Chitinase-Producing Bacteria

This invention relates to novel man-made bacterial strains which produce chitinase, an enzyme which degrades chitin. This invention further relates to the use of such strains as a means to inhibit soil fungi and nematodes and to enhance plant growth by biological control of plant pathogens.

The soil contains a wide variety of life forms which can interact with plants, including bacteria, fungi and nematodes. These life forms are especially abundant in the rhizosphere, the area of the soil that surrounds and is influenced by the plant roots. As used herein the term rhizosphere embraces the rhizoplane, the root-soil interface including the surface of the root. The term rhizobacteria, as used herein, refers to bacteria adapted to the rhizosphere. The interactions between these soil inhabiting life forms are complex, some being antagonistic and others being mutually beneficial.

The interactions between plants and the various soil life forms are similarly complex, in some instances helpful to the plant and in other instances deleterious to the plant. Fungi harmful to plants (fungal pathogens) include fungal species from a wide variety of genera, including Fusarium, Pythium, Phytophthora, Verticillium, Rhizoctonia, Macrophomina, Thielaviopsis, Sclerotinia and numerous others. Plant diseases caused by fungi include pre- and post-emergence seedling damping-off, hypocotyl rots, root rots, crown rots, vascular wilts and a variety of other forms of symptom development. Nematodes harmful to plants (nematode pathogens) include nematode species from the genera Meloidogyne, Heterodera, Ditylenchus, Pratylenchus. Plant diseases caused by nematodes include root galls, root rot, lesions, "stubby" root, stunting, and various other rots and wilts associated with increased infection by pathogenic fungi. Some nematodes (e.g. Trichodorus, Longidorus, Xiphenema) can serve as vectors for virus diseases in a number of plants including Prunus, grape, tobacco and tomato.

Various approaches are available for attempting to control deleterious fungi and nematodes. One method, long known in the art, is chemical treatment of soil or plants with fungicides or nematicides. Another method is application of certain naturally occuring bacteria which inhibit or interfere with fungi or nematodes. See, in general, K. F. Baker and R. J. Cook, Biological Control of Plant Pathogens, Freeman and Co. (1974) for a description of fungi and nematodes and their interaction with plants, as well as a description of means for biological control of fungal and nematode pathogens.

One approach to biocontrol of fungal and nematode pathogens is based on the widespread presence of chitin as an integral part of the cell walls of fungi and the outer covering of nematodes or nematode eggs or nematode cysts. Chitin is an unbranched polysaccharide polymer consisting of N-acetyl-D-glucosamine units. It is insoluble in water, dilute mineral acids and bases but can be broken down enzymatically by chitinase, the degradation products being soluble monomers or multimers of N-acetyl-D-glucosamine. Chitinase is produced by certain naturally occurring bacteria and fungi and there have been reports of the role of chitinase in the suppression of pathogens.

R. Mitchell and M. Alexander, "The Mycolytic Phenomenon and Biological Control of Fusarium in Soil", Nature, 190, 109-110 (1961) describes naturally occurring mycolytic, or fungi-lysing, soil bacteria (genera Bacillus and Pseudomonas) which suppress soil Fusarium by means of chitinase activity.

B. Sneh, "Use of Rhizosphere Chitinolytic Bacteria for Biological Control", Phytopath. Z., 100, 251-56 (1981) discloses naturally occurring chitinolytic isolates identified as Arthrobacter sp. and Serratia liquifaciens. Sneh also discloses introduction of a chitinolytic bacterial strain from the genus Arthrobacter into the rhizosphere to protect carnation seedlings from Fusarium wilt.

A. H. Michael and P. E. Nelson, "Antagonistic effect of soil bacteria on Fusarium roseum culmorum", Phytopathology, 62, 1052-1056 (1972) discloses similar control with a naturally occurring Pseudomonas species.

J. Monreal and E. T. Reese, "The Chitinase of Serratia marcescens", Canadian Journal of Microbiology, 15, 689-696 (1969) describes a Serratia marcescens bacterial strain (QMB1466) selected as the most active chitinase producer out of a number of naturally occurring bacterial and fungal strains tested. Other strains tested which displayed some chitinase activity included bacterial strains from the genera Enterobacter and Streptomyces, and fungal strains from the genera Aspergillus, Pencillium and Trichoderma. Chitinase is characterized as an induced enzyme system in strain QMB1466, i.e. the yields of chitinase produced by the strain were higher when chitin was present. Monreal et al. reports at p. 692 that chitinase production on a chitin medium is repressed by the addition of other carbon-containing metabolites, e.g. sugars, to the medium. The Serratia marcescens enzyme system is described as extracellular and including endochitinase, a chitobiase and a "factor" for hydrolysis of "crystalline" chitin.

The naturally occuring Serratia marcescens chitinase system is further described in R. L. Roberts and E. Cabib, "Serratia Marcescens Chitinase: One-Step Purification and Use for the Determination of Chitin", Analytical Biochemistry, 127, 402-412 (1982).

J. D. Reid and D. M. Ogrydziak, "Chitinase-Over-producing Mutant of Serratia marcescens", Applied and Environmental Microbiology, 41, 664-669 (1981) describes work with a mutant of Serratia marcescens, strain IMR-1E1, obtained by mutation of strain QMB1466. The mutant had increased chitinase activity compared to strain QMB1466, as measured by zones of clearing on chitin-agar plates. On page 664 Reid et al. refers to the "high rate of reversion of IMR-1E1to decreased levels of chitinase production."

C. I. Kado and P. F. Lurquin, "Prospectus for Genetic Engineering in Agriculture", Phytopathogenic Prokaryotes, Vol. 2, M. S. Mount and G. H. Lacy eds., 309 (1982), while not discussing the role of chitinase in controlling chitin-containing pathogens, notes the possibility of a different approach to controlling fungi, namely, inserting into bacteria genes coding for compounds which inhibit chitin synthase in fungi. That is, the compound chitin synthase, necessary for production of chitin in fungi, would be inhibited by the bacterial compounds.

P. M. Miller and D. C. Sands, "Effects of Hydrolytic Enzymes on Plant-parasitic Nematodes", Journal of Nematology, 9, 192-197 (1977) describes the effect of chitinase, obtained from a commercal supplier, on certain nematodes. Miller et al. discloses that chitinase hydrolytic enzymes are toxic to certain nematodes, in particular Tylenchorhynchus dubius, the toxicity being greater in aqueous solution than in soil.

There are a number of limiting factors and disadvantages with respect to work to date on biological control of plant pathogens using chitinase-producing bacteria introduced into the soil. First is the inability to regulate the production of chitinase in the introduced bacteria in such a way that proper amounts of chitinase are produced. Second is the limited ability of many of such bacteria to colonize and persist in the rhizosphere of host plants, a key consideralion for effective biocontrol. Particularly important in this respect is the ability of biocontrol bacteria to colonize the roots of host plants effectively, the roots being the site of much plant-pathogen interaction. Third is that chitinase production is repressed in the presence of other carbon sources, e.g. metabolites released by the root. Another problem, at least as to mutants, is reversion to forms exhibiting decreased levels of chitinase production.

The present invention comprises novel man-made bacteria, in particular rhizobacteria, which have the ability to produce chitinase as the result of introduction into the bacteria of foreign DNA encoding for chitinase activity. The foreign DNA is isolated from a foreign source, bacterial or otherwise, or is substantially homologous to such DNA. The novel bacteria can be prepared using various means, including the use of appropriate vectors, for introducing the foreign DNA, and thus the capacity to produce chitinase, into a bacterial cell or a parent of a bacterial cell, under conditions where chitinase activity is expressed. The vectors are used to clone and introduce the foreign DNA which encodes for chitinase activity. The novel bacteria can be used to introduce chitinase into the soil, particularly into the soil rhizosphere, thereby providing a means of inhibiting chitin-containing or chitinase-sensitive plant pathogens, including fungi and nematodes, and thereby also providing a means of enhancing the growth and well being of plants sensitive to such pathogens.

The present invention provides a means of overcoming limitations of the prior art methods of bacterial control of chitinase-sensitive plant pathogens. The invention provides a means of introducing sufficient chitinase production capacity into a strain. The invention also provides a means to introduce chitinase capacity into strains best suited to function in the soil and rhizosphere, in particular root colonizing rhizobacteria strains. Further, in accordance with the invention, the problem of reversion of modified strains to wild type is overcome in that the novel strains of the invention result from the actual introduction of genetic material, rather than from mutation. Additionally, the invention provides the means to overcome the problem of repression of chitinase activity in the presence of root exudates or other carbon sources, in that regulatory systems can be employed which render the bacterial cell insensitive to such repression.

The novel bacterial cells of the invention are made by introduction of foreign DNA, or heterologous DNA, which codes for production of, or expression of, the enzyme chitinase. The term "chitinase" is used herein to mean chitin-degrading enzyme, the term "chitin-degrading" embracing both chitin-hydrolyzing and chitin-solubilizing. The term "chitinase DNA" is used herein to mean DNA which encodes for chitinase, and embraces foreign chitinase DNA obtained directly or indirectly from a source organism, including bacteria, fungi and plants, as well as DNA which regardless of source is substantially homologous to such foreign chitinase DNA. "Chitinase activity", or "chitinolytic activity", as used herein, means the ability or capacity of a bacterial cell to produce chitinase. Such chitinase can be secreted by the bacteria into the immediate environment.

Chitinase DNA can be obtained from a wide variety of naturally occurring bacteria which are known to or can be shown to produce chitinase, including bacteria from the genera Serratia, Bacillus, Pseudomonas, Arthrobacter, Enterobacter, and Streptomyces. Bacterial strains containing chitinase DNA have been known and available from laboratories or collections for years. For instance, chitinase-producing Serratia marcescens strain QMB1466, which was described by Monreal et al. in 1969 and by Reid et al. in 1981, (in each case the reported source of the strain being the U.S. Army Natick Laboratory Culture Collection) is available from a number of sources, including the American Type Culture Collection at Rockville, Maryland (ATCC 990). Chitinase-containing bacterial strains are also readily obtainable by known techniques by virtue of their widespread distribution in nature. Such strains in general are found in soil, on plants, on insects and in water systems, as well as in other places where chitin is present. For example, chitinolytic bacteria can be isolated from the rhizosphere of a wide variety of plants including sugar beet, cotton, bean or carnation. Chitinase-producing bacteria can also be obtained from root surfaces, fungal resting structures (e.g. sclerotia, chlamydospores), nematode egg masses, insect or arthropod exo-skeleton and irrigation water.

Isolation of bacterial strains containing chitinase DNA can be accomplished by a number of techniques, including direct isolation on chitin-containing media, enrichment or baiting with chitin or fungal hyphae. These techniques are common and known to those skilled in the art. Chitinase-producing fungi can be isolated from sources such as those stated above for chitinolytic bacteria, again using standard techniques for plating fungi. See, in general, J. Tuite, Plant Pathological Methods: Fungi and Bacteria, Burgess Publishing Co. (1969) with respect to techniques for isolation of bacteria and fungi.

Foreign chitinase DNA for conferring chitinase activity on a host, or recipient, bacterium can be obtained directly from a source organism, e.g. bacteria, fungi, yeast, insect or plant, using techniques of genome fragmentation and DNA isolation known to those skilled in the art. For instance, for a bacterial source organism, isolated as explained above, total bacterial DNA (that is, the entire genome including chromosomal DNA and extra-chromosomal DNA) is isolated by standard techniques, e.g. lysis of bacteria in the presence of appropriate detergents, proteases and chelating agents, followed by phenol and chloroform extractions and precipitation with ethanol. The isolated DNA is partially digested to various degrees with an appropriate restriction enzyme or enzymes selected on the basis of appropriate sites in the cloning vector which is to be used. The products of the digestion process are fractionated by standard techniques, for instance on a glycerol gradient Fractions containing DNA in an appropriate size range, e.g. about 22 to about 32 kb (kilo bases), are selected for insertion into an appropriate vector using known techniques, for instance as described below, thus yielding a genomic library (consisting of cosmid clones, in the case of a cosmid vector).

An alternative to obtaining chitinase DNA directly by genome fragmentation of a source organism is obtaining chitinase DNA indirectly by isolating, from the source organism, messenger RNA (mRNA) corresponding to chitinase DNA. A cDNA (copy DNA) library can be prepared from the mRNA, using reverse transcriptase in accordance with techniques known to those skilled in the art, and inserted into an appropriate cDNA expression vector, such that clones encoding chitinase activity could be detected by clearing of chitin on plates.

The choice of particular vector turns on a number of considerations known to those skilled in the art, including the size of the fragment, nature of the host, number and position of restriction sites desired, and selection marker or markers desired. Techniques for introduction of DNA into a vector and subsequent introduction of the vector into the host bacteria are known to those skilled in the art. See in general T. Maniatis et al, "Molecular Cloning, a Laboratory Manual," Cold Spring Harbor Publications, 1982 (hereinafter Maniatis) with respect to techniques for insertion of DNA fragments into a host bacterium (as well as with respect to general techniques for fragmentation and fractionation of a genome.)

Introduction of foreign DNA into the host bacteria results in the creation of a bank of modified, i.e. transformed or transduced, host bacteria which can be screened for chitinase DNA. In many cases the host bacteria will be E. coli. The screening can be carried out by plating the host strains on a medium which contains chitin, e.g. colloidal chitin. The development of zones of clearing in the chitin around a colony is evidence that the colony is chitinolytic. Microscopic examination showing dissolution of surrounding chitin particles is further evidence. Alternative means to screen will be apparent to those skilled, e.g. plating on a fungal lawn, or chemical tests to show the presence of chitinase.

For those bacteria shown by screening to exhibit chitinase activity, there can be optionally employed a subsequent subcloning to reduce the quantity of cloned DNA which is not involved in the coding for chitinase. An appropriate enzyme digestion is carried out and the digestion products ligated to another more convenient cloning vector, e.g. one with high copy number, and the ligation products are again transformed into E. coli bacteria by known techniques. Transformants are screened for chitinase production as described above.

After the cloning and any subcloning, if desired, the chitinase DNA can be transferred from the first host (transferor or donor) bacterial cell into a second host (transferee or recipient) bacterial cell. This transfer can be accomplished using known techniques, for instance by conjugation using helper plasmids to mobilize the plasmid into the transconjugant cell, the specifics depending on the transferor bacterium, the recipient bacterium, and the cloning vector which is used to propagate the chitinase DNA. For instance, if the chitinase DNA is cloned on an IncP (incompatibility group P) type plasmid derivative, such as pLAFR1, transfer to a second host strain in many instances can be accomplished by conjugation, e.g. using a helper plasmid such as pRK2013. See in general G. Ditta et al, "Broad Host Range DNA cloning system for Gram-negative bacteria: Construction of a gene bank of Rhizobium meliloti", Proc. Natl. Acad. Sci., 77, 7347-7351 (1980) (hereinafter Ditta), with respect to conjugation using helper plasmids. Where the intended use of the bacteria modified to have chitinase ability is in control of plant pathogens residing in the soil, the bacteria of choice will normally be rhizobacteria. In that event chitinase DNA is transferred from the first host, normally E. coli, into the second host rhizobacterial strain.

Depending on the systems and circumstances involved in transferring a vector containing chitinase DNA from one bacterial cell to another, various techniques known to those skilled in the art may be used to ensure proper expression of the chitinase DNA in the host. For instance, an effective regulatory or promoter system will be necessary to bring about proper expression, that is, to ensure that the production of chitinase, encoded for by foreign chitinase DNA, can be brought about under conditions where chitinase production is desired. If the promoter from the source organism (i.e, the promoter which normally works in the source organism with the foreign chitinase DNA) is not effective in the host, it may be necessary to incorporate into the vector a regulatory system different from that which controlled the foreign DNA in the source organism. A promoter system of choice may be one which allows the bacterial cell to produce chitinase in a manner insensitive to the presence of carbon sources, e.g. root metabolites, in the immediate environment. That is, the cell can be made to produce chitinase constitutively. Various other techniques to enhance chitinase activity in the modified cell may be employed as well, e.g. multicopy vectors or means to enhance secretion of the chitinase from the cell.

Plasmids containing chitinase DNA, i.e. clones or chimaeric plasmids, can be introduced into a bacterial host by transformation, e.g. using CaCl.sub.2, the transformed cell being called the transformant. The plasmid may be a cosmid vector containing chitinase DNA, i.e. a cosmid clone, and if so it can also be introduced into the bacterial cell by transduction, the product cell being the transductant.

The particular adaption of rhizobacterial cells to the rhizosphere is related to their ability to multiply and compete at the root-soil interface or the root surface, or in the intercortical cell spaces. Root colonizing rhizobacteria typically reach population densities of 10.sup.4 or greater colony forming units (cfu) per mg of root tissue, from low initial populations, during the first several weeks of plant growth. Various rhizobacteria have been described, including strains from the genera Pseudomonas (in particular P. fluorescens and P. putida), Agrobacterium, Enterobacter and Alcaligenes. See in general T. Suslow, "Role of Root-Colonizing Bacteria in Plant Growth", Phytopathogenic Prokaryotes, Vol. 1, M. S. Mount and G. H. Lacy eds., 187-223 (1982) for a discussion of root colonizing rhizosphere bacteria and their properties. The choice of root colonizing strain to receive chitinase DNA will turn on the plant to be protected, the pathogen or pathogens to be protected against, the method of application, and the cultural practices related to the crop of interest.

For bacterial strains which already have some chitinase activity, introduction of chitinase DNA in accordance with the present invention serves to enhance chitinase activity in the host. Other bacteria already have anti-fungal (fungicidal) or anti-nematodal (nematicidal) capacity by some mechanism other than chitinase activity, in which case introduction of chitinase DNA confers chitinase activity and enhances anti-pathogen ability.

The present invention can also be used in combination with the introduction of some other foreign DNA, that is foreign DNA other than chitinase DNA, into a bacteria. For instance, in the case of rhizobacteria, such other foreign DNA could provide the host with some other form of anti-pathogen activity or with some other means to allow it to enhance the soil environment to the benefit of the plant.

The present invention is of agricultural use as a means for the production of chitinase, including the production of chitinase as an antibiotic for the purpose of degrading or otherwise inhibiting, repelling or killing plant pathogens harmful to a wide variety of agricultural crops. The invention has particular utility for inhibiting chitinase-sensitive fungi or nematodes (that is, fungi or nematodes which are inhibited, repelled or destroyed in the presence of chitinase), where such fungi or nematodes or their activities in soil are harmful to plants. Regardless of the mechanism by which such pathogens are injurious to plants, their inhibition serves to enhance plant growth and health.

Bacteria, and particularly rhizobacteria, modified in accordance with the present invention and grown to sufficient proportions, e.g., by fermentation, can be used to combat chitin-containing soil pathogens by application of the bacteria to soil, seeds, vegetative plant parts or irrigation water. For example, mycolytic bacteria created in accordance with the invention can be used in such ways to attack or inhibit fungi. The modified bacteria can be applied in various formulations containing agronomically acceptable adjuvants or carriers in dosages and concentrations chosen to maximize the beneficial effect of the rhizobacteria.

For application to soil, to soil mixes, or to artificial plant growth media, the modified bacteria may be applied as a powder or granule in a suitable carrier. Alternatively, the modified bacteria may be applied as a suspension or dispersion, e.g. as an aqueous suspension with a suitable protectant such as methycellulose, dextran, dextrin, alginate, magnesium silicate. The modified bacteria may also be applied as a wettable powder.

For application to seeds, the modified bacteria may be applied as part of a seed coating composition, for instance mixed with xanthan gum, magnesium silicate, methylcellulose, gum arabic, polyvinyl pyrollidone, dextrins or dextrans. In addition, small amounts of partially hydrolyzed chitin may be added to the pelleting mix, dust, granule, suspension, or wettable powder to enhance chitinase production. See in general T. Suslow et al., "Rhizobacteria of sugar beets: effects of seed application and root colonization on yield", Phytopathology, 72, 199-206 (1982); and, J. Kloepper et al., "Development of a powder formulation for inoculation of potato seed pieces", Phytopathology, 71, 590-592 (1981), for a discussion of rhizobacteria and seed coating compositions.

Bacteria into which chitinase capability has been introduced by this invention may also be applied to the above-ground surface of a plant, e.g., the leaf or stem surface, either to permit the modified bacteria to travel or spread to the roots or to inhibit chitinase-sensitive pathogens which may be present on blossoms or plant surfaces, for instance, fungal pathogens such as Botrytis, Monilinia, Alternaria, and Cercospora. Blossoms of Prunus sp., in particular, provide an ideal environment for the growth of epiphytic bacteria, e.g. Pseudomonas syringae or Erwinia herbicola, that have the ability to produce inhibitory levels of chitinase.

The method of the invention can also be used for introduction of chitinase genes into species of Rhizobium which enter into a nitrogen-fixing symbiosis within the nodules of leguminous plants. The nodules are frequently the point of entry of pathogenic fungi and nematodes.

The method of the invention additionally provides a means to introduce chitinase DNA into a bacteria, e.g. Agrobacterium, which is used to transfer the foreign DNA to plants. Such transfer results in a direct means for the plant to inhibit chitinase-sensitive plant pathogens, either alone or in conjunction with bacteria modified to have chitinase ability. A particularly attractive form of such transfer is one where the chitinase DNA is expressed by the plant only at the site of pathogen attack, e.g. only in the root cells.

Both of the above applications (introduction of chitinase activity into Rhizobium or plants) would involve subcloning the chitinase genes and bringing them under the control of different regulatory sequences from those which act in the source organism. For example, elevated expression in E. coli could be brought about by using the lac Z system (B-galactosidase structural gene promoter). In nodules elicited by Rhizobium a nitrogenase promoter could be used, and in plant leaves the promoter of a highly expressed leaf gene could be used.

The overall procedure was to construct a set of random cosmid clones from the Serratia marcescens genome which would cover the entire genome several times over in such a way that statistically there was at least a 99% chance of covering every DNA sequence in the genome. Clones carrying an entire chitinase gene were inserted in E. coli, which is quite closely related, taxonomically, to S. marcescens. The work involved in isolating clones which carry chitinase DNA had the following steps, as explained in detail below.

(a) Isolating total S. marcescens DNA.

(b) Partial digesting of S. marcescens DNA.

(c) Purifying a fraction of the partial DNA digest in which the DNA fragment size was 22 kb-32 kb.

(d) Ligating the purified DNA to a cosmid cloning vector.

(e) In vitro packaging into lambda phage.

(f) Transfecting E. coli cells with lambda phage and selection.

(g) Carrying out small scale plasmid isolations on tetracycline resistant colonies and digesting to check that foreign DNA had been cloned.

(h) Plating and screening for colonies which clear chitin.

(i) Characterizing cosmid clones conferring chitinase activity.

(a) Isolation of total S. marcescens DNA

Cells of Serratia marcescens QMB1466 were removed from culture storage and streaked on agar media to form single isolated pure colonies.

A single colony was inoculated into 5 mls of 1% bactotryptone, 0.5% yeast extract and 0.5% NaCl (hereinafter LB) liquid medium and grown overnight with shaking at 28.degree. C. 1 ml aliquots were spun down in 1.5 ml Eppendorf tubes and resuspended in 0.3 ml 20 mM Tris, 10 mM EDTA (pH 8.0). 0.1 ml of 5% SARKOSYL and 0.1 ml of 5 mg/ml pronase were added and the cells were incubated at 37.degree. C. for lysis to proceed for two hours. After this incubation, the solution was passed through a 19 gauge needle to shear the DNA slightly and thus to reduce the viscosity of the solution. Next, 0.5 ml of phenol (pH adjusted to 8.0 with Tris) was added and the mixture shaken in the Eppendorf tube prior to centrifugation. This step was repeated three times, with the supernatant from one centrifugation being re-extracted with fresh phenol. Then the supernatant was extracted three times with 0.8 ml of a one-to-one mixture of phenol and chloroform/isoamyl alcohol (24:1) and once with 0.8 ml of chloroform/isoamyl alcohol. The supernatant from this final spin was brought to 0.3 M sodium acetate and the DNA precipitated by addition of 2.5 volumes of ethanol. After centrifugation to pellet the DNA precipitate, the DNA was redissolved in 0.1 ml of 10 mM Tris/1 mM EDTA (hereinafter TE). An aliquot was taken and diluted into 0.5 ml for measurement of the optical density at 260 nm in order to find out the concentration of nucleic acid. Typically this procedure permitted the isolation of 100-200 micrograms of DNA.

The procedure adopted for establishment of appropriate DNA to enzyme ratios for correct partial digestion was the widely used method described in Maniatis pp. 282-283. The objective was to establish conditions where the maximum fluorescence of the partial digestion products occurred in the size range 40-50 kb. 10 ug (microgram) of DNA was incubated in 150 ul (microliter) of the restriction enzyme buffer specified by the manufacturer (New England Biolabs) and dispensed in 15 ul aliquots except for one tube which contained a 30 ul aliquot. 10 units of EcoR1 were added to the 30 ul aliquot, the contents of the tube were mixed and a 15 ul aliquot withdrawn, added to the next tube and the contents mixed, and the procedure repeated down the series of tubes. After a one hour incubation at 37.degree. C. the reaction was terminated with 3 ul of 0.25 M EDTA/50% glycerol/0.01% bromophenol blue, and the digestion products run on a 0.4% agarose gel which was stained with 0.5 ug/ml ethidium bromide and examined by fluorescence in short wave uv light. The migration of the partial digestion products in the gel was compared to size markers of known size. Once conditions were established for partial digestion of chromosomal DNA to the appropriate degree, 200 ug of DNA was digested to this degree in an appropriately scaled-up volume. Partial digests giving weight average sizes at the maximum fluorescence position of 40 kb and 20 kb were mixed and fractionated on a glycerol gradient.

The digestion was terminated by addition of enough 0.5 M EDTA to bring the final EDTA concentration to 10 mM followed by incubation of the reaction at 65.degree. C. for 10 minutes, and then was kept on ice until the gradient was loaded. An aliquot was checked for the degree of digestion being appropriate by running on a 0.3% agarose gel with DNA fragment size markers of appropriate size (e.g., digests of lambda DNA).

Linear gradients of 10-40% glycerol were prepared in 38 ml polyallomer tubes. The 10% or 40% glycerol stock solutions were made up in 1 M sodium acetate, 5 mM EDTA. 0.5-1.5 ml aliquots of partial digests containing 100-300 ug of partial digest were loaded on top of the gradients which were then spun at 25000 rpm for 16 hours.

At the end of the centrifugation the tubes were punctured at the bottom, 1 ml aliquots were dripped out and the DNA in them analyzed by agarose gel electrophoresis in 0.3% gels. Fractions containing DNA in the size range 22-32 kb were chosen for further work.

Fractions of this size range were pooled and dialyzed against 10 mM Tris 1, mM EDTA (pH 8.0) for 24 hours with three buffer changes.

These fractions were then concentrated by isobutanol extraction (Maniatis, p. 463) to about 0.3 ml, brought to 0.3 M sodium acetate and the DNA was precipitated by addition of 2.5 volumes of ethanol. The precipitated DNA was redissolved in 10 ul of TE and the quantity of DNA recovered estimated by measuring the optical density at 260 nm of a dilution of an aliquot of this DNA.

The vector used for cloning the Serratia DNA was pLAFR1. As described by A. Friedman et al., "Construction of a broad host range cosmid cloning vector and its use in the genetic analysis of Rhizobium meliloti", Gene, 18, 289-96 (1982), pLAFR1 has a single Eco RI site, a cos site from lambda phage for in vitro packaging, and a tetracycline resistance marker. The vector pLAFR1 selects DNA inserts of about 22 to about 32 kb in length. The vector can be mobilized to other genera of bacteria where it can replicate.

5 ug of pLAFR1 DNA was digested to completion with EcoR1, and the DNA phenol extracted and ethanol precipitated. The precipitated DNA was redissolved in 20 ul of TE.

Test ligations were carried out on both the pLAFR1 DNA and the size-fractionated DNA which was to be cloned to verify that the ends were ligatable.

In a typical ligation of pLAFR1 DNA to size-fractionated Serratia DNA, a 5-fold molar excess of Serratia DNA was adopted. A typical ligation contained in 10 ul 0.4 ug of pLAFR1 DNA and 3 ug of Serratia DNA. The reaction was 66 mM Tris (pH 7.5), 10 mM MgCl.sub.2, 1 mM ATP, 15 mM dithiothreitol, 0.05% BSA, 0.5 mM spermidine and 20 units/ul T4 DNA ligase (New England Biolabs). The reaction was carried out overnight at 15.degree. C.

Packaging extracts were prepared as described in Maniatis p. 264-267. Freeze thaw lysate was frozen at -80.degree. C. in 10 ul aliquots, and sonic extract was frozen away in 15 ul aliquots. One tube of each was thawed on ice and the freeze thaw was added to the sonic extract and mixed gently. Then 5 ul of the ligation was added to the mixture and after gentle mixing, the packaging reaction was allowed to proceed at 25.degree. C. for one hour. The reaction was diluted with 500 ul of 10 mM MgCl.sub.2 10 mM Tris (pH 7.5), 10 mM NaCl (hereinafter SM), and 500 ul chloroform were added. The mixture was inverted five times in the capped Eppendorf tube and spun for five minutes in an Eppendorf bench centrifuge.

E. coli strain DH1 (ATCC #33849) displays no detectable chitinase activity (see Table II). The strain was grown to saturation in LB containing 0.4% maltose. A 0.2 ml aliquot was withdrawn and mixed with 0.1 ml of SM and 10 ul of the diluted packaging mix. After mixing gently, phage absorption was allowed to proceed for 20 minutes at 37.degree. C.

The transfection was added to 1.7 ml of LB in a tube and the cells permitted to grow out for 40 minutes at 37.degree. C. In the first experiment 20, 100, 500, and 1100 ul aliquots were plated on LB plates containing 1.5% agar and 10 mg/l tetracycline to investigate the colony forming units derived from the packaging. The plates were incubated at 37.degree. C. for 16-20 hours. In a typical experiment a 10 ul aliquot of packaging dilution would contain 1000 colony forming units.

Single tetracycline resistant colonies were picked from plates at the end of stage (f) above into 8 ml aliquots of LB containing 0.5 g/l uridine and incubated with aeration for 12-20 hours at 37.degree. C. Cells were spun down and resuspended in 0.2 ml 50 mM glucose, 20 mM Tris, 10 mM EDTA (pH 8.0). The cells were then lysed in 0.4 ml 0.2M NaOH, 1% SDS. This was neutralized with 0.3 ml 3M potassium acetate which had been brought to pH 5.0 with acetic acid and incubated on ice for five minutes. After thorough mixing, the mixture was spun at 8000 rpm for ten minutes. The supernatant was precipitated with 0.6 volumes of isopropanol and the DNA recovered by centrifugation in a bench top Eppendorf centrifuge. The pelleted DNA was redissolved in 0.3 ml TE and extracted with equal volumes of phenol and chloroform. After centrifugation the supernatant was brought to 0.3 M sodium acetate and the DNA precipitated with 2.5 volumes of ethanol. After centrifugation the DNA was redissolved in 0.05 ml of TE and 5 ul aliquots were used for restriction enzyme digestion followed by gel electrophoresis.

In a typical experiment eight independent clones were grown up and 50% of them contained inserts.

Theoretically about 500 independent clones of a genome the size of E. coli should give a 99% chance of getting any particular sequence among the clones. It is desirable to independently isolate any clone with chitinase activity at least once. Five thousand colonies were plated out at about 250 colonies per plate on LB medium containing 2.0% colloidal chitin and 10 mg/l tetracycline. This concentration of chitin had been previously shown to clearly evidence the chitinase activity of S. marcescens QMB1466.

After about seven days at 32.degree. C., certain colonies gave rise to clear zones around them. Altogether about twenty different colonies gave rise to convincing clear zones in their vicinity. See Table II and step i below. In Table II, DH1 is the original E. coli (step f); DH1/pLAFR1 is E. coli DH1 containing the cosmid vector pLAFR1 but without insert; DH1/C3 is E. coli DH1 containing cosmid vector pLAFR1 with one chitinase size class insert; and DHl/C12 is E. coli DH1 containing cosmid vector pLAFR1 with a second chitinase size class insert (this strain is deposited with the American Type Culture Collection in Rockville, Maryland as ATCC No. 67152).

Ten of the twenty colonies were inoculated into 8 ml of LB tet medium and plasmid DNA prepared as above. This DNA was analyzed for the DNA sequences in the plasmid by digestion with EcoR1. Each of the ten plasmid DNAs fell into one of two distinct size classes after EcoR1 digestion. Seven out of the ten cosmid clones showed one large EcoR1 fragment of about 25 kb, in addition to the vector band of 21.6 kb. Of these, one (C3) was chosen for further characterization (see Table II and associated text). Three of the plasmids showed insert fragments of 3 kb, 9.5 kb and 17 kb, in addition to the vector band of 21.6 kb. Of these, one (C12) was chosen for further characterization (see Table II and associated text).

The phenotype of chitinase production was shown to be plasmid borne by reintroduction of the plasmid into E. coli bacteria by transformation of the bacteria with plasmid DNA. 1 ul out of the 50 ul of plasmid DNA prepared as in (h) was incubated with 0.1 ml of competent E. coli cells prepared essentially by the method of M. Dagert and S. D. Ehrlich, Gene, 6, 23-28 (1979). After a 20 minute incubation on ice and a two minute heat shock at 37.degree. C., bacteria were grown out in LB medium for one hour and plated on LB tetracycline chitin plates. Bacteria which acquired tetracycline resistance all acquired the capacity to make chitinase.

The above experiments provide evidence that S. marcescens QMB1466 contains two independent chitinase genes which have been isolated on distinct cosmid clones. The means is therefore presented to express these genes either together or separately in a recipient organism.

Pseudomonas fluorescens strain NZ130 and Pseudomonas putida strain MK280 were isolated from radish roots, and soybean roots, respectively, by plating on King's Medium B from serial dilutions of root washings. See in general T. Suslow, "Role of Root-Colonizing Bacteria in Plant Growth," Phytopathogenic Prokaryates, Vol. 1, M. S. Mount and G. H. Lacy eds., 187-223 (1982) for details of fluorescent Pseudomonas isolation and characterization for colonizing ability and plant growth promotion. Strain NZ130 has been identified as P. fluorescens Biotype D (Pseudomonas chlororaphis in some taxonomies) and strain MK280 as P. putida. Their characteristics include the following:

Strain NZ130 has plant growth promoting characteristics on a number of crops including potato, radish, soybean, cotton, and sugar beet. Strain NZ130 also has biological control characteristics with respect to Pythium sp., but no measurable chitinase activity. Root colonization data collected for NZ130, in general, reach average population densities of 5.5.times.10.sup.4 colony-forming units (cfu) per mg root tissue (dry weight) on radish, soybean and cotton.

Strain MK280 has been shown to increase the emergence of soybean and to effectively colonize roots of soybean and sugar beets. Population densities, in general, reach as high as 1.2.times.10.sup.6 cfu/mg root tissue.

pLAFR1 is a mobilizable cloning vector derived from pRK290 (Ditta). It can be mobilized into other genera of bacteria using a helper plasmid pRK2013 in a three-way mating process. Two Pseudomonas strains were chosen as recipients for these matings; these were NZ130r (NZ130, rifampcin resistant) and MK280r (MK280, rifampicin resistant). The donor (transferor) strain was E. coli DH1 or HB101 containing one of the two chitinase cosmid clones, and the helper strain was HB101 containing pRK2013, all of which materials are commonly available.

Donor, recipient and helper strains were grown up to mid-log phase, without selection, in LB. 0.05 ml aliquots from each strain were added to each other and the mixture put out as a 0.15 ml aliquot on an LB plate for 12-16 hours at room temperature. After the conjugation, a loop was run through the cells. Cells from the loop were resuspended in 10 mM MgSO.sub.4 and the mixture of cells plated at various dilutions on minimal sucrose tetracycline (10 mg/l ) plates. This procedure selected against E. coli, which cannot grow on minimal sucrose plates, and selects for Pseudomonas cells which acquire the tetracycline resistance gene. Exconjugant cells were obtained and were tested for chitinase activity as above. See Table II. In Table II, TS031 is NZ130r with cosmid vector pLAFR1 but without insert; TSO43 is NZ130r with cosmid vector pLAFR1 with one chitinase size class insert (corresponding to the insert of DH1/C12); TSO35 is MK280r with cosmid vector pLAFR1 but without insert; and TSO44 is MK280r with cosmid vector pLAFR1 with one chitinase size class insert (corresponding to the insert of DH1/C3). TSO44 has been deposited with the American Type Culture Collection in Rockville, Maryland (ATCC No. 39637). Note that Table II also lists data for two naturally occurring, chitinase producing strains which are not root colonizers: Serratia marcescens strain ATCC 990 and a strain of naturally occurring Arthrobacter sp.