In recent years we carried out genomic research on Actinomycetes using DNA microarrays of the complete Streptomyces coelicolor genome in order to analyze genome evolution in the Streptomyces and spore producing Actinomycetales. This enabled the in vitro mapping of a conserved region present in the linear genome of Streptomyces coelicolor across the Actinomycetales, which was confirmed by recent in silico genomic analysis of new Actinomycete genomes. Furthermore, the presence of two genus specific conserved region on either side of this core was also found, which expands the information obtained by in silico complete genome comparisons. This research enabled the pinpointing of Streptomyces rimosus as a Streptomyces species with a novel genome structure. This was done in collaboration with the Streptomyces rimosus group at the University of Strathclyde, Scotland, UK.

The major limitation of DNA microarrays as a tool is that genes novel to a target species are not detected or explored and producing new microarrays for new species is expensive. Therefore, as the cost of genome sequencing as decreased and with the help of the NYMU Genome center, we have moved on to targeting whole genome sequencing. The first genomes targeted were strains of Streptomyces rimosus again in collaboration with the University of Strathclyde, Scotland, UK. We sequenced a production strain of Streptomyces rimosus at NYMU Genome Center using 454 technology and this was followed by the wild-type strain and three further production strains. This is the first time to our knowledge that industrial production strains have been genome sequenced as a group. Streptomyces rimosus produces oxytetracycline and the titer increase from the wild-type to the final production strain isolated in the 1980's is about 2000 fold or more, depending on the culture conditions. All the changes in these strains were introduced by mutation and selection.

Our genome sequencing of Streptomyces rimosus confirmed that the sequence divergence of this species from Streptomyces coelicolor, the microarray strain, was high. However, Streptomyces rimosus still retained a linear genome structure outline above that is present in all known sequenced Streptomyces and many other Actinomycetes. This confirms the limitation of the DNA microarray approach, although it remains much less labor intensive than whole genome sequencing, which requires extensive annotation to get good comparisons between genomes. We have published the Streptomyces rimosus wild-type genome and are still working on analyzing in parallel the four genome sequences from wild-type and production strains in order to determine why production of oxytetracyline is increased. Previous hypotheses to explain how such large increases in antibiotic production occur are 1) duplication of the production operon for the antibiotic; 2) deletion of other secondary metabolic pathways to allow increased energy flow into the desired pathway; and 3) blocking of specific pathways to shunt precursors into the desired pathway. Comparison of the genomes fro m the wild-type and production strains showed that, although there were limited deletions that fitted hypothesis 2, mainly secondary metabolic pathways remained untouched by deleted; furthermore, deletions did not seem to be affecting metabolite flow either, as is proposed by hypothesis 3. Finally, no duplications of the oxytetraclycine operon were found. The increase in oxytetracycline productions seems to rely on point mutations, possibly affecting promoter activity.

We next expanded out interest into another unusual group of Streptomyces. Almost all Streptomyces are non-pathogenic. In terms of plants, Streptomyces scabies and related species are well known as pathogens of dicotyledonous plants. In terms of human and animal pathogenesis, only Streptomyces somaliensis, Streptomyces sudanensis and a few other related species have been identified as affecting humans. These species cause actinomycetoma or more specifically streptomycetoma in many dry regions of the world, which is very difficult to treat and results in the amputation of limbs from many patients. Our aim was that by obtaining the genome sequences of these pathogens, we will be able to develop new approaches to preventing and curing streptomycetomas. We have successfully sequenced Streptomyces somaliensis and Streptomyces sudanensis by 454 sequencing. These closely related but by no means identical species have the smallest genomes of any linear genome Streptomyces at about 5 Mb. Most Streptomyces have linear genomes of 7 to 8 Mb, with some being as large as 10+ Mb (Streptomyces scabies being a case in point). Linear has not been proved biochemically, but the presence of Tpg, Tap and Ttr genes very close to the end of their linear genomes is highly indicative of linearity. The unusual truncated length of these two genomes fits the idea that these are species that are regular pathogens of mammals and not just opportunistic pathogens. Genome shortening is characteristic of pathogens, especially obligatory pathogens. Interestingly, unlike the situation in Streptomyces scabies, no pathogenic island or islands could be identified. This created a problem in terms of identifying genes involved in human pathogenesis. Such genes ought to be present in both Streptomyces somaliensis and Streptomyces sudanensis, but absent from other Streptomyces. Unfortunately this list was long at >500 genes. In order to reduce this number we next genome sequenced Streptomyces fradiae, the closest available relative to Streptomyces somaliensis and Streptomyces sudanensis, and also the production strain for neomycin. The genome size of this species is relatively normal at about 7Mb and shows high sequence similarity to the pathogenic species. Analysis is continuing at present. The opportunity arose to carry out 3rd generation sequencing at a very good price and we did this with our remain streptomycetoma isolated stain which lacks a species title. This was very successful and we obtain a single 5.3 Mb contig plus two minor contigs (30kb and 16 kb, possibly plasmids). We are now using this single contig to order the contigs available for Streptomyces somaliensis and Streptomyces sudanensis. We hope in the future to 3rd generation sequencing to close these contigs.