line indicates S . cerevisiae 2μm DNA; the open arrow (=*) indicates the ADCJ promoter (the ADCI gene codes for S . cerevisiae alcohol dehydrogenase or ADCI, formerly called ADHI) ; the open diamond (O) indicates the ADCI transcriptional terminator; the rectangular block indicates the LEU2 gene; and the hatched arrow indicates the D-arabitol dehydrogenase gene. Figure 3 shows the construction of plasmid pJDB(AX)-

16. XYL2 is the xylitol dehydrogenase gene from Pichia stipitis . dalD is the D-arabitol dehydrogenase gene. ADCI is the transcriptional regulation area (promoter) of the

ADCI gene that precedes and is operably linked to the dalD coding sequence. The symbols are not the same as in Figure

4. On the plasmid diagrams, the single line (-) indicates bacterial sequences and 2μm DNA where noted; the closed arrow indicates the ADCI promoter; the shaded diamond (♦) indicates the ADCI transcriptional terminator; the rectangular block indicates the LEU2 marker gene; the hatched arrow indicates the XYL2 gene; and the blocked rectangle indicates the dalD gene.

Figure 4 shows the construction of the E . coli-Z. rouxii shuttle vector pSRT(AX)-9. The symbols are as in Figure 3.

Figure 5 shows the restriction map of the cloned T. Candida rDNA fragment.

Figure 6 shows the construction of the plasmid pTC(AX) . Figure 7 shows the restriction map of the cloned

T. Candida rDNA fragment.

Figure 7a shows the construction of the plasmid pCPU(AX) .

Figure 8 shows the cloning of the ZWF1 and gnd gene. Figure 8a shows the construction of the PAAH(gnd) plasmid.

Figure 9 shows the construction of plasmid pSRT(ZG) . Figure 10 shows the cultivation of the strain Z . rouxii ATCC 13356 [pSRT(AX)-09] in a fermentor. Figure 11 shows the cultivation of the mutant derived from strain _.. rouxii ATCC 13356 [pSRT(AX)-9] in a fermentor.

Detailed Description of the Preferred Embodiments

I. Definitions

In the description that follows, a number of terms used in recombinant DNA technology are extensively utilized. In order to provide a clear and consistent understanding of the specification and claims, including the scope to be given such terms, the following definitions are provided. Carbon source other than xylose or xylulose. As used herein, by a "carbon source other than D-xylose and D- xylulose" is meant a carbon substrate for xylitol production other than D-xylose and D-xylulose or polymers or oligomers or mixtures thereof (such as xylan and hemicellulose) . The carbon source preferably supports growth of the generically engineered microbial hosts of the invention, and fermentation in yeast hosts. Many cheap and readily available compounds can be used as carbon sources for the production of xylitol in the microbial hosts of the present invention, including D-glucose, and various D- glucose-containing syrups and mixtures of D-glucose with other sugars. Other sugars assimilable by the hosts of the invention, including yeast and fungi, such as various aldo- and ketohexoses (for example, D-fructose, D-galactose, and D-mannose) , and oligomers and polymers thereof (for example, sucrose, lactose, starch, inulin and maltose) are intended to be included in this term. Pentoses other than xylose and xylulose and non-carbohydrate carbon sources such as glycerol, ethanol, various plant oils or hydrocarbons (preferably n-alkanes containing 14-16 carbon atoms) are also intended to be included in this term. The spectrum of carbon sources useful as substrates for the production of xylitol by the hosts of the present invention will vary depending on the microbial host. For example, glucose and glucose-containing syrups are the preferred carbon source for xylitol production with the genetically manipulated Zygosaccharomyces rouxii of the invention, while n-alkanes, preferably having 14-16 carbon atoms, are the preferred carbon source for modified Candida tropicalis strains.

Gene. A DNA sequence containing a template for a RNA polymerase. The RNA transcribed from a gene may or may not code for a protein. RNA that codes for a protein is termed messenger RNA (mRNA) and, in eukaryotes, is transcribed by RNA polymerase II. A gene containing a RNA polymerase II template (as a result of a RNA polymerase II promoter) wherein a RNA sequence is transcribed which has a sequence complementary to that of a specific mRNA, but is not normally translated can also be constructed. Such a gene construct is herein termed an "antisense RNA gene" and such a RNA transcript is termed an "antisense RNA." Antisense RNAs are not normally translatable due to the presence of translational stop codons in the antisense RNA sequence. A "complementary DNA" or "cDNA" gene includes recombinant genes synthesized by, for example, reverse transcription of mRNA, thus lacking intervening sequences (introns) . Genes clones from genomic DNA will generally contain introns.

Cloning vehicle. A plasmid or phage DNA or other DNA sequence which is able to carry genetic information, specifically DNA, into a host cell. A cloning vehicle is often characterized by one or a small number of endonuclease recognition sites at which such DNA sequences can be cut in a determinable fashion without loss of an essential biological function of the vehicle, and into which a desired DNA can be spliced in order to bring about its cloning into the host cell. The cloning vehicle can further contain a marker suitable for use in the identification of cells transformed with the cloning vehicle, and origins of replication that allow for the maintenance and replication of the vehicle in one or more prokaryotic or eukaryotic hosts. Markers, for example, are tetracycline resistance or ampicillin resistance. The word

"vector" is sometimes used for "cloning vehicle." A "plasmid" is a cloning vehicle, generally circular DNA, that is maintained and replicates autonomously in at least one host cell.

Expression vehicle. A vehicle or vector similar to a cloning vehicle but which supports expression of a gene that has been cloned into it, after transformation into a host. The cloned gene is usually placed under the control of (i.e., operably linked to) certain control sequences such as promoter sequences, that can be provided by the vehicle or by the recombinant construction of the cloned gene. Expression control sequences will vary depending on whether the vector is designed to express the operably linked gene in a prokaryotic or eukaryotic host and can additionally contain transcriptional elements such as enhancer elements (upstream activation sequences) and termination sequences, and/or translational initiation and termination sites.

Host. A host is a cell, prokaryotic or eukaryotic, that is utilized as the recipient and carrier of recombinant material. Host of the Invention. The "host of the invention" is a microbial host that does not naturally produce xylitol in significant amounts during fermentation from common carbon sources other than D-xylose or D-xylulose, or polymers or oligomers or mixtures thereof, but has been engineering to do so according to the methods of the invention. By a "significant amount" is meant an amount which is suitable for isolation of xylitol in pure form or an amount that can be reliably measured by the analytical methods normally used for the analysis of carbohydrates in the microbial fermentation broth.

Arabitol Dehvdroσenase. There are two types of D- arabitol dehydrogenases: D-xylulose-forming (EC 1.1.11) and D-ribulose-forming. D-ribulose-forming dehydrogenases are found in wild type yeasts and fungi. D-xylulose-forming arabitol dehydrogenases are known only in bacteria. Unless otherwise stated, it is the D-xylulose-forming arabitol dehydrogenase that is intended herein and referred to herein as arabitol dehydrogenase.

Oxidative Branch of the Pentose-Phosphate Pathway. By the "oxidative branch of the pentose-phosphate pathway" is meant to include that part of the pentose-phosphate shunt that catalyzes oxidative reactions, such as those reactions catalyzed by D-glucose-6-phosphate dehydrogenase

(EC 1.1.1.49) and 6-phospho-D-gluconate dehydrogenase. (EC 1.1.1.44), and that utilizes hexose substrates to form pentose phosphates. The "non-oxidative" part of the pentose-phosphate pathway (which also catalyzes the net formation of riboεe from D-glucose) is characterized by non-oxidative isomerizations such as the reactions catalyzed by ribose-5-phosphate isomerase, D-ribulose-5- phosphate-3-epimerase and transaldolase. See Biological

Chemistry, H.R. Mahler & E.H.Cordes, Harper & Row, publishers, New York, 1966, pp. 448-454.

Functional Derivative. A "functional derivative" of a protein or nucleic acid, is a molecule that has been chemically or biochemically derived from (obtained from) such protein or nucleic acid and which retains a biological activity (either functional or structural) that is a characteristic of the native protein or nucleic acid. The term "functional derivative" is intended to include "frag- ments," "variants," "analogues," or "chemical derivatives" of a molecule that retain a desired activity of the native molecule.

As used herein, a molecule is said to be a "chemical derivative" of another molecule when it contains additional chemical moieties not normally a part of the molecule. Such moieties can improve the molecule's solubility, absorption, biological half life, etc. The moieties can decrease the toxicity of the molecule, or eliminate or attenuate any undesirable side effect of the molecule, etc. Moieties capable of mediating such effects are disclosed in Reming¬ ton 's Pharmaceutical Sciences (1980). Procedures for coupling such moieties to a molecule are well known in the art. Fragment. A "fragment" of a molecule such as a protein or nucleic acid is meant to refer to a portion of the native amino acid or nucleotide genetic sequence, and in particular the functional derivatives of the invention. Variant or Analog. A "variant" or "analog" of a protein or nucleic acid is meant to refer to a molecule substantially similar in structure and biological activity to either the native molecule, such as that encoded by a functional allele.

II. Construction of Metabolic Pathways for Xylitol Biosynthesis

According to the invention, the native metabolic pathways of a microbial host are manipulated so as to decrease or eliminate the utilization of carbon into purposes other than xylitol production. All of the hosts of the invention produce xylitol in one fermentation step. In one embodiment, a hosts of the invention can possess xylitol dehydrogenase (EC 1.1.1.9) activity sufficient for xylitol production. However, as described below, in those hosts wherein it is desired to overproduce xylitol dehydrogenase activity, recombinant genes encoding xylitol dehydrogenase can be transformed into the host cell.

In the practical realization of the invention, all of the hosts of the invention are characterized by the ability to synthesize xylitol from structurally unrelated carbon sources such as D-glucose and not just from D-xylose and/or D-xylulose. The hosts of the invention are also capable of secreting the synthesized xylitol into the medium.

Specifically, in the exemplified and preferred embodiments, the hosts of the invention are characterized by one of two pathways. First, a pathway in which arabitol is an intermediate in xylitol formation and second, a pathway in which xylulose-5-phosphate is directed into xylitol formation through dephosphorylation and reduction reactions. Accordingly, the hosts of the invention are characterized by at least one of the following genetic alterations: (l) a gene encoding a protein possessing D-xylulose- forming D-arabitol dehydrogenase activity (EC 1.1.1.11) has been cloned into the host—thus providing for the conversion of D-arabitol to D-xylulose (characteristic of pathway I) ; and/or (2) the native host gene encoding transketolase activity has been inactivated (characteristic of pathway II).

In addition a variety of further modifications to the hosts can be performed, so as to enhance the xylitol producing capabilities of such hosts. For example, the hosts as described in (1) and (2) can be further modified such that:

(3) a gene encoding a protein possessing xylitol dehydrogenase (EC 1.1.1.9) activity has been cloned into the host; (3) the native host gene encoding D-xylulokinase (EC 2.7.1.17) has been inactivated;

(4) a gene encoding a protein possessing D-glucose- 6-phosphate dehydrogenase (EC 1.1.1.49) activity has been cloned into the host;

(4) a gene encoding a protein possessing 6-phospho- D-gluconate dehydrogenase (EC 1.1.1.44) activity has been cloned into the host;

(5) a gene encoding a protein possessing D-ribulose- 5-phosphate-3-epimerase (EC 5.1.3.1) activity has been cloned into the host;

In a preferred embodiment, the hosts of the invention possess more than one of the above-described genetic alterations. For example, in a preferred embodiment, carbon flows from D-arabitol "directly to"

(that is, in one step) D-xylulose, and from D-xylulose

"directly to" xylitol. Accordingly, in such embodiment, the host of the invention has been altered such that a gene encoding a protein possessing D-xylulose-forming D-arabitol dehydrogenase activity and a gene encoding a xylitol dehydrogenase (EC 1.1.1.9) have been cloned into the host.

It should be noted that while, in many embodiments, D- arabitol is internally synthesized from other carbon sources by the hosts of the invention, D-arabitol could also be externally added directly to the medium.

In another preferred embodiment, the xylitol biosynthesis pathway does not incorporate arabitol as an intermediate. Rather, the carbon flow is from D-xylulose-5- phosphate to D-xylulose further to xylitol. When D-glucose is used as the carbon source, the flow of carbon would be through the oxidative portion of the pentose phosphate pathway, from D-glucose to D-glucose-6-phosphate to 6- phospho-D-gluconate to D-ribulose-5-phosphate. The D- ribulose-5-phosphate would further epimerized to D- xylulose-5-phosphate, dephosphorylated to D-xylulose and reduced to xylitol. Accordingly, a host of the invention for utilization of this embodiment would include a host in which:

(al) a gene encoding a protein possessing D-glucoεe-6- phosphate dehydrogenase (EC 1.1.1.49) activity has been cloned into the host or the native gene of the host is overexpressed; and/or (a2) a gene encoding a protein possessing 6-phospho-D- gluconate dehydrogenase (EC 1.1.1.44) activity has been cloned into the host or the native gene of the host is overexpressed; and/or (a3) a gene encoding a protein possessing D-ribulose-5- phosphate-3-epimerase activity has been cloned into the host or the native gene of the host is overexpressed; and/or

(a4) a gene encoding a protein possessing xylitol dehydrogenase (EC 1.1.1.9) activity has been cloned into the host or the native gene of the host is overexpressed; (b) the native tranεketolase gene has been inactivated; and/or (c) the native host gene encoding xylulokinase (EC

2.7.1.17) activity has been inactivated.

The dephosphorylation step (D-xylulose-phosphate to D- xylulose conversion) is the only step catalyzed by an enzyme that has not been characterized in pure form.

However, the enzyme activity responsible for the similar step (D-ribulose-5-phosphate to D-ribulose) in the native

D-arabitol-forming pathway of osmophilic yeast was previously shown to be non-specific and capable also of catalyzing the dephosphorylation of xylulose-5-phosphate

(Ingram, J.M. and W.A. Wood, J . Bacteriol . 89:1186-1194

(1965)). The mutation of transketolase and overexpression of the two dehydrogenases of the oxidative pentose phosphate pathway serve a dual purpose. First, they can increase the efficiency of pathway I by increasing the amount of ribulose-5-phosphate in the cell and consequently the production of arabitol and xylitol. Secondly, the over- accumulation of xylulose-5-phosphate which is necessary for the operation of pathway II should also result from the same combination of modifications.

Therefore, methods utilizing the naturally occurring pathway leading to the formation of D-arabitol from various carbon sources and extending this pathway by two more reactions to convert D-arabitol into xylitol are not the only possible pathway within the invention. Other pathways leading to xylitol as a final metabolic product and not involving D-arabitol as an intermediate can be constructed. Thus, a pathway to xylitol from D-ribulose-5-phosphate, can be realized through more than one chain of reactions. D- ribulose-5-phosphate can efficiently be converted to D- xylulose-5-phosphate by D-ribulose-5-phosphate-3-epimerase and if further conversion of D-xylulose-5-phosphate is prevented by a mutation in the transketolase gene, the accumulated D-xylulose-5-phosphate can be dephosphorylated by the same non-specific phosphatase as D-ribulose-5- phosphate (Ingram, J.M. et al . , J. Bacteriol . 89:1186-1194 (1965)) and reduced into xylitol by xylitol dehydrogenase. Realization of this pathway can further require the inactivation of D-xylulokinase gene in order to minimize the energy loss due to the futile loop: D-xylulose-5- phosphate → D-xylulose → D-xylulose-5-phosphate. An additional genetic change - introduction and (over)- expression of the D-ribulokinase gene (EC 2.7.1.47) could minimize simultaneous D-arabitol production by such strains by trapping the D-ribulose produced by the unspecific phosphatase. The D-ribulose will be converted back into the D-ribulose-5-phosphate and further into D-xylulose-5- phosphate. III. Construction of the Hosts of the Invention

The process for genetically engineering the hosts of the invention, according to the invention, is facilitated through the isolation and partial sequencing of pure protein encoding an enzyme of interest or by the cloning of genetic sequences which are capable of encoding such protein with polymerase chain reaction technologies; and through the expression of such genetic sequences. As used herein, the term "genetic sequences" is intended to refer to a nucleic acid molecule (preferably DNA) . Genetic sequences which are capable of encoding a protein are derived from a variety of sources. These sources include genomic DNA, cDNA, synthetic DNA, and combinations thereof. The preferred source of genomic DNA is a yeast genomic library. The preferred source of the cDNA is a cDNA library prepared from yeast mRNA grown in conditions known to induce expression of the desired mRNA or protein.

The cDNA of the invention will not include naturally occurring introns if the cDNA was made using mature mRNA as a template. The genomic DNA of the invention may or may not include naturally occurring introns. Moreover, such genomic DNA can be obtained in association with the 5' promoter region of the gene sequences and/or with the 3' tran- scriptional termination region. Further, such genomic DNA can be obtained in association with the genetic sequences which encode the 5' non-translated region of the mRNA and/or with the genetic sequences which encode the 3' non- translated region. To the extent that a host cell can recognize the transcriptional and/or translational regulatory signals associated with the expression of the mRNA and protein, then the 5' and/or 3' non-transcribed regions of the native gene, and/or, the 5' and/or 3' non- translated regions of the mRNA, can be retained and employed for transcriptional and translational regulation. Genomic DNA can be extracted and purified from any host cell, especially a fungal host, which naturally expresses the desired protein by means well known in the art (for example, see Guide to Molecular Cloning Techniques , S.L. Berger et al . ,. eds., Academic Press (1987)). Preferably, the mRNA preparation used will be enriched in mRNA coding for the desired protein, either naturally, by isolation from cells which are producing large amounts of the protein, or in vitro , by techniques commonly used to enrich mRNA preparations for specific sequences, such as sucrose gradient centrifugation, or both.

For cloning into a vector, such suitable DNA preparations (either genomic DNA or cDNA) are randomly sheared or enzymatically cleaved, respectively, and ligated into appropriate vectors to form a recombinant gene (either genomic or cDNA) library.

A DNA sequence encoding a desired protein or its functional derivatives can be inserted into a DNA vector in accordance with conventional techniques, including blunt- ending or staggered-ending termini for ligation, restriction enzyme digestion to provide appropriate termini, filling in of cohesive ends as appropriate, alkaline phosphatase treatment to avoid undesirable joining, and ligation with appropriate ligases. Techniques for such manipulations are disclosed by Maniatis, T. , (Maniatis, T. et al ., Molecular Cloning (A Laboratory Manual ) , Cold Spring Harbor Laboratory, second edition, 1988) and are well known in the art.

Libraries containing sequences coding for the desired gene can be screened and the desired gene sequence identified by any means which specifically selects for a sequence coding for such gene or protein such as, for example, a) by hybridization with an appropriate nucleic acid probe(s) containing a sequence specific for the DNA of this protein, or b) by hybridization-selected translational analysis in which native mRNA which hybridizes to the clone in question is translated in vitro and the translation products are further characterized, or, c) if the cloned genetic sequences are themselves capable of expressing mRNA, by i munoprecipitation of a translated protein product produced by the host containing the clone.

Oligonucleotide probes specific for a certain protein which can be used to identify clones to this protein can be designed from the knowledge of the amino acid sequence of the protein or from the knowledge of the nucleic acid sequence of the DNA encoding such protein or a related protein. Alternatively, antibodies can be raised against purified forms of the protein and used to identify the presence of unique protein determinants in tranεformants that express the desired cloned protein. The sequence of amino acid residues in a peptide is designated herein either through the use of their commonly employed three-letter designationε or by their εingle-letter deεignationε. A liεting of theεe three-letter and one- letter designations can be found in textbooks such as Biochemistry, Lehninger, A., Worth Publisherε, New York, NY (1970) . When the amino acid sequence is liεted horizontally, unless otherwise stated, the amino terminus is intended to be on the left end and the carboxy terminus iε intended to be at the right end. Similarly, unleεε otherwise stated or apparent from the context, a nucleic acid sequence is presented with the 5' end on the left.

Because the genetic code is degenerate, more than one codon can be used to encode a particular amino acid (Watεon, J.D. , In: Molecular Biology of the Gene , 3rd Ed., W.A. Benjamin, Inc., Menlo Park, CA (1977), pp. 356-357). The peptide fragmentε are analyzed to identify sequences of amino acids which can be encoded by oligonucleotides having the lowest degree of degeneracy. This iε preferably accomplished by identifying sequences that contain amino acidε which are encoded by only a single codon.

Although occaεionally an amino acid εequence can be encoded by only a εingle oligonucleotide sequence, frequently the amino acid sequence can be encoded by any of a set of εimilar oligonucleotideε. Importantly,- whereaε all of the memberε of this set contain oligonucleotide sequenceε which are capable of encoding the εame peptide fragment and, thuε, potentially contain the εame oligonucleotide εequence as the gene which encodes the peptide fragment, only one member of the set contains the nucleotide sequence that is identical to the exon coding sequence of the gene. Because this member is preεent within the set, and is capable of hybridizing to DNA even in the presence of the other members of the set, it is possible to employ the unfractionated set of oligonucleotides in the same manner in which one would employ a single oligo¬ nucleotide to clone the gene that encodes the peptide.

Using the genetic code, one or more different oligonucleotides can be identified from the amino acid sequence, each of which would be capable of encoding the desired protein. The probability that a particular oligo¬ nucleotide will, in fact, constitute the actual protein encoding sequence can be estimated by considering abnormal base pairing relationships and the frequency with which a particular codon iε actually used (to encode a particular amino acid) in eukaryotic cells. Using "codon usage rules," a single oligonucleotide sequence, or a set of oligonucleo¬ tide sequences, that contain a theoretical "most probable" nucleotide εequence capable of encoding the protein sequences iε identified.

The suitable oligonucleotide, or set of oligonucleotides, which is capable of encoding a fragment of a certain gene (or which iε complementary to such an oligonucleotide, or set of oligonucleotides) can be εyn- thesized by means well known in the art (see, for example, Synthesis and Application of DNA and RNA, S.A. Narang, ed. , 1987, Academic Presε, San Diego, CA) and employed aε a probe to identify and iεolate a clone to εuch gene by techniqueε known in the art. Techniques of nucleic acid hybridization and clone identification are discloεed by Maniatiε, T. , et al . , in: Molecular Cloning, A Laboratory Manual , Cold Spring Harbor Laboratorieε, Cold Spring Harbor, NY (1982)), and by Hames, B.D. , et al . , in: Nucleic Acid Hybridization, A Practical Approach , IRL Presε, Waεhington, DC (1985)). Thoεe memberε of the above- deεcribed gene library which are found to be capable of εuch hybridization are then analyzed to determine the extent and nature of coding εequenceε which they contain. To facilitate the detection of a deεired DNA coding εequence, the above-deεcribed DNA probe is labeled with a detectable group. Such detectable group can be any material having a detectable physical or chemical property. Such materials have been well-developed in the field of nucleic acid hybridization and in general most any label useful in such methods can be applied to the preεent invention. Particularly uεeful are radioactive labels, such as ³²P, ³H, ¹⁴C, ³³S, ¹²⁵I, or the like. Any radioactive label can be employed which provides for an adequate signal and has a sufficient half-life. If single stranded, the oligonuc¬ leotide can be radioactively labelled using kinase reactions. Alternatively, polynucleotides are also useful as nucleic acid hybridization probes when labeled with a non-radioactive marker such as biotin, an enzyme or a fluo- rescent group.

Thuε, in summary, the elucidation of a partial protein sequence, permits the identification of a theo¬ retical "most probable" DNA sequence, or a set of such sequenceε, capable of encoding εuch a peptide. By construc- ting an oligonucleotide complementary to this theoretical sequence (or by constructing a set of oligonucleotides com¬ plementary to the set of "most probable" oligonucleotides) , one obtains a DNA molecule (or set of DNA molecules) , capable of functioning as a probe(ε) for the identification and iεolation of cloneε containing a gene.

In an alternative way of cloning a gene, a library iε prepared uεing an expreεεion vector, by cloning DNA or, more preferably cDNA prepared from a cell capable of expreεεing the protein into an expreεεion vector. The library iε then εcreened for memberε which expreεε the deεired protein, for example, by screening the library with antibodies to the protein.

The above discuεεed ethodε are, therefore, capable of identifying genetic sequences which are capable of encoding a protein or biologically active or antigenic fragmentε of this protein. In order to further characterize such genetic sequenceε, and, in order to produce the recombinant protein, it iε deεirable to express the proteins which these sequenceε encode. Such expreεεion identifieε thoεe cloneε which expreεε proteinε possesεing characteriεticε of the deεired protein. Such characteris¬ tics can include the ability to specifically bind antibody, the ability to elicit the production of antibody which are capable of binding to the native, non-recombinant protein, the ability to provide a enzymatic activity to a cell that is a property of the protein, and the ability to provide a non-enzymatic (but specific) function to a recipient cell, among others.

A DNA sequence can be shortened by means known in the art to isolate a desired gene from a chromosomal region that contains more information than necessary for the utilization of this gene in the hoεtε of the invention. For example, reεtriction digeεtion can be utilized to cleave the full-length sequence at a desired location. Alternatively, or in addition, nucleases that cleave from the 3'-end of a DNA molecule can be used to digest a certain sequence to a shortened form, the desired length then being identified and purified by gel electrophoresis and DNA εequencing. Such nucleases include, for example, Exonucleaεe III and Bal31. Other nucleaεes are well known in the art.

In the practical realization of the invention the oεmophilic yeast Z . rouxii has been employed aε a model. Z . rouxii iε compatible with food production since it is traditionally used in Japan for the manufacture of εoy sauce. The yeast has been deεcribed for instance in: The

Yeasts, A Taxonomic Study, Kreger-van Rij (ed.), Elsevier

Science publishers B.V., Amsterdam 3984, wherein this yeast is described on pages 462-465. Other D-arabitol-producing yeastε like Candida polymorpha , Torulopsiε Candida , Candida tropicalis , Pichia farinoεa, Torulaspora hansenii, etc., aε well aε D-arabitol producing fungi like Dendryphiella salina or Schizophyllum commune can alεo be uεed aε host organiεmε for the purpoεeε of the preεent invention.

The enzymes oxidizing D-arabitol into D-xylulose (EC

1.1.1.11) are known to occur in bacteria but not in yeast or fungi. For the purposes of the present invention

Klebsiella terrigena is the preferred source of the D- arabitol dehydrogenase (D-xylulose forming) gene since it iε a nonpathogenic soil bacterium and it has a high inducible D-arabitol dehydrogenase activity. The Klebsiella terrigena strain Phpl used in the examples was obtained from K. Haahtela, Helsinki University. The isolation of the strain is described in Haahtela et al . , Appl . Env .

Microbiol . 45: 563-510 (1983)). The cloning of the D- arabitol dehydrogenaεe gene can be conveniently achieved by constructing a genetic library of the K. terrigena chromosomal DNA in a suitable vector, for instance well known, and commercially available, plasmid pUC19. This library is transformed into one of many E. coli strains which are able to utilize D-xylulose but not D-arabitol as a sole carbon source. E. coli strain SCSI available from Stratagene is an example of a suitable strain. The tranεformantε are then plated on a medium containing D- arabitol aε a εole carbon εource and the clones able to grow on this medium are isolated. The coding region of the K. terrigena D-arabitol dehydrogenase can be conveniently iεolated in a form of 1.38 kb Bcll-Clal fragment and fuεed with appropriate promoter and tranεcription terminator sequences. The Saccharomyces cerevisiae ADCI promoter and transcription terminator are examples of transcriptional regulatory elements εuitable for the purposes of the present invention when the yeast Z . rouxii is used as a host organism. The sequence of ADCI is available from GenBank.

Although the majority of yeastε and fungi poεεess the xylitol dehydrogenase (EC 1.1.1.9) gene, overexpresεion of the said gene will typically be necesεary for the implementation of the present invention. The cloning of the

Pichia εtipitis XYL2 gene encoding xylitol dehydrogenase

(EC 1.1.1.9) can conveniently be achieved by polymerase chain reaction technology using the published information on the nucleotide sequence of the XYL2 gene (Kδtter et al . , Curr . Genet . 18:493-500 (1990)). The gene can be introduced into other yeast species without any modifications and expressed under control of its own promoter or the promoter can be exchanged for another strong yeaεt promoter.

Genetically εtable transformants can be constructed with vector syεtemε, or transformation syεte s, whereby a desired DNA iε integrated into the hoεt chromoεome. Such integration can occur de novo within the cell or be assiεted by transformation with a vector which functionally inserts itself into the host chromoεome, for example, with phage, retroviral vectors, transposons or other DNA elementε which promote integration of DNA εequences in chromoεomeε.

The geneε coding for D-arabitol dehydrogenase and xylitol dehydrogenase (EC 1.1.1.9) under control of suitable promoterε can be combined in one plaεmid construc¬ tion and introduced into the host cells of an D-arabitol producing organiεm by tranεformation. The nature of the plaεmid vector will depend on the host organism. Thus, for Z . rouxii vectors incorporating the DNA of the pSRl cryptic plasmid (Ushio, K. et al . , J . Ferment . Technol . 66:481-488 (1988)) are used in the preferred embodiment of the preεent invention. For other yeast or fungal specieε for which autonomouεly replicating plaεmidε are unknown, integration of the xylitol dehydrogenase (EC 1.1.1.9) and D-arabitol dehydrogenase genes into the hoεt'ε chromosome can be em¬ ployed. Targeting the integration to the ribosomal DNA (DNA encoding riboεomal RNA) locus of the host is the preferred method of obtaining the high copy-number integration and high level expresεion of the two dehydrogenaεe genes such targeting can be achieved by providing recombinant DNA εequences on the recombinant construct sufficient to direct integration to this locus. The genetic markers used for the transformation of the D-arabitol-producing microorganismε are preferably dominant markerε conferring reεiεtance to variouε antibioticε such as gentamicin or phleomycin or heavy metals, εuch aε copper, or the like. The εelectable marker gene can either be directly linked to the DNA gene sequences to be expresεed, or introduced into the same cell by co-transformation. Besideε introduction of D-arabitol dehydrogenaεe and xylitol dehydrogenaεe (EC 1.1.1.9) genes, other genetic modificationε can be uεed for constructing novel xylitol- producing strains. Thus, the genes coding for the enzymes of the oxidative pentose phosphate pathway can be overexpresεed in order to increaεe the rate of εyntheεiε of D-arabitol precursor D-ribulose-5-phosphate. Alεo, the gene coding for tranεketolaεe—the enzyme catalyzing the catabolism of pentulose-5-phoεphateε or pentoεe-5- phosphates may be inactivated by conventional mutagenesis or gene disruption techniques leading to increased accumulation of five-carbon sugar phosphates. Inactivation of the D-xylulokinase gene can increase xylitol yield by eliminating the losε of D-xyluloεe due to phoεphorylation. A combination of an inactivating tranεketolaεe mutation with the overexpreεsion of D-ribuloεe-5-epimerase can be used for creating a different type of xylitol production pathway in which D-arabitol is not used as an intermediate. To expresε a deεired protein and/or itε active derivativeε, tranεcriptional and tranεlational εignalε recognizable by an appropriate hoεt are neceεεary. The cloned coding εequenceε, obtained through the methodε deεcribed above, and preferably in a double-εtranded form, can be operably linked to εequenceε controlling transcriptional expression in an expresεion vector, and introduced into a hoεt cell, either prokaryote or eukaryote, to produce recombinant protein or a functional derivative thereof. Depending upon which εtrand of the coding εequence iε operably linked to the sequenceε controlling tranεcriptional expression, it iε also posεible to express antiεenεe RNA or a functional derivative thereof.

Expression of the protein in different hosts can result in different post-translational modifications which can alter the properties of the protein. Preferably, the preεent invention encompasses the expresεion of the protein or a functional derivative thereof, in eukaryotic cellε, and especially in yeast.

A nucleic acid molecule, such as DNA, is said to be

"capable of expreεεing" a polypeptide if it containε expression control εequenceε which contain tranεcriptional regulatory information and such sequences are "operably linked" to the nucleotide sequence which encodes the polypeptide.

An operable linkage is a linkage in which a sequence is connected to a regulatory sequence (or sequences) in such a way aε to place expression of the sequence under the influence or control of the regulatory sequence. Two DNA sequences (such as a coding sequence and a promoter region εequence linked to the 5' end of the coding sequence) are said to be operably linked if induction of promoter function results in the transcription of mRNA encoding the desired protein and if the nature of the linkage between the two DNA sequences does not (1) alter the reading frame of a coding sequence, (2) interfere with the ability of the expreεεion regulatory sequences to direct the expression of the protein, antisenεe RNA, or (3) interfere with the ability of the DNA template to be transcribed. Thus, a promoter region would be operably linked to a DNA sequence if the promoter was capable of effecting transcription of that DNA sequence.

The precise nature of the regulatory regions needed for gene expression can vary between species or cell types, but shall in general include, as necessary, 5 ' non- transcribing and 5' non-translating (non-coding) sequences involved with initiation of transcription and translation respectively, such as the TATA box, capping sequence, CAAT sequence, and the like. Especially, such 5' non- transcribing control sequences will include a region which contains a promoter for tranεcriptional control of the operably linked gene. Such tranεcriptional control εequenceε can alεo include enhancer sequenceε or upεtream activator εequences, as desired.

Expression of a protein in eukaryotic hosts such as yeast requires the use of regulatory regions functional in such hosts, and preferably yeast regulatory systems. A wide variety of tranεcriptional and tranεlational regu¬ latory εequenceε can be employed, depending upon the nature of the hoεt. Preferably, theεe regulatory signals are asεociated in their native εtate with a particular gene which iε capable of a high level of expression in the host cell.

In eukaryotes, where tranεcription iε not linked to translation, such control regions may or may not provide an initiator methionine (AUG) codon, depending on whether the cloned εequence containε εuch a methionine. Such regionε will, in general, include a promoter region εufficient to direct the initiation of RNA εyntheεiε in the host cell. Promoters from yeast genes which encode a mRNA product capable of translation are preferred, and especially, strong promoters can be employed provided they also function as promoterε in the host cell. Preferred strong yeast promoters include the GAL1 gene promoter, glycolytic gene promoters such as that for phosphoglycerolkinaεe (PGK) , or the conεtitutive alcohol dehydrogenaεe (ADCI ) promoter (Ammerer, G. Meth . Enzymol . 101C-. 192-201 (1983); Aho, FEBS Lett . 291:45-49 (1991)).

Aε is widely known, translation of eukaryotic mRNA is initiated at the codon which encodes the first methionine. For this reason, it is preferable to ensure that the linkage between a eukaryotic promoter and a DNA sequence which encodes the deεired protein, or a functional derivative thereof, doeε not contain any intervening codonε which are capable of encoding a methionine. The preεence of such codons reεultε either in a formation of a fusion protein (if the AUG codon is in the same reading frame aε the protein-coding DNA εequence) or a frame-εhift mutation

(if the AUG codon iε not in the same reading frame as the protein-coding sequence) .

Tranεcriptional initiation regulatory εignalε can be εelected which allow for repreεεion or activation, εo that expression of the operably linked genes can be modulated. Of interest are regulatory signals which are temperature- sensitive so that by varying the temperature, expression can be represεed or initiated, or are εubject to chemical regulation, e.g., metabolite. Translational signals are not necesεary when it is desired to express antisense RNA sequences.

If desired, the non-transcribed and/or non- translated regions 3' to the sequence coding for a desired protein can be obtained by the above-described cloning methods. The 3 '-non-transcribed region can be retained for its transcriptional termination regulatory sequence elementε; the 3-non-tranεlated region can be retained for itε translational termination regulatory sequence elements, or for those elementε which direct polyadenylation in eukaryotic cells. Where the native expression control sequences signals do not function satisfactorily in a host cell, then εequences functional in the hoεt cell can be εubstituted. The vectors of the invention can further comprise other operably linked regulatory elements such as DNA elements which confer antibiotic resistance, or origins of replication for maintenance of the vector in one or more host cellε. In another preferred embodiment, especially for Z . rouxii , the introduced sequence is incorporated into a plasmid vector capable of autonomous replication in the recipient hoεt. Any of a wide variety of vectors can be employed for this purpose. Factors of importance in selecting a particular plaεmid or viral vector include: the eaεe with which recipient cellε that contain the vector can be recognized and εelected from thoεe recipient cells which do not contain the vector; the number of copies of the vector which are deεired in a particular host; and whether it is desirable to be able to "shuttle" the vector between host cells of different species.

Preferred yeaεt plaεmidε will depend on the hoεt. For Z . rouxii vectors based on the native cryptic plasmidε pSRl (Toh, E. et al . , J. Bacteriol . 151:1380-1390 (1982)), pSBl, pSB2, pSB3 or pSB4 (Toh, E. et al . , J. Gen . Microbiol . 130 : 2521 -253 (1984)) are preferred. Plaεmid pSRT303D (Jearnpipatkul, A., et al . , Mol . Gen . Genet . 206:88-84 (1987)) iε an example of uεeful plasmid vector for Zygosaccharomyces yeast.

Once the vector or DNA sequence containing the con¬ struct(ε) iε prepared for expreεεion, the DNA conεtruct(ε) iε introduced into an appropriate hoεt cell by any of a variety of εuitable means, including transformation. After the introduction of the vector, recipient cells are grown in a selective medium, which εelects for the growth of vector-containing cellε. Expreεεion of the cloned gene εequence(ε) reεultε in the production of the desired protein, or in the production of a fragment of this protein. This expresεion can take place in a continuous manner in the transformed cells, or in a controlled manner, for example, by induction of expression.

To construct the hostε of the invention that have been altered εuch that they can no longer expreεε a certain gene product, εite-directed mutageneεiε can be performed using techniques known in the art, such as gene disruption

(Rothstein, R.S., Meth . Enzymol . 101C-. 202-211 (1983)).

IV. Production of Xylitol

When recombinant arabitol producing yeaεt, preferably osmophilic, are used aε hoεtε of the invention, they can be grown in high osmotic presεure medium, for example medium containing 10-60% D-glucoεe, and preferably 25% D-glucoεe. ("Normal" medium uεually containε only 2-3% glucose.) High osmotic presεure medium induceε D-arabitol formation in wild type εtrainε of oεmophilic yeasts such as Z . rouxii . The culture medium of the recombinant and control (wild type) strainε iε analyzed according to methodε known in the art, at different cultivation timeε, for the preεence of xylitol. In cultivation conditionε not optimized for maximum D-arabitol yield, the experimental strain Z . rouxii ATCC13356[pSRT(AX) -9] produced and secreted into the culture media both xylitol and D- arabitol. Only D-arabitol was detected in the culture medium cf the control strain. The yield of xylitol in the first trials [see Table 4 in example 4] was approximately 7.7 g/1 after 48 hours of cultivation.

Xylitol can be purified from the medium of the hoεtε of the invention according to any technique known in the art. For example, US 5,081,026, incorporated herein by reference, described the chromatographic separation of xylitol from yeaεt cultureε. Thus, from the fermentation step, xylitol can be purified from the culture medium uεing chromatographic steps as described in US 5,081,026, followed by crystallization.

Having now generally described the invention, the same will become better understood by reference to certain specific examples that are included herein for purpoεeε of illustration only and are not intended to be limiting unless otherwise specified.

Examples Example 1 Cloning of the bacterial D-arabitol dehydrogenase gene

Klebsiella terrigena Phpl (obtained from K.

Haahtela, Helεinki Univerεity, see Haahtela et al . , Appl .

Env. Microbiol . 45 : 563-510 (1983)) was grown in 1 liter of LB medium (1% tryptone, 0.5% yeaεt extract, 1% NaCl) overnight at 30° C. Bacterial cellε (approximately 5 g) were collected by centrifugation, waεhed once in TE (10 mM triε-HCl, 1 mM EDTA, pH 7.5) and reεuεpended in TE containing 1% εodium dodecyl sulfate and 200 μg/ml proteinase K. The suspension was incubated at 37°C for 30 min and then extracted once with an equal volume of phenol and two times with chloroform. 3 M sodium acetate (1/10 of volume) and ethanol (3 volumes) were added and the precipitated nucleic acids collected by centrifugation and redissolved in 5 ml of TE. The RNA was removed by centrifugation of the solution through 25 ml of 1 M NaCl overnight at 30,000 rpm in a Beckmann Ti50.2 rotor. The K. terrigena chromosomal DNA obtained by the above method had an average fragment size of more then 50,000 base pairs (bp) . The DNA waε then digeεted by the reεtriction endonucleaεe Sau3A in the supplier's (Boehringer's) buffer at an enzyme:DNA ratio of 5U/mg until the average DNA fragment size was reduced to approximately 5-10 kb (assessed by agarose gel electrophoresiε) . The digest was fractionated by electrophoresis through a 20 x 10 x 0.6 cm 0.6% agarose gel in TBE buffer (0.09M tris-boric acid, 1 mM EDTA, pH 8.3) at 5 V/cm overnight, a well waε cut in the agaroεe εlab at a poεition correεponding to a fragment εize of approximately 5 kb, a piece of dialyεiε membrane waε fixed along the well and electrophoreεiε waε continued until eεεentially all the DNA fragmentε larger then 5 kb were adεorbed on the membrane. The plasmid DNA of pUC19 (purchased from Pharmacia) was digested with the restriction endonuclease BamKL and bacterial alkaline phoεphataεe uεing the εupplier'ε buffer and reaction conditionε. The linear form of pUC19 waε purified by preparative gel electrophoreεiε using the membrane electroelution method described above and ligated with the 5-15 kb fraction of the Klebsiella terrigena chromosomal DNA. The ligation mixture waε used to transform competent cells of E . coli SCSI (purchased from Stratagene) to ampicillin resistance. In this experiment, approximately 10,000 recombinant clones were obtained. The pooled cells from the tranεformation plates were spread onto minimal medium plateε containing D-arabitol (1%) aε the εole carbon source. After two days of incubation at 37°C εeveral colonieε were obtained. Plaεmid DNA waε iεolated from two (faεteεt growing) of theεe cloneε and used to retransform E . coli εtrain HB101, which iε unable to catabolize D- arabitol but able to uεe D-xyluloεe. All the tranεformantε proved to be D-arabitol-utilization positive on McConkey

(obtained from Difco) agar-D-arabitol plateε while all control cloneε (the εame εtrain tranεformed with pUC19) were negative. Reεtriction analyεiε haε εhown that the two iεolated clones contained identical plaεmidε. One of the two iεolated plasmids (named pARL2) waε uεed for further characterization. A reεtriction map of the cloned 9.5 kb (approximately) fragment of K. terrigena DNA in pARL2 bearing the D-arabitol dehydrogenaεe gene is represented in Figure 1.

Example 2

Expresεion of the bacterial D-arabitol dehydrogenase gene in yeast From the original K. terrigena 9.5 kb DNA clone containing the D-arabitol dehydrogenase gene an approximately 1.8 kb Sad-HindiII fragment was εubcloned using conventional recombinant DNA techniques and found to contain the D-arabitol dehydrogenaεe gene. The plaεmid containing this DNA fragment in the pUC19 vector (pADH, where ADH mean D-arabitol dehydrogenase) waε iεolated from E . coli strain JM110, digested with Bell and Clal restriction endonucleases and a 1.38 kb DNA fragment was isolated by preparative agarose gel electrophoresiε. Thiε DNA fragment waε treated with the Klenow fragment of DNA- polymerase I in the presence of all four deoxynucleotide triphosphateε and ligated with a yeast expreεεion vector pAAH5 (Ammerer, Meth . Enzymol . 101:192-203 (1983)) that had been cut with HindHI and treated with Klenow fragment (Figure 2) . The reεulting expreεεion plasmid, pYARD, is a shuttle E . coli-Saccharomyceε cerevisiae vector containing a bacterial (E. coli) origin of replication and ampicillin resistance gene, a yeast (S. cerevisiae) origin of replication from 2 μm DNA, and the yeast (S . cerevisiae) LEU2 gene for selection in yeaεt. The expression cassette includes a yeast alcohol dehydrogenaεe I (ADCI) promoter and (ADCI) transcription terminator flanking and operably linked to the K . terrigena D-arabitol dehydrogenase gene. Saccharomyces cerevisiae εtrain GRF18 (MAT , leu2-3 , 122, his3-ll , 15) and S . cerevisiae strain DBY746 (ATCC 44773; MATα, leu2-3 , 112 his3-Al ura3-52 trpl-289) were both used as the hoεtε for tranεformation with D-arabitol dehydrogenase expression vector pYARD described above, with the same results. The transformation waε performed by the εtandard lithium chloride procedure (Ito et al . , Bacteriol . 153:163-160 (1983)) uεing the LEU2 marker of pYARD for tranεformant selection. The transformants were grown in liquid culture in minimal medium: 0.67% yeast nitrogen base ("Difco") , 2% D-glucose, 100 mg/1 of histidine and tryptophane at 30°C overnight with εhaking. Cellε were collected by centrifugation, εuspended in a minimal volume of 0.1 M potasεium phoεphate buffer pH 6.8, containing ImM NAD⁺ and diεrupted with 0.5 mm glaεε beadε in a Bead beater apparatuε ("Bioεpec productε") for 6 minutes with ice cooling. D-arabitol dehydrogenase activity was measured as described above. An D-arabitol-grown K. terrigena cell extract waε used as a positive control in theεe experimentε and DBY746 transformed with pAAH5 as negative control. The resultε preεented in Table 2 show that D-arabitol dehydrogenaεe gene iεolated from K. terrigena iε expreεεed efficiently in yeast.

Example 3

Construction of yeast vectors for overexpresεion of xylitol dehydrogenaεe and D-arabitol dehydrogenaεegeneε The known nucleotide sequence of a yeast Pichia εtipitis) gene, XYL2 , encoding xylitol dehydrogenase (Kδtter et al . , Curr. Genet . 18:493-500 (1990)) was used to synthesize oligonucleotides for the cloning of this gene by the polymerase chain reaction. The two oligonucleotideε: CGAATTCTAGACCACCCTAAGTCGTCCC (5'-oligonucleotide) [SEQ ID No. :1: ] and TTCAAGAATTCAAGAAACTCACGTGATGC (3'-oligo¬ nucleotide) [SEQ ID No. :2:] were designed to incorporate convenient restriction εiteε XJal and EcoRI at the 5'- and 3'-termini of the PCR product. The 5'-oligonucleotide anneals at position 1-24 of XYL2 and the 3'-nucleotide anneals at position 1531-1560, according to the numbering used in Kδtter et al . , Curr . Genet . 28:493-500 (1990).

Pichia εtipitiε CBS6054 (Centraalbureau voor Schim elcultures, Ooεterεtraat 1, PO Box 273, 3740 AG

Baarn, The Netherlands) was grown overnight in YEPD medium

(1% yeast extract, 2% peptone, 2% D-glucose) , the cells were collected by centrifugation, washed once with 1 M sorbitol εolution containing 1 mM EDTA, pH 7.5, reεuεpended in the εame εolution and digeεted with Lyticaεe (Sigma) . The digeεtion waε controlled by monitoring the optical denεity at 600 nm of a 1:100 dilution of the cell εuεpenεion in 1% SDS. The digestion was terminated when this value dropped to approximately one seventh of the original. The spheroplaεt εuspension waε then waεhed four timeε with 1M εorbitol εolution. The spheroplastε were lyεed in 1% SDS and treated with 200 μg/ml proteinaεe K at 37°C for 30 min. After one phenol and two chloroform extractionε, the nucleic acidε were ethanol precipitated and rediεεolved in a small volume of TE buffer. The integrity of the chromosomal DNA was checked by agarose gel electrophoresis. The average DNA fragment size was higher then 50 kb. PCR was performed using Taq DNA polymerase (Boehringer) in the εupplier's buffer. The thermal cycle was 93°C - 30 sec, 55°C - 30 sec, 72°C - 60 sec. The PCR product was chloroform extracted, ethanol precipitated and digeεted with EcoRI and Xbal under εtandard conditionε. After agaroεe gel purification, the DNA fragment waε cloned into Xbal and -EcoRI cut pUC18 (plaεmid pUC(XYL2)) . Subεequent reεtriction analysis confirmed that the restriction map of the cloned fragment corresponds to the nucleotide sequence of P. εtipitiε XYL2 gene.

Yeast plasmids for overexpressing D-arabitol dehydrogenase and xylitol dehydrogenase were constructed as illustrated by Figure 3 and Figure 4. To change the flanking restriction sites of the D-arabitol dehydrogenase expresεion casεette of plasmid pYARD, the whole casεette waε exciεed by BamHI digeεtion and cloned into the BamHI cleaved pUClδ. The reεulting plaεmid pUC(YARD) was digested with Sail and 2-σoRI and the sole 2.0 kb DNA fragment waε iεolated by preparative gel electrophoreεiε. This fragment was ligated with a 1.6 kb Hinάlll-EcoRI DNA fragment isolated from the plasmid pUC(XYL2) and the 6.6 kb fragment of E . coli yeast shuttle vector pJDB207 (Beggε, J.D. Nature 275:104-109 (1978)) digeεted with tfindlll and Sail. Plaεmid pJDB(AX)-16 waε iεolated after tranεformation of E. coli with the above ligation mixture. Thiε plaεmid is capable of replicating in both E. coli and Saccharomyces cerevisiae . In S. cerevisiae it directε the high level synthesiε of both D-arabitol dehydrogenaεe and xylitol dehydrogenaεe. Plasmid pSRT(AX)-9 was synthesized by ligation of the 4.7 kb Sail fragment from the plasmid pJDB(AX)-9 and the linear form of the plasmid pSRT303D (Jearnpipatkul et al . , Mol . Gen . Genet . 206:88-94 (1987)) obtained by partial hydrolysiε with Sail .

Example 4

Construction of yeast strains secreting xylitol Zygosaccharomyces rouxii ATCC 13356 was transformed with the plasmid pSRT(AX)-9 by a εlight variation of a previouεly deεcribed method (Uεhio, K. et al . , J . Ferment . Technol . 66:481-488 (1988)). Briefly, Z . rouxii cellε were grown overnight in YEPD medium (giving a culture with optical denεity at 600 nm of 3-5) , collected by low-εpeed centrifugation, waεhed twice in 1 M εorbitol, 1 mM EDTA εolution pH 7.5, resuspended in 1/5 of the original culture volume of the same solution containing 1% 2-mercaptoethanol and digested at room temperature with lyticase (Sigma) . The digeεtion waε followed by diluting a εuitable aliquot of the cell εuεpenεion into 1% SDS εolution and meaεuring the optical denεity of the diluted εample at 600 nm. When this value dropped to 1/7 of the original, the digestion was terminated by cooling the suεpenεion on ice and waεhing (by a 10 min, 1000 rpm centrifugation at 0°C) with the εorbitol solution until the ercaptoεthanol smell could no longer be detected. The spheroplasts were washed once with cold 0.3 M calcium chloride solution in 1 M sorbitol and resuεpended in the εame εolution in about 1/4 of original culture volume. 200 μl aliquotε of thiε εuεpenεion and 10-20 μg of plaεmid DNA were mixed and incubated at 0°C for 40 min. 0.8 ml of ice-cold 50 % PEG-6000 εolution containing 0.3 M calcium chloride waε added to the spheroplaεt εuspension and incubation in the cold waε continued for 1 h. The εpheroplaεts were concentrated by centrifugation at 4000 rpm for 10 min in a table-top centrifuge, resuspended in 2 ml of YEPD containing 1 M sorbitol and left for regeneration overnight at room temperature. The regenerated cells were plated onto YEPD plates containing 50-100 μg/ml of gentamicin and incubated at 30°C for 4-6 days. The transformants were grown in liquid YEPD medium, cell extracts prepared aε described for S. cereviεiae (Example 2) , and the activities of D-arabitol dehydrogenase and xylitol dehydrogenase measured. The resultε of theεe meaεurementε are compared with εimilar meaεurementε made in other organisms in Table 3. They show that both genes are expresεed efficiently in Z . rouxii .

The cellε of Z . rouxii ATCC 13356 tranεformed with pSRT(AX)-9 were grown for two days in 50 ml of YEPD containing 50 μg/ml of gentamicin, collected by centrifugation and used to inoculate 100 ml of YEPD containing 25 wt% D-glucose and 50 μg/ml of gentamicin.

This culture waε grown in a 1000 ml flaεk on a rotary εhaker (200 rpm) at 30°C. In another experiment (Experiment

2) the same medium without yeast extract (low phosphate medium) was used. Xylitol content in the culture broth was analyzed by standard HPLC and gas chromatography. The reεultε are preεented in Table 4. No xylitol waε detected in the culture medium of untranεformed Z . rouxii grown in the same media (without gentamicin) . Table 3

Activities of D-arabitol dehydrogenase and xylitol dehydrogenase in different organisms and strains

D-arabitol xylitol

Organism

Plasmid (*) dehydrogenase dehydrogenase and strain (IU/mg protein) (IU/mg protein)

Klebsiella terrigena Phpl

Table 4

Production of xylitol by the Z rouxii ATCC 13356 carrying the plasmid pSRT(AX)-9

(g/1)

Uεing a similar approach, strains producing xylitol from other carbon sources can be constructed. For example, Candida tropicaliε is capable of converting n-alkanes into D-arabitol (Hattori, K. and Suzuki T. , Agric . Biol . Chem . 38:1875-1881 (1974)) in good yield. The 4.7 kb Sail fragment from the plasmid pJDB(AX)-9 can be inserted into the plasmid pCUl (Haas, L. et al ., J. Bacteriol . 172:4571- 4577 (1990) ) and used to transform C. tropicaliε strain SU- 2 (ura3) . Alternatively, the same expreεεion caεεette can be tranεformed into a prototrophic C. tropicaliε εtrain on a plasmid vector bearing a dominant selective marker.

Example 5

Construction of an integrative dominant selection vector for the expresεion of arabitol dehydrogenaεe and xylitol dehydrogenaεe and tranεformation of Torulopεiε Candida

Chro oεomal DNA waε iεolated from T. Candida (ATCC 20214) by a procedure εimilar to the εtandard procedure used for S. cerevisiae . However, preparation of T . Candida spheroplastε required a high concentration of Lyticase (Sigma) - approx. 50,000 U per 10 g of cells and long incubation time - from several hours to overnight incubation at room temperature to achieve efficient cell wall lysis. The buffer for spheroplast preparartion waε ImM

EDTA, pH 8, containing 1 M sorbitol and 1% mercaptoethanol.

The spheroplaεtε were waεhed three times with the same buffer (without mercaptoethanol) and lysed in 15 ml of 1%

SDS. Two phenol extractionε were performed immediately after cell lysis and the DNA was precipitated by addition of 2 volumes of ethanol and centrifugation (5 min at 10,000 rpm) . The DNA waε waεhed twice with 70% ethanol and dried under vacuum. The DNA waε diεεolved in 5 ml of 10 mM tris-

HC1 buffer containing 1 M EDTA, RNAεe A waε added to 10 μg/ml concentration and the εolution was incubated for 1 h at 37°C. The integrity of the DNA was confirmed by agarose gel electrophoresiε using uncut lambda DNA as a molecular weight reference.

200 μg of the chromoεomal DNA of T. Candida was cut with Hindlll and EcoRI , the digest was applied into a 6 cm wide well on an 0.7% (8x15x0.8 cm) preparative agarose gel and separated by electrophoresiε. A 1 cm wide strip of the gel was cut out of the gel and blotted onto a positively charged nylon membrane (Boehringer 1209 299) . The blot waε probed with a lOkb DNA fragment of Zygoεaccharomyceε bailii rDNA excised with Sail from the plasmid pAT68 (K. Sugihara et al . , Agric . Biol . Chem . 50 (6) :1503-151 (1986)). The probe was labeled and the blots were developed using DIG DNA Labeling and Detection Kit (Boehringer 1093 657) according to the manufacturer's instructionε. Three hybridization bandε were obεerved corresponding to DNA fragments of approximtely 4.5, 2.7 and 1.1 kb. Using the blot as a reference, a band corresponding to the largest (4.5 kb) hybridizing DNA fragment was cut out of the remaining portion of the preparative gel, the DNA was electroeluted and ligated with pUC19 cut by EcoRI and Hindlll . The ligation mixture waε uεed to tranεform E . coli . The tranεformed bacteria were plated onto a charged nylon membrane (Bio-Rad 162-0164) laying on the agar εurface of a plate containing LB medium with ampicillin. After 24 h incubation at 37°C, the membrane was lifted and in situ lysis of bacterial colonies was performed according to the manufacturer's inεtructionε. No replica plateε were needed εince E . coli penetrates thiε type of membrane and after the filter iε lifted there is a visible trace of every bacterial colony on the agar surface. The membrane was probed with the same Z . bailii rDNA fragment using the same DIG detection kit as above. A number of poεitive cloneε were identified (approximately 2-5% of all cloneε) . A restriction analysis of the plasmid mini-preparations from 8 hybridization-poεitive cloneε and 4 hybridization- negative clones was performed using a mixture of EcoRI , Hindlll and EcoRV . All hybridization-poεitive cloneε produced identical reεtriction fragment patternε (with characteriεtic fragmentε of 0.55 and 1.5 kb) while the εame patterns of the hybridization-negative clones were all different. The plaεmid DNA from one of hybridization- poεitive clones waε isolated on preparative scale and name pTCrDNA. It was concluded that the cloned piece was a fragment of T . Candida rDNA because 1) it hybridized εtrongly with rDNA of Z . bailii and 2) it was cloned from the partially enriched T . Candida chromosomal DNA digest with high frequency (rDNA is known to be represented by about 100 copieε in yeaεt) . A partial reεtriction map of the cloned DNA fragment is shown in Figure 5. A plasmid combining the rDNA fragment of T . Candida with a dominant selection marker was conεtructed as follows. Plasmid pUT332 (Gatignol, A., et al . , Gene 92:35- 41 (1990)) was cut with Hindlll and Kpnl , the 1.3 kb DNA fragment waε isolated by agarose electrophoresis, and ligated with the 4.5 kb Hindlll-EcoRI fragment from pTCrDNA and pUC19 digested with EcoRI and -Kpnl (Figure 6) . The ligation mixture was transformed into E. coli and a clone bearing the plaεmid pTC(PHLE) waε identified by reεtriction analysis. PTC(PHLE) was partially digeεted with Sail and the linear form of the plaεmid was purified by agarose gel electrophoresiε. It waε ligated with a 5 kb Sail fragment iεolated from pJDB(AX)-16 (Example 2). Plaεmid pTC(AX) waε identified among the cloneε obtained after tranεformation of thiε ligation mixture into E . coli (Figure 6) . This plaεmid contains the following functional elements: a) bacterial phleomycin-resistance gene under con¬ trol of a yeaεt promoter and transcription terminator en¬ abling the direct selection of the yeast cells tranεformed by thiε plasmid; b) a piece of T. Candida rDNA providing a target for homologous recombination with T. Candida chromosome and improving the efficiency of transformation; c) the expresεion cassette for arabitol-dehydroge- nase and xylitol dehydrogenaεe geneε providing for the syntheseis of the two enzymes of arabitol-xylitol conversion pathway.

Example 6

Transformation of T. Candida and analysis of xylitol production

The plasmid pTC(AX) was used to transform the Torulopεiε Candida strain ATCC 20214. T. Candida was grown for 36 h in YEPD medium containing 10% glucose. The cells were collected by centrifugation (2000 rpm for 10 min at 4°C) and washed three timeε with εterile 1 M εorbitol. The cell pellet waε εuεpended in an equal volume of cold 1 M εorbitol, 200 μl aliquoteε were mixed with pUT(AX) DNA (20- 100 μg) and then tranεferred into ice-cold 2 mm electrode gap electroporation cuvetteε and electroporated uεing Invitrogen ElectroPorator apparatuε with the following εettingε: voltage 1800 V, capacitance 50μF, parallel resistance 150 Ω. The cells were transferred into 2 ml of YEPD containing 1 M εorbitol and incubated ovenight at 30°C on a εhaker. The tranεformed cellε were collected by low speed centrifugation, and plated onto plateε containing YEPD medium titrated to pH 7.5 and containing 30 μg/ml of phleomycin. The plateε were incubated at 30°C for 7-10 days. Most of the yeast colonies that developed during thiε time were background mutantε εince εimilar number of colonieε appeared alεo on the control plateε (which contain cellε treated similarly but without addition of DNA) . To distinguiεh true transformants from spontaneous mutants, the chromosomal DNA was isolated from 72 individual yeast colonies by a scaled down procedure for isolating T . Candida chromosomal DNA deεcribed above. 10 μg of each of theεe DNA preparationε waε cut with a mixture of EcoRI and BamHI . The digeεtε were εeparated on a 1% agaroεe gelε and then blotted onto a poεitively charged nylon membrane aε deεcribed in Example 5. The blots were probed with DNA from the plasmid pADH (Example 2) which contains arabitol- dehydrogenaεe εequences and pUC sequences but no DNA fragments of yeast origin (to avoid hybridization between possible homologouε yeast sequenceε) . The probe was labeled and the blots were developed using DIG DNA Labeling and Detection Kit (Boehringer 1093 657) . Only one clone (T. Candida : :pTC(AX) ) with a hybridization signal compatible with the structure of the transforming plasmid (three bands in the 2-3 kb region) was discovered indicating a very low transformation efficiency. A positive hybridization signal was detected for one more clone, however, the position of the only hybridizing band (about 5-7 kb) indicated that either only a fragment of the pTC(AX) haε integrated into yeast chromosome or some rearrangement occurred at the integration εite. We aεεumed that the plasmid had integrated into the T. Candida chromosome. This asεumption iε compatible with the obεervation that after growth in non-εelective medium and cloning, all cloneε of T. Candida : :pTC(AX) retain the phleomycin reεiεtant phenotype.

The T . Candida : :pTC(AX) tranεfor ant was grown in YEPD medium containing 10% glucose for 36 h, and the arbitol dehydrogenase and xylitol dehydrogenase activities were measured as deεcribed in Example 4. The reεults are presented in Table 5.

The activity of xylitol-dehydrogenase was not significantly increaεed over the activity level of endogenous T. Candida xylitol dehydrogenase. The activity of plaεmid-encoded arabitol-dehydrogenase (EC 1.1.1.11) was difficult to separate from the activity of endogeneous εynonymouε but different (ribuloεe-forming, EC 1.1.1.) arabitol dehydro¬ genase. The only definitive conclusion from this experiment was that the expreεεion level of both enzymeε waε much lower than in Z . rouxii . Attemptε to increaεe the plaεmid copy number and the expreεεion level of the two dehydrogenaεeε of the integrated plasmid by cultivating T . Candida : :pTC(AX) on media with increasing concentrations of phleomycin were not succeεsful.

The xylitol production by the T. Candida : :pTC(AX) was tested after growing it on YEPD containing 10% glucoεe for 5-7 dayε. In three εeparate experimentε, the transformant produced 1.1; 1.6; and 0.9 g/1 xylitol, while no xylitol was detected in the culture medium of the wild type T. Candida by HPLC. The detection limit of the analytical method we employed is lower then 0.1 g/l. Therefore, it is possible to conclude that xylitol production by T. Candida : :pTC(AX) is in fact determined by the plaεmid.

Example 7 Tranεformation of Candida polymorpha with arabitol-dehydrogenaεe and xylitol-dehydrogenaεe geneε.

In order to introduce arabitol dehydrogenaεe and xylitol-dehydrogenaεe genes into Candida polymorpha , a mutant in the orotidine phosphate decarboxylase gene (hereinafter called alεo URA3) waε iεolated uεing a modification of the method of Boeke et al . (Boeke, J.D., et al . , Mol . Gen . Genet . 197:345-346 (1984)). C. polymorpha εtrain ATCC 20213 waε grown for 24 h in YEPD, the cellε were collected by centrifugation (2000 rpm, 10 min) , waεhed with water two times and suspended in three volumes of sterile 0.1 M sodium phosphate buffer pH 7.0. Ethyl methaneεulfonate waε added to 1% concentration and the cells were incubated for 2h at room temperture. The reaction was stopped by transferring the cells into 0.1M sodium thiosulfate solution and washing them three times with sterile water. Mutagenized cells were transferred into 0.5 liters of YEPD and grown at 30°C with shaking for two days. The yeaεt waε collected by centrifugation, waεhed two timeε with water and tranεferred into 1 liter of medium containing 0.7% Yeaεt Nitrogen Base (Difco) , 2% glucose (SC medium) and incubated on a rotary shaker for 24 h at 30°C. 1 mg of nystatin was added to the culture and the incubation continued for 4 hours. Nyεtatin-treated cellε were εeparated from the medium by centrifugation, waεhed two timeε with water, and tranεferred into 1 liter of SC medium containing 50 mg/liter of uracil. The cells were incubated on a rotary shaker for 5 days and then plated on SC medium plates containing 50 mg/liter uracil and 1 g/liter of fluoroorotic acid. After incubating the plateε for two weekε at 30°C, approximately 400 fluoroorotic acid reistant colonies were obtained and all of them were tested for uracil auxotrophy. Five uracil-dependent clones were isolated. However, three cloneε did grow on uracil-free medium, although at a reduced rate. Two clones (named C. polymorpha U-2 and C. polymorpha U-5) which had a clear uracil-dependent phenotype were used for tranεformation experimentε.

Cloning of the C . polymorpha URA3 gene waε achieved by a conventional εtrategy. The chromoεomal DNA was isolated by the method described in Example 5. The DNA was partially cut with Sau3A and fractionated on agaroεe gelε.

Several fractionε were collected and their molecular εize diεtribution was checked by analytical gel electrophoresis. The fractions in the range of 5 to 10 kb were uεed for cloning experimentε. The vector uεed for construction of the library, pYEpl3 (Broach et al . , Gene 8:121-133 (1979)), contains the S. cerevisiae LEU2 gene, 2μ origin of replication and a unique restriction site for BamHI . The vector waε cut with BamHI , purified by agaroεe gel electrophoreεiε and dephoεphorylated with bacterial alkaline phoεphataεe. Several independent vector preparationε and ligation conditions varying the vector to insert ratio and reaction volume) were tested in small εcale experimentε to optimize for the largeεt number of tranεformantε and highest percentage of recombinant clones

(analysed by restriction analysis of plasmid minipreps from random cloneε) . The large εcale ligations were performed using the optimized conditionε and tranεformed into E . coli εtrain XLl-BLUE. The conεtructed library included about 15,000 primary tranεformantε approximately 90% of which were inεert-containing.

Yeaεt εtrain DBY746 (MATalpha, leu2-3,112, his3-Al, trpl-289, ura3-52) was transformed with a C. polymorpha gene library. Each library pool was transformed into S. cerevisiae separately using about 20 μg of plaεmid DNA and plating the tranεformation on one plate supplemented with uracil, tryptophan and histidine (i.e. using only leucine selection). 3,000-10,000 yeast tranεformantε per plate were obtained. The yeaεt tranεformantε were then replica plated on plates with mininal medium supplemented with tryptophan and histidine (uracil minus plates) . Control replicas on histidine-hiεtidine (uracil minuε plateε) . Control replicas on histidine- inus and tryptophan-minus plates were also made. One to four uracil- independent clones appeared on almoεt every plate. The histidine-independent and tryptophan-independent isolates were also obtained, however the number of isolates waε 2-3 timeε lower then for URA3 iεolateε. Plasmids from six of the uracil-independent clones were rescued into E . coli , isolated on a preparative scale and used to re-transform DBY746 to leucine prototrophy. 10- 20 random colonies from each transformation were then checked for uracil dependence. Five out of six rescued plasmids transformed DBY746 to the Leu⁺ Ura⁺ phenotype. Restriction analysis of the five plasmids named revealed very similar restriction patterns. Those patterns were too complex to produce an unambiguouε restriction map of the cloned fragment. However, it was found that Hindlll digestion generates in all clones a fragment covering most of the DNA insert (approximately 4.5 kb) . This fragment from one of the C. polymorpha URA3 isolateε - pCP291 - waε purified by agaroεe gel electrophoreεiε and εubcloned into pJDB(AX) partially digeεted with Hindlll. The εtructure of the plaεmid pCPU(AX) iεolated aε a reεult of thiε cloning experiment iε shown in Figure 7A.Z

Transformation of both C. polymorpha U-2 and C. polymorpha U-5 was attempted using the same electroporation conditions as deεcribed in Example 6. The electroporated cellε, after overnight εhaking in YEPD containing 1M εorbitol, Zwere plated on SC medium plateε. After 10 dayε of incubation at 30°C, approximately 100 colonies were observed on the plate containing C. polymorpha U-2 transformed with pCPU(AX) , while the number of colonies on the control plate (containing similarly treated cells of this strain without added DNA) waε only 14. C. polymorpha U-5 failed to demonεtrate a εignficant effect in a transformation experiment over the no-DNA-control. Three random clones from the C. polymorpha U-2 transformation plate were firεt εtreaked on a freεh SC medium plate and theεe εtreakε uεed to inoculate 100 ml cultureε of YEPD containing 15% glucoεe. The control culture waε inoculated with C. polymorpha U-2. After incubation on a rotary εhaker (200 rpm) at 30°C for 10 dayε, the xylitol content in the culture medium was analysed by HPLC. The reεults of this experiment are presented in Table 6.

Considerable variation in the level of xylitol production between different clones was observed. It may be a consequence of integration of the pasmid pCPU(AX) at different loci of the C. polymorpha chromosome. Strain C. polymorpha U-2: :pCPU(AX)-2 is probably a revertant and not a true transformant. However, these experiments clearly demonstrated that the arabitol-xylitol pathway may also be introduced into C. polymorpha .

Example 8 Cloning of the enzymes of oxidative pentose phosphate pathway and their overexpresεion in osmophilic yeast

The first enzyme of the oxidative pentose phoεphate pathway - D-glucose-6-phosphate dehydrogenaεe iε coded in S. cerevisiae by the ZWF1 gene. The εequence of this gene is known (Nogae T. , and Johnston, M. Gene 96:161-19 (1990); Thomaε D. et al ., The EMBO J. 20:547-553 (1991)). The gene including the complete coding region, 600 bp of the 5'- noncoding region and 450 bp of the 3'-noncoding region has been cloned by PCR using the two oligonucleotides: CAGGCCGTCGACAAGGATCTCGTCTC (5'-oligonucleotide) [SEQ ID No. :3 : ] and AATTAGTCGACCGTTAATTGGGGCCACTGAGGC (3'-oligo¬ nucleotide) [SEQ ID No.:4:]. The 5'-oligonucleotide anneals at positionε 982-1007 and the 3'-oligonucleotide annealε at poεition 3523-3555 in the numbering of D-gluccεe-6- phosphate dehydrogenase as described in Nogae T. , and Johnston, M. Gene 96:161-19 (1990). The chromosomal DNA was isolated from S. cerevisiae strain GRF18 by the method described in Example 3. The PCR parameters were the same as in Example 3. The amplified DNA fragment containing the ZWF1 gene was digested with Sail and cloned into pUC19 digested with the same restrictaεe reεulting in plasmid pUC(ZWF) . The identity of the cloned gene was checked by restriction analysis. The second enzyme of the pentose phosphate pathway, 6-phosphogluconic acid dehydrogenase iε coded in E . coli by the gnd gene. The nucleotide εequence of thiε gene is known (Nasoff, M.S. et al ., Gene 27:253-264 (1984)). In order to clone the gnd _.gene from E . coli , the chro oεomal DNA was iεolated from the E. coli εtrain HB101 by a method identical to the method uεed for iεolation of the Klebsiella terrigena DNA (Example 1) . The oligonucleotideε (GCGAAGCTTAAAAATGTCCAAGCAACAGATCGGCG [SEQ ID No.:5:] and GCGAAGCTTAGATTAATCCAGCCATTCGGTATGG [SEQ ID No. :6:]) for the PCR amplification of the gnd gene were deεigned to amplify only the coding region and to introduce Hindlll εiteε immediately upεtream of the initiation codon and immediately downstream of the stop codon. The 5'- oligonucleotide anneals at positions 56-78 and the 3'- oligonucleotide anneals at position 1442-1468 in the numbering of 6-phosphogluconic acid dehydrogenase as described in Nasoff, M.S. et al . , Gene 27:253-264 (1984) . The amplified DNA fragment was digested with Hindlll and ligated with the Hindlll digeεted vector pUC19. Ten independent apparently identical cloneε of the resulting plasmid pUC(gnd) were pooled. This was done in order to avoid poεεible problemε aεεociated with the sequence errors which might be introduced during PCR amplification. The coding region of the gnd gene was fuεed with the S. cerevisiae ADCI promoter and transcription terminator by transferring the 1.4 kb Hindlll fragment from the pUC(gnd) pool into the expreεεion vector pAAH5 (Ammerer, Meth . Enzymol . 202:192-203 (1983)). Several independent clones of the resulting plasmid pAAH(gnd) were transformed into S. cerevisiae strain GRF18..by the lithium chloride procedure (Ito et al . , J. Bacteriol . 153:163-168 (1983)) and the activity of 6-phospho-D-gluconate dehydrogenase was measured in the tranεformants. In all the transformants, the activity of the 6-phospho-D-gluconate dehydrogenaεe waε elevated several times relative to the untransfor ed hoεt indicating that the bacterial 6-phospho-D-gluconate dehydrogenase can be efficiently expresεed in yeast. The clone of pAAH(gnd) which produced the highest activity in yeaεt waε chosen for further constructions. The cloning of the ZWF1 and gnd genes as well as the construction of the pAAH(gnd) plasmid are illustrated by Figure 8 and 8a.

In order to overexpreεε simultaneously both D- glucose-6-phosphatedehydrogenase and 6-phospho-D-gluconate dehydrogenase genes in an osmophilic yeast host, the plasmid pSRT(ZG) was constructed. The method for constructing this plasmid is illuεtrated by Figure 9. Briefly, the 6-phoεpho-D-gluconate dehydrogenaεe expreεεion caεεette from the plaεmid pAAH(gnd) waε transferred as a 3.1 kb BamEI DNA fragment into BamHI cut pUC19. The reεulting plasmid pUC(ADHgnd) was cleaved with Sacl and -j al and a 3.1 kb DNA fragment was purified by agarose gel electrophoresiε. Thiε fragment waε εimultaneouεly ligated with two other DNA fragments: 2.5 kb fragment of the pUC(ZWF) obtained by digestion with Sacl and partial digestion with Pstl and a vector fragment of the plasmid pSRT(AX)-9 (Example 3) digested with Pstl and Xbal . The structure of the resulting plaεmid pSRT(ZG) waε confirmed by reεtriction analysis. After transforming Z . rouxii ATCC 13356 with pSRT(ZG) (by the method described in Example 4) the activity of D-glucose-6-phosphate dehydrogenase and 6- phoεpho-D-gluconate dehydrogenaεe waε measured in the transformed strain. Both enzymes had approximately twenty times higher activities in the transformed strain than in the untranεformed control (Table 7) . Similar methodε can be uεed to achieve overexpression of the two genes coding for enzymes of the oxidative part of the pentose phosphate pathway in other yeast specieε. However, εince there are no vectorε capable of being maintained aε extrachromoεomal plasmids in most yeaεt εpecieε other than Saccharomyceε or Zygoεaccharomyceε - integrative tranεformation iε the only useful method for such hosts. The preferred type of integrative vectors are the vectorε targeted for integration at the riboεomal DNA locuε εince vectorε of thiε type provide for high copy number integration and conεequently for higher expreεεion level (Lopeε, T.S. et al . , Gene 79:199-206 1989)).

Table 7

Activity of the D-glucose-6-phosphate dehydrogenase and

6-phospho-D-gluconate dehydrogenase in Z. rouxii ATCC 13356 transformed with pSRT(ZG) and untransformed control

(μmole/min*mg of protein)

Yeast strain D-glucose-6-P 6-P-Gluconate dehydrogenase dehydrogenase

Z rouxii 0.33 0.45 (untransformed)

Z rouxii 7.3 7.7 [pSRT(ZG)]

Example 9

Construction of the transketolaεe mutantε in yeaεt The tranεketolaεe mutantε in yeaεt can be obtained oεt conveniently by a εite directed gene diεruption method although conventional methods of chemical mutagenesis are also applicable. A homologous tranεketolaεe gene cloned from the yeaεt species in which the mutation is desired will generally be necessary to apply the gene disruption technology although sometimeε a heterologouε clone from a very closely related species can also be used. The sequence of the transketolaεe gene from S. cereviεiae iε known (Fletcher, T.S., and Kwee, I.L., EMBL DNA εequence library, ID:SCTRANSK, accession number M63302) . Uεing this sequence, the S. cereviεiae transketolase gene was cloned by PCR. Oligonucleotides AGCTCTAGAAATGACTCAATTCACTGACATTGATAAGCTAGCCG [SEQ. ID No. :7: ] and GGAGAATTCAGCTTGTCACCCTTATAGAATGCAATGGTCTTTTG [SEQ ID No.:8:] and the chromoεomal DNA from S. cereviεiae (isolated as described above) were used for the amplification of the DNA fragment containing the transketolase gene. The 5'-oligonucleotide annealε at poεitionε 269-304 and the 3'-oligonucleotide annealε at poεition 2271-2305 in the numbering of transketolase as described in (Fletcher, T.S., and Kwee, I.L., EMBL DNA se- quence library, ID:SCTRANSK, accession number M63302) . The fragment was digested with Xbal and -EcoRI and cloned into the pUCl9 cleaved with the same enzymes. The restriction analyεis of the resulting plasmid pUC(TKT) confirmed the identity of the cloned DNA fragment. This plasmid waε cut with Bgrlll and Clal and the large DNA fragment waε purified by agaroεe electrophoresis. Thiε fragment waε ligated with two DNA fragmentε iεolated from the plaεmid pUT332 (Gatignol, A. et al ., Gene 92:35-41 (1990)) : a Clal-Pstl fragment bearing the URA3 gene and the 3'-part of the bleomycin reεiεtance gene and a BamHI-Pstl DNA fragment bearing the yeaεt TEFl (tranεcription elongation factor 1) promoter and the 5'-part of the bleomycin reεistance gene coding sequence. After transformation of E . coli with the above ligation mixture and restriction analyεiε of the plaεmid clones, the plasmid pTKT(BLURA) was iεolated. Thiε plasmid contains the coding sequence of S. cereviεiae transketolase gene in which the 90 bp fragment of between the Clal and Bglll siteε iε εubstituted with a fragment of pUT332 containing two markers selectable in S. cereviεiae - the URA3 gene and the bleomycin resistance gene under control of the TEFl -. promoter and CYCl transcription terminator. The plaεmid waε uεed to tranεform S. cereviεiae εtrain DBY746 (ATCC 44773; MATα his3Al leu2-3 , 112 ura3-52 trpl-289) to phleomycin reεiεtance (10 μg/ml phleomycin in YEPD medium) by the lithium chloride method (Ito et al . , J. Bacterial . 153:163-168 (1983)). The transformants were tested for uracil prototrophy and for the ability to grow on D-xylulose. Five URA3 clones, three of which displayed reduced growth on D-xylulose, were grown in 100 ml cultureε, cell extractε were prepared by εhaking with glaεε beads and the transketolaεe activity waε meaεured in the crude extractε. The assay was performed in 0.1 M glycyl- glycine buffer pH 7.6 containing 3 mM magnesium chloride, 0.1 mM thiamine pyrophosphate, 0.25 mM NADH, and 0.2 mg/ml bovine εerum albumin. Immediately before measurement 3 μl of a solution containing 0.5 M D-xyluloεe-5-phoεphate and 0.5 M ribose-5-phosphate were added to 1 ml of the above buffer followed by 7.5 U of triose phoεphate iεo eraεe and 1.5 U of cf-glycerophosphate dehydrogenase (both from Sigma) . The reaction was initiated by adding a suitable aliquot of the crude extract and followed by recording the decrease of optical density at 340 nm. The cellular extract of the strain DBY746 was used as a control. The transketolase activity in the crude extract of DBY746 and the two tranεformantε with unretarded growth on D-xyluloεe waε readily measurable at approximately 0.25 U/mg protein. The three transformants with reduced growth on D-xyluloεe had tranεketolase activity below the detection limit of our method (at least 20 times lower than wild-type) . The activities of D-glucoεe-6-phosphate dehydrogenase and 6- phospho-D-gluconate dehydrogenase were also measured as a control for possible enzyme inactivation during the preparation of the cell extracts. The activities of these two enzymes were very εimilar in all εix εtrainε. Therefore, it waε concluded that the three cloneε which grew poorly on D-xylulose contained the mutation in the transketolase gene. The growth of the strains with the disrupted transketolaεe gene waε alεo εomewhat retarded on a εynthetic medium lacking aromatic amino acids phenylalanine and tyrosine although the effect was smaller than the effect of this mutation on D-xylulose utilization. Transketolase genes from other yeast specieε can conveniently be cloned by complementation of the transketolase mutation in the S. cereviεiae strains described above. Preferably, the cloning can be performed by constructing a gene library of the non-Saccharomyceε yeast strain in an appropriate vector (for example, the well known plasmid YEpl3) , transforming this library into a S. cereviεiae εtrain bearing transketolase mutation (for example, the mutants obtained by gene disruption described above) and selecting for tranεformantε with reεtor d growth on D-xyluloεe. The plaεmid DNA can be reεcued from εuch D- xylulose-positive tranεformantε and uεed to tranεform the same recipient εtrain. All the cloneε from thiε transformation εhould be able to grow well on D-xyluloεe. Transketolase activity can be measured in the tranεformants and its reappearance at a εignificant level can εerve aε proof of the identity of the cloned gene. Additional and final proof can be obtained by εequencing εhort εtretcheε of cloned DNA and finding pieceε of εequence homologous to the S. cereviεiae transketolaεe gene εequence or by demonεtrating hybridization of the cloned DNA fragment with the authentic tranεketolaεe clone from S. cereviεiae . Alternatively, the cloning procedure can be baεed on the DNA hybridization aε primary method for selecting the clones containing the transketolase gene from the gene library of a non-Sacc.haro_7.yces yeast. A fragment of the S. cereviεiae transketolase gene can be used aε the probe for a colony or plaque hybridization experiment and the cloneε which give the strongest hybridization signal can be further analyzed by partial εequencing.

Whatever method iε uεed for the cloning of tranεketolaεe gene from a chosen yeast specieε the εubεequent εtepε for obtaining a mutation in the tranεketolase gene in this yeast are the same. The cloned DNA fragment should be characterized by constructing a partial reεtriction map and preferably localizing the coding region of the tranεketolaεe gene. Then a piece of DNA which can function aε a εelectable marker in the choεen yeaεt iε inserted into the DNA fragment containing the transketolase gene not closer than several hundred bp from either of the termini of this fragment. The cassette containing the bacterial phleomycin gene under control of a strong yeast promoter, such as the above-described cassette from the plasmid pUT332, could for example, be used for many yeast species as a dominant selective marker. It is esεential that the inεertion of the DNA fragment bearing the εelectable marker iε done in εuch a faεhion that the coding region of the tranεketolaεe gene is either disrupted by the inserted DNA or, preferably, the inserted DNA fragment subεtitutes (part of) the coding region. Such a DNA construct can then be used to disrupt the chromosomal copy of the transketolaεe gene in the selected yeast by a method similar to the method described above for obtaining the transketolaεe mutation in S. cereviεiae . Any suitable transformation method can be employed, the preferred methods are protoplast transformation and electroporation. The selection of the cloneε bearing the diεrupted tranεketolaεe gene can be done εimilarly to the method deεcribed above for S. cereviεiae . Alεo, the analyεiε of the εtructure of the transketolase chromosomal region by Southern hybridization can be uεed aε an alternative method or in addition to other methods.

Example 10

Cloning of the D-ribulokinase gene

The preferred way to clone the D-ribulokinase gene is similar to the method described in Example 1 for the cloning of the D-arabitol dehydrogenase. It iε known that D-ribulokinase gene in several bacteria such as E. coli or Klebsiella aerogeneε is a part of the ribitol utilization operon (Loviny, T. et al . , Biochem . J. 230:579-585 (1985)). It iε also known that E . coli B strainε do not contain thiε operon and are therefore incapable of utilizing ribitol as a carbon source. Thus, an E. coli B strain (such as common laboratory εtrains HB101 and JM103 or strains which can be transformed with high efficiency such as SCSI or XLl-Blue from Stratagene) can be transformed with a gene library of a ribitol-utilizing bacteria constructed in any suitable vector, preferably pUC19. Non-pathogenic bacterial species such as Klebεiella terrigena are the preferred source organismε for iεolation of the D-ribulokinaεe gene. The E. coli transformants which are capable of growth on minimal medium containing ribitol as the sole carbon source can then be selected. The plaεmid DNA from εuch ribitol- poεitive clones can be isolated and used to retransform an E . coli B strain. All transformants from such retransformation εhould be able to grow on ribitol aε the εole carbon εource. A reεtriction map of the cloned inεert can then be constructed. Using this map various deletion derivatives of the original clone can be prepared and analyzed for the retention of ribitol operon by above- mentioned functional test. Several successive deletions can be performed in order to minimize the size of the DNA fragment bearing the ribitol operon to 3.5 - 4 kb (the size of this operon in K. aerogeneε) . Finally, (partial) nucleotide sequence of the D-ribulokinase gene can be determined and used to exciεe the coding region of thiε gene either uεing suitable naturally occurring reεtriction εiteε or uεing known PCR techniqueε for introduction of such siteε. The D-ribulokinaεe gene can be expressed in other hosts, preferably yeaεtε, by a method that includeε εtandard εtepε such as fusing the coding region of the D- ribulokinaεe gene to a εuitable promoter and tranεcription terminator, transferring the expresεion caεεette to a vector εuitable for the tranεformation of the choεen hoεt, obtaining the transformants and, finally verifying the efficiency of D-ribulokinase expression.

Example 11

Cloning and overexpression of the D-ribulose-5- phosphate-3-epimerase gene

The method for isolating homogeneous D-ribuloεe-5- phoεphate-3-epimeraεe from baker's yeast (industrial Saccharomyceε cereviεiae yeast) is known (Williamson, W.T. et al . , Meth . Enzymol . 9:605-608 (1966)). The enzyme can be isolated and the N-terminal as well as partial internal aπino acid sequences determined by the generally known methods. Thuε obtained partial amino acid εequenceε can then be used to generate, by a procedure known as reverse translation, the sequenceε of oligonucleotides which then can be used to prime the polymerase chain reaction. The DNA fragments generated by PCR can be used as hybridization probes to screen a yeast gene library for a full length copy of the D-ribulose-5-phoεphate-3-epimeraεe gene. The preferred way to overexpreεs the D-ribulose-5-phosphate-3- epimeraεe gene in other yeaεt hosts is to clone it into a vector which haε a high copy number in the deεired hoεt (for example, pSRT303D vector for Z . rouxii ) . An alternative and more efficient way of overexpreεsing the gene is to determine at least a partial nucleotide sequence of the D-ribulose-5-phoεphate-3-epimerase gene around the translation start codon and uεe this information for isolating the coding sequence of the D-ribulose-5- phoεphate-3-epimeraεe gene and fusing it to a promoter known to function efficiently in the chosen host. Example 12

Cloning of the D-xylulokinase gene and construction of D-xylulokinase mutants

Methods for cloning of the D-xylulokinase (EC 2.7.1.17) gene from different yeast species have been described (Ho, N.W.Y. et al . , Enzyme Microbiol . Technol . 12:417-421 (1989); Stevis, P.E. et al., Applied and Environmental Microbiol . 53:2975-2977 (1987)) . Also, a method for conεtructing the D-xylulokinaεe mutation in S. cereviεiae by gene disruption has been described (Steviε, P.E. et al . , Appl . Biochem . Biotechnol .20:327-334 (1989)). Similar methodε can be uεed for conεtructing D-xylulokinaεe mutantε in other yeaεtε. For the yeaεt εpecieε other than S. cereviεiae, the genetic arkerε uεed for the disruption of D-xylulokinase gene are preferably dominant antibiotic resistance markers (see Example 9) . Alternatively, clasεical mutant conεtruction methodε baεed on chemical (for example, treatment with ethyl methane sulfonate or acriflavine) or physical (ultraviolet light, X-rays) mutagenesis can be employed. The mutant enrichment can be performed by growing the mutagenized cellε on D-xyluloεe aε the εole carbon εource in the presence of antibiotic (such as nystatin) which killε only growing cells. The inability of D-xylulokinase mutants to utilize D-xylulose as the sole carbon source for growth can be uεed for the εelection of mutants.

Example 13

Strains producing xylitol via D-arabitol with improved yield Example 4 describeε the method for the construction of a yeast strain capable of producing xylitol from structurally unrelated carbon εourceε εuch as D-glucose by a pathway which utilizeε D-arabitol as the key intermediate. To improve xylitol yield in fermentations with the εtrainε utilizing this "D-arabitol pathway" - the D-arabitol yield must be improved. The pathway leading from D-glucose to D-arabitol in D-arabitol-producing yeasts has been described (Ingram, J.M. et al . , J. Bacteriol . 89:1186- 1194 (1965)) . D-arabitol is produced from D-ribulose-5- phoεphate via dephoεphorylation and reduction with a NADPH- linked D-ribulose reductase. Formation of D-ribulose-5- phosphate from D-glucose 6-phosphate by two εucceεεive irreverεible dehydrogenation steps with D-glucose-6- phosphate dehydrogenase and 6-phospho-D-gluconate dehydrogenase is a universally occurring pathway known as the oxidative branch of the pentose phosphate pathway (or hexose monophoεphate εhunt) . In the non-oxidative branch of the pentoεe phosphate pathway, D-ribulose-5-phoεphate iε reverεibly iεomerized into riboεe-5-phoεphate and D- xyluloεe-5-phoεphate. Riboεe-5-phoεphate and D-xyluloεe-5- phoεphate are further metabolized by tranεketolaεe. Therefore, tranεketolase can be mutated in an D-arabitol- producing microbial strain and the fraction of D-ribulose- 5-phoεphate converted into D-arabitol will be increaεed. Example 9 deεcribes the method for obtaining the transketolase mutants. Further increase of the D-arabitol yield can be achieved if the rate of D-ribulose-5-phosphate bioεyntheεiε iε maximized through overexpreεεion of the two genes coding for the enzymes of the oxidative branch of the pentoεe phosphate pathway as described above (Example 8) . The strains optimized by this method with respect to the D- arabitol yield can then be further transformed with recombinant DNA conεtructionε bearing the xylitol dehydrogenaεe and D-arabitol dehydrogenaεe geneε (Exampleε 3 and 4) reεulting in εtrainε with improved efficiency of xylitol production. Example 14

Strains producing xylitol by alternative pathways The method according to Examples 4 and 13 are the most straightforward methodε for the conεtruction of microbial strains capable of converting D-glucoεe and other carbon sourceε into xylitol. Theεe methodε utilize the naturally occurring pathway leading to the formation of D- arabitol from variouε carbon εourceε and extend thiε pathway by two more reactionε to convert D-arabitol into xylitol. However, thiε pathway iε not the only possible pathway. Other pathways leading to xylitol aε a final metabolic product and not involving D-arabitol aε an intermediate can be conεtructed. Thuε, a pathway to xylitol from the same precursor - D-ribulose-5-phosphate can be realized through a different chain of reactions. D- ribulose-5-phosphate can efficiently be converted to D- xylulose-5-phosphate by D-ribuloεe-5-phosphate-3-epimerase (Example 11) and if further converεion of D-xyluloεe-5- phoεphate iε prevented by a mutation in the transketolase gene, the accumulated D-xylulose-5-phoεphate can be dephoεphorylated by the εame non-εpecific phoεphatase as D- ribulose-5-phosphate (Ingram, J.M. et al . , J. Bacteriol . 89:1186-1194 (1965)) and reduced into xylitol by xylitol dehydrogenase (Example 3) . Realization of this pathway can further require the inactivation of D-xylulokinase gene (Example 12) in order to minimize the energy loεε due to the futile loop: D-xyluloεe-5-phosphate -> D-xylulose → D- xyluloεe-5-phoεphate. An additional genetic change introduction and (over)-expression of the D-ribulokinase gene (E.C. 2.7.1.47) could minimize simultaneous D-arabitol production by such strains by trapping the D-ribulose produced by the unspecific phosphatase. The D-ribulose will be converted back into the D-ribulose-5-phosphate and further into D-xylulose-5-phoεphate. Example 15

Stability of the recombinant Z . rouxii strain and production of xylitol under conditions of fermentor cultivation The stability of xylitol production during extended cultivation was checked in both selective conditions (using the selective medium: YEPD containing 50 mg/literε G418 and 30% glucoεe) and non-εelective conditionε (uεing the same medium without G418) . A single freshly obtained transformant of Z . rouxii ATCC 13356 [pSRT(AX) -9) ] waε grown in a 200 ml volume of G418 containing YEPD. The cellε were tranεferred into 50% glycerol εolution and frozen at - 70°C in 1 ml aliquoteε. Four frozen aliquotes of Z . rouxii [pSRT(AX) -9) ] were used to inoculate two 50-ml cultures in selective medium and two in non-selective medium. After the cultureε reached the stationary phase of growth (50-60 h at 30°C and 200 rpm) a sample waε taken for the HPLC analysis of pentitol content and 1 ml of the culture waε uεed to inoculate another 50 ml of the εame (either εelective or non-εelective) medium. The growth-dilution cycle waε repeated four more times. The conditions of this experiment approximate the propagation of the recombinant strain from a εtandard frozen inoculum in a large εcale fermentation. The reεultε of this experiment are presented in Table 8. Predictably, the stability of the recombinant strain is higher on the εelective medium. However, even under non- εelective medium the decline in xylitol yield was only detected after approximtely 20 generations. Under εelective conditionε, the xylitol production was stable for approximately 30 generationε.

An aliquote of the frozen εtock of the tranεformed

Z . rouxii εtrain waε used to inoculate a 2 liter fermentor containing 1 liter of medium having the following composition (per liter): 0.1 g NaCl , 6.8 g potaεεium phosphate, 0.5 g ammonium sulphate, 20 g of yeast extract and 400 g of glucose, 50 mg of G481, pH 6.0. The cultivation conditionε were: aeration rate, 0.5 v/min; agitation, 400 rpm; temperature, 30°C.

Figure 10 εhowε the time courεe of the glucoεe conεumption and xylitol accumulation in this fermentation. The concentration of dissolved oxygen which reflects the respiritory activity of the yeast culture is also shwon. An apparent biphasic growth was observed: in the first phase, a plateau in glucose and xylitol concentration was reached in about forty hours (less than half of the available glucose was conεumed at that point) , the εecond phaεe waε obεerved after approximately 200 hours of cultivation when the glucose conεumption and xylitol production reεumed. The final xylitol concentration was 15 g/liter, almost two times higher than the concentration obtained in the flask fermentations. The biphasic growth with a long lag period indicated that a spontaneouε mutant waε εelected (preεumably having a higher alcohol tolerance than the parent strain) . To check this hypothesiε, a εingle clone was isolated from the culture at the endpoint of the fermentor run. This isolate waε grown in a fermentor under the conditionε identical to thoεe described above. The results of this experiment (Figure 11) confirmed that a mutant capable of complete assimilation of 400 g/liter glucose in about 60 hours waε indeed isolated. This experiment also shows that the xylitol-producing Z . rouxii strain can also aεsimilate xylitol when all the glucose in the culture medium is consumed.

All referenceε are incorporated herein by reference. Having now fully deεcribed the invention, it will be understood by those with skill in the art that the scope can be performed with a wide and equivalent range of concentrations, parameterε, and the like, without affecting the spirit or scope of the invention or any embodiment thereof.

SEQUENCE LISTING

(1) GENERAL INFORMATION:

(i) APPLICANT: Harkki, Anu M.

Myaεnikov, Andrey N. Apajalahti, Juha H.A. Paεtinen, Oεεi A.

(ii) TITLE OF INVENTION: Manufacture of Xylitol

(iii) NUMBER OF SEQUENCES: 8

(iv) CORRESPONDENCE ADDRESS:

(A) ADDRESSEE:

(B) STREET:

(C) CITY:

(D) STATE:

(E) COUNTRY:

(F) ZIP:

(v) COMPUTER READABLE FORM:

(A) MEDIUM TYPE: Floppy diεk

(B) COMPUTER: IBM PC compatible

(C) OPERATING SYSTEM: PC-DOS/MS-DOS

(D) SOFTWARE: Patentln Release #1.0, Version #1.25

(vi) CURRENT APPLICATION DATA:

(A) APPLICATION NUMBER: PCT (to be asεigned)

(B) FILING DATE: herewith

(C) CLASSIFICATION:

(vi) PRIOR APPLICATION DATA:

(A) APPLICATION NUMBER: US 08/110,672

(B) FILING DATE: 24-AUG-1993

(vi) PRIOR APPLICATION DATA:

(A) APPLICATION NUMBER: US 07/973,325

(B) FILING DATE: 05-NOV-1992

(viii) ATTORNEY/AGENT INFORMATION:

(A) NAME:

(B) REGISTRATION NUMBER:

(C) REFERENCE/DOCKET NUMBER:

(ix) TELECOMMUNICATION INFORMATION:

(A) TELEPHONE:

(B) TELEFAX: (2) INFORMATION FOR SEQ ID N0:1:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 28 baεe pairε

(B) TYPE: nucleic acid

(C) STRANDEDNESS: both

(D) TOPOLOGY: both

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:l

CGAATTCTAG ACCACCCTAA GTCGTCCC 28

(2) INFORMATION FOR SEQ ID NO:2:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 29 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: both

(D) TOPOLOGY: both

( i) SEQUENCE DESCRIPTION: SEQ ID NO:2

TTCAAGAATT CAAGAAACTC ACGTGATGC 29

(2) INFORMATION FOR SEQ ID NO:3:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 26 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: both

(D) TOPOLOGY: both

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:3:

CAGGCCGTCG ACAAGGATCT CGTCTC 26

(2) INFORMATION FOR SEQ ID NO:4:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 33 baεe pairε

(B) TYPE: nucleic acid

(C) STRANDEDNESS: both

(D) TOPOLOGY: both (xi) SEQUENCE DESCRIPTION: SEQ ID NO:4

AATTAGTCGA CCGTTAATTG GGGCCACTGA GGC 33

(2) INFORMATION FOR SEQ ID NO:5:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 35 baεe pairε

(B) TYPE: nucleic acid

(C) STRANDEDNESS: both

(D) TOPOLOGY: both

( i) SEQUENCE DESCRIPTION: SEQ ID NO:5:

GCGAAGCTTA AAAATGTCCA AGCAACAGAT CGGCG 35

(2) INFORMATION FOR SEQ ID NO:6:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 34 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: both

(D) TOPOLOGY: both

(xi) SEQUENCE DESCRIPTION: SEQ ID NO: 6:

GCGAAGCTTA GATTAATCCA GCCATTCGGT ATGG 34

(2) INFORMATION FOR SEQ ID NO:7:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 44 base pairs

(B) TYPE: nucleic acid

(C) STRANDEDNESS: both

(D) TOPOLOGY: both

(xi) SEQUENCE DESCRIPTION: SEQ ID NO:7:

AGCTCTAGAA ATGACTCAAT TCACTGACAT TGATAAGCTA GCCG 44 (2) INFORMATION FOR SEQ ID NO:8:

(i) SEQUENCE CHARACTERISTICS:

(A) LENGTH: 44 base pairε

(B) TYPE: nucleic acid

(C) STRANDEDNESS: both

(D) TOPOLOGY: both

(Xi) SEQUENCE DESCRIPTION: SEQ ID NO:8:

GGAGAATTCA GCTTGTCACC CTTATAGAAT GCAATGGTCT TTTG 44

Claims

What Is Claimed Is:

1. A method for the production of xylitol from a recombinant host, wherein said method comprises:

(a) constructing within a microbial host, a novel metabolic pathway, said pathway leading to the syntheεiε of xylitol aε an end product from a carbon εource other than D-xyloεe, D-xyluloεe, ixtureε of D-xyloεe and D-xyluloεe, and polymerε and oligo erε containing D-xyloεe or D-xyluloεe aε major co ponentε;

(b) growing εaid recombinant host of step (a) under conditions that provide for said synthesis of said xylitol using said pathway and on a carbon εource other than D-xyloεe, D-xylulose, mixtures of D-xylose and D-xylulose, and polymerε and oligomerε containing D-xylose or D-xylulose as major components; and.

(c) recovering said xylitol produced in step (b) .

2. The method of claim 1, wherein arabitol is an intermediate in said pathway.

3. The method of claim 2, wherein said novel metabolic pathway extends and/or modifies the metabolic pathway of the native hoεt that leadε to arabitol aε an end product in the native host. 4. The method of claim 3, wherein the construction of said novel metabolic pathway comprises transforming an arabitol-producing microbial host with a DNA encoding D-arabitol dehydrogenase (EC 1.1.1.11).

5. The method of claim 4, wherein the construction of said novel metabolic pathway further comprises transforming said recombinant hoεt constructed with a DNA encoding xylitol dehydrogenaεe (EC 1.1.1.9). 6. The method of claim 1, wherein said native host is either arabitol-producing yeast or arabitol-producing fungus.

7. The method of claim 6, wherein said host does not expreεε D-xylulokinaεe (EC 2.7.1.17).

8. The method of claim 6, wherein εaid hoεt does not express transketolaεe (EC 2.2.1.1).

9. The method of claim 6, wherein εaid hoεt is further transformed with one or more coding sequences selected from the group consisting of DNA encoding xylitol dehydrogenase, D-glucose-6-phoεphate dehydrogenaεe (EC 1.1.1.49), 6-phospho-D-gluconate dehydrogenase (EC 1.1.1.44), and D-ribulose-5-phosphate-3-epimerase (EC 5.1.3.1) . ιo. The method of any of claims 1-9, wherein said yeast is selected from the group consiεting of Z . rouxii and Candida polymorpha , Torulopεiε Candida , Pichia farinoεa , Torulaεpora hanεenii , and εaid funguε iε εelected from the group consisting of Dendryphiella εalina and Schizophyllum commune .

11. The method of claim 10, wherein εaid yeaεt is Z . rouxii .

12. The method of claim 1, wherein xylitol is formed by conversion of D-xylulose-5-phosphate to D-xylulose followed by reduction of D-xylulose to xylitol.

13. The method of claim 12 wherein εaid hoεt iε further tranεformed with a conεtruct encoding one or more enzymeε εelected from the group consisting of D-glucoεe-6- phoεphate dehydrogenaεe (EC 1.1.1.49), 6-phoεpho-D- gluconate dehydrogenase (EC 1.1.1.44), D-ribulose-5- phoεphate-3-epimeraεe (EC 5.1.3.1), D-ribulokinase (EC 2.7.1.47) and xylitol dehydrogenase (EC 1.1.1.9).

14. The method of claim 12, wherein said hoεt does not expreεε tranεketolaεe (EC 2.2.1.1). 15. The method of claim 14 wherein said host is further transformed with a construct encoding one or more enzymes selected from the group consisting of D-glucose-6- phoεphate dehydrogenase (EC 1.1.1.49), 6-phospho-D- gluconate dehydrogenase (EC 1.1.1.44), D-ribulose-5- phoεphate-3-epimeraεe (EC 5.1.3.1), D-ribulokinaεe (EC 2.7.1.47) and xylitol dehydrogenaεe (EC 1.1.1.9).

16. The method of claim 12, wherein εaid hoεt doeε not expreεε D-xylulokinaεe (EC 2.7.1.17). 17. The method of claim 16 wherein εaid host is further tranεformed with a construct encoding one or more enzymes selected from the group consisting of D-glucose-6- phoεphate dehydrogenase (EC 1.1.1.49), 6-phospho-D- gluconate dehydrogenase (EC 1.1.1.44), D-ribulose-5- phoεphate-3-epimeraεe (EC 5.1.3.1), D-ribulokinaεe (EC 2.7.1.47) and xylitol dehydrogenaεe (EC 1.1.1.9).

18. The method of claim 12, wherein εaid host does not expresε tranεketolaεe (EC 2.2.1.1) and D-xylulokinase (EC 2.7.1.17) . 19. The method of claim 18 wherein said host is further transformed with a construct encoding one or more enzymes selected from the group consisting of D-glucose-6- phoεphate dehydrogenase (EC 1.1.1.49), 6-phospho-D- gluconate dehydrogenase (EC 1.1.1.44), D-ribulose-5- phosphate-3-epimerase (EC 5.1.3.1), D-ribulokinase (EC 2.7.1.47) and xylitol dehydrogenase (EC 1.1.1.9).

20. The method of any of claims 12-19, wherein said yeast is selected from the group consisting of Zygoεaccharomyceε rouxii , Candida polymorpha , Torulopεiε Candida , Pichia farinoεa , Torulaεpora hanεenii , and εaid fungus is selected from the group conεiεting of Dendryphiella εalina and Schizophyllum commune .

21. The method of claim 20, wherein said yeast is Z . rouxii . 22. A recombinant microbial hoεt, εaid hoεt being capable of εynthesizing xylitol in a single fermentation from a carbon source other than D-xylose, D-xylulose, mixtures of D-xylose and D-xyluloεe, or polymers or oligomers containing D-xylose and/or D-xylulose as major components, said syntheεiε being greater than that of the correεponding non-recombinant microbial hoεt.

23. The recombinant host of claim 22, wherein the native metabolic pathway leading to arabitol as an end product in the non-recombinant host has been extended or modified in a manner that increaseε εaid εyntheεiε of said xylitol from said recombinant hoεt.

24. The recombinant hoεt of claim 23, wherein said pathway has been extended or modified by the transformation of DNA encoding D-arabitol dehydrogenase (EC 1.1.1.11) into said host.

25. The recombinant host of claim 24, wherein said pathway has been extended or modified by the transformation of DNA encoding xylitol dehydrogenase (EC 1.1.1.9) into said host.

26. The recombinant host of claim 22, wherein said host does not express D-xylulokinase (EC 2.7.1.17).

27. The recombinant host of claim 22, wherein said host does not expresε tranεketolaεe (EC 2.2.1.1). 28. The recombinant hoεt of claim 22, wherein εaid hoεt does not express D-xylulokinase (EC 2.7.1.17) or transketolaεe (EC 2.2.1.1).

29. The recombinant host of claim 22, wherein said hoεt iε tranεformed with a gene encoding xylitol dehydrogenase.

30. The recombinant host of claim 22, wherein εaid host is transformed with a gene encoding D-glucoεe-6- phoεphate dehydrogenaεe (EC 1.1.1.49). 31. The recombinant hoεt of claim 22, wherein εaid host is transformed with a gene encoding 6-phospho-D- gluconate dehydrogenase (EC 1.1.1.44).

32. The recombinant host of claim 22, wherein said host is tranεformed with a gene encoding D-ribulose-5- phoεphate-3-epimeraεe (EC 5.1.3.1).

33. The recombinant hoεt of claim 22, wherein said host iε tranεformed with a conεtruct encoding xylitol dehydrogenaεe. 34. The recombinant hoεt of any of claims 22-33, wherein said non-recombinant microbial host is an arabitol- producing yeaεt or an arabitol-producing funguε.

35. The recombinant host of claim 34, wherein said yeast is selected from the group consisting of Zygosaccharomyceε rouxii , Candida polymorpha , Torulopεiε Candida , Pichia farinoεa , Torulaεpora hanεenii , and said fungus is selected from Dendryphiella εalina and Schizophyllum commune .

36. The recombinant hoεt of claim 35, wherein said yeast is Z . rouxii .