signature=59574dfb91a8e544c2d3dc0e03300a1e,TARGETED GENE DELETIONS FOR POLYSACCHARIDE SLIME FORMERS

CROSS REFERENCES TO RELATED APPLICATIONS

This is a divisional application of U.S. application Ser. No. 11/347,341 filed Feb. 3, 2006, allowed, which claims the benefit of U.S. Provisional Application No. 60/649,559, filed Feb. 4, 2005. The prior applications are incorporated herein by reference in their entirety.

FIELD OF THE INVENTION

This invention is related to the area of sphingan polysaccharide production. In particular, it relates to site-directed genetic methods for improving sphingan-producing strains.

BACKGROUND OF THE INVENTION

Sphingomonasstrains, such as ATCC 53159 and ATCC 31461, produce copious amounts of capsular polysaccharide. While under some conditions polysaccharide may be released from the cell [5, 6], during growth with abundant carbon source as in fermentation, the polysaccharide is firmly attached to the cell surface. Attempts to increase productivity of fermentations for diutan and gellan may be limited by the capsular nature of the polysaccharide, which may impair uptake of nutrients. Also, if there are a limited number of sites for biosynthesis of the polysaccharide, there may be a maximum amount of polysaccharide that can be produced by each cell. The polysaccharide gellan has been observed to be involved in cell clumping since mutants that do not make any polysaccharide grow uniformly in suspension [3]. These cell clumps may interfere with techniques such as determination of cell number by optical density, centrifugation of cells, e.g., for isolation of DNA or protein, and separation or lysis of cells for polysaccharide purification.

The mechanism of attachment and the genes involved in attachment of polysaccharide to the cell surface in Sphingomonads have not been previously determined. Induced mutants of Sphingomonasstrains ATCC 31461, ATCC 31555, ATCC 31554, and ATCC 21423 that produce polysaccharide in a slime form have been isolated, but the genes mutated were not determined, and the methods of inducing and selecting the mutants were not disclosed [10]. Genes for biosynthesis of gellan [3, 8], diutan [1] and sphingan S-88 [9] have been isolated. The functions of many of these genes were assigned by biochemical tests or by homology to genes of known functions in databases such as GenBank. For example, genes have been identified that are involved in assembly of the tetrasaccharide repeat unit [7, 8], and in synthesis of the precursor dTDP-L-rhamnose [3, 9]. It would be expected that genes affecting only attachment of polysaccharide to the cell surface would still have the polysaccharide producing phenotype (i.e., mucoid colonies on solid media and viscous broth).

A cluster of 18 genes for gellan biosynthesis spanning 21 kb was described, in addition to four genes for gellan synthesis not in the cluster [3]. The DNA sequences were deposited in GenBank in June 2003 (Accession number AY217008). Among the genes in the cluster were gelM, gelN, and gelI. A deletion of most of adjacent genes gelM and gelN was constructed. The gelI gene was inactivated by an insertion. The gelM-gelN deletion strain and the gelI mutant were shown to produce somewhat reduced amounts of gellan and more fluid broths, and the gellan produced was shown to have the same composition as gellan from the wild-type strain. The attachment of the polysaccharide to the cell was not reported

The Sphingomonas elodeagelR, gelS, and gelG genes appear to be in an operon in the same order as in the S-88 sps gene cluster, but not adjacent to the genes in the cluster of 18 genes [3]. The GelR protein was somewhat smaller than its S-88 homolog (659 vs. 670 amino acids) with 49% identity, and had homology to surface layer proteins and other membrane proteins. The DNA sequences of gelR, gelS and gelG genes were deposited in GenBank in June 2003 (Accession number AY220099). No mutation in gelR was constructed in this report [3]. Yamazaki et al. report that strains with mutations in gene spsR were still mucoid, indicating that they produce polysaccharide, but the polysaccharide was not characterized as to rheology or attachment to the cell [9, 12].

Yamazaki described classical mutants of four Sphingomonasthat produce polysaccharide as slime rather than attached to the cell [10]. Yamazaki did not describe how to screen mutagenized cultures for the slime phenotype. Yamazaki did not identify which gene or genes were mutated.

Sa-Correia reviewed work done on isolation of genes for gellan synthesis [8]. Sa-Correia described partial sequencing of some genes including urf32 and urf26 (equivalent to gelM and gelN described in Harding et al. [3]). The complete sequences of these genes were deposited in GenBank in April 2003 (GenBank Accession number AY242074). No function of these genes is reported. In the GenBank submission, genes urf32 and urf26 were merely designated as putative membrane protein and putative exported protein, respectively. No sequence for gelI or gel R was deposited.

Coleman describes the isolation of genes for diutan biosynthesis and investigation of some gene functions [1]. The dpsM and dpsN genes, which were designated by Coleman as orf3 and orf4, were described, but functions were not indicated.

A cluster of genes for biosynthesis of the S-88 polysaccharide from Sphingomonasstrain ATCC 31554 was described [9, 12]. The functions of genes urf32 and urf26 (homologs of dpsM, gelM and dpsN, gelN), and spsl (homolog of gelI, dpsI) were not described. Gene spsR (homolog of gelR, dpsR) was described as encoding a protein remotely similar to bacterial and fungal polysaccharide lyases. The DNA sequences were deposited in GenBank (Accession number U51197).

There is a continuing need in the art to improve methods of making industrially useful sphingans and the properties of the sphingans.

SUMMARY OF THE INVENTION

According to one embodiment of the invention, a method is provided of making a bacterium. The bacterium is of the genus Sphingomonasand comprises a mutation in one or more genes selected from the group consisting of genes M, N, I, or R of the sphingan polysaccharide biosynthetic gene cluster. The M and N genes are also referred to in some publications as genes urf32, urf26, respectively, for unknown reading frame [8, 9]. A segment of genomic DNA of a first bacterium of the genus Sphingomonasis isolated. The segment comprises all or part of genes M and/or N, or I, or R of the sphingan polysaccharide biosynthetic gene cluster. A mutation in the segment is induced to form a mutated segment. The mutated segment is introduced into a second bacterium of the genus Sphingomonas. The second bacterium comprises wild-type genes M and/or N, or I, or R of the sphingan polysaccharide biosynthetic gene cluster. A progeny of the second bacterium in which the mutated segment has integrated in the genome and replaced wild-type genes M and/or N, or I, or R of the sphingan polysaccharide biosynthetic gene cluster of the second-bacterium isolated. The Sphingomonasbacterium may or may not be S. elodea.

According to another embodiment of the invention, another method is provided of making a bacterium of the genus Sphingomonaswhich comprises a mutation in one or more genes selected from the group consisting of genes M, N, I, and R of the sphingan polysaccharide biosynthetic gene cluster. Two non-contiguous segments of genomic DNA of a first bacterium of the genus Sphingomonasare isolated. The segments flank or include genes M and N of the sphingan polysaccharide biosynthetic gene cluster. Similarly, segments flanking gene I or gene R can be isolated. The two non-contiguous segments are ligated together. The ligated non-contiguous segments are introduced into a second bacterium of the genus Sphingomonas. The second bacterium comprises wild-type genes M and/or N, or I, or R of a sphingan polysaccharide biosynthetic gene cluster. A progeny of the second bacterium in which the ligated segment has integrated in the genome and replaced wild-type genes M and/or N, or I, or R of a sphingan polysaccharide biosynthetic gene cluster of the second bacterium is isolated. The Sphingomonasbacterium may or may not be S. elodea.

According to yet another embodiment of the invention, a composition is provided. The composition comprises a native gellan polysaccharide with gel strength greater than that of an equivalent weight of native gellan from a capsular strain.

According to yet another embodiment of the invention, a composition is provided. The composition comprises a diutan polysaccharide which imparts to a fluid an increased viscosity relative to an equivalent weight of diutan produced by strain ATCC 53159.

According to another embodiment of the invention, an isolated and purified bacterium of the genus Sphingomonasis provided. The bacterium comprises a deletion in one or more genes selected from the group consisting of genes M, N, I, and R of the sphingan polysaccharide biosynthetic gene cluster. The bacterium can be cultured in a culture medium under conditions suitable for producing sphingan polysaccharide to produce sphingan polysaccharide in the culture medium. The culture broth of the bacterium can be used directly as a viscosifier or gelling agent, or after precipitation with alcohol. Alternatively, the culture broth can be subjected to a procedure to remove bacteria from the culture broth prior to use as a viscosifier or gelling agent or recovery from the broth. The Sphingomonasbacterium may or may not be S. elodea.

These and other embodiments which will be apparent to those of skill in the art upon reading the specification provide the art with new methods, strains, and compositions for making viscosifiers and gelling agents.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1. Comparison of gene clusters for polysaccharide biosynthesis in Sphingomonasstrains ATCC 31554, ATCC 31461 and ATCC 53159.

FIG. 2. Slime forming characteristics of S60WTC gelM-gelN mutants

FIG. 3. Slime forming characteristics of S60WTC gelN and gelI Mutants

FIG. 4. Sequence of DNA at the site of deletion of dpsN (SEQ ID NO: 19), and amino acid sequence of the fusion peptide (SEQ ID NO: 20).

FIG. 5A-5C. Slime forming characteristics of dpsN mutants

FIG. 6A-6B. Slime forming characteristics of dpsM mutants

DETAILED DESCRIPTION OF THE INVENTION

Genes have been identified that control the attachment of polysaccharide to bacterial cells in two Sphingomonasstrains. Deletion of either one or both genes gelM (dpsM) and gelN (dpsN) or inactivation of geld results in polysaccharide being released into the medium as slime rather than attached to the cell surface as capsular polysaccharide. Formation of slime form of polysaccharide eases handling of bacterial cultures, improves mixing during fermentation, may increase production, and in some cases improves rheology of the polysaccharide. Site directed mutagenesis is advantageous over random mutagenesis and screening for slime-forming mutants for many reasons, including speed and avoidance of unrelated mutations. Inactivation of the gene gelR was found to improve the rheology (gel strength) of the slime form of gellan polysaccharide.

Orthologs of dpsM, dpsN, gelM, gelN, and gelI can be inactivated in any Sphingomonasstrain to obtain the slime-forming phenotype. Orthologs of gelR can be inactivated to prevent degradation of the polysaccharide resulting in improved rheology. Suitable Sphingomonads include without limitation those which make rhamsan (ATCC 31961), welan (ATCC 31555), gellan (ATCC 31461), and diutan (ATCC 53159) and strains making polysaccharides S7 (ATCC 21423), S88 (ATCC 31554), S198 (ATCC 31853) and NW 11 (ATCC 53272). The ATCC numbers refer to the deposit numbers of the strains at the American Type Culture Collection. These are exemplified by S. elodeaATCC 31461 and Sphingomonassp. ATCC 53159, but other strains can be used. Suitable Sphingomonads which can be used include Sphingomonas adhaesiva, Sphingomonas aerolata Sphingomonas alaskensis, Sphingomonas aquatilis, Sphingomonas aromaticivorans, Sphingomonas asaccharolytica, Sphingomonas auraniaca, Sphingomonas capsulata, Sphingomonas chlorophenolica, Sphingomonas chungbukensis, Sphingomonas cloacae, Sphingomonas echinoides, Sphingomonas elodea, Sphingomonas faeni, Sphingomonas herbicidovorans, Sphingomonas koreensis, Sphingomonas macrogoliabidus, Sphingomonas malii, Sphingomonas melonis, Sphingomonas natatoria, Sphingomonas parapaucimobilis, Sphingomonas paucimobilis, Sphingomonas piluitosa, Sphingomonas prune, Sphingomonas rasa, Sphingomonas sanguinis, Sphingomonassp., Sphingonas stygia, Sphingomonas subarctica, Sphingomonas suberifaciens, Sphingomonas subterranea, Sphingomonas taefonensis, Sphingomonas terrae, Sphingomonas trueperi, Sphingomonas ursincola, Sphingomonas wittichii, Sphingomonas xenophaga, Sphingomonas yabuuchiae, and Sphingomonas yanoikuyae. Orthologs can be identified on the basis of gene location and organization in a sphingan biosynthetic gene cluster, on the basis of overall homology, and/or on the basis of domain homology. Typically, the level of overall homology will be greater than 44%, often greater than 55%, 66%, 77%, 88%, or 98% with one of the dpsM, dpsN, gelM, gelN, gelI , or gelR genes. An ortholog desirably has homology greater than 80% with at least one of these four genes.

Site directed mutagenesis can be used to make mutations in a desired known target gene or genomic region. This eliminates the trial-and-error nature of random induced mutagenesis or spontaneous mutagenesis. Formation of deletions insures that the mutations will not revert, as is possible with point (substitution) mutations and insertion mutations, for example. Deletions also have the benefit of not employing exogenous DNA, such as drug resistance markers or other environmentally undesirable markers.

An isolated segment of genomic DNA comprising the M and/or N, I, or R of the sphingan biosynthetic gene cluster or flanking DNA is DNA that is not connected to genomic DNA to which it is normally attached. Isolated DNA can be obtained by purification from natural sources, by synthesis, or by amplification, as non-limiting examples. The isolated DNA will typically be on a fragment of DNA in vitro, but isolated DNA could also be on a vector, such as a plasmid or transducing phage, which contains the desired portion of the Sphingomonasgenome. Flanking DNA is typically from the genomic regions immediately adjacent to the M and/or N, I, or R within about 500 bp of the genes, or within about 1-2 kb of the genes.

Any method known in the art can be used to introduce a mutation into an isolated segment comprising all or part of genes M and/or N, I, or R of the sphingan biosynthetic gene cluster. A deletion can be introduced using restriction endonucleases, for example, and rejoining formerly non-contiguous nucleotides. A deletion can be formed by amplifying and joining two non-contiguous segments of the genes or two non-contiguous segments of DNA flanking the target gene. An insertion can be made in an isolated segment using endonuclease digestion and ligation. Chemical mutagenesis can be used on an isolated segment of genomic DNA. Any mutagenesis method can be selected and used according to the particular circumstances.

After mutations have been induced, the segment of genomic DNA can be reintroduced into a recipient bacterium. Typically, but not necessarily, the recipient will be of the same species as the donor of the segment. Any method known in the art for introducing exogenous DNA into a bacterium can be used. Suitable methods include without limitation electroporation, conjugation, spheroplast formation, calcium chloride precipitation and transformation, liposomes, and viral transduction. Any nucleic acid introduction method can be selected and used according to the particular circumstances.

If the segment of mutated genomic DNA introduced into the recipient bacterium does not have a means of replicating itself, then it must integrate into a replicon in the recipient bacterium in order to be maintained. Typically such an integration event will integrate the entire incoming plasmid. One can detect a marker on the introduced DNA to identify that the DNA has integrated. In order to detect resolution of the integrate, one can screen or select for loss of a marker on the introduced DNA. Suitable markers for accomplishing this are known in the art, and any can be used as the circumstances dictate. To determine the isolates in which the introduced version of the sphingan genes replaces the wild-type version in the recipient, the size or sequence of the DNA can be determined, for example, by PCR.

As demonstrated below, the slime form of sphingan produced for example by a sphingan biosynthetic gene cluster gene M and/or N, mutant may have improved rheological properties over the form which is attached to bacterial cells. Such improved rheological properties are reflected in the ability of the same weight of material to provide more viscosifying power. Such improvement may be modest, such as at least 5% 10%, 15%, 20% or 25%, or it can be more substantial, with an improvement of at least 30%, 40%, 50%, 60%, 70%, 80%, or 90% relative to the sphingan produced by the capsule-forming parent. Rheologically properties can be measured using any technique which is known in the art. Suitable techniques include without limitation the measurement of Low Shear Rate Viscosity (“LSRV”) in tap water solutions and the measurement of Sea Water Viscosity (“SWV”) in high salt solutions.

The slime form of gellan, produced, for example, by a gelN mutant in combination with a mutation in the putative lyase gene, gelR, results in formation of gellan of high gel strength. The gel strength will typically be greater than 1000, whereas the capsular strain typically produces a gellan with gel strength of 700-900, but less than 1000.

Purified bacteria according to the present invention are those which have been microbiologically purified, for example using liquid dilution techniques or streaking on solid media to form single colonies. Standard techniques known in the art of microbiology can be used for this purpose.

Mutants according to the present invention can be cultured and grown using the same or similar techniques as used for the parental strains. Liquid culture properties of the mutants may be improved, permitting increased aeration and mixing. The culture broth of the mutant may also provide more efficient recovery than with the attached form of polysaccharide. In addition, the mutants may also provide a product with improved clarity relative to the attached form of polysaccharide. Bacteria may optionally be removed from the polysaccharide produced by the mutant by filtration, centrifugation, or by sedimentation. The culture broth can be chemically, enzymatically, or thermally (hot or cold) treated before or after bacteria removal, as desired.

The genes from S. elodeaATCC 31461 involved in gellan attachment to the cell surface are gelM and gelN (FIG. 1; SEQ ID NO: 13) and gelI (FIG. 1, SEQ ID NO: 25). A strain has been constructed that has a deletion of most of genes gelM and gelN, resulting in the slime-forming phenotype. A specific deletion of gelN has also been constructed, and an insertion in gene gelI. Both of these mutations result in the slime-forming phenotype. The coding sequences of gelM and gelN are at nucleotides 501-1382 and 1366-2064, respectively, in SEQ ID NO: 13. The encoded amino acid sequences are shown in SEQ ID NOs: 16 and 15, respectively. The coding sequences of gelI is at nucleotides 501 to 1403, respectively, in SEQ ID NO: 25. The encoded amino acid sequences are shown in SEQ ID NO: 26. A deletion of gene gelR was found to result in improved gel strength for gellan in the slime form. The coding sequences of gelR is at nucleotides 478 to 2457, respectively, in SEQ ID NO: 27. The encoded amino acid sequences are shown in SEQ ID NO: 28.

The genes from Sphingomonassp. ATCC 53159 involved in diutan attachment to the cell surface are dpsM and dpsN (FIG. 1; SEQ ID NO: 14), and presumably dpsI based on homology to gen. Deletions of each of genes dpsM and dpsN have been constructed and both result in the slime-forming phenotype. The coding sequences of dpsM and dpsN are at nucleotides 456-1337 and 1321-2019, respectively, in SEQ ID NO: 14. The encoded amino acid sequences are shown in SEQ ID NOs: 18 and 17, respectively.

It will be apparent to those skilled in the art that the same or similar methods used for gellan synthesis may also he used for diutan synthesis. Thus, mutations in genes dpsI and dpsR could readily be constructed. The coding sequences of dpsI is at nucleotides 501-1472, respectively, in SEQ ID NO: 29. The encoded amino acid sequences are shown in SEQ ID NO: 30. The coding sequences of dpsR is at nucleotides 501-2498, respectively, in SEQ ID NO: 31. The encoded amino acid sequences are shown in SEQ ED NO: 32.

The above disclosure generally describes the present invention. All references disclosed herein are expressly incorporated by reference. A more complete understanding can be obtained by reference to the following specific examples which are provided herein for purposes of illustration only, and are not intended to limit the scope of the invention.

EXAMPLES

Example 1

Production of Gellan Slime-Forming Mutants

For construction of mutants of Sphingomonas elodea, a derivative of ATCC 31461 designated S60 wtc1was used, which has improved uptake of DNA. This strain can be readily made by one skilled in the art. PCR amplification was used to amplify regions flanking the gelM-gelN genes [3]. The amplified fragments were cloned into the pLO2 vector and introduced into S60 wtc by conjugation to replace the gelM and gelN genes on the genome with the deletion, by double crossover homologous recombination. Vector pLO2 does not replicate in S60 wtc, so initial selection for kanamycin resistance selects for those colonies in which the plasmid has integrated into the chromosome by homologous recombination. The vector also contains the gene sacB. This gene confers sensitivity to sucrose. Thus, selection on sucrose can be used to detect isolates that have lost the plasmid and retain one copy of the deletion or wild-type genes.

S. elodeaATCC 31461 has a low efficiency of uptake of DNA, particularly large plasmids (about 10−7). Spontaneous mutants of ATCC 31461 with increased efficiency of DNA uptake were isolated. It was suspected that those few cells that were successful recipients of a plasmid, such as the broad-host-range plasmid pLAFR3, represented mutants in the recipient population with an increased ability to take up this plasmid DNA. To allow loss of the plasmid, three transconjugants containing pLAFR3 were grown under nonselective conditions (i.e., without tetracycline antibiotic) with serial passage for about 30 generations. Three independent plasmid-cured strains (i.e., tetracycline-sensitive derivatives from each of the initial transconjugants) were tested and all three exhibited increased conjugation frequency (4.2×10−3, 0.6×10−2, and 1.5×10−2), representing a 105-fold increase compared to the wild-type strain. This increased conjugation frequency was stable and reproducible. One of these strains was designated S60 wtc [3].

A plasmid containing the gelM-gelN deleted region was introduced into S60 wtc by tri-parental conjugal mating, using pRK2013 to provide transfer functions, and transconjugants selected on YM-Sm (25 ug/ml)-Km (7.5 ug/ml) medium. Streptomycin prevents growth of the E colistrains. Kanamycin resistant plasmid integrants were isolated. Sucrose sensitivity was used to select for a second recombination event which eliminated the vector. Five isolates were passed two times under non-selective conditions, i.e., without antibiotic. Aliquots were then plated on medium with 8% sucrose. Sucrose resistant colonies were isolated and tested for kanamycin sensitivity. Genomic DNA was isolated and PCR was used to determine which Kms isolates had retained the deletion. An amplified fragment of the expected size for a deletion resulted from the genomic DNA from four strains. These four deletion strains were purified on YM medium. All four strains appeared less mucoid, softer, flatter and darker yellow than the wild type.

Example 2

Characterization of GELM-GELN Deletion Strains

The gelM-gelN deletion isolates were evaluated in shake flask fermentations. The ΔgelM-gelN culture broth was fluid and smooth compared to the more solid, viscous S60 wtc broth. Precipitation with isopropyl alcohol produced longer, thicker, fibers from the mutant strains compared to S60 wtc fibers. However, the deletion mutants had 22% reduction in yield of total precipitable material and produced only 30% of the broth viscosity of wild-type. The gellan produced had a normal composition of sugars, and glyceryl and acetyl substituents.

The mutants were evaluated for slime forming characteristics, using several techniques, including microscopic evaluation, cell clumping, cell pellet formation, and hot settling test, as shown in FIG. 2.

The hot settling test consisted of heating the gellan broth in the autoclave for ten minutes to melt the gellan, then transferring the hot broth to a large test tube and incubating overnight at 95° C. (to maintain broth as liquid). With a capsular strain, the cells are attached to the polysaccharide and remain suspended. For slime-formers, the cells are not attached and settle during overnight incubation. The gelM-gelN deletion strains were shown to be slime formers by this test.

For the centrifugation test, the strains were grown overnight in DM2 media containing 1% glucose and centrifuged at maximum speed in the Eppendorf centrifuge. Inactivation of gelM-N genes results in complete loss of attachment of the polysaccharide from the cell surface such that the cells can be pelleted by centrifugation.

By microscopic evaluation, most of the S60 wtcΔgelM-N cells are free and motile, whereas the S60 wtc are in the gum matrix. In cell culture, the S60 wtcΔgelM-N cells grow in suspension, whereas S60 wtc cells form clumps.

Example 3

Construction of Gellan Slime-Forming Mutants

A deletion was constructed of gelN for gellan biosynthesis. PCR primers were designed to amplify DNA fragments upstream (500 bp) and downstream (401 bp) of the gelN gene [3]. Primers used are shown in Table 1.

TABLE 1Primers for construction of gelN deletion mutant.PrimerSequencePurpose

SacI-GelN primer 15′ TGGAGCTC-GGTGCTGTGGTTGTTCTT 3′Amplifies 500 by upstream

(SEQ ID NO: 1)of gelN

XbaI-GelN primer 25′ GGTCTAGA-GTCAGGCCGGTTGAACAT 3′

(SEQ ID NO: 2)

XbaI-GelN primer 35′ AGTCTAGA-GCCTGAACGCCGAAAGGG 3′Amplifies 401 bp

(SEQ ID NO: 3)downstream of gelN

SphI-COlN primer 45′ CTTCCATGC-GGTGATGGTGGAGAATGG 3′

(SEQ ID NO: 4)

Primers SacI-GelN primer] and XbaI-GelN primer2 were used to amplify a 500 bp fragment from the gelM gene as a SacI-XbaI fragment (total 516 bp). Primers XbaI-GelN primer3 and SphI-GelN primer4 were used to amplify a 401 by fragment from the atrD gene as an XbaI-SphI fragment (total 418 bp). Since the end of the gelM gene overlaps the start of the gelN gene by 17 bp, the stop codon of gelM and the start codon of gelN were retained, as well as the natural stop codon of gelN. The PCR fragments were ligated sequentially into the polylinker of plasmid vector pLO2 [4], resulting in clone pLO2-gelNdeln#1 carrying the deletion of gelN.

Plasmid pLO2-gelNdeln#1 was then used to transfer the deletion to strain S60 wtc3by conjugation and homologous recombination. Strain S60 wtc is a strain derived from ATCC 31461 as a spontaneous mutant with increased ability to take up plasmid DNA [2]. A chromosomal integrant was selected by kanamycin resistance. Subsequent growth for approximately 30 generations in the absence of antibiotic allowed for excision of the plasmid. Recombinants that had lost the plasmid were then selected by sucrose (8%) tolerance, due to loss of the plasmid-encoded sacB gene, and then colonies were screened for kanamycin sensitivity. The sacB gene encodes an enzyme levansucrase for synthesis of levan from sucrose. Levan is toxic to the cells. Cells that have lost the sacB gene can grow on sucrose. The sucrose tolerant isolates can be wild-type or deletion. Genomic DNA was prepared from several isolates to identify those isolates that had retained the gelN deletion versus the wild-type gene, as determined by PCR. Isolates with the gelN deletion had softer, more watery colonies compared to the hard colonies of the wild-type gelN+ isolates (See above discussion regarding use of mutant with increased ability to take up DNA).

Example 4

Characterization of GELN Deletion Mutants

The gelN deletion mutants had similar properties to the gelM-gelN deletion mutants. Cells were readily pelleted by centrifugation. In cell culture, the gelN deletion mutants grew in suspension, whereas the wild-type cells formed clumps. Thus, inactivation of the gelN gene can result in the slime phenotype as shown in FIG. 3.

Five individual isolates of gelN deletion mutants were evaluated in shake flask fermentations. The average yield (total precipitable material, TPM) for the gelN mutants (1.10 g/100 ml) was comparable to that of the S60 wtc control (1.08 g/100 ml). Selected gelN mutants were evaluated in 20 L Applikon fermentors using media containing organic and inorganic nitrogen, salts and corn syrup. Gellan polysaccharide was precipitated with isopropyl alcohol, dried and weighed. Average yield of total precipitable material for the mutants was 94% of that of the wild-type control, however, the broth viscosity was decreased by about 40%. This decrease in broth viscosity facilitates mixing in the fermentors.

TABLE 2Fermentation characteristics of gelN mutantsTPMBroth Viscosity

Straing/LCP

Wild-type12.805500

gelN #112.043850

gelN #212.152700

gelN #411.803400

average12.003317

Viscosity was measured in the Brookfield LVF viscometer using the No. 4 spindle at 60 rpm.

Example 5

Construction of Mutants Producing Gellan Slime with Improved Quality

The slime mutants of gellan had lower broth viscosity, as described in Example 4, which facilitates mixing in the fermentors. Gellan polysaccharide forms a gel after heating and cooling, and gellan is used in various food applications due to its unique textural and rheological properties. Therefore, the gel strength of the gellan produced by the slime mutants was evaluated. The gel strength is determined by the break or fracture of a prepared gel surface with a plunger.

The gellan fermentation broth was adjusted to pH 4.6 (to prevent deacylation) and pasteurized by heating to about 100° C. for several minutes. Gellan product was precipitated by addition of three times volume of isopropyl alcohol, and the fibers dried at 60° C. for two hours and milled.

A calcium solution was prepared by adding 2 ml of a 0.3 M CaCl2.2H2O stock solution to 295 ml of deionized water in a tared 600 ml stainless steel beaker. While stirring the solution at 700 rpm, 3.0 g of gellan product was added and allowed to disperse for 1 to 2 minutes. The beaker was placed into a preheated water bath at 94-95° C. for four minutes, covered and heated for 15 minutes, then stirred for 3 minutes. Solution weight was adjusted to 300 g with heated deionized water, mixed and then left standing at 94-95° C. Solution was transferred into six ring molds (0.5 inch height, 1.375 inch outer diameter, 0.125 inch wall thickness) and covered with a plastic cover plate and allowed to cool at room temperature (20-21° C.) for 20 to 24 hours. The disc was removed from the mold onto a plexiglass plate. Gel strength, or force to break (recorded in g/cm2) was determined in a TA-TX2 Texture Analyzer with a Kobe plunger (TA-19) at 1.0 mm/s.

Gellan from the slime mutants had lower gel strength than that from the wild-type capsular stain, as shown in Table 4. This result was in contrast to mutants of ATCC 53159 that produce the slime form of diutan, which had improved rheology as described in Example I 1. It was considered possible that the slime form of gellan may be degraded by a gellan lyase enzyme, produced by S. elodea. Therefore, a strain was constructed that has a deletion of a gene, gelR, which produces a protein with homology to polysaccharide degrading proteins, e.g. lyases.

The gelR deletion was constructed in S. elodeastrain S60 wtc and strain GBAD-1 [11]. PCR primers were designed to amplify DNA fragments upstream (502 bp) and downstream (435 bp) of the gelR gene. PCR primers used are shown in Table 3.

TABLE 3Primers used for construction of gelR deletionPrimerSequencePurpose

SacI-GelR primer 15′ ACGAGCTCAGATCAGCCGCAACCTCCT 3′Amplifies 486 bp

(Seq ID No: 21)upstream of gelR

XbaI-GelR primer 25′ GCTCTAGA-CGCCGCCATGTTAATCACC 3′

(Seq ID No: 22)

XbaI-GelR primer 35′ GCTCTAGA-GATGCGTTCCACGCCTGAC 3′Amplifies 419 bp

(Seq ID No: 23)downstream of gelR

SphI-GelR primer 45′ ATGCATGC-CGATCGCGCTCATCAGGGT 3′

(Seq ID No: 24)

Primers SacI-GelR primer I and XbaI-Ge1R primer 2 were used to amplify a 498 bp fragment upstream of gelR as a SacI-XbaI fragment (total 502 bp). Primers XbaI-GelR primer 3 and SphI-GelR primer 4 were used to amplify a 419 by fragment downstream of gelR as an XbaI-SphI fragment (total 435 bp). The PCR fragments were digested with restriction enzymes and ligated sequentially into the polylinker of plasmid vector pLO2 [4] resulting in clone pLO2-gelRdeletn#4, carrying the deletion of gelR.

Plasmid pLO2-gelRdeletn#4, which cannot replicate in Sphingomonas, was transferred into S. elodeastrains S60 wtc and GBAD-1 by conjugation from E. coliDH5α, using helper plasmid pRK2013 that supplies transfer functions [2]. Chromosomal integrants were selected by kanamycin resistance on yeast extract-malt extract (YM) medium with kanamycin and streptomycin (to counterselect E. coli). Subsequent growth of the Sphingomonasintegrants for approximately 30 generations in the absence of antibiotic allowed for the excision of the plasmid. Recombinants that had lost the plasmid were selected by sucrose tolerance due to loss of the plasmid encoded sacB gene, and colonies screened for kanamycin sensitivity. PCR was used to test which isolates had retained the gelR deletion.

The gelN deletion was than transferred into the gelR deletion mutant of the GBAD-1 strain as described above in Example 3. Plasmid pLO2-gelNdeln#1 was used to transfer the gelN deletion into GBAD gelR by conjugation and homologous recombination. A chromosomal integrant was selected by kanamycin resistance on YM agar with Km (20 ug/ml) and Sm (25 ug/ml). Subsequent growth in the absence of antibiotic allowed for excision of the plasmid. Recombinants that had lost the plasmid were selected by sucrose tolerance due to loss of the plasmid-encoded sacB gene, and then colonies were screened for kanamycin sensitivity.

The gelR deletion mutants exhibited different colony morphology than the wild-type strains. The gelR deletion strains had smaller rough gummy colonies compared to larger smooth gummy colonies with transparent edges for the gelR+ of S60 wtc or GBAD-1. The gelN-gelR deletion mutants had colony morphology similar to the gelN slime mutants.

Example 6

Characterization of GELN-GELR Mutants

These strains were evaluated in 20 L Applikon fermentors using media containing organic and inorganic nitrogen, salts and corn syrup. Gellan polysaccharide was precipitated with isopropyl alcohol, dried and weighed. Gel strength was determined by the method described in Example 5. The GBAD-1 gelN-gelR strain produced gellan of higher gel strength than the gellan produced from the gelN slime mutants or the wild-type capsular strains.

TABLE 4Rheological characterization of gellan from mutantsAver TPMAver.

% ofBroth ViscAver

StrainPhenotypen =wild-typecPGel strength

S60wtccapsule3—7292411

GRAD1capsule1914600629

GelN mutantsslime7932900132

GBAD1gelRgelNslime39650831447

Example 7

Construction of GELI Mutant

An insertion mutation in gene gelI was constructed. PCR primers were designed to amplify an internal fragment of the gelI gene [3]. The amplified fragment was cloned into the pLO2 plasmid vector and introduced into S60 wtc by conjugation, selecting on YM-Sm (25 μg/ml)-Km (7.5 μg/ml) medium. Selection for kanamycin resistance selects for those transconjugants that have the plasmid inserted by homologous recombination into the gelI gene, thus inactivating this gene. The gelI mutant had altered colony morphology, similar to that of the gelM-gelN and the gelN deletion strains, i.e. mucoid but softer colonies.

Example 8

Characerization of GelI Mutant Strain

The gel! mutant was evaluated in shake flask fermentation. The mutant had less viscous broth compared to the wild-type strain and about a 20% reduction in yield of total precipitable material. The gellan produced had a normal composition of sugars and glyceryl and acetyl substituents [3].

The gelI mutant was evaluated for slime forming characteristic using several techniques including microscopic evaluation, cell clumping, cell pellet formation and hot settling test as described above. The gelI insertion mutant had similar characteristics to the gelM-gelN and the gelN deletion mutants. Microscopic evaluation showed that the cells were free and motile. In cell culture, the gelI mutants grew in suspension rather than clumps. Cells were readily pelleted from DM2medium by centrifugation. Cells also settled well in the hot settling test. Thus, the mutation in the gelI gene also results in the slime phenotype, as shown in FIG. 3.

Example 9

Production of Diutan Slime-Forming Mutants

Sphingomonassp. ATCC 53159 (S-657) produces a polysaccharide (diutan) with a structure similar to that of gellan (i.e., it has a glucose-glucuronic acid-glucose-rhamnose repeat unit), but with a side chain of two rhamnose residues attached to one glucose residue. Diutan has two acetyl substituents, and lacks glyceryl groups. Diutan is useful as a viscosifier in oil field and cement applications. Sphingomonasstrains produce polysaccharides as capsules firmly bound to the cell surface. The exact mechanism of attachment is not known. The capsule may limit productivity by impairing oxygen uptake. The functionality of the polysaccharide may be hindered by its being attached to the cell rather than free in solution.

Strain Construction:

Deletions of the corresponding genes dpsM and dpsN of Sphingomonassp. ATCC 53159, which produces diutan (S-657), were constructed. Each gene was deleted independently and the effect on capsule to slime determined. Briefly, PCR was used to amplify two fragments homologous to DNA flanking the target gene. These fragments were cloned into a narrow-host-range plasmid pLO2 that cannot replicate in Sphingomonasand contains two selective markers, kanRand sacB. Selection for kanamycin resistance selects for cells in which the plasmid has integrated into the chromosome in one of the homologous regions. The kanamycin resistant strain was then grown under nonselective conditions to allow loss of the plasmid by a second recombination. Loss of plasmid was selected by tolerance to sucrose. The sacB gene encodes an enzyme levansucrase for synthesis of levan from sucrose. Levan is toxic to the cells. Cells that have lost the sacB gene can grow on sucrose. The sucrose tolerant isolates can be wild-type or deletion. Presence of the deletion was confirmed by PCR. Mutants were tested for slime or capsule production. No foreign DNA, plasmid, or antibiotic resistance genes remained in the final strain.

Detailed Construction of dpsN and dpsM Deletions Strains

Deletions of dpsM and dpsN were constructed on a plasmid and transferred to the genome of ATCC 53159, using a gene replacement strategy similar to that described for S. elodeadeletion mutants [3]. PCR was used to amplify DNA regions flanking the target gene and then the fragments cloned into plasmid pLO2 [4], which was then used to exchange the deletion for the target gene in the chromosome. Primers used for the PCR are shown in Table 5. Restriction sites for cloning (shown in italics) were added to the ends of the primers.

TABLE 5Primers for construction of deletion mutations.PrimerSequencePurpose

SacI-Dpsn primer 15′ TTGAGCTC-GCTGTGGCTGTTCTTCCT 3′Amplifies 497 bp

(SEQ ID NO: 5)upstream of dpsN

XbaI-DpsN primer 25′ CGTCTAGA-GTCACGCCGGTTGAACAT 3′

(SEQ ID NO: 6)

XbaI-DpsN primer 35′ TCTCTAGA-CTCGGTCACCAGGTCTGAA 3′Amplifies 396 bp

(SEQ ID NO: 7)downstream of dpsN

SphI-DpsN primer 45′ CTCGCATGC-CGGTAAAGGTGAAG 3′

(SEQ ID NO: 8)

SacI-DpsM primer 15′ TTGAGCTC-GATCGGCGTTAAGACTGC 3′Amplifies 474 bp

(SEQ ID NO: 9)upstream of dpsN

XbaI-DpsM primer 25′ CGTCTAGA-TCATCGCGGTCGCTGCCAT 3′

(SEQ ID NO: 10)

XbaI-DpsM primer 35′ CCTCTAGA-CGTCGGAGGCATCATGTTC 3′Amplifies 509 bp

(SEQ ID NO: 11)downstream of dpsM

SphI-DpsM primer 45′ TCGCATGC-TCTGCTGATTGCCGTTCT 3′

(SEQ ID NO: 12)

Deletion constructions were designed to leave the remaining genes for diutan synthesis intact. For the dpsN deletion, primers SacI-DpsN primerl and XbaI-DpsN primer2 were used to amplify a 497 by fragment from the dpsM gene as a SacI-XbaI fragment (total 513 bp). Primers XbaI-DpsN primer3 and SphI-DpsN primer4 were used to amplify a 396 by fragment from the atrD gene as an XbaI-SphI fragment (total 413 bp). Since the end of the dpsM gene overlaps the start of the dpsN gene by 17 bp, the stop codon of dpsM and the start codon of dpsN were retained, as well as the natural stop codon of dpsN. Thus this construction may result in formation of a small peptide of 13 amino acids, as shown in FIG. 4. The PCR fragments were ligated sequentially into the polylinker of plasmid vector pLO2 [4], resulting in clone pLO2-dpsNdeln#3 carrying the deletion of dpsN.

This plasmid, pLO2-dpsNdeln#3 which cannot replicate in Sphingomonas, was transferred into the Sphingomonasstrain ATCC 53159 by conjugation from E. coliDH5α using a helper plasmid pRK2013 that supplies transfer functions [2]. Chromosomal integrants were selected by kanamycin resistance on YM medium with 7.5 μg/ml kanamycin, and 25 μg/ml streptomycin (to counterselect E. coli). Subsequent growth of the Sphingomonasstrains for approximately 30 generations in the absence of antibiotic allowed for excision of the plasmid. Recombinants that had lost the plasmid were then selected by sucrose (8%) tolerance, due to loss of the plasmid-encoded sacB gene, and then colonies screened for kanamycin sensitivity. Genomic DNA was prepared from several isolates to identify those isolates that had retained the dpsN deletion versus the wild-type gene, as determined by PCR.

Similarly, a deletion of dpsM was constructed. Primers SacI-DpsM primer] and XbaI-DpsM primer2 were used to amplify a 474 by fragment from the dpsE gene as a SacI-XbaI fragment (total 490 bp). Primers XbaI-DpsM primer3 and SphI-DpsM primer 4 were used to amplify a 509 by fragment from the dpsN gene as a XbaI-SphI fragment (total 525 bp). Since the end of the dpsM gene overlaps the start of the dpsN gene by 17 bp, the stop codon of dpsM and the start codon of dpsN were retained. A stop codon was incorporated within the XbaI cloning site. A 7-amino acid peptide may be formed from the dpsM start site. The PCR fragments were ligated sequentially into the polylinker of plasmid vector pLO2 [4], resulting in clone pLO2-dpsMdeln#1 carrying the deletion of dpsM. This plasmid was transferred by conjugation into ATCC 53159 selecting for kanamycin resistant integrants, followed by growth in the absence of antibiotic and detection of sucrose tolerant, kanamycin sensitive recombinants. Genomic DNA was isolated from selected recombinants and screened by PCR for presence of the deletion.

Example 10

Characterization of Diutan Slime-Forming Mutants

Results of several tests showed that both the dpsM and dpsN deletions result in a change from capsule former to slime former, as shown in FIGS. 3 and 4.

1. Microscopic evaluation of two dpsN deletion mutants (#3 and #5) and two dpsM deletion mutants (#1 and #5) grown about 16 hours in high carbon fermentation medium indicated that cells from these mutants did not form the large cell aggregates characteristic of the Sphingomonascapsular strain, S-657 (FIG. 5A, and FIG. 6A).

2. Wild-type ATCC 53159 cells grown in defined medium (DM2) with 1% glucose for 24 hours and diluted ten-fold formed visible clumps, where as the dpsM and dpsN slime mutants form uniform suspensions similar to that of a non-mucoid strain, DPS I (FIG. 5C for dpsN).

3. Centrifugation of 24-hour cultures grown in DM2 medium with 1% glucose showed that the cells from the dpsM and dpsN slime mutants could be pelleted, whereas those from wild-type ATCC 53159 (S-657) remained attached to the polysaccharide, and thus did not pellet (FIG. 5B and FIG. 6B).

Six independent isolates of dpsN deletion mutants exhibited an average 5.4% increase in total precipitable material compared to the wild-type control, in shake flask fermentations. Selected dpsN and dpsM mutant isolates were evaluated in 20 L Applikon fermentors using media containing organic and inorganic nitrogen, salts and different carbon concentrations (3-5%). Polysaccharide was precipitated with isopropyl alcohol, dried and weighed. The dpsN mutants consistently exhibited a slight increase in total precipitable material compared to the wild-type capsular control strain. The dpsM mutants gave more variable and generally lower productivity as shown in Table 6.

TABLE 6

Increase in yield of polysaccharide with dps mutants

5% carbon sourcedpsN #3n = 35.9%

dpsN #5n = 23.9%

dpsM #1n = 1−30.2%

dpsM #5n = 19.3%3% carbon sourcedpsN #3n = 24.2%

dpsM #1n = 2−10.1%

dpsM #5n = 4−2.7%

Example 11

Characterization of Diutan Slime-Form Polysaccharide

Rheological properties of diutan recovered from these fermentations by precipitation with isopropyl alcohol was determined, as shown in Table 7. Both dpsM and dpsN slime mutations resulted in improved viscosity of diutan.

TABLE 7Rheological properties of diutan from slime mutants%0.06 s−1%%

StrainSWV3increaseviscosityincreaseLSRVIncrease

wild-n = 526.727,7602010

type

dpsNn = 535.332%37,92037%387393%

#3

dpsNn = 237.340%41,40049%4075103%

#5

dpsMn = 340.552%37,73336%390594%

#1

dpsMn = 539.146%3944042%372085%

#5

aver.42%aver.41%aver.94%

It was also observed that fiber quality, e.g., length, was improved with the slime mutants. Since the polysaccharide molecules are free in solution rather than attached to the surface of the cell, the precipitation of these molecules may be facilitated.

Low Shear Rate Viscosity Measurement.

Low shear rate viscosity is the viscosity of a 0.25% solution of diutan at 3 rpm. Standard or synthetic tap water was prepared by dissolving 10 g of NaCI and 1.47 g of CaCl2.2H2O in 10 liters of deionized water. 4.5 g of Polyethylene Glycol (PEG) 200 was weighed directly in a 400-ml tall form beaker. A 0.75 g aliquot of diutan product was weighed, and dispersed in the PEG 200 to form a consistent slurry. 299 ml of synthetic tap water was added to the beaker and the mixture stirred at 800±20 rpm for approximately 4 hours. The beaker was removed from the stirring bench and placed in a 25° C. water bath and allowed to stand for 30 min. The viscosity was measured using a Brookfield LV Viscometer with the No. 2 spindle at 3 rpm.

Seawater Viscosity Measurement.

Seawater viscosity was determined . using the following procedure. Seawater solution was prepared by dissolving 41.95 g of sea salt (ASTM D-1141-52, from Lake Products Co., Inc. Maryland Heights, Mo.) per 980 g deionized water, with pH adjusted to 8.2 with HCl or NaOH as needed. 307 g of seawater solution was transferred to a mixing cup; 0.86 g of diutan product was slowly added over 15-30 seconds to the mixing cup and allowed to mix at 11,500 rpm for 45 minutes in the Fann Multi-Mixer, Model 9B5 (Fann Instruments, Inc, Houston, Tex.). Three drops of Bara Defoam (NL Baroid/NL Industries, Inc., Houston, Tex.) was added and stirring was continued for an additional 30 seconds. The mixing cup was removed from the mixer and immersed in chilled water to lower the fluid's temperature, then placed in a constant temperature bath at 25° C. The solution was transferred to a 400-ml tall form beaker.

Fann viscosity (Fann Viscometer, Model 35A) was measured while mixing at low speed (3 rpm). The shear stress value was read from the dial and recorded as the SWv value at 3 rpm.

The viscosity was also determined on the Brookfield LV DV-II or DV-II viscometer with the LV-2C spindle. The 0.06 sec−1reading was measured at 0.3 rpm.

Example 12

Materials and Methods

Medium. YM contains per liter, 3 g yeast extract, 5 g peptone, 3 g malt extract, and 10 g glucose. DM2 medium contains per liter, 2.68 g K2HPO4, 1.31 g KH2PO4, 2.0 g NH4SO4, 0.1 g MgSO4.7H2O, 15 mg CaCl2.2H2O, 8.34 mg FeSO4.7H2O, 0.05 mg MnCl2.4H2O, 0.03 mg CoCl2.6H2O, 0.8 mg CuSO4.5H2O, 0.02 mg Na2MoO4.2H2O, 1.0 mg ZnSO4.7H2O, 0.2 mg H3BO3and 10 g glucose. Gellan shake flask fermentation medium contains per liter, 0.23 g NaCl, 0.165 g CaCl2.2H2O, 2.8 g K2HPO4, 1.2 g KH2PO4, 1.9 g NaNO3, 1.0 g N—Z-Amine type EKC (Sheffield Products), 36.46 g Star-Dri corn syrup, 2.5 mg FeSO4.7H2O, 24 μg CoCl2.6H2O and 0.1 g MgSO4.7H2O.

Centrifugation test for slime. Strains were grown approximately 24 hours at 30° C. in DM2 medium containing 1% glucose, with shaking at 350 rpm and then centrifuged at maximum speed (10,000 rpm) for 5 minutes in the Eppendorf centrifuge.

Hot settling test. Strains were grown in gellan shake flask fermentation medium. Fermentation broth was heated in the autoclave for 10 minutes to liquefy gellan. The hot broth was then transferred to a large test tube and allowed to settle overnight at 95° C. (to maintain broth as liquid). With a capsular strain the cells are attached to the polysaccharide and remain suspended. For slime-formers, the cells are not attached and precipitate during overnight incubation.

PCR amplification. The high fidelity PCR enzyme “PfuUltra hot start DNA polymerase” from Stratagene (LaJolla, Calif.) was used.

REFERENCES

The disclosure of each reference cited is expressly incorporated herein.1. Coleman R C. 2001. Cloning and analysis of Sphingonionassp. ATCC 53159 polysaccharide genes. San Diego State University MS thesis

2. Ditta G. S Stanfield, D Corbin and D R Helinski. 1980. Broad host range DNA cloning system for Gram-negative bacteria: construction of a gene bank of Rhizobium meliloti. Proc Natl Acad Sci USA 77: 7347-7351.

3. Harding N E, Y N Patel and R J Coleman. 2004. Organization of genes required for gellan polysaccharide biosynthesis in Sphingomonas elodeaATCC 31461. J Ind Microbiol Biotechnol 31:70-82.

4. Lenz, O., E. Schwartz, J. Demedde, M. Eitinger and B. Friedrich. 1994. The Alcaligenes eutrophus H116 hoxX gene participates in hydrogenase ‘regulation. J. Bacteriol. 176:4385-4393.

5. Matthews T D. 2004. Identification of genes involved in phenotypic phase shifting of Sphingomonassp. ATCC 53159 San Diego State University MS thesis

6. Pollock T J and R W Armentrout. 1999. Planktonic/sessile dimorphism of polysaccharide-encapsulated Sphingomonads. J Ind Microbiol Biotechnol 23: 436-441.

7. Pollock, T J, W A T van Workum, L Thome, M Mikolajczak, M Yamazaki, J W Kijne and R W Armentrout. 1998. Assignment of biochemical functions to glycosyl transferase genes which are essential for biosynthesis of exopolysaccharides in Sphingomonasstrain S88 and Rhizobium leguminosarum. J Bacteriol 180: 586-593.

8. Sa-Correia I, A M Fialho, P Videira, L M Moreira, A R Marques and H Albano. 2002. Gellan gum biosynthesis in Sphingomonas pancimobilisATCC 31461: Genes, enzymes and exopolysaccharide production engineering. J Ind Microbiol Biotechnol. 29: 170-176.

9. Yamazaki M, L Thome, M Mikolajczak, R W Armentrout and T J Pollock. 1996. Linkage of genes essential for synthesis of a polysaccharide capsule in Sphingomonasstrain S88. J Bacteriol 178: 2676-2687.

10. U.S. Pat. No. 6,605,461.

11. U.S. Pat. No.7,361,754.

12. U.S. Pat. No. 5,854,034.