Cloning Streptomyces genes for antibiotic production

Cloning Streptomyces genes for antibiotic production

42 Trends in Biotechnology, Vol. 1, No. 2, 1983 BamHI and PstI, are particularly Cloning Streptomycesgenes for antibiotic production David A. Hopwo...

1MB Sizes 2 Downloads 52 Views

42

Trends in Biotechnology, Vol. 1, No. 2, 1983

BamHI and PstI, are particularly

Cloning Streptomycesgenes for antibiotic production David A. Hopwood, Mervyn J. Bibb, Celia J. Bruton, Keith F. Chater, Jerald S. Feitelson and Jos6 A. Gil The cloning and recombination o f the genes o f Streptomyces bacteria offer a method o f increasing antibiotic yields and generating new antibiotics. Novel vectors, both plasmids and phages, have been developed for use with Streptomyces. This article describes s o m e of these vectors and relevant cloning and screening techniques. The need to develop cheaper and more effective antibiotics and the spread of resistance to many commonly used antibiotics have ensured that Streptomyces bacteria continue to be investigated intensively. Streptomyces produce the great majority of the several thousand known antibiotic compounds 1, including many which are of great medical importance such as tetracyclines. Some species of Streptomyces may carry more than 100 expressed genes dedicated to antibiotic production - over 1% of the genome. A better understanding of the structure and regulation of genes involved in antibiotic production should assist in the development of more rational approaches to increasing antibiotic yields. Furthermore, gene cloning and the recombination of genes from different species may yield novel and valuable antibiotics. Techniques for the cloning of Streptomyces genes (reviewed in Refs 2-4) have been made possible by the availability of several Streptomyces plasmids and temperate bacteriophages with potential as DNA cloning vectors; and by the discovery that plasmid or phage DNA is taken up quite efficiently by Streptomyces protoplasts treated with polyethylene glycol (PEG) s-7. The principles of gene cloning are outlined in the diagram in the centre of this issue and additional terms are defined in the glossary on page 43. The first Streptomyces genes to be

cloned using these vectors were those responsible for resistance to the antibiotics methylenomycinS; neomycin and thiostreptong; and viomycin1°. The last three of these, and a cloned gene for tyrosinase which catalyses the conversion of tyrosine to black melanin pigment (E. Katz, C. J. Thompson and D. A. Hopwood, submitted), have been combined with plasmid or phage replicons to produce a series of cloning vectors that are now being used to isolate genes for antibiotic biosynthesis.

Three useful cloning vectors Three vectors, each with a different set of applications, can serve to illustrate the range of currently available cloning vectors for Streptomyces.

useful because they lie within the coding sequence of the aph gene: insertion of DNA at either of these sites therefore abolishes neomycin resistance. In shotgun cloning experiments to isolate new genes, plJ61 is often cleaved with BamHI, and donor DNA is cleaved with any member of a set of restriction enzymes - BamHI, Bali, BgllI, MboI, Sau3A or XhoII - which all generate 5' G A T C cohesive tails ('sticky ends') which can join with each other. (This is because their recognition site is either GATC, MboI) or inCTAG, cludes this sequence (e.g. GGATCC CCTAGG for BamHI). Usually the vector is treated with alkaline phosphatase to remove 5' phosphate groups and so render the ends of the vector molecules unable to ligate with each other to reconstitute the vector, the vector and donor DNA are mixed, treated with DNA ligase, and added to S. lividans protoplasts in the presence of PEG ~°. The protoplasts are plated on regeneration medium and, after about 18 h incubation (to allow expression of the tsr' gene), the plates are overlaid with agar containing thiostrepton to select transformed cells (trausformants) which have taken up the plasmids containing thiostrepton-resistance genes. These

Plasmid plJS1. pIJ61 is derived from SLP1.2, an autonomous conjugative plasmid, with a copy number of 4-5, found in Streptomyces lividans 66 after mating with Streptomyces coelicolor A3(2). SLP1.2 comes from a chromosomal DNA sequence of S. coelicolor which 'loops out' from the chromosome and is transferred by conjugation to S. lividans ~'. plJ61 (Re£ 12), which has similar features to pBR322 in Eschen'chia coh~3, was constructed from SLP1.2 by the addition of a neomycin-resistance gene (aph, for aminoglycoside phosphotransferase) from Streptomyces fradiae ATCC 10745 and a gene (tsr) for thiostrepton resistance from Streptomyces azureus ATCC 14921, and deletion of two segments of nonessential SLP1.2 DNA. The resulting plasmid (Fig. 1) has several sites suitable for DNA insertion but two, for

The authors are all at the John Innes Institute, Norwich NR4 7UH, UK. Jos~ A. Gll's present address is Departmento de Microbiologia Facultad de Biologia, Universidad de Le6n, Le6n, Spain. ©1983,ElaewecSoeacePubhshersB.V.,Amtterdam0166 9430/83/$0100

I

'Sstl

Fig. 1. Simplified restrictionmap ofpIJ61 (modified from Re£ 12) showing a selection of available clot~ng ~tes. Cl~vage by Sstl or KgnI remov~ a fragment of non-

essential DNA, whereas all the other sites are unique and allow simple insertion of cloned DNA. The BamHI and PstI sites are within the neomycin-resistance gene; clones containing foreign DNA inserted at these sites can be recognized by their susceptibility to neomycin (insertional inactivation), aph = aminoglycoside phosphotransferase gem (for neomycin resistance); tsr = thiostrepton-resistance gene.

Trends in Biotechnology, tlol. 1, No. 2, 1983

cells develop into colonies over the next few days. Routinely, one obtains about 1(P- 10s transformants per/~g of vector D N A in such a cloning experiment (compared with I0~- 107per/~gwith un-

43

cleaved vector DNA). The transformant colonies can be replicated to neomycin-containing plata to verify that a high proportion (in practice 20-90%) are neomycin-sensitive and

thereforeshould carryinsertsof foreign D N A . With cloned fragments in the size range 5-15 kilobascpairs(kb) any desired clone is normally found among only a few thousand transformants;this

Glossary auxotroph - a mutant microorganism that requires a specific organic nutrient for growth; auxotrophs are usually denvecI from a prototroph (a strain that grows on a chemically defined medium without complex organic nutrients) by mutation in a gene for one of the enzymes of the nutrient's biosynthetic pathway, thereby blocking the pathway. blocked mutant - a mutant microorganism that has lost the ability to make an antibiotic (analogous to an auxotroph, except that growth of the mutant is not impaired).

from the host DNA, replicates, and produces progeny viruses which burst (lyse) the host cell. In order to enter the prophage state, the phage DNA must have a repressor gene (c +) to switch off most phage genes, and a short segment of DNA with the same sequence of basesas a homologous segment in the host chromosome. The two homologous sequences can then pair with each other, break at corresponding points, and rejoin In a new configuration (i.e. form a crossover), thus:

("

conjugative plasmid - some bacteria wdl mate with each other (conjugate) and exchange DNA. Conjugation is controlled by genes carnecl on certain plasmids, which are said to be conjugative.

cosmid - a plasmid carrying an artifioally added segment of DNA (the cos sequence) which is recognlzed by the head proteins when phage lambda DNA is packaged to form a complete virus. Cosmlds, carrying foreign DNA, can therefore be packaged outside E. cob and introduced into it. Cosmids are good vectors for making a gene library because they allow large segments of foreign DNA (up to 40 000 base pairs) to be doned ettloently, thereby reducing the size of the hbrary.

cosynthesis - production of antibiotic by a pair of blocked mutants cultured together. A mutant blocked late in the biosynthetic pathway accumulates a precursor of the antibiotic which can be converted by a mutant blocked at an eaHier point in the pathway into the complete antibiotic. Cosynthesis tests are used to classify a set of blocked mutants. lysogen - when a temperate bacteriophage (such as lambda phage of E.

col0 infects its host bacterium, the DNA of the phage may become integrated into the host's chromosome and the vegetative functions of the phage are repressed; the phage DNA (the prophage) is then passed on to subsequent generations of bacteria. These bacteria are said to be tysogens because occasionally the benign relationship between prophage and host is upset; the prophage detaches itself

DNA

~ . ~

circular DNA

1 rl////lllllllllJ Normally, the homologous sequences are specific attachment sites carried on the host chromosome and on the phage DNA; in special phage cloning vectors, the attachment site is deleted and instead a piece of cloned host chromosomal DNA provides a homologous region for crossing-over to occur. plaque - a circular 'hole' produced by phage infection in a confluent culture of bacteria growing on an agar medium. A virus reproduces in its host to produce progeny viruses which attack other nearby bacteria, and the process is repeated until all the bacteria in a small area (the plaque) are destroyed. Plaque formation is used to count.viruses, using a sensitive, or indicator, strain of bacteria.

moter and with a cloning site just on its 5' side. Fragments of foreign DNA able to act as promoters are recognized when, after insertion at the cloning site, they lead to expression of the marker gene; expression is said to occur by readthrough from the foreign promoter.

protopiasts - cells of bactena, fungi or plants from wNch the rigid walls have been enzymaticalty removed to expose the cytoplasmic membrane. Under the influence of polyethylene glycol, protoplasts will fuse together or will take up DNA. On suitable media, protoplasts will regeneratethe cell wall to produce a new culture. shotgun cloning - the cloning of unselected (random) segments of the DNA of a donor cell to form a library of bacterial colonies among which the desired clone has to be sought. Southern transfer - a technique, named after its inventor E. M. Southern, in wNch fragments of DNA of specific sequence are recognized in a complex mixture, such as restriction fragments of the entire genome of an organism. The mixture of DNA fs fractionatecl by gel electrophoresis, transferred to a nitrocellulose filter and separated into single strands. A sample of probe DNA, complementary to part of the desired fragment, is radio. actively labelled (usually with ~2p) and allowed to find the corresponding sequences amongst the DNA on the filter and to hybridize with them. Autoradiography of the filter on X-ray film then reveals the presence and position of such DNA. sub-cloning - the further cloning of a piece of foreign DNA from an original done, onto the same or a different vector. Subcloning may be done to remove unwanted DNA segments from a clone, to clone a specific part of a gene such as its promoter, etc.

promoter - the sequences of DNA that

transformation (or transfection) - the

lie just before the beginning of a gene 0.e. on its 5' side) to which the transcnbing enzyme (RNA polymerase) binds and initiates transcription of the messenger RNA.

process whereby microorganisms (or their protoplasts) take up DNA from outside (the term transfection Is used when the DNA is that of a virus).

promoter-probe vector - a vector

Seealsocentrefold diagram for outhneof the pnnoples of gene cloning.

carrying a marker gene lacking a pro-

44

Sphl4/

Trendsin Biotechnology, VoL 1, No. 2, 1983

nl ~ ?

\

Bg[II-~a|r'~7kh" ) Ss~l~ Fig. 2. Simplified restriction map ofpU702 (modified from Katz et a/. submitted) showing a selection of available cloning sites. The SphI, BglII and SstI sites are within the tyrosinase gene, allowing clone recognition by imertional inactivation (see Fig. 3). tsr = thiostrepton resistance gene; me/ = tyrosinase (melanin production) gene.

would be expected from the known genome size (c. 104 kb) of streptomycetes. Plasmid plJ702 pIJ702 (Katz et aL submitted) differs from plJ61 in having a high copy number (40-300 per chromosome) and a broad host-range within the genus Streptomyces, since it is derived from the broad host-range, multicopy plasmid plJ101 (Ref. 14). pIJ702 (Fig. 2) lacks the conjugative functions of plJ101 as a result of in vivo and in vitro deletions. It carries the tsr gene for vector selection and the mel gene, coding for tyrosinase, from Streptomyces antibioticus I M R U 3720 (Katz et al. submitted). The cloning sites normally used - those for SphI, SstI or BglII - are all within the mel gene (if the BGIII site is chosen the donor DNA can be cut with any of the set of enzymes used with the BamHI site of pIJ61). The function of the melgene is destroyed when DNA is inserted into it at any of the restriction sites and thus transformed colonies can easily be identified by their colour. Transformants harbouring the unaltered vector plasmid (Fig. 3) form black colonies when grown on tyrosine-contalning regeneration plates because they can convert tyrosine into melanin; recombinant plasmids produce colonies of white cells.

+C31. In KC400 some non-essential ~?C31 DNA has been removed and replaced by the viomycin phosphotransferase (vph) gene (which gives viomycin resistance) from Streptomyces vinaceus NCIB 8852 (Ref. 10) and the whole of the E. coli plasmid pBR322. The vector can therefore be propagated either as a phage in Streptomyces or as a plasmid (plJS05) in E. coll. The phage lacks its attachment site and so cannot lysogenize Streptomyces hosts (Fig. 4). For use as a vector, KC400 (or plJ505) DNA is treated with PstI to remove a segment including most of pBR322 bounded by two PstI sites (digestion also with PvulI cleaves the fragment, giving blunt-ended pieces, essentially preventing its re-insertion into the vector: Fig. 4); PstI fragments of donor DNA are then ligated with the vector DNA, and the mixture is used to transfect protoplasts of S. lividans, a convenient general host 16. The resulting plaques are replica plated to plates coated with spores of an indicator strain (usually the strain from which the cloned DNA was derived). Phages contalning inserted DNA can lysogenize the indicator strain, rendering it resistant to viomycin, provided the inserted fragments are homologous with sequences in the recipient DNA, allowing the cloned DNA to cross over with host DNA. I f the cloned DNA is

an internal segment of a gene which lacks both a promoter sequence at the beginning and coding sequences at its 3' end which are essential for the production of a functional gene product, then the function of that gene is abolished by the single crossover event that integrates the phage DNA into the recipient's genome to form the lysogen (Fig. 5). Insertion of such a fragment within an operon would almost certainly produce effects on genes downstream from the cloned fragment. In either case a mutant phenotype results. This provides a useful general strategy for cloning antibiotic genes (see below), since ~C31 can infect many antibioticproducing species.

Three procedures for cloning antibiotic biosynthetic genes Genes for antibiotic resistance, or those for easily detected single gene products like tyrosinase, are comparatively easy to isolate since the desired clones can readily be detected in a foreign host such as S. lividam, either by direct selection (for antibiotic resistance) or by a simple screening procedure. Antibiotic biosynthetic genes are in principle more difficult to isolate, not only because there is not usually a direct selection procedure, but also because a typical antibiotic is made by step-wise synthesis involving a path-

,,,

Phage +C31 KC400 +C31 KC400 (Ref. 15) is one of a series of cloning vectors derived from the temperate Streptomyces phage

,

,

Fig. 3. Colonies ofStreptomyces lizidans carrying pIJ702. Black colonies (which are able to produce melanin from tTrosine) carry m o d i f i e d pIJT02; white colonies (melanin nonprodudng)carrypIJ702withforeignDNAiraertedintheBglIIsiteofthemdgene(Katzet a/. submitted).

Trends in Biotechnology, VoL 1, No. 2, 1983

45

"-'-' PvulI BamHl

J,I



$C31 KC400

I

I PstI

40.'Ikb

I ~tI

Fig. 4. Structure of phage vector KC400 derived from/pC31, c+ = repressor gene. The black box represents pBR322 DNA, and the white box S. vinaceus DNA including the viomycin-resistance gene (labelled vph). The ends of the phage genome carry complementary singlestranded (cos)sequences which, when ligated, give rise to a circular plasmid (pIJS05) capable of replicating in E. coil Cleavageby PstI and PvuII removes, and effectivelydestroys, the indicated segment of DNA, which can then be replaced by PstI fragments (2-6 kb) of DNA to be cloned. way of perhaps 10-30 enzyme- sample of genes coding for secondary catalysed steps. Therefore, one would metabolic functions which may reveal not expect to isolate the genes by general features of the regulation of cloning them into S. lividans, or any gene expression for antibiotic producother 'standard' host, and detecting tion; and (2) serve as model studies to antibiotic production, since this would show how cloning can be applied to the require the simultaneous cloning and isolation of genes for the biosynthesis of expression of a complete set of bio- industrially important antibiotics. synthetic genes. We have approached the problem of A step in candicidin biosynthesis cloning some of the genes of antibiotic Candicidin, a polyene macrolide antibiosynthesis in three different ways, biotic (Fig. 6), is made in Streptomyces each using a different one of the three griseus I M R U 3570 by polyketide synvectors described above. We hope that thesis on an aromatic p-aminoacetothese approaches will: (1).provide a phenone starter unit synthesized from

resfricfion fragmenfs being ctoned

(a)

I

1

promoter

ferminafor

A BIC I (b)

- -

truncated mRNA (c)

.

.

.

,

,

- -

comptefe mRNA .

. no mRNA

truncated mRNA (d)

1 ,

'J

no mRNA

comptete mRNA Fig. 5. The use ofJ?C31KC400 for mutational cloning23.(a) Shows a wild-type operon, with three restriction fragments (A, B and C), which can be cloned into +C31 KC400; the recombinant clones are selected as viomycin-resistant transductants which arise through single crossovers between the cloned fragment and the homologous DNA of the host. If, as in segments A and C, the cloned segment contains either the promoter or the last essential coding sequences (taken here, for simplicity, to be the transcription terminus), the resulting lysogen retains the wild phenotype (diagrams b and d). If, as in segment B, the cloned segment contains neither, the resulting lysogen will have a mutant phenotype (c).

chorismic acid by the enzyme p-aminobenzoic acid synthetase (PABAsynthetase). This enzyme may play an important role in the regulation ofcandicidin synthesis, since its activity rises from being barely detectable during vegetative growth to high levels at the onset of candicidin production. Moreover, there is a close correlation between the activity of PABA-synthetase and candicidin titre in wild-type S. gr/seus under various levels of phosphate repression, and in mutants with different degrees ofimpairment of candicidin productionIL Repression of antibiotic productivity by inorganic phosphate is a common phenomenon in antibiotic fermentations. Here, then, was an opportunity to clone a gene involved in antibiotic biosynthesis without having to take into account the whole metabolic pathway. The gene was cloned (J. A. Gil and D. A. Hopwood, submitted) in S. lividans, using the B a m H I site ofpIJ41 (a derivative of SLP1.2, identical with pIJ61 except that it has an extra 100 bp which contains a second PstI site), in two ways: (1) by selecting a clone carrying a segment of S. griseus DNA which could restore prototrophy to a PABA-requiring (pab) auxotroph oYS. lividans; and (2) by selecting a sulphonamide-resistant S. lividans clone carrying a segment of DNA from a sulphonamide-resistant mutant of S. griseus (which presumably owes its resistance to over-production of PABA, thereby overcoming competitive inhibition by sulphonamide). Each experiment yielded a single clone (out of a few thousand transformants), and both clones carried a 4.5 kb B a m H I fragment ofS. griseus DNA (apparently the same segment in each clone). We are now investigating the effects on the regulation ofcandicidin biosynthesis of returning the cloned DNA to the S. griseus donor at high copy number,

46

Trends in Biotechnology, Vol. 1, No. 2, 1983

H2N~ I

II

c.=c.=c,,= o

o.

o

o,,.° ~

.o

x 0

' 0

~ ~ OH OH OH 0

)

OH

ko.

o.

CO0H

Fig. 6. Structure ofcandiddin. after cloning into pIJ702. The effects of mntagenesis in vitro of the pab gene on candicidin biosynthesis are also being studied.

An O-methyl transferase involved in undecylprodigiosin biosynthesis Provided one has a set of blocked mutants for the biosynthesis of an antibiotic of interest and a reliable screening procedure to detect antibiotic production, a general strategy for cloning the biosynthetic genes is to seek antibiotic-producing colonies amongst transformants of a mutant after introducing segments of DNA from the wild-type, antibiotic-producing strain in the expectation that the correct fragment of wild-type DNA will make good or 'complement' the mutation in the blocked mutant. In this way individual genes may be isolated. This approach has been used for the red pigmented antibiotic undecylprodigiosin (Fig. 7) of S. coelicolor A3(2) (Ref. 18). A series of mutants defective in antibiotic biosynthesis (red mutants) had been previously isolated and grouped into five classes (redd-~3 by cosynthesis tests; representatives of each class had been mapped to a cluster of closely-linked sites.on the S. coelicolor chromosomal9. In the first cloning experiment, the BglII site of pIJ702 was used, with S. coelicolor wild-type DNA cut by Bali, and the recipient was a red mutant of the E class (redE60). A single red colony (producing the antibiotic) was found out of 740 transformants. The clone (plJ750) carried a segment of 4.7 kb ofS. codicolor DNA. It also restored the antibiotic production to another red//mutant (redE8) but not to red mutants of classes A-D; thus it probably carried only the redE gene. The cloned DNA was used as a 32p. labelled DNA hybridization probe against a library of 800 E. coli colonies carrying cosmid clones2°prepared from S. codicolor DNA; several hybridizing colonies were revealed. From one of

these, carrying a 25 kb EcoRI segment ofS. coelicolor DNA, a 9.7 kb SstI fragment was sub-cloned into pIJ702 to give a clone (pIJ752) which restored antibiotic production m redB as well as red//mutants. More recently, a further primary clone was isolated, carrying an 8.9 kb PstI fragment of S. coelicolor DNA, which complemented both red// and red//mutants (F. Malpartida and J. S. Feitelson, personal communication). Thus at least the red/l, B and E genes have been cloned and confirmed to be very close together, possibly contiguous. These data agree with the earlier results obtained from genetic mapping in vivo ~9. Further analysis of cosmid and lambda phage gene librariesin E. coli, using the presently availablered clones as DNA hybridiT~tion probes should lead to the isolationof the entiregroup of biosyntheticgenes for thisantibiotic, together with their control sequences, so that their organization and expression can be studied in vivo and in vitro. Meanwhile, it has been established that the redE mutations block a late step in the synthesis ofundecylprodigiosin, an O-methyltransferase reaction. This was demonstrated in two ways. One approach involved cosynthesis tests between the S. coelicolor red mutants and a group of mutants of the Gram-negative Serratia marcescens blocked in the synthesis of the closely similar antibiotic, prodigiosin2k One of the S. marcescens mutants (OF) was known to be blocked in the correspondhag O-methylation ofnorprodigiosin to prodigiosin; the OF mutant and the redE80 mutant of S. coelicolor showed nearly identical cosynthesis reactions with other mutants of both organisms. Another line of evidence involved assays in vitro for methylation of undecylnorprodigiosin to undecylprodigiosin. Cell-free extracts of redEO0 lacked significant O-methyhransferase activity, but the clones carrying pIJ750 and pIJ752 in the redEO0 mutant did

possess the activity. Interestingly, the original clone (pIJ750) had low concentrations of O-methyltransferase, somewhat below the wild-type activity, while pIJ752 had an e~hanced activity (about five times that oftbe wild-type). These findings suggest that pIJ750 carries a segment of cloned DNA lackhag its own promoter (or possibly even that it has an incomplete redE structural gene) so that antibiotic production by the clone occurs only after crossing-over between the cloned DNA and the corresponding mutant chromosomal gene, to give a recombinant functional gene in the chromosome (at a single copy), pIJ752, on the other hand, would carry an intact gene and probably its own promoter, so that expression could occur from the multiple copies ofthe cloned DNA itself giving a higher level of enzyme activity.

Genes for methylenomycin biosynthesis Methylenomycin A (Fig. 8) was the first antibiotic whose structural biosynthetic genes were shown to be naturally carried on a plasmid; they are on the SCP1 plasmid of S. coelicolor A3(2) (Re£ 22). Unfortunately, SCP1 is hard to isolate and it has so far not been feasible to clone the methylenomycin genes directly by starting from purified plasmid DNA which should have made the clones much easier to fred. Instead, some of the methylenomycin genes have been cloned from total DNA of Streptomyces parvulus ATCC 12434 carrying SCP1, which had been transferred to this strain by conjugation (K. F. Chater and C. J. Bruton, submitted). The donor DNA was cleaved with PstI and ligated into the ~C31 phage vector KC400. Protoplasts orS. lividans were transfected to give phage plaques, which were replica plated to lawns of an S. lividans strain carrying SCP1; after incubation until sporulation, the S. lividans cultures were replicated to viomycin to detect lysogens. These, by the argument outlined above, would carry inserted D N A segments homo-

~

CH3

(CH2)IoCH3

Fig. Z Structure ofundecylprodigiosin (the methyl group in largetype is added by an Omethyltramferase to the precursor, undecylnorprodigiosin)ms.

47

Trendsin Biotechnology, Vol.1, No. 2, 1883 Iogous with sequences in the recipient. The chromosomes ofS. pamulus and S. lividam are not homologous, so a high proportion of the viomycin-resistant colonies carried prophages with inserted PstI fragments ofSCP1 DNA, since SCP1 was the only DNA sequence common to both donor and recipient (the experiment was aided by the fact that the SCPI DNA was unexpectedly amplified to multiple copies in S. parvulus, but even then not in an isolatable form). When 278 of the viomycin-resistant transductants were tested for methylenomycin production, nine were nonproducers; they represented at least two classes of mutations because five of them co-synthesized methylenomycin with one of the non-producing mutants originally isolated by Kirby and Hopwood2L In agreement with this, three of the five contained phages with a cloned 2.85 kb PstI fragment in common, and phages from the other two had a different fragment (2.1 kb) in common: all four non-cosynthesizers yielded phages with a 2.3 kb fragment in common. Thus at least 7 kb of DNA needed for methylenomycin biosynthesis has been cloned. This strategy for isolating clones carrying (parts of) antibiotic biosynthetic (or other) genes has been called 'mutational cloning m and should prove useful in other situations. Even species for which transformation and transfection systems have not been developed and for which there are no mutants which do not produce antibiotics can be subjected to this approach, since the initial cloning of DNA is carried out in S. lividam and the DNA is subsequently introduced into the antibiotic-producing host by natural phage infection. The resulting clones, carrying only part of the gene or operon of interest, can then be used as very specific radioactive probes both to isolate complete genes from clone libraries (as described above for the undecylprodigiosin genes) and to identify mRNAs. Such probes can also be used for Southern transfer analysis, which has shown that the three classes of cloned methylenomycin production DNA are not only located very close to each other in SCP1 DNA, but also to the methylenomycin resistance gene (mmr, the first Streptomyces gene to be cloned8) (K. F. Chater and C. J. Bruton, unpublished).

What next? Our main interest in the clones now being isolated is to use them to understand the organizational and regulatory features of the genes for secondary metabolites: do controls occur at the leveloftranscription or translation, and how do they operate?

0

H3C~'~CH2 oLI H3C~ ~C02H Fig. 8. Structure ofmethylenomycin A. The DNA sequences characteristic of Streptomyces promoters are not yet identified, even though a number of segments ofStreptomyces DNA having promoter activity have been isolated by the use of promoter-probe vectors derived from SLP1.2 (Ref. 24). Most Streptomyces promoters appear to be different from those of E. coli. Thus, the fragments of Streptomyces DNA containing promoters isolated by Bibb and Cohen ~4 did not function as promoters in E. coli, whereas several E. coli promoters (and promoters from Serratia and Bacillus) did lead to transcription in S. lividans. Moreover, several Streptomyces genes are expressed in E. coli, but (at a significant level) only when orientated in such a way that transcriptional readthrough can occur from a known E. coli promoter: they include endoglycosidase H (Ref. 26); vph (Ref. 14); aph (Ref. 27); and pab (J. A. Gil and D. A. Hopwood, submitted). There is also evidence that, like Bacillus subtilis (Ref. 28), S. coelicolor possesses several different forms of RNA polymerase that can transcribe different classes of promoters (vegetative or spo.rglation-specific B. subtilis promoters have been used in in vitro assays for these RNA polymerases, J. Westpheling and R. Losick, personal communication). By using a combination of promoter-probe vectors and mutational cloning in its analytical mode to localize promoters, transcription assays in vivo and in vitro, and DNA sequencing, along with sitespecific mutagenesis and deletion analysis, it should soon be possible to discern some of the features ofclifferential gene expression in Streptomyces: what determines the onset of antibiotic synthesis at a certain stage in the

development of a colony or culture, and how the normal controls on expression can be modulated or overcome.

Industrial applications In the past, the yield of antibioticproducing cultures has been increased mainly by empirical methods. On the one hand, the culture medium has been optimized, sometimes by the addition of specific antibiotic precursors, or nutrients which tend to favour antibiotic biosynthesis (for example, oils which yield acetyl coenzyme-A for polyketide synthesis); and fermentation parameters such as pH or methods of aeration have been tuned to the needs of the process. On the other hand, the organism itself has been changed, by successive rounds of random mutation and screening, to increase its capacity to synthesize a particular antibiotic. By such strategies the titre of antibiotic in some streptomycete fermentations (and in the production of penicillin by the fungus Penicillium chrysogenum) has been increased a thousand-fold over a period of 30-40 years. The study of cloned genes involved in antibiotic production promises, for the first time, to make the task of yield improvement more rational by leading to an understanding oftbe regulation of these genes. Does the typically late onset of antibiotic biosynthesis always reflect late transcription of the genes? If so, what controls it, and could transcription be brought forward by substituting 'vegetative' for 'late' promoters or by some other trick of genetic engineering? Are the mRNA transcripts of antibiotic biosynthetic genes capable (unlike those of most bacterial genes) of a stable existence in the cell and, if not, could mRNA stability be increased? What is the basis of carbon and nitrogen catabolite repression of antibiotic biosynthesis, and of phosphate repression, and how can these brakes on productivity be removed? These and many other questions will doubtless be addressed and answered as more and more genes controlling antibiotic biosynthesis are cloned. Although several thousand antibiotics have been discovered new ones are still required; to overcome natural or acquired resistance in human bacterial pathogens; to attack fungi, viruses and cancers; to eliminate infestations by insects, mites and parasitic worms; to promote growth in farm animals; to control plant diseases; and for applica-

48 tions as yet unthought of. New antibiotics are still being found by more or less traditional means - the isolation and screening of microorganisms, mainly actinomycetes, from a wide variety ofnatural habitats. However, as more antibiotics are isolated, it becomes harder to discover something new. The use of gene cloning techniques promises to lead to greater novelty. The knowledge on which the rational production of new antibiotics will be based is not yet available - since it can only come from the cloning and study of the genes which, as we have seen, is just beginning. While we wait for this knowledge, and the cloned genes, an empirical approach can already be applied. The idea is to make clone libraries of the DNA of particular streptomycetes (the donors) into other streptomycetes (the recipients) and screen for interesting new activities. These may be expected for two reasons~9. The first, and more obvious, is that new combinations of antibiotic biosynthetic genes will be created, which do not occur in nature; sometimes an enzyme from a donor pathway may act on a substrate or intermediate in a recipient pathway to create novel compounds. This will require a certain, but by no means totally unexpected, lack of specificity on the part of the biosynthetic enzymes. (It will also, of course, depend on transcription and translation of the genes in the new host, but this is unlikely to be a problem- we have not encountered interspecific barriers in our work and indeed, as we saw above, streptomycetes seem to be rather indiscriminating in their expression of DNA from other bacteria.) The second, and perhaps on the face of it more speculative, reason is that cloning of antibiotic biosynthetic genes into a new cellular environment will sometimes lead to their derepression to a level capable of yielding a detectable product when none was seen before.

Trends in Biotechnology, Vol. 1, No. 2, 1983

Several examples (cited in Ref. 29), in which natural mating was used to generate recombinants (cloning techniques were not available) which produced hitherto unsuspected antibiotics, give grounds for believing in this approach. In conclusion, we believe that gene cloning in relation to antibiotic biosynthesis is about to enter an exciting stage. The vectors and detection procedures for clones are already adequate for the task in some strains, and the methods will certainly be refined and generalized as the work progresses. Unexpected and useful results must surely follow.

J. M. and Hopwood, D. A. (1982) Gene 20, 51-62 13 Bolivar, F., Rodriguez, R. L., Greene, P. J., Betlach, M. V., Heynecker, H. L., Boyer, H. W., Crosa, J. H. and Falkow, S. (1977) Gene 2, 95-113 14 Kieser, T., Hopwood, D. A., Wright, H. M. and Thompson, C.J. (1982)Mol. Gen. Genet. 185, 223-238 15 Harris, J. E., Chater, K. F., Bruton, C. J. and Piret, J. M. (1983) Gene (in press) 16 Rodicio, M. R. and Chater, K. F. (1982)ff. Bacteriol. 151, 1078-1085 17 Gil, J. A., Naharro, G., Villaneuva, J. R. and Martin, J. F. (1980) in Advances in Biotechnology (Vezina, C. and Singh, K., eds), Vol. III, pp. 141-146, Pergamon Press 18 Feitelson, J. S. and Hopwood, D. A. (1983) Mol. Gen. Genet. (in press) 19 Rudd, B. A. M. and Hopwood, D. A. References (1980)J. Gen. MicrobioL 119, 333-340 1 B&dy, J. (1980) Process Biochem. 20 Groffeld, F. G., Dalai, H-H. M., Oct/Nov. 28-35 deBoer, E. and FlaveU, R. A. (1981) 2 Chater, K. F., Hopwood, D. A., Kieser, Gem 13, 227-237 T. and Thompson, C. J. (1982) Curt. 21 Williams, R. P. and Qadri, S. M. H. in Top. MicrobioL lmmunoL 96, 69-95 The Genus Serratia (von Graevenitz, A. 3 Hopwood, D. A. and Chater, K. F. and Rubin, S. J., eds), pp. 31-75, CRC (1982) in Genetic Engineering (Setlow, Press J. K. and Hollaender, A., eds), Vol. 4, 22 Kirby, R. and Hopwood, D. A. (1977) pp. 119-145, Plenum Press ft. Gen. Microbiol. 98, 239-252 4 Bibb, M. J., Chater, K. F. and Hopwood, D. A. (1983) in Experi- 23 Chater, K. F. (1983) in Proceedings of the 4th International Symposium on mental Manipulation of Gene Expression Genetics of Industrial Microorganisms, (Inouye, M. ed), Academic Press (in Kyoto (in press) press) 5 Bibb, M. J., Ward, J. M. and 24 Bibb, M. J. and Cohen, S. N. (1982) Hopwood, D. A. (1978) Nature 274, Mol. Gen. Genet. 187, 265-277 398-400 25 Thompson, C. J. and Gray, G. S. (1983) 6 Suarez, J. E. and Chater, K. F. (1980) Proc. Nat. Acad. Sci. USA (in press) J. Bacteriol. 142, 8-14 26 Robbins, P. W., Wirth, D. F. and 7 Krllgel, H., Fiedler, G. and Noack, D. Hering, C. (1981)ff. Biol. Chem. 256, (1980)Mol. Gen. Genet. 177, 297-300 10640-10644 8 Bibb, M. J., Schottel, J. L. and Cohen, 27 Schupp, T., Toupet, C., StalhammarS. N. (1980) Nature 284, 526-531 C.arlemalm, M. and Meyer, J. (1983) 9 Thompson, C. J., Ward, J. M. and MoL Gen. Genet. 189, 27-33 Hopwood, D. A. (1980) Nature 286, 28 Losick, R. and Pero, J. (1981) Cell 25, 525-527 582-584 10 Thompson, C. J., Ward, J. M. and 29 Hopwood, D. A. (1981) in [$-lactam Hopwood, D. A. (1982) J. Bacteriol. Antitn'otics (M. R. J. Salton and G. D. 151, 668-672 Schockman, eds), pp. 585-598, Aca11 Bibb, M. J., Ward, J. M., Kieser, T. demic Press and Hopwood, D. A. (1981)Mo/. G-on. 30 Hopwood, D. A., Wright, H. M., Bibb, Genet. 184, 230-240 M. J. and Cohen, S. N. (1977) Nature 12 Thompson, C. J., Kieser, T., Ward, 268, 171-174