SsgA plays an important role in the control of sporulation-specific cell division and morphogenesis of streptomycetes, and ssgA null mutants have a rare conditionally non-sporulating phenotype. In this paper we show that transcription of ssgA and of the upstream-located ssgR, an iclR-type regulatory gene, is developmentally regulated in Streptomyces coelicolor and activated towards the onset of sporulation. A constructed ssgR null mutant was phenotypically very similar to the ssgA mutant. The absence of ssgA transcription in this mutant is probably the sole cause of its sporulation deficiency, as wild-type levels of sporulation could be restored by the SsgR-independent expression of ssgA from the ermE promoter. Binding of a truncated version of SsgR to the ssgA promoter region showed that ssgA transcription is directly activated by SsgR; such a dependence of ssgA on SsgR in S. coelicolor is in clear contrast to the situation in S. griseus, where ssgA transcription is activated by A-factor, and its control by the SsgR orthologue, SsfR, is far less important. Our failure to complement the ssgR mutant with S. griseus ssfR suggests functional differences between the genes. These observations may explain some of the major differences in developmental control between the phylogenetically divergent species S. coelicolor and S. griseus, highlighted in a recent microreview (Chater and Horinouchi (2003) Mol Microbiol 48: 9–15). Surprisingly, transcription of ssgA and ssgR is not dependent on the early whi genes (whiA, whiB, whiG, whiH, whiI and whiJ ).
Streptomycetes are soil-dwelling Gram-positive bacteria that have an unusually complex life cycle, which makes them particularly interesting for the study of bacterial development and evolution (Chater and Losick, 1997). During its life cycle, Streptomyces undergoes two apparently different events of cell division (reviewed in Flärdh and van Wezel, 2003). Initially, cell division results in the formation of semipermeable septa in the vegetative hyphae (‘crosswalls’) that delimit the multinucleoid hyphal cells. In solid-grown cultures, the reproductive phase is initiated by the formation of an aerial mycelium, with initially aseptate hyphae; in a later sporulation-programmed stage many septa are simultaneously formed, eventually resulting in mono-nucleoid spores (Chater, 2001). Recently, the genome sequences of S. coelicolor and S. avermitilis were elucidated, taking Streptomyces research into the genomics era (Bentley et al., 2002; Ikeda et al., 2003).
Genes involved in the transition from vegetative to aerial mycelium are called bald (bld) genes, characterized by the bald appearance of mutants as a result of their failure to produce aerial hyphae, and those involved in the subsequent processes leading to sporulation, are white (whi) genes, characterized by the white appearance of mutants resulting from failure to complete sporulation. Six whi loci, designated whiA, whiB, whiG, whiH, whiI and whiJ, identified by Chater (1972), are essential for the sporulation process (Flärdh et al., 1999). The best characterized of these sporulation genes are whiG, encoding an RNA polymerase sigma factor (Chater et al., 1989); whiB, encoding a transcription factor with many homologues in streptomycetes and mycobacteria (Soliveri et al., 2000), and whiH, encoding a GntR-family transcription factor (Ryding et al., 1998). The whiH mutant is phenotypically similar to a mutant in which the developmental ftsZ promoter had been inactivated (Flärdh et al., 2000).
Recently, Chater and Horinouchi (2003) compared the developmental regulatory cascades in Streptomyces coelicolor and Streptomyces griseus. These two organisms probably diverged from a common ancestor around 200 million years ago (Embley and Stackebrand, 1994) and an important difference between them is that S. griseus sporulates in submerged culture. The signal for the onset of this still poorly understood process is the production of the γ-butyrolactone A-factor. In S. griseus, A-factor plays a more direct role in developmental control than the highly similar γ-butyrolactones (called SCBs) found in S. coelicolor: Whereas A-factor non-producing mutants of S. griseus are defective in development and antibiotic production (reviewed in Horinouchi, 2002; Chater and Horinouchi, 2003), these processes seem barely affected in γ-butyrolactone-deficient mutants of S. coelicolor (Takano et al., 2001).
One of the key targets of A-factor in S. griseus is ssgA (Yamazaki et al., 2003), a gene encoding a protein unique to sporulating actinomycetes that was originally identified as an effecter of cell division in S. griseus (Kawamoto and Ensign, 1995). SsgA plays an activating role in the production of sporulation septa, as its enhanced expression in S. coelicolor resulted in fragmentation of the mycelia in submerged cultures, producing spore-like compartments at high frequency (van Wezel et al., 2000a). ssgA mutants of S. coelicolor and S. griseus are defective in sporulation, but form apparently normal vegetative septa (Jiang and Kendrick, 2000; van Wezel et al., 2000a). In total seven ssgA-like genes (ssgA-G) occur in S. coelicolor and six in S. avermitilis (Flärdh and van Wezel, 2003). The ssgB gene was recently identified as a novel whi gene and a null mutant produced large non-sporulating colonies (Keijser et al., 2002).
Upstream of ssgA lies ssgR, a member of the family of iclR-type regulatory genes. This family includes IclR itself, the repressor for the isocitrate lyase gene (Sunnarborg et al., 1990) and the acetate utilization operon (Galinier et al., 1990) in E. coli, and the glycerol regulon repressor GylR in S. coelicolor (Hindle and Smith, 1994). The IclR-type proteins, most often repressors, are characterized by an N-terminally situated helix–turn–helix (DNA binding) domain and a C-terminal substrate binding domain, which is also important for oligomerization of the protein (Zhang et al., 2002). None of the other ssgA-like genes is situated near an iclR-type regulatory gene. The SsgR homologues of S. coelicolor, S. avermitilis and S. griseus (called SsfR) are highly similar, although the S. griseus homologue has a predicted N-terminal extension of 84 residues. Insertional inactivation of ssfR in S. griseus is phenotypically similar to the ssgA mutant, although ssgA transcription does not depend on SsfR S. griseus (Yamazaki et al., 2003).
In this work, we study the transcriptional regulation of ssgA and ssgR in S. coelicolor A3(2) strain M145 and its early sporulation mutants whiA, whiB, whiG, whiH, whiI and whiJ, together with the transcriptional dependence of ssgA on ssgR. We identified significant differences in the regulation of ssgA between S. coelicolor and S. griseus, and propose that this is one of the determinants of the morphological and developmental divergence between the two microorganisms.
ssgR is important for sporulation of S. coelicolor
The ssgR gene (S. coelicolor database reference SCO 3925) encodes a 241 amino acid IclR-type regulatory protein. Fourteen putative iclR-like genes could be identified within the S. coelicolor genome. An interesting feature of SsgR is its predicted transmembrane (TM) helix, YALGTVCAAIPITVGTTAATM (residues 185–205; high probability, as predicted by the TMPred server). Orientation is most likely N-terminus inside, which is consistent with the presence of a predicted helix–turn–helix (HTH) DNA binding domain in the N-terminal section of the protein (aa 19–40). Highly similar TM domains are also found in the SsgR homologues from S. avermitilis (SAV4268) and S. griseus (SsfR; AAF61237). Another IclR-type regulator occurs in S. coelicolor with a highly similar putative TM domain (SCO2832, predicted TM sequence YAVGTV CAAVPITAGSAVGCL), which is the closest relative of SsgR in S. coelicolor (40% overall amino acid identity). However, its HTH domain is very different from that of SsgR.
To study the possible role of ssgR in S. coelicolor M145, an in frame deletion mutant of S. coelicolor ssgR was created, as described in the Experimental procedures section. This removed the approximately 280 bp NcoI-SphI fragment of ssgR, resulting in an in frame deletion of the gene corresponding to aa 78–172. This mutant was designated GSR1. As shown in Figure 1, the mutant had a phenotype similar to that of the ssgA mutant GSA3; similarly, GSR1 formed aerial hyphae, but failed to produce spores on rich media such as R2YE or MM with glucose as the sole carbon source. Spores were produced on particular media, notably on MM with mannitol or on SFM, although at reduced levels and after prolonged incubation (Fig. 1). Such an unusual conditionally White phenotype is also typical of the ssgA mutant (van Wezel et al., 2000a).
ssgR is transcribed from a single developmentally regulated promoter
For transcriptional analysis of ssgR and precise localization of the transcriptional start sites, transcript mapping experiments were performed on RNA isolated from S. coelicolor M145 grown on MM agar plates with mannitol as the sole carbon source. RNA was isolated at 12–24 h intervals during 5 days, so as to provide representative samples to analyse transcription. The developmental stage of the samples was monitored using phase-contrast microscopy.
Considering the low expression level of ssgR (see below) we used an end-to-end 32P-labelled RNA probe for transcript mapping. RNA protection analysis using probe ssgR-T7 (−358/+1, relative to the ssgR translational start; Fig. 2A) resulted in one major protected band, with a length of approximately 210 nt (Fig. 3). Transcription of ssgR was upregulated after approximately 64 h, corresponding to the onset of sporulation (few spores were observed by phase-contrast microscopy at this point). The experiment was repeated several times, also using RNA derived from cultures grown on SFM instead of MM agar plates. Although the mycelium grew significantly faster on SFM than on MM, the developmental dependence of ssgR transcription was very similar in all experiments (not illustrated). Promoter probing experiments using the redD reporter system (van Wezel et al., 2000c) confirmed the presence of promoter activity immediately upstream of ssgR (data not illustrated), ruling out the possibility that the transcripts had arisen from processing of an upstream-located promoter.
The exact transcriptional start point of the ssgR transcript was identified by co-migration of the protected RNA probe together with a DNA sequencing ladder. The most likely transcriptional start site was identified as one of two A residues around 210 nt upstream of the translational start of ssgR (Fig. 3B). This transcriptional start site is preceded by the sequence TAGAGT, which fully conforms to the consensus − 10 sequence (TAGAPuT) for promoters recognized by the major RNA polymerase σ factor (σhrdB) of Streptomyces (Strohl, 1992). However, we failed to identify a plausible − 35 sequence. Transcriptional analysis of the S. griseus ssgR orthologue, designated ssfR, showed that it is also transcribed from a single promoter, with a promoter sequence that is similar to that of ssgRp; the −10 sequences are almost identical (TAGAGT for S. coelicolor, TACAGT for S. griseus) and of the −50/−8 regions (relative to the transcriptional start sites), 62% of the nucleotides are identical (S. Horinouchi, pers. commun.). Upstream of the SsgR homologue of S. avermitilis we identified a sequence that is highly similar to the ssgR promoter of S. coelicolor and this sequence therefore constitutes a likely ssgR promoter. This putative promoter (presumed − 50/−1 region is 80% identical to that of S. coelicolor ssgRp) would start around nt position − 240 relative to the start of S. avermitilis ssgR.
Transcription of ssgA is dependent on ssgR
High-resolution mapping of ssgA transcripts was performed on the same RNA as was used for the analysis of ssgR transcription, with RNA probe ssgA-T7 (−195/+41, relative to the ssgA translational start; Fig. 2A and B). Two RNA-protected bands of approximately 125 nt and 110 nt were observed (p1 and p2, respectively; Fig. 4A, left panel). The bands appear as double bands, with one nt difference. Such a duplication is often seen, e.g. for the ftsZ promoters (Flärdh et al., 2000) and the sigF promoter (Kelemen et al., 1998), probably as the result of an experimental artefact. Whereas in all independent experiments it was difficult to detect ssgR-derived transcripts, ssgA transcripts could be readily detected. SsgA transcription was induced after approximately 80 h, a time point where sporulation was well underway, while ssgR transcription was induced one time point earlier, corresponding to the onset of sporulation. Thus, transcription of ssgR is activated at least several hours prior to transcription of ssgA. We did not observe full-length protection of the probe in these experiments, which was confirmed by experiments using probe ssgA-S1 in Fig. 5 (below), which carries a 50 nt non-homologous extension at its 3′ end.
Co-migration of a DNA sequencing ladder showed that the transcriptional start sites corresponded to the approximate nt positions − 84 and − 69 relative to the translational start of ssgA, respectively (Fig. 4B). The transcriptional start sites (full sequences in Fig. 2B) are preceded by the sequences 5′-TTGTGA−18 bp–CAAGAT-3′ (for p1) and 5′ -TTGAGC−15 bp–TTAGAG-3′ (for p2), which show limited similarity to the consensus − 35 and − 10 sequences (5′-TTGACN-16–18 bp–TAGAPuT-3′; (Strohl, 1992) for promoters recognized by the major RNA polymerase holoenzyme of Streptomyces.
Because ssgR and ssgA mutants are phenotypically highly similar, a possible dependence of ssgA transcription on SsgR was investigated. For this purpose, transcriptional analysis was performed using RNA isolated from surface-grown MM mannitol cultures of GSR1, the congenic ssgR mutant of S. coelicolor M145. We repeatedly failed to detect significant levels of ssgA transcripts in these samples, suggesting that ssgA transcription is directly or indirectly dependent on SsgR in S. coelicolor (Fig. 4A, right panel). The integrity of the RNA was confirmed by mapping the transcript of ssgD (SCO6722), one of the six ssgA-like genes in S. coelicolor, using the same RNA as in the experiments for mapping ssgA transcripts. The ssgD gene is expressed in all growth phases and in an ssgR-independent manner, from a single promoter (B. Traag and G. P. van Wezel, unpublished data). We observed no difference between M145 and GSR1 (data not illustrated). The absence of ssgA-derived transcripts in the ssgR mutant was confirmed by RT-PCR experiments (see below).
Expression of SsgA restores sporulation to an ssgR mutant
To further assess the dependence of ssgA transcription on ssgR, we analysed the morphological effect of ssgR-independent expression of ssgA in the ssgR mutant. For this purpose, two constructs were introduced into GSR1, one with ssgA preceded by its own (putatively SsgR-dependent) regulatory sequences and one with ssgA positioned behind the SsgR-independent and constitutive ermE promoter. In a control experiment, we also introduced ssgR expression constructs in the ssgA mutant. In the latter case, no effect was expected. As additional controls, ssgR and ssgA mutants were complemented by wild-type copies of ssgR and ssgA respectively. The results are shown in Fig. 6A.
As was anticipated, the ssgA and ssgR mutants could be complemented by the introduction of wild-type ssgA (on pGWS10, giving transformant GSA4) and ssgR (on pGWR1, transformant GSR2) respectively. This underlines that the non-sporulating phenotype of the ssgR and ssgA mutants is solely the result of the respective gene deletions. Interestingly, morphological differentiation of the ssgR mutant could be fully restored by the expression of ssgA under the control of the SsgR-independent ermE promoter (GSR4 in Fig. 6A). Therefore, the sporulation deficiency of the ssgR mutant is due to the lack of sufficient SsgA. In contrast, introduction of multiple copies of ssgA behind its own promoter did not complement the ssgR mutant (GSR3 in Fig. 6A), providing additional evidence that the natural ssgA promoters are both inactive in an ssgR mutant background. Finally, as expected, introduction of pGWR1 in the ssgA mutant (transformant GSA5) had no apparent effect on development. As a control, all plasmids were transformed to the parental strain M145; these transformants showed normal sporulation, indicating that none of the plasmids had a negative effect on sporulation.
To test if S. griseus ssfR could also restore sporulation to the ssgR mutant, we introduced pGWR5 into GSR1. pGWR5 is essentially the same plasmid as pGWR1, except that it contains ssfR from S. griseus B2682 instead of ssgR. The non-sporulating phenotype of the resulting transformants shows that introduction of ssfR fails to complement the S. coelicolor ssgR mutant (Fig. 6B).
SsgR binds directly and specifically to the ssgA promoter region
To analyse if SsgR could directly bind to the upstream region of ssgA, expression constructs were made to express and purify sufficient quantities of the protein for DNA binding assays (Experimental procedures). Construct pGWR11 was designed to produce full-length SsgR (241 aa, called SsgR-241) in E. coli, containing an N-terminal His6-tag for purification using Ni-NTA chromatography. However, the protein was fully insoluble and, after purification of the protein from inclusion bodies using denaturing procedures, renaturation resulted in complete precipitation of the protein with no detectable soluble protein (as judged by SDS–PAGE). The TMPred program identified a likely transmembrane (TM) region in SsgR, encompassing amino acids 185–205 (also present in the homologues of S. avermitilis and S. griseus). To remove the putative TM domain from the protein, we designed construct pGWR12 to express a shortened SsgR protein (155 aa, called SsgR-155). This resulted in soluble protein, which eluted from a Ni-NTA column in buffer containing 150 mM imidazole.
The protein fraction containing pure SsgR-155 was tested for DNA binding activity in a mobility shift assay. Binding was observed to the DNA fragment SsgA-S1, which contains the − 195/+45 section relative to the start of ssgA (Fig. 2A and B), producing a single and discrete DNA–protein complex (band C in Fig. 7). Complete binding (lane 2) was observed with around 25 ng (approximately 1 pmol) of purified protein; at a 1 : 12 dilution (lane 4) virtually all DNA was in the unbound state, although some residual binding activity could be observed. The mobility shift could be fully reversed by the addition of excess of cold probe (Fig. 7, lane 5), whereas the presence of a large excess of pBR322 did not affect binding. A fragment containing a shorter part of the ssgA promoter region (up to −75 relative to the start of ssgA) bound SsgR with similar affinity, narrowing down the SsgR binding site to the region between − 195 and − 75 relative to the start of ssgA. Specificity of SsgR for the ssgA promoter was underlined by its failure to bind to the SsgR promoter or to DNA fragments harbouring the start and middle part of the ssgR gene (not shown).
Transcription of ssgA and ssgR in sporulation (whi) mutants of S. coelicolor
As ssgA controls the formation of sporulation septa and expression of both ssgA and ssgR is induced during sporulation, a possible dependence of their transcription on the S. coelicolor sporulation genes whiA, whiB, whiG, whiH, whiI and whiJ was investigated. RNA from the corresponding whi mutants (Table 1) grown on MM agar plates with mannitol as the sole carbon source, was analysed by S1 nuclease mapping. Transcription of hrdB, which encodes the principal, essential σ factor of S. coelicolor, is expressed at a relatively constant level in S. coelicolor and was monitored as an internal control (data shown in Kelemen et al., 1998).
S1 nuclease mapping of transcripts with DNA probe ssgA-S1 (−195/+82; Fig. 2A and B) revealed both ssgA transcripts in all whi mutants, showing that none of these genes is essential for ssgA transcription (Fig. 5A). Expectedly, transcripts were absent in RNA isolated from vegetative mycelium. Developmental regulation of ssgA was also not significantly affected in any of the whi mutants, although ssgA transcription was slightly but reproducibly upregulated in the whiH mutant, which is developmentally stalled in a phase immediately before the onset of sporulation-specific cell division (Ryding et al., 1998).
Transcriptional analysis of ssgR in the whi mutants was done with DNA probe ssgR-S1, encompassing the − 338/+47 region relative to the ssgR translational start site (Fig. 2A). Similar to ssgA, ssgR appeared slightly upregulated in the whiH mutant, but was otherwise not significantly affected in any of the whi mutants (Fig. 5B).
RT-PCR analyses of developmental genes
In an independent set of experiments, we performed transcript analysis on RNA isolated from surface-grown MM Mannitol cultures by RT-PCR (Fig. 8). To establish whether ssgR is important for earlier stages of aerial development, we analysed the transcription of the crucial regulatory gene whiG. The outcome of was very similar for M145 and the ssgR mutant, with approximately constant transcript levels of whiG relative to the 16S rRNA in both strains, showing that whiG transcription was not significantly affected by the deletion of ssgR (Fig. 8). Expectedly, there was a strong increase in ssgA transcript levels in RNA samples from surface-grown M145 towards the onset of sporulation (48 and 96 h), while we failed to detect ssgA-derived transcripts in the ssgR mutant. These results correspond well to the transcript mapping experiments presented in Fig. 4A. Considering the dependence of ssgA on adpA (bldH) in S. griseus, we also analysed ssgA transcription in adpA mutant M851, which was characterized previously (Takano et al., 2003). Interestingly, ssgA-derived transcripts were readily detected in this mutant; however, presence of ssgA transcripts in the 24 h sample (corresponding to vegetative growth) indicates that its regulation may be affected in the adpA mutant.
ssgR and ssgA are transcribed in a growth-phase dependent manner
Our results show that ssgA and ssgR are sporulation genes, transcribed in a growth phase-dependent manner on solid-grown cultures. A constructed ssgR mutant had a phenotype very similar to that of the ssgA mutant published previously (van Wezel et al., 2000a), showing a conditional White phenotype (no sporulation except on mannitol-containing media). The ssgR gene was transcribed from a single growth phase-dependent promoter (ssgRp) in solid-grown cultures of S. coelicolor M145. Whereas steady-state transcript levels of ssgR were low under the conditions tested, the activation of ssgRp coincided with the onset of aerial mycelium formation and transcript levels were strongly upregulated at the onset of sporulation. A very similar promoter was responsible for transcription of the ssgR orthologue in S. griseus (S. Horinouchi, pers. commun.). The sequence TAGAGT separated by 5–6 nt from the transcriptional start site, constitutes a − 10 promoter consensus sequence that may be recognized by the principal sigma factor σhrdB, but a − 35 consensus sequence for this probable promoter could not be identified.
ssgA is transcribed from two promoters (ssgA p1 and ssgA p2), separated by approximately 15 nt, suggesting overlap between them. The absence of read-through indicates that on solid-grown cultures, ssgA is transcribed only from these two promoters. Earlier experiments with liquid-grown cultures showed that ssgA is expressed at a low level and only after nutritional downshift and co-regulated with ssgR from a full-length transcript (van Wezel et al., 2000a). This experiment was repeated, again showing read-through from the ssgR promoter (data not illustrated). This discrepancy remains unexplained, but could reflect interesting differences between liquid- and solid-grown cultures. Both promoters were developmentally controlled and maximal transcript levels were reached one time point later than those of ssgR transcripts, at a time corresponding to sporulation. Such a sporulation-specific expression conforms to a role in the activation of sporulation-specific cell division. Despite the similar developmental regulation and strength of ssgA p1 and p2, there is no obvious similarity between the respective promoter sequences.
Comparison to the S. griseus ssgA promoters revealed almost complete conservation between ssgASc p1 and ssgASg p2 (Fig. 4C). This sequence is also highly conserved in S. avermitilis. However, ssgASc p2 shows no sequence similarity to ssgASg p1. Apparently, some elements of ssgA regulation are shared between S. coelicolor and S. griseus, whereas others are different (see the final paragraph).
Expression of ssgA is activated by SsgR
Our failure to detect ssgA transcripts in the ssgR mutant indicated that the growth-phase-dependent induction of ssgA transcription is dependent on SsgR. This is in accordance with the observation that ssgR transcription is strongly induced one time point earlier than that of ssgA, at a time between aerial hyphae formation and the onset of sporulation; transcripts of ssgA itself were induced when sporulation was initiated. Interestingly, ssgA transcripts were more abundant than those of ssgR, as seen in multiple independent experiments. In further support of activation of ssgA transcription by SsgR, introduction of a plasmid expressing ssgA from the constitutive and SsgR-independent ermE promoter restored a wild-type phenotype to the ssgR mutant of S. coelicolor, whereas a similar plasmid harbouring ssgA with its natural promoter failed to complement the mutant. This suggests that whereas it is possible that other genes are regulated by ssgR, the sporulation deficiency of the ssgR mutant is solely due to lack of SsgA and that ssgA transcription fully depends on SsgR. This is also supported by the observation that fragmentation of S. coelicolor in submerged culture, typical of transformants expressing ssgA, was induced by the introduction of pGWR3, a multicopy plasmid harbouring only ssgR (data not illustrated). Interestingly, introduction of a plasmid harbouring S. griseus ssfR did not restore sporulation to the ssgR mutant, indicating significant functional differences between the respective gene products. SsgR of S. coelicolor belongs to the family of IclR-like transcriptional regulators (Zhang et al., 2002). These proteins are characterized by an N-terminal DNA binding domain and a C-terminal ligand binding domain, which is also important for oligomerization of the protein. Typically, binding of the substrate induces a conformational change, releasing the protein from its target sequence (Zhang et al., 2002; Yamamoto and Ishihama, 2003). As an exception, the Streptomyces SsgR/SsfR proteins have a predicted transmembrane helix. Our experiments suggest that cleavage of this putative TM domain is required for its solubility, as only a truncated version of SsgR (155 amino acids long) was soluble. Whether SsgR also binds a substrate is unclear, but considering that on several occasions it was shown that the C-terminal section of IclR-type proteins is required for interaction between monomers, its removal is expected to affect the mode of binding of SsgR. The truncated SsgR-155 produced a single mobility shift on a DNA fragment containing the ssgA promoter region, indicating that the SsgR dependence of the transcription of ssgA is mediated through direct trans-activation by SsgR. The binding site was narrowed down to the − 195/−75 section relative to the ssgA gene. The region around the stop codon of ssgR constitutes a possible binding site, as it har-bours a for streptomycetes unusually A/T-rich sequence (TGAAAACTCACTCC) that shows significant similarity to the consensus sequence TGAAAA(A/T)NNTTTPyPy for IclR-type binding sites (Pan et al., 1996; Zhang et al., 2002). Typically, IclR-type regulators bind to multiple IclR boxes, although we failed to observe additional binding by SsgR. We cannot rule out the possibility that this different behaviour results from the absence of the C–terminal (interaction) part of the protein, which we had to remove to obtain soluble SsgR protein. We are currently analysing the mode of action and the activation of SsgR in more detail.
Interdependence of ssgRA and other developmental genes
In S. coelicolor whiA, whiB, whiG, whiI and whiJ mutants, the level and timing of ssgA and ssgR transcription are comparable to those found in the parental wild-type S. coelicolor A3(2) as well as those in S. coelicolor M145. Considering that ssgRA transcripts are found in all ‘early’whi mutants analysed, this strongly suggests that the gene cluster is not controlled by the classical whi genes. In a reverse experiment, we also showed that whiG is not significantly affected in an ssgR mutant, providing further support for the mutual independence of the ssgRA cluster on the one hand, and the early whi genes on the other. This independence is apparently supported by the observation that the White phenotype of the ssgA and ssgR mutants is medium-dependent and that these genes as well as most of the ssgA-like genes are under carbon catabolite control, which is not the case for other whi genes (manuscript in preparation). Further transcriptional analysis is required to assess whether the observed up-regulation of whiH in the ssgA and ssgR mutants is significant. A possible explanation for the necessity of whi gene-independent expression of ssgRA is that the genes are also involved in the activation of sporulation-specific cell division under conditions where an aerial mycelium is not produced, in particular during submerged sporulation. Indeed, ssgA is essential for this process in S. griseus (Kawamoto et al.,1997) and its overexpression results in hyperseptation and a low level of submerged sporulation in S. coelicolor (van Wezel et al., 2000a).
Differential regulation of ssgRA in S. coelicolor and S. griseus
How does the situation in S. coelicolor compare to that in S. griseus? In the latter organism, ssgA is dependent on AdpA, an A-factor-dependent transcriptional activator that is essential for development and streptomycin production (Ohnishi et al., 2002). In contrast, S. coelicolor scbA mutants fail to produce the A-factor-like γ-butyrolactone SCB1, but show normal sporulation (Takano et al., 2001) and adpA mutants sporulate normally on mannitol-containing media (Takano et al., 2003). As such sporulation is not possible in the absence of ssgA, its transcription is most probably not dependent on SCB1 or AdpA, which was confirmed by our observation that ssgA transcripts could be readily detected in an S. coelicolor adpA mutant. Rather, in S. coelicolor both ssgA promoters are directly dependent on activation by SsgR. Whereas in S. griseus the activity of ssgASg p2 (virtually identical to ssgASc p1, Fig. 4C) also depends on SsfR, that of ssgASg p1 does not, which results in significant expression of ssgA in an ssfR mutant (Yamazaki et al., 2003). Therefore, regulation is clearly different in S. griseus, also illustrated by the significantly higher ssgA transcript levels in this organism, which is a prerequisite for submerged sporulation (G. P. van Wezel, unpublished data).
In summary, several important differences exist between the regulation of the ssgA orthologues in S. coelicolor[indicated with (Sc)] and that in S. griseus (Sg):
issgASc fully depends on activation by SsgR, whereas activity of only one of the ssgASg promoters is reduced in an ssfR mutant.
ii The SsgRSc and SsfRSg proteins may be functionally different, as the gene from S. griseus fails to complement the S. coelicolor ssgR mutant.
iii It is unlikely that the A-factor-like molecule SCB1 plays a role in the regulation of ssgASc, whereas A-factor is essential for the regulation of ssgASg (through AdpA). This may at least partially explain the different impact of A-factor on the development of these organisms, because SCB1 (and on mannitol also AdpA) mutants of S. coelicolor sporulate normally, whereas A-factor is essential for sporulation of S. griseus (Takano et al., 2001).
The expression level of ssgA has a major impact on mycelial morphology of both organisms (Kawamoto et al., 1997; van Wezel et al., 2000a,b) and the different expression levels of ssgA in these organisms thus provides a possible explanation for their strong morphological differences in submerged cultures, and this may be one of the main reasons why S. griseus is able to sporulate in submerged culture but not S. coelicolor.
We are currently investigating the exact roles of SsgA and SsgR in the sporulation process, with focus on their cellular localizations, protein structures and molecular modes of action.
Streptomyces coelicolor A3(2) and its derivative M145, as well as the developmental (whi) mutants (Table 1), were obtained from the John Innes Centre strain collection and adpA mutant M851 from E. Takano (Tübingen, Germany). M145 was used for transformation and propagation of Streptomyces plasmids. Preparation of media, protoplast preparation and transformation were performed according to Kieser et al. (2000). SFM (Soy flour agar plates; Kieser et al., 2000) medium was used to make spore suspensions. Minimal Medium (MM) agar plates containing 0.5% (w/v) mannitol, were used for RNA isolation; R2YE agar plates were used for regenerating protoplasts and, after addition of the appropriate antibiotic, for selecting recombinants. For standard cultivation of Streptomyces and for plasmid isolation, YEME or TSBS [tryptone soy broth (Difco) containing 10% (w/v) sucrose], were used.
Plasmids and constructs
The plasmids and constructs described in this paper are summarized in Table 2 and a map of the ssgRA gene cluster is shown in Fig. 2A.
Table 2. . Plasmids and constructs. Nucleotide numbering is relative to the start of the respective genes.
Streptomyces/E. coli shuttle vector (5–10 and around 100 copies per genome, respectively)
pSET152 harbouring 1.2 kb ssgA fragment (−625/+540 relative to the translational start site of ssgA)
pET15b-based construct for the expression and purification of full-sized SsgR-241
pET15b-based construct for the expression and purification of truncated SsgR-155
General cloning vectors.
pIJ2925 (Janssen and Bibb, 1993) is a pUC19-derived plasmid used for routine subcloning. For cloning in Streptomyces we used the shuttle vectors pHJL401 (Larson and Hershberger, 1986), pWHM3 (Vara et al., 1989) and pSET152 (Bierman et al., 1992). All three vectors have the E. coli pUC19 origin of replication; maintenance in streptomycetes occurs via the SCP2*ori (Lydiate et al., 1985) (five copies per chromosome) on pHJL401, the pIJ101 ori (50–100 copies per chromosome) on pWHM3 and the attP sequence (allowing integration at the attachment site of bacteriophage ϕC31) on pSET152. Plasmid DNA was isolated from ET12567 prior to transformation to Streptomyces. For selection of plasmids in E. coli ampicillin was used, except for pSET152 (apramycin); chloramphenicol was added for ET12567 transformants. For selection in S. coelicolor we used thiostrepton for pHJL401 and pWHM3 and apramycin for pSET152.
Construction of pGWR2 for in frame deletion of ssgR.
To create a construct for in frame deletion of S. coelicolor ssgR, a 1400 bp BglII-BamHI fragment containing ssgR and part of ssgA was inserted into BamHI-digested pIJ2925 and the approximately 280 bp NcoI-SphI segment of ssgR (Fig. 2A) was removed to create an in frame deletion in the ssgR gene on the plasmid. For this purpose, the DNA was digested with NcoI and SphI and the protruding ends were filled in (NcoI) or removed (SphI) using T4 DNA polymerase and dNTPs, followed by ligation and transformation. To ascertain we had created an in frame deletion, the DNA sequence was determined and a 279 bp deletion was confirmed. Subsequently, the apramycin resistance cassette aac(C)IV (Kieser et al., 2000) was inserted into the EcoRI site of the construct (outside the ssgRA insert), producing pGWR2. After transformation of the non-replicating construct to S. coelicolor M145, initial integrants (apramycin resistant) were selected, allowed to sporulate on SFM plates without antibiotics and replicated non-selectively to allow a second recombination event to take place and plated for single colonies. The latter were replicated to SFM containing apramycin, to screen for double recombinants, which should have lost the plasmid and hence have become sensitive to apramycin. About 30% of all apramycin sensitive colonies were sporulation mutants. A check of four sporulating and four non-sporulating double recombinants by PCR revealed that all sporulation mutants carried the expected 300 bp in frame deletion, while sporulating colonies had a wild-type ssgR gene. One of the mutant colonies was selected and designated GSR1. The location of the deletion is shown in Fig. 2A.
Constructs for complementation experiments.
For complementation of the ssgR mutant, plasmid pGWR1 was designed. This low-copy-number pHJL401-based vector harbours the PCR-generated − 350/+950 region (relative to the ssgR translational start), including ssgR itself and approximately 300 bp promoter sequences; the oligonucleotides used for PCR were Q16 and Q11 (Table 3), designed such as to add EcoRI and BamHI restriction sites at the ssgR upstream and downstream end, respectively. A multicopy derivative of pGWR1, designated pGWR3, contains the same insert in pWHM3. Plasmid pGWR5 is essentially the same construct as pGWR1, only with the corresponding region of S. griseus ssfR instead of S. coelicolor ssgR. Plasmids GWS7 and pGWS10 were used for expression of ssgA. pGWS7 (van Wezel et al., 2000a) contains ssgA behind the constitutive ermE promoter in the integrative vector pSET152, while pGWS10 is a pHJL401 derivative, harbouring a 1.2 kb insert with ssgA preceded by its natural upstream (promoter) region. The insert of pGWS10 was generated by PCR with oligonucleotides Q18 and Q6 (Table 3) and corresponded to the − 625/+540 section relative to the start of ssgA.
Table 3. . Oligonucleotides. Added T7 promoter sequences are underlined, restriction sites used for cloning presented in bold face. Restriction sites: gaattc, EcoRI; aagctt, HindIII; ggatcc, BamHI; catatg, NdeI. The BamHI site in Q11 occurs naturally in ssgA, other sites were designed. Location of 5′ end of oligonucleotides (T7 promoter sequences not included) is relative to the start of the gene presented in the last column. The database accession number for whiG is SCO5621, for rrnASCO4123.
Location 5′ end
Constructs for the expression of (His)6-tagged SsgR.
The SsgR expression plasmid pGWR11 was constructed by amplifying ssgR from the S. coelicolor genome with Pfu DNA polymerase and oligonucleotides Q2 and Q3(Table 3) and cloned as an NdeI-HindIII fragment into pET15b, allowing the production of N-terminally (His)6-tagged SsgR; the full sized product was designated SsgR-241. For the expression of a truncated version of SsgR (called SsgR-155), containing the N-terminal 155 amino acids of the protein and lacking the TM signature, we made the pET15b-based expression construct pGWR12. The procedure was the same as for pGWR11, except that oligonucleotides Q2 and R155Bam were used (Table 3), the latter introducing a stop codon immediately downstream of the CTC codon for Leu155. The PCR product was cloned as an NdeI-BamHI fragment into pET15b.
Expression and purification of N-terminally (His)6-tagged SsgR
Plasmids pGWR11 and pGWR12 were used for the expression of full-sized SsgR (241 aa) and truncated SsgR (155 aa), respectively. For protein expression and purification, BL21 codonplus cells (Stratagene) transformed with either pGWR11 or pGWR12 were grown in LB broth at 30°C to an OD600 of about 0.5, SsgR expression was induced by the addition of 0.1 mM IPTG and the culture was incubated for a further 4–6 h. Cells were collected by centrifugation at 4°C and pellets resuspended in resuspension buffer (50 mM Tris-HCl pH 7.5, 5 mM MgCl2, 300 mM NaCl and 10 mM imidazole). After sonication, the samples were spun down and the supernatant and pellet analysed on a 10–12% SDS–PAGE gel. Gels were stained with Coomassie Brilliant blue in 7% ethanol/12% acetic acid.
Soluble SsgR-155 was purified using Ni-NTA affinity chromatography according to the Novagen protocol and eluted with 200 mM imidazole. Samples were dialysed against standard buffer (50 mM Tris pH 7.5, 40 mM NH4Ac, 10 mM MgCl2, 1 mM DTT).
Inclusion bodies containing SsgR-241 were harvested by centrifugation at 15 000 g, solubilized and purified using the Pierce 6xHis B-Per kit, according to the manufacturer's instructions. After purification of SsgR-241 under denaturing conditions using Ni-NTA affinity chromatography, the purified protein was dialysed against standard buffer.
Binding of SsgR to target DNA was studied in the following binding buffer: 20 mM Tris-HCl (pH 7.4), 100 mM KCl, 5 mM MgCl2, 1 mM EDTA, 1 mM DTT and 5% glycerol. To minimize non-specific interactions, 50 ng of pBR322 and 100 µg ml−1 BSA were added. Reaction mixtures (20 µl) contained 3 nM [32P]-labelled DNA probe and 2 µl of diluted protein fraction (between 1 and 20 ng). After 10 min incubation at 30°C, 2 µl dye (50% glycerol, 0.25% xylene cyanol and 0.25% bromophenol blue) was added and samples were analysed on a non-denaturing 6% polyacrylamide gel in 1× TBE buffer. The DNA probe used was produced by PCR using oligonucleotides Q10 and [32P]-labelled Q11 (Table 3) and corresponded to the − 195/+41 region relative to the ssgA translational start, similar to ssgA-S1 (Fig. 2A). Electrophoresis was carried out at 4°C (4 h at 15 V cm−1). After drying, gels were analysed in a phosphor-imager (Bio-Rad).
PCRs were performed in a minicycler (MJ Research, Watertown, MA), using Pfu polymerase (Stratagene, La Jolla, LA) and the buffer provided by the supplier, in the presence of 5% (v/v) DMSO, with annealing temperature of 58°C. For oligonucleotides see next section and Table 3.
Transcript mapping and probes
Mycelium was grown on MM or SFM agar plates with mannitol (0.5% w/v) as the carbon source, by plating spores onto presterilised cellophane discs. The RNA was purified from the surface-grown mycelium using the Kirby-based protocol (Kieser et al., 2000), except that DNaseI treatment was used in addition to salt precipitation to fully eliminate DNA from the nucleic acid preparations. Phase-contrast light microscopy was used to assess the developmental stage of the surface-grown mycelium prior to harvesting and RNA isolation. Appropriate primers were labeled at their 5′ ends with (γ-32P)-ATP by using T4 polynucleotide kinase before DNA probes were produced by PCR, after which high-resolution S1 nuclease mapping was carried out according to previously described protocol (Kieser et al., 2000). Alternatively (α-32P)-UTP-radiolabelled RNA probes were produced with T7 RNA polymerase according to the Maxiscript kit (Ambion) and RNA protection assays were carried out using the RPAIII kit (Ambion). For each RNA protection assay, excess of probe was hybridized to 30 µg of RNA. Protected fragments were analysed on denaturing 6% polyacrylamide gels, where desired alongside a DNA sequencing ladder, produced using the T7 sequencing kit (Amersham Pharmacia Biotech), with as sequencing primers the downstream primers used for generating the PCR-based ssgA and ssgR probes.
Probes used for transcript mapping were produced by PCR, using oligonucleotides described in Tables 3, (i) ssgR-T7, a T7 RNA polymerase-generated RNA probe with incorporated (α-32P)-UTP, produced from a PCR fragment made using oligonucleotides T7-RF and T7-Rrev; the probe was designed against the − 358/+1 region (relative to the ssgR translational start); (ii) ssgR-S1, a DNA probe encompassing the − 338/+47 region (relative to the ssgR translational start), generated by PCR with oligonucleotides Q16 and 32P-labelled Q17; (iii) ssgA-T7, a T7 RNA polymerase-generated RNA probe with incorporated (α-32P)-UTP, produced from a PCR fragment made using oligonucleotides T7-AF and Q10; the probe was designed against the − 195/+41 region relative to the ssgA translational start; (iv) ssgA-S1 was generated by PCR on plasmid pGWS6, using 32P-labelled Q11 and the 17-mer universal (‘forward’) pUC primer and corresponded to the − 195/+82 region (relative to the ssgA translational start); the probe contains an approximately 50 nt non-homologous 3′ extension to discriminate between full-length protection by RNA and experimental artefacts due to probe reannealing.
RT PCR analyses
RT-PCR analyses were carried out using the SuperScript III one-step RT-PCR System (Invitrogen) for the analysis of RNA. RNA was isolated from mycelium grown on SFM agar plates with a cellophane overlay, on which all strains sporulated well (M145, M851) or a bit (ssgR mutant). The samples were prepared after 24 h, 48 h and 96 h, corresponding to vegetative growth, aerial growth and sporulation, respectively. For each RT-PCR reaction 1 µg of RNA was used together with 0.5µM (final concentration) of each primer. The programme used was as follows: 30 min cDNA synthesis at 55°C, followed by 40 cycles of: 15 s at 94°C (denaturing), 30 s at 56°C (annealing) and 60 s at 68°C (elongation). The reaction was completed by 5 min incubation at 68°C. Samples were tested on a 2% agarose gel in TAE buffer and stained with ethidium bromide. The following combinations of oligonucleotides were used (Table 3): (i) ssgA-RT-for and ssgA-RT-rev for ssgA; (ii) whiG-RT-for and whiG-RT-rev for whiG; (iii) 16S-RT-for and 16S-RT-rev for 16S rRNA (to check the integrity of the RNA preparations). RT-PCR experiments without prior reverse transcription were performed on all RNA samples to assure exclusion of DNA contamination. Data were verified in several independent experiments.
We are very grateful to S. Horinouchi for sharing unpublished data and for providing valuable comments on the manuscript, to E. Takano for providing adpA mutant M851, to R. Amons (LUMC, Leiden) for N-terminal amino acid sequencing of SsgR, and to K. Flårdh, B. Keijser, B. Kraal, E. Takano and E. Vijgenboom for stimulating discussions. This work was supported by a grant from the Royal Netherlands Academy of Sciences (KNAW) to G.V.W., and from the Dutch applied research council (STW) to B.T.