Streptomyces coelicolor has nine SigB-like RNA polymerase sigma factors, several of them implicated in morphological differentiation and/or responses to different stresses. One of the nine, SigN, is the focus of this article. A constructed sigN null mutant was delayed in development and exhibited a bald phenotype when grown on minimal medium containing glucose as carbon source. One of two distinct sigN promoters, sigNP1, was active only during growth on solid medium, when its activation coincided with aerial hyphae formation. Transcription from sigNP1 was readily detected in several whi mutants (interrupted in morphogenesis of aerial mycelium into spores), but was absent from all bld mutants tested, suggesting that sigNP1 activity was restricted to the aerial hyphae. It also depended on sigN, thus sigN was autoregulated. Mutational and transcription studies revealed no functional significance to the location of sigN next to sigF, encoding another SigB-like sigma factor. We identified another potential SigN target, nepA, encoding a putative small secreted protein. Transcription of nepA originated from a single, aerial hyphae-specific and sigN-dependent promoter. While in vitro run-off transcription using purified SigN on the Bacillus subtilis ctc promoter confirmed that SigN is an RNA polymerase sigma factor, SigN failed to initiate transcription from sigNP1 and from the nepA promoter in vitro. Additional in vivo data indicated that further nepA upstream sequences, which are likely to bind a potential activator, are required for successful transcription. Using a nepA–egfp transcriptional fusion we located nepA transcription to a novel compartment, the ‘subapical stem’ of the aerial hyphae. We suggest that this newly recognized compartment defines an interface between the aerial and vegetative parts of the Streptomyces colony and might also be involved in communication between these two compartments.
Morphological development in Streptomyces coelicolor is initiated by germination of the unigenomic spores and followed by the formation of long, multigenomic and branching hyphal filaments that generate a dense network of vegetative hyphae growing on the surface and into the solid medium. A major step in the progression of Streptomyces development is the formation of hyphal branches that emerge into the air and later develop into chains of unigenomic spores (Kelemen and Buttner, 1998; Chater and Horinouchi, 2003; Elliot and Talbot, 2004). Although the transition from vegetative to aerial growth is a widely studied topic, little is known in detail about what triggers this process. Nutrient limitation, encounters with stresses and intercellular signalling are all considered as cues that prompt development of the aerial hyphae. Mutants that fail to make aerial filaments lack the fluffy appearance of the wild-type Streptomyces colony and exhibit a ‘bald’ phenotype (Hopwood, 1967; Merrick, 1976). Many such bld genes have been cloned and characterized, and they include three main functional classes: regulatory genes (the major class); genes involved in signalling; and genes encoding structural elements of the aerial hyphae (Kelemen and Buttner, 1998; Chater, 2001).
An intriguing characteristic of several pairs of these bld mutants is that they exhibit the phenomenon called ‘extracellular complementation’ by restoring aerial development, in a unidirectional manner, when grown in close proximity on rich media (Willey et al., 1991; 1993; Kelemen and Buttner, 1998). The cascade of bldJ→bldK→bldA/bldH→bldG→bldC→bldD has been established, where each mutant acts as a donor and restores aerial development of the mutants to its left. Because this ‘complementation’ did not require direct contact of the mutants, but reflected a ‘co-synthesis’ of the type originally described for mutants blocked in antibiotic biosynthetic pathways (Delic et al., 1969), it was suggested that a cascade of at least five extracellular signals are transmitted to finally produce SapB, a morphogenetic surfactant (Willey et al., 1991; 2006). The signals themselves are uncharacterized, except for a small, 655 Da, serine- and glycine-containing oligopeptide shown to induce aerial development of bldJ in a bldK-dependent manner (Nodwell and Losick, 1998). Although the structure of this oligopeptide has not been determined, it was suggested that, because it is taken up by the BldK oligopeptide permease, it might in fact be signal 1 of the cascade shown above. Another signalling molecule with a different chemical structure, the γ-butyrolactone A-factor, plays a major role in both morphological differentiation and antibiotic production of Streptomyces griseus as reviewed in Chater and Horinouchi (2003). In S. coelicolor, the role, if any, of the family of γ-butyrolactones in aerial development is less clear. However, adpA of S. coelicolor, encoding a homologue of the A-factor-dependent central regulator of S. griseus, has been shown to be identical to bldH, a member of the bld signalling cascade (Chater and Horinouchi, 2003; Takano et al., 2003).
Not all bld mutants fit into the hierarchy of the signalling cascade. This applies not only to some of the original bld mutants, bldB and bldF, but also to mutants in the adenylate cyclase gene, cya (Susstrunk et al., 1998) and mutants in the recently reported dasR, encoding a GntR-type regulator involved in N-acetylglucosamine metabolism (Rigali et al., 2006), both of which exhibit bald phenotypes. In addition, carbon utilization of several of the bld mutants is de-regulated (Pope et al., 1996), suggesting that progress of the developmental programme is closely linked to the general physiology of the mycelium.
Recently, significant developments were made in our understanding of the physical requirements for the emergence of hyphal filaments into the air during colony development. The structure of SapB, exogenous application of which restored aerial development to all the bld mutants of the cascade (Willey et al., 1991; 1993), has been solved and revealed a lantibiotic-like modified oligopeptide (Kodani et al., 2004). In addition to SapB, two new classes of proteins, the rodlins and chaplins, were shown to be part of a hydrophobic sheath covering the aerial hyphae and hence the spores. Although mutations in the rodlins did not affect development of the S. coelicolor colony, multiple knockouts of the chaplin family resulted in a bald phenotype (Claessen et al., 2002; 2003; 2004; Elliot et al., 2003a; Elliot and Talbot, 2004).
Although SigF of S. coelicolor has not been implicated in any form of stress response, several other members of the SigB-like sigma family have been shown to be involved both in morphogenesis and in responses to specific stresses, suggesting a connection between stress and development in S. coelicolor. Transcription of sigH was shown, on the one hand, to be induced by heat, ethanol and osmotic stresses, and on the other hand to be directly targeted by BldD, a developmental repressor (Kelemen et al., 2001; Sevcikova et al., 2001; Viollier et al., 2003). Production of another member of the SigB-like sigma family, SigB [= SigJ (Kelemen et al., 2001; Viollier et al., 2003)], was induced during osmotic stress (Cho et al., 2001; Viollier et al., 2003) and a constructed sigB mutant was also impaired in formation of aerial hyphae when grown on solid medium with a high osmolite concentration (Cho et al., 2001).
Interestingly, one of the nine SigB-like sigma factor genes, sigN, is located immediately downstream of sigF on the chromosome of S. coelicolor, raising the possibility that the biological role of sigN is connected to that of sigF. Two sporulation specific sigma factors, SigE and SigG, are encoded by adjacent genes, spoIIGB and spoIIIG, on the B. subtilis chromosome (Masuda et al., 1988; Karmazyn-Campelli et al., 1989). While SigE controls early sporulation in the mother cell, SigG, expression of which depends on SigE, directs transcription in the spore compartment at late stages of sporulation (reviewed in Errington, 1993).
Here we address the biological role of sigN and its potential link to sigF in S. coelicolor. In doing so we discovered that SigN was involved in controlling aerial development and we have defined, for the first time, a novel basal compartment, the ‘subapical stem’, located at the interface between the aerial hyphae and the vegetative mycelium.
A sigN null mutant is conditionally defective in development
In order to identify the role of sigN we generated a sigN knockout mutant using PCR targeting (Gust et al., 2003). The oligonucleotide primers were designed in such way that, apart from the first 13 amino acids (aa) and the last 9 aa, the complete coding sequences for sigN were removed in the mutant allele. A representative sigN null mutant was designated K100.
The phenotype of the sigN mutant was indistinguishable from the wild type when grown on a wide range of media including minimal medium containing 0.5% mannitol (MM-mannitol), soya flour medium containing 1% mannitol (SFM) and during osmotic, heat, cold or oxidative stress. However, K100 exhibited a delay in development when grown on glucose as carbon source. K100 had a ‘pale’ grey appearance on soya flour containing glucose (SF-glucose) instead of mannitol (Fig. 1A), and scanning electron microscopy (SEM) confirmed that the development of K100 was delayed compared with the wild type (Fig. 1B). The sigN mutant spores produced on SF-glucose were thinner and longer than wild-type spores, although after 5 to 6 days growth the spores became rounded, similar to the wild type. Moreover, the sigN mutant colonies were bald, even when grown for long periods, on MM medium containing 1% glucose (MM-glucose), a medium on which the classical bld mutants were identified (Fig. 1C). However, unlike many classical bld mutants, the sigN mutant did not fail to produce the pigmented antibiotics, actinorhodin and undecylprodigiosin, and K100 was indistinguishable from the wild-type S. coelicolor M145 on the rich medium, R2YE, a medium on which classical bld mutants produce no aerial mycelium. The sigN phenotype was complemented when a single copy of the sigN gene was introduced in trans into K100 using pAT20.
The fact that sigN lies downstream of sigF on the chromosome raised the question of a potential functional link between these two sigma factors. To assess such a link, we created a sigF mutant (K101) and a sigF-sigN double mutant (K102) using PCR targeting with appropriate primers. As expected, the phenotype of K101 (pale greenish-grey aerial mycelium) was indistinguishable from that of a previously described sigF mutant, J1978 (Kelemen et al., 1998). Both K101 and J1978 were generated in the M145 background and, while in K101 the complete sigF sequence was replaced by the apramycin resistance cassette (apramycin resistance gene and oriT), in J1978 part of sigF was replaced by the thiostrepton resistance gene. The pale greenish-grey phenotype of K101 (sigF) was observed on a wide variety of media (SFM, MM-mannitol, R2YE, SF-glucose or MM-glucose) and was preserved even after prolonged growth (6–7 days). In contrast, the delayed development of K100 (sigN) was exclusive to media containing glucose and the most severe impairment, lack of emergence of aerial hyphae, was observed only on MM-glucose medium. The sigF-sigN double mutant, K102, exhibited the characteristic sigF phenotype when grown on MM-mannitol or SFM medium. On SF-glucose medium K102 developed more slowly than K101 or the wild-type M145 strain, but after 5 days had a grey-green appearance and further analysis by light microscopy confirmed that it showed the characteristic features of sigF, such as small, round spores often with irregular septation (Potuckova et al., 1995). On MM-glucose medium, however, K102 (sigF-sigN) was indistinguishable from K100 (sigN) and had a bald phenotype.
Transcription of sigN is initiated from two promoters
For the initial transcription analysis S1 nuclease protection assays were performed using RNA from S. coelicolor A3(2) grown on solid MM-mannitol medium together with a radioactively labelled sigN-specific probe. The uniquely labelled 5′ end of the probe was internal to the sigN coding sequences. The same RNA samples were previously used to monitor expression of several developmental genes, including sigF, whiG (Kelemen et al., 1996), sigH (Kelemen et al., 2001), whiH (Ryding et al., 1998) and hrdB (Kelemen et al., 1996). Monitoring transcription of hrdB, encoding the principal sigma factor in S. coelicolor, was used as a control.
sigN transcription was found to be initiated from two distinct promoters, sigNP1 and sigNP2, and both were upregulated during aerial hyphae formation (Fig. 2A). Because sigN lies immediately downstream of sigF and is oriented in the same direction, we had to exclude the possibility of sigN being co-transcribed from a promoter upstream or within sigF. S1 nuclease analysis using a probe with non-specific DNA at its 3′ end confirmed that no sigN transcription originated upstream of sigNP2 under the conditions tested (growing on SFM medium at 30°C, for 1–4 days; data not shown). High-resolution S1 mapping was performed to identify the two transcriptional start points (Fig. 2B). The sigN promoter sequences of S. coelicolor are shown on Fig. 2C.
Because both sigN promoters were upregulated during aerial development (Fig. 2A), we wanted to test whether transcription from either of the sigN promoters was restricted to the aerial hyphae. We therefore monitored sigN transcription during growth in liquid medium, where no morphological differentiation of wild type occurs. While sigNP2 activity was readily detectable, no transcription from sigNP1 was identified throughout liquid growth (Fig. 3A), suggesting that sigNP1 activity might be specifically confined to the aerial hyphae. We also monitored the two sigN promoters in several whi and bld mutants. Neither promoter depended on any of the early whi genes tested: whiG, whiH, whiA and whiB (data not shown). While transcription from sigNP2 was apparent in all the samples, sigNP1 activity was absent in all the bld mutants tested (Fig. 3A), including some of the original point mutants (bldD-J774, bldC-J660 and bldG-WC103) together with constructed bldC and bldD null mutants (Elliot et al., 2003b; Hunt et al., 2005). The dependence of sigNP1 on all bld mutants tested, the coincidence of sigNP1 transcription with aerial hyphae formation on solid medium and the absence of sigNP1 activity during growth in liquid medium where S. coelicolor fails to differentiate, all suggested that sigNP1 transcription was specific to the aerial hyphae.
Transcription of sigN is autoregulated but does not depend on sigF
Transcription of sigB of B. subtilis is initiated from a promoter controlled by SigB itself (Kalman et al., 1990). In S. coelicolor, autoregulation controls transcription of at least two other members of the SigB-like sigma factor subfamily, sigBP1 (Cho et al., 2001) and the P3 promoter of sigH (Sevcikova et al., 2001). In addition, two SigB-like sigma factors from Streptomyces ambofaciens, HasL and HasR, control their own promoters (Roth et al., 2004). The promoter of sigNP2 resembles hrdB-dependent promoters, albeit with a slightly longer spacer of 20 bp, whereas the sigNP1 sequences show high similarity to the consensus promoter sequence recognized by SigB of B. subtilis (Boylan et al., 1991; Haldenwang, 1995; Price, 2000; Figs 2C and 5C). In order to test whether sigN was autoregulated we performed S1 nuclease protection assays using RNA from the wild-type strain M145 and the isogenic sigN mutant, K100 (Fig. 3B). Samples were taken at regular intervals representing distinct phases during morphological differentiation from cultures grown on solid SFM medium. Because the sigN mutant lacks sequences downstream of the first 13 aa, the probe was generated using a labelled oligonucleotide (SigN7) internal to sequences present in the sigN mutant. Transcription from both sigNP1 and sigNP2 was readily detectable in the samples from M145 but no sigN transcription was detected in the sigN mutant. The absence of sigNP1 activity in K100, together with the sigNP1 promoter sequences (Figs 2C and 5C), suggests that SigN directs transcription from the sigNP1 promoter and sigN is therefore autoregulated. It is less straightforward to explain why sigNP2 was affected in the sigN mutant. Previous analysis of the promoters sigHp1 and sigHp2 showed similar apparent ‘co-upregulation’ in the bldD mutant (Kelemen et al., 2001), where only one of the promoters, sigHp1, was a direct target for repression by BldD. Transcription of hrdB was readily detectable in all samples tested, confirming that the absence of the sigN-specific signal was not due to poor RNA preparations.
In order to test whether expression of the two adjacent sigma factor genes, sigN and sigF, were interdependent, we monitored transcription of sigN and sigF in the sigF and sigN mutants respectively (Fig. 4). Transcription of sigN in the wild-type strain was indistinguishable from that in the sigF mutant. Similarly, transcription of sigF was unaffected in the sigN mutant, implying that despite their proximity on the chromosome, there is no interdependence between sigN and sigF expression.
nepA, encoding a small extracellular protein, depends on sigN
Preliminary microarray analysis (K.A. Dalton and G.H. Kelemen, unpubl. data) using the PCR-based S. coelicolor microarrays (Bucca et al., 2003; http://www.surrey.ac.uk/SBMS/Fgenomics/) and the same RNA samples used to monitor the sigN promoters (Fig. 3B) identified SCO4002 as a potential SigN target. In order to confirm these preliminary results we performed S1 nuclease protection assays to monitor SCO4002 transcription in the wild-type M145 and sigN mutant grown on solid SFM medium (Fig. 5A). Transcription of SCO4002 was initiated from a single promoter and coincided with the formation of aerial mycelium during colony development. No SCO4002 transcription was observed during vegetative growth or at late stages of sporulation. This suggested that the transient expression of SCO4002, like that of sigNP1, was restricted to the aerial hyphae. Importantly, transcription of SCO4002 depended on sigN (Fig. 5A) because no SCO4002 transcription was detectable in RNA samples from the sigN mutant. In order to confirm that the lack of SCO4002 transcription was due to the absence of sigN in K100, we showed that SCO4002 promoter activity was restored in K100 containing pAT20 carrying the wild-type sigN allele in trans (Fig. 5B). The promoter sequences of nepA and sigNP1 are compared in Fig. 5C. Most of the RNA used in the S1 nuclease protection assay was taken from mycelium grown on mannitol-containing medium routinely used in the laboratory, allowing comparisons with previous experiments monitoring sigF or hrdB transcription. We tested transcription of sigN and nepA using two independent RNA samples from wild-type M145, and sigN and bldC mutants grown on solid medium containing glucose. The transcription patterns of sigN and nepA were independent of the carbon source (data not shown), so we did not conduct an extensive analysis of RNA extracted from mycelium grown on glucose containing medium.
SCO4002 encodes a putative secreted protein of 108 aa. Searches for homologous proteins using psi-blast (http://www.ncbi.nlm.nih.gov/BLAST/) revealed proteins only among streptomycetes: SAV4213 in Streptomyces avermitilis (http://avermitilis.ls.kitasato-u.ac.jp/) and a yet unannotated protein derived from the translation of sequences from Streptomyces scabies (http://www.sanger.ac.uk/Projects/S_scabies/). SAV4213 contains only 78 aa and the S. scabies protein 101 aa. Alignment of the three protein sequences (Fig. 6A) revealed that the first 31 aa and the last 47 aa are highly conserved among all three proteins. Between these highly conserved regions lies an extension of 29 aa and 25 aa exclusive to S. coelicolor and to S. scabies respectively. signalp (http://www.cbs.dtu.dk/services/SignalP/; Bendtsen et al., 2004) predicted the highly conserved N-terminal 31 aa as a signal peptide with probability 1.000, and a cleavage site between positions 29 (A) and 30 (D) with probability 0.97. This suggests secreted proteins of 49 aa (S. avermitilis), 72 aa (S. scabies) and 79 aa (S. coelicolor). However, the highly conserved residues within these sequences comprise 5–26 aa, which might suggest further processing, generating an even shorter, secreted oligopeptide. Based on the bioinformatic analysis and the fact that SCO4002 depended on sigN, we designated SCO4002 as nepA (sigma N-dependent extracellular oligopeptide).
Using PCR targeting (Gust et al., 2003) we replaced nepA with the apramycin resistance cassette in S. coelicolor M145. A representative nepA mutant was designated as K106 and was further characterized. The morphology of the K106 mutant was monitored on several solid media (including SFM, MM-mannitol, MM-glucose and R5 media) and its phenotype resembled that of the sigN mutant. The progression of aerial development of K106, as compared with the wild-type strain, was delayed when grown on MM-glucose, although this delay was less severe than that of the sigN mutant, allowing aerial hyphae formation at the edges of the colonies. Similarly to the sigN phenotype, this developmental delay was most prominent when K106 was grown as single colonies (Fig. 6B).
SigN directs transcription from the ctc promoter of B. subtilis in vitro, but not from sigNP1 or nepAP
In order to demonstrate the sigma factor activity of SigN in vitro, we overexpressed SigN with an N-terminal His extension in Escherichia coli using the vector pET28a (Novagen). During the purification trials His-SigN was found in the insoluble fraction and was purified using denaturing conditions of the Ni-NTA kit (Qiagen) as shown on Fig. 7A. The apparent molecular weight estimated on SDS-PAGE was higher than that predicted from the His-SigN sequence (36.6 kDa). This aberrant migration is characteristic of many sigma factors and is believed to result from their acidic pI (Haldenwang, 1995). In B. subtilis ctc transcription is initiated by SigB (Igo and Losick, 1986) and the fact that ctc promoter activity does not require any additional transcription factors has made this promoter a commonly used template when testing the activities of SigB-like sigma factors, such as SigF (Kelemen et al., 1998) or SigH (Viollier et al., 2003) of S. coelicolor. Core RNA polymerase from E. coli together with His-SigN was sufficient to produce a specific transcript from Pctc (promoter of ctc) in an in vitro run-off transcription assay (Fig. 7B), demonstrating the sigma factor activity of His-SigN. However, when we used template DNA carrying either the sigN or the nepA promoters, His-SigN failed to initiate transcription in vitro (data not shown). Moreover, no transcription was detected from either the sigN or the nepA promoters when in vitro run-off transcription was attempted using unfractionated RNA polymerase holoenzyme (kindly provided by Luis Servin-Gonzalez and Mark Buttner) purified from liquid cultures of S. coelicolor (data not shown). This suggested either that the holoenzyme that initiates transcription from the sigN and nepA promoters was absent from the preparation, or that an additional missing transcription factor was required for successful initiation at the sigN and nepA promoters.
The location of nepA transcription defines a novel compartment, the ‘subapical stem’, within the aerial hyphae
Transcription of nepA was transiently induced during aerial hyphae formation (Fig. 5A), so we tested whether it depended on some of the bld genes, bldC, bldD and bldG. Transcription of nepA was absent in all the bld mutants tested (Fig. 3A), suggesting that, like sigNP1, nepAP expression was likely to be restricted to the aerial hyphae. To confirm this, we created a nepA–egfp transcriptional fusion plasmid, pK10, by introducing a PCR-generated nepA promoter fragment into pIJ8660 (Sun et al., 1999). pK10 was introduced into S. coelicolor M145 by conjugation and one of the representative exconjugants, K103, was monitored by fluorescence microscopy for green fluorescence during different stages of development on solid SFM or MM-mannitol medium containing apramycin.
Transcription of nepA was completely absent during vegetative growth (Fig. 8) and was exclusive to the aerial compartment. High-resolution images taken from samples grown alongside a coverslip slide indicated that nepA transcription was not present even in the early aerial hyphae prior to sporulation. Moreover, fluorescence from the nepA–egfp fusion was detectable only in the part of the aerial hyphae immediately adjacent to the spore chains and enclosed by the nearest septa at the interface between the vegetative and aerial parts of the colony (Fig. 8B and C). Previously, aerial hyphae-specific transcription has been shown for sigHp2 (Kelemen et al., 2001) and some of the chaplin genes, such as chpH (Claessen et al., 2003; Elliot et al., 2003a). However, both sigHp2- and chpH-specific green fluorescence was localized to the apical part of the aerial hyphae, including, at later stages, the spore chains. The location of nepA transcription specifies a novel compartment we designated as the ‘subapical stem’, which for the first time enables us to define the interface between aerial and vegetative hyphae. To test whether enclosure of the stem depends on sporulation, we monitored nepAP-specific fluorescence in one of the whi mutants, whiG (J2400). Even in a non-sporulating strain, such as the whiG mutant, nepAP activity was not detected in the entire length of the aerial hyphae but was restricted to a compartment enclosed by septation (Fig. 8D). This suggests that neither nepA transcription nor formation of the stem requires sporulation, per se.
The fact that, although nepA depended on sigN, SigN was not sufficient for transcription in vitro from the nepA promoter raised the possibility that initiation of nepA transcription required some activator besides the RNA polymerase holoenzyme. Generating further egfp constructs, pK11 and pK12, enabled us to test the promoter activities of DNA fragments of different length upstream of the transcriptional start point (tsp). While pK10 carried 360 bp upstream of nepA tsp, pK11 and pK12 contained 126 bp and 55 bp upstream sequences respectively. Transcription of nepA judged by the green fluorescence of K104, carrying pK11, was indistinguishable in strength and location from that of K103, carrying pK10. However, despite the fact that in pK12 the 55 bp upstream sequences contained both the putative −10 and −35 promoter sites, no nepA-specific fluorescent signal was detectable in S. coelicolor M145 carrying pK12 (data not shown). Probably the extra 71 bp in pK11 permits the binding of one or more additional proteins that are needed for promoter activation. This would be consistent with the negative results of the in vitro run-off transcription experiments, in which such additional factors were not expected to be present.
We attempted to dissect the sigN promoters by generating egfp fusions carrying only sigNP1, or sigNP2 or both promoters (details in Experimental procedures and Fig. 2C). Because sigNP1 is 40 nt downstream of sigNP2, the construct carrying sigNP1 on its own contained 46 bp upstream of the P1 tsp, including the putative promoter site. After introducing the generated constructs into S. coelicolor M145 no fluorescence signal was detected when sigNP1–egfp was monitored. This suggested that one or more additional transcription factor(s) is required for transcription from the sigNP1 promoter. Fluorescence signals specific to sigNP2 and sigNP1+2 were detectable, albeit extremely weak (Fig. S1). The weak fluorescence signal of sigNP2–egfp was present in both the substrate and aerial compartments, while the sigNP1+2–egfp fusion clearly indicated an increased signal in the subapical stem compartment of the aerial hyphae. This suggested that transcription of nepA and sigNP1 coincided not only temporally, as shown by the S1 nuclease analysis (Figs 3B and 5A), but also spatially to the subapical stem of the aerial hyphae.
sigN, a new member of the bld gene family
Similarly to the classical bld mutants, mutation in sigN resulted in a bald phenotype when colonies were grown on MM-glucose as carbon source, and the bld phenotype was ‘rescued’ when sigN colonies were grown on medium containing the less well-utilized carbon source, mannitol. However, unlike the classical bld mutants, which are also blocked in production of aerial hyphae on a rich medium, R2YE, the sigN colonies were indistinguishable from wild type on R2YE medium. Mutations in the bld genes often result in the failure to produce secondary metabolites, some of which are coloured blue (actinorhodin) or red (undecylprodigiosin). Although we observed some slight differences in pigmentation of the sigN mutant when grown on media containing glucose (SF-glucose or MM-glucose; Fig. 1), colonies of the sigN mutant never lacked pigmentation.
One of the classical bld genes, bldG, encodes a putative anti-anti-sigma factor homologous to RsbV or SpoIIAA of B. subtilis that could, in principle, control the activity of an RNA polymerase sigma factor (Bignell et al., 2000). Activation of SigB and SigF in B. subtilis are the most studied examples of post-translational sigma factor activation via protein–protein interactions. The sigma factor (SigB or SigF) is held inactive by binding to an antagonist protein or anti-sigma factor (RsbW or SpoIIAB). In turn, the anti-anti-sigma factor (RsbV or SpoIIAA) can abolish this interaction by making a complex with the anti-sigma factor and releasing the sigma factor to initiate transcription. Complex formation between the anti-anti-sigma factor and the anti-sigma factor depends on phosphorylation of RsbV or SpoIIAA by RsbW or SpoIIAB respectively (Benson and Haldenwang, 1993; Duncan and Losick, 1993; Min et al., 1993; Alper et al., 1994; Duncan et al., 1996). Next to bldG there is an open reading frame that could function as a sigma factor antagonist, and phosphorylation of BldG has been shown both in vivo and in vitro (Hesketh et al., 2002; Bignell et al., 2003). However, while in B. subtilis the three genes encoding SpoIIAA, SpoIIAB and SigF or RsbV, RsbW and SigB are adjacent on the chromosome, there is no sigma factor gene in the vicinity of S. coelicolor bldG (Bignell et al., 2000).
Could SigN be controlled by an anti-sigma factor, which in turn is controlled by BldG? Analysis of the gene organization around the nine SigB-like sigma factor genes identified sigF, sigN, sigK and sigG with no adjacent putative anti-sigma, or anti-anti-sigma genes. The presence of more than 40 genes encoding putative anti-sigma factors and at least 17 genes encoding putative anti-anti-sigmas (C. Reynold, pers. comm.) in the S. coelicolor genome suggests that the physical proximity of genes encoding partner proteins might not be critical. Furthermore, some degree of promiscuity is inevitable: some sigmas might be controlled by multiple anti-sigmas, or conversely some anti-sigmas might control the activity of multiple sigmas. Similarly, multipartnerships can also be expected between the anti-sigmas and anti-anti-sigmas. The choice of a partnership might be highly dependent on the availability and activity of the partners, which, in turn, could be controlled by physiological state, nutrient availability or environmental stimuli. While the polycistronic gene organization of the spoIIA operon of B. subtilis allows equimolar production of the interacting proteins, the potential multipartnership in S. coelicolor might not require a 1:1 ratio for any single partner and a different, more complex mechanism is expected to control the availability of any given sigma factor.
The bald phenotype of the sigN mutant is compatible with the hypothesis that BldG might be an antagonist of an anti-SigN protein under the physiological condition of growth in the presence of glucose. If so, SigN might not be the only sigma factor regulated by BldG, because, besides being bald, the bldG mutant failed to produce the pigmented antibiotics, a characteristic not shared with the sigN mutant. Further experiments are needed to establish the link, if any, between SigN and BldG.
SigN is not sufficient to initiate transcription at sigNP1 and nepA promoters
Comparison of the promoter sequences of sigNP1 and nepA with the consensus SigB-dependent promoters (Fig. 5C; Haldenwang, 1995) displayed strong similarity in the −35 sequences and a somewhat lower but significant similarity in the −10 sequences, implying that these promoters could be targets for a SigB-like sigma factor. Transcription from both sigNP1 and nepA depended on sigN in vivo, suggesting that SigN might control these promoters directly. However, run-off transcription using RNA polymerase holoenzyme containing SigN failed to initiate transcription at these promoters in vitro, although successfully initiated transcription from the ctc promoter of B. subtilis under the same conditions. In addition, no transcripts were generated from the sigN and nepA templates when we used unfractionated RNA polymerase holoenzyme purified from liquid cultures of S. coelicolor, a condition in which SigN is expected to be produced via the sigNP2 promoter. All the above leads to two possible conclusions: (i) that SigN does not directly interact with these promoters but indirectly controls their expression, or (ii) that SigN is necessary but not sufficient for transcription initiation from the sigNP1 and nepA promoters. Indeed, monitoring fluorescence using transcriptional egfp fusions of nepA and sigN demonstrated no in vivo promoter activity from DNA fragments carrying little more than the −10 and −35 sequences of sigNP1 or nepA, suggesting that transcription from these promoters requires some additional upstream DNA, perhaps to interact with transcription factor(s). Only when these transcription factors are identified can we begin to address which conclusion above is correct.
The striking similarities within the promoter recognition domains of the nine SigB-like sigma factors (Fig. S2) raise some fundamental questions. First, do these sigma factors recognize distinct sequences at their specific target sites? Although computer searches (Lee et al., 2004b) identified several putative S. coelicolor promoters predicted to be recognized by SigB [= SigJ (Viollier et al., 2003)], only two promoters, sigBP1 and catBP, have been shown both in vitro and in vivo to be targets for SigB (Lee et al., 2004b). Also, the B. subtilis ctc promoter was transcribed in vitro by SigF (Kelemen et al., 1998), SigH (Viollier et al., 2003), SigN (shown in this article), SigM, SigJ and SigI (unpublished results from G.H. Kelemen's laboratory) – in fact by all the SigB-like sigma factors of S. coelicolor that have been tested. This leads to the second fundamental question: how can these sigma factors with potentially overlapping promoter specificities exhibit distinct biological functions? Besides regulating the activity of the sigma factors both temporally and spatially, one possible answer could entail ‘weak’ target promoters that require sigma factor-specific transcriptional activators for successful transcription initiation. Besides SigN, this could also explain the way in which SigF controls its targets, as SigF controls whiEP2 in vivo, but failed to transcribe from whiEP2 in vitro (Kelemen et al., 1998; Homerova et al., 2000; Novakova et al., 2004).
What is the role of the subapical stem compartment during colony development?
Correct localization of proteins controlling bacterial development is essential to their biological function (Shapiro and Losick, 1997; 2000). Previous transcriptional fusions to egfp localized the spatial patterns of expression of several Streptomyces genes largely to two distinct ‘tissues’, the aerial and the vegetative hyphae. Transcription specific to aerial hyphae has been shown for sigHp2 (Kelemen et al., 2001) and for some of the chaplin genes, such as chpH (Claessen et al., 2003; Elliot et al., 2003a). In contrast, transcription of redD, encoding a pathway-specific activator for production of the red-pigmented antibiotic, undecylprodigiosin, was confined to vegetative hyphae (Sun et al., 1999). Interestingly, both sigHp2- and chpH-specific green fluorescence was localized to the apical part of the aerial hyphae, including, at later stages, the spore chains. Sporulating aerial hyphae were also the location for transcription of sigF (Sun et al., 1999) and one of the two ftsZ promoters (Flardh et al., 2000). In this article we have shown that transcription of nepA was exclusive to a novel, developmentally distinct location, the subapical stem of the aerial hyphae. The subapical stem compartment is defined by our analysis as a hyphal segment that is enclosed by septa, at least one of which is the septum separating the spore chains from the non-sporulating hyphal filament in wild-type colonies (Fig. 9). Although we have never detected nepA-specific fluorescence prior to sporulation in the wild-type strain, the presence of nepA transcription in a non-sporulating mutant, whiG, confirmed that nepA transcription was not dependent on sporulation, per se. While the three bld genes, bldC, bldD and bldG, were required for nepA or sigNP1 transcription and, presumably, for the emergence of the stem, the earliest acting of known essential sporulation genes, whiG, was not needed for the formation of the stem compartment.
It is not known exactly how transformation of the aerial hyphae into chains of spores begins, although it was speculated that the formation of a special septum could enclose the part of the aerial hyphae that is converted into spores (Flardh et al., 1999; Kwak et al., 2001). This special septum would divide the aerial hyphae into two subcompartments, the stem and the tip, with distinct fates. The fate of the tip compartment is transformation into spore chains during normal development, while the fate of the subapical stem is still unknown. Early characterization of the organization of the aerial mycelium has described the spore-bearing cells as ‘fertile’ and the non-sporulating hyphal fragments as ‘sterile’ (Wildermuth, 1970); the latter, we believe, might correspond to the subapical stem. Here, for the first time, we have assigned a specific function to this compartment, the upregulation of sigN expression together with the production of NepA. Because nepA encodes a putative secreted protein or peptide it is conceivable that the subapical stem, at least in part, is involved in signalling between the aerial tip compartment and the vegetative mycelium. In Streptomyces, signalling has been shown to control both development (the bld signalling cascade) and antibiotic production (γ-butyrolactones) and signalling in bacteria is implicated in a wide range of biological phenomena such as competence, quorum sensing or biofilm formation (see review by Bassler and Losick 2006). C-factor, a 17 kDa protein, is involved in cell-to-cell communication essential for fruiting body formation and sporulation of Myxococcus xanthus (Kim and Kaiser, 1990a,b; Lobedanz and Sogaard-Andersen, 2003). Meanwhile, the complex programming of sporulation in B. subtilis relies on cross-talk between the mother cell and fore-spore compartments, where communication is achieved via secreted proteins, SpoIIR and SpoIVD (reviewed in Losick and Stragier, 1992; Rudner and Losick, 2001).
Could NepA be part of the bld signalling cascade? The first signal in the hierarchy of extracellular communication between a set of bld mutants was predicted to be a peptide (Nodwell and Losick, 1998). However, the nepA mutant restored aerial hyphae formation to the bldJ mutant (data not shown) that is predicted to be blocked in production of signal1. The nepA mutant was indistinguishable from the wild-type strain and exhibited extracellular complementation when grown next to bldJ, bldA, bldC, bldD or bldH on R2YE medium (data not shown). This suggests that either nepA is located to the ‘right’ of bldD or it is not part of the originally described bld cascade.
According to the early paper by Wildermuth (1970) the ‘sterile’ aerial fragments, likely to be identical to the subapical stem compartment described in this article, frequently showed signs of lysis. Recently, two rounds of orderly process of cell disintegration have been demonstrated in Streptomyces antibioticus: the first round taking place in the young substrate mycelium while the second round affecting the aerial mycelium (Manteca et al., 2006). It is conceivable to speculate that SigN might be involved in the ordered programme of cell death in S. coelicolor : the sigNP2-specific expression controlling events in the substrate mycelium, whereas the stem-specific, sigNP1-driven expression affecting the breakdown of the subapical stem compartment. Further microarray analysis of the nepA and sigN mutants grown on MM-glucose will be required to further elucidate the role of nepA and sigN during development together with the distinct developmental programme of the subapical stem compartment.
Bacterial strains and growth conditions
Escherichia coli BW25113 (Datsenko and Wanner, 2000) carrying pIJ790 was the host for recombination between the tailored apramycin resistance cassette and the target gene to generate knockout mutants. ET12567 (MacNeil et al., 1992) containing pUZ8002 (Kieser et al., 2000) was used to transfer plasmids or cosmids from E. coli into S. coelicolor by conjugation. E. coli DH5α (Hanahan, 1983) was used for routine cloning. The strains of S. coelicolor used in this work are listed in Table 1. Streptomyces strains were grown at 30°C on SFM, MM-glucose or MM-mannitol solid media or in tryptic soya broth liquid medium (Kieser et al., 2000). In order to test the sigN mutant phenotype we also used a SF-glucose solid medium prepared like SFM, except that mannitol was replaced by 1% glucose. pIJ82 is a derivative of pSET152 (Bierman et al., 1992) in which the apramycin resistance gene is replaced with the hygromycin resistance gene (H.M. Kieser, pers. comm.). pIJ82 was used to generate the construct for complementing K100 (sigN::apr).
Mutant alleles of sigN and sigF were generated in cosmid D57 (a gift from H.M. Kieser) because in this cosmid, both sigN and sigF lie more than 7 kb from the cosmid ends. D57 was part of the original cosmid library, but did not become a member of the ordered set, where sigN is represented by two overlapping cosmids, 2St10A7 and 2StD60 (Redenbach et al., 1996). To create the nepA (SCO4002) null allele we used cosmid 2St10A7. Knockout mutants were generated using PCR targeting (Gust et al., 2003). The apramycin resistance cassette (apramycin resistance gene and oriT) was PCR-amplified using appropriate primers: for the sigN mutant, D13sigN1 and D13sigN2; for the sigF mutant, SigFknock.1 and SigFknock.2; for the sigF-sigN double mutant, SigFknock.1 and D13sigN2; for the nepA mutant, 4002-knock1 and 4002-knock2 (see Table 2). The PCR cassettes were then introduced into cosmid D57 or 2St10A7 in BW25113 carrying pIJ790 (Datsenko and Wanner, 2000) to generate D57/sigN::apr; D57/sigF::apr; D57/sigFN::apr and 2St10A7/nepA::apr mutants in E. coli. After the mutant alleles were passaged through the dcm-dam- ET12567 strain (MacNeil et al., 1992) containing pUZ8002 they were introduced into S. coelicolor M145 by conjugation. The double-cross-over exconjugants were tested for resistance to apramycin and sensitivity to kanamycin. To confirm the mutants, chromosomal DNA was isolated and analysed by Southern blot hybridization.
Table 2. List of oligonucleotides used in this work.
(A) Primers used to generate knockout mutations. Nucleotides corresponding to the apramycin resistance template are shown in bold. (B) Primers for generating egfp transcriptional fusions. Restriction endonuclease sites are shown in bold (B and D). (C) Primers used to produce probes during S1 nuclease protection assays. (D) Primers used to overexpress SigN.
A ∼1.4 kb BamHI–KpnI DNA fragment containing the entire sigN gene including its promoter sequences but no adjacent genes was first introduced into pIJ2925 (Janssen and Bibb, 1993) to generate pAT19. The sigN fragment was liberated from pAT19 using BglII, and introduced into the BamHI site of pIJ82, to create pAT20. pAT20 was introduced into S. coelicolor M145 and K100 by conjugation and hygromycin-resistant exconjugants, carrying pAT20 integrated to the ΦC31 attachment site on the chromosome, were examined for their phenotype.
S1 nuclease mapping
Mycelium grown on solid medium covered with cellophane discs or in liquid medium was collected, and initial extracts were generated by grinding in liquid nitrogen. From these extracts RNA was prepared and S1 nuclease protection assays were performed according to Kieser et al. (2000) using 40 μg of RNA in each assay. Probes labelled with 32P at their single 5′ end were generated using PCR with two oligonucleotides, one labelled at its 5′ end with [γ-32P]-ATP using T4 polynucleotide kinase. Probes were generated using the following oligonucleotide pairs: sigN probes using SIGN4 and either SIGN5* or SIGN7*; sigF probes using SIGF15 and SIGF16*; hrdB probes using HRDB1* and HRDB2; nepA probes using 4002.1* and 4002.2 (asterisks mark the labelled oligonucleotides). The probe for hrdB, encoding the principal sigma factor in S. coelicolor, was used as a control to assess the RNA preparations. DNA fragments protected by RNA were separated on 6% sequencing gels. Transcriptional start points of sigN and SCO4002 were identified against a dideoxy-sequencing ladder (Amersham Pharmacia biotech; T7 sequencingTM kit) produced using the same labelled oligonucleotide used to generate the probes for S1 analysis.
Generation of the nepA–egfp and sigN–egfp transcriptional fusions
Two appropriate primers, 4002-Kpn and 4002-Bam, were used to generate a PCR product containing KpnI and BamHI sites, together with sequences 55 bp downstream and 360 bp upstream of the transcriptional start point of nepA. This PCR product was introduced into pIJ8660 (Sun et al., 1999) as a BamHI–KpnI fragment generating pK10, which was introduced into S. coelicolor M145 by conjugation. Apramycin-resistant exconjugants were selected and one of the representative strains, K103, was used to monitor nepA transcription. In a similar way, we generated pK11 and pK12, carrying truncated nepA fragments containing 126 bp and 55 bp sequences upstream of the transcriptional start respectively. The truncated sequences were generated by PCR using 4002-Bam2 with 4002-Kpn and 4002-Bam3 with 4002-Kpn oligonucleotides. The positions of the oligonucleotides compared with that of the nepA promoter are shown in Fig. 5D. pK11 and pK12 were introduced into S. coelicolor M145, generating K104 and K105 respectively. As pIJ8660 integrates into the ΦC31 attB site in the S. coelicolor chromosome, to maintain selection for the presence of the plasmids, we propagated K103, K104 and K105 on medium containing apramycin.
To generate sigN–egfp fusions we used DNA fragments created by PCR using the following oligonucleotides: P1sigN1 and P1sigN2 to carry only sigNP1; P2sigN1 and P2sigN2 to carry sigNP2 exclusively; or P1sigN1 and P2sigN2 to carry both sigNP1 and sigNP2 promoters. These PCR products were introduced into pIJ8660 as BamHI–KpnI fragments and the constructs generated were moved to S. coelicolor M145 as described above. Fluorescence was monitored using a Leica TCS SP2 laser-scanning confocal microscope with a 63×, 1.4NA oil-immersion objective.
Overexpression and purification of His-SigN
The sigN gene was introduced into pET28a (Novagen) as an NdeI–EcoRI fragment generated by PCR using the primers SigNover1 and SigNover2 followed by the appropriate restriction digests. One of the clones confirmed by sequencing was designated pK20 and moved into E. coli BL21 (DE3)pLysS (Novagen). Cultures of BL21(DE3)pLysS/pK20 were grown at 37°C to an OD600 = 0.4 and, after induction with 1 mM IPTG, for a further 4 h. After harvesting the mycelium, His-SigN was purified under denaturing conditions (8 M urea) using the Ni-NTA Spin Kit (Qiagen) according to the manufacturer's instructions. Fractions eluted from the Ni-NTA column with a buffer of low pH (4.5) were dialysed against 50 mM NaH2PO4 and 50 mM NaCl, pH = 8.0 buffer at 4°C. The His-SigN protein was analysed on an 8% SDS-PAGE and stored at −20°C in the presence of 20% glycerol.
In vitro run-off transcription was performed as described by Buttner et al. (1987) using 0.5 μg of His-SigN, core RNA polymerase from E. coli (Cambio), [α-32P]-CTP (3000 Ci mmol−1; Amersham Biosciences) and an appropriate DNA template. The 340 bp EcoRI–BamHI fragment of pMI340 (Igo and Losick, 1986) was used to test transcription from the B. subtilis ctc promoter. Templates carrying the sigN and nepA promoters were generated by PCR using the SIGN4, SIGN7 and 4002.1, 4002.2 oligonucleotide pairs respectively.
Scanning electron microscopy
Single colonies grown on solid medium were immobilized to the aluminium SEM stub using O.C.T. compound (BDH Laboratory Supplies, Poole, UK). The stub was then immediately plunged into liquid nitrogen slush at approximately −210°C to cryo-preserve the material. The sample was transferred onto the cryo-stage of a CT1500HF cryo-transfer system (Gatan, Oxford, UK) attached to a Philips XL30 FEG scanning electron microscope (Philips Electron Optics, FEI UK, Cambridge, UK). Sublimation of surface frost was performed at −95°C for 3 min before sputter-coating the sample with platinum for 2 min at 10 mA, below −110°C. After sputter-coating, the sample was moved onto the cryo-stage in the main chamber of the microscope, held at approximately −140°C. The sample was viewed at 3 kV and digital TIFF files were stored.
We thank Kim C. Findlay for assistance with the scanning electron microscopy, Dylan Sweetman for the stereomicroscopy images of Streptomyces colonies, Colin Smith and his laboratory for training and help in transcriptomics, Luis Servin-Gonzalez and Mark Buttner for RNA polymerase holoenzyme preparations from S. coelicolor, and Kate Bowdrey for her contribution to the overexpression of SigN. We are very grateful to David Hopwood, Keith Chater, Matthew Hutchings and Andrew Johnston for their valuable comments on the manuscript. This work was supported by BBSRC Grant G15940 to G.H.K.