Lack of Constant Genome Ends in Streptomyces Coelicolor

was recently shown (HOPWOOD 1965b) that all known markers of Strepto‘“,Ces coelicolor are linked on a circular map. The question then arises: what is the topology of the genome? We may consider three possibilities: (1 ) The genome is a closed loop; (2) The genome is open, but its ends vary in position, so that individual chromosomes are circular permutations of one another ( STREISINGER, EDGAR and DENHARDT 1964); (3) The genome is open, with ends at constant positions. In the absence of special assumptions ( 1 ) and (2) would lead to map circularity, as in Escherichia coli and bacteriophage T 4 respectively. (3) would also do so if the products of even numbers of crossovers were recovered, but not those of odd numbers ( STAHL and STEINBERG 1964). Now, there is evidence (HOPWOOD, manuscript in preparation) that the zygotes of S. ccrelicolor are partial diploids, consisting of a complete chromosome from one parent and an incomplete one from the other. Such merozygotes, as in other bacterial species, give rise to recombinant progeny with complete haploid genomes only by even numbers of crossovers; therefore when a sample of haploid recombinants from a cross is analysed, map circularity results, and would do so even if the genome were open, with ends in constant positions. This situation, then, prevents us from deducing lack of constant genome ends from the finding of map circularity. The present paper describes an analysis of individual heteroclones ( SERMONTI, MANCINELLI and SPADA-SERMONTI 1960), which are heterozygous colonies originating from spores containing partial diploid nuclei ( HOPWOOD, SERMONTI and SPADA-SERMONTI 1963). It is probable that each of the nuclei is derived, by an odd number of crossovers, from a merozygote containing a chromosomal fragment of a certain length, and that the length of the diploid region in the heteroclone corresponds to the length of this fragment ( HOPWOOD, in preparation), but this is not a necessary part of the argument in the present paper. Heteroclones are obtained by plating the products of a cross on a medium selective for wildtype alleles, one from each parent, at two closely linked nutritional loci. The nuclei of the heteroclones are diploid at these closely linked loci, and the diploid region extends, for distances which vary from one heteroclone to another, on either side of these loci. Markers within the diploid region are (with rare exceptions) heterozygous, and those outside it are hemizygous, so that the length of the diploid region in each heteroclone can be determined by analysing the genotypes of the haploid segregants to which the heteroclone gives rise. The present