SapB and the chaplins: connections between morphogenetic proteins in Streptomyces coelicolor


*E-mail; Tel. (+1) 905 525 9140; Fax (+1) 905 522 6066.


Morphogenesis in the streptomycetes features the differentiation of substrate-associated vegetative hyphae into upwardly growing aerial filaments. This transition requires the activity of bld genes and the secretion of biosurfactants that reduce the surface tension at the colony–air interface enabling the emergence of nascent aerial hyphae. Streptomyces coelicolor produces two classes of surface-active molecules, SapB and the chaplins. While both molecules are important for aerial development, nothing is known about the functional redundancy or interaction of these surfactants apart from the observation that aerial hyphae formation can proceed via one of two pathways: a SapB-dependent pathway when cells are grown on rich medium and a SapB-independent pathway on poorly utilized carbon sources such as mannitol. We used mutant analysis to show that while the chaplins are important, but not required, for development on rich medium, they are essential for differentiation on MS (soy flour mannitol) medium, and the corresponding developmental defects could be suppressed by the presence of SapB. Furthermore, the chaplins are produced by conditional bld mutants during aerial hyphae formation when grown on the permissive medium, MS, suggesting that the previously uncharacterized SapB-independent pathway is chaplin dependent. In contrast, a bld mutant blocked in aerial morphogenesis on all media makes neither SapB nor chaplins. Finally, we show that a constructed null mutant that lacks all chaplin and SapB biosynthetic genes fails to differentiate in any growth condition. We propose that the biosurfactant activities of both SapB and the chaplins are essential for normal aerial hyphae formation on rich medium, while chaplin biosynthesis and secretion alone drives aerial morphogenesis on MS medium.


The developmental life cycle of the sporulating, filamentous, Gram-positive soil bacterium Streptomyces coelicolor is a valuable model for the study of prokaryotic differentiation (as reviewed in Chater and Horinouchi, 2003). Growth of S. coelicolor begins with spore germination and hyphal outgrowth, which leads to the establishment of a vegetative, or substrate, mycelium. The sensing of nutrient deprivation is thought to stimulate reproductive growth, which results in the emergence of aerial hyphae known collectively as an aerial mycelium, in which chains of spores develop. Much of the research into S. coelicolor development has focused on the complex regulatory networks that govern the differentiation of vegetative hyphae into upwardly growing aerial hyphae (Kelemen and Buttner, 1998; Chater and Horinouchi, 2003). In contrast, little was known about the actual mechanics of aerial hyphae formation until very recently (Elliot and Talbot, 2004; Claessen et al., 2006; Willey et al., 2006). Importantly, it appears that the surface tension between the aqueous milieu of the colony and the air must be reduced if nascent hyphae are to grow into the air, and this can be accomplished by two classes of surfactant-like molecules: the SapB peptide (Tillotson et al., 1998; Kodani et al., 2004) and the chaplin proteins (Claessen et al., 2003; Elliot et al., 2003a).

SapB is an amphiphilic, ‘lantibiotic-like’ lanthionine-containing peptide (Kodani et al., 2004). Lantibiotics are ribosomally synthesized oligopeptide antibiotics produced by Gram-positive bacteria that are translated as inactive prepeptides, and undergo extensive post-translational modification before being cleaved to yield the mature peptide (Sahl and Bierbaum, 1998). Unlike most lantibiotics, SapB does not appear to have antibiotic activity. Instead, it is secreted onto the colony surface where it is thought to function as a biosurfactant, facilitating the upward growth of emerging aerial hyphae. A similar lantibiotic-like structure and role have been ascribed to SapT, produced by Streptomyces tendae (Kodani et al., 2005). The surfactant activities of these peptides are evidenced by their capacity to restore aerial development to Streptomyces mutants blocked in aerial hyphae formation [the so-called bald (bld) mutants] when added extracellularly (Willey et al., 1991; Tillotson et al., 1998; Kodani et al., 2005).

The SapB structural gene, ramS, encodes a 42-amino-acid peptide. The C-terminal half of the peptide undergoes post-translational modification involving the dehydration of four alanine residues and the introduction of two lanthionine bridges. The modified peptide is exported and the leader peptide is cleaved to yield the 21-amino-acid mature SapB (Kodani et al., 2004; Willey et al., 2006). ramS is part of an operon that includes three other genes: ramC, which encodes an enzyme thought to dehydrate and cyclize the RamS translation product (Kodani et al., 2004), and ramA and ramB, which encode components of an ABC transporter likely involved in SapB export (Ma and Kendall, 1994; Keijser et al., 2000; Kodani et al., 2004; Willey et al., 2006). Transcription of the ramCSAB gene cluster is positively regulated by the product of the convergently transcribed ramR gene, which encodes a response regulator-like protein that binds to the ramCSAB promoter region (Keijser et al., 2002; Nguyen et al., 2002; O'Connor et al., 2002; O'Connor and Nodwell, 2005). There is no cognate sensor kinase genetically linked to ramR, and nothing is currently known about the regulatory mechanisms that directly activate ramR gene expression at the onset of aerial mycelium formation. bld mutants fail to make SapB, but there has yet to be a direct connection established between any bld gene products and the ram gene cluster in S. coelicolor. In wild-type S. coelicolor, SapB is produced only during growth on rich medium, and it is completely absent during aerial development on minimal medium (Willey et al., 1991).

The chaplins are a group of eight secreted, surface-active proteins that are necessary for aerial hyphae formation (Claessen et al., 2003; Elliot et al., 2003a). All chaplins have an N-terminal Sec secretion signal, and they share a conserved hydrophobic domain called the ‘chaplin domain’. Five of the chaplins, ChpD–H, consist solely of a signal peptide and a single-chaplin domain, and are referred to as the ‘short’ chaplins, while the remaining three ‘long’ chaplins (ChpA–C) have a signal peptide and two chaplin domains followed by a C-terminal sorting signal that targets them for covalent attachment to the cell wall by the sortase enzyme (Claessen et al., 2003; Elliot et al., 2003a; Marraffini et al., 2006). Deletion of individual chaplin genes has no discernible effect on colony morphology, and at least four chaplin genes must be knocked out before a notable delay in morphogenesis is detectable. Unlike SapB, chaplin production does not appear to be medium dependent, as the chaplin genes are transcribed both on rich medium (Elliot et al., 2003a) and on minimal medium (Claessen et al., 2003). Like SapB, however, the chaplins are not expressed by bld mutants grown on non-permissive rich media (Elliot et al., 2003a) and, as is the case for ramR, nothing is known about their upstream regulation. The chaplins and SapB also share substantial surfactant characteristics, both having the capacity to dramatically reduce the surface tension of water (Tillotson et al., 1998; Claessen et al., 2003). In vivo, the chaplins are also proposed to contribute to an intricate ‘rodlet’ ultrastructure found on the surface of aerial hyphae and spores (Claessen et al., 2004), and in vitro, the chaplins are capable of self-assembling into insoluble filaments that are reminiscent of amyloid-like fibres (Claessen et al., 2003).

The presence of two separate pathways leading to the erection of aerial hyphae in S. coelicolor is suggested by the observation that many bld mutants are conditionally defective in their ability to raise aerial hyphae (Kelemen and Buttner, 1998; Nguyen et al., 2002; Chater and Horinouchi, 2003). That is, these bld mutants are blocked in aerial hyphae formation when grown on rich medium, but can undergo complete morphological differentiation when grown on minimal medium in which glucose has been substituted with an alternative carbon source such as mannitol or arabinose. SapB production is blocked in all bld mutants when grown on rich medium, with the exception of bldM and bldN (where SapB production is concomitant with delayed aerial hyphae formation; J.M. Willey, unpubl. data). In contrast, SapB is not made by any strain (wild type or bld mutant) when grown on minimal medium (Willey et al., 1991). It is thought that the elimination of the cohesive forces between water molecules covering filaments within a mycelium is prerequisite to the emergence of aerial hyphae, thus a different surface active molecule must fulfil the biosurfactant role in the SapB-independent developmental pathway. A strong candidate for this role is the chaplins.

Until now, little was known about the redundancy or functional interplay between SapB and the chaplins. In this study, we demonstrate that SapB and the chaplins overlap in their functional roles of enabling aerial hyphae formation, that the chaplins are key contributors to the ‘SapB-independent’ pathway of aerial hyphae formation on minimal medium, but that the chaplins and SapB are both required for normal aerial hyphae formation on rich medium.


Deletion of all chaplin genes results in a conditionally bald mutant phenotype

It was previously demonstrated that deletion of four or more chaplin genes results in delayed aerial hyphae formation (Claessen et al., 2003; Elliot et al., 2003a), and that loss of all eight chaplin genes in an S. coelicolor M145 genetic background results in severe attenuation of aerial hyphae formation when grown on MS (soy flour mannitol) medium (Claessen et al., 2004). Similarly, we found that deletion of seven or all eight chaplin genes in an M600 genetic background (strains J3149 and J3150 respectively) resulted in a severe morphological defect on MS medium. However, these strains were conditionally bald: they initially appeared bald on MS medium, and were capable of raising only a very sparse aerial mycelium on MS medium after 6–7 days (as opposed to the wild-type strain where aerial hyphae were detectable after only 2 days) (Fig. 1A). In contrast, on rich medium (R2YE), however, the chaplin mutants raised an aerial mycelium with similar growth kinetics to the wild-type strain, but with the overall abundance of hyphae significantly reduced (Fig. 1B). An identical phenotype was observed for chaplin mutants grown on R2YE lacking sucrose (data not shown). As observed previously for the M145 8×chp mutant (Claessen et al., 2004), the surface of the aerial hyphae of the M600 8×chp mutant was devoid of the rodlet ultrastructure characteristic of the wild type and, despite its lack of aerial hyphae on MS medium, was still capable of making occasional spore chains in prostrate surface hyphae (data not shown). While the primary role of the chaplins in aerial development is predicted to be a structural one, the reduced sporulation observed in the 8×chp mutant compared with the wild-type strain suggests there may be additional signalling events that are blocked in the chaplin mutant.

Figure 1.

ram mutants and chp mutants are differentially affected in their abilities to form aerial hyphae. Phenotypic comparison of ΔramR (J3287), ΔramCSAB (J3288), 8×chp mutant (J3150) and their congenic parental strain (M600) grown on (A) MS medium (7 days) or (B) R2YE medium (4 days). Note the obverse defects in aerial hyphae formation of the ram mutants versus the chp mutant.

Comparison of chp and ram null mutant phenotypes

While various ram mutant strains have been reported (Keijser et al., 2000; Nguyen et al., 2002; O'Connor et al., 2002), direct comparisons between chp and ram mutant phenotypes have not been possible because of their differing genetic backgrounds. To address this, we used the S. coelicolor M600 wild-type strain, and constructed two ram mutant strains for comparison with the full 8×chp mutant constructed in the same genetic background. In one strain we deleted the ramCSAB operon, which includes the SapB structural gene, ramS, and in the other we deleted ramR, which encodes the ramCSAB transcriptional activator. As expected, the SapB-deficient ramCSAB and ramR mutant strains (J3288 and J3287 respectively) had very similar phenotypes: when grown on rich medium (R2YE), they were initially bald, but formed a substantial aerial mycelium upon extended incubation, with J3288 (ΔramCSAB) developing an aerial mycelium slightly earlier than J3287 (ΔramR) (Fig. 1B). However, when grown on minimal-mannitol or MS agar, they were phenotypically indistinguishable from the wild-type strain (Fig. 1A). This is in contrast to the observations made by O'Connor et al. (2002), who report that a ramC knockout in an M145 background appears to be developmentally delayed on both rich and minimal-mannitol media, and Keijser et al. (2000), who describe a ramRAB mutant, also in an M145 background, that was morphologically identical to its wild-type parent regardless of media composition.

The full 8×chp mutant (J3150) exhibited an obverse conditional phenotype to that of the ram mutants: as described above, on MS medium, it raised a very sparse aerial mycelium only after prolonged incubation, but formed an aerial mycelium on rich medium (R2YE) with wild-type growth kinetics. Thus, J3150 had a conditional bald phenotype, but suppression of the mutant phenotype was contrary to that of the ram mutants and the classical bld mutants, which appear wild type when utilizing alternative carbon sources and are bald on rich medium.

Overexpression of ramR restores aerial hyphae formation to a chp mutant irrespective of growth media

Restoration of aerial hyphae formation in most bld mutants can be achieved by overexpressing ramR, which drives unregulated overproduction of SapB such that SapB is produced on minimal medium by wild-type strains (Nguyen et al., 2002). Given the conditional bald phenotype of the 8×chp mutant (J3150), we were interested to determine if ramR overexpression could also restore wild-type levels of aerial hyphae formation when this strain was grown on a variety of media/carbon sources. ramR was placed on a high-copy-number plasmid (pIJ486), and was introduced into J3150. Multiple copies of ramR in J3150 grown on rich medium (R2YE) restored the abundance of aerial hyphae to near wild-type levels (Fig. 2A). Interestingly, high-copy ramR also restored a modest level of aerial hyphae formation to J3150 when grown on MS medium (Fig. 2B). This suggested that the defects in aerial development exhibited by the chp mutant could be partially overcome by the overproduction of a molecule under RamR control, with SapB the most likely candidate.

Figure 2.

Overexpression of ramR induces aerial hyphae formation in J3150, the 8×chp mutant. ramR cloned into the multicopy Streptomyces vector pIJ486 increased aerial hyphae abundance of J3150 on both (A) rich R2YE medium (3 days) and (B) MS medium (7 days) relative to J3150 containing pIJ486 alone. Shown for comparison are the wild-type M600 and M600 containing pIJ486.

SapB application rescues both ram and chp mutant phenotypes

Additional evidence that SapB could mediate aerial hyphae formation in the absence of either endogenously synthesized SapB or the chaplins was obtained through extracellular complementation experiments using exogenous addition of SapB. It is well established that exogenously applied SapB restores the capacity of bld mutants to form aerial hyphae (Willey et al., 1991) by releasing undifferentiated hyphae from the aqueous colony surface (Tillotson et al., 1998). To determine whether the application of SapB to either a ramCSAB or a ramR mutant had the same effect, the ram mutants were spread on R2YE or MS media. SapB was applied to the surface of each mutant strain, and identical results were seen after 2 days of growth: zones of aerial differentiation (hyphae, and later spores) were visible in the spots where SapB had been applied, while the rest of the mycelium remained bald (Fig. 3A). These results are similar to those obtained when SapB is applied to a ramS null strain (Kodani et al., 2005). When a similar experiment was performed on MS medium to test the response of the chaplin-deficient strain (J3150), a robust aerial mycelium developed where SapB was applied, again after 2 days of growth (Fig. 3B). In all instances, resultant aerial hyphae gave rise to spore chains, as detected through coverslip impressions (Fig. 3C). This suggested that the 8×chp mutant lacked only sufficient surfactant molecules to raise aerial hyphae, and that the addition of SapB overcame both this deficiency, and any signalling block associated with knockout of all chaplin genes. Taken together with the ramR overexpression results, this suggests that SapB, and not some other molecule under RamR control, can drive aerial hyphae formation in the absence of chaplins.

Figure 3.

A and B. Exogenous application of purified SapB accelerates aerial hyphae formation of (A) J3288 (ΔramCSAB) and (B) J3150 (Δ8×chp) on MS medium after 3 days of growth. For each strain, 1, 2.5 and 5 μl aliquots of purified SapB (0.5 μg μl−1) were applied to the surface after 2 days of growth. DMSO (5 μl) was applied as a negative control. C. Coverslip impressions taken after exogenous application of SapB to J3150 show sporulating aerial hyphae.

There is a direct correlation between aerial hyphae formation and the production of the chaplins/SapB

Given the apparent ability of purified SapB to fulfil the role of biosurfactant in the absence of the chaplins, and considering the reciprocal phenotypes of the ram and chp mutant strains, we were curious to determine whether the conditional bald phenotypes of each mutant strain could be correlated with production of the remaining surfactant molecules (either the chaplins or SapB). Using the wild-type M600 strain and the congenic derivatives J3288 (ΔramCSAB), J3287 (ΔramR) and J3150 (8×chp mutant), we followed the production of the five short chaplins (ChpD–H) and SapB over 6 days using MALDI-ToF mass spectrometry and immunoblot analysis respectively. The wild-type strain demonstrated very similar growth kinetics on both rich (R2YE) and MS media, with aerial hyphae starting to appear 2 days after inoculation, and sporulation initiating after 3 days. Cell surface extracts were collected from strains at 24 h intervals, beginning 2 days after inoculation. In the wild-type strain, chaplin proteins could be detected after 3 days of growth, and were still present in sporulating cultures after 7 days on both rich (R2YE) and MS media (Table 1). We also followed the surface localization of the chaplins in the two ram mutant strains when they were grown on R2YE medium. After 2 days, few aerial hyphae could be detected, but at 3 days there was a substantial increase in aerial hyphae abundance, and this coincided with measurable chaplin protein production (Table 1). Thus, the timing of chaplin production in the ram mutant strains, as well as chaplin abundance, appeared to be identical to the wild-type strain on rich medium.

Table 1.  Correlation between aerial hyphae formation and chaplin/SapB detection for wild type (M600) and the chp and ram null mutant strains.
  1. Aerial hyphae: –, none; +, sparse; ++, abundant.

  2. +C, chaplins detected by MALDI-ToF mass spectrometry; +S, SapB detectable by Western blot; –S, no SapB detectable by Western blot.

 R2YE+++++++++Aerial hyphae
+C; +S+C; +S+C; +S+C; +SChp/SapB observations
 MS+++++++++Aerial hyphae
+C; –S–S–S+C; –SChp/SapB observations
 R2YE+++++Aerial hyphae
+S+S+SChp/SapB observations
 MS+Aerial hyphae
–S–S+SChp/SapB observations
 R2YE+++++++Aerial hyphae
+C+C+C+CChp/SapB observations
 MS++++++++Aerial hyphae Chp/SapB observations

Immunoblot analysis was performed to explore SapB production in the three mutant strains. Predictably, the ram null mutant strains failed to make SapB (data not shown), while J3150 (8×chp mutant) produced greater than wild-type levels of the peptide on rich medium (Fig. 4). Intriguingly, J3150 also produced SapB when grown for an extended period on MS medium, with SapB production concomitant with the appearance of sparse aerial hyphae (Fig. 4). The capacity of the 8×chp mutant to produce SapB on a medium where its synthesis is normally suppressed suggests that the repression of ramR expression seen in wild-type strains can be partially alleviated in a chp mutant background.

Figure 4.

Immunoblot analysis of SapB isolated from aerial hyphae and spores of wild-type M600 and 8×chp mutant J3150 after 3 or 6 days of growth on both R2YE and MS media. Unexpectedly, J3150 produced SapB after 6 days of growth on MS medium, in contrast to all previously examined S. coelicolor strains.

Conditional bld mutants make the chaplins on mannitol-containing media

The agents involved in the ‘SapB independent’ pathway of aerial hyphae formation have not been identified, but given the opposing ram versus chp mutant phenotypes, it was logical to consider the chaplins as likely candidates. To examine this possibility, we tested an assortment of bld mutants for chaplin production on rich (R2YE) and MS media. Based on previous RNA analyses (Elliot et al., 2003a), we knew that at least two of the chaplin genes were very poorly expressed in a number of bld mutants (bldN, bldM, bldH, bldG and bldC) when grown on rich medium. What was not known, however, was whether the chaplins were expressed when bld mutants differentiate on MS medium, and whether chaplin expression could be in any way correlated with the appearance of aerial hyphae in the conditional bld mutants.

Several bld mutants (in various genetic backgrounds) were examined for chaplin gene expression, including bldH, bldD and bldA (conditional mutants that sporulate when grown on MS medium), as well as bldB, which is unable to raise aerial hyphae on any medium. The mutants were grown on R2YE and MS media for 7 days, after which cell surface proteins were isolated, and the presence or absence of chaplin proteins was assessed. As shown in Fig. 5A and B, peaks corresponding to the chaplin proteins were negligible in the extract from the R2YE-grown culture of bldH, but were readily detectable in the extract from the MS-grown culture. Identical results were obtained for bldD and bldA (data not shown). In contrast, chaplins were not detectable in any extract isolated from a bldB mutant, irrespective of the growth medium (Fig. 5C and D), showing a correlation between aerial mycelium formation and chaplin production.

Figure 5.

MALDI-ToF mass spectrometry for bldH and bldB. MALDI-ToF mass spectra of the short chaplins in ΔbldH (A and B) and ΔbldB (C and D) mutant strains grown on MS (A and C) and R2YE (B and D) media for 7 days. Chaplins D, E, G and H are indicated. The x-axis shows the mass (m)/charge (z), where z = 1. An identical scale was used for comparing the results of each strain on the two different media; however, to emphasize the lack of chaplin production in the bldB mutant, a higher scale was used than for examination of chaplin production in the bldH mutant.

A complete chp/ram mutant is bald on all media

Given the contrasting phenotypes of the chp and ram mutant strains on rich and MS media, the ‘SapB-dependent’ and ‘SapB-independent’ pathways for aerial hyphae formation, and the similar surfactant capabilities of SapB and the chaplins, we wanted to determine whether deletion of both the chp and the ram genes would result in a mutant strain unable to raise an aerial mycelium on any medium (a bldB-like unconditional phenotype). A full chp/ram mutant was created using a viomycin resistance cassette to replace either ramR (J3289) or the ramCSAB (J3290) gene cluster in the 8×chp mutant background. Immunoblot and MALDI-ToF analyses confirmed that these strains produced neither SapB nor the chaplins (data not shown). The chp/ram null mutant strains failed to raise any aerial hyphae on MS medium (a phenotype similar to that of the 8×chp mutant), and were only capable of raising an extremely sparse aerial mycelium on rich R2YE medium after extended incubation (Fig. 6). These sparse aerial hyphae were undetectable by coverslip impressions, and unlike typical aerial structures, were highly hydrophilic (data not shown). These results suggest that aerial hyphae formation requires the activity of either the chaplins (MS medium) or SapB and the chaplins (rich medium), and in the absence of both molecules, a robust aerial mycelium fails to develop on all media tested.

Figure 6.

Phenotype of a full chp/ram mutant strain (J3289), relative to J3150 (8×chp mutant), J3287 (ramR null mutant) and M600 (wild type). The full chp/ram mutant makes (A) no detectable aerial hyphae on MS medium after 7 days, and (B) few hyphae on R2YE medium after 4 days.


Surfactant requirement is medium specific

Many bld mutants are defective in both aerial hyphae formation and SapB production when grown on rich medium, but are able to raise an aerial mycelium in the absence of SapB on minimal media. This has brought about the idea of two separate pathways leading to the erection of aerial hyphae: a bld gene/SapB-dependent pathway and a bld gene/SapB-independent pathway (Kelemen and Buttner, 1998; Nguyen et al., 2002; Chater and Horinouchi, 2003). Both SapB (Tillotson et al., 1998) and the chaplins (Claessen et al., 2003; Elliot et al., 2003a) are believed to act as surfactants during aerial morphogenesis, reducing the surface tension at air–aqueous interfaces, enabling nascent hyphae to grow into the air. The chaplins are also localized to the surface of aerial hyphae and spores where they could act to prevent the desiccation of aerial filaments following emergence from the aqueous vegetative environment. This would be consistent with a model for chaplin activity put forward by Chater and Chandra (2006), where they propose a role for the chaplins in maintaining an aqueous environment for cellular development during aerial hyphal growth. The observation that ram mutants (and thus SapB mutants) are not fully bald but are merely delayed in aerial hyphae formation on rich medium indicates that SapB is not the only molecule capable of facilitating aerial hyphae development under these conditions. Our results suggest that growth of the aerial mycelium on rich medium is dependent on both the chaplins and SapB. In contrast, aerial mycelium formation during SapB-independent morphogenesis on minimal medium appears to require only the chaplins, where they were produced by both conditional bld mutants and wild-type strains. The direct correlation between chaplin production on minimal media and aerial hyphae formation is strengthened by the observation that the non-conditional bld mutant, bldB, was blocked in chaplin production under all conditions examined.

Regulating the production of surfactant molecules

The obverse phenotypes of ram and chp mutants, and the differing expression profiles of SapB and the chaplins, reveal a sophisticated regulatory network governing surfactant production in S. coelicolor. These regulatory differences have resulted in two separate, but not completely independent pathways of aerial hyphae formation: the SapB-dependent pathway of aerial hyphae formation on rich (R2YE) medium, to which the chaplins contribute, and the chaplin-dependent pathway (or SapB-independent pathway) of aerial hyphae formation on MS medium, to which SapB makes no contribution (Fig. 7). The production of both SapB and the chaplins on rich medium requires the bld genes; however, it appears that chaplin expression on MS medium is largely bld gene independent, as conditional bld mutants produce significant levels of chaplin proteins on MS medium, and are consequently able to raise a substantial aerial mycelium.

Figure 7.

Regulation of aerial hyphae formation in S. coelicolor. Two separate pathways allow for aerial hyphae formation in S. coelicolor. On rich medium, aerial hyphae formation requires the activity of the bld genes, as well as that of SapB and the chaplins, while on minimal medium, only the chaplins are essential, with neither SapB nor the majority of the bld genes (with the exception of bldB) being necessary.

In S. coelicolor, the ramCSAB operon is positively controlled by RamR (Keijser et al., 2002; Nguyen et al., 2002; O'Connor et al., 2002; O'Connor and Nodwell, 2005). In Streptomyces griseus, the orthologous amfTSAB operon (Ueda et al., 1993) is similarly governed by the RamR orthologue, AmfR, but is negatively regulated by BldD (Ueda et al., 2005). ramCSAB regulation by BldD does not, however, appear to be conserved in S. coelicolor (B.K. Leskiw and J.R. Nodwell, pers. comm.). While ramR gene expression is obviously sensitive to media composition (repressed on minimal and MS media), the direct regulators of this phenomenon in S. coelicolor are unknown. Similarly, there is currently nothing known about the direct regulators of chaplin gene expression.

One particularly intriguing observation is that the 8×chp mutant ultimately produces SapB on MS medium, enabling it to raise a sparse aerial mycelium. As wild-type strains do not produce SapB on MS medium, it is clear that the repression of ramR expression seen on MS medium in wild-type strains is alleviated in the 8×chp mutant. This may represent a compensatory response to the absence of normal chaplin/surfactant production under these conditions. Similarly, the production of SapB in the 8×chp mutant on rich (R2YE) medium is increased relative to the wild type, likely contributing to the formation of the observed aerial mycelium. Gaining insight into the mechanism by which the normal regulation of ramR can be overridden under these circumstances presents an additional challenge for the future.

Multiple surfactants in the streptomycetes and filamentous fungi

It is apparent from this work that SapB and the chaplins fulfil very similar roles in Streptomyces development, and that contributions by both are essential for aerial hyphae formation under particular conditions. Their importance is underscored by the conservation of the ram and chp genes in other Streptomyces species (Elliot et al., 2003a; Willey et al., 1991; 2006). A question that remains unanswered is why streptomycetes have evolved and maintained two distinct, yet functionally overlapping surfactant systems, with one used in a nutrient-poor environment, and both used in a nutrient-rich environment. Sporulation is the primary means of reproduction and dispersal for Streptomyces, and biosurfactant production is necessary for this process to occur efficiently in solid culture. Given that soil is a highly variable environment, it seems likely that streptomycetes have evolved the capacity to produce surfactant molecules that are optimized for different environmental conditions.

It is conceivable that, in addition to SapB and the chaplins, there are as yet undiscovered surfactants that make a minor contribution to aerial hyphae formation. This possibility is strengthened by San Paolo et al. (2006), who identified a new RamR-regulated gene cluster that, when overexpressed, stimulates aerial hyphae formation in a SapB-independent, and likely chaplin-independent, manner.

Like Streptomyces, the filamentous fungi raise aerial structures with the aid of surfactant molecules termed hydrophobins (as reviewed by Wösten, 2001). A number of fungi have multiple hydrophobin genes; however, the resulting proteins are usually differentially expressed and have distinct functions, rather than overlapping functions as seen for SapB and the chaplins. The production of partially redundant surfactant species is not unprecedented, however, as the fungus Schizophyllum commune produces several different hydrophobins (Wessels et al., 1991), as well as an unrelated protein that can mediate the emergence of aerial hyphae (Lugones et al., 2004). In addition, the plant pathogen Ustilago maydis raises aerial hyphae through the activity of proteins unrelated to the hydrophobins, termed ‘repellants’ (Wösten et al., 1996). This implies that, like streptomycetes, filamentous fungi can use different surfactant molecules to raise aerial structures.

Surfactants and aerial development

If the importance of SapB and the chaplins in the erection of aerial hyphae in S. coelicolor were exclusively due to their surfactant properties, then it might be expected that any surfactant could substitute for them and fulfil this role. Supporting this prediction is the observation that the S. tendae lanthionine-containing peptide, SapT (Kodani et al., 2005) and the hydrophobin protein SC3 from the fungus S. commune (Tillotson et al., 1998) are able to stimulate aerial hyphae formation in S. coelicolor bld mutants when applied exogenously. In contrast to this, however, is the finding that the Bacillus subtilis surfactant peptide surfactin (Richter et al., 1998), which is required for development of aerial structures in B. subtilis, strongly inhibits aerial hyphae formation when applied to S. coelicolor, without apparent detriment to vegetative filaments (Straight et al., 2006). Intriguingly, the surfactin-mediated block in aerial development is alleviated by the addition of either purified SapB or the chaplins (Straight et al., 2006). These results demonstrate that microbial surfactants are not necessarily interchangeable, and that they exhibit specificity in their mode of action. They also suggest that in natural microbial communities, such compounds may have functions beyond their obvious roles as surface active molecules. We may therefore not fully understand the roles of SapB and the chaplins until they are analysed in more natural, complex microbial populations.

Experimental procedures

Bacterial strains, plasmids and growth conditions

Streptomyces strains (Table 2) were cultured at 30°C on R2YE, MS, minimal and DNA agar media, or in a 1:1 mixture of tryptone soya broth (TSB) and yeast extract-malt extract (YEME) supplemented with MgCl2 and glycine liquid media (Kieser et al., 2000). Escherichia coli strain BW21153/pIJ790 was grown at 30°C and was used for PCR-targeted gene disruptions (Datsenko and Wanner, 2000; Gust et al., 2003), while DH5α (used for plasmid construction and routine subcloning) and ET12567/pUZ8002 (used for generation of methylation-free DNA; Kieser et al., 2000) were grown at 37°C. Plasmids used were pGEM-T (Promega), pIJ2925 (Bierman et al., 1992), pIJ486 (Ward et al., 1986) and pKN22 (Nguyen et al., 2002), which is a derivative of pHM11a (Motamedi et al., 1995) containing ramR under the control of the constitutive ermE promoter.

Table 2. Streptomyces coelicolor strains.
M600SCP1 SCP2Chakraburtty and Bibb (1997)
J1501hisAl uraAl strAl Pgl SCP1 SCP2Chater et al. (1982)
J3150M600 ΔchpADΔchpCH::aadAΔchpBΔchpFΔchpGΔchpE::aprThis work
J3287M600 ΔramR::vphThis work
J3288M600 ΔramCSAB::vphThis work
J3289M600 ΔchpADΔchpCH::aadAΔchpBΔchpFΔchpGΔchpE::aprΔramR::vphThis work
J3290M600 ΔchpADΔchpCH::aadAΔchpBΔchpFΔchpGΔchpE::aprΔramCSAB::vphThis work
C109J1501 bldHChampness (1988)
J1700J1501 bldA39Piret and Chater (1985)
ΔbldDM600 bldD::aprElliot et al. (2003b)
J669bldB43 mthB2 cysD18 agaA7 SCP1NF SCP2*Merrick (1976)

Construction of chp and/or ram null mutants

Null mutants were constructed as described in Elliot et al. (2003a) according to the methods of Gust et al. (2003). The construction of a sevenfold (7×) and an eightfold (8×– complete chaplin knockout) chp mutant strain followed on from the fivefold chp mutant (J3145) created by Elliot et al. (2003a), where chpAD was replaced with an apramycin resistance cassette which was then removed to generate an in-frame deletion (and to allow recycling of the apramycin resistance cassette), chpCH was replaced by a spectinomycin/streptomycin resistance cassette, and chpB was replaced with an apramycin resistance cassette. The apramycin resistance cassette was then removed from chpB by a second targeting event, to generate J3145A. As chpF and chpG are separated by only six genes, they were both knocked out of cosmid SCC61A (chpF was first knocked out using an apramycin resistance cassette, which was then removed and used to replace chpG), and both mutations were introduced simultaneously into the fivefold mutant, creating J3149. The apramycin resistance cassette was then removed from chpG (to create J3149A, the 7× mutant), and was used to replace chpE on cosmid I5, which was then introduced into the 7× mutant by conjugation to create the 8× mutant (J3150).

Null mutants in the ram gene cluster were constructed by replacing the entire ramR gene or ramCSAB gene cluster with a cassette carrying the viomycin resistance gene (vph) and oriT from RK2 by PCR targeting (Gust et al., 2003). Cosmid SC6G3 was introduced into E. coli BW21153 carrying plasmid pIJ790. ramR and ramCSAB were disrupted separately by electroporation of the cells with the PCR-amplified vph-oriT cassette, generated using primers ramR Fwd and ramR Rev, or ramC Fwd and ramB Rev (Table 3). The resulting cosmid, carrying the viomycin resistance cassette in place of either ramR or ramCSAB, was then introduced into the methylation-deficient E. coli ET12567 carrying pUZ8002 before being transferred into S. coelicolor M600 or J3150 by conjugation (Kieser et al., 2000). Viomycin-resistant, kanamycin-sensitive exconjugants were identified, isolated, and confirmed by PCR and Southern blot hybridization (see Table 2). Complementation of the ramR mutant was achieved using the plasmid pTO11 (O'Connor et al., 2002), while the ramCSAB mutant was complemented by the introduction of the ramCSAB gene cluster on plasmid pIJ6902 (Huang et al., 2005). The ramCSAB cluster was excised from cosmid 6G3 using EcoRI and MscI, before being cloned into pBluescript KS (Stratagene) digested with EcoRI and EcoRV. The cluster was then excised using EcoRI and XhoI, and was cloned into pIJ6902 digested with the same enzymes.

Table 3.  Oligonucleotide sequences used in this study.
Oligo nameSequence (5′ to 3′)

Construction of the ramR overexpression plasmid

ramR was PCR amplified using S. coelicolor M600 chromosomal DNA as template, Expand High Fidelity DNA polymerase, and primers ramR for and ramR down (Table 3). The resulting product was introduced into pGEM-T (Promega), and ramR clones were verified by digesting with SphI and SpeI. Using these enzymes, ramR was excised, cloned into pIJ2925 digested with SphI and XbaI, and the correct insertion was confirmed by PCR. pIJ2925+ramR was then passaged through E. coli ET12567, to generate methylation-free DNA. Non-methylated ramR was excised as an EcoRI and HindIII fragment and was cloned into EcoRI- and HindIII-digested and dephosphorylated pIJ486 before the entire ligation reaction was introduced into S. coelicolor J3150 by protoplast transformation (Kieser et al., 2000). Thiostrepton was used for positive transformant selection. Correct constructs were verified by PCR using primers 486 oligo and Fd univ (Table 3). As a negative control, pIJ486 was also introduced into wild-type S. coelicolor M600 and the 8×chp mutant J3150 by protoplast transformation, and thiostrepton-resistant transformants were selected.

Purification of SapB and immunoblot analysis

Extraction of cellular proteins from S. coelicolor strains and the detection of SapB production by immunoblot were performed as described previously (Guijarro et al., 1988). Purified SapB, used as a positive control for the immunoblots and for the exogenous application to lawns of J3150, J3287 and J3288, was isolated from spores of S. coelicolor J1501/pKN22, as reported in Kodani et al. (2004).

Purification and analysis of the chaplin proteins

The chaplin proteins were isolated from cell wall extracts as described previously (Elliot et al., 2003a), with several modifications. S. coelicolor wild-type and mutant strains were grown on cellophane discs on the surface of MS or R2YE solid media. After 2–7 days, cells were scraped from the surface of the cellophane, and were re-suspended in 10 ml of Tris buffer (pH 8.0). The suspension was sonicated on ice for 3–15 s before spinning at 8000 g for 10 min. The supernatant was removed, the remaining pellet was re-suspended in 3 ml of 2% SDS and was heated to 100°C for 10 min. After cooling briefly, the suspension was centrifuged at 8000 g for 10 min, and the SDS-heat treatment was repeated. The suspension was transferred to 2 ml Eppendorf tubes, and was spun at 16 000 g before discarding the supernatant and flash-freezing the pellet in an ethanol-dry ice bath and drying it by rotary evaporation. Trifluoroacetic acid (400–800 μl) was added to the dried pellet, and the soluble fraction removed and dried using a stream of air. MALDI-ToF mass spectrometry conditions were as described in Elliot et al. (2003a).


The authors would like to thank Ms Shuko Kodani and Mr Francois Blues for their technical assistance, and Dr Justin Nodwell and Stephanie Au-Young for their generous gift of complementation plasmids. This work was supported by the Canada Research Chairs programme (to M.A.E.), the Canadian Institutes of Health Research (to M.A.E., Grant No. MOP-77553), the National Institutes of Health (to J.M.W., Grant No. GM069398-01) and the Biotechnology and Biological Sciences Research Council of the United Kingdom (to M.J.B., Grant No. 208/EGH16080).