Cloning and Physical Mapping of the EcoRI Fragments of the ...EcoRI probe to the partial EcoRI digest revealed bands of 4.3, 13, 27, 33, 46, and 47 kb (Fig. 2B). This result determined

JOURNAL OF BACTERIOLOGY,0021-9193/98/$04.0010

May 1998, p. 2796–2799 Vol. 180, No. 10

Copyright © 1998, American Society for Microbiology

Cloning and Physical Mapping of the EcoRI Fragmentsof the Giant Linear Plasmid SCP1

MATTHIAS REDENBACH,1,2 KAZUYA IKEDA,1 MASAYUKI YAMASAKI,1

AND HARUYASU KINASHI1*

Department of Molecular Biotechnology, Graduate School of Engineering, Hiroshima University,Higashi-Hiroshima 739-8527, Japan,1 and Genome Research Unit, Department

of Genetics, Kaiserslautern University, D-67653 Kaiserslautern, Germany2

Received 25 November 1997/Accepted 10 March 1998

A cosmid library was constructed for the 350-kb giant linear plasmid SCP1 and aligned on a successive lin-ear map. Only a 0.8-kb gap has remained uncloned in the terminal inverted repeats close to both ends. Partialdigestion of the aligned cosmids with EcoRI and hybridization with the flanking fragments of the vector enabledphysical mapping of all of the EcoRI fragments. On this map, the methylenomycin biosynthetic gene cluster,the insertion sequence IS466, and the sapCDE genes coding for spore-associated proteins were localized.

Linear plasmids have been found in a variety of eucaryoticspecies, but have rarely been identified in procaryotic organ-isms. However, the filamentous soil bacteria Streptomyces spp.possess frequently linear plasmids, which seem therefore to bea significant feature of this genus (12). The plasmids vary insize between 12 kb (28) and 1 Mb (24a). They contain terminalinverted repeats (TIRs) which range between 44 bp in SLP2 ofStreptomyces lividans (3) and 95 kb in pPZG101 of Streptomycesrimosus (6). The 59 ends of all of the plasmids investigated wereshown to be blocked by a protein.

Lin et al. (20) first demonstrated that S. lividans has an 8-Mblinear chromosome. The linear chromosome topology has beenproved for Streptomyces griseus (19), Streptomyces ambofaciens(18), Streptomyces coelicolor A3(2) (26), and S. rimosus (25).The close structural similarities between Streptomyces chromo-somes and linear plasmids suggest a hypothetical idea thatlinear plasmids were generated by recombination of linearchromosomes, or circular chromosomes were linearized by in-tegration of a linear plasmid.

SCP1 from S. coelicolor A3(2) is genetically the best-studiedlinear plasmid. The element long remained unisolated but waspostulated on the basis of classical genetic analysis (27). It wasshown to carry genes for the biosynthesis of and resistance tothe antibiotic methylenomycin (2, 17). With the introduction ofpulsed-field gel electrophoresis (PFGE), SCP1 was revealed tobe a 350-kb giant linear plasmid containing 80-kb-long TIRs(13, 15).

Considerable attention has been paid to SCP1 because ofits interaction with the host chromosome. Freely replicatingSCP1 mobilizes chromosomal markers randomly, as seen forall fertility plasmids in Streptomyces. In addition, Hopwood andcolleagues isolated mutants carrying an SCP1-chromosome hy-brid structure which promotes a directional DNA transfer dur-ing conjugation (9, 10). Free SCP1-prime plasmids containinga certain chromosomal DNA stretch were found as well as mu-tants possessing SCP1 integrated in a central region of theS. coelicolor chromosome. Those SCP1-integrated strains showedeither unidirectional or bidirectional DNA transfer with re-spect to the SCP1 integration site. Preliminary structural anal-

ysis of the integrated copies of SCP1 has been carried out forS. coelicolor A3(2) strain 2612 (an NF strain) (8) and strainsA608 (a pabA donor) and A634 (an NF-like donor strain) (16).

To reveal in more detail the genetic organization of SCP1and its role in conjugative DNA transfer, we established anordered contig of cosmid clones covering most parts of theelement.

Generation of an ordered cosmid library. A cosmid librarywas prepared for total DNA of S. coelicolor 1147, a wild-typeA3(2) strain, which carries both SCP1 and a 31-kb circularplasmid, SCP2 (22). Total DNA was isolated and partially di-gested with Sau3AI to an average size of 40 to 60 kb. As de-scribed previously (26), the DNA was ligated to the vector Su-percos-1 (5) (Stratagene, La Jolla, Calif.) and packaged intolambda phages which were subsequently used to transfect Esch-erichia coli Sure. Two thousand cosmid clones representing thetotal genome of S. coelicolor A3(2) were screened with theSCP1 DNA, which was isolated by contour-clamped homoge-neous electric fields (CHEF) (4) and nonradioactively labeledwith dig-11-dUTP (Boehringer Mannheim, Mannheim, Ger-many). A total of 192 clones which hybridized strongly with theSCP1 DNA were obtained.

Supercos-1 possesses T3 and T7 promoters flanking theBamHI cloning site. DNAs of 25 randomly selected cloneswere isolated and restricted with SalI. The dig-11-dUTP-la-beled transcripts of the cosmid ends were generated from theSalI-digested DNAs by using T3 and T7 RNA polymerases.The labeled end probes were hybridized against colony blots ofSCP1 clones to identify neighboring cosmid clones. Four ad-ditional rounds of hybridizations, including a total of 30 cos-mids, were performed to obtain 11 cosmids which could beunambiguously aligned and cover most parts of SCP1 (Fig. 1).Cosmid 31 is completely located within the TIR and was there-fore used for the alignment at both ends of SCP1. It was notpossible to find a clone among the 192 SCP1 cosmids thatreaches closer than 5.0 kb to the very end, because no cosmidhybridized with plasmid pSCP201, which contains the 4.1-kbend SpeI fragment of SCP1 (14).

Construction of the EcoRI fragment map. To establish aprecise restriction map of SCP1, we chose EcoRI as an appro-priate enzyme. EcoRI generated a suitable number (about 40)of observable SCP1 fragments and could excise the insert fromthe vector. Since SCP1 and Supercos-1 contain only two AseIsites, the 4.3- and 0.9-kb AseI-EcoRI fragments of the latter

* Corresponding author. Mailing address: Department of MolecularBiotechnology, Graduate School of Engineering, Hiroshima Univer-sity, Higashi-Hiroshima 739-8527, Japan. Phone: 81-824-7869. Fax:81-824-24-7869. E-mail: [email protected].

2796

on March 22, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

could be used as probes to determine the order of EcoRIfragments. Each cosmid clone was first completely digestedwith AseI and then partially digested with EcoRI. The partialdigest was separated by CHEF gel electrophoresis and hybrid-ized with the 4.3- and 0.9-kb AseI-EcoRI probes separately.This method also confirmed the direction of the insert, becausethe 4.3- and 0.9-kb fragments carry the T7 and T3 promoters,respectively.

In Fig. 2, the analysis of cosmid 53 is shown as an example.The complete restriction digest of cosmid 53 with AseI andEcoRI gave seven fragments (Fig. 2A): four from the SCP1DNA (14.0, 13.0, 8.6, and 6.5 kb) and three from the vectorDNA (4.3, 1.7, and 0.9 kb). Hybridization of the 4.3-kb AseI-EcoRI probe to the partial EcoRI digest revealed bands of 4.3,13, 27, 33, 46, and 47 kb (Fig. 2B). This result determined theorder of EcoRI fragments from the T7 promoter (right-handed

FIG. 1. Ordered cosmid and restriction maps of SCP1. In the upper part, 12 cosmids covering almost the entire region of SCP1 are aligned on a continuous linearmap. The open and solid circles indicate T3 and T7 promoters, respectively. pSCP201, which carries the 4.1-kb terminal SpeI fragment of SCP1 (14), is included at bothends. A gap has remained uncloned between pSCP201 and cosmid 31. The restriction map of SCP1 is drawn in the center. Sites for EcoRI and rare cutting enzymesare shown above and below the SCP1 line. Previous sequence analysis of pSCP201 revealed that five EcoRI sites are present in the 342 nucleotides at the end of SCP1.Among them, four EcoRI sites, the total size of whose fragments is 0.2 kb, were omitted here to avoid complexity, but they are included in Fig. 3A. The sizes offragments are indicated in kilobases. The positions of the mmr, IS466, and sapCDE genes and two TIRs, left (TIR-L) and right (TIR-R), are shown at the bottom. Thetranscription direction of the mmr gene (24) is indicated by an arrow. Af, AflII; As, AseI; V, EcoRV; Sp, SpeI; Ss, SspI; X, XbaI.

FIG. 2. Restriction analysis of cosmid 53. (A) Conventional agarose gel electrophoresis of the complete digest of SCP1 with AseI and EcoRI. (B and C) Southernhybridization of the partial digest of SCP1. The SCP1 DNA was completely digested with AseI and then partially digested with 3 U of EcoRI in 20 ml of reaction bufferfor 10 and 20 min. The restricted DNAs were separated by CHEF gel electrophoresis and transferred to nylon membranes. Hybridization was carried out with twofragments of the vector as a probe: the 4.3-kb AseI-EcoRI fragment (B) and the 0.9-kb AseI-EcoRI fragment (C). (D) Determination of the real size of the rightmostEcoRI fragment of cosmid 53. The SCP1 DNA was digested with EcoRI, separated by conventional agarose gel electrophoresis, and probed with the 13.0-kb right-endfragment of cosmid 53. The EcoRI digests of cosmids 53 and 31 were used as references. M1, lDNA digested with HindIII; M2, lDNA-Mono Cut Mix (New EnglandBiolabs, Inc., Beverly, Mass.); E, EcoRI; As, AseI.

VOL. 180, 1998 NOTES 2797

on March 22, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

direction) as (4.3)-8.6-14.0-6.5-13.0-(0.9 kb). The same analy-sis was carried out with the 0.9-kb AseI-EcoRI probe, whichshowed hybridizing bands of 0.9, 14, 20, 34, 43, and 47 kb (Fig.2C). The order in the opposite direction from the T3 promoteragreed with the result described above.

The sizes (8.6 and 13.0 kb) of two EcoRI fragments locatedat the left and right ends of cosmid 53 did not reflect real sizes,because both fragments had been cut by Sau3AI before clon-ing. The leftmost fragment was detected in the EcoRI digest ofthe left-hand cosmid 63 and was determined to be 10.5 kb long(data not shown). The size of the rightmost fragment wasdeduced to be 14.0 kb by hybridization of the 13.0-kb right-endfragment of cosmid 53 to the complete EcoRI digest of SCP1itself (Fig. 2D). This was confirmed by direct observation of theidentical fragment in the EcoRI digest of cosmid 11 due to theTIR (data not shown). All of the aligned cosmids except cos-mid 32 were analyzed in this way, and their EcoRI fragmentmaps were constructed.

Cosmid 32 contains one AseI site but no EcoRI site; thus, alarge EcoRI fragment extends over three cosmids, 17, 32, and39. This large fragment was directly observed by CHEF gelelectrophoresis of the EcoRI digest of SCP1 and was deter-mined to be 66 kb long (data not shown).

Although cosmid 31 carries a DNA fragment of the TIRregion, it does not reach the 4.1-kb end SpeI fragment clonedin plasmid pSCP201 (14). A detailed restriction map at the endof SCP1 was constructed by using pSCP201 and cosmid 31,which revealed that the size of the gap between these cloneswas 0.8 kb (Fig. 3A). To close the gap, we tried to clone thefollowing five fragments by using three cloning vectors, pUC19,pBR322, and pACYC184, with different copy numbers: the1.6-kb SpeI-EcoRI fragment, the 2.3-kb EcoRI fragment, the3.8-kb SpeI-EcoRV fragment, the 7.6-kb XbaI-EcoRV frag-ment, and the 20-kb BamHI fragment (Fig. 3A). However, wehave not succeeded in cloning any of these fragments. Thissuggests the possibility that the region around the 0.8-kb gaphas a special structure unclonable in any of the three vectors orthat its protein product has a deteriorative effect on the host.

As a result, the entire region of the SCP1 DNA except fortwo 0.8-kb gaps close to both ends has been cloned, and all ofthe EcoRI fragments were aligned on a linear physical map

as shown in Fig. 1. We previously reported a map of EcoRVfragments of SCP1 (13). (The number of the EcoRV fragmentswas corrected to 16, because the 20-kb band was revealed to bea doublet.) All of the EcoRV sites were now located preciselyon a linear map by restriction analysis of the ordered cosmids.In addition, all of the recognition sites for AflII, AseI, SpeI,SspI, and XbaI were localized on the same map. The total sizeof the EcoRI fragments was calculated to be 363.1 kb, whichwas a little bit larger than the reported size (350 kb) of theintact SCP1 (13). We again tested the size of the intact SCP1by using Saccharomyces cerevisiae AB972 chromosomes as sizemarkers. The intact SCP1 moved at the same rate with chro-mosome III (data not shown), whose size is 350 kb according toLink and Olson (21).

Location of the mmy, IS466, and sapCDE genes. Severalgenes were reported to be present on SCP1. Chater and Bru-ton (2) constructed the restriction maps for the methylenomy-cin biosynthetic gene (mmy) clusters on plasmids SCP1 andpSV1; the latter was suggested to be a 170-kb circular plasmid(1), in contrast to the linear structure of the former. Hybrid-ization experiments with pIJ518 as a probe, which carries the7.5-kb mmy fragment of SCP1 (2) (Fig. 3B), revealed that theresistance gene (mmr) is located on the 10-kb EcoRI fragmentof cosmid 73. The restriction map of the mmy region on cosmid73 agreed with that of Chater and Bruton (2).

We previously reported that the insertion sequence IS466(11) is present just at the inside end of the right TIR of SCP1(14). Analysis of cosmid 63 determined its location on theEcoRI map as shown in Fig. 3C. Guijarro et al. (7) isolatedspore-associated proteins SapA and SapB from S. coelicolorA3(2). The sapA gene coding for the former protein has beencloned (7) and was located on AseI fragment F of the chromo-some (26). McCormick et al. (23) cloned additional sap genes,sapCDE, which, however, seemed not to be essential for sporeformation, because they were mapped to SCP1, and SCP12

strains show no effect on sporulation. Two sets of the sapCDEgenes were located in each of both TIR regions, and theirrestriction maps were constructed (22a). Analysis of cosmid 53could locate the sapCDE genes on the EcoRI map (Fig. 3C).

Ordered clone libraries have proved to be an excellent toolfor analysis of large genomic DNAs. For the 8-Mb S. coelicolor

FIG. 3. Detailed restriction maps of three regions of SCP1. (A) Terminal region of SCP1. Four EcoRI sites at the extreme end of SCP1 are included in this map.The gap between the SpeI site at the right end of pSCP201 and the Sau3AI site at the left end of cosmid 31 was deduced from this map to be 0.8 kb. (B) mmy region.The positions of the mmr gene and plasmid pIJ518 are shown. The map direction is opposite to that of Chater and Bruton (2). (C) Location of IS466, the right TIR(TIR-R), and the sapCDE genes. E, EcoRI; Ba, BamHI; Cl, ClaI; Nd, NdeI; Sp, SpeI; Ss, SspI; V, EcoRV; X, XbaI.

2798 NOTES J. BACTERIOL.

on March 22, 2021 by guest

http://jb.asm.org/

Dow

nloaded from

A3(2) chromosome, an ordered cosmid library has been estab-lished and used extensively for gene mapping (27), and it isnow being used for the genome project in collaboration withthe Sanger Centre and the John Innes Centre. Here we re-ported the generation of an ordered cosmid contig of the giantlinear plasmid SCP1, leaving only two 0.8-kb gaps uncloned. Adetailed EcoRI restriction map was constructed, and the gene-tic markers IS466, mmy, and sapCDE were assigned to specificSCP1 fragments. These results will be of great value for studyof the structure and function of SCP1 and analysis of the in-teraction of SCP1 with the chromosome of S. coelicolor A3(2).

We thank Keith Chater for plasmid pIJ518; John Cullum for pMT644,which carries IS466; and Joe McCormick for pRC2 and pRD2, whichcarry a part of sapC and sapD, respectively.

M. Redenbach has been supported by a scholarship from the JapanSociety for the Promotion of Sciences.

REFERENCES

1. Aguilar, A., and D. A. Hopwood. 1982. Determination of methylenomy-cin A synthesis by the pSV1 plasmid from Streptomyces violaceus-ruberSANK95570. J. Gen. Microbiol. 128:1893–1901.

2. Chater, K. F., and C. J. Bruton. 1985. Resistance, regulatory and productiongenes for the antibiotic methylenomycin are clustered. EMBO J. 4:1893–1897.

3. Chen, C. W., T.-W. Yu, Y.-S. Lin, H. M. Kieser, and D. A. Hopwood. 1993.The conjugative plasmid SLP2 of Streptomyces lividans is a 50 kb linearmolecule. Mol. Microbiol. 7:925–932.

4. Chu, G., D. Vollrath, and R. W. Davis. 1986. Separation of large DNA moleculesby contour-clamped homogeneous electric fields. Science 234:1582–1585.

5. Evans, G. A., K. Lewis, and B. E. Rothenberg. 1989. High efficiency vectorsfor cosmid microcloning and genomic analysis. Gene 79:9–20.

6. Gravius, B., D. Glocker, J. Pigac, K. Pandza, D. Hranueli, and J. Cullum.1994. The 387 kb linear plasmid pPZG101 of Streptomyces rimosus and itsinteractions with the chromosome. Microbiology 140:2271–2277.

7. Guijarro, J., R. Santamaria, A. Schauer, and R. Losick. 1988. Promoterdetermining the timing and spatial localization of transcription of a clonedStreptomyces coelicolor gene encoding a spore-associated polypeptide. J. Bac-teriol. 170:1895–1901.

8. Hanafusa, T., and H. Kinashi. 1992. The structure of an integrated copy ofthe giant linear plasmid SCP1 in the chromosome of Streptomyces coelicolor2612. Mol. Gen. Genet. 231:363–368.

9. Hopwood, D. A., and T. Kieser. 1993. Conjugative plasmids of Streptomyces,p. 293–311. In D. B. Clewell (ed.), Bacterial conjugation. Plenum Press, NewYork, N.Y.

10. Hopwood, D. A., and H. M. Wright. 1976. Interaction of the plasmid SCP1with the chromosome of Streptomyces coelicolor A3(2), p. 607–619. In K. D.MacDonald (ed.), Second International Symposium on the Genetics of In-dustrial Microorganisms. Academic Press, London, United Kingdom.

11. Kendall, K., and J. Cullum. 1986. Identification of a DNA sequence asso-ciated with plasmid integration in Streptomyces coelicolor A3(2). Mol. Gen.Genet. 202:240–245.

12. Kinashi, H. 1994. Linear plasmids from actinomycetes. Actinomycetologica8:87–96.

13. Kinashi, H., and M. Shimaji-Murayama. 1991. Physical characterization ofSCP1, a giant linear plasmid from Streptomyces coelicolor. J. Bacteriol. 173:1523–1529.

14. Kinashi, H., M. Shimaji-Murayama, and T. Hanafusa. 1991. Nucleotidesequence analysis of the unusually long terminal inverted repeats of giantlinear plasmid, SCP1. Plasmid 26:123–130.

15. Kinashi, H., M. Shimaji, and A. Sakai. 1987. Giant linear plasmids inStreptomyces which code for antibiotic biosynthesis genes. Nature 328:454–456.

16. Kinashi, H., M. Murayama, H. Matsushita, and O. Nimi. 1993. Structuralanalysis of the giant linear plasmid SCP1 in various Streptomyces coelicolorstrains. J. Gen. Microbiol. 139:1261–1269.

17. Kirby, R., and D. A. Hopwood. 1977. Genetic determination of methyleno-mycin synthesis by the SCP1 plasmid of Streptomyces coelicolor A3(2).J. Gen. Microbiol. 98:239–252.

18. Leblond, P., G. Fischer, F.-X. Francou, F. Berger, M. Guerineau, and B.Decaris. 1996. The unstable region of Streptomyces ambofaciens includes210-kb terminal inverted repeats flanking the extremities of the linear chro-mosomal DNA. Mol. Microbiol. 19:261–271.

19. Lezhava, A., T. Mizukami, T. Kajitani, D. Kameoka, M. Redenbach, H.Shinkawa, O. Nimi, and H. Kinashi. 1995. Physical map of the linear chro-mosome of Streptomyces griseus. J. Bacteriol. 177:6492–6498.

20. Lin, Y.-S., H. M. Kieser, D. A. Hopwood, and C. W. Chen. 1993. Thechromosomal DNA of Streptomyces lividans 66 is linear. Mol. Microbiol. 10:923–933.

21. Link, A. J., and M. V. Olson. 1991. Physical map of the Saccharomycescerevisiae genome at 110-kilobase resolution. Genetics 127:681–698.

22. Lydiate, D. J., F. Malpartida, and D. A. Hopwood. 1985. The Streptomycesplasmid SCP2*: its functional analysis and development into useful cloningvectors. Gene 35:223–235.

22a.McCormick, J. R., and R. Losick. Personal communication.23. McCormick, J. R., R. Santamaria, and R. Losick. 1991. The genes for three

spore-associated proteins are encoded on linear plasmid SCP1 in Streptomy-ces coelicolor, abstr. P1-125. Abstracts of the 8th International Symposium onBiology of Actinomycetes, Madison, Wis.

24. Neal, R. J., and K. F. Chater. 1987. Nucleotide sequence analysis revealssimilarities between proteins determining methylenomycin A resistance inStreptomyces and tetracycline resistance in eubacteria. Gene 58:229–241.

24a.Pandza, K., and J. Cullum. Personal communication.25. Pandza, K., G. Pfalzer, J. Cullum, and D. Hranueli. 1997. Physical mapping

shows that the unstable oxytetracycline gene cluster of Streptomyces rimosuslies close to one end of the linear chromosome. Microbiology 143:1493–1501.

26. Redenbach, M., H. M. Kieser, D. Denapaite, A. Eichner, J. Cullum, H.Kinashi, and D. A. Hopwood. 1996. A set of ordered cosmids and a detailedgenetic and physical map for the 8 Mb Streptomyces coelicolor A3(2). Mol.Microbiol. 21:77–96.

27. Vivian, A. 1971. Genetic control of fertility in Streptomyces coelicolor A3(2):plasmid involvement in the interconversion of UF and IF strains. J. Gen.Microbiol. 69:353–364.

28. Wu, X., and K. L. Roy. 1993. Complete nucleotide sequence of a linearplasmid from Streptomyces clavuligerus and characterization of its RNA tran-scripts. J. Bacteriol. 175:37–52.

VOL. 180, 1998 NOTES 2799

on March 22, 2021 by guest

http://jb.asm.org/

Dow

nloaded from