SCP1, coding for the methylenomycin biosynthesis genes in Streptomyces coelicolor, was shown to be a giant linear plasmid of 350 kb with a copy number of about four by analysis with pulsed-field gel electrophoresis. A detailed physical map of SCP1 was constructed by extensive digestion with six restriction endonucleases, by DNA hybridization experiments, and finally by cloning experiments. SCP1 has unusually long terminal inverted repeats of 80 kb on both ends and an insertion sequence at the end of the right terminal inverted repeat. Analysis by pulsed-field gel electrophoresis in agarose containing sodium dodecyl sulfate revealed that a protein is bound to the terminal 4.1-kb SpeI fragments derived from both ends of SCP1. Treatment with k exonuclease or exonuclease III and SpeI digestion also indicated that the 5' ends of SCP1 are attached to a protein.The application of pulsed-field gel electrophoresis (PFGE) (6, 37) to Streptomyces DNA enabled the detection of giant linear plasmids of 90 to 590 kb from antibiotic-producing strains such as Streptomyces lasaliensis, S. violaceoruber, S. fradiae, S. parvulus, S. venezuelae, and S. rochei (23). This method also revealed that SCP1, a plasmid in S. coelicolor which had been sought for a long time as a genetic determinant of methylenomycin biosynthesis (25, 26), was a giant linear plasmid (24). Southern blot analysis with pIJ601 (1), a plasmid containing the methylenomycin resistance gene, confirmed that SCP1 carries the methylenomycin biosynthetic gene cluster (7).Extensive genetic studies by Hopwood and his collaborators revealed that SCP1 is present in various states in S. coelicolor strains (for reviews, see reference 8), namely as an autonomous replicating plasmid, as an autonomous replicating plasmid containing a chromosomal fragment (SCP1'-cysB), and integrated into the chromosome (NF strains). We confirmed these various states of SCP1 by PFGE analysis in a preliminary investigation (22) and further demonstrated that S. violaceoruber JCM4979 contains a series of plasmids (410 to 590 kb) with a size difference of about 30 kb (22). These plasmids were suggested to be formed by integration of a circular sex plasmid, SCP2 (4, 27), into SCP1 and by subsequent amplification and deletion. S. violaceoruber JCM4979 was derived from Erikson's S. coelicolor A3(2) strain (11), with which Hopwood started the genetic study (16). Therefore, S. violaceoruber JCM4979 and S. coelicolor 1147, a wild-type culture of S. coelicolor A3(2) at the John Innes Institute, are apparently identical. We thus believed at first that strain 1147 had the same series of plasmids as had strain JCM4979. This was not found to be the case, however; strain 1147 has only SCP1 and not the plasmid ladder, which is also a characteristic of strain M138 (24).To study the details of the structures of various states of SCP1 in S. coelicolor strains and to elucidate the mechanism of its dynamic changes, we undertook the physical charac-* Corresponding author.
