As is not surprising, a number of proteins are required for spore germination that are not present in growing cells (Paidhungat and Setlow 2002; Setlow 2003; Paredes-Sabja et al. 2011; Xiao et al. 2011). These proteins are synthesized only during sporulation, some in the developing spore and some in the mother cell, and this synthesis is regulated at the transcriptional level. This developmental and compartment-specific transcription during sporulation is driven by modification of RNA polymerase specificity by changes in sigma factors and other regulatory proteins associated with RNA polymerase at various times in development and in the forespore and mother cell compartments (de Hoon et al. 2010; Higgins and Dworkin 2012). There are five unique types of germination proteins, and these include GerP proteins, GerD, GR proteins, SpoVA proteins and CLEs as described below.
GerD is an approx. 180-residue protein present in spores of Bacillus species and is synthesized only in the forespore in parallel with GR and SpoVA proteins 1–2 h prior to forespores' CaDPA accumulation (Paidhungat and Setlow 2002; Setlow 2003). GerD is almost certainly a peripheral IM protein in the spore that is largely held in the IM by a diacylglycerol anchor (Igarashi et al. 2004; Pelczar and Setlow 2008; Mongkolthanaruk et al. 2009). GerD exhibits no clear sequence similarity to known proteins. However, this protein is predicted to be largely hydrophilic and is on the outer surface of the IM where it is largely if not completely in a single small focus termed the germinosome where the GR proteins are also found (Griffiths et al. 2011; Korza and Setlow 2013). GerD is essential for GR assembly in the germinosome in spores, although GRs are present in the IM of ΔgerD spores. In the absence of GerD, rates of GR-dependent germination are greatly decreased (Pelczar et al. 2007; Gupta, S., and Christie, G., personal communication), although why this is the case is not clear. However, the essential role of GerD in GR-dependent germination appears confined to the Bacillales, as GerD is not present in spores of the Clostridiales (Paredes-Sabja et al. 2011; Xiao et al. 2011). In B. subtilis, there are approx. 4000 GerD molecules/spore and its level is constant in different sporulation media, although may vary between individual spores in populations (Griffiths et al. 2011; Ramirez-Peralta et al. 2012a; Stewart and Setlow 2013).
The GRs are what recognize and respond to nutrient germinants, with a variety of different germinants recognized by GRs as described above. Spores of Bacillus species invariably have multiple GRs with different and sometimes overlapping germinant specificity (Ross and Abel-Santos 2010; Paredes-Sabja et al. 2011; Xiao et al. 2011). As GerD, GR proteins are also synthesized in the forespore and in parallel with GerD and SpoVA proteins. GRs contain three proteins termed A, B and C, and these are most likely subunits that interact to form a functional GR; a likely fourth GR D subunit has also been recently identified in Bacillales and Clostridiales species (see below). It is common that all three GR subunits are encoded in a single operon. In these cases, individual A, B and C proteins are associated primarily with a single GR, but there are clearly many exceptions to this, especially because some GR subunits are expressed monocistronically. Indeed, in at least one case, a GR protein from a monocistronic gene has been shown to interact at least functionally with GR proteins expressed from a rather far removed genetic locus, and there is other evidence for at least some functional interaction between individual proteins of different GRs (Igarashi and Setlow 2005; Christie et al. 2008; Stewart et al. 2012). There is clear sequence homology between A subunits of different GRs both within and across species and between B subunits and C subunits. However, there is minimal sequence homology with proteins in available databases. Most Clostridiales spores also have GRs that are clearly closely related to those of Bacillales, including putative D subunits, although a few Clostridium species have no GRs yet still germinate well with specific small molecule signals (Paredes-Sabja et al. 2011; Xiao et al. 2011). The most notable of the latter species is C. difficile whose spores germinate with specific bile salts, undoubtedly reflecting this species' preferred niche in the mammalian large intestine. Where it has been studied, all three GR subunits are required for GR function in Bacillus species, although as loss of any one subunit often renders the others unstable, this conclusion is not as definitive as one would like. Indeed, there is some evidence with spores of Clostridium perfringens that individual GR subunits may retain some germination-triggering activity (Banawas et al. 2013, in press).
Bioinformatic analysis suggests that the GR A proteins have 4–6 transmembrane (TM) segments, plus a large N-terminal hydrophilic segment and a small hydrophilic C-terminal segment (Wilson et al. 2012; Korza and Setlow 2013). The A protein has been localized in spores' IM by a variety of methods, and as noted above, GRs are present largely or completely with GerD in the germinosome. A large fraction of the A subunit of B. subtilis spores' GerA GR, presumably the N-terminal hydrophilic segment, is on the IM's outer surface (Wilson et al. 2012; Korza and Setlow 2013). The GRs' B subunits are predicted to be largely hydrophobic, with 10–12 TM segments, plus some small hydrophilic loops. The hydrophobic character of the B subunits has greatly slowed studies on these proteins, although they are almost certainly also in the IM along with the GRs' A and C subunits. As GerD, GRs' C subunits are IM lipoproteins, with these proteins on the IM's outer leaflet (Ross and Abel-Santos 2010; Paredes-Sabja et al. (2011); Xiao et al. 2011; Wilson et al. 2012; Korza and Setlow 2013). The structure of most of the B. subtilis GerB GR's C subunit has been determined by X-ray crystallography (Li et al. 2010), and the protein has a novel structure, so the structure did not give specific insight GR C protein function. As expected, mutations of highly conserved residues in the A, B and C subunits most often reduce or eliminate the function of the affected GR, even if the other GR proteins are still stable (Christie et al. 2008; Cooper and Moir 2011; Li et al. 2011a; Mongkolthanaruk et al. 2011). However, loss of one GR's subunit by a suitable deletion and even some point mutations can result in loss of the GRs' other subunits, presumably by their degradation. However, in some cases, analysis of specific mutants in GR proteins has given some hints as to locations of germinant binding regions in the GR proteins (Christie et al. 2010a; Mongkolthanaruk et al. 2011).
Very recently evidence was presented for a fourth GR protein that modifies GR function, termed ‘D subunit’, in spores of at least B. megaterium and B. subtilis (Paredes-Sabja et al. 2011; Ramirez-Peralta et al. 2013). Genes for these putative D subunits are present within or just upstream of operons encoding other GR subunits, are transcribed in parallel with the adjacent or encompassing GR operon and encode small approx. 75 residue proteins with two predicted TM domains, but no obvious sequence homology. In B. megaterium and B. subtilis spores, loss of these putative D-genes has either positive or negative effects on GR function, although not on levels of GR A, B and C subunits. Similar putative GR D subunits are encoded in many other Bacillales species' GR coding regions and in Clostridiales species as well.
Levels of GRs in spores vary significantly depending on sporulation conditions, with rich media leading to higher GR levels than with poor media (Hornstra et al. 2006; Ramirez-Peralta et al. 2012b). In rich media, levels of individual GRs in B. subtilis range from 600 to 1100 molecules per spore (Stewart and Setlow 2013), and these levels can be 3- to 8-fold lower in spores made in a poor medium. In B. subtilis, regulation of GR subunit genes is complex, with regulation by medium composition in some unknown fashion and by at least two auxiliary transcription factors, SpoVT and YlyA, all of which precisely control levels of GR subunits in individual sporulating cells (Ramirez-Peralta et al. 2012a; Traag et al. 2013). Probably as a result of this complex regulation, GR levels appear to vary significantly not only between spore populations made differently, but also between individual spores in a population, with stochastic effects on regulatory protein levels and GR expression levels likely also contributing to the wide variations in GR levels in individuals in spore populations (Griffiths et al. 2011; Setlow et al. 2012). The precise level of specific GRs in spores is almost certainly the major factor determining the rate of germination of spores of different populations, and also of different individuals in a single population (Zhang et al. 2010, 2011, 2012a, 2013) as discussed further below under SD spores.
Bacillus subtilis spores' GRs have been by far the best studied and these comprise five potential GRs, all encoded by tricistronic operons, although in one case, a putative GR D subunit is also a part of the operon (Setlow 2003; Ramirez-Peralta et al. 2013). In B. subtilis, two of the five GR subunits have no known germinants although the genes are expressed, and loss of these two GR operons does not affect spore germination with any known germinant. Of the other three GRs, GerA responds to l-alanine or l-valine alone and is strongly inhibited by d-alanine. As d-alanine can be generated by a sporulation-specific alanine racemase present in spores' outer layers, this enzyme has been suggested to modulate spore germination during sporulation by d-alanine synthesis as noted above. While GerA alone can trigger B. subtilis spore germination, neither of the other two important B. subtilis GRs, GerB and GerK, alone trigger germination efficiently. Rather these two GRs cooperate to respond to a germinant mixture termed ‘AGFK’, composed of l-asparagine (or l-alanine), D-glucose, D-fructose and K+. It appears likely that GerB binds the amino acid and probably the D-fructose, and GerK binds the D-glucose, while both GerB and GerK GRs probably bind K+. In the presence of D-glucose and K+, the GerK GR can also stimulate GerA function significantly, as can GerB plus l-asparagine, D-fructose and K+. This cooperation between GRs is a general feature and can also be seen in the synergy between low concentrations of germinants that target different GRs that respond to individual germinants, as these germinant mixtures generally give much more rapid germination than would be predicted by the activities with the individual germinants alone, although negative effects of one GR on the activity of another have also been seen (Atluri et al. 2006; Luu et al. 2011; Yi et al. 2011; Stewart et al. 2012; Gupta et al. 2013). However, the mechanism of this positive or negative cooperativity between GRs is not known, although it could be due to GR–GR association such as seen in the interaction between sensory proteins in bacterial chemotaxis.
As would be expected, loss of all GRs from Bacillus subtilis spores greatly reduces their germination with nutrient germinants, although not with CaDPA or dodecylamine. However, GR-less B. subtilis spores still exhibit a very slow rate of germination of approx. 0·1% per day, although the mechanism of this spontaneous germination is not clear (Paidhungat and Setlow 2002; Setlow 2003). Perhaps some component downstream of GRs in the signalling process fires occasionally. At the other end of the spectrum, forespore-specific overexpression of GRs 5–12-fold in B. subtilis results in more rapid spore germination, although the increased germination rates are by no means linear with GR levels (Cabrera-Martinez et al. 2003). Interestingly, if GerA or GerK GR expression is under too strong a forespore-specific promoter, the developing spore appears to germinate within the sporangium soon after GR synthesis begin, although why this happens is not clear.
The SpoVA proteins are a group of up to seven proteins that are expressed in the developing forespore in parallel with GRs and GerD (Paidhungat and Setlow 2002; Setlow 2003; Paredes-Sabja et al. 2011; Xiao et al. 2011). In B. subtilis, there are seven proteins, SpoVAA, B, C, D, Eb, Ea and F, which are encoded in a single operon in this order. Other Bacillales and Clostridiales species can have fewer proteins (although SpoVAC, SpoVAD and SpoVAEb are always present) and in some cases with more than one transcriptional unit. Bioinformatic analyses predict that SpoVA proteins other than SpoVAD and SpoVAEa are membrane proteins, generally integral membrane proteins, and at least SpoVAD and SpoVAEa, the two hydrophilic SpoVA proteins, have been localized to spores' IM and on this membrane's outer surface and are present at 6–8000 molecules per spore (SpoVAD) or 750 molecules per spore (SpoVAEa; Korza and Setlow 2013; Perez-Valdespino, A., 2013, unpublished results; Stewart and Setlow 2013).
In B. subtilis, loss of any SpoVA protein, except SpoVAEa or SpoVAF, gives an asporogenous phenotype, as while DPA is synthesized normally in sporulation, it is not taken up and the DPA-less spores germinate in the sporangium and lyse (Errington 1993; Perez-Valdespino, A. 2013, unpublished results). A C. perfringens ΔspoVA strain also does not accumulate DPA in spores, although these DPA-less spores are stable (Paredes-Sabja et al. 2008). Several temperature-sensitive spoVA B. subtilis mutants have also been isolated, and these accumulate DPA relatively normally at the permissive temperature, but DPA release from these spores during nutrient germination at the nonpermissive temperature is slowed considerably (Vepachedu and Setlow 2004; Wang et al. 2011a). These results led to the suggestion that SpoVA proteins make up the channel for DPA accumulation during sporulation, perhaps as CaDPA, as well as CaDPA release in spore germination. This suggestion was strengthened recently by a number of additional results in B. subtilis including the demonstration that (i) SpoVAD specifically binds the natural isomer of CaDPA in vitro, and this binding is eliminated as is CaDPA uptake in sporulation by mutation of conserved residues in the likely CaDPA binding pocket (Li et al. 2011b), (ii) SpoVAC expressed in E. coli acts as a mechanosensitive channel (Velasquez-Guzman et al. 2012), (iii) dodecylamine may well directly activate a SpoVA protein channel (Vepachedu and Setlow 2007b), and (iv) loss of SpoVAEa greatly reduced GR-dependent germination, but with no effects on CaDPA or dodecylamine germination (Perez-Valdespino, A., 2013, unpublished results).
While the results given above are certainly consistent with SpoVA proteins forming a CaDPA channel in the spores' IM, there is no information on the precise structure or organization of this channel, how it is gated, how the energy requiring process of CaDPA uptake into developing spores takes place, and how the same channel can act in both CaDPA uptake in sporulation that will require energy to accumulate CaDPA against a concentration gradient, as well as the likely energy independent CaDPA efflux in spore germination.
Spores of Bacillus species have two redundant CLEs, CwlJ and SleB, and both require the cortex-specific modification MAL for recognition and cleavage of PG (Setlow 2003; Giebel et al. 2009; Setlow et al. 2009; Heffron et al. 2010; Paredes-Sabja et al. 2011). SleB appears to be a lytic transglycosylase, and CwlJ has been predicted to have a similar catalytic activity, although this has not been demonstrated either in vivo or in vitro (Li et al. 2012, 2013). CwlJ is synthesized in the mother cell compartment of the sporulating cell and is located near the coat–cortex boundary. The enzyme is synthesized in its mature form and is thus present in the spore in a form that could be active. CwlJ localization in the spore requires its partner protein GerQ (originally YwdL) that in some Bacillus species is encoded as a cistron of a bicistronic operon with cwlJ, but in other Bacillus species, including B. subtilis, gerQ and cwJ are monocistronic. Most if not all SleB is synthesized in the forespore and is found in the IM (Boland et al. 2000; Chirakkal et al. 2002), however, some SleB is also found at the coat–cortex boundary. As CwlJ, SleB is present in spores in its mature form and also requires a partner protein, YpeB, for its localization in spores. YpeB has a likely membrane insertion sequence, and most of both YpeB and SleB are associated with the spore's IM, with both proteins on the outer surface of the IM (Korza and Setlow 2013). However, association of SleB and YpeB has not been demonstrated in vitro.
The crystal structure of the catalytic domain of SleB from B. cereus and B. anthracis has been determined at high resolution, and the structures are almost identical and very similar to those of lytic transglycosylases, including a large cleft for PG binding plus a catalytic glutamate residue (Jing et al. 2012; Li et al. 2012). SleB also has a PG-binding domain that is important for catalytic activity in vivo but not in vitro (Heffron et al. 2011; Li et al. 2013). Interestingly, YpeB can significantly inhibit SleB activity in vitro, and this has been suggested to be important for regulation of SleB activity in vivo (Li et al. 2013). CwlJ lacks a PG-binding domain, but its catalytic domain has extensive sequence similarity to SleB, and this similarity extends to CwlJ's predicted structure plus a putative catalytic glutamate residue that is essential for CwlJ function in vivo (Li et al. 2013).
A variety of evidence has shown that CwlJ is activated during germination by the CaDPA released in Stage I of germination (Setlow 2003). CwlJ is also activated by exogenous CaDPA, but only with the natural DPA isomer. The germination by exogenous CaDPA requires high CaDPA concentrations, while low concentrations are rather ineffective. This likely precludes triggering of the germination of nearby dormant spores by CaDPA released from one germinating spore (Zhang et al. 2011). In contrast to CwlJ, CaDPA alone does not activate SleB, and its mechanism of activation in germination is not known. It has been suggested that SleB's activity is determined by changes in strain on cortex PG due to changes in core water content (Setlow 2003), but the evidence for this suggestion is relatively weak. This is certainly a major unknown about CLEs, as is why CwlJ does not become active in sporulation when CaDPA is taken up or why SleB does not act on cortex PG shortly after the enzyme is made, as this is well before CaDPA uptake.
In contrast to spores of Bacillales species, most Clostridiales spores do not have CwlJ and SleB as CLEs, but rather have a single CLE termed SleC (Makino and Moriyama 2002; Paredes-Sabja et al. 2011; Xiao et al. 2011). SleC is synthesized in the mother cell and is in spores' outer layers where it is readily extracted from spores during decoating treatments (as is CwlJ in Bacillus spores). While there is no evidence for a partner protein for SleC, more important is the fact that SleC is synthesized as a preproprotein. The presequence is cleaved shortly after synthesis and is presumably crucial for translocation of SleC made in the mother cell across the outer forespore membrane to its location below the outer spore membrane where it is present in the dormant spore as the proprotein that is inactive. This proprotein is activated by proteolytic cleavage by a CspB protease, of which spores can have one or several, in the first min of spore germination. The CspB(s) are subtilisin-like proteases that are also synthesized in the mother cell and are localized in the same region of the spore as is pro-SleC. Presumably CspB action is triggered somehow by Stage I germination events, but in most cases how CspB is activated is not clear. The exception is with spores of C. difficile in which it appears likely that CspB is directly activated by bile salts (Francis et al. 2013), thus obviating the need to have any GRs recognizing bile salts for the germination of spores of this and similar species that lack GRs.