Introduction

Streptomyces are a group of aerobic high %G+C Gram positive bacteria that undergo complex differentiation to form filamentous mycelium, aerial hyphae and spores. In addition, they produce a broad range of secondary metabolites including antibiotics, antiparasitic agents, herbicides, anti-cancer drugs and various enzymes of industrial importance. At present, two Streptomyces species have had their complete genome sequences published, namely the model organism Streptomyces coelicolor (%G+C=72.1) and avermictin producer Streptomyces avermitilis (%G+C=70.7) (Bentley et al. 2002; Ikeda et al. 2003), although a number of other sequences are close to publication. Two important aspects of the genomes structures of Streptomyces have been supported by sequence data. Firstly, that the genome size of Streptomyces is large compared to other bacteria; 8,667,507 basepairs for S. coelicolor (7825 protein coding genes) and 9,025,608 basepairs (7,577 protein coding genes) for S. avermitilis. Secondly, that the genomes of these two species are linear and both ends contain unique terminal inverted repeats that probably covalently bind a terminal protein. (Lin et al. 1993; Chen et al. 2002; Huang et al. 1998). Over 2500 Streptomyces strains are present in the Ribosomal Database Project (www.rdp.cme.msu.edu), over 1500 are available at the American Type Culture Collection (http://www.atcc.org/) and many more are held in both public and private culture collections throughout the world. Analysis of the small subunit ribosomal RNA gene sequences of Streptomyces confirms that they form a monophyletic clade, but one with considerable diversity. In addition, there is significant gene diversity at the interspecies level across the genomes of both completely sequenced Streptomyces with 2291 gene unique to S. avermitilis and 2307 genes unique to S. coelicolor. The comparative genomics of three species of the Streptomyces and one species of Kitasatospora have been analyzed by DNA/DNA microarray analysis and this showed that firstly the microarray data agreed well with the genome sequence data of S. avermitilis and secondly that the four species analyzed produced genome patterns that was congruent with each other (Hsiao and Kirby 2007). The available bioinformatics and microarray data suggest that Streptomyces and closely related genera have a common core region of the genome that is also present in many other Actinobacteria including Mycobacterium and Corynebacterium (Bentley et al. 2002). Outside of this core region there are two regions containing genus (or perhaps wider) specific genes. Finally, at the terminal regions of the linear chromosomes are species-specific regions that contain only a limited number of conserved genes, some of which are involved in terminal replication, including tpgA and tapA.

In this study we expanded the analysis to include a wider range of Streptomyces including Streptomyces rimosus (Petkovic et al. 2006) as well as two non-Streptomyces myceliate Actinomycetales, Streptosporangium roseum and Saccharomonospora viridis. All tested species except for Streptomyces rimosus showed congruency of the microarray pattern, suggesting that S. rimosus may have a novel genome lineage compared to other Streptomyces.

Materials and methods

16S phylogeny

This was carried out on the small subunit 16S ribosomal RNA gene sequences (bp 91–447, S. coelicolor A3(2) 16S DNA sequence AL939108)obtained from Ribosomal Database Project-II Release 9 (http://www.rdp.cme.msu.edu/index.jsp) and PCR cloning/sequencing of various S. rimosus strains. These were aligned using CLUSTALX (Thompson et al. 1997). The analysis was carried out using Neighbor-Joining algorithm (Gascuel and Steel 2006) from the same program.

Arrays

PCR arrays covering about 97% of the complete genome of Streptomyces coelicolor A3(2) (www.surrey.ac.uk/SBMS/Fgenomics/Microarrays/index.html) were used in this study. The Surrey microarray is made up of 7758 unique PCR amplified sequences, 7563 from the chromosome and 195 from SCP1, the large linear plasmid found in S. coelicolor that encodes methylenomycin (Kirby et al. 1975; Kirby and Hopwood 1977; Bentley et al. 2004). There are an additional 376 non-unique, alternative and cross-hybridizing sequences that are also spotted on to the array together with no probe spots and control spots. The Surrey array does not include a number of transposition element related genes. The sequences of the PCR products are not available due to intellectual property protection requirements.

Strains and growth conditions

The following species were used for the interspecific comparative genomics part of this study: Streptomyces coelicolor A3(2) (SCP1+) 104, Streptomyces antibioticus (ATCC15848), Streptomyces argenteolus (ATCC11009), Streptomyces avermitilis (ATCC31267), Streptomyces bikiniensis (ATCC11062), Streptomyces cattleya (ATCC35852), Streptomyces clavuligerus (ATCC27064), Streptomyces lipmanii (ATCC27357), Streptomyces rimosus subsp. rimosus G7 (type strain isolated by Pfizer Ltd and purportedly ATCC10970), Streptomyces scabies (BCRC13682), Streptomyces tanashiensis (ATCC23967), Streptomyces virginiae (ATCC12630), Streptomyces aureofaciens (ATCC10762), Streptosporangium roseum (ATCC12428) and Saccharomonospora viridis (ATCC15345). For the intraspecific comparative genomics aspects of this study, the following S. rimosus strains were used, G37, 23383, M4018 and S15883, in addition to S. rimosus G7. These strains are all linear descendants of G7 over about 30 years and have undergone mutagenesis and selection affecting various aspects of oxytetracycline production and other phenotypic characteristics. Table 1 includes descriptions of all the Actinobacteria used in this study. Fresh spores were collected and mycelium cultured in TSB liquid medium with 0.5% glycine at 30°C for 3 days.

Table 1 Species and strains of Actinobacteria used for bioinformatical and experimental purposes in this study
Full size table

Preparation of labeled DNA

Genomic DNA from a stationary phase culture was purified by the salting out procedure (Pospiech and Neumann 1995) and had been sonicated to <2Kb. Four to six micrograms of sonicated genomic DNA were used as template and this was denatured in the presence of 12 μg of 72%-GC-content random hexamers in a total volume of 25 μl at 100°C for 10 min. The mixture was then snap-cooled on ice before adding the remaining reaction components: About 1.5 μl of Cy3-dCTP or Cy5-dCTP (Amersham Pharmacia Biotech), 4 μl Klenow fragment (NEB #212), 5 μl Klenow buffer, 0.5 μl dNTP (4mM dATP, 4 mM dTTP, 10 mM dGTP, and 0.2 mM dCTP), and 14 μl ddH2O. The random primed labeling reaction was carried out for 2–3 h at 37°C. Buffer exchange, purification and concentration of the DNA products was accomplished by three cycles of diluting the reaction mixture in 0.5 ml TE buffer (10 mM Tris and 1 mM EDTA pH 8.0) and filtering though a Microcon-30 microconcentrators (Millipore).

Microarray hybridization and data analysis

In all cases array hybridizations were carried out in duplicate. The two DNA pools to be compared were mixed and applied to an array in a hybridization mixture that contained 3.68X SSC, 0.18% SDS, and 1 μg yeast tRNA (total 16.3 μl), which had been heated at 100°C for 5 min before being applied to array. Hybridization took place under a glass coverslip sealed by glue in a humidified Omnislide (Thermo Hybaid) at 60°C for 12–14 h. The slides were washed, dried and scanned for fluorescence using a GenePix TM 4000B scanner (Axon instruments). Average signal intensity and local background measurements were obtained for each spot on each array using GenePixPro software. The dataset was screened for aberrant spots and these were eliminated from the analysis after manual checking. The signal from each gene spots was inputed into the computer program ScanAlyze (Eisen et al. 1998; Gollub et al. 2003). The data was then processed into a mean Log2 Cy3/Cy5 ratio format. The dataset was normalized for each array separately and outputed to Excel where after checking the alignment of the datasets from each array, a mean signal for each common gene was calculated. Based on Bentley et al. 2002, the mean signal and standard deviation for the core region of genes from SCO2050 to SCO5800 was calculated. The standard deviation was used to set a cut-off for gene absence at 2SD below the core mean signal. The microarray data is presented relative to the S. coelicolor standard in two ways. This is either as a color plot of the genes where green presents a negative hybridization signal, black represents an equal hybridization signal and red indicates a positive hybridization signal using the program Treeview (Eisen et al. 1998), or as numeric values for the signal from each gene. The microarray data for the S. rimosus species described here can be accessed at NCBI Gene Expression Omnibus (GSE10126).

Genome comparison of the actinobacteria

The program Mauve (Darling et al. 2004; http://www.gel.ahabs.wisc.edu/mauve) is a system for efficiently constructing multiple genome alignments in the presence of large-scale evolutionary events such as rearrangement and inversion. Multiple genome alignment provides a basis for research into comparative genomics and the study of evolutionary dynamics. Mauve was used at its default settings to compare the genomes of selected Actinobacteria with complete genome sequences against the linear genome of S. coelicolor. This is done in order to confirm the presence of a core conserved syntenous region across most Actinobacteria. These Actinobacteria were selected based on the fact they had the largest genomes within a genus and that they covered a wide range across the Actinobacteria; they were Bifidobacterium longum, Frankia alni, Leifsonia xyli, Mycobacterium vanbaalenii, Nocardia farcinicia, Nocardiodes sp., Rhodococcus sp. RHA1 and Thermobifidia fusca. The choice of the largest genome within a genus was in order to minimize any effects due to gene loss, particularly in pathogens and symbiotes. It should be noted that all the genomes analyzed in this part of the study are believed to be circular with the notable exception of Rhodococcus sp. RHA1 (Mcleod et al. 2006).

Results and discussion

Phylogeny of Streptomyces rimosus

Figure 1 shows a SSU rRNA gene phylogenetic tree of the species that were analyzed by DNA/DNA genomic microarray analysis in this study together with a range of Actinobacteria that have complete genome sequences and a number of other selected Actinobacteria including other Streptomyces spp. It is clear that the Streptomyces as a genus form a distinct clade with Kitasatospora species with a good bootstrap value of 898/1000. This clade includes both various Streptomyces rimosus subsp. rimosus strains, Streptomyces rimosus subsp. paromomycinus and Streptomyces platensis, all of which are closely related to Streptomyces rimosus G7 and its derivatives. These strains also form a clade with quite good bootstrap support (610/1000) with only the presence of Kitasatospora azaticus disrupting the group. Furthermore, the situation is made more complex by the presence of two further clades near the root of the Streptomyces/Kitasatospora clade. These are a Kitasatospora subclade, which supports the distinctness of Kitasatospora, and another containing Streptomyces cattleya and Streptomyces griseocarneus. Together with Streptomyces roseoverticillatus on the other side of the S. rimosus group, these suggest that there may be something novel about the species close to the base of the Streptomyces/Kitasatospora clade. In terms of the organisms included in this study, Streptosporangium roseum and Saccharomonospora viridis are both outside of the clade containing the Streptomyces and Kitasatospora. When the whole tree is examined, Saccharomonospora viridis groups with Amycolatopsis azureus in a sister clade to Streptomyces/Kitasatospora with good bootstrap support and Streptosporangium roseum is found in a quite separate clade with Thermobifidia fusca and Frankia alni at some distance from the Streptomyces/Kitasatospora clade, also with good bootstrap support. The Streptomyces species analyzed in this study by microarray genomotyping are spread evenly across the Streptomyces/Kitasatospora clade. Thus, while SSU rRNA gene phylogenetic analysis places S. rimosus at the edge of the Streptomyces clade, S. cattleya, does branch outside of S. rimosus, retains the genome profile found in the other Streptomyces found in the main part of the clade. The molecular phylogeny of S. cattleya conforms well with the phylogeny proposed earlier based on phenotypic characters (Williams et al. 1983a, b). Furthermore, this phylogeny also places S. cattelya and S. rimosus at the boundary of the Streptomyces clade.

Fig. 1
figure 1

SSU rRNA gene phylogenetic tree of selected Actinobacteria including those analyzed in this study by microarray comparative genomics. The sequences were aligned using CLUSTALX and the tree produced by Neighbor-Joining using the same program

Full size image

Genome analysis by DNA/DNA microarray hybidization

The initial species analyzed by DNA/DNA genomic microarray analysis in this study include 12 Streptomyces species, one of which is Streptomyces rimosus subsp. rimosus G7, type strain of this species obtained from Pfizer & Co. Inc. and purported to be the origin of ATCC10970; in addition a Streptosporangium species and a Saccharomonospora species were included. Figure 2 shows a pseudo-color display of the microarray data covering 7083 genes from the S. coelicolor genome sequence. It is clear from this that all the species analyzed except for S. rimosus show a relatively similar pattern of DNA conservation. This include S. cattleya and the two none Streptomyces species, Saccharomonospora and Streptosporangium, although the overall signal for Streptosporangium is distinctly weaker than for the other species. Bentley et al. (2002) identified using the S. coelicolor genome sequence that there was a conserved core of genes within the S. coelicolor chromosome from about SCO2050 to SCO5800. These genes are present in a number of other Actinobacteria genomes. Supplementary Fig. 1 shows analysis of a wider range of Actinobacteria genomes than that available to Bentley et al. analyzed at the DNA using the program Mauve and this confirms that this core region is quite ubiquitous across many Actinobacteria. In terms of the phylogenetic tree in Fig. 1, this core structure only seems to break down once the clades containing Bifidobacterium longum is reached and includes both linear and circular genomes. Thus, the result whereby the majority of species analyzed by microarray comparison in this study have a conserved genome pattern is not that unexpected and it is suggested that a core genome structure in terms of gene synteny may be the norm in the Actinobacteria. It should be noted that microarray analysis does not imply that these species have a similar linear genome structure to S. coelicolor, S. avermitilis, S. scabies and Rhodococcus sp. RHA1, but the results do suggest this as a first approximation if linear replication associated genes seem to be present. Another specific characteristic of linear Actinobacteria genomes would seem to be the absence in the available sequences of xerC and xerD, the gene products of which are a site-specific recombinase involved in the resolution of circular dimers and recB. However, the comparative genomic approach used here cannot detect genes that are already absent in S. coelicolor. Thus it is not unexpected that S. cattleya, Saccharomonospora and Streptosporangium retain a conserved pattern by microarray analysis. What is unexpected is that Streptomyces rimosus G7, the type strain for S. rimosus, does not.

Fig. 2
figure 2

Analysis of the genomes by comparative whole genome DNA microarray analysis of Streptomyces antibioticus, Streptomyces argenteolus, Streptomyces aureofaciens, Streptomyces avermitilis, Streptomyces bikiniensis, Streptomyces cattleya, Streptomyces clavuligerus, Streptomyces lipmanii, Streptomyces rimosus G7, Streptomyces scabies, Streptomyces tanashiensis, Streptomyces virginiae, Streptosporangium roseum and Saccharomonospora viridis using a Streptomyces coelicolor genome microarray against Streptomyces coelicolor A3(2) 104. The arrays are normalized using the core region of Streptomyces coelicolor; green shows a lack of hybridization, black shows equivalent hybridization and red shows a stronger average hybridization. The five regions of the Streptomyces coelicolor genome described in Hsiao and Kirby 2007 are also indicated

Full size image

An earlier study by our laboratory using DNA/DNA genomic microarray analysis of four species confirmed that the terminal regions of these species seemed to be poorly conserved (SCO0001 to SCO01100 and SCO7600 to SCO7850). Furthermore, between the core regions and these terminal regions, two regions of conservation within the Streptomyces, perhaps with even higher conservation than the core region, could be detected and these regions show synteny when the genome sequences of other Streptomyces are compared to S. coelicolor (Chater, KF, unpublished and analysis of S. coelicolor, S. avermitilis and S. scabies genome sequences by Artemis (http://www.sanger.ac.uk/Software/Artemis)). If we examine these genus specific regions in Fig. 2, it will be seen that the results for most strain support a higher level of gene conservation in these two regions as expressed by a red shift, this is most clear in the left region. This structural pattern is also present S. cattleya, Saccharomonospora viridis and Streptosporangium roseum, although the divergence of Streptosporangium roseum has weakened the effect in this species. This further supports a similar genome structure across these strains.

Gene presence and absence

As a result of the fact that the one clear exception to this congruent pattern in Fig. 2 is S. rimosus, the major horizontally transferred regions (HTRs) within the S. coelicolor genome postulated by Bentley et al. (2002) and expanded by Hsiao and Kirby (2007) were examined. These regions were also identified by microarray analysis between S. coelicolor and the very closely related species Streptomyces lividans by Jayapal et al. (2007). Taking the slightly different criteria used to define an HTR in the three publications, there is almost a 100% agreement between them. Hsaio and Kirby (2007) and Jayapal et al. identify more HTRs than Bentley et al. (2002), but with excellent inclusivity and excluding the terminal regions, only 4/22 unique to their study HTRs are postulated by Hsiao and Kirby and 3/21 by Jayapal et al. (Fig. 3). Based on this good agreement, Supplementary Fig. 2 shows examples of some of the HTRs (the possible HTRs in S. coelicolor may be as many as 27 or more when the larger microarray dataset used here is examined). In all eight cases, the genes within the HTR are absent from most or all of the species analyzed with the exception of S. rimosus. S. rimosus, on the other hand, shows high signal levels within the HTR regions for more but by no means all genes, with perhaps HTR-L, HTR-T and HTR-V being the best examples. This implies the presence of some genes that are present in S. coelicolor but seem to be generally absent from most Streptomyces. The implication of this is not that the HTRs as a whole are present in S. rimosus, but that unlike the main group of Streptomyces, S. rimosus contains some genes that have been only found as yet in the S. coelicolor HTRs, which may have origins outside of the main group of Streptomyces. These genes represent common ancestral genes that have been lost by most Streptomyces, but have re-enter S. coelicolor as HTRs. Notwithstanding what their origin may be in either S. coelicolor or S. rimosus, the absence of clearly defined HTRs found by both Hsaio and Kirby and Japayal et al. further supports the novel genome structure for S. rimosus earlier proposed based on the overall microarray hybridization pattern.

Fig. 3
figure 3

A scale analysis of the open reading frames of S. coelicolor showing the horizontally transferred regions or HTRs identified by Hsiao and Kirby (2007) (HTR B-W; upper dark grey) and the large and small genetic islands identified by Jayapal et al. (GI 1-5 and Gi 1-18, respectively; lower light grey). Note that the terminal regions defined in this paper using the larger dataset than Hsiao and Kirby includes HTR A within the terminal region. Furthermore, the chromosome ends defined by Jayapal et al. are not the same as the terminal regions defined by Hsiao and Kirby. The latter were defined on lack of gene signal across four Streptomyces, while the former show that the terminal regions of S. coelicolor resemble those of S. lividans much more closely than those present in the four Hsiao and Kirby strains. The “*” indicates a HTR identified in Bentley et al. (2002) and all are confirmed by Hsiao and Kirby and by Jayapal et al. There are some discrepancies in start positions, end position and size between the HTRs and the GIs/Gis. This is to be expected, not least because the hybridization conditions used for the two microarray analysis datasets were not identical

Full size image

In order to avoid the possibility of a systematic error, the SSU rRNA gene sequence of S. rimosus G7 and its selected derivatives G37, 23383, M4018 and S15883, were amplified by PCR and sequenced. The sequence was 99.5% identical to the sequence available from Ribosomal Database Project-II Release 9 over about 360bp. The initial analysis using two microarrays of S. rimosus G7 was then repeated using freshly isolated DNA from a new single colony by a different scientist. The results were 97% in agreement with the previous microarray data. The above S. rimosus strains were derived by mutation and selection from S. rimosus G7 and represent 30 years of selection. When these strains were compared to S. rimosus G7 by genomotyping, they gave almost identical microarray profiles (Supplementary Fig. 3). Variations in signal for individual genes was less the 3% and no cases of gene deletion or gene duplication were found between these five strains.

The dataset from two specific groups of genes were analyzed in more detail. The first was the ribosomal proteins in order to confirm the validity of the dataset. Table 2 shows that Log2 signal levels for these genes and it is clear that they show a high level of conservation as would be expected. The exceptions are few in number. Two each are scored as absent in Streptosporangium and Saccharomonspora, four in S. cattleya, but eleven in S. rimosus . This supports the divergent nature of the S. rimosus genome compared to even certain non-Streptomyces mycelate Actinomycetales. One particular ribosomal protein, SCO0569 (L36) is known to have a duplicated elsewhere on the S. coelicolor chromosome and therefore is probably a pseudogene in S. coelicolor. SCO0569 is scored as absent in many Streptomyces species Interestingly, as shown in Table 2, the main copy of L36, SCO4726 is scored as absent in S. cattleya, but SCO0569 is scored as present in S. cattleya, which contrasts with Streptosporangium, Saccharomonspora and S. rimosus; this suggests perhaps that SCO0569 is a horizontally transferred pseudogene from a distant species.

Table 2 Microarray genomotyping data for Streptosporangium roseum, Saccharomonospora viridis, Streptomyces rimosus G7 and Streptomyces cattley a
Full size table

Certain genes can be specifically associated with the Streptomyces linear chromosomal model and these include TapA, TpgA and TtrA. TapA and TpgA have been show to be involved in terminal capping of the linear chromosome and linear replication (Bao and Cohen 2001; Bey et al. 2000; Yang et al. 2002), while TtrA seems to have a function associated with genetic exchange of linear DNA between Streptomyces strains (Bey et al. 2000). Table 3 shows the microarray data for these three genes across the species studied here and almost all the species show a signal level that implies all three genes are present; this includes S. rimosus. This cutoff is based on 2SD below the mean core signal level for a strain (Hsiao and Kirby 2007). Interestingly, however, the two Streptomyces at the extreme edge of the Streptomyces clade, S. cattleya and S. rimosus do show the lowest signal for tpgA, and this is very close to the boundary for 2SD. The signals for tapA are significantly higher implying the low tpgA signal may be due to evolutionary divergence. The exception to the conservation of tpgA and tapA is Saccharomonospora viridis, where the signal suggests that tpgA is absent but tapA is present; in this case too, gene divergence might be the explanation. In S. scabies, S. virginiae and Streptosporangium roseum, the signals suggest that ttrA might be absent. This is especially unexpected for S. scabies as this gene seems to be present based on Blast analysis of the preliminary genome sequence from http://www.sanger.ac.uk/Projects/S_scabies/. However, the S. scabies strain used in this study is not the same as used for genome sequencing and therefore a difference is not inexplicable, particularly when the diversity of the S. scabies-like pathogens is taken into account. It is also important to note that ttrA is found at the extreme end of the linear chromosome of all known chromosomal terminal regions, namely S. coelicolor, Streptomyces lividans, S. avermitilis, S. scabies and S. ambofaciens and therefore is particularly vulnerable to circularization, which can occur at relatively high frequencies in Streptomyces (Lin and Chen 1997; Choulet et al. 2006a; Choulet et al. 2006b).

Table 3 Microarray data for the genes tpgA, tapA and ttrA for Streptomyces antibioticus, Streptomyces argenteolus, Streptomyces aureofaciens, Streptomyces avermitilis, Streptomyces bikiniensis, Streptomyces cattleya, Streptomyces clavuligerus, Streptomyces lipmanii, Streptomyces rimosus G7, Streptomyces scabies, Streptomyces tanashiensis, Streptomyces virginiae, Streptosporangium roseum and Saccharomonospora viridis
Full size table

Failure to detect major differences after significant mutation and selection over 30 years

The lack of divergence between the S. rimosus strains analyzed in this study is most interesting and unexpected (Supplementary Fig. 3). These strains have undergone mutation and selection for tetracycline production over a period of about 30 years and show phenotype variation associated with this selection (data not shown). Although microarray genomotyping of the type used here does not allow study of genes novel to the strain being analyzed (S. rimosus in this case and an example would be the oxytetracycline biosynthesis cluster), the lack of changes in genes conserved between S. coelicolor and S. rimosus suggests that point mutations may play an important role in any changes to central metabolism induced during selection for higher antibiotic production. In other words, while it is possible that gene duplication/deletion does occur with respect to metabolic pathway not detected by the microarray approach, changes to core metabolic functions would seem to be less drastic, even though changes in the metabolic flux within the cell are probably essential to provide enough appropriate metabolites for high antibiotic production.

Conclusions

Based on the microarray results presented in this study, we suggest that the genome of S. rimosus ATCC10970 has a different structure from the majority of Streptomyces. Interestingly, the authors were unable to find any pattern to the genes that seemed to be conserved between S. coelicolor and S. rimosus, suggesting that the genome differences are major and not just vertical evolutionary divergence, the possibility remains that, based on the evidence for conservation of tapA, tpgA and ttrA, the chromosome of S. rimosus is linear like the chromosomes of other Streptomyces that have been investigated, but that the genes making up the chromosome are highly divergent from those of other Streptomyces. In addition, it is likely that the synteny found across the Actinobacterial core region, as well as the Streptomyces genus specific regions defined in Hsiao and Kirby (2007), are not present in S. rimosus. Alternatively, if the Actinobacterial core region is present, it would seem to be highly disrupted and/or highly divergent from other Streptomyces that have been analyzed. We conclude that S. rimosus may represent a distinct evolutionary lineage making it a worthy and important candidate for genome sequencing in the future. Furthermore, based on the phylogenetic clustering in Fig. 1, an extension of the DNA/DNA microarray genomotyping to various Kitasatospora spp., Streptomyces platensis, Streptomyces griseocarnus, Streptomyces roseovertillatus and Streptomyces rimosus paromomycinus would help to define how wide spread this structure might be. Finally, point mutations between the various high production strains derived from S. rimosus G7 would seem to be the most probable source of variation in the core metabolic genes that allow increased oxytetracycline biosynthesis.