Mutations in fts genes partially or completely block both vegetative cell division and sporulation septation in the filamentous bacterium Streptomyces coelicolor A3(2). Using a novel screen, we independently isolated two double-mutant strains, each containing a spontaneous suppressor mutation, which partially restores division to an ftsQ-null mutant. Genetic complementation experiments revealed that the suppressor mutations alone confer no observable defect in sporulation. The suppressor mutations were genetically mapped to regions of the chromosome, distinct from each other and the division and cell wall cluster containing ftsQ. Therefore, the genes identified by the suppressor mutations were named sqnA and sqnB (suppressor of ftsQ-null) and may be representatives of a novel class of genes involved in cell division or the regulation of cell division in this mycelial organism.
Streptomyces coelicolor A3(2) is a Gram-positive, filamentous soil bacterium, which normally grows as a mycelium. During vegetative growth, crosswalls are infrequent and widely spaced, resulting in multinucleoid hyphal compartments. As colonies mature, aseptate aerial hyphae grow away from the substrate mycelium surface. Upon growth completion, the aerial filaments are synchronously divided into evenly-proportioned, unigenomic cells that further metamorphose into spores . Thus, temporally and spatially distinct cell division events result in either sporadic vegetative crosswalls or regularly-spaced sporulation septa during the developmental cycle of this organism.
Cell division is essential for growth and viability in Escherichia coli, Bacillus subtilis and other organisms where it has been studied , but not in S. coelicolor, in which ftsZ- or ftsQ-deletion mutants, although defective in cell division, remain viable [3,4]. In an S. coelicolor ftsQ-null mutant, vegetative crosswalls are present in the substrate mycelium at an approximately 10-fold lower frequency than wild-type strains, and very few spores form in aerial filaments. Like ftsZ mutants, the ftsQ mutant colonies are surrounded by blue ‘halos’ on minimal glucose medium, due to the production of copious amounts of actinorhodin, the diffusible blue-pigmented antibiotic [3,4]. In contrast, wild-type colonies produce only light pigmentation in the adjacent agar, providing a convenient visual distinction between strains that are division proficient and those at least partially blocked for septation.
We independently isolated two derivatives of an ftsQ-null mutant that no longer exhibited the characteristic blue-halo phenotype of a division mutant. Each double mutant contains a suppressor mutation at a location distant from ftsQ, at least partially restoring cell division in aerial filaments. These novel genes may be involved directly in cell division or in the developmental regulation of the process.
2Materials and methods
2.1Bacterial strains and growth conditions
All bacterial strains used are listed in Table 1. E. coli strain TG1 was used for standard plasmid manipulation. To bypass the methyl-specific restriction system of S. coelicolor, the dam−dcm−E. coli strain ER2-1 (an F′ derivative of GM2163; ) was used for preparation of unmodified plasmid DNA.
Streptomyces strains were grown at 30°C. YEME (liquid), R2YE (agar) and minimal medium MM (agar), for growth of S. coelicolor, were as described previously [4,6]. Glucose was added to 0.5% (w/v) in MM. Final concentrations of antibiotics used for Streptomyces were streptomycin at 20 μg ml−1 in MM and 50 μg ml−1 in R2YE, spectinomycin at 100 μg ml−1 in MM and 200 μg ml−1 in R2YE, apramycin at 25 μg ml−1, and thiostrepton at 50 μg ml−1. E. coli strains were propagated in Luria broth. Final concentrations of antibiotics used for E. coli were ampicillin at 100 μg ml−1, carbenicillin at 100 μg ml−1, apramycin at 100 μg ml−1 and spectinomycin at 50 μg ml−1.
2.2DNA and transformation techniques
S. coelicolor chromosomal DNA was prepared according to Hopwood et al. . DNA restriction and modifying enzymes were used according to the manufacturer's recommendations. Plasmid DNA was prepared from E. coli using the QIAprep Spin Kit (Qiagen). The QIAquick Gel Extraction Kit (Qiagen) was used to purify DNA fragments fractionated by agarose gel electrophoresis. Standard procedures for transformation of S. coelicolor were used [4,6]. Double-stranded plasmid DNA was alkaline-denatured before introduction into S. coelicolor by transformation to stimulate homologous recombination .
2.3Construction of a plasmid for genetic complementation
pJR137, containing the wild-type ftsQ gene (Fig. 1), was constructed by adding the apramycin-resistance gene aac(3)IV as an Xba I fragment from pOJ427 (B. Schoner) to Xba I-digested pJR130 .
2.4Construction of genetic mapping strains
pJA2, a plasmid identical to pJR137, except that it contains a copy of the ftsQ-null allele (ΔftsQ::aadA) instead of wild-type ftsQ, was used to transform 2709 to spectinomycin resistance. DU56 was a spectinomycin-resistant, apramycin-sensitive 2709 transformant, in which the wild-type copy of ftsQ has been replaced with an ftsQ-null allele. HU41 was a thiostrepton-resistant, glucose-utilizing recombinant from a cross between 2709 and J1668Δ3.4. DU66 was isolated as a thiostrepton-resistant, prototrophic recombinant from a cross between HU202 and HU41.
S. coelicolor genetic crosses and statistical analyses were performed as described by Hopwood et al. , except that the mycelia from the crosses were harvested and mechanically macerated in saline (0.85% NaCl) and used immediately, or stored at −80°C after the addition of glycerol (final concentration of approximately 18%). Mycelia were harvested instead of spores, because the ftsQ-null mutants sporulate and plate less efficiently than wild-type. For genetic mapping, the blue-halo phenotype of cell division-defective strains on minimal glucose medium was used to distinguish suppressor-free from suppressor-containing strains. The aadA gene in ftsQ-null-containing strains does not confer resistance to streptomycin even at low concentrations in minimal glucose agar, allowing the use of the chromosomal strA1 marker for the selection of streptomycin-resistant recombinants in genetic crosses involving ΔftsQ::aadA.
Methods for microscopy of MM-grown colonies were essentially those of McCormick and Losick .
3Results and discussion
3.1Identification of two strains containing a suppressor mutation
Cell division gene ftsQ is located in a large cluster of genes (dcw) required for division and cell wall synthesis (Fig. 1). S. coelicolor strains deleted for ftsQ overproduce actinorhodin when grown on MM, resulting in colonies surrounded by blue halos (HU151, Fig. 2). HU201 (ΔftsQ::aadA sqn-1) and HU202 (ΔftsQ::aadA sqn-2) were independently isolated as spontaneous suppressor-containing strains (suppressor of ftsQ-null) from ftsQ-null mutant HU151, because they no longer displayed a blue-halo phenotype on MM (Fig. 2).
3.2Division and sporulation phenotype as observed by phase-contrast examination
Phase-contrast microscopy showed that sporulation septation was restored to the suppressor-containing strains (Fig. 3). Wild-type strain M145 sporulated well, producing long chains of oval spores by 3 days of growth on MM, while the ftsQ mutant HU151 produced long, straight aerial filaments (Fig. 3A,B), which rarely differentiated to form coils (such as those produced by certain developmental whi mutants) or spores .
Each suppressor-containing strain displayed an enhanced ability to sporulate in comparison to the ftsQ-null strain HU151. The majority of HU201 aerial filaments were delayed for sporulation by several days compared to wild-type strain M145, and most spore chains were much shorter than those of the wild-type strain (Fig. 3C). These shorter chains mainly consisted of very elongated spores, presumably generated by subdividing an aerial filament less frequently than normal. In contrast, HU202 displayed a sporulation (and division) phenotype that closely resembles wild-type strain M145 (Fig. 3D). The aerial hyphae of HU202 contained normal-sized spore compartments at the tips, but a higher than normal frequency of elongated spores at the stalk end of many aerial filaments.
3.3Phenotypes conferred by the suppressor mutations revealed by genetic complementation of the ftsQ-null allele
In order to determine the phenotypes of strains containing each suppressor mutation alone, we introduced a wild-type copy of ftsQ (ftsQ+) by transformation into the dcw cluster of the double-mutant strains, as well as into HU151 (ΔftsQ). A single homologous recombination event resulted in the integration of pJR137, a non-replicating plasmid containing ftsQ+, as well as other genes both upstream and downstream of ftsQ (Fig. 1). The new heterozygous merodiploid strains, containing a deletion of ftsQ complemented by a closely-linked copy of wild-type ftsQ, were compared to an equivalent homozygous wild-type construct using phase-contrast microscopy to determine if the suppressor mutation alone confers a defect in cell division.
Spore formation in aerial filaments of all of these new strains appeared visually identical to wild-type M145 containing no integrated plasmid (Fig. 3E–H). This implies that the gene products produced by the suppressor genes, in the complemented strains, do not impede the role of wild-type FtsQ. However, because uncomplemented HU202 sporulates at an efficiency which approaches that of wild-type strains, it is difficult to distinguish between HU202 and strains of HU202 complemented with ftsQ+, suggesting the possibility that in the presence of sqn-2, the role of FtsQ might be dispensable.
3.4Genetic mapping of suppressor mutation sqn-1
We crossed mapping strain DU56 with HU201/pIJ922 to genetically map sqn-1. An equal probability of recombination for all markers was expected for this cross, because it used an SCP2*-derived fertility plasmid [8,9]. However, markers for this cross did not segregate as well as in crosses involving division-proficient strains, and we could not unambiguously map sqn-1 even after testing 808 recombinants (data not shown).
To obtain a map position for sqn-1, we carried out a further cross that used NF (fertility plasmid SCP1 integrated at the nine o'clock position of the chromosome) to mobilize the chromosome. This type of cross should result in a much higher frequency of recombination . HU41, an NF mapping strain with wild-type ftsQ, was crossed with HU201, selecting for spectinomycin (ΔftsQ::aadA) and streptomycin resistance (strA1). Every recombinant, therefore, contained the ftsQ-null allele and the suppressor mutation or the ftsQ-null allele alone. A probable map location for sqn-1 was established, suggesting that the suppression phenotype is caused by a single mutation (Fig. 4). A modified contingency χ-square test showed that sqn-1 was closely linked to cysD, but not the other markers (Fig. 4). The gene identified by the sqn-1 mutation was called sqnA.
3.5Genetic mapping of suppressor mutation sqn-2
As was observed above, little segregation of markers occurred with an SCP2*-based cross (HU202 × DU56) despite testing 434 recombinants (data not shown). HU202 was next mated with NF mapping strain HU41, because this type of cross had been more successful for mapping sqn-1. The mapping data implicated linkage to both proA and argA when 318 recombinants were analyzed (Fig. 5). The data clearly eliminated the lower half of the chromosome as a potential map position for sqn-2, indicating that this mutation is distinct from sqnA. Allele frequencies placed sqn-2 near either argA or proA.
Few recombinants possessing the ftsQ-null mutation without sqn-2 were obtained in the crosses described above, making it difficult to determine linkage. Although the formal possibility existed that the suppressive phenotype observed for HU202 could be the result of more than one mutation in the upper arc of the chromosome, the simplest interpretation of the data is the presence of a single mutation either near proA or argA. Therefore, a further cross was performed using a NF version of HU202 named DU66. DU66 was crossed with ftsQ-null mapping strain DU56, selecting for markers outside of the upper arc of the chromosome (Fig. 5). This time a useful segregation of markers was obtained, and a higher proportion of recombinants possessed only the ftsQ-null mutation. Probability values for this cross favor a position for sqn-2 clockwise to argA (Fig. 5), though a position counter-clockwise to proA cannot be ruled out by the data. Presumably, the difficulty in obtaining unambiguous mapping data results from the fact that both the ftsQ-null and suppressor mutations are located in the upper arc of the chromosome, influencing the outcome of the cross. Nevertheless, note that the data clearly indicated that sqn-2 is distinct from sqnA, which maps near the cysD locus. It is also distinct from the ftsQ-null mutation and the other fts genes in the division and cell wall cluster (located near hisA), because it readily segregated from the ftsQ-null allele (Fig. 5). The gene identified by the sqn-2 mutation was therefore called sqnB.
3.6A strategy for cloning the new genes
The availability of an ordered overlapping cosmid library of the S. coelicolor genome (336 cosmids) provides a means for cloning the suppressor genes . Cosmids representing the regions identified by genetic mapping can be introduced by transformation into each suppressor-containing strain . Our genetic mapping data is sufficiently accurate to reduce the number of cosmids to be tested; pooling cosmids in the initial transformation experiments will further reduce the labor. Experiments to clone the sqn genes by transformation with pools of cosmids are underway. The mutants could be potentially interesting, providing novel insights for the role of FtsQ in septum formation.
In summary, we have utilized a novel screen to identify genes affecting cell division in a model mycelial organism. Our data clearly establish the utility of this approach. Furthermore, the fact that the two spontaneous mutants possess mutations at different loci suggests that saturation of this class of mutations might not have been attained. We have therefore recently isolated several additional (spontaneous and induced) suppressor mutants. Finally, the introduction of point mutations in ftsQ, created by site-directed mutagenesis, could lead to the isolation of different classes of suppressor mutations than those found for a deletion mutation.
The mutants described in this article were isolated in Richard Losick's laboratory. We thank Keith Chater for providing a strain, the Sanger Centre for providing S. coelicolor genome sequence data, and Keith Chater, Klas Flärdh, Richard Losick, William Margolin and Justin Nodwell for critical reading of this manuscript. This work was supported by a Grant to J.R.M. from the National Institutes of Health (GM 56915).