Correspondence: Haruyasu Kinashi, Department of Molecular Biotechnology, Graduate School of Advanced Sciences of Matter, Hiroshima University, 1-3-1 Kagamiyama, Higashi-Hiroshima 739-8530, Japan. Tel./fax: +81 82 428 7579; e-mail: firstname.lastname@example.org
Streptomyces linear chromosomes frequently cause deletions at both ends spontaneously or by various mutagenic treatments, leading to chromosomal circularization and arm replacement. However, chromosomal circularization has not been confirmed at a sequence level in the model species, Streptomyces coelicolor A3(2). In this work, we have cloned and sequenced a fusion junction of a circularized chromosome in an S. coelicolor A3(2) mutant and found a 6-bp overlap between the left and right deletion ends. This result shows that chromosomal circularization occurred by nonhomologous recombination of the deletion ends in this species, too. At the end of the study, we discuss on stability and evolution of Streptomyces chromosomes.
The saprophytic filamentous soil bacteria, from the genus of Streptomyces, are well known to produce many clinically useful antibiotics. It is also a characteristic feature of this genus to contain an 8–9 Mb linear chromosome in place of a usual circular bacterial chromosome (Bentley et al., 2002; Ikeda et al., 2003; Ohnishi et al., 2008). The linearity of Streptomyces chromosomes was first proved for Streptomyces lividans (Lin et al., 1993), and the complete linear genome sequence of the model species, Streptomyces coelicolor A3(2), was determined in 2002 (Bentley et al., 2002)
Streptomyces chromosomes are unusually unstable and are often subject to deletion and amplification spontaneously or by various mutagenic treatments (Volff & Altenbuchner, 1998). The size of chromosomal deletions reaches up to 2 Mb in some Streptomyces species (Fischer et al., 1997; Chen et al., 2010). Studies of this genetic instability of Streptomyces faced great difficulty previously, because Streptomyces chromosomes had been considered to be circular for a long time. We now have a correct idea of the linear structure of Streptomyces chromosomes and a powerful method, pulsed-field gel electrophoresis, for physical analysis of their rearrangements. We know at present that chromosomal deletions occur from both ends. However, instability of Streptomyces chromosomes has not been clarified well. For example, an amplifiable unit of DNA (AUD) was tandemly amplified several hundred times to form amplified DNA sequence (ADS) in S. lividans mutants (Altenbuchner & Cullum, 1985), but the gross structures of the mutant chromosomes have not been clarified.
The most frequent destination of Streptomyces chromosomes following terminal deletion is circularization. Chromosomal circularization was indicated by detection of a macrorestriction fragment in deletion mutants of S. lividans (Lin et al., 1993; Redenbach et al., 1993) and S. ambofaciens (Leblond et al., 1996). It was finally confirmed in Streptomyces griseus by cloning and sequencing of the fusion junctions of the circularized chromosomes (Kameoka et al., 1999; Inoue et al., 2003). No sequence homology was found between the left and right deletion ends in two mutants, and only 1-bp and 6-bp homology was found in two other mutants. Accordingly, it was proposed that nonhomologous recombination between the left and right deletion ends caused chromosomal circularization (Inoue et al., 2003). Microhomology was also detected at the fusion points of circularized chromosomes of Streptomyces avermilitis mutants (Chen et al., 2010).
Chromosomal arm replacement is another destination of deleted chromosomes. When one chromosomal arm is deleted and the left and right arms carry a homologous sequence in an inverted orientation, homologous recombination between them causes arm replacement, which recovers a telomere and generates longer terminal inverted repeats (TIRs) at both ends. This phenomenon was first reported for S. ambofaciens (Fischer et al., 1998) and was followed by S. griseus (Uchida et al., 2003) and S. coelicolor A3(2) (Widenbrant et al., 2007). Even the long TIRs formed by arm replacement suffer terminal deletion. When an inverted repeat sequence is present at the deletion end inside the long TIR, it could form a hairpin structure, which invades the opposite TIR strand during replication leading to a circular chromosome with an extremely large palindrome (Uchida et al., 2004). Similar various rearrangements were observed in Streptomyces linear plasmids when deletions were introduced at specific locations within telomeres (Qin & Cohen, 2002).
In spite of these extensive analyses, chromosomal circularization has not been proved for the model species S. coelicolor A3(2) at a sequence level. In this study, we first report the cloning and sequencing of a fusion junction of a circularized chromosome of S. coelicolor A3(2) and show that chromosomal circularization occurred by nonhomologous recombination in this species, too. At the end of the study, we discuss on stability and evolution of Streptomyces chromosomes.
Materials and methods
Bacterial strains, plasmids, cosmid libraries, and medium
Streptomyces coelicolor A3(2) strain No. 4 used in this study is an eshA (named for a defect of extension of sporogenic hyphae; Kwak et al., 2001) mutant obtained by cultivation of the wild-type strain 1147 at a high temperature (Kawamoto et al., 2001). Strain M145 for which the genome project has been carried out (Bentley et al., 2002) was used as a reference strain for comparison. The cosmid libraries of S. coelicolor A3(2) used in this study have been constructed and ordered by Redenbach et al. (1996) and Zhou et al. (2012). Escherichia coli XL1-Blue and pUC19 were used for cloning and sequencing of DNA fragments. Plasmid pLUS221 used as a probe for hybridization carries the 1.3-kb BamHI fragment at the extreme end of the S. coelicolor A3(2) chromosome (Huang et al., 1998). Glucose-meat extract-peptone (GMP) medium contains 10 g of glucose, 4 g of peptone, 2 g of meat extract, 2 g of yeast extract, 5 g of NaCl, and 0.25 g of MgSO47H2O L−1 (pH 7.0).
DNA manipulation and Southern hybridization
Streptomyces coelicolor A3(2) wild-type and mutant strains were reciprocally grown in liquid GMP medium in Sakaguchi flasks at 28 °C for 3 days. DNA manipulation for Streptomyces (Kieser et al., 2000) and E. coli (Sambrook et al., 1989) was carried out according to standard procedures. Total DNA was digested with restriction enzymes, separated by conventional agarose gel electrophoresis, and transferred to nylon membrane filters by the capillary method. DNA probes were labeled with digoxigenin-11-dUTP (Roche Diagnostics, Mannheim, Germany) using random primers, and hybridization was carried out overnight at 70 °C according to the manufacturer's protocol. After hybridization, washing was carried out twice for 5 min each in 2× wash solution at room temperature, and then twice for 15 min each in 0.1× wash solution at 70 °C.
PCR and nucleotide sequencing
Two primers for PCR amplification, del-L, 5′-CACCGAATTCTGAGCGATGGTCGTCGTGA-3′ (the EcoRI site is underlined) and del-R, 5′-ATACGGATCCTTCGCGATCGTCCCGCTGA-3′ (the BamHI site is underlined), were designed based on Southern hybridization analysis of the left and right deletion ends of mutant No. 4. PCR was performed on a 2720 Thermal Cycler (Applied Biosystems, Foster city, CA) with KOD-Plus DNA polymerase (Toyobo, Osaka, Japan). Nucleotide sequencing was performed by the dideoxy termination method, using BigDye Terminator v3.1/v1.1 Cycle Sequencing Kits (Life Technologies, Carlsbad, CA) and a 3130xl Genetic Analyzer (Life Technologies).
Results and discussion
Analysis of chromosomal deletion in mutant No. 4
Streptomyces coelicolor A3(2) strain No. 4 used in this study is an eshA mutant of the wild-type strain 1147 and shows several defective phenotypes (Kawamoto et al., 2001). The eshA gene encodes a nucleotide-binding protein, disruption of which caused a loss of actinorhodin production due to a reduced level of ppGpp (Saito et al., 2006). The eshA gene was identified as SCO7699 (nt 8535532–8536947 of the genome sequence) located at 131 kb from the right end (Bentley et al., 2002), suggesting that at least the right chromosomal arm was deleted beyond this locus. To clarify whether only the right telomere was deleted or both telomeres were, Southern blot analysis was carried out using labeled pLUS221, which contains the 1.3-kb end BamHI fragment of the chromosome (Huang et al., 1998). As shown in Fig. 1a, the reference strain M145 showed a positive signal at 1.3 kb, whereas no hybridizing signal was observed for mutant No. 4. As the S. coelicolor A3(2) chromosome has the 24-kb TIRs at the left and right ends, this result indicates that both telomeres were deleted in mutant No. 4.
To determine the range of deletions in the left and right arms, we used the original cosmid library constructed by Redenbach et al. (1996) as probe for hybridization, but could not identify the deletion-end cosmids. Therefore, we used another cosmid library recently constructed by Zhou et al. (2012) and finally identified the left deletion-end cosmid, 3–14, and the right deletion-end cosmid, 8–65. As shown in Fig. 1b, cosmid 3–14 (covers nt. 230,501-274,102 of the chromosome) and cosmid 8–64 (nt. 7,790,510-7,825,517) gave fewer hybridizing signals for mutant No. 4 compared with strain M145. Referring to the restriction maps of cosmids 3–14 and 8–64, the left and right deletion ends were delimited. It was revealed that the 4.2-kb BamHI fragment of cosmid 3–14 and the 13.6-kb BamHI fragment extending over the right end of cosmid 8–65 were fused to generate a 3.3-kb BamHI fragment (Figs 1b and 2a). Similarly, the 3.9-kb SphI fragment of cosmid 3–14 and the 6.2-kb SphI fragment of cosmid 8–65 were fused to generate a 4.2-kb SphI fragment. It should be noted that cosmids 3–14 and 8–65 hybridized to the same 3.3-kb BamHI fragments, and the hybridization intensity of the former was stronger than the latter. This result indicates that the 3–14 sequence in the 3.3-kb BamHI fragment is larger than the 8–65 sequence.
Based on these results, two primers, del-L and del-R, were synthesized and used for PCR amplification, which gave a 1.7-kb amplified fragment (data not shown). This fragment was digested with BamHI and EcoRI and cloned into pUC19 to give plasmid pOPP. As expected, the pOPP probe hybridized to both the left and right deletion-end fragments of strain M145 and to the fusion fragment of mutant No. 4 (Fig. 1c).
Sequence analysis of the fusion junction
Nucleotide sequencing of plasmid pOPP determined the fusion junction of the circularized chromosome of mutant No. 4. The sequences around the left and right deletion ends of strain M145 and the fusion junction of mutant No. 4 are aligned and compared in Fig. 2b. Between the left and right deletion ends, a 6-bp microhomology was identified, which is much shorter than the minimum size (20 bp) of homology required for homologous recombination (Watt et al., 1985). This result again supports nonhomologous recombination of two deletion ends, which was proposed for chromosomal circularization of other Streptomyces species. At the left deletion end, a putative sporulation control protein, SpoOM (SCO247), is encoded. On the other hand, at the right deletion end, a possible binding-protein-dependent transport protein (SCO7030) is encoded. The generated fusion gene encodes for a protein, in which due to frame coincidence, the N-terminal 16-aa of the SCO247 protein were replaced by the N-terminal 9-aa of SCO7030. As this protein carries most (325 aa/341 aa) of the SCO247 sequence, it may function as the SCO247 protein does in the parent strain 1147.
Mutant No. 4 lost a total of 1088 kb DNA (237 kb from the left end and 851 kb from the right end) during terminal deletion and circularization. Although mutant No. 4 shows several defective phenotypes, it grows normally (Kawamoto et al., 2001). Thus, many genes located in the deleted terminal regions are not essential for survival. As the eshA gene is located at 131 kb from the right end, it is possible that deletion of other genes rather than eshA caused some of the defective phenotypes found in mutant No. 4.
Circularized Streptomyces chromosomes are stably maintained
There have been contradictory issues on stability of circularized Streptomyces chromosomes. Lin & Chen (1997) and Volff et al. (1997) independently constructed artificially circularized chromosomes of S. lividans and studied their genetic instability. In both cases, the circularized chromosomes showed higher frequencies of deletion and amplification. Thus, they claimed that the circularized chromosomes were more unstable than the parent linear chromosomes. However, it should be noted that in both cases, the deleted sequences in the artificially circularized chromosomes were restricted to the left and right TIR regions of about 30 kb each. Therefore, deletable genes such as the chloramphenicol resistance gene (cmlR) and the arginine biosynthetic gene (argG) and an amplifiable sequence such as AUD were still retained. Fischer et al. (1997) also prepared circularized chromosomes of S. ambofaciens with extremely large deletions (more than 2 Mb) by mutagenic treatments and reported their genetic instability, too. In this case, it may be possible that the extremely large deletions eliminated regions important for stable maintenance of the circularized chromosome.
In contrast to the examples described above, our and other groups obtained stably maintained circular chromosomes by mutagenic treatments of S. griseus (Kameoka et al., 1999; Inoue et al., 2003), S. coelicolor A3(2) (this work), and S. avermitilis (Chen et al., 2010). In these cases, the sizes of deletions were 480 kb (130 kb at the left end + 350 kb at the right end), 580 kb (30 kb + 550 kb), and 300 kb (130 kb + 170 kb) for S. griseus, 1088 kb (237 kb + 851 kb) for S. coelicolor A3(2), and 1939 kb (1611 kb + 328 kb) for S. avermitilis, respectively. Thus, the deletable sizes seem to have some relation to the chromosomal size of each species: 7.8 Mb (calculated from AseI fragments) for S. griseus strain 2247 (Lezhava et al., 1995), 8668 kb for S. coelicolor A3(2) (Bentley et al., 2002), and 9026 kb for S. avermilitis (Ikeda et al., 2003). Larger chromosomes may contain larger terminal regions dispensable for survival (Kirby, 2011). We speculate that terminal deletions proceed progressively until to appropriate points, where circularized chromosomes reach a stable state and are stably maintained. Thus, mutant chromosomes with small deletions or amplification may be intermediates in this process, which are finally converted to more stable circular chromosomes.
Linear replicons always have a problem of terminal replication, namely incomplete replication of the 5′ ends. Streptomyces linear chromosomes solve this problem as follows. Linear chromosomes are replicated bidirectionally from an internal origin (Musialoski et al., 1994). This leaves single-strand overhangs at the 3′ ends (Chang & Cohen, 1994), which are filled by a novel patching synthesis primed by terminal protein (Qin & Cohen, 1998). It was suggested that Streptomyces linear chromosomes were generated by integration of a linear plasmid into a circular chromosome (Volff & Altenbuchner, 2000). If circularized chromosomes are stably maintained, why have Streptomyces kept a linear chromosome in the evolutionary history. The following advantages could be raised for linear chromosomes. (1) The genome sizes of Streptomyces are about two times larger than those of bacteria such as E. coli and Bacillus subtilis. Therefore, Streptomyces linear chromosomes could accommodate many genes that are indispensable for sophisticated morphological differentiation and adaptation to environmental changes of this soil-living genus. The size of circular chromosomes is limited, because large super-twisted circular replicons are difficult to be unwound and resolved to single strands during replication. (2) A single crossover with another linear replicon near the end of chromosome leads to end exchange (Pandza et al., 1998), by which Streptomyces linear chromosomes could obtain genes necessary for secondary metabolism and environmental adaptation. In addition, when a single crossover occurs near the center of chromosome, it could give two chimeric chromosomes. This event actually occurred in S. coelicolor A3(2) strain 2106 (Yamasaki & Kinashi, 2004), which was considered as a model of chromosomal multiplication.
Structural comparison of Streptomyces linear replicons and studies of their genetic instability will give us important hints how circular chromosomes have been converted to linear chromosomes in the evolutionary history (Volff & Altenbuchner, 2000; Chen et al., 2002; Kirby, 2011).
We thank Carton Chen for providing the telomere probe pLUS221.