Correspondence: Ángel Manteca, Departamento de Biología Funcional and IUBA, Facultad de Medicina, Universidad de Oviedo, 33006 Oviedo, Spain. Tel.: + 34985103555; fax: +34 985 103534; e-mail: firstname.lastname@example.org
Streptomycetes comprise very important industrial bacteria, producing two-thirds of all clinically relevant secondary metabolites. They are mycelial microorganisms with complex developmental cycles that include programmed cell death (PCD) and sporulation. Industrial fermentations are usually performed in liquid cultures (large bioreactors), conditions in which Streptomyces strains generally do not sporulate, and it was traditionally assumed that there was no differentiation. In this work, we review the current knowledge on Streptomyces pre-sporulation stages of Streptomyces differentiation.
Streptomycetes are gram-positive, mycelium-forming, soil microorganisms that play important roles in mineralization processes in nature. They have great socio-economic relevance, as they produce several clinically relevant secondary metabolites (antibiotics, antitumorals, immunosuppressants, etc.) (Hopwood, 2007). Streptomycetes have complex developmental cycles that resemble filamentous fungi, forming hyphae and mycelia. They also have sporulation and programmed cell death (PCD) processes and are considered multicellular prokaryotic models.
Differentiation and development of Streptomyces in solid cultures
The traditional Streptomyces developmental cycle mainly focused on the sporulation phases occurring in solid cultures. After spore germination, a completely viable vegetative mycelium (substrate) grows on the surface and inside agar until it differentiates to a reproductive (aerial) mycelium that grows into the air, producing spores at the end of the cycle (reviewed in Flärdh & Buttner, 2009). This developmental model has been refined with respect to the stages preceding aerial mycelium formation and sporulation (Fig. 1). A young, compartmentalized mycelium (MI) was reported to die early on, following a highly ordered sequence (Manteca et al., 2005, 2006a). Subsequently, the viable segments of this mycelium differentiate into a multinucleated second mycelium (MII). MII grows inside the culture medium (substrate mycelium) until it starts to express hydrophobic covers and grows into the air (aerial mycelium) and ends by forming spores (Manteca et al., 2007). Prior to sporulation, there is a second round of PCD affecting substrate and aerial mycelium (Wildermuth, 1970; Mendez et al., 1985; Miguelez et al., 1999). MI, MII, and PCD were mainly described in Streptomyces antibioticus ATCC11891 (Miguelez et al., 1999; Manteca et al., 2005) or Streptomyces coelicolor M145 (Manteca et al., 2007). However, the existence of these developmental stages can be considered general to the Streptomyces genus, as they were observed in all the streptomycetes analysed: Streptomyces griseus IFO 13350, Streptomyces avermitillis MA-4680, Streptomyces cinereoruber ATCC19740, as well as hundreds of unclassified streptomycetes, examined during the screening experiments aimed at discovering novel secondary metabolites (Yagüe P, Genilloud O, Manteca A, unpublished results).
Spore germination constitutes the first step of Streptomyces development (Fig. 2). However, the mechanisms activating germination remain somewhat vague. Spore germination comprises a sucession of distinctive steps, which were organized nicely by Hardisson et al. (1978) into three stages: darkening, swelling, and germ tube emergence. Darkening only required exogenous divalent cations (Ca2+, Mg2+ or Fe2+) and spore energy reserves. Calcium was reported to accumulate in the spore covers and be released during germination (Eaton & Ensign, 1980; Salas et al., 1983). Wang et al. (2008) demonstrated that calcium regulation could be mediated, at least in part, by cabC, a gene encoding an EF-hand calcium-binding protein. Trehalose was demonstrated to be consumed during the early stages of germination (Hey-Ferguson et al., 1973; McBride & Ensign, 1987). The second stage, swelling, needed an exogenous carbon source, and the last stage, germ tube emergence, required additional carbon and nitrogen sources.
Spore germination is highly regulated and can be externally modified. Hirsch & Ensign (1976) reported that the latency preceding germination of S. viridochromogenes spores was eliminated by gentle heat shock, a method that is routinely used to synchronize spore germination in Streptomyces (Kieser et al., 2000). Guijarro et al. (1983) revealed the existence of a protein fraction that rapidly degrades during germination and that might be regulating this process. Mikulík et al. (1984) demonstrated that RNA and protein synthesis began in the first 5 min following spore inoculation, a fact that was later confirmed by Strakova et al. (2013). Ribosomes were described as being complex, with melanine-type pigments forming insoluble aggregates, rendering them inactive in the dormant spores (Mikulík et al., 1984). Haiser et al. (2009) demonstrated the importance of cell wall hydrolases in both spore formation and spore germination. The existence of germination inhibitors excreted by germinating spores was discovered in Streptomyces viridochromogenes by Grund & Ensign (1985) and its chemical nature was subsequently characterized by Petersen et al. (1993). These inhibitors were also identified in S. coelicolor (Song et al., 2006). Cyclic AMP is involved in the regulation of germination (Süsstrunk et al., 1998); this regulation is mediated, at least partially, by the cyclic AMP receptor protein (Crp) (Derouaux et al., 2004; Piette et al., 2005). NepA has been described as a structural cell wall protein involved in maintaining spore dormancy in S. coelicolor (de Jong et al., 2009). Noens et al. (2007) identified SsgA as a protein marking cell wall sites where germination takes place.
Overall, important information has already been obtained concerning Streptomyces germination. However, there is still much to discover to fully understand the biochemical pathways regulating this important process.
Primary compartmentalized mycelium (MI)
MI is completely compartmentalized and is different from substrate and aerial mycelia, which are multinucleated (Manteca et al., 2005). Compartmentalization of this mycelium was studied by fluorescence microscopy, using membrane (FM 4-64, Cell Mask) and cell wall (WGA, vancomycin) fluorescent stains, as well as electron microscopy (Manteca et al., 2005; Manteca & Sánchez, 2009) (Fig. 1). MI septa membranes did not generally display thick cell walls; moreover, they were curved, probably due to the osmotic cellular pressure which could not be supported by their thin cell walls (Manteca et al., 2005; Manteca & Sánchez, 2009) (Fig. 1b). The function of MI thin septa remains unknown. They may facilitate intercellular communication inside Streptomyces hyphae.
Jakimowicz & van Wezel (2012) described the existence of two different septa in Streptomyces: substrate and aerial mycelia septa and spore septa. The formation of the two types of septa is regulated by different mechanisms (Willemse et al., 2011). MI septa would constitute a third type of thin septa that is structurally different from substrate/aerial and sporulation septa. This would make Streptomyces a very unusual organism, with three distinct septa associated with different developmental stages. Mechanisms regulating MI septa formation have yet to be discovered. FtsZ is one of the key proteins involved in cell division in bacteria. FtsZ was proven to participate in the formation of substrate/aerial/sporulation septa and its mutation gives rise to a non-sporulating syncytial mycelium having no septa (McCormick et al., 1994). Surprisingly, FtsZ mutant tolerated strong mechanical breakage (McCormick et al., 1994); the reason for this resistance remains unknown. One possibility could be the existence of some kind of septa in the FtsZ mutant similar to those with thin cell walls present in MI.
Secondary multinucleated mycelium (MII)
MI mycelial segments which remained viable after the first round of PCD started to grow as multinucleated hyphae (MII), whereas dead segments were progressively dismantled (Manteca et al., 2006b) (Fig. 1). Cellular debris generated by MI dead cells was present in the extracellular medium (cytosolic proteins; diaminopimelic acid, d-alanine, and other amino acids originating from cell wall degradation; nucleolytic activities; DNA or RNA fragments, etc.) (Manteca et al., 2006a). They were also observed under the electron microscope: García (1995) reported that substrate mycelium in S. antibioticus was ‘embedded among an intercellular material’ and Manteca et al. (2005) described the complete disorganization of MI dying cells. MII growth is completely viable on the surface and inside agar (substrate mycelium) until it undergoes a new round of PCD (Wildermuth, 1970; Mendez et al., 1985; Miguelez et al., 1999; Manteca et al., 2006a). The remaining viable MII hyphae start to form hydrophobic covers (chaplin-rodlin layer) (reviewed in Claessen et al., 2006) and grow into the air (aerial mycelium). Substrate mycelium was considered the vegetative mycelium, whereas aerial mycelium hyphae were considered specialized hyphae destined to sporulate (Chater, 1984). Aerial mycelium would re-use nutrients released by the substrate mycelium during the second round of PCD (a kind of cannibalism) (Mendez et al., 1985) and antibiotics would be produced by the substrate and/or aerial mycelium to prevent competition with other microorganisms during sporulation.
Substrate and aerial mycelia (MII) are multinucleated (reviewed in Jakimowicz & van Wezel, 2012). The existence of these multinucleated hyphae is very unusual and its biological relevance has yet to be revealed. Other well characterized filamentous bacteria, such as Cyanobacteria, are not multinucleated (reviewed in Singh & Montgomery, 2011). The obvious advantage to being multinucleated would be to facilitate the distribution of nutrients and biochemical signals, but with a very important risk in nature, as any damage would spread to the whole colony. Other mycelial microorganisms, such as fungi, also have multinucleated mycelia at temporary specific stages which are usually related with reproduction (reviewed in Glass & Kaneko, 2003). As discussed below, when Streptomyces development was analysed in soils, MI was the predominant mycelium and MII was a transitory phase preceding sporulation, which suggests that MI may be the true vegetative mycelium in nature.
The transition from substrate to aerial mycelium was extensively studied (Fig. 2). Streptomyces coelicolor mutant strains defective in different stages of hydrophobic cover formation (aerial mycelium) were used for the genetic and biochemical analysis of Streptomyces differentiation. Bald mutants (defective in aerial growth) regulate the ‘sky-pathway’, which activates the expression of genes related with hydrophobic cover formation (Rdls Chps, SapB) (Fig. 2). Elegant revisions of the state of the art of these developmental pathways have already been published (Claessen et al., 2006; McCormick & Flärdh, 2012).
The mechanisms regulating the absence of septa in MII or their presence in MI remain unknown (Fig. 2). Some authors have described the potential of substrate hyphae to septate and form spores prior to aerial mycelium differentiation, a feature known as ‘ectopic sporulation’ or ‘de-programmed sporulation’. Kelemen et al. (1995) described ectopic sporulation in a mutant strain lacking a DNA fragment near glkA in S. coelicolor. Kwak & Kendrick (1996) and Ohnishi et al. (2002) have described the same process in class III bald and NP4 mutants of S. griseus. Sporulation was also reported in substrate mycelium of wild Streptomyces carpinensis strain (Miguelez et al., 1997). Ohnishi et al. (2002) postulated the existence of unknown, specific mechanisms inhibiting septa formation in substrate hyphae. Manteca et al. (2010a, b) analysed differences between MI and MII proteomes, identifying several putative regulatory proteins differentially expressed in both types of mycelia. These experiments were recently extended to MI and MII transcriptomes (Yagüe et al., 2013). Further work will be necessary to characterize the biochemical pathways controlling the transition from MI to MII (Fig. 2).
Compartmentalization of tip ends of aerial mycelium and sporulation
The last stage of Streptomyces development in solid cultures corresponds to hypha septation and spore formation (Fig. 1). Streptomyces whi mutants defective in different stages of sporulation were used for the genetic and biochemical analyses of these developmental stages (Fig. 2). Sporulation is beyond the scope of this review. Revisions of the state of the art of genes and proteins regulating sporulation already exist (Claessen et al., 2006; Flärdh & Buttner, 2009; Jakimowicz & van Wezel, 2012; McCormick & Flärdh, 2012).
Differentiation and development of Streptomyces in liquid cultures
Most Streptomyces species do not sporulate in liquid cultures and it was widely accepted that no morphological differentiation took place in these conditions. Secondary metabolites would be produced by the substrate mycelium at the stationary phase after a transient growth arrest (Granozzi et al., 1990; Neumann et al., 1996; Novotna et al., 2003; Zhou et al., 2005; Chouayekh et al., 2007). Despite that, sporulation was reported in liquid cultures for several streptomycetes, such as Streptomyces venezuelae (Glazebrook et al., 1990), S. griseus (Kendrick & Ensign, 1983), Streptomyces chrysomallus (Kuimova & Soina, 1981), S. antibioticus ETHZ7451 (Novella et al., 1992), Streptomyces albidoflavus SMF301 (Rho & Lee, 1994), or Streptomyces brasiliensis (Rueda et al., 2001). Sporulation was also seen to be activated in several Streptomyces species under nutritional downshifts, including the model strain S. coelicolor (Koepsel & Ensign, 1984; Daza et al., 1989), and was also observed in several streptomycetes liquid cultures during the screening experiments aimed at discovering novel secondary metabolites (Yagüe P, Genilloud O, Manteca A, unpublished results).
New aspects regarding Streptomyces development (MI, MII, PCD) in solid cultures were extended to liquid cultivation (Manteca et al., 2008) (Fig. 1a). Similar to solid cultures, there was a young, compartmentalized mycelium (MI) that differentiated to a multinucleated mycelium (MII). The MII emergence was preceded by a transient growth arrest, which was the consequence of MI PCD. The only mycelial phases present in liquid were MI and MII without hydrophobic layers (Fig. 1). It was demonstrated that MII is the antibiotic-producing mycelium. This was the first time that antibiotic production was associated with differentiation in liquid cultures (Manteca et al., 2008). The lifespan of MI in liquid cultures was longer than in solid media (around 17 h in solid vs. 48 h in liquid) (Manteca et al., 2007, 2008). MI compartmentalization correlated well with the traditionally accepted existence of a specific phase at the beginning of the development – ‘the middle of the exponential phase’ in which protoplasts could be formed in Streptomyces liquid cultures (Okanishi et al., 1974). Protoplast formation by MII multinucleated mycelium was almost non-existent, a feature that can, in fact, be used to fractionate MI and MII mycelia (Manteca et al., 2010a).
Proteomic (Manteca et al., 2010b) and transcriptomic (Yagüe et al., 2013) analyses demonstrated that differentiation in liquid was much more similar to solid cultures than might be expected within the context of the classical Streptomyces developmental model. Proteins and transcripts involved in primary metabolism were up-regulated in MI, whereas proteins and genes involved in secondary metabolite biosynthesis were up-regulated in MII. The most remarkable differences between MII from solid and liquid cultures involved proteins regulating the hydrophobic cover formation and sporulation (Manteca et al., 2008, 2010b). Differentiation of MII after mycelia growth arrest is not enough to guarantee secondary metabolite production, as it can also be regulated by environmental signals, including components of the culture medium, such as nitrogen (Aharonowitz, 1980), carbon (Sánchez et al., 2010) and phosphate (Chouayekh & Virolle, 2002; Martín, 2004).
Streptomyces development in conditions resembling nature (soils)
The significance of the first compartmentalized mycelium was obscured by its short lifespan in usual laboratory culture conditions (Manteca et al., 2005, 2008). This might be attributable to the relatively high cell densities attained in laboratory culture conditions, which provoked massive cell death, differentiation, and sporulation. Natural growth conditions imply discontinuous growth and limited colony development (Williams, 1985). When Streptomyces development was analysed in conditions resembling nature (soils inoculated with poor spore inocula), a new developmental cycle emerged in which MI was the predominant mycelium (Manteca & Sánchez, 2009) (Fig. 3). Spore germination was a very slow, non-synchronous process that commenced at about 7 days and lasted for at least 21 days. The mycelium did not clump into dense pellets and remained in the MI compartmentalized mycelium phase for a long time. Even after 1 month of incubation, PCD, MII or sporulation were not detected. It is clear that in nature, cell death and sporulation must take place at the end of the long vegetative phase (Wellington et al., 1990; Anukool et al., 2004) when the nutrient imbalance gives rise to bacterial differentiation. As already commented above, the absence of compartmentalization in the vegetative Streptomyces mycelium (substrate) was unique in filamentous bacteria and difficult to understand due to the fragility of a multinucleated mycelium in nature. If we consider development in conditions resembling nature, compartmentalized MI would in fact be the dominant stage and multinucleated MII would be a transient antibiotic-producing structure, facilitating nucleic acid division and preceding sporulation (Fig. 3).
Streptomyces programmed cell death
Bacterial PCD can be defined as any type of genetically controlled cell dismantling involving the activation of specific cell death transducers, regulators, and effectors (Engelberg-Kulka et al., 2006). PCD was described in bacteria from different taxa, such as Bacillus and Escherichia coli (Engelberg-Kulka et al., 2006), Anabaena (Ning et al., 2002), Caulobacter (Hochman, 1997; Bos et al., 2012), Streptococcus (Guiral et al., 2005), Staphylococcus (Chatterjee et al., 2010), and Myxobacteria (Søgaard-Andersen & Yang, 2008). With few exceptions, such as the toxin-antitoxin modules from E. coli, the competence-sporulation processes from Bacillus subtilis (both reviewed in Engelberg-Kulka et al., 2006) and the competence processes of Streptococcus pneumonia (Guiral et al., 2005), the biochemical pathways controlling bacterial PCD, as well as the biological role of this process, are poorly understood (reviewed in Engelberg-Kulka et al., 2006). These three, well characterized bacterial PCDs are regulated in distinct ways, and there is no general biochemical model applicable to all bacterial PCD.
Miguelez et al. (1999) and Manteca et al. (2006a) demonstrated that Streptomyces death phenomena associated with development present the characteristics of programmed cell death. Biochemical parameters, such as the degradation of the cell wall and membrane, DNA/RNA degradation, corroborated the existence of a highly regulated, active cellular suicide that entails the activation of specific degradative enzymes (Manteca et al., 2006a). Among these enzymes there was a precursor of sequence non-specific nucleases involved in massive chromosomal degradation (Nicieza et al., 1999) and the sequence-specific nuclease (endoG) that produced chromosomal bands analogous to those that appear in the programmed cell death of eukaryotic cells (apoptosis) (Cal et al., 1996; Samejima & Earnshaw, 2005). A proteomic analysis revealed that PCD in S. coelicolor was accompanied by the appearance of enzymes involved in the degradation of cellular macromolecules, regulatory proteins, and stress-induced proteins (Manteca et al., 2006b). Sevillano et al. (2012), identified the first functional toxin-antitoxin system in Streptomyces that could be related to PCD. Bacteria having complex life cycles (streptomycetes, cyanobacteria, etc.) harbour several eukaryotic signalling domains and are considered to be the evolutive origin of these domains (Zhang, 1996; Aravind et al., 1999; Koonin & Aravind, 2002; Petrickova & Petricek, 2003). In all, 244 genes (3% of all Streptomyces ORFs) harbour these kinds of domains (Table 1). Further work will be necessary to characterize the biochemical regulation of Streptomyces PCD and to determine whether the genes described above, including those encoding for proteins harbouring eukaryotic type signalling domains, are involved in its regulation.
Table 1. Genes harbouring eukaryotic type signalling domains in the Streptomyces coelicolor genome according to the Conserved Domain Database
The biological function of Streptomyces PCD remains somewhat unclear. It was reported to be involved in the generation of nutrients to be consumed by the aerial/sporulating mycelium, a kind of cannibalism (Mendez et al., 1985; Miguelez et al., 1999). If we consider that the best known bacterial PCDs, those occurring in Streptococcus and Bacillus, are involved in competence (taking fragmented DNA by transformation) (Guiral et al., 2005; Engelberg-Kulka et al., 2006), and that in the case of Bacillus, this process precedes sporulation and antibiotic production, an analogous process might also be happening in Streptomyces: appropriate DNA fragments would be produced by specific nuclease activities (Cal et al., 1996) and the lysis of MI mycelium (Manteca et al., 2006a) and incorporated by the multinucleated MII followed by recombination and the formation of a huge battery of variable spores. Several authors have hypothesized about the existence of horizontal gene transmission (HGT) phenomena in Streptomyces and other actinomycetes (Wiener et al., 1998; Ueda et al., 1999; Egan et al., 2001; Metsä-Ketelä et al., 2002; García-Vallve et al., 2003; Kawase et al., 2004; Nishio et al., 2004; Doroghazi & Buckley, 2010) but the mechanisms generating this HGT remains poorly understood. Conjugative plasmids (reviewed in Thoma & Muth, 2012) and transduction (Burke et al., 2001) may contribute in some way to this HGT, but competence/transformation may also occur. Streptomyces PCD precedes MII differentiation and sporulation and it was postulated that components released during the degradation of these dying cells could be producing diffusible signals in the form of amino acids/peptides (Sánchez & Braña, 1996) or N-acetylglucosamine (Rigali et al., 2006), thereby inducing differentiation. Further work will need to delve into the biological significance of Streptomyces PCD.
Conclusions and future perspectives
Streptomyces growth in nature differs substantially from that observed in ordinary laboratory cultures, a fact that must be borne in mind when development is analysed. MI is the vegetative mycelium and predominates in nature. Under stress conditions (nutrient/oxygen limitation etc.) it suffers a PCD and differentiates to a multinucleated mycelium (MII) that forms spore chains at the end of the cycle. Multinucleated MII would facilitate rapid growth and nucleoid division prior to sporulation. MII produces antibiotics that are decisive in helping the bacterium compete with other microorganisms.
Streptomyces research has classically focused on the aerial mycelium formation and sporulation phases taking place in solid cultures. By contrast, pre-sporulation stages, including differentiation in liquid cultures, have been largely ignored. The new insights regarding pre-sporulation stages of Streptomyces in combination with future work aimed at understanding the biochemical regulation of these processes will be key to comprehending and optimizing hyphae differentiation in industrial fermentations, as well as improving the screening for new secondary metabolites from natural Streptomyces strains.
This research was funded by an ERC Starting Grant (Strp-differentiation 280304) and by grant BIO2010-16303 from the Subdirección General de Proyectos de Investigación, (DGI), Ministry of Science and Innovation (MICINN). We also thank Priscilla A. Chase for proofreading the text.