1. FIELD OF THE INVENTION The present invention relates to a novel approach for drug discovery. More particularly, the invention relates to a system for improving the process of lead optimization and development of compounds, when these compounds are natural products produced by microorganisms belonging to the order Actinomycetales or chemical derivatives of these compounds.

The invention relates to a system for transferring the capability to produce a natural product from a microorganism belonging to the order Actinomycetales into a defined host, where said natural product can be optimally produced and its biosynthetic pathway suitably modified.

2. BACKGROUND ART Natural products are complex molecules with important uses in medicine. Examples include: antibacterial agents, such as erythromycin, teicoplanin, tetacycline; antitumor compounds, such as dauxorubicin; antihelmintic compounds, such as avermectin; immunosuppressive agents, such as cyclosporin and FK506; antifungal compounds, such as amphotericin and nystatin; etc. Natural products are produced as secondary metabolites by a wide range of living organisms. Although many secondary metabolites have been identified, there remains the need to obtain novel structures with new activities or enhanced properties.

Current methods of obtaining such molecules include screening of natural isolates and chemical modification of existing ones. Random screening of natural products from

disparate sources has resulted in the discovery of many important drugs and is still employed for seeking for novel activities. This process, which consists in exposing a miniaturized biological system to tens or hundreds of thousands of different compounds, in order to find those few that exhibit a desired property, is designated high throughput screening, or HTS.

One of the used sources widely in HTS is a collection of natural products produced by small-scale fermentation of newly isolated microorganisms. A natural product may have one or more potential therapeutic properties, including but not limited to antibacterial, antifungal, antiviral, antitumor, immunomodulating or other pharmacological properties. Natural products have long constituted a source of interesting, structurally original and "imaginative"molecules endowed with potent biological activities. In addition, recent observations indicate that only a small fraction of the microbial flora present in environmental samples, ranging from 0.01 to 1% according to the estimates, is related to known species. Microorganisms belonging to the order Actinomycetales represent thus far the group of producers unsurpassed for chemical and biological diversity. However, more than 15,000 natural products produced by microorganisms have been described, and the chances of finding new structures are relatively small, unless efforts are directed towards those classes of microorganisms that have been little exploited in the past.

Poorly characterized actinomycete genera can thus constitute a useful source of novel structures. With proper methodologies, unusual genera can be isolated from environmental samples and some of these isolates will produce interesting activities. These could either represent completely new entities, or known molecules

acting on a novel target or in a previously unreported way.

Many of these products will have original structures and potent biological activities. However, newly discovered secondary metabolites will be produced for the most part by microorganisms which have been isolated for the characteristic of being unusual and selected for their ability to produce a given bioactivity. Consequently, little will be known about the best conditions for growth, productivity and storage. Often the microorganism does not produce a single bioactive compound, and other, unrelated activities must be completely removed for a meaningful evaluation of the properties of the lead compound.

Furthermore, rarely is a secondary metabolite produced as a single, bioactive molecule, but is often present as a "complex"of several, closely related compounds, only some of which may possess the desired biological or chemical properties. Therefore, physiological conditions, such as nutrient and cofactor supply, that allow obtaining a "controlled"complex need to be established empirically by a trial and error approach. Finally, the natural product may need structural modification, and this can be achieved only by chemical means. In essence, the scarce knowledge available on the physiology and genetics of the producing strain will severely hamper the lead optimization and development processes.

Chemical modification of preexisting natural products has been successfully employed to generate derivatives of natural products, but it still suffers from practical limitations to the type of compounds obtainable. Many natural products are often structurally complex molecules, with relatively large molecular weights. Due to their structural complexity, total synthesis of natural products is often prohibitive for the number of necessary steps and

the overall yield; furthermore, selective modification of a natural product can often be efficiently performed only on limited portions of the molecule. This difficulty of generating structural derivatives by conventional medicinal chemistry slows down the process of lead optimization and supply. Microorganisms employ intricate biosynthetic machineries to make natural products: for example, synthesis of the macrolide antibiotic erythromycin, a secondary metabolite in the medium-range structural complexity, requires the participation of over 40 different enzymatic activities (Katz and Donadio, 1995, Macrolides, in Genetics and Biochemistry of Antibiotic Production, Vining and Stuttard eds., Butterworth-Heinemann, Boston CT, p. 385-420). Biosynthetic pathways can often be redirected through manipulation of the fermentation conditions or of the biosynthesis genes, in order to produce desired analogs of the original structure. The availability of genes involved in the formation of secondary metabolites has been exploited for the formation of derivatives of natural products obtained after genetic manipulation of the producing organism (Hopwood, 1997, Chem. Rev. 99: 0-39).

These manipulations have resulted in novel molecules, many of which would be extremely hard if not impossible to produce by chemical derivatization of the parent compound.

The obvious economical and environmental benefits resulting from the formation of the desired structure in one fermentation step constitute an additional stimulus for the application of pathway engineering for the rational design of novel structures. The compounds obtained in this way are amenable evaluation of their biological properties as well as being substrates for further derivatization by chemical or biological means.

In summary, the supply of a natural product produced

by a newly discovered microorganism, the optimization of the complex composition, and the process of lead optimization will all benefit from a detailed knowledge of the genetics and physiology of the producing strain. The present invention describes a general method for transferring the capability to produce any secondary metabolite from the original actinomycete producer to an established and genetically manipulatable production host.

The general concept of the invention is illustrated in Fig.

1. Conditions for optimal growth, metabolite production and maintenance need therefore to be developed for one host. In addition, the availability of the cloned genes in a genetically manipulatable and well characterized host allows the utilization of all the genetic tools developed for these strains for the creation of novel derivatives of the natural product after genetic intervention.

3. SUMMARY OF THE INVENTION The present invention provides a system for producing and manipulating natural products produced by a large group of bacteria for the purpose of drug discovery, development and production. The method of the invention transfers the ability to produce a secondary metabolite from an actinomycete that is the original producer of the natural product, to another production host that has desirable characteristics.

In one embodiment, the invention involves the construction of a library from a donor organism, the producer of a natural product, in an Artificial Chromosome that can be shuttled between a convenient, neutral cloning host, such as the bacterium Escherichia coli, and a production host, such as the actinomycetes Streptomyces lividans or Streptomyces coelicolor. The clones directing

the synthesis of the natural product are identified in said library, transferred into the production host where said natural product is synthesized.

In another embodiment, the invention involves the reconstruction of a large segment that directs the synthesis of a natural product, starting from smaller DNA fragments cloned from the genome of a donor organism. This reconstruction occurs in an Artificial Chromosome that can be maintained in a convenient neutral host, such as the bacterium Escherichia coli, and subsequently transferred into an actinomycete production host. The reconstructed genomic segment in the Artificial Chromosome is transferred into the production host where said natural product is synthesized.

The present invention also relates to Escherichia coli-Streptomyces Artificial Chromosomes, recombinant DNA vectors useful for shuttling the genetic information necessary to synthesize a given natural product between a donor actinomycete producer and a production host.

3.1 DEFINITIONS As used herein, the following terms will have the meaning indicated.

An"Escherichia coli-Streptomyces Artificial Chromosome", or ESAC, is a recombinant DNA construct that can maintain very large DNA inserts in an Escherichia coli host and that can be introduced and maintained in an actinomycete production host.

An"Escherichia coli-Streptomyces Artificial Chromosome"library, or ESAC library, is a library of different recombinant constructs carrying very large DNA inserts that can be maintained in an Escherichia coli host SUBSTITUTE SHEET (RULE 26)

and introduced and maintained in an actinomycete production host.

A pESAC is a vector used to construct an"Escherichia coli-Streptomyces Artificial Chromosome"or an ESAC library.

A"natural product"is a secondary metabolite made by a microorganism through a series of biosynthetic steps. This natural product may or may not have any useful biological activity.

A"complex"is the mixture of related natural products with similar properties and biological activity that are often produced by the same biosynthetic pathway.

A"donor organism"is the original producer of a natural product, where the synthesis of said compound is governed by a defined number of genetic elements.

A"gene cluster", a"cluster", a"biosynthesis cluster" all designate a contiguous segment of the donor organism's genome that contains all the genes required for the synthesis of a natural product.

A"production host"is a microorganism where the formation of a natural product is directed by a gene cluster derived from a donor organism.

As used in the present invention, the following abbreviations are employed: °C (Celsius degree); h (hour); min (minute); kb (kilobase); pl (microliter); ml (milliliter); mm (millimeter); mg (milligram); pg (microgram); ng (nanogram); M (molar); Mb (megabase); UV (ultraviolet); kV (kilovolt); Q (Ohm); mFa (millifaraday).

In addition, the following abbreviations are used: Ab, antibiotic; Ap, ampicillini attB, chromosomal attachment site ; attP, phage or plasmid attachment site; bp, base pair ; ca., circa (i. e. "about"); Cm, chloramphenicol; E.,

Escherichia; ESAC, E. coli-Streptomyces Artificial Chromosome ; GC, guanosine + cytosine; HTS, high throughput screening; Km, kanamycin ; int, integrase encoding gene; LB, Luria Broth ; LMP, low melting point; P., Planobispora ; PCR, polymerase chain reaction ; PFGE, Pulsed Field Gel Electrophoresis; R, resistance; rpm, rounds per minute; S., Streptomyces; s, sensitive ; Sac., Saccharopolyspora ; sacB, gene conferring sensitivity to sucrose ; SDS, sodium dodecyl sulfate ; Tc, tetacycline; TE, TrisHCl EDTA buffer ; tet, tetacycline resistance gene ; Th, thiostrepton ; ts, temperature sensitive ; tsr, thiostrepton resistance gene ; U, units; vol, volume; wt, weight; YEME, yeast extract malt extract medium.

4. BRIEF DESCRIPTION OF THE DRAWINGS Figure 1. Scheme of the invention. The general concept of the invention, whereby the gene cluster required for the synthesis of a natural product in a donor organism is established as an ESAC in an Escherichia coli host, and then transferred into a desired production host, where it integrates into the chromosome and directs production of the secondary metabolite. The hexagon represents the natural product, the twisted thin line the bacterial chromosomes, and the thick line the desired gene cluster.

The pESAC episome is represented by a circle.

Figure 2. E. coli-Streptomyces Artificial Chromosome vectors. Vectors pPAC-S1 and pPAC-S2 differ solely for the orientation of the int-tsr cassette. Relevant features of the vectors are illustrated. Kmr indicates resistance to kanamycin; sacB indicates sensitivity to sucrose. Suitable cloning sites are shown as: B, BamHI; S, ScaI ; X, XbaI. The replicating function of bacteriophage P1 are indicated by

the thick bars.

Figure 3. General scheme of the invention, top-down approach. High molecular weight DNA from the donor organism is cloned into a pESAC. The resulting library in E. coli is screened with the required probes, and the relevant ESACs are identified. These are introduced into the desired production host strain, where they integrate site- specifically into the host chromosome. Symbols and abbreviations are as in Fig. 1.

Figure 4. General scheme of the invention, bottom-up approach. A cosmid library is prepared with DNA from the donor organism and screened with the required probes. The overlapping inserts from the positive cosmids, which consitute the correct contig, are assembled into a pESAC via homologous recombination in E. coli. The reconstructed ESAC is introduced into the desired production host, where it integrates site-specifically into the host chromosome.

Symbols and abbreviations are as in Fig. 1.

Figure 5. Scheme of assemblage. The figure illustrates a hypothetical genomic segment from a donor organism that is covered by the inserts from three overlapping clones.

The relevant fragments A and D, which denote the ends of the segment, and B and C, which represent regions of overlap, are indicated with their relative orientation (thick side on the fragment rectangle). The bottom part illustrates the reconstructed ESAC.

Figure 6. Constructs required for cluster assemblage.

The plasmids indicated are generated by routine in vitro DNA manipulations. Fragments A, B, C and D are as in Fig.

5. Fragment pairs are in this example separated by a marker, indicated as AbR for antibiotic resistance.

Selective markers present on the two compatible replicons

are, as an example: CmR and KmR.

Figure 7. Interplasmid insert exchange. Each of the CmR derivatives, as of Fig. 6, is introduced in the same E. coli cell as the cognate clone of Fig. 5 (for example a cosmid that carries a KmR marker). Formation and then resolution of the cointegrate leads to the transfer of the cosmid's insert, indicated here by a looping line, in the CmR replicon.

Figure 8. Sequel of assembling steps. A series of interplasmid cointegration and resolution events is conducted. Only the growing ESAC is indicated. The starting construct (Fig. 6) is recombined with plasmid pAB2 (Fig.

7), leading to the insertion of the segment flanked by fragments A and B. Next, the AbR marker from pBC1 (Fig. 6) is introduced between fragments B and C, and subsequently replaced by the insert from pBC2 (Fig. 7). Finally, the AbR marker from pCD1 (Fig. 6) is introduced between fragments C and D, and subsequently replaced by the insert from pCD2 (Fig. 7).

Figure 9. A gene cluster from Planobispora rosea. The extension of a gene cluster from P. rosea ATCC 53733 is reported, together with the cosmids pRP16, pRP31 and pRP58.

The fragments A, B, C and D used for assemblage are highlighted. Restriction sites are abbreviated as: M, SmaI ; P, PstI; S, Sstl.

Figure 10. Site-specific integration of an ESAC. PFGE analysis of S. lividans ZX7 transformed with ESAC-70. Lanes 1 and 2: S. coelicolor M145; lane 3: S. lividans ZX7 DNA; lane 4: ZX7 attB:: ESAC-70 DNA, colony 1; lane 5: ZX7 attB:: ESAC-70 DNA, colony 2; lane 6: 50-kb ladder, size marker. All DNAs in lanes 1-5 are digested with DraI.

Conditions for PFGE are: 200 Volts, 70 s switching for 7 15

h, 120 s switching for 11 h.

Figure 11. Characterization of S. lividans transformants. Southern hybridization of S. lividans attB:: PAD6, grown with (lane 1) or without (lane 2) thiostrepton. P. rosea DNA is shown as control (lane 3).

Lane 4 contains 1-kb ladder. All DNAs are digested with BamHI and probed with labeled PAD6.

5. DETAILED DESCRIPTION OF THE INVENTION In its broadest sense, the present invention entails a general procedure for constructing a Streptomyces host producing any natural product after selective transfer of the relevant genes from the original actinomycete producer, the donor strain. This general procedure is outlined in Fig. 1. The present invention can be applied with only limited information on the structure of the natural product and very little knowledge of the original producer's genetics. The present invention has a substantial impact on the process of drug discovery involving natural products or their structural derivatives. The transfer of the producing capability to a well characterized host can substantially improve several portions of the process of lead optimization and development: the titer of the natural product in the producing strain can be more effectively increased; the purification of the natural product can be carried out in a known background of possible interfering activities; the composition of the complex can be more effectively controlled; altered derivatives of the natural product can be more effectively produced through manipulation of the fermentation conditions or by pathway engineering. In order to better understand the value of the present invention, a brief description is reported below of

the current methods for optimizing the productivity of the producing strain, for purifying a natural product, for controlling the composition of a complex, and for producing derivatives of a natural product.

The production of a natural product is controlled by several mechanisms, few of which have been established in detail. Generally, the level of production of a natural product depends on the composition of the growth medium; on the presence of appropriate precursors or on the absence of specific inhibitors; on the expression level and timing of genes controlling the biosynthetic pathway and competing routes; and on the level and specific activity of key enzymes in the pathway. Because of this complexity, the productivity of the original strain is usually increased by an empirical process, which may include, among other things, one or more of the following steps: strain purification, selection of phenotypic variants arising spontaneously or after mutagenic treatment of the strain, variation in the fermentation medium or in the fermentation parameters; genetic engineering of the producing strain.

Fundamental knowledge about the physiology of the producing strain and the variables affecting titer must be achieved for an effective improvement of productivity. This knowledge is very scant in a newly identified producer strain.

During the discovery and development phase, sufficient quantities of a natural product must be available for an evaluation of its properties and/or for the generation of analogs. Because of its uniqueness, a specific purification process must be developed for each natural product.

However, it is highly desirable to have the natural product as free as possible of compounds that may interfere with the biological activity of the molecule. Contaminating

impurities must be characterized analytically and biologically. In a poorly characterized producer, little information is available on the relevance of contaminating impurities.

A natural product may be produced by a microorganism as a complex of a few or tens of molecules with minor structural differences, designated congeners. Although most of the congeners are usually biologically active, only one or a few may represent the desired product: for example, one congener may be substantially more active than the others; it may possess better physico-chemical properties; or it may be a better substrate for chemical modification.

The composition of a complex can be somehow controlled by intervening on the fermentation parameters. However, the most effective way is usually the altered expression of selected genes by genetic engineering (e. g. Sezonov et al., 1997, Nature Biotechnol. 15: 349-353).

Chemical modification of natural products represents the most commonly used means of obtaining novel structures.

This approach has been successfully employed, but it still suffers from practical limitations to the number and type of compounds obtainable. The structural complexity of many natural products makes their total synthesis often too lengthy and expensive to be of any practical use. This same structural complexity, with either the presence of several closely related functional groups or their absence, limits modification of a natural product to selected portions of the molecule. Methods of combinatorial synthesis need an initial scaffold as the starting building block, and this can be often generated only through a low yield degradation of the natural product. However, derivatives of natural products that would be very hard if not impossible to produce by chemical means have been obtained after genetic

alteration of the biosynthetic pathway. Examples include the introduction of additional genetic information (Epp et al., 1989, Gene 85: 293-301), the targeted inactivation of selected genes or portion thereof (Donadio et al., 1993, Proc. Natl. Acad. Sci. USA 90: 7119-7123), the"mixing and matching"of genes or portions thereof from different pathways (McDaniel et al., 1994, Nature 375: 549-554).

All the above activities are important for the process of lead optimization and for the development of selected lead structures. They can all benefit, to different extent, from a detailed knowledge of the physiology of the producing strain, and from the possibility of genetically manipulating it. The process by which a given organism is genetically manipulated in order to alter the type, quality or quantity of a natural product is referred to as pathway engineering. The ability to perform pathway engineering in a newly isolated microorganism producing a bioactive molecule with promising characteristics can therefore considerably expedite the optimization of a lead structure and the development process. Pathway engineering can be schematized as a sequel of three steps: a) isolation of the genes of interest; b) performing on selected gene (s) the manipulations required by the specific objective; and c) introduction of the modified gene (s) in suitable form in an appropriate host.

Isolation of the genes of interest from most actinomycetes can be achieved quite easily. The genes for primary metabolism are usually well conserved, and they can be easily accessed in any microorganism by using suitable hybridization probes or by the PCR. The genetic elements governing the biosynthesis of the major classes of secondary metabolites have been also described, and many genes can similarly be identified. Since natural product

biosynthesis is governed by clusters, one needs to identify just a few genes in order to have them all. However, synthesis of the vast majority of natural products requires a considerable extent of genetic information. For examples, biosynthesis of the natural products erythromycin (an antibiotic), avermectin (an antihelmintic agent) and rapamycin (an immunosuppressant) requires 55,90 and 95 kb, respectively, of genetic information (Katz and Donadio, 1993, Annu. Rev. Microbiol. 47: 875-912; MacNeil, 1995, Avermectins, in Genetics and Biochemistry of Antibiotic Production, Vining and Stuttard eds., Butterworth- Heinemann, Boston CT, p. 421-442; Schwecke et al., 1995, Proc. Natl. Acad. Sci. USA 92: 7839-7843). Other natural products may require even larger extent of genetic information. Therefore, in order to isolate an entire cluster in a single piece, cloning vectors capable of accepting and maintaining large DNA segments are necessary.

The manipulation of the isolated genes is generally best performed in a convenient cloning host, such as E. coli. Manipulations relevant to pathway engineering can include some or all of the following: site directed mutagenesis, gene inactivation, gene fusions, modification of regulatory sequences, etc. Techniques for the in vitro manipulation of DNA and for the propagation of the mutated alleles in E. coli are well developed and can be applied to DNA from virtually any source (Sambrook et al., 1989, In Molecular Cloning: A laboratory Manual, 2nd edn, Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press).

The final step in pathway engineering requires the introduction of modified or heterologous gene (s), in suitable form, in a strain where these genes can be appropriately expressed. This strain is often the strain

producing the natural product whose quantity, quality or type one wants to alter. The genes of interest must be carried on appropriate vectors: according to the particular objective of pathway engineering, one may need, among others, vectors that can be stably maintained as single or multicopy episomes; that can insert into the host chromosome at a fixed location; that allow replacement of an endogenous gene with an in vitro modified allele; that allow deletion of selected genes from the host chromosome.

In addition, for each strain one must have means for introducing heterologous DNA and selecting for its presence. Therefore, in order to genetically manipulate a given producer, one must establish conditions for rendering the bacterial cell capable of receiving incoming DNA; for selecting the incoming DNA; and develop vectors and methodologies for the various types of manipulations exemplified above. Low-and high copy-number, integrative, non-replicating vectors must be developed with appropriate selection markers. Thus, for each producing strain, specific gene transfer tools and conditions must be developed, starting in most cases from extremely poor knowledge about the microorganism. In addition, techniques developed for one species do not necessarily apply to a new species from the same genus, and often not even to a new strain. It is then no wonder that, among the thousands of strains described as producers of interesting natural products, gene transfer systems have been developed only for a limited number of species, which serve either as model organisms for genetic and physiological studies, or produce a commercially important molecule. The present invention provides tools for the general manipulation of any secondary metabolite pathway, and overcomes the difficulties of developing ad hoc conditions for a new

producer.

Naive hosts have been shown to produce the appropriate natural product or its intermediate (s) when the relevant DNA was introduced into them (Malpartida and Hopwood, 1984, Nature 309: 462-464; Hong et al., 1997, J. Bacteriol.

179: 470-476; Kao et al., 1994, Science 265: 509-512; McGowan et al., 1996, Mol. Microbiol. 22: 415-426; Kealey et al., 1998; Proc. Natl. Acad. Sci. USA 95: 505-509). However, the examples reported thus far represent special cases. Indeed, they include the introduction of relatively small DNA segments into a production host; or the transfer of gene clusters within members of the same bacterial genus; or they have required the careful engineering of specific biosynthesis genes under the control of appropriate genetic elements that direct their expression. Furthermore, the Streptomyces vectors currently available have an upper limit of ca. 40 kb (Hopwood et al., 1987, Methods Enzymol.

153: 116-167).

Until now, it was not established that DNA fragments exceeding 100 kb, derived from the high GC genome of actinomycetes, could be cloned and stably maintained in an E. coli host. Nor was any report of the introduction of large DNA segments into a Streptomyces host. The unexpected finding described herein is that these cloning tasks can be achieved according to the principles and methodologies of the present invention. Furthermore, the genetic elements required for the synthesis of a natural product in the original producer are genetically stable in a heterologous host, where they can direct the synthesis of the desired molecule. It was also unexpected and unprecedented that this heterologous stability and expression can occur when the donor organism and the production host belong to different bacterial genera.

The present invention rests on the fact that the genes required for the formation of a natural product are found as gene clusters of a defined size; that these gene clusters can be conveniently isolated, manipulated and transferred among different actinomycete strains; that they are expressed in a heterologous host; and on the fact that all the primary metabolite precursors required for the formation of a particular natural product are either produced by selected enzymes encoded by cluster-specific genes, or are present and available in the heterologous host at the time of formation of the natural product. The present invention addresses also the crucial aspect of natural product formation in actinomycetes: i. e. synthesis of many natural products may require over 100 kb of genetic information. To be generally applicable, transferring all the genes necessary for the production of any natural product requires cloning vectors capable of accommodating fragments as large as 150 kb, and possibly more. An object of the present invention is therefore represented by vectors capable of accommodating such large fragments which are also capable of being stably maintained in a suitable microbial host, such as a Streptomyces host.

Examples of these vectors are designated with the generic name pESAC. They are derived from bacterial artificial chromosomes (Shizuya et al., 1992, Proc. Natl.

Acad. Sci. USA 89: 8794-8797; Ioannou et al., 1994, Nature Genet. 6: 84-89) and can carry inserts up to 300 kb, or more.

As a general example of the broad applicability of the principles and methodologies described in the present invention, the Examples reported below describe how a convenient Streptomyces host can be engineered to carry a large gene cluster in order to produce a desired natural product through the use of an appropriate ESAC. The

exemplary organism chosen as the donor organism is the actinomycete P. rosea, belonging to one of the lesser characterized genera of actinomycetes (Goodfellow, 1992, In The Prokaryotes, 2nd edn., Balows, Trueper, Dworkin, Harder and Schleifer eds, Springer-Verlag, New York, NY, USA).

This organism produces the natural product GE2270 (Selva et al., 1991, J. Antibiotics 44: 693-701), an antibacterial agent. This particular case therefore describes the general applicability of the present invention, since very little information is available on the donor organism, on its genetics and physiology, and on the gene clusters present in its genome. Further examples described herein illustrate the application of the principles and methodologies of the present invention to other gene clusters described in the literature.

The present invention, relating to a general method for transferring the capability to produce any natural product from the original actinomycete to an established and genetically manipulatable Streptomyces host, can be schematized in a series of passages summarized as: 1) design of suitable vectors; 2) construction of a large- insert library in said vectors; 3) selection of the desired clones with appropriate probes; 4) insertion of the selected clones into a convenient Streptomyces host; and 5) growth of the recombinant strain under appropriate conditions to produce the natural product.

Actinomycetes produce a large number of natural products with important applications. However, other important classes of microbial producers are known, and newer ones are likely to be discovered in the upcoming years, as more microbial sources are screened for potential new drugs. Important classes of microbial producers include, among others, filamentous fungi, bacilli,

mixobacteria, pseudomonas and cyanobacteria. The series of passages described above can therefore be applied to other important classes of microbial producers, provided that two requisites are met: the synthesis of the desired natural product is governed by a gene cluster; suitable production host (s) exist; and appropriate selective marker (s) and maintenance function (s) are introduced into the Artificial Chromosome.

Furthermore, the series of passages summarized above and described in detail in the Examples, involve the use of a neutral cloning host. This host, as described in the present invention, is the bacterium Escherichia coli. In a preferred example of such a host, a high cloning efficiency can be obtained, and many of the analyses of the ESACs can be quickly performed. However, it is evident to one of ordinary skill in this art that any other host that allows high cloning efficiency can be used as neutral cloning host. Additionally, the use of such a host is not a conditio sine qua non for the applicability of the present invention. In fact, when it is possible to establish directly a library in a production host, there is no need for an intermediate neutral cloning host.

In summary, the present invention consists of a method for transferring the production of a natural product from an actinomycete donor organism that is the original producer of said natural product to a different actinomycete host, where this transfer is achieved by means of an E. coli-Streptomyces Artificial Chromosome that carries a gene cluster governing the biosynthesis of said natural product derived from said donor organism. This method comprises the steps of : (a) isolating large fragments of chromosomal DNA of the actinomycete donor organism of a size which encompasses

the gene cluster that directs the biosynthesis of the natural product; (b) constructing a suitable vector capable of accomodating said large fragments of chromosomal DNA and of introducing and stably maintaining said large fragments of DNA into an E. coli host; (c) constructing an E. coli-Streptomyces Artificial Chromosome by inserting said large fragments of chromosomal DNA of step (a) into the above said vector of step (b) and selecting the E. coli-Streptomyces Artificial Chromosome comprising the entire gene cluster construct that directs the biosynthesis of the above said natural product; (d) transforming an actinomycete host different from the donor actinomycete host with the E. coli-Streptomyces Artificial Chromosome of step (c) that carries the gene cluster governing the biosynthesis of said natural product wherein the actinomycete host carries a region which is specific for the integration of the E. coli-Streptomyces Artificial Chromosome.

6. GENERAL METHODS Plasmids, Bacterial Strains and Growth Conditions Plasmids pUCBM20, pUCBM21, pBR322 and pUC18 are obtained from Boheringer Mannheim; plasmid pIJ39 and OC31 DNA have been described (Hopwood et al., 1985, Genetic Manipulation of Streptomyces: A Laboratory Manual, The John Innes Foundation, Norwich, UK) and are available from prof. David Hopwood, The John Innes Centre, Norwich, UK; plasmid pCYPAC2 has been described (Ioannou et al., 1994, Nature Genetics 6: 84-89) and is available from prof. Pieter de Jong, Roswell Park Cancer Institute, Buffalo, NY, USA; plasmid pMAK705 has been described (Hamilton, et al., 1989,

J. Bacteriol, 171: 4617) and is available from prof. Sidney Kushner, University of Georgia, Athens, USA; cosmid Lorist6 has been described (Gibson et al., 1987, Gene, 53: 283-286) and is from prof. Stewart Cole, Pasteur Institute, Paris, France. E. coli strains are obtained from commercial sources: DH5a (Life Technologies), DH10B (Life Technologies), C600 (E. coli Genetic Stock Center), DH1 (Life Technologies) and XLlblue (Stratagene). S. coelicolor M145 and S. lividans ZX7 have been described (Hopwood et al., 1985, Genetic Manipulation of Streptomyces: A Laboratory Manual, The John Innes Foundation, Norwich, UK) and are available from prof. David Hopwood, The John Innes Institute, Norwich, UK. Planobispora rosea ATCC 53733, Streptomyces hygroscopicus ATCC 29253, Amycolatopsis mediterranei ATCC 13685 and Saccharopolyspora erythraea NRRL2338 are from the ATCC culture collection. All other materials are from commercial sources. Media for cultivation of E. coli (Sambrook et al., 1989, In Molecular Cloning: A laboratory Manual, 2nd edn, Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press) and Streptomyces (Hopwood et al., 1985, Genetic Manipulation of Streptomyces: A Laboratory Manual, The John Innes Foundation, Norwich, UK) have been described. The JM medium for S. coelicolor has been described (Puglia et al., 1995, Mol. Microbiol. 17: 737-746).

DNA Manipulations DNA manipulations are performed following described procedures, using the appropriate E. coli strains as cloning hosts (Sambrook et al., 1989, In Molecular Cloning: A laboratory Manual, 2nd edn, Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press). Genomic DNA from actinomycetes is prepared as described (Hopwood et al., 1985, Genetic Manipulation of

Streptomyces: A Laboratory Manual, The John Innes Foundation, Norwich, UK). A cosmid library of P. rosea DNA is constructed in the cosmid vector Lorist6 following published procedures (Sambrook et al., 1989, In Molecular Cloning: A laboratory Manual, 2nd edn, Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press).

Amplification by the PCR are performed following published guidelines (Innis, Gelfand, Sninsky and White, eds., 1990, PCR Protocols: A guide to Methods and Applications, Academic Press, San Diego, CA, USA).

Hybridizations Probes Pep6 and Pep8 are derived from conserved motifs in peptide synthetase gene sequences (Turgay and Marahiel, 1994, Pept. Res. 7: 238-241).

Oligonucleotide probe Pep6 consists of an equimolar mixture of 5'-GCSTACATCATCTACACSTCSGGSACSACS-GGSAAGCCSAAGGG- 3'(SEQID N°1)(SEQID N°1) and 5'- GGSTACATCATCTACACSAGCGGSACSACSGGSAAGCCSAAGGG-3' (SEQID N°2).

Oligonucleotide probe Pep8 consists of an equimolar mixture of 5'-AKGCTGTCSCCSCCSAGSNNGAAG-AAGTYGTCGTCGATSCC-3' (SEQID N°3) and 5'-AKGGAGTCSCCSCCSAGSNNGAAGAAGTYGTCGTCGATSCC- 3'(SEQID N°4).(SEQID N°4). [S indicates G or C; K indicates G or T ; Y, C or T; and N, any base]. Hybridizations are performed with a hybridization stringency set at 2xSSC, 55 °C, and a final wash set at the same stringency.

Preparation of high molecular weight DNA Procedures for the preparation of high molecular weight DNA from actinomycetes for PFGE have been described (Dyson, 1993, Trends Genet. 9: 72; Kieser et al., 1992, J. Bacteriol.

174: 5496-5507). They are modified for constructing libraries as described in the Examples.

7. EXAMPLES

The present invention consists in a series of passages, involving the design of suitable vectors; the introduction of large DNA inserts in said vectors employing genomic DNA from the donor organism; the selection of clones carrying the cluster specifying the synthesis of the desired natural product; the introduction of selected clone (s) into the appropriate production host; and the growth of the recombinant strain under appropriate conditions for metabolite production. These passages are described in detail in the Examples reported herein. These Examples outline the steps necessary to accomplish each passage, for the overall purpose of the present invention: the production of a natural product in a different host. They serve to illustrate the principles and methodologies of the present invention, and are not meant to restrict its scope to the Examples specified herein.

7.1 Cloning vectors Bacterial Artificial Chromosomes are circular plasmids that can be easily propagated in and prepared from E. coli cells by standard miniprep methods (Shizuya et al., 1992, Proc.

Natl. Acad. Sci. USA 89: 8794-8797; Ioannou et al., 1994, Nature Genet. 6: 84-89). In order to adapt Bacterial Artificial Chromosomes to a Streptomyces host, they need to be endowed with a selectable marker and maintenance functions. Site-specific integration, mediated by the action of an integrase encoded by the int gene, allows the stable incorporation of episomal elements into the host genome, at a defined locus designated attB. The episomal element needs to carry the cognate attP site and it may lack replicative functions. In addition, int-mediated excision of the integrated element from the chromosome via

reversal of the integration event can be prevented through selection of the resistance marker carried by the integrated episome; or, if necessary, after site-specific integration has occurred, the int gene on the integrated episome can be inactivated. Site-specific integration therefore allows the introduction of foreign DNA in single copy at a defined genetic locus. Several systems capable of directing site-specific integration of incoming circular DNA into the chromosome of a Streptomyces host have been described. A convenient system that can be used in the present invention is for istance the int-attP system derived from the temperate bacteriophage OC31 (Kuhstoss and Rao, 1991, J. Mol. Biol. 222: 897-908), which directs, during lysogen formation, integration of the 41-kb phage genome at the attB site, located in a stable segment of the S. coelicolor chromosome (Redenbach et al., 1996, Mol.

Microbiol. 21: 77-96). Several selectable markers have been described that can be used for Streptomyces (Hopwood et al., 1985, Genetic Manipulation of Streptomyces: A Laboratory Manual, The John Innes Foundation, Norwich, UK).

The tsr gene, conferring resistance to the antibiotic thiostrepton (Thompson et al., 1982, Gene 20: 51-62), is used in the present invention. The pESAC vectors, pPAC-S1 and pPAC-S2, described in the present invention, are depicted in Fig. 2. Their relevant features are: ability to accommodate DNA inserts up to 300 kb; low copy number in E. coli for increased stability; ease of propagation in E. coli because of the pUC19 stuffer segment; BamHI, XbaI or ScaI cloning sites, with positive selection of inserts for resistance to sucrose; T7 and SP6 promoters flanking the cloning site; KmR or ThR for selection in E. coli or actinomycetes, respectively; site-specific integration at

the OC31 attB site into the Streptomyces genome. Vectors pPAC-S1 and pPAC-S2 are 22 kb in size and differ solely for the orientation of the int-tsr cassette. After release of the stuffer pUC19 segment, the vector size is reduced to 19.7 kb. When cloning in the BamHI site, the vector can be released by digestion with DraI, resulting in vector fragments of 7.4,4.2 and 0.6 kb. The additional 7.5 kb of vector DNA will be associated with the insert. DraI rarely cuts in the high-GC genome of actinomycetes, so that the insert size can be easily calculated.

Example 1 Isolation of the int region from OC31 Two pairs of PCR primers, 5'-TTTTTGGTACCTGACGTCCCGAAGG- CGTG-3'(SEQID N°5)(SEQID N°5) and 5'-CAGCTTGTCCATGGCGGA-3' (SEQID N°6); and 5'-TCTGTCCGCCATGGACAAGC-3' (SEQID N°7) and 5'- TTTTTGGATCCGGCTAACTAACTAAACCGAGA-3' (SEQID N°8), are used to amplify the int-containing fragments of 1.3 and 0.9 kb, respectively. The template is OC31 DNA. The amplified fragments are digested with KpnI + NcoI and NcoI + BamHI, respectively, and recovered from an agarose gel.

Example 2 Construction of plasmid pINT The 1.3 and 0.9 kb fragment, prepared as described in Example 1, are ligated to pUCMB21, digested with KpnI + BamHI. The resulting mixture contains the desired plasmid pINT.

Example 3 Construction of E. coli K12 DH5a/pINT Approximately 10 ng of plasmid pINT, prepared as described

in Example 2, are used to transform E. coli DH5a and a few of the resulting ApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pINT, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 4.0 and 0.9 kb after digestion of the plasmid with NcoI + BamHI.

Example 4 Construction of plasmids pUIT1 The 1.8 kb BamHI fragment containing the tsr gene is isolated from pIJ39 and ligated to pINT, prepared as described in Example 3 and previously digested with BamHI.

The resulting mixture contains the desired plasmids pUIT1.

Example 5 Construction of E. coli K12 DH5a/pUITl Approximately 10 ng of plasmid pUIT1, prepared as described in Example 4, are used to transform E. coli DH5a and a few of the resulting ApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUIT1, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 4.9 and 1.8 kb after BamHI digestion of the plasmid.

Example 6 Construction of plasmid pUIT3 The 3.7 kb ApaI fragment, containing the int-tsr cassette, is isolated from plasmid pUIT1, prepared as described in Example 5, and ligated to pUCBM21 digested with ApaI. The resulting mixture contains the desired plasmid pUIT3.

Example 7

Construction of E. coli K12 DH5a/pUIT3 Approximately 10 ng of plasmid pUIT3, prepared as described in Example 6, are used to transform E. coli DH5a and a few of the resulting ApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUIT3, as verified by the observation, upon agarose gel-electrophoresis, of fragments of of 4.2 and 2.2 kb after BamHI digestion of the plasmid.

Example 8 Construction of plasmid pUIT4 The BamHI site present in the int-tsr cassette of plasmid pUIT3 is eliminated as follows. Plasmid pUIT3, prepared as described in Example 7, is partially digested with BamHI, followed by filling-in of the resulting ends, and treated with DNA ligase. The resulting mixture contains the desired plasmid pUIT4.

Example 9 Construction of E. coli K12 DH5a/pUIT4 Approximately 10 ng of plasmid pUIT4, prepared as described in Example 8, are used to transform E. coli DH5a and a few of the resulting ApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUIT4, as verified by the observation, upon agarose gel-electrophoresis, of a 6.4 kb fragment after BamHI digestion of the plasmid.

Example 10 Construction of plasmid pPAC-S1 and pPAC-S2 The 3.7 kb ApaI fragment from pUIT4, prepared as described in Example 9, is mixed with pCYPAC2, previously digested

with NheI. After filling-in of the ends, DNA ligase is added. The resulting mixture contains the desired plasmids pPAC-Sl and pPAC-S2.

Example 11 Construction of E. coli K12 DHlOB/pPAC-Sl and DH10B/pPAC-S2 Approximately 10 ng of plasmids pPAC-S1 and pPAC-S2, prepared as described in Example 10, are used to transform E. coli DH10B and a few of the resulting KmR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pPAC-S1, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 8.1,4.8,4.6,2.2,2.2,0.5 and 0.1 kb after EcoRI digestion of the plasmid. Another colony is found to carry pPAC-S2, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 8.1,7.8,2.2,2.2, 1.5,0.5 and 0.1 kb after EcoRI + BamHI digestion of the plasmid.

Although the present invention is described in the Examples listed above in terms of preferred embodiments, they are not to be regarded as limiting the scope of the invention. The above Examples serve to illustrate the principles and methodologies for constructing Bacterial Artificial Chromosomes that can be introduced in a Streptomyces host. It will occur to those skilled in the art that selectable markers different from the tsr gene can be employed for selection in Streptomyces. Other useful markers are described in detail in laboratory manuals (Hopwood et al., 1985, Genetic Manipulation of Streptomyces: A Laboratory Manual, The John Innes Foundation, Norwich, UK) and include but are not limited to: genes conferring resistance to apramycin, kanamycin,

erythromycin, hygromycin, viomycin. It will also occur to those skilled in the art that functions other than those specified by OC31 can be used for directing site-specific integration in the Streptomyces chromosome. These functions are described in recent literature (Hopwood and Kieser, 1991, Methods Enzymol. 204: 430- 458) and include but are not limited to those derived from pSAM2, SLP1, IS117.

Bacterial Artificial Chromosomes derived from the E. coli F plasmid have been described (Shizuya et al., 1992, Proc.

Natl. Acad. Sci. USA 89: 8794-8797). It will occur to those skilled in the art that, using the principles and methodologies described above, the int-tsr cassette from pUIT4, prepared as described in Example 9, could be inserted into a unique site of pBAC108L (Shizuya et al., 1992, Proc. Natl. Acad. Sci. USA 89: 8794-8797) or of suitable derivatives of this vector, leading to the formation of a BAC-based series of pESAC. It will occur to those skilled in the art that other pESACs differing, for example, in their size, in the E. coli replicon they carry, in the selectable marker for E. coli, in the cloning sites, can also be used in the present invention. Other differences and variations in the technical aspects of the present invention could be employed. These include but are not limited to: different methods and sources for obtaining selectable markers and integrative functions; different cloning sites and methodologies; different orientation of the insert; different E. coli hosts for amplifying the recombinant constructs. All these variations fall within the scope of the present invention.

7.2 Construction of large inserts in pESAC Two distinct methodologies for introducing large DNA

fragments into the vectors described in Section 7.1 fall within the scope of the present invention. The first methodology can be referred to as the top-down approach and is depicted in Fig. 3. It consists of directly cloning the desired gene cluster into a pESAC through the construction of a genomic library of DNA fragments of average size of 100 kb, or more. The library is then screened with suitable probes (Section 7.3) in order to identify the desired cluster. The second methodology can be considered a bottom- up approach and is illustrated in Fig. 4. It consists of assembling the desired gene cluster from pre-existing smaller segments of cloned, overlapping DNA, through the iterative use of homologous recombination in E. coli. The desired overlapping clones encompass the desired gene cluster and are identified as described in Section 7.3.

Both methodologies fall within the scope of this invention.

Depending on factors such as previous knowledge about the biosynthesis cluster, the extent of characterization of the producing strain, the existence of other natural products of interest produced by the original microorganism, one methodology may be preferred over the other. However, the two methodologies are not mutually exclusive.

7.2.1 Preparation of a large insert library In order to prepare a large-insert library, particular care must be taken in the preparation of genomic DNA from the actinomycete strain of choice. Although several procedures have been described for the isolation of genomic DNA, few are suitable for obtaining sufficient yields of high molecular weight DNA. The strain of choice is grown in a medium that allows dispersed growth to facilitate lysis of the cells. Examples of suitable growth media for different genera of actinomycetes can be found in the literature

(Balows, Trueper, Dworkin, Harder and Schleifer eds., 1992, The Prokaryotes, 2nd edn., Springer-Verlag, New York, NY, USA). The growth time should allow formation of a sufficient quantity of biomass; however, long incubation times should be avoided, since mycelia are generally more resistant to lysis as they age. The mycelium is pelleted, washed and embedded in agarose for the subsequent lytic steps. Lysis of the cells is achieved by a combination of enzymatic (e. g., incubation with lysozyme and/or achromopeptidase) and mild physical treatments (e. g., SDS).

The concentrations of reagents and the incubation times need to be optimized for each strain. A good starting point is represented by the conditions described in Example 12.

The quality of the DNA preparation is checked by PFGE under appropriate conditions. Once a suitable preparation is obtained, the DNA can be digested as described in Example 13. The exact incubation time and the units of restriction endonuclease are adjusted to the particular DNA preparation for optimizing the size and yield of the bulk of digested DNA, which should exceed 150 kb. The partially digested DNA is size-fractionated on a PFGE gel, without exposure to ethidium bromide or UV light, in order to avoid damage to the DNA. The gel slice containing the desired DNA fraction is localized by staining the marker-containing portion of the gel and cut. All subsequent manipulations are performed with great care (Birren and Lai, 1993, Pulsed Field Gel Electrophoresis: A Practical Guide, Academic Press, New York, NY). The size-selected DNA is ligated to an appropriately prepared pESAC (see Example 14) employing a high molar excess of vector to insert (ca. 10: 1) in order to minimize the formation of chimeric clones (i. e. those constituted by the religation of two uncontiguous inserts).

Subsequent steps are performed using published procedures

for the cloning in Bacterial Artificial Chromosomes, as described in Examples 16 and 17.

The genome size of actinomycetes is around 8 Mb.

Consequently, a 10-genome equivalents library consisting of 800 clones with an average insert size of 100 kb has >99.9% probability of containing the desired clone (Sambrook et al., 1989, In Molecular Cloning: A laboratory Manual, 2nd edn, Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press). Therefore, the average clone in the library will have a 10-kb segment (8,000 kb divided by 800 clones = 10 kb/clone) of unique DNA, i. e. DNA not found in any other clone. Consequently, a 90 kb cluster will have a high chance of being exactly contained within one or two 100-kb clones in a 800-clone library. The number of clones to be screened and the average insert size to be looked for in the ESAC library depends on the expected size of the biosynthesis gene cluster. The larger the difference between the average insert size and the expected size of the gene cluster, the smaller the number of clones to screen in order to identify an entire gene cluster in a single clone. ESAC DNA is prepared from a representative number of clones obtained after electroporation of a ligation mixture and analyzed for determining the frequency of insert-carrying clones and their average size. If necessary, all insert containing clones can be analyzed by miniprep procedure (Birren and Lai, 1993, Pulsed Field Gel Electrophoresis: A Practical Guide, Academic Press, New York, NY, USA) and clones carrying inserts below a certain threshold can be discarded. Alternatively, the number of clones carrying insert of the appropriate size can be estimated after analysis of a representative number of ESACs. The quality of the library can be evaluated by probing with cloned genes from the strain (if available),

or from highly conserved"housekeeping"genes from a strain with a similar GC content, such as S. coelicolor.

Example 12 Preparation of high molecular weight chromosomal DNA S. coelicolor strain M145 is grown in YEME medium containing 0.5% (wt/vol) glycine for 40 h at 30°C on an orbital shaker (ca. 200 rpm). The mycelium is pelleted by centrifugation, washed with 10.3% sucrose and the chromosomal DNA is extracted from the mycelium embedded in 0.75% LMP agarose by treatment with 1 mg/ml lysozyme and with 1 mg/ml proteinase K in 0.1% SDS for 40 h at 50°C.

Example 13 Preparation of partially digested chromosomal DNA S. coelicolor M145 chromosomal DNA, prepared as described in Example 12 and embedded in LMP agarose plugs, is partially digested by limiting the magnesium concentration for 20 min with 4 U of Sau3AI. The resulting DNA fragments are resolved by PFGE and the size-selected genomic DNA fraction (larger than 100 kb) is recovered and released from the agarose gel by digestion with gelase.

Example 14 Preparation of pPAC-S1 for library construction The vector pPAC-Sl, prepared as described in Example 11, is cut with ScaI and then treated with calf intestinal phosphatase. The recovered DNA is then digested with BamHI and treated with an excess of calf intestinal phosphatase.

The short ScaI-BamHI linker fragments are removed by spin dialysis.

Example 15 Construction of the ESAC library Size selected genomic DNA, prepared as described in Example 13, is ligated to pPAC-S1, prepared as described in Example 14, employing 300 Molecular Biology Units of T4 DNA ligase in a 50 tl final volume and using a ca. 10: 1 molar ratio of vector to insert. The resulting ligation mixture contains the desired ESAC library, consisting of fragments S. coelicolor DNA inserted into the pPAC-S1 vector.

Example 16 Introduction of the library into E. coli K12 DH10B The ligation mixture, prepared as described in Example 15, is drop-dialyzed against 0.5 X TE for 2 h using 0.025 mm type VS membranes (Millipore) and a few Ll are used to electroporate 40 pu ouf electrocompetent E. coli DH10B cells. The electroporation conditions are: 2.5 kV, 100 Q and 25 mFa employing the Biorad Gene Pulser II. The cells are plated on LB-agar plates containing 25 g/ml Km and 5% sucrose to select for recombinant cells harboring insert- carrying pPAC-S1. Individual colonies are picked into 0.1 ml of LB broth containing 25 pg/ml Km in 96-well microtiter plates, where they are stored at-80 °C after overnight incubation and addition of glycerol to 20% (v/v).

Example 17 Preparation of recombinant ESACs Individual colonies, prepared as described in Example 16, are inoculated into 5 ml of LB broth containing 25 g/ml Km and grown overnight. ESAC DNA is isolated using the alkaline extraction procedure (Sambrook et al., 1989, In Molecular Cloning: A laboratory Manual, 2nd edn, Cold

Spring Harbor, New York: Cold Spring Harbor Laboratory Press)) without the phenol extraction step. The DNA is analyzed, after digestion with DraI, by PFGE. Three bands of 7.4,4.2 and 0.6 kb are common to all clones and represent vector DNA.

The examples described above illustrate the principles and methodologies of constructing a large-insert library of S. coelicolor DNA in a pESAC. Although the present invention is described in the Examples listed above in terms of preferred embodiments, they are not to be regarded as limiting the scope of the invention. The above descriptions serve to illustrate the principles and methodologies for constructing a large-insert DNA library in a pESAC. It will occur to those skilled in the art that other Streptomyces strains can be used as a source of DNA for constructing the library. For example, an ESAC library of the rapamycin producer Streptomyces hygroscopicus ATCC 29253 can be constructed, employing the procedures reported for PFGE analysis (Ruan et al., 1997, Gene 203: 1-9) and applying the principles and methodologies described in Examples 12 through 17.

It will also occur to those skilled in the art that strains from actinomycete genera other than Streptomyces can be used as a source of DNA for constructing an ESAC library. These strains can belong to any genus of the order Actinomycetales, which include but are not limited to the genera reported in Table 1. As another example, an ESAC library of the erythromycin producer Saccharopolyspora erythraea can be constructed, employing the procedures reported for PFGE analysis (Reeves et al., 1998, Microbiology 144: 2151-2159) and applying the principles and methodologies described in Examples 12 through 17. Those

skilled in the art understand that bacterial taxonomy is a rapidly evolving field and new genera may be described while old genera may be reclassified. Therefore, the list of bacteria genera related to actinomycetes is very likely to change. Nonetheless, the principles and methodologies of the present invention can be applied to any donor organism related to the actinomycetes.

It will also occur to those skilled in the art that different actinomycete strains will require growth media different from those reported in Example 12. Furthermore, alternative media and conditions for growth can be employed for obtaining mycelia for DNA preparation; that alternative methods of lysis of mycelia can be utilized; that restriction endonucleases other than Sau3AI can be equally effective for constructing a library; that other methods for fragmenting DNA can be employed. In addition, it will occur to those skilled in the art that pESAC other than pPAC-S1, which include but are not limited to the possible vectors described in Section 7.1, can be used for constructing a library. Alternative methods for ligating DNA, for introducing the library in E. coli cells, and hosts other than DH10B are well described in the literature and can be employed in the present invention. All the above variations in strains, reagents and methodologies that can be employed for preparing a large-insert library of actinomycete DNA into a pESAC fall within the scope of the present invention.

Table1 List of exemplary genera of Actinomycetales Acidothermus Cellulomonas Kineococcus Actinobispora Chainia Kineosporia Actinocorallia Clavibacter Kitasatoa Actinokineospora Coriobacterium Kitasatosporia Actinomadura Corynebacterium Kocuria Actinomyces Couchioplanes Kutzneria Actinoplanes Cryobacterium Kytococcus Actinopolyspora Curtobacterium Lentzea Actinopycnidium Dactylosporangium Luteococcus Actinosporangim Demetria Microbacterium Actinosynnema Dermabacter Microbispora Aeromicrobium Dermacoccus Micrococcus Agrococcus Dermatophilus Microellobosporia Agromyces Dietzia Microlunatus Ampullariella Elytrosporangium Micromonospora Amycolata Excellospora Microsphaera Amycolatopsis Exiguobacterium Micro raspora Arcanobacterium Frankia Microthrix Arthrobacter Friedmanniella Mobiluncus Atopobium Gardnerella Mycobacterium Aureobacterium Geodermatophilus Nesterenkonia Bifidobacterium Glycomyces Nocardia Blastococcus Gordona Nocardioides Bogoriella Herbidospora Nocardiopsis Brachybacterium Intrasporangium Oerskovia Brevibacterium Janibacter Pelczaria Catellatospora Jonesia Phenylobacterium Catenuloplanes Kibdelosporangium Pilimelia Pimelobacter Rhodococcus Streptomyces Planobispora Rothia Streptosporangium Planomonospora Rubrobacter Streptoverticillium Planopolyspora Saccharomonospora Terrabacter Planotetaspora Saccharopolyspora Terracoccus Prauseria Saccharothrix Thermoactinomyces Promicromonospora Sanguibacter Thermocrispum Propionibacterium Skermania Thermomonospora Propioniferax Spirilliplanes Tropheryma Pseudonocardia Spirillospora Tsukamurella Rarobacter Sporichthya Turicella Rathayibacter Stomatococcus Renibacterium Streptoalloteichus

7.2.2 Assemblage by homologous recombination The bottom-up strategy of assembling large fragments from a set of pre-existing smaller segments of partially overlapping DNA cloned from the genome of the actinomycete donor organism, is described in this section. This methodology makes use of the same pESAC described in the present invention under Section 7.1. The desired cluster is assembled from existing partially overlapping clones by the iterative use of homologous recombination in E. coli. In the example of Fig. 5, three overlapping clones, designated 1,2 and 3, and derived from the genome of a donor organism, encompass the desired biosynthesis cluster. These clones include leftward fragment"A"unique to clone 1; fragment"B"common to clones 1 and 2; fragment"C"common to clones 2 and 3 ; and rightward fragment"D"unique to clone 3. These fragments can range from a few hundred bp to a few kb, and are thus amenable to routine in vitro DNA manipulations. The number of overlapping clones encompassing the cluster may vary. However, if n is the number of overlapping clones that cover the desired genomic segment, the number of fragments to consider will be equal to n + 1. In the example illustrated in Fig. 5, four fragments are required. The cluster of Fig. 5 is reconstructed into a pESAC through the use of successive rounds of homologous recombination in E. coli. Fragments A and B are cloned in a ts vector, as shown in Fig. 6, which carries a selectable marker, CmR as exemplified in Fig. 6.

The same is done with fragment pairs B-C and C-D (Fig. 6).

The relative orientation of each fragment pair in the ts vector must be the same as in the gene cluster. The fragments in each pair may be separated by a selectable marker, designated AbR in Fig. 6, to monitor interplasmid

insert exchange. Therefore, three constructs in the ts vector, designated pAB1, pAB2 and pAB3, are required. The A-B-C-D four-fragment cassette is cloned in a pESAC (Fig.

6). The relative orientation of the four fragments in the pESAC must be the same as in the gene cluster. Again, a selectable marker may separate any of two fragments to monitor interplasmid insert exchange. The original clone (for example, a cosmid, which carries a selectable marker, KmR as exemplified in Fig. 7) containing part of the cluster and the cognate ts construct (Fig. 7) are introduced into the same E. coli cell. The interplasmid cointegrate between the original clone and the ts construct is selected at the non-permissive temperature for the ts replicon. This occurs through a single, reciprocal homologous recombination mediated by either one of the two fragments in the A-B, B-C or C-D pairs. The cointegrate is then resolved at the permissive temperature, leading to insert exchange between the two replicons (Fig. 7). The presence in the ts replicon of the genomic segment comprised between fragments A and B can be monitored by the appearance of CmR Abs colonies. This is done for clone 1 and pAB1, resulting in pAB2 ; for clone 2 and pBC1, resulting in pBC2 ; and for clone 3 and pCD1, resulting in pCD2. Each insert from the original overlapping clones (Fig. 5) is thus transferred into the ts replicon, as outlined in Fig. 7. Subsequently, the inserts from clone 1, now present in the ts plasmid pAB2, is introduced into the pESAC construct carrying the entire A- B-C-D cassette. This is done by selecting for the interplasmid cointegrate between pAB2 and the pESAC construct at the non-permissive temperature, and then resolving the cointegrate at the permissive temperature, selecting for KmR Abs colonies. This leads to insert exchange between the two replicons (as shown in Fig. 8).

Next, a selectable marker is introduced in the growing ESAC between the next fragment pair, again through the use of two rounds of single, reciprocal homologous recombination mediated by plasmid pBC1, leading to the appearance of KmR AbR colonies. Subsequently, the interplasmid exchange with pBC2 leads to the introduction of the genomic segment comprised between fragments B and C. Finally, the use of pCD1 first and subsequently of pCD2 leads to the reconstruction of the genomic segment into the pESAC.

Therefore, through the use of alternating steps, the AbR marker first and the genomic segment later are introduced between any fragment pair, as schematized in Fig. 8. This iterative procedure results in the reconstruction of the original chromosomal region in the pESAC.

A series of examples described herein illustrate how a 90-kb gene segment from the actinomycete P. rosea is assembled from three pre-existing cosmids via homologous recombination. The cosmids, designated pRP16, pRP31 and pRP58, are identified in a cosmid library constructed in the vector Lorist6 by the use of selective hybridization probes. The relevant information about the cluster is reported in Fig. 9. The reconstruction of the cluster results in the formation of the intermediate derivatives pPAD1, PAD2, PAD4 and PAD6, carrying inserts of 10,39,68 and 89 kb, respectively. The examples reported herein serve to illustrate the principles and methodologies of the present invention and are not meant to restrict its scope.

Example 18 Isolation of cosmid clones pRP16, pRP31 and pRP58 A cosmid library of P. rosea DNA prepared in the vector Lorist6 is screened with oligonucleotide probes Pep6 and Pep8, according to the conditions described under Section

6. Among the positive colonies identified, several cosmids were found to span the ca. 90 kb segment of the P. rosea chromosome reported in Fig. 9. Signature sequences close to the left and right ends of this segment are reported as SEQID N°9 and SEQID N°10, respectively. Three cosmids are chosen for further studies. Cosmids pRP16, pRP31 and pRP58 exhibits, after digestion with BamHI and resolution by agarose gel-electrophoresis, fragments of 7.5,7.2,5.6, 5.2,2.7,2.0,1.9,1.9,1.8,1.6,1.4,0.9 and 0.7 kb; of 10.5,6.2,3.1,2.8,2.6,2.5,2.1,1.9,1.9,1.5,1.4, 1.2,1.0,1.0,0.9,0.9,0.7,0.6,0.5,0.5,0.1 and 0.1 kb; and of 10.0,7.6,6.7,6.2,3.4,3.0,2.8,2.1,1.0, 1.0,0.9,0.9,0.7,0.6,0.5,0.5,0.3 and 0.1 kb ; respectively.

Example 19 Construction of plasmid pUA1 The 0.9 kb SmaI-SstI fragment, comprised between map coordinates 2.0-2.9 kb of Fig. 9, is obtained from cosmid pRP16, prepared as described in Example 18, and ligated to pUC18 previously digested with SstI and SmaI. The resulting mixture contains the desired plasmid pUAl.

Example 20 Construction of E. coli K12 XLlblue/pUA1 Approximately 10 ng of plasmid pUAl, prepared as described in Example 19, are used to transform E. coli XLlblue and a few of the resulting ApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUAl, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 2.7 and 0.9 kb after digestion of the plasmid with BamHI + SstI.

Example 21 Construction of plasmid pUA2 The 0.9 kb BamHI-SstI fragment from pUAl, prepared as described in Example 20, is ligated to pUCBM20 previously digested with BamHI and SstI. The resulting mixture contains the desired plasmid pUA2.

Example 22 Construction of E. coli K12 XLlblue/pUA2 Approximately 10 ng of plasmid pUA2, prepared as described in Example 21, are used to transform E. coli XLlblue and a few of the resulting ApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUA2, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 2.7 and 0.9 kb after digestion of the plasmid with EcoRI + SstI.

Example 23 Construction of plasmid pUB1 The 1.8 kb SstI-BamHI fragment, comprised between map coordinates 33.4-35.2 of Fig. 9, is obtained from cosmid pRP16, prepared as described in Example 18, and ligated to pUC18 previously digested with SstI + BamHI. The ligation mixture contains the desired plasmid pUB1.

Example 24 Construction of E. coli K12 XLlblue/pUB1 Approximately 10 ng of plasmid pUB1, prepared as described in Example 23, are used to transform E. coli XLlblue and a few of the resulting ApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to contain plasmid pUB1 as verified by the observation, upon agarose gel electrophoresis, of fragments

2.7 and 1.8 kb after digestion with SstI + XbaI.

Example 25 Construction of plasmid pUC1 The 6.2 kb BamHI fragment, comprised between map coordinates 54.2-60.4 kb of Fig. 9, is obtained from cosmid pRP58, prepared as described in Example 18, and ligated to pUC18 previously digested with BamHI. The ligation mixture contains the desired plasmid pUCl.

Example 26 Construction of E. coli K12 XLlblue/pUC1 Approximately 10 ng of plasmid pUC1, prepared as described in Example 25, are used to transform E. coli XLlblue and a few of the resulting ApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUC1, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 4.9 and 4.0 kb after digestion of the plasmid with PstI.

Example 27 Construction of plasmid pUD1 Synthetic oligonucleotides 5'-GATCTAAGCTTGGGGG-3' (SEQID N°11) and 5'-CCCCCAAGCTTA-3' (SEQID N°12) are annealed and ligated to the 1.5 kb PstI-BamHI fragment, comprised between map coordinates 89.5-91.0 kb of Fig. 9 and obtained from cosmid pRP58, prepared as described in Example 18. The ligation mixture is digested with HindIII and ligated to pUC18 previously digested with PstI + HindIII. The resulting mixture contains the desired plasmid pUD1.

Example 28 Construction of E. coli K12 XLlblue/pUD1

Approximately 10 ng of plasmid pUD1, prepared as described in Example 27, are used to transform E. coli XLlblue and a few of the resulting ApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to contain plasmid pUD1 as verified by the observation, upon agarose gel-electrophoresis, of fragments of 2.7 and 1.5 kb after digestion with PstI + HindIII.

Example 29 Construction of plasmid pUAB1 The 0.9 kb EcoRI-SstI fragment from plasmid pUA2, prepared as described in Example 22, and the 1.8 kb SstI-BamHI fragment from pUB1, prepared as described in Example 24, are ligated to pUC18 previously digested with EcoRI + BamHI. The ligation mixture contains the desired plasmid pUAB1.

Example 30 Construction of E. coli K12 XLlblue/pUAB1 Approximately 10 ng of plasmid pUAB1, prepared as described in Example 29, are used to transform E. coli XLlblue and a few of the resulting ApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUAB1, as verified by the observation, upon agarose gel-electrophoresis, of two fragments of 2.7 kb after digestion of the plasmid with EcoRI + XbaI.

Example 31 Isolation of the tet fragment The 1.6 kb fragment containing the tet gene is isolated after PCR amplification of pBR322 DNA using oligonucleotides 5'-GAGCTCTCATGTTTGACAGCT-3'(SEQID N°13) and 5'-GAGCTCTGACTTCCGCGTTTCCAG-3' (SEQID N°14) as primers,

followed by digestion with SstI.

Example 32 Construction of plasmid pUAB2 Plasmid pUAB1, prepared as described in Example 30, is digested with SstI and ligated to the tet fragment prepared as described in Example 31. The ligation mixture contains the desired plasmid pUAB2.

Example 33 Construction of E. coli K12 DH5a/pUAB2 Approximately 10 ng of plasmid pUAB2, prepared as described in Example 32, are used to transform E. coli DH5a and a few of the resulting TcRApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUAB2, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 4.3 and 2.7 kb after digestion of the plasmid with EcoRI + XbaI.

Example 34 Construction of plasmid pUBC1 The 1.8 kb SstI-XbaI fragment obtained from plasmid pUB1, prepared as described in Example 24, and the 4.0 kb XbaI- PstI fragment obtained from plasmid pUC1, prepared as described in Example 26, are ligated to pUC18 previously digested with SstI + PstI. The ligation mixture contains the desired plasmid pUBC1.

Example 35 Construction of E. coli K12 XLlblue/pUBC1 Approximately 10 ng of plasmid pUBC1, prepared as described in Example 34, are used to transform E. coli XLlblue and a

few of the resulting ApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUBCl, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 5.8 and 2.7 kb after digestion of the plasmid with EcoRI + HindIII.

Example 36 Construction of plasmid pUBC2 Plasmid pUBC1, prepared as described in Example 35 and previously digested with XbaI, and the tet fragment, prepared as described in Example 31, are treated with T4 DNA polymerase and T4 DNA ligase. The ligation mixture contains the desired plasmid pUBC2.

Example 37 Construction of E. coli K12 DH5a/pUBC2 Approximately 10 ng of plasmid pUBC2, prepared as described in Example 36, are used to transform E. coli DH5a and a few of the resulting TcRApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUBC2, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 5.6 and 4.5 kb after digestion of the plasmid with HindIII.

Example 38 Construction of plasmid pUCD1 The 4.0 kb XbaI-PstI fragment obtained from plasmid pUC1, prepared as described in Example 26, and the 1.5 kb PstI- HindIII fragment isolated from plasmid pUD1, prepared as described in Example 28, are ligated to pUC18 previously digested with XbaI and HindIII. The mixture contains the desired plasmid pUCD1.

Example 39 Construction of E. coli K12 XLlblue/pUCD1 Approximately 10 ng of plasmid pUCD1, prepared as described in Example 38, are used to transform E. coli XLlblue and a few of the resulting ApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUCD1, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 5.5 and 2.7 kb after digestion of the plasmid with XbaI + HindIII.

Example 40 Construction of plasmid pUCD2 Plasmid pUCD1, prepared as described in Example 39 and previously digested with PstI, and the tet fragment prepared as described in Example 31, are treated with T4 DNA polymerase and T4 DNA ligase. The ligation mixture contains the desired plasmid pUCD2.

Example 41 Construction of E. coli K12 DH5a/pUCD2 Approximately 10 ng of plasmid pUCD2, prepared as described in Example 40, are used to transform E. coli DH5a and a few of the resulting TcRApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUCD2, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 6.7 and 3.1 kb after digestion of the plasmid with HindIII.

Example 42 Construction of plasmid pUAD1 The 4.3 kb EcoRI-XbaI fragment obtained from plasmid pUAB2,

prepared as described in Example 33, and the 5.5 XbaI- HindIII fragment from plasmid pUCD1, prepared as described in Example 39, are ligated to pUC18, previously digested with EcoRI + HindIII. The ligation mixture contains the desired plasmid pUADl.

Example 43 Construction of E. coli K12 DH5a/pUAD1 Approximately 10 ng of plasmid pUAD1, prepared as described in Example 42, are used to transform E. coli DH5a and a few of the resulting TcRApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUAD1, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 8.9 and 3.6 kb after digestion of the plasmid with HindIII.

Example 44 Construction of plasmid pMAB1 The 4.3 kb EcoRI-XbaI fragment obtained from plasmid pUAB2, prepared as described in Example 33, is treated with T4 DNA Polymerase and ligated to pMAK705 previously digested with HincII. The ligation mixture contains the desired plasmid pMAB1.

Example 45 Construction of E. coli K12 C600/pMAB1 Approximately 10 ng of plasmid pMAB1, prepared as described in Example 44, are used to transform E. coli C600 and a few of the resulting CmRTcR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pMAB1, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 4.1,3.4,

1.4 and 0.9 kb after digestion of the plasmid with HindIII + EcoRI.

Example 46 Construction of plasmid pMBC1 The 7.1 kb fragment from plasmid pUBC2, prepared as described in Example 37, is obtained after partial digestion with PstI, treated with T4 DNA polymerase and ligated to pMAK705 previously digested with HincII. The ligation mixture contains the desired plasmid pMBC1.

Example 47 Construction of E. coli K12 C600/pMBC1 Approximately 10 ng of plasmid pMBC1, prepared as described in Example 46, are used to transform E. coli C600 and a few of the resulting CmRTcR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pMBC1, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 9.5,1.5, 1.3 and 0.3 kb after digestion of the plasmid with BamHI.

Example 48 Construction of plasmid pMCD1 The 7.1 kb fragment from plasmid pUCD2, prepared as described in Example 41, is obtained by complete digestion with XbaI and partial digestion with HindIII, treated with T4 DNA polymerase and ligated to pMAK705, previously digested with HincII. The ligation mixture contains the desired plasmid pMCD1.

Example 49 Construction of E. coli K12 C600/pMCD1 Approximately 10 ng of plasmid pMCD1, prepared as described

in Example 48, are used to transform E. coli C600 and a few of the resulting CmRTcR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pMCD1, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 8.6 and 4.3 kb after digestion of the plasmid with BamHI.

Example 50 Construction of plasmid pPAD1 The 10.0 kb EcoRI-NdeI fragment from plasmid pUAD1, prepared as described in Example 43, is ligated to pPAC-S1, prepared as described in Example 11 and previously digested with ScaI. The ligation mixture contains the desired plasmid pPAD1.

Example 51 Construction of E. coli K12 C600/pPAD1 Approximately 10 ng of plasmid pPAD1, prepared as described in Example 50, are used to transform E. coli C600 and a few of the resulting KmRTcR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pPAD1, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 19.7, 5.8,3.1 and 1.2 kb after digestion of the plasmid with BamHI. After digestion with DraI and resolution by PFGE, pPAD1 yields fragments of 17.4,7.4,4.2 and 0.6 kb.

Example 52 Construction of E. coli K12 C600/pMABl:: pRP16 E. coli C600/pMAB1, prepared as described in Example 45, is transformed with ca. 50 ng of pRP16, prepared as described in Example 18. The CmRKmR colonies that appear at 30 °C on

the LB-agar plates are grown at 30°C in LB broth containing Km and Cm, aliquots are withdrawn at various times and appropriate dilutions plated. Few of the CmRKmR colonies that appear on the LB-agar plates after overnight incubation at 44°C are grown in LB broth containing Km and Cm for 16 h at 44°C and analyzed for their plasmid content.

One colony is found to carry pMABl:: pRP16, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 34,10.7,1.7,1.6,1.5,1.2 and 0.6 kb after digestion of the plasmid with EcoRI.

Example 53 Construction of E. coli K12 C600/pMAB2 Several colonies of E. coli C600/pMABl:: pRP16, prepared as described in Example 52, are grown individually in LB broth containing Cm for 24 h at 30°C, diluted 1: 100 and incubated for further 8 h. Appropriate dilutions are plated. Few of the resulting CmRKmSTcs colonies that appear at 30°C are analyzed for their plasmid content. One colony is found to carry pMAB2, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 37 and 1.5 kb after digestion of the plasmid with EcoRI.

Example 54 Construction of E. coli K12 DH1/pMBCl:: pRP31 Approximately 50 ng of pRP31, prepared as described in Example 18, are used to transform E. coli DH1 cells.

Competent cells are prepared from one of the resulting KmR colonies and transformed with ca. 10 ng of plasmid pMBC1, prepared as described in Example 47. The CmRKmR colonies that appear at 30°C on the LB-agar plates are grown at 30°C in LB broth containing Km and Cm, aliquots are withdrawn at

various times and appropriate dilutions plated. Few of the CmRKmR colonies that appear on the LB-agar plates after overnight incubation at 44°C are grown in LB broth containing Km and Cm for 16 h at 44°C and analyzed for their plasmid content. One colony is found to carry pMBCl:: pRP31, as verified by the observation, upon agarose gel- electrophoresis, of fragments of 22.2,14.1,14.0 and 6.0 kb after digestion of the plasmid with EcoRV.

Example 55 Construction of E. coli K12 DH1/pMBC2 Several colonies of E. coli DH1/pMBCl:: pRP31, prepared as described in Example 54, are grown individually in LB broth containing Cm for 24 h at 30°C, diluted 1: 100 and incubated for further 8 h. Appropriate dilutions are plated. Few of the resulting CmRKmSTcs colonies that appear at 30°C are analyzed for their plasmid content. One colony is found to carry pMBC2, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 14.4,14.1 and 1.5 kb after digestion of the plasmid with EcoRI.

Example 56 Construction of E. coli K12 DH1/pMCDl:: pRP58 Approximately 50 ng of pRP58, prepared as described in Example 18, are used to transform E. coli DH1 cells.

Competent cells are prepared from one of the resulting KmR colonies and transformed with ca. 10 ng of plasmid pMCD1, prepared as described in Example 48. The CmRKmR colonies that appear at 30°C on the LB-agar plates are grown at 30°C in LB broth containing Km and Cm, aliquots are withdrawn at various times and appropriate dilutions plated. Few of the CmRKmR colonies that appear on the LB-agar plates after

overnight incubation at 44°C are grown in LB broth containing Km and Cm for 16 h at 44°C and analyzed for their plasmid content. One colony is found to carry pMCDl:: pRP58, as verified by the observation, upon agarose gel- electrophoresis, of fragments of 39,16,1.7,1.6,1.5,1.2 and 0.6 kb after digestion of the plasmid with EcoRI.

Example 57 Construction of E. coli K12 DH1/pMCD2 Several colonies of E. coli DH1/pMCDl:: pRP58, prepared as described in Example 56, are grown individually in LB broth containing Cm for 24 h at 30°C, diluted 1: 100 and incubated for further 8 h. Appropriate dilutions are plated. Few of the resulting CmRKmSTcs colonies that appear at 30°C are analyzed for their plasmid content. One colony is found to carry pMCD2, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 42 and 1.5 kb after digestion of the plasmid with EcoRI.

Example 58 Construction of E. coli K12 C600/pMAB2:: pPAD1 E. coli C600/pMAB2, prepared as described in Example 53, is transformed with ca. 50 ng of plasmid pPAD1, prepared as described in Example 51. The CmRKmR colonies that appear at 30°C on the LB-agar plates are grown at 30°C in LB broth containing Km and Cm, aliquots are withdrawn at various times and appropriate dilutions plated. Few of the CmRKmR colonies that appear on the LB-agar plates after overnight incubation at 44°C are grown in LB broth containing Km and Cm for 16 h at 44°C and analyzed for their plasmid content.

One colony is found to carry pMAB2:: pPAD1, as verified by the observation, upon agarose gel-electrophoresis, of

fragments of 19.7,7.2,5.6,5.6,5.5,5.2,3.1,2.7,1.9, 1.9,1.8,1.8,1.6,1.4,1.2,0.9,0.9 and 0.7 kb after digestion of the plasmid with BamHI.

Example 59 Construction of E. coli K12 C600/PAD2 Several colonies of E. coli C600/pMAB2:: pPAD1, prepared as described in Example 58, are grown individually in LB containing Km for 24 h at 30°C, diluted 1: 100 and incubated for further 8 h at 44°C. Appropriate dilutions are plated.

Few of the resulting KmRCmSTcs colonies that appear at 37°C are analyzed for their plasmid content. One colony is found to carry PAD2, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 19.7,7.2,5.6,5.5, 5.2,2.7,1.9,1.9,1.8,1.8,1.6,1.4,0.9 and 0.7 kb after digestion of the plasmid with BamHI. After DraI digestion and resolution by PFGE, PAD2 yields fragments of 45,7.4,4.2 and 0.6 kb.

Example 60 Construction of plasmid pMCD3 The 1.4 kb KpnI-XhoII fragment obtained from plasmid pCYPAC2 after digestion with XhoII, treatment with T4 DNA polymerase and digestion with KpnI, and the 7.1 kb XbaI- HindIII fragment from pUCD2, prepared as described in Example 40 and obtained after partial digestion with HindIII, complete digestion with XbaI and treatment with T4 DNA polymerase, are ligated to pMAK705, previously digested with KpnI + HincII. The ligation mixture contains the desired plasmid pMCD3.

Example 61

Construction of E. coli K12 C600/pMCD3 Approximately 10 ng of plasmid pMCD3, prepared as described in Example 60, are used to transform E. coli C600 and a few of the resulting CmRTcR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pMCD3, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 9.8 and 4.3 kb after digestion of the plasmid with BamHI.

Example 62 Construction of E. coli K12 C600/PAD2:: pMCD3 E. coli C600/PAD2, prepared as described in Example 59, is transformed with ca. 10 ng of plasmid pMCD3, prepared as described in Example 61. The CmRKmR colonies that appear at 30°C on the LB-agar plates are grown at 30°C in LB broth containing Km and Cm, aliquots are withdrawn at various times and appropriate dilutions plated. Few of the CmRKmR colonies that appear on the LB-agar plates after overnight incubation at 44°C are grown in LB broth containing Km and Cm for 16 h at 44°C and analyzed for their plasmid content.

One colony is found to carry PAD2:: pMCD3, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 19.7,9.8,7.2,5.6,5.5,5.2,4.3,2.7,1.9, 1.9,1.8,1.8,1.6,1.4,0.9 and 0.7 kb after digestion of the plasmid with BamHI.

Example 63 Construction of E. coli K12 C600/PAD3 Several colonies of E. coli C600/PAD2:: pMCD3, prepared as described in Example 62, are grown individually in LB broth containing Km for 24 h at 30°C, diluted 1: 100 and incubated for further 8 h at 44°C. Appropriate dilutions are plated.

Few of the resulting KmRCmSTcR colonies are analyzed for their plasmid content. One colony is found to carry PAD3, as verified by the observation, upon agarose gel- electrophoresis, of fragments of 21.5,7.2,5.6,5.2,4.3, 2.7,1.9,1.9,1.8,1.8,1.6,1.4,0.9 and 0.7 kb after digestion of the plasmid with BamHI.

Example 64 Construction of E. coli K12 C600/PAD3:: pMCD2 E. coli C600/PAD3, prepared as described in Example 63, is transformed with ca. 50 ng of plasmid pMCD2, prepared as described in Example 57. The CmRKmR colonies that appear at 30°C on the LB-agar plates are grown at 30°C in LB broth containing Km and Cm, aliquots are withdrawn at various times and appropriate dilutions plated. Few of the CmRKmR colonies that appear on the LB-agar plates after overnight incubation at 44°C are grown in LB broth containing Km and Cm for 16 h at 44°C and analyzed for their plasmid content.

One colony is found to carry PAD3:: pMCD2, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 21.5,10,9,0,7.6,7.2,6.2,5.6,5.2,4.3, 8,2.7,1.9,1.9,1.8,1.8,1.6,1.4,1.0,0.9,0.9, 3 and 0.1 kb after digestion of the plasmid with BamHI.

Example 65 Construction of E. coli K12 C600/PAD4 Several colonies of E. coli C600/PAD3:: pMCD2, prepared as described in Example 64, are grown individually in LB broth containing Km for 24 h at 30°C, diluted 1: 100 and incubated for further 8 h at 44°C. Appropriate dilutions are plated.

Few of the resulting KmRCmSTcs colonies are analyzed for

their plasmid content. One colony is found to carry PAD4, as verified by the observation, upon agarose gel- electrophoresis, of fragments of 22,10,7.6,7.2,6.2, 5.6,5.2,3.1,2.8,2.7,1.9,1.9,1.8,1.8,1.6,1.4,1.0, 0.9,0.9,0.9,0.7,0.5,0.3 and 0.1 kb after digestion of the plasmid with BamHI. After DraI digestion and resolution by PFGE, PAD4 yields fragments of 79,4.2 and 0.6 kb.

Example 66 Construction of E. coli K12 C600/PAD4:: pMBC1 E. coli C600/PAD4, prepared as described in Example 65, is transformed with ca. 10 ng of plasmid pMBC1, prepared as described in Example 47. The CmRKmR colonies that appear at 30°C on the LB-agar plates are grown at 30°C in LB broth containing Km and Cm, aliquots are withdrawn at various times and appropriate dilutions plated. Few of the CmRKmR colonies that appear on the LB-agar plates after overnight incubation at 44°C are grown in LB broth containing Km and Cm for 16 h at 44°C and analyzed for their plasmid content.

One colony is found to carry PAD4:: pMBC1, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 22,10,9.6,7.6,7.2,6.2,5.6,5.2,3.1, 2.8,2.7,1.9,1.9,1.8,1.8,1.6,1.5,1.4,1.3,1.0,0.9, 9,0.7,0.5,0.3,0.3 and 0.1 kb after digestion of the plasmid with BamHI.

Example 67 Construction of E. coli K12 C600/PAD5 Several colonies of E. coli C600/PAD4:: pMBC1, prepared as described in Example 66, are grown individually in LB broth containing Km for 24 h at 30°C, diluted 1: 100 and incubated for further 8 h at 44°C. Appropriate dilutions are plated.

Few of the resulting KmR CmS TcR colonies are analyzed for their plasmid content. One colony is found to carry PAD5, as verified by the observation, upon agarose gel- electrophoresis, of fragments of 22,10,7.6,7.2,6.2, 7,1.9,1.9,1.8,1.8,1.6,1.4,1.3, 1.0,0.9,0.9,0.9,0.7,0.5,0.3,0.3 and 0.1 kb after digestion of the plasmid with BamHI.

Example 68 Construction of E. coli K12 C600/PAD5:: pMBC2 E. coli C600/PAD5, prepared as described in Example 67, is transformed with ca. 50 ng of plasmid pMBC2, prepared as described in Example 55. The CmRKmR colonies that appear at 30°C on the LB-agar plates are grown at 30°C in LB broth containing Km and Cm, aliquots are withdrawn at various times and appropriate dilutions plated. Few of the CmRKmR colonies that appear on the LB-agar plates after overnight incubation at 44°C are grown in LB broth containing Km and Cm for 16 h at 44°C and analyzed for their plasmid content.

One colony is found to carry PAD5:: pMBC2, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 6,4.7,3.4,2.8,2.1,1.2,1.2,1.0 and 0.4 kb after digestion of the plasmid with HindIII.

Example 69 Construction of E. coli K12 C600/PAD6 Several colonies of E. coli C600/PAD5:: pMBC2, prepared as described in Example 68, are grown individually in LB broth containing Km for 24 h at 30°C, diluted 1: 100 and incubated for further 8 h at 44°C. Appropriate dilutions are plated.

Few of the resulting KmRCmSTcs colonies are analyzed for their plasmid content. One colony is found to carry the

correct ESAC, designated PAD6, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 47,46,8.1,4.6,2.2,0.5 and 0.1 kb after digestion of the plasmid with EcoRI. After digestion with DraI and resolution by PFGE, PAD6 yields fragments of 102,4.2 and 0.6 kb.

Although the present invention is described in the Examples listed above in terms of preferred embodiments, they are not to be regarded as limiting the scope of the invention. The above Examples serve to illustrate the principles and methodologies for assembling pre-existing overlapping segments of DNA into pESAC.

It will occur to those skilled in the art that the cluster of Fig. 9 can be assembled using A-B-C-D fragments other than those specified in the Examples. Any A fragment, such that no useful genes are present to its left (using the orientation of Fig. 9) can be used for assembling the cluster. Similarly, any D fragment, such that no useful genes are present to its right (using the orientation of Fig. 9) can also be used. Furthermore, any fragment common to pRP16 and pRP31, or to pRP31 and pRP58, can be used in place of the fragments B and C, respectively, described above. It will also occur to those skilled in the art that other methods for obtaining these fragments, such as use of different segments from the cluster of Fig. 9, of different restriction endonucleases, or of the PCR, can be used for achieving equivalent results. In addition, intermediate vectors, other than the pUC-series used in the above Examples, can be used for subcloning fragments A through D, and the use of these intermediate vectors is merely instrumental to the transfer of the fragment pairs into the ts vector. Some or all of the fragment pairs could

therefore be cloned directly into a ts vector.

It will also occur to those skilled in the art that cosmids other than pRP16, pRP31 and pRP58 can be used to achieve equivalent results, provided that they encompass the entire gene cluster and they have overlapping segments.

It will also occur to those skilled in the art that pMAK705, Lorist6 and pPAC-S1, are merely examples of ts, cosmid and pESAC, respectively. Any of the several cosmid vectors described in the literature, other ts replicons derived from pMAK705 or other sources, and any of the pESAC other than pPAC-S1, which include the possible vectors described in Section 7.1, can be used for obtaining equivalent results.

Those skilled in the art understand that the purpose of a ts replicon is to select for interplasmid cointegrates at the non-permissive temperature. However, cointegrate formation can occur between any two replicons, and cointegrate can be isolated after transformation of suitable hosts with a plasmid preparation made from an E. coli cell harboring both replicons. Selection for the antibiotic resistance markers carried by both replicons can lead to the isolation of cointegrates from the resulting transformants.

Furthermore, it will occur to those skilled in the art that the inclusion of the tet marker between the A-B, B-C and C-D fragment pairs serves solely the scope of recognizing insert exchange after resolution of the interplasmid cointegrate. Selectable markers other than tet can be equally effective, as long as they are not present in the vectors. Those skilled in the art understand that the presence of a selectable marker within the fragment cassettes is not absolutely necessary, as insert exchange can be observed by other methods, such as selective

hybridization or PCR. Similarly, different E. coli hosts other than those used in the above Examples can be also employed.

It will also occur to those skilled in the art that, as described in Examples 58 through 69, interplasmid insert exchange can be obtained in a sequel independent of the order of the overlapping cosmid clones in the genomic contig. Indeed, the schematic of Fig. 8 illustrates the sequel of interplasmid exchanges A-B, followed by B-C and then by C-D, while Examples 58 through 69 describe the sequel A-B, C-D and last B-C. Furthermore, technical variations on the methodologies employed here can produced equivalent results. All these variations fall within the scope of the present invention.

It will occur to those skilled in the art that the principles and methodologies described in Sections 7.2.1 and 7.2.2 are not mutually exclusive. For example, a construct equivalent to PAD6 can be directly isolated by subjecting the producer strain P. rosea to the principles and methodologies described in Section 7.2.1. Similarly, selected cosmids from the described S. coelicolor library (Redenbach et al., 1996, Mol. Microbiol. 21: 77-96) can be used for assembling a large chromosomal segment into pPAC- S1, following the principles and methodologies described in Section 7.2.2. Furthermore, it will occur to those skilled in the art that the principles and methodologies of Section 7.2.1 and 7.2.2 can complement each other. For example, after constructing an ESAC library of P. rosea DNA, inserts from individual ESACs may be enlarged by applying the principles and methodologies of Section 7.2.2, using, for example, cosmids overlapping the cognate ESACs.

Those skilled in the art understand that the principles and methodologies described in Section 7.2.2 and

illustrated in schematic form in Fig. 4 are general enough that they can be applied to other strains and clusters.

Methods for preparing high molecular weight DNA, for constructing and propagating in E. coli an ESAC library can be developed from the principles and methodologies described in Examples 12 through 17. Methods for preparing the appropriate combinations of fragment pairs to yield the starting plasmids described in Fig. 6, can be developed for other clusters following the principles and methodologies described in Examples 19 through 51; methods for assembling an entire cluster into a pESAC can be developed following the principles and methodologies described in Examples 52 through 69. In order to illustrate how the principles and methodologies described in Section 7.2 can be extended to other actinomycete strains producing different natural products, the constructions of ESACs carrying large gene clusters from different producer strains are reported herein. The Examples describe, for each cluster, the selection of the appropriate fragments A, B, C and D; and the construction of the starting plasmids, equivalent to those reported in Fig. 6. These plasmids can then be used to to reassemble each cluster according to the scheme of Fig. 8.

The rapamycin gene cluster from S. hygroscopicus is contained within three overlapping cosmids designated cos58, cos25 and cos2 (Schwecke et al., 1995, Proc. Natl.

Acad. Sci. USA 92: 7839-7843). The Examples described herein report the preparation of the appropriate fragments A, B, C and D; the construction of the plasmids containing the A-B, B-C and C-D cassettes; and the cloning approach to obtain constrcuts equivalent to those reported in Fig. 6.

Example 70

Preparation of the rapamycin fragments A, B, C and D Primers 5'-TTTTTGAATTCGGTACCAGCCGACGGCGA-3' (SEQID N°15) and 5'-TTTTTGGATCCCTGTTCCACCAGCGCACC-3' (SEQID 16) are used to amplify a 1.2 kb fragment from cos58 ; primers 5'- TTTTTTCTAGACCGTCGTCGGTGGTTCTC-3'(SEQID N° 17) and 5'- TTTTTGGATCCAGGAAGCCCTGTGCTGTC-3'(SEQID N°18)(SEQID N°18) a 1.2 kb fragment from cos58; primers 5'- TTTTTTGTAGAGGTCAAGATCCGGGGCAT-3' (SEQID N°19) and 5'- TTTTTCTGCAGGACAGCGCCCTTGAGGTG-3' (SEQID N°20) a 1.2 kb fragment from cos25; and primers 5'-TTT- TTCTGCAGGCGACGAAGAGGGGC-3' (SEQID N°21) and 5'- TTTTTAAGCTTAGCGCGACCGGGGCGGT-3'(SEQID N°22)(SEQID N°22) a 0.9 kb fragment from cos2. Fragment A, B, C and D are then digested with EcoRI + BamHI, BamHI + XbaI, XbaI + PstI, and PstI + HindIII, respectively.

Example 71 Construction of plasmid pUR1 Fragments A and B, prepared as described in Example 70, are ligated to pUC18 digested with EcoRI + XbaI. The resulting mixture contains the desired plasmid pUR1.

Example 72 Construction of E. coli K12 DH1/pUR1 Approximately 10 ng of plasmid pUR1, prepared as described in Example 71, are used to transform E. coli DH1 and a few of the resulting ApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUR1, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 2.7 and 2.4 kb after digestion of the plasmid with EcoRI + XbaI.

Example 73

Construction of plasmid pUR2 Fragments B and C, prepared as described in Example 70, are ligated to pUC18 digested with BamHI + PstI. The resulting mixture contains the desired plasmid pUR2.

Example 74 Construction of E. coli K12 DH1/pUR2 Approximately 10 ng of plasmid pUR2, prepared as described in Example 73, are used to transform E. coli DH1 and a few of the resulting ApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUR2, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 2.7 and 2.4 kb after digestion of the plasmid with BamHI + PstI.

Example 75 Construction of plasmid pUR3 Fragments C and D, prepared as described in Example 70, are ligated to pUC18 digested with XbaI + HindIII. The resulting mixture contains the desired plasmid pUR3.

Example 76 Construction of E. coli K12 DH1/pUR3 Approximately 10 ng of plasmid pUR3, prepared as described in Example 75, are used to transform E. coli DH1 and a few of the resulting ApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUR3, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 2.7 and 2.1 kb after digestion of the plasmid with EcoRI + HindIII.

Example 77 Construction of plasmid pUR11

Plasmid pUR1, prepared as described in Example 72 and previously digested with BamHI, and the tet fragment, prepared as described in Example 31, are treated with T4 DNA Polymerase and DNA ligase. The ligation mixture contains the desired plasmid pUR11.

Example 78 Construction of E. coli K12 DH1/pUR11 Approximately 10 ng of plasmid pUR11, prepared as described in Example 77, are used to transform E. coli DH1 and a few of the resulting TcRApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUR11, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 3.9 and 2.8 kb after digestion of the plasmid with HindIII.

Example 79 Construction of the plasmid pUR21 Plasmid pUR2, prepared as described in Example 74 and previously digested with XbaI, and the tet fragment, prepared as described in Example 31, are treated with T4 DNA Polymerase and DNA ligase. The ligation mixture contains the desired plasmid pUR21.

Example 80 Construction of E. coli K12 DH1/pUR21 Approximately 10 ng of plasmid pUR21, prepared as described in Example 79, are used to transform E. coli DH1 and a few of the resulting TcRApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUR21, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 3.9 and 2.8 kb after digestion of the plasmid with HindIII.

Example 81 Construction of the plasmid pUR31 Plasmid pUR3, prepared as described in Example 76 and digested with PstI, and the tet fragment, prepared as described in Example 31, are treated with T4 DNA Polymerase and DNA ligase. The ligation mixture contains the desired plasmid pUR31.

Example 82 Construction of E. coli K12 DH1/pUR31 Approximately 10 ng of plasmid pUR31, prepared as described in Example 81, are used to transform E. coli DH1 and a few of the resulting TcRApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUR31, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 3.9 and 2.5 kb after digestion of the plasmid with HindIII.

Example 83 Construction of plasmid pUR13 The 4.0 kb EcoRI-XbaI fragment obtained from plasmid pUR11, prepared as described in Example 78, and the 2.1 kb XbaI- HindIII fragment obtained from plasmid pUR3, prepared as described in Example 76, are ligated to pUC18 digested with EcoRI + HindIII. The ligation mixture contains the desired plasmid pUR13.

Example 84 Construction of E. coli K12 DH1/pUR13 Approximately 10 ng of plasmid pUR13, prepared as described in Example 83, are used to transform E. coli DH1 and a few of the resulting TcRApR colonies that appear on the LB-agar

plates are analyzed for their plasmid content. One colony is found to carry pUR13, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 4.9 and 3.9 kb after digestion of the plasmid with HindIII.

Those of ordinary skill in the art understand that the plasmids constructed above can be used for transfering the two-fragment cassettes present in pUR11, pUR21 and pUR31 into a ts vector. This can be achieved by recovering the 4.0 kb insert from pUR11, the 4.0 kb insert from pUR21, and the 3.7 kb insert from pUR31, after digestion with EcoRI + XbaI, EcoRI + PstI, and EcoRI + NdeI, respectively.

Similarly, those of ordinary skill in the art understand that the 6.1 kb four-fragment cassette present in plasmid pUR13 can be easily transfered into pPAC-S1 after digestion with EcoRI + NdeI. These subcloning experiments lead to the formation of plasmids equivalent to those reported in Fig.

6.

As another application of the principles and methodologies of the present invention, the Examples reported below describe the preparation of the appropriate fragments A, B, C and D from the Sac. erythraea erythromycin gene cluster. This cluster has been described and is contained within a series of overlapping clones (Tuan et al., 1990, Gene 90: 21-29; Donadio et al., 1993, In Industrial Microorganisms: Basics and Applied Genetics, Baltz, Hegeman and Skatrud eds., ASM, Washington, DC, pp. 257-265; Pereda et al., 1997, Gene 193: 65-71). The construction of the plasmids containing the A-B, B-C and C- D cassettes and the cloning approach to obtain constructs equivalent to those reported in Fig. 6 are also described.

Example 85

Preparation of the erythromycin fragment A Synthetic oligonucleotides 5'-CATGGGAATTCGGGGG-3' (SEQID N°23) and 5'-CCCCCGAATTCC-3' (SEQID N°24) are annealed and ligated to the 1.2 kb NcoI-BamHI fragment isolated from cosmid p3B2. The resulting mixture is digested with EcoRI + BamHI.

Example 86 Preparation of the erythromycin fragments B, C and D Primers 5'-TTTTTGGATCCGGGGCAGCGGTTGGTTCC-3' (SEQID N°25) and 5'-TTTTTTCTAGAAGGCAGCTCCAGATGATC-3' (SEQID N°26) are used to amplify a 1.0 kb fragment from cosmid p3B2; primers 5'-TTTTTCTAGACCGGACTCGGCCGGCTCG-3'(SEQID N°27)(SEQID N°27) and 5'- TTTTTCTGCAGCCGCACGCCTCGGTGGTC-3' (SEQID N°28) a 1.1 kb fragment from cosmid pSI; and primers 5'- TTTTTCTGCAGGGACCCTGAGTGCAGGTG-3' (SEQID N°29) and 5'- TTTTTAAGCTTCAGTAGCCGTCGCTGAGC-3'(SEQID N°30)(SEQID N°30) a 1.1 kb fragment from plasmid pEB6. Fragments B, C and D are then digested with BamHI + XbaI, XbaI + PstI, and PstI + HindIII, respectively.

Example 87 Construction of plasmid pUE1 Fragment A, prepared as described in Example 85, and fragment B, prepared as described in Example 86, are ligated to pUC18 digested with EcoRI + XbaI. The resulting mixture contains the desired plasmid pUE1.

Example 88 Construction of E. coli K12 DH1/pUE1 Approximately 10 ng of plasmid pUE1, prepared as described in Example 87, are used to transform E. coli DH1 and a few of the resulting ApR colonies that appear on the LB-agar

plates are analyzed for their plasmid content. One colony is found to carry pUR1, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 2.7 and 2.2 kb after digestion of the plasmid with EcoRI + XbaI.

Example 89 Construction of plasmid pUE2 Fragments B and C, prepared as described in Example 86, are ligated to pUC18 digested with BamHI + PstI. The resulting mixture contains the desired plasmid pUE2.

Example 90 Construction of E. coli K12 DH1/pUE2 Approximately 10 ng of plasmid pUE2, prepared as described in Example 89, are used to transform E. coli DH1 and a few of the resulting ApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUE2, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 2.7 and 2.1 kb after digestion of the plasmid with BamHI + PstI.

Example 91 Construction of plasmid pUE3 Fragments C and D, prepared as described in Example 86, are ligated to pUC18 digested with XbaI + HindIII. The resulting mixture contains the desired plasmid pUE3.

Example 92 Construction of E. coli K12 DH1/pUE3 Approximately 10 ng of plasmid pUE3, prepared as described in Example 91, are used to transform E. coli DH1 and a few of the resulting ApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony

is found to carry pUE3, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 2.7 and 2.2 kb after digestion of the plasmid with EcoRI + HindIII.

Example 93 Construction of plasmid pUE11 Plasmid pUE1, prepared as described in Example 88 and previously digested with BamHI, and the tet fragment, prepared as described in Example 31, are treated with T4 DNA Polymerase and DNA ligase. The ligation mixture contains the desired plasmid pUE11.

Example 94 Construction of E. coli K12 DH1/pUE11 Approximately 10 ng of plasmid pUE11, prepared as described in Example 93, are used to transform E. coli DH1 and a few of the resulting TcRApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUE11, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 3.9 and 2.6 kb after digestion of the plasmid with HindIII.

Example 95 Construction of the plasmid pUE21 Plasmid pUE2, prepared as described in Example 90 and previously digested with XbaI, and the tet fragment, prepared as described in Example 31, are treated with T4 DNA Polymerase and DNA ligase. The ligation mixture contains the desired plasmid pUE21.

Example 96 Construction of E. coli K12 DH1/pUE21 Approximately 10 ng of plasmid pUE21, prepared as described

in Example 95, are used to transform E. coli DH1 and a few of the resulting TcRApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUE21, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 3.7 and 2.7 kb after digestion of the plasmid with HindIII.

Example 97 Construction of the plasmid pUE31 Plasmid pUE3, prepared as described in Example 92 and digested with PstI, and the tet fragment, prepared as described in Example 31, are treated with T4 DNA Polymerase and DNA ligase. The ligation mixture contains the desired plasmid pUE31.

Example 98 Construction of E. coli K12 DH1/pUE31 Approximately 10 ng of plasmid pUE31, prepared as described in Example 97, are used to transform E. coli DH1 and a few of the resulting TcRApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUE31, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 3.8 and 2.7 kb after digestion of the plasmid with HindIII.

Example 99 Construction of plasmid pUE13 The 3.8 kb EcoRI-XbaI fragment obtained from plasmid pUE11, prepared as described in Example 94, and the 2.2 kb XbaI- HindIII fragment obtained from plasmid pUE3, prepared as described in Example 92, are ligated to pUC18 digested with EcoRI + HindIII. The ligation mixture contains the desired plasmid pUE13.

Example 100 Construction of E. coli K12 DH1/pUE13 Approximately 10 ng of plasmid pUE13, prepared as described in Example 99, are used to transform E. coli DH1 and a few of the resulting TcRApR colonies that appear on the LB-agar plates are analyzed for their plasmid content. One colony is found to carry pUE13, as verified by the observation, upon agarose gel-electrophoresis, of fragments of 4.8 and 3.9 kb after digestion of the plasmid with HindIII.

Those of ordinary skill in the art understand that the plasmids constructed above can be used for transfering the two-fragment cassettes present in pUE11, pUE21 and pUE31 into a ts vector. This can be achieved by recovering the 3.8 kb insert from pUE11, the 3.7 kb insert from pUE21, and the 3.8 kb insert from pUE31, after digestion with EcoRI + XbaI, EcoRI + PstI, and EcoRI + NdeI, respectively. Similarly, those of ordinary skill in the art understand that the 6.0 kb four-fragment cassette present in plasmid pUE13 can be easily transfered into pPAC-S1 after digestion with EcoRI + NdeI. These subcloning experiments lead to the formation of plasmids equivalent to those reported in Fig. 6.

The Examples reported above describe the principle and methodologies for assembling the erythromycin gene cluster into the pPAC-S1. It will occur to those skilled in the art that the principles and methodologies illustrated in Fig. 7 and described in Examples 52 through 69 can be applied to the erythromycin gene cluster, employing the pMAK705 derivatives constructed according to the principles described above and the eryhtomycin cosmids.

As a further example, the preparation of the

appropriate fragments A, B, C and D from the A. mediterranei rifamycin gene cluster is illustrated below.

This cluster has been described and is contained within a series of overlapping clones (August et al., 1998, Chem.

Biol. 5: 69-79).

Example 101 Preparation of the rifamycin fragments A, B, C and D Primers 5'-TTTTTGAATTCTGCAGACCGCCGAGGAAG-3' (SEQID N°31) and 5'-TTTTTGGATCCGGAGTCGTAGCTGACGAC-3' (SEQID N°32); 5'- TTTTGGATCCCGACCACGCGGGGACGTC-3' (SEQID N°33) and 5'- TTTTTTCTAGACCAGGGAACCCGTGCTGC-3'(SEQID N°34) (SEQID N°34) 5'- TTTTTTCTAGACGGAAGCTCGCCGCGATC-3' (SEQID N°35) and 5'- TTTTTCTGCAGGTCCGTAGCCCGGACACC-3'(SEQID N°36) (SEQID N°36) and 5'- TTTTTCTGCAGTTCGGGCGACAGTTCCTT-3' (SEQID N°37) and 5'- TTTTTAAGCTTCAACAAGCCATCCGGGTC-3' (SEQID N°38), are used to amplify fragments of 1.2,1.2,1.2 and 1.0 kb, respectively, from A. mediterranei genomic DNA. Fragments A, B, C and D are then digested with EcoRI + BamHI, BamHI + XbaI, XbaI + PstI, and PstI + HindIII, respectively.

Those of ordinary skills in the art understa nd that the fragments generated from the rifamycin gene cluster contain the same restriction sites as those generated from the rapamycin and erythromycin gene clusters, so that the same cloning strategies for generating the pUC18 derivatives containing the A-B, B-C and C-D cassettes, described above in Examples 72-77 for the rapamycin cluster and 88-93 for the erythromycin cluster, can also be applied to the rifamycin cluster. In addition, the rifamycin fragments A, B, C and D have been selected so that the same cloning methodologies described above for inserting tet within the A-B, B-C and C-D cassettes from the rapamycin

and erythromycin clusters, described in Examples 78-83 and 94-99, respectively, can be applied in this instance as well. Furthermore, the construction of the four-fragment cassette can also make use of the same cloning strategy.

Therefore, following the same principles and methodologies described in detail for the rapamycin and erythromycin clusters in Examples 72-85 and 88-101, respectively, plasmids equivalent to those reported in Fig. 6 can be constructed for assembling the rifamycin cluster into the pPAC-S1. It will occur to those skilled in the art that the principles and methodologies illustrated in Fig. 7 and described in Examples 52-69 for the P. rosea cluster can be applied to any gene cluster, once the appropriate pMAK705 derivatives have been constructed, employing available overlapping clones.

Thus, as schematized in Fig. 7, interplasmid insert exchange can be conducted between any plasmid containing the desired region and the cognate ts construct., Plasmids corresponding to pAB2, pBC2 and pCD2 can therefore be derived from any cluster. Similarly, the principles and methodologies illustrated in Fig. 8 can be applied employing the appropriate A-B-C-D cassette and the cognate pMAK705 derivatives prepared according to the scheme of Fig. 7. The principles and methodologies illustrated in Fig. 7 and Fig. 8 and described in Examples 52-69 can therefore be extended to other clusters.

It will occur to those skilled in the art that, although illustrated in Fig. 5 through 8 by three overlapping clones and described in the Examples 58 through 69 by the use of five rounds of interplasmid insert exchange, the principles and methodologies described in this section of the present invention can be extended to a different number of overlapping clones. If n is the number

of overlapping clones that encompass the desired genomic segment, n will also be the number of homologous recombination rounds that introduce cluster DNA into the pESAC. If an AbR marker is used to facilitate monitoring insert exchange, the total number of rounds of homologous recombination will be equal to 2n-1. Interplasmid homologous recombination has been described to introduce large DNA segments into a desired vector (O'Connor et al., 1989, Science 244: 1307-1312; Kao et al., 1994, Science 265: 509-512) or to target a smaller segment into a large episome (Yang et a., 1997, Nature Biotechnol. 15: 859-865).

However, it was not be anticipated that these procedures could be applied iteratively for the precise reconstruction of very large DNA segments.

7.3 Identification of positive clones The principles and methodologies described in Section 7.2 for obtaining an entire gene cluster in a pESAC rely on the identification of the desired genomic segment. When using the principles and methodologies described in Section 7.2.1, the desired clones are identified by screening an ESAC library with one of the possible strategies described below. When using the principles and methodologies described in Section 7.2.2, the desired clones are identified in a genomic library, such as a cosmid library, with one of the possible strategies described below, and then assembled into pESAC. The principles and methodologies for identifying the genes responsible for the biosynthesis of natural products are well described in the literature and are reported here solely to illustrate the fact that they represent a necessary step in the overall scope of the present invention.

The genes involved in the biosynthesis of natural

products in actinomycetes are invariably found as gene clusters in the chromosome of the producing organism, often associated with one or more resistance determinants.

Consequently, identifying one gene allows ready access to all the others. One or more genes responsible for the biosynthesis of a natural product could have been described, or the entire cluster could be known. Several biosynthesis clusters from actinomycetes have been reported and other clusters are likely to be described in the future. Suitable probes from the cluster extremities can be derived from published clusters, when available. Thus, fragments A and D, described in Example 70, can be used as probes to screen an ESAC library prepared from S. hygroscopicus DNA. ESACs positive to both probes will contain the rapamycin cluster. Similar strategies can be applied to ESAC libraries prepared from Sac. erythraea and A. mediterranei DNA, screened with fragments A and D, prepared as described in Examples 85-86 and 101, respectively.

If no biosynthesis genes are known, different strategies for identifying them can be applied. These strategies are well described in the literature and are summarized below. One possible strategy involves the isolation of the resistance gene (s) after cloning in a heterologous host that is sensitive to that natural product (for example, Stanzak et al., 1986, Bio/Technol. 4: 229- 232). Another possible strategy is based on reverse genetics: a particular biosynthetic enzyme is purified, and from its partial protein sequence (s) the corresponding gene is isolated via PCR or hybridization (for example, Fishman et al., 1987, Proc. Natl. Acad. Sci. USA 84: 8248-8252).

Another approach relies on the complementation of mutants blocked in one or more biosynthesis steps, after

introduction of a DNA library constructed in a suitable vector into the wild type strain (for example, Malpartida and Hopwood, 1984, Nature 309: 462-464). Another approach involves the construction of an expression library in a suitable vector in an appropriate host, where the gene product is sought after using specific antibodies or looking for a particular enzymatic activity (for example, Jones and Hopwood, 1984, J. Biol. Chem. 259: 14151-14157).

Another possible approach makes use of heterologous probes derived from biosynthesis, resistance or regulatory genes.

Natural products can be broadly grouped into classes according to their biosynthetic origin, and for many of them suitable probes are available. For example, genes encoding aromatic or modular polyketide synthases can be effectively identified through the use of heterologous hybridization probes (Malpartida et al., 1987, Nature 325: 818-821; Schwecke et al., 1995, Proc. Natl. Acad. Sci.

USA 92: 7839-7843); suitable probes have been reported for peptide synthetase genes (Turgay and Marahiel, 1994, Pept.

Res. 7: 238-241); for genes involved in the formation or attachment of modified sugars (Decker et al., 1996, FEMS Microbiol. Lett. 141: 195-201). As the understanding of the genetics of natural product biosynthesis increases, other heterologous probes will become available.

The size of clusters can be estimated from those of known clusters involved in the synthesis of structurally similar natural products. For examples, synthesis of macrolides is expected to require clusters in the 60-70 kb range (Katz and Donadio, 1993, Annu. Rev. Microbiol.

47: 875-912; Kuhstoss et al., 1996, Gene 183: 231-236); synthesis of glycopeptides, clusters in the 70 kb range (van Wageningen et al., 1998, Chem. Biol. 5: 155-162). In instances where no clusters have been described for the

same structural class of natural products, the size of the relevant cluster can be estimated from considerations about its known or likely biosynthesis route. Once the desired cluster has been identified, its extent can be established by analysis of the DNA sequence of the cloned cluster or of parts thereof. Comparison of the DNA sequence to databases can allow the identification of the likely borders of the gene cluster.

Employing the above mentioned approaches, the desired gene cluster can be identified in any library. If an ESAC library is used, the identified cluster is ready for transfer into the production host. If a smaller fragment library is employed, the cluster can be assembled into a pESAC.

Those skilled in the art understand that, when an ESAC library from a donor organism is constructed, any ESAC can be selected from said library and transferred into a production host. Therefore, a single donor organism can be utilized as the source of several biosynthesis clusters that can be mobilized into a production host. Similarly, an ESAC library needs not be constructed from a single donor organism.

7.4 Transformation of a Streptomyces host Once the desired gene cluster has been introduced into a pESAC, one or more ESACs are introduced into a suitable Streptomyces host. This is accomplished by employing published procedures for transformation of Streptomyces.

Only minor modifications from established procedures (Hopwood et al., 1985, Genetic Manipulation of Streptomyces: A Laboratory Manual, The John Innes Foundation, Norwich, UK) are required for obtaining a sufficient number of transformants. Because transformations

are performed with single, purified ESACs, transformation efficiencies do not need to be particularly high. The Examples reported below illustrate the principles and methodologies for introducing ESACs into S. lividans. They serve to describe the present invention and are not meant to restrict its scope. Streptomyces transformants are selected for ThR, specified by the tsr marker present in the pESAC. Since the incoming DNA is incapable of self- replication in Streptomyces, site-specific integration occurs at the chromosomal attB site, mediated by the int- attP function specified by the pESAC. That integration has occurred at the proper site can be verified by Southern hybridization or by PFGE analysis of the transformants.

Fig. 10 illustrates a PFGE separation of a S. lividans derivative carrying an ESAC with a 70 kb insert integrated into its chromosome.

Example 102 Introduction of ESACs into S. lividans ZX7 A few hundred ng of three individual ESACs, prepared as described in Example 17 and carrying inserts of S. coelicolor DNA of 70,120 and 140 kb (designated ESAC-70, ESAC-120, and ESAC-140, respectively), are used to transform protoplasts of S. lividans ZX7. The colonies that appear on the R2YE plates, after overlaying with Th, are analyzed for their ThR by streaking them on fresh R2YE plates.

Example 103 Cultivation and preservation of S. lividans ZX7/ESAC Individual colonies of S. lividans ZX7 transformants with the individual ESACs, prepared as described in Example 102,

are grown for several passages in solid medium without and with Th. Spore suspension, or mycelium prepared after cultivation in JM or YEME medium with Th, are stored at- 80°C after addition of glycerol to 20% (v/v).

Example 104 Characterization of S. lividans ZX7 attB:: ESAC-70 Individual colonies of S. lividans ZX7 attB:: ESAC-70, prepared as described in Example 102, are grown in YEME and total genomic DNA is prepared. The DNA is digested with BamHI, resolved by agarose gel-electrophoresis, and transferred onto a membrane. Hybridization to labeled pPAC- S1 DNA, prepared as described in Example 11, reveals the appearance of three bands of approximately 16,8 and 2.7 kb. PFGE analysis of genomic DNA reveals the disappearance of a 2.5 Mb DraI fragment present in ZX7 and the appearance of two fragments of 1.4 and 1.1 Mb (Fig. 10).

Although the present invention is described in the Examples listed above in terms of preferred embodiments, they are not to be regarded as limiting the scope of the invention. The above Examples serve to illustrate the principles and methodologies for introducing ESACs into S. lividans, for cultivating the resulting transformants and for confirming their genotype. The above Examples serve to illustrate the principles and methodologies for the transformation of S. lividans with ESACs carrying DNA inserts from a different species. It will occur to those skilled in the art that additional ESACs, either containing different inserts of S. coelicolor DNA, prepared as described in Example 17, or carrying DNA inserts from other actinomycetes can be used to transform S. lividans ZX7. As

another example transfer of large DNA segments, the transformation of S. lividans with a P. rosea gene cluster is illustrated below. Confirmation of the correct genotype of the resulting transformants is illustrated in Fig. 11.

Example 105 Construction S. lividans ZX7 attB:: PAD6 A few hundred ng of PAD6, prepared as described in Example 69, are used to transform protoplasts of S. lividans ZX7.

The colonies that appear on the R2YE plates, after overlaying with Th, are analyzed for their ThR by streaking them on fresh R2YE plates.

Example 106 Characterization of S. lividans ZX7 attB:: PAD6 Individual colonies of S. lividans ZX7 attB:: PAD6, prepared as described in Example 105, are grown in YEME medium and total DNA is prepared. The DNA is digested with BamHI, resolved by agarose gel-electrophoresis and transferred onto a membrane. Hybridization to labeled PAD6 il illustrated in Fig. 11. Bands of 22,10,7.6,7.2,6.2, 5.6,5.2,3.1,3.0,2.8,2.7,2.6,2.5,2.1,1.9,1.9,1.8, 1.6,1.5,1.4,1.2,1.0,1.0,0.9,0.9,0.9,0.7,0.6,0.5, 0.5,0.3 and 0.1 kb. The profile of P. rosea DNA is shown for comparison.

Those skilled in the art understand that S. lividans ZX7 attB:: PAD6 contains the expected number and size of bands expected from transfer of the cluster of Fig. 9 via PAD6. In analogy to the above Examples, the rapamycin, erythromycin and rifamycin clusters assembled in pESAC, according to the principles and methodologies described in Section 7.2.2, can be used to transform S. lividans. It

will occur to those skilled in the art that other S. lividans strains can be equally used as hosts for transformation with ESACs. Furthermore, OC31 can lysogenize other Streptomyces species, in addition to S. lividans.

These include but are not limited to the species reported in Table 2. Furthermore, a OC31 attB site may be engineered into Streptomyces species or other actinomycetes that are not naturally lysogenized by phage OC31. Therefore, any ESAC, prepared according to the principles and methodologies of Section 7.2, and any natural or engineered actinomycete host, fall within the scope of the present invention.

It will occur to those skilled in the art that alternative methods for introducing DNA into an actinomycete host can be employed. These include but are not limited to electroporation (MacNeil, 1989, FEMS Microbiol. Lett. 42: 239-244) and conjugation from E. coli (Mazodier et al., 1989, J. Bacteriol. 171: 3583-3585). It will also occur to those skilled in the art that alternative media and growth conditions can be employed for cultivating the transformants, and that they can be analyzed by different methods than those described above.

Technical variations on the methodologies described above can produced equivalent results. All these variations fall within the scope of the present invention.

Table 2 List of exemplary species of Streptomyces and other genera of Actinomycetales allowing attP-mediated integration of OC31 (Hopwood et al., 1985, Genetic Manipulation of Streptomyces: A Laboratory Manual, The John Innes Foundation, Norwich, UK; Lomovskaya et al., 1997, Microbiol. 143: 875-883; Kuhstoss et al., 1991, Gene 97: 143- 146; Soldatova et al., 1994, Antibiot. Khimioter. 39: 3-7).

Streptomyces coelicolor Streptomyces lividans Streptomyces hygroscopicus Streptomyces bambergiensis Streptomyces ambofaciens Streptomyces griseofuscus Streptomyces lipmanii Streptomyces thermotolerans Streptomyces clavuligerus Streptomyces fradiae Saccharopolyspora spinosa

7.5 Growth of the recombinant Streptomyces and metabolite production When an ESAC, introduced into a production host according to the principles and methodologies described in Section 7.4, carries the entire biosynthesis gene cluster derived from a donor organism, the recombinant strain produces the relevant natural product. Naive actinomycete hosts have been shown to produce the appropriate natural product or its intermediate (s) when the relevant DNA was introduced into them (Malpartida and Hopwood, 1984, Nature 309: 462- 464; Hong et al., 1997, J. Bacteriol. 179: 470-476; Kao et al., 1994, Science 265: 509-512). Thus, transformants of Streptomyces and other actinomycete species carrying the relevant biosynthesis clusters are expected to produce the corresponding natural product. The recombinant production hosts are cultivated in a suitable medium and the presence of the relevant metabolites is determined following appropriate procedures, which may include biological assays, chromatographic properties, MS, NMR, etc.

It will occur to those skilled in the art that ESACs, containing the relevant biosynthesis cluster derived from any donor actynomycete, can be used to transform S. lividans. The resulting transformants will produce the corresponding natural product. For example, an ESAC carrying the rapamycin, erythromycin or rifamycin cluster, prepared according to the principles of Section 7.2, can be used to transform S. lividans and rapamycin, erythromycin or rifamycin, respectively, can be produced by the resulting recombinant strain. Furthermore, it will occur to those skilled in the art that other Streptomyces or actinomycete strains that naturally contain or have been engineered to contain a phage OC31 attB site, can be used

as production hosts for desired natural products.

Therefore, any natural product produced after introduction of the relevant cluster carried on ESAC into a suitable production host, falls within the scope of the present invention.

The present invention describes principles and methodologies for optimizing and speeding up the process of lead optimization and development in drug discovery. These can be applied since the early stages of drug discovery as briefly summarized herein. A natural product produced by a donor organism has an interesting property, such as antibacterial, antifungal, antitumor, antihelmintic, herbicidal, immunosuppressive, or other pharmacological activity. The potential is seen for increasing the productivity of the producing organism, and/or for improving the biological or physico-chemical properties of said natural product after manipulating its structure. The biosynthetic pathway for the natural product is inferred from its chemical structure. This leads to a hypothesis on the genes involved, including the approximate size of the corresponding cluster. A large insert library is constructed in the pESAC vectors described herein using genomic DNA prepared from the donor organism. Through a judicious choice of hybridization probes and PCR primers, the desired cluster is identified in the library.

Alternatively, the cluster is assembled into the pESAC vectors described herein from overlapping cosmid clones identified by hybridization as above. The selected clone (s) are transferred into S. lividans, S. coelicolor or other suitable strain, and the resulting transformants are evaluated for production of the natural product.

Once production is obtained, the desired genetic, physiological and technological manipulations can be

performed on the production host, employing well-developed methodologies. The bioactive molecule is purified from a known host, amid a background of known metabolites. If necessary, ad hoc mutations can be conveniently introduced in the production host to eliminate unwanted, interfering products. Because of the deeper knowledge on the physiological processes and regulatory networks for secondary metabolism in the production host compared to the donor organism, targeted approaches to strain improvement, using classical and molecular techniques, can be applied.

Furthermore, well-characterized mutant strains are available for the production host, and the desired ESAC could be easily introduced into different genetic backgrounds. In addition, the biosynthetic pathway can be easily manipulated, because of the availability of the cloned genes and the efficient genetic tools for the production host. Finally, additional specialized genes or even entire clusters can be introduced into the production host, further expanding the possible applications of the present invention.

As it is apparent from the above description, a further object of this invention is to provide a process for the procduction of a natural product by cultivating an actinomycete strain capable of producing said natural product in the presence of a nutrient medium, isolating and purifying said natural product, characterized in that the actinomycete strain capable of producing said natural product (production host) is an actinomycete strain modified by means of an E. coli-Streptomyces Artificial Chromosome that carries a gene cluster governing the biosynthesis of said natural product derived from an actinomycete donor organism which is the original producer of said natural product, according to the method described

herein. Preferably, said modified actinomycete strain shall be a Streptomyces lividans or Streptomyces coelicolor strain.

Finally, even in a case where the natural product may not be made by the production host after transfer of the relevant cluster, appropriate tools are available to remedy that situation. Lack of production of the expected natural product might be due to several possibilities: absence of required gene (s); DNA, gene product or natural product instability; inadequate levels of gene expression or of appropriate precursors. In a well-defined production host, each of these possible causes may be directly investigated and remedied.

Therefore, the present invention provides significant advantages over the existing process of drug discovery and development, including production. It exploits the fact that the host where the natural product will be produced is an organism commonly used for process development and genetic manipulation, with substantial information available, including safety for handling it.