OUP user menu

Chromosome diversity and similarity within the Actinomycetales

Ralph Kirby
DOI: http://dx.doi.org/10.1111/j.1574-6968.2011.02242.x 1-10 First published online: 1 June 2011

Abstract

Many chromosomes from Actinomycetales, an order within the Actinobacteria, have been sequenced over the last 10 years and the pace is increasing. This group of Gram-positive and high G+C% bacteria is economically and medically important. However, this group of organisms also is just about the only order in the kingdom Bacteria to have a relatively high proportion of linear chromosomes. Chromosome topology varies within the order according to the genera. Streptomyces, Kitasatospora and Rhodococcus, at least as chromosome sequencing stands at present, have a very high proportion of linear chromosomes, whereas most other genera seem to have circular chromosomes. This review examines chromosome topology across the Actinomycetales and how this affects our concepts of chromosome evolution.

Keywords
  • actinomycete
  • chromosomes
  • linear
  • topology
  • Streptomyces
  • Rhodococcus

Introduction

The Actinomycetales are a major order within the high percentage of G+C Gram-positive bacteria and fall within the class Actinobacteria. The order Actinomycetales is made up of 13 suborders covering many species that are important pathogens, relevant to biotechnology and ecologically significant (Zhi et al., 2009). Because of their importance to humans and the environment, many genomes of class Actinobacteria (251), subclass Actinobacteridae (234) and order Actinomycetales (201) have been completely sequenced in the last 10 or so years (as of 8 December 2010 and including draft assemblies; http://www.ncbi.nlm.nih.gov). Thus the genome sequences available for members of the Actinomycetales consist of about a 10th of the available genomes from Bacteria. The importance of these organisms to many fields seems to have focused genome research in the direction of the Actinomycetales.

It is noteworthy that only 36 other chromosomes from the class Actinobacteria have been sequenced. Many, if not most, of the genera making up the Actinomycetales undergo differentiation to a greater or lesser extent (Flärdh & Buttner, 2009). The Actinobacteria are characterized by a unique molecular synapomorphy whereby there is a homologous insertion of about 100 nucleotides between helices 54 and 55 of the 23S rRNA gene (Chater & Chandra, 2006). Furthermore, the Actinomycetales are a coherent clade when analysed phylogenetically using 16S sequences (Fig. 1). This poses the question as to whether the complete chromosome sequences available for the Actinomycetales, when compared with each other, are consistent with the accepted phylogeny evolution. In general, the principal events that shape a bacterial chromosome are gene duplication, horizontal gene transfer, gene loss and chromosomal rearrangements (Andersson & Hughes, 2009). Of these, gene duplication seems to contribute only modestly, horizontal gene transfer seem to be quite important, and gene deletion and genetic drift, which are countered by positive selection, probably vary with ecological niche and the type of chromosome rearrangements. Of these three contributions, it is likely that gene deletion and genetic drift are the most related to evolutionary time because such events are largely dependent on repeated sequences and mobile elements (Ventura et al., 2007). However, up to the present, no reliable method of tracing the evolutionary development of chromosomes in terms of these various events has been successful. Nonetheless, there is evidence to suggest that the Actinomycetales might have enough coherence across their chromosomes to allow some insights into this problem.

1

The genomes of various Actinomycetales were aligned with the 8.67Mb genome of Streptomyces coelicolor (not shown) using mauve with identical settings (http://asap.ahabs.wisc.edu/software/). The coloured blocks represent syntenous regions. These were then compared with a 16S phylogenetic tree of the same species together with other genera of Actinomycetales. The outgroup for the 16S analysis was Bifidobacterium longum and this is included at the top of the alignment. The genera in bold and underlined are those where there is some evidence for chromosome linearity. The genome sequences were obtained from the GenBank database.

Chromosome diversity and similarity within the Actinomycetales are made more interesting because of the topological diversity of their chromosomes; specifically, some families seem to have a preference for linear chromosomes, whereas the majority prefer circular chromosomes (Lin et al., 1993; Reeves et al., 1998; Redenbach et al., 2000; Bentley et al., 2002; Goshi et al., 2002; Ikeda et al., 2003; Bentley & Parkhill, 2004; McLeod et al., 2006; Ohnishi et al., 2008). In fact, the frequency of linear chromosomes within the Actinomycetales is high compared with all other orders in the kingdom Bacteria. What evolutionary factors lead to a linear vs. a circular chromosome remain open to debate (Chen, 1996; Chen et al., 2002; Qin & Cohen, 2002), but it is important to realize that linearity vs. circularity does not seem to affect chromosome structure dramatically. Here, we will examine the chromosome diversity and similarity of the Actinomycetales, as displayed by the complete chromosome sequences available, and suggest that changes vary across the chromosome (Ventura et al., 2007; Hsaio & Kirby, 2008; Kirby et al., 2008). As the number of chromosome sequences available for the Actinomycetales increases and the genera from which they come broadens, it becomes important to try and understand how chromosome evolution in this order has occurred and is occurring. This is not least because over 80% of the world's antibiotics originally were identified as being produced by a member of the Actinomycetales (Hopwood, 2006).

Chromosome linearity and circularity

The majority of prokaryote chromosomes are believed to be circular. However, it can also be stated that biochemical proof of the circularity of many of these chromosomes is lacking and that they are circular by default. This remains true for the Actinobacteria and the Actinomycetales. In the case of the actinobacterial chromosome sequences in particular, evidence for the existence of linearity is based on one or more of three factors. First, direct isolation and analysis of the end of the linear chromosome with its covalently attached terminal protein by biochemical means is definitive (Lin et al., 1993; Goshi et al., 2002). Secondly, an analysis of the gene topology by pulsed-field gel electrophoresis (PFGE) is highly suggestive (Rednenbach et al, 2000). Finally, identification of genes associated with chromosome linearity, such as tpg (gene encoding the terminal protein that is covalently linked to the end of the linear chromosome), tap (gene encoding a telomere-associated protein that seems to be essential to linear chromosome replication and is usually closely linked with tpg on the chromosome) and ttr (gene encoding a protein that is present very close to ends of most linear chromosomes and seems to be involved in linear genome mobilization), implies linearity is present or was present at some point in the past (Goshi et al., 2002; Huang et al., 2007; Suzuki et al., 2008; Kirby & Chen, 2011). However, the absence of homologues of one or all of the tpg, tap and ttr trio does not confirm circularity because there is significant diversity in the terminal replication mechanism of linear chromosomes and plasmids of Actinomycetales (Huang et al., 2007; Suzuki et al., 2008). The problems of defining linearity other than by definitive biochemical means, which is laborious, can be illustrated in a number of ways.

Using PFGE, Saccharopolyspora erythraea NRRL 2338 was suggested to be linear based on analysis of the absence and presence of chromosome bands before and after proteinase K treatment (Reeves et al., 1998). However, by chromosome sequencing, Oliynyk et al. (2007) indicated that the chromosome of this species is circular. Analysis at the gene level of the chromosome sequence does not identify any homologues of the tpg, tap and ttr trio or the presence of terminal repeats, which supports the latter conclusion. Notwithstanding the missed restriction sites pinpointed by the chromosome sequencing, the entry of the 8Mb chromosome into the PFGE gel after proteinase K digestion, and the failure of the untreated chromosome to enter the gel under identical circumstances, supports directly the presence of bound terminal protein at the ends of a linear chromosome. Furthermore, Oliynyk et al. (2007) provide indirect evidence to support circularity, for example on the basis of the detection by gel electrophoresis of a fragment overlapping both proposed termini of the linear chromosome. The question remains somewhat open, but perhaps biased towards circularity. In the case of other Actinomycetales chromosome sequences, there is even less evidence to support circularity. Fragments having covalently linked protein at one end will be difficult to clone and the terminal repeats at the end of a linear chromosome can easily be closed to give a circular sequence by automatic contig computer software analysis if they are long enough (Huang et al., 2007).

Secondly, PFGE experiments have suggested that Actinoplanes philippinensis, Amycolatopsis orientalis, Micromonospora chalcea, Nocardia asteroides, Rhodococcus opacus and Streptoverticillium abikoense have linear chromosomes (Redenbach et al., 2000). Linearity is supported by sequencing in the case of R. opacus (http://www.expasy.ch/sprot/hamap/RHOOB.html) and Rhodococcus jostii (McLeod et al., 2006), whereas Rhodococcus erythropolis (http://www.expasy.ch/sprot/hamap/RHOE4.html), Amycolatopsis mediterranei (Zhao et al., 2010), Nocardia farcinica (Ishikawa et al., 2004) and many other species are described as circular based on chromosome sequencing. These findings indicate that chromosome linearity in the Actinomycetales is not limited to the Streptomyces, as was suggested might be the case by Oliynyk et al. (2007), but that there is heterogeneity in some genera; this includes Rhodococcus and Nocardia at least and perhaps many other genera.

Thirdly, if the available information on the chromosome sequences of Actinomycetales is examined (Table 1), a number of trends can be identified, even though many of the sequences are not fully annotated. Most Streptomyces have homologues of tpg, tap and ttr, which are genes directly or indirectly associated with chromosomal linearity (Bey et al., 2000; Bao & Cohen, 2001, 2003; Yang et al., 2002). This implies that the chromosomes with these genes are linear or have been linear in the recent past. Circularization of linear Streptomyces chromosomes is a relatively common occurrence in the laboratory and is effectively nonreversible, except possibly if another linear plasmid or another linear chromosome becomes involved (Volff et al., 1997). The absence of recognized terminal repeat sequences in the unpublished Streptomyces chromosomes is not unexpected, as a special approach is needed to obtain the sequences at the ends of the linear chromosome due to the presence of the covalently bound terminal protein that inhibits cloning. Furthermore, even in the absence of a cloning step, whole genome sequencing by the Roche 454 sequencing system will not obtain sequences from fragments that are covalently bound to a protein or peptide. It is expected that if and when these sequences are completely finished, most if not all will have recognized terminal repeats to which the Tpg protein will be covalently attached.

View this table:
1

Details of various actinomycete genomes

OrganismLinearCoding sequences (ORFs)Size (Mb)% GCtpgtapttrTerminal repeatsGenBank number
Streptomyces avermitilisYes75809.0370.7+++YesBA000030
S. bingchenggensis BCW-1Yes1002511.9370.8+++YesCP002047
S. coelicolorYes77698.6772.2+++YesAL645882
S. griseusYes71368.5572.2??YesAP009493
S. scabieiYes910710.1571.4+++YesFN554889
S. albusNE60266.6273.1ABYC00000000
S. clavuligerusLinear?64486.7371.1+++ABJH00000000
S. ghanaensisLinear?77508.2272.2+++ABYA00000000
S. griseoflavusLinear?71257.3671.7++ACFA00000000
S. hygroscopicusLinear?949810.4771.2++ACEX00000000
S. lividansYes75478.1972.2+++ACEY00000000
S. pristinaespiralisLinear?57416.1570.4++ABJI00000000
S. roseosporusLinear?70567.7671.4++ABYB00000000
Streptomyces sp. CNE76727.9272.6ACEW00000000
Streptomyces sp. Mg1Linear?69757.1171.2+++ABJF00000000
Streptomyces sp. SPB74Linear?49335.0171.9++ABJG00000000
S. sviceusNE82658.4770.1ABJJ00000000
S. viridochromogenesLinear?77398.5571.1++ACEZ00000000
Amycolatopsis mediterranei U32No936610.2470.9NoCP002000
Frankia alni ACN14aNo67117.5072.8NoCT573213
Frankia sp. CcI3No44995.4370.1NoCP000249
Frankia sp. EAN1pecNo71918.9871.2NoCP000820
Kitasatospora setae NBRC 14216Yes75698.7874.2+++YesAP010968
Micromonospora aurantiacaNo62227.0372.0NoCP002162
Rhodococcus equiNo45255.0468.7NoFN563149
Rhodococcus erythropolis PR4No60306.5262.3NoAP008957
R. jostii RHA1Yes72117.8067.0YesCP002162
R. opacus B4Yes72467.9167.9YesAP011115
Saccharomonospora viridisNo38384.3167.3NoCP001683
Saccharopolyspora erythraea?71978.2171.0NoAM420293
Salinispora arenicolaNo49175.7969.5NoCP000850
Salinispora tropicaNo45365.1869.5NoCP000667
Streptosporangium roseumNo943010.1270.8NoCP001814
  • * Yes, molecular biological evidence that strongly supports a linear chromosome; NE, no evidence to support either a linear genome or a circular genome; Linear?, no strong molecular biological evidence for a linear genome but a linear genome is suggested by other evidence; No, genome is accepted as circular in the literature; ?, some evidence suggests that both circular and linear genomes may occur in this species, although a circular chromosome is the accepted structure for the whole published genome sequence.

  • +, Homologue present by blast analysis >1e-10; ?, potential homologues are atypical and highly divergent in terms of protein sequence from those found in many Streptomyces; –, no homologue can be identified.

  • Yes, terminal repeats have been found. No, no terminal repeats have been found; –, no evidence of terminal repeats but genome sequencing is incomplete.

  • § Genome linearity is supported in Ohnishi et al. (2008) but the terminal repeat sequences and the possible tpg and tap homologues are very different from other Streptomyces. There is also an absence of a ttr homologue.

The exceptions within the Streptomyces that lack tpg and tap are Streptomyces albus, Streptomyces sp. C and Streptomyces sviceus (Table 1). There are three possibilities with these strains: (1) the sequencing is incomplete, particularly in the terminal regions where the tpg, tap and ttr genes generally reside; (2) these chromosomes are circular and therefore tpg, tap and ttr are absent; or (3) the tpg and tap homologues are highly divergent from the typical proteins encoded by these genes in most Streptomyces. The third option is not unlikely because two atypical tpg and tap gene pairs have already been found in Streptomyces griseus and SCP1 (Goshi et al., 2002; Bentley et al., 2004; Huang et al., 2007). Besides the Streptomyces, only R. jostii and R. opacus have been found to be linear chromosomes by sequencing (McLeod et al., 2006; http://www.expasy.ch/sprot/hamap/RHOOB.html). Both species have similar terminal repeats that are distinct from those of the typical Streptomyces and from S. griseus or SCP1. Neither has an identified tpg or tap gene on the linear chromosome, although tentative tpg homologues have been identified on plasmids in these species. In Letek et al. (2010), which describes the circular chromosome of pathogenic Rhodococcus equi, it is suggested that chromosome topology is not correlated with phylogeny among the rhodococci and is related rather to genome size. This agrees with the ideas being put forward here whereby linearization events via linear plasmids can produce new linear genomes again and again.

Based on the above considerations, it seems that linear chromosomes are not confined to the Streptomyces and that pinpointing linear chromosomes may be quite difficult unless special care is taken with genome sequencing because a significant terminal repeat sequence could easily result in the assembly of a circular chromosome if misinterpreted. Furthermore, the difficulty identifying tpg and tap homologues in chromosomes that are distant from the Streptomyces, or even within the genus Streptomyces if they are atypical, means that the apparent absence of these linear replication genes does not necessarily imply a circular chromosome. Nonetheless, there is a lack of a clear phylogenetic relationship between the Actinomycetales clade structure and the presence of linear chromosomes. This supports the hypothesis that linear chromosomes are a late development and that their origin within the Actinomycetales has probably occurred multiple times, even within the genus Streptomyces. That there has been more than one linearization event within the Streptomyces is supported by two findings. First, the arms outside of the syntenous central chromosome regions of certain Streptomyces are asymmetric, Streptomyces avermitilis being one example (Fig. 2). This asymmetry could arise in two ways: by uneven extension of the arms by gene addition or through the creation of a new linear chromosome by insertion of a linear plasmid at a site distinct from that of previous linear chromosomes. Secondly, both a standard type of terminal repeat structure as seen in Streptomyces coelicolor and many other Streptomyces, which may represent the original linearization, and a novel terminal repeat structure such as that of S. griseus, which may represent a more recent linearization events by a novel plasmid, are present. The presence of both types of linear terminal structure supports the idea that a linear chromosome may be advantageous when the chromosome is large and has a high G+C content.

2

mauve (http://asap.ahabs.wisc.edu/software/) comparison of Streptomyces coelicolor and Streptomyces avermitilis. The coloured blocks represent syntenous regions. The asymmetric nonsyntenous terminal regions are marked as left and right nonsyntenous arms (LNS arm and RNS arm). Note that the reverse complement of the S. avermitilis sequence was used for this analysis.

The question as to why a linear chromosome might be advantageous to members of Actinomycetales, particularly Streptomyces and Rhodococcus species, remains open. In this context, it would seem that genome linearity is associated with one obvious factor – chromosome size. Although not an absolute relationship, the linear chromosomes and the potentially linear chromosomes are generally larger than 7Mb in size, whereas many circular chromosomes in the Actinomycetales are smaller than 6Mb. For example, the chromosome of Kitasatospora setae, a member of a genus closely related to the Streptomyces, is linear, based on its chromosome sequence and has a genome size of 8.78Mb (Ichikawa et al., 2010). Further, the genome sizes of the linear chromosome of Rhodococcus spp. are 7.80Mb (R. jostii) and 7.91Mb (R. opacus). The circular genome R. erythropolis has a genome size of 6.52Mb. Two exceptions stand out, S. erythraea at 8.21Mb and Streptosporangium roseum at 10.12Mb. As indicated earlier, some strains of the former may be linear and the chromosome sequence of the latter is not complete. If a large chromosome size is associated with linearity, two possible hypotheses for a selective advantage can be proposed. First, the modular structure of the linear chromosome with a central core region, with regions on either side of this containing genes associated with being a highly complex organism that undergoes complex morphogenetic changes and then two terminal regions that seem to be completely unique to each species, may lend itself to easily increasing in size without disrupting essential functions found in the central core. Alternatively, on genetic transfer, linear chromosomes may generally be able to eliminate circular chromosomes by recombination, which in a myceliate organism would be highly advantageous to the linear chromosome.

Chromosome structure in the Actinomycetales

Figure 1 shows the alignment of various complete actinomycete chromosomes against the chromosome of S. coelicolor as a standard. It is immediately noticeable, with the exception of the outgroup Bifidobacterium longum, which is not a member of the Actinomycetales but an Actinobacteria, that there is significant similarity and synteny across most of the species analyzed. This gene conservation is mostly concentrated in the centre of the chromosomes and corresponds to the previously identified core region of the Streptomyces (Bentley et al., 2002; Hsiao & Kirby, 2008). The similarity in the core region has been supported broadly by many chromosome sequences, including those not present in Fig. 1, such as A. mediterranei (Zhao et al., 2010) and K. setae (Ichikawa et al., 2010).

The core region contrasts clearly with the terminal regions of the chromosomes, where very little similarity or gene conservation can be found in any of the Actinomycetales investigated. There are a few exceptions to the lack of gene conservation in the terminal regions, the three most apparent being Frankia, Nocardia and Rhodococcus at the left ends of the chromosomes (Fig. 1), but even these regions are relatively small. The three regions contain many hypothetical and conserved hypothetical genes as well as genes encoding a number of σ factors, antibiotic biosynthetic clusters and other secondary metabolic genes, such as chitinases. Notwithstanding these gene similarities, there is no obvious evolutionary basis for gene conservation between these species and S. coelicolor in the 7900000–8400000-bp region of the latter's chromosome to the right of the chromosomes in Fig. 1.

Between the terminal regions and the core region there are two other distinct regions, one to the left and one to the right of the core region. In Fig. 3, where the chromosomes of Streptomyces are compared in a similar manner to those of the Actinomycetales in Fig. 1, it can be seen that these two regions are conserved, perhaps even to a higher degree than the core region, especially the one on the left. Originally these were suggested to be regions of the chromosome found only in members of the genus Streptomyces, based on the synteny of the core region with various Actinobacteria such as Mycobacterium and Corynebacterium. Those species show no or very limited morphological development and have very little gene similarity outside of the core region of the Streptomyces chromosome. However, when Fig. 3 is compared with Fig. 1 it is clear that the left and right regions between the terminal regions and the core region are distinct. The left regions, here termed the left Actinomycetales-specific region, seems to be more highly conserved in the Streptomyces compared with the right region and this syntenous conservation is also present in many Actinomycetales to a significant degree. This contrasts with the right region, termed the right Streptomyces-specific region in Figs 1 and 3. This region is quite well conserved in Streptomyces, but is rather more poorly conserved in Actinomycetales. These regions are supported by Fig. 4, where the five regions are compared in terms of gene conservation using DNA/DNA comparative microarray analysis against S. coelicolor across a number of Streptomyces and non-Streptomyces Actinomycetales species. The left terminal region shows the highest divergence across both Streptomyces and non-Streptomyces, in contrast to the left Actinomycetales-specific region, which shows consistently low divergence across all Actinomycetales. The core region shows higher divergence than the left Actinomycetales region, possibly due to the horizontally transferred regions that are present within this region (Jayapal et al., 2007). The right terminal region shows a trend towards higher divergence, although not to the same extent as the left terminal region, suggesting that the two terminal regions are quite distinct. The right Streptomyces-specific region shows significant variability in terms of divergence, which supports the concept that, in both the Streptomyces and non-Streptomyces species, this region contains species-specific genes. In some species, the two right-hand regions are distinct, whereas in others they are less so.

3

The genomes of various Streptomyces were aligned with the 8.67Mb genome of Streptomyces coelicolor (not shown) using mauve with identical settings (http://asap.ahabs.wisc.edu/software/). The coloured blocks represent syntenous regions. These were then compared with a 16S phylogenetic tree of the same species. The outgroup for the 16S analysis was Frankia alni. At the top of the alignment are indicated the known horizontally transferred regions within the S. coelicolor genome.

4

The genomes of various Streptomyces and non-Streptomyces Actinomycetales were compared using DNA/DNA microarray comparative genomics against the Streptomyces coelicolor genome. Genes with hybridization >2SD lower than the genome mean were identified as highly divergent and very likely to have no homologue. The frequency of such genes was compared across the five regions of the Streptomyces chromosome as defined in Fig. 3.

The above results thus suggest distinct evolutionary origins for the left Actinomycetales-specific region and the right Streptomyces-specific region. One possibility is that the left actinomycete-specific region is an early evolutionary acquisition to the core chromosome found in the simple Actinomycetales, whereas the right Streptomyces-specific region is a later addition to the already expanded chromosome that occurred when the Streptomyces began to evolve as a distinct clade. The diversity of the latter region may represent the diversity across the Streptomyces, whereas the greater similarity of the former region within the Streptomyces and the Actinomycetales may be associated with what makes a sporoactinomycete different from the simple Actinomycetales such as Mycobacterium and Corynebacterium.

Gao et al. (2006) published a list of signature proteins that are distinctive characteristics of Actinobacteria as a class. In Table 2, these signature proteins are presented with their homologues from S. coelicolor (only five do not have such a homologue) together with the positions of the Streptomyces genes in terms of the above five regions with the linear chromosome of S. coelicolor. All except two of these actinobacterial signature proteins (SCO0908 and SCO6030) are found either in the core region or the Actinomycetales-specific region and are absent from the two terminal regions and the Streptomyces-specific region. This supports the proposed regional nature of the Streptomyces chromosome and adds weight to the hypothesis that the Actinomycetales-specific region has an earlier evolutionary origin than the Streptomyces-specific region and, obviously, the two terminal regions. SCO0908 is very close to the boundary of the left terminal region, which might suggest that the present position of its boundary, as shown in Fig. 3, which is defined by the left edge of HTR GI-1, should be moved to about SCO0900, as defined by the left-most block of S. avermitilis homologues (Fig. 3). Similarly, SCO6030 is very close to the boundary between the core region and the Streptomyces-specific region, which might support a similar minor change in this boundary to the left edge of HTR Gi-16.

View this table:
2

Distribution of actinobacterial signature proteins as identified by Gao et al. (2006) across the genome of Streptomyces coelicolor

Streptomyces coelicolor geneSignature gene numberStreptomyces coelicolor regions as indicated in Figs 1 and 3
SCO0908Tfu_0365Left terminal region
SCO1084ML2156Left Actinomycetales-specific region
SCO1372Lxx16410Left Actinomycetales-specific region
SCO1383ML2075Left Actinomycetales-specific region
SCO1381ML2073Left Actinomycetales-specific region
SCO1421ML1439Left Actinomycetales-specific region
SCO1480ML0540Left Actinomycetales-specific region
SCO1653ML1312Left Actinomycetales-specific region
SCO1662ML1306Left Actinomycetales-specific region
SCO1664ML1300Left Actinomycetales-specific region
SCO1665ML1299Left Actinomycetales-specific region
SCO1929ML0589Left Actinomycetales-specific region
SCO1938ML0580Left Actinomycetales-specific region
SCO1997ML1009Left Actinomycetales-specific region
SCO2078ML0921Left Actinomycetales-specific region
SCO2097ML0904Left Actinomycetales-specific region
SCO2105ML0898Left Actinomycetales-specific region
SCO2153ML2446Left Actinomycetales-specific region
SCO2154ML2154Left Actinomycetales-specific region
SCO2169ML0869Left Actinomycetales-specific region
SCO2196ML0857Left Actinomycetales-specific region
SCO2197Lxx10090Left Actinomycetales-specific region
SCO2391ML1781Core region
SCO2460Tfu_1340Core region
SCO2557Lxx08190Core region
SCO2634ML1485Core region
SCO2893ML0169Core region
SCO2915ML1166Core region
SCO2916ML1165Core region
SCO2947ML1016Core region
SCO3001TFu_2483Core region
SCO3011ML0775Core region
SCO3016Tfu_2498Core region
SCO3030ML0762Core region
SCO3034ML0760Core region
SCO3094ML0257Core region
SCO3095ML0256Core region
SCO3097ML2030Core region
SCO3327ML2428ACore region
SCO3349ML2435Core region
SCO3375ML0234Core region
SCO3576Lxx03620Core region
SCO3547ML0284Core region
SCO3822ML0115Core region
SCO3854ML00133Core region
SCO3872ML0007Core region
SCO3888ML2705Core region
SCO3902ML2687Core region
SCO4043Tfu_0030Core region
SCO4084ML2207Core region
SCO4088ML2204Core region
SCO4197ML2200Core region
SCO4205ML2442Core region
SCO4287ML1927Core region
SCO4579ML2064Core region
SCO4590Tfu_1240Core region
SCO5167Tfu_0515Core region
SCO5173ML0816Core region
SCO5199ML0642Core region
SCO5240ML0804Core region
SCO5145ML1067Core region
SCO5414ML1176Core region
SCO5493ML1706Core region
SCO5601ML1610Core region
SCO5697Tfu_0751Core region
SCO5721ML1544Core region
SCO5766ML0986Core region
SCO5829ML1029Core region
SCO5855ML2137Core region
SCO5864ML1026Core region
SCO5866ML1027Core region
SCO6030ML1041Right Streptomyces-specific region
  • Five signature proteins from Gao et al. (2006) do not have homologues in S. coelicolor: ML0561, ML0591, ML0899, ML2473 and ML2570.

Interestingly, in general, the overall chromosome similarity by mauve [multiple alignment of conserved genomic sequence with rearrangements; a software package that attempts to align orthologous and xenologous regions among two or more genome sequences that have undergone both local and large-scale changes (http://asap.ahabs.wisc.edu/software/)] conforms to the 16S phylogeny at a gross level (Figs 1 and 3). This is further supported by the close similarity of Streptomyces lividans (http://www.ncbi.nlm.nih.gov/nuccore?Db=genomeprj&DbFrom=nuccore&Cmd=Link&LinkName=nuccore_genomeprj&IdsFromResult=224184466) chromosome sequence to that of S. coelicolor, which are sister species (Fig. 3). This suggests that these two regions may as a whole and in their gene complement represent the chromosome gain steps and evolutionary branch points that have resulted in distinct genera. Thus the core region contains the original basic gene structure of the Actinomycetales and also other members of the Actinobacteria. The left Actinomycetales-specific region may contain the genes needed to be a specific genus with the Actinomycetales, whereas the right Streptomyces-specific region defines members of the genus Streptomyces. Finally, the two terminal regions contain many of the genes that are species specific within the Streptomyces.

This is a simplification, and horizontal transfer of regions in all species, which are shown in Fig. 3 (top) specifically for S. coelicolor, is also undoubtedly important in defining each species. Nonetheless, the above analysis suggests that specific exploration of the two regions immediately to the right and left of the core chromosome may help identify genes and gene groups that are important to specific genera and also help us understand how the Actinobacteria evolved from unicellular nondifferentiating Gram-positive organisms into multicellular filamentous organisms that undergo complex differentiation.

Unfortunately, the above analysis does little to help answer the question posed earlier, namely, what drives chromosome linearity in the Actinomycetales and Streptomyces. Most of the chromosomes shown in Fig. 1 and Table 1 are circular. Those with some evidence of one or another type of linearity are indicated. This contrasts with Fig. 3, where all of the chromosomes probably should be regarded as linear. If there is an exception it is S. albus, which has the smallest chromosome size and where no homologues of tpg, tap or ttr have been identified. However, there are two trends that might help us. The first is that the potentially linear chromosomes cluster around the Streptomyces, which suggests that the chromosome linearity has only evolved a few times. In other words, the functional mechanisms that allow a linear chromosome to exist have only evolved on rare occasions. This does not mean that the change from a circular to a linear chromosome is a rare event. Once a mechanism for linear replication has evolved and exists on plasmids and chromosomes, then linearization is only one recombination event away (Chen, 1996; Chen et al., 2002). This is simply because when a single homologous or nonhomologous recombination event occurs between a linear replicon and a circular replication, the resulting molecule is always linear. Thus a small linear plasmid can linearize a large circular genome while retaining the machinery for linear terminal replication. Linear plasmids are common in the Actinomycetales and thus, as mentioned earlier, linearization of circular Streptomyces chromosomes seems to occur regularly. Chromosome arm asymmetry in the Streptomyces supports this. Specifically, the chromosome origin of replication in genome-sequenced Streptomyces moves significantly with respect to its position relative to the ends of the linear genome (see S. avermitilis and S. coelicolor). The second trend is that the groups with potentially linear chromosomes generally have chromosomes of a larger size, most being larger than 6.5Mb. This suggests that if you need to increase your chromosome size evolutionarily, linearity may be an advantage.

Footnotes

  • Editor: Simon Silver

References

View Abstract