Role of SafA in spore coat assembly.

A: Main stages of sporulation: pre-divisonal cells (a); sporangia that have undergone polar division (b), at the beginning of engulfment (c), following engulfment completion (d), during the last stages in synthesis of the spore cortex and the coat and crust layers (d-f) and when spores are released (g). The localization of SafA-YFP is represented in green (panels c to e). The cell type-specific sigma factors and their window of activity is indicated. B: In a targeting step, soon after engulfment begins, SafA is recruited to the forespore by the SpoIVA ATPase. SpoIVA localization at or close to the forespore outer membrane in turn relies on SpoVM. In a second step, encasement of the forespore by SafA depends on an interaction with SpoVID. C: SafA is the Hub for inner coat assembly; it recruits proteins produced both under the control of σE and σK, such as Tgl. Assembly of SafA is auto-regulatory as the recruitment of Tgl results in cross-linking of SafA (two way arrow). CotE, required for outer coat assembly and CotZ, required for crust assembly, are not represented for simplicity. D: Tgl reaction cycle. In a first step, a Gln donor substrate approaches the enzyme from the Q-side and forms an acyl-enzyme covalent intermediate. In a second step, a Lys substrate (acyl acceptor) engages the enzyme from the K-side and the cross-link is formed. The reaction occurs in a hydrophobic tunnel that harbors the catalytic residues including Cys116. The ceiling of the tunnel is thought to open to release the cross-linked product. The mostly hydrophobic residues that line the Q- and K-side entries of the tunnel are indicated.

Domain organization and properties of SafA.

A: SafAFL has an N-terminal LysM domain which together with a region termed A forms a localization signal. This signal is required for targeting SafA to the spore surface via an interaction with SpoIVA, and for encasement, through an interaction with SpoVID. The LysM domain is also important for the interaction of SafA with the cortex peptidoglycan while C30 is the morphogenetic domain of the protein, responsible for the recruitment of the inner coat proteins. C30 is also made independently through internal translation of the safA mRNA initiating at methionine codons 161 or 164. A third form of the protein, SafAN21, which includes the LysM domain, region A and the linker, has no known function. A 6-mer short acidic peptide (CAM6) at the C-terminal of SafAFL or C30 is also indicated. C30 relies on an interaction with the C30 domain of SafAFL for its localization to the coat. B: SafA is an intrinsically disordered protein (IDP). Disorder propensities, calculated using Metapredict (blue) (51) are compared to the structural confidence calculated by AlphaFold 2 (AF-2; red), as measured by the predicted local distance difference test (pLDDT). The two predictions are in anti-phase: Metapredict indicates high disorder in regions where AlphaFold assigns low structural confidence. These findings support a model in which SafA has a well-structured LysM domain, followed by highly disordered regions, particularly in region A and the C30 domain. C: Structure of SafA predicted with AF-3. The left panel shows the structure colored based on SafA’s molecular architecture (LysM, region A, linker, C30, and CAM6, as in panel A). D: Diagram of states partitioned into five distinct conformational classes based on the fractions of positively and negatively charged residues. SafA is found within the region associated with compact conformations, suggesting that it may adopt a more condensed, globule-like structure despite its disordered nature. E: Self-intermolecular interaction maps (inter-maps) between the intrinsically disordered regions (IDRs) of SafA reveal potential sub-regions of intermolecular interactions. These interactions were predicted using the chemical physics encoded within the Mpipi-CG (52) and CALVADOS 2.0 (53) force fields using FINCHES (Ginell et al., 2025). Attraction and repulsion are displayed using a purple-to-green gradient, indicating varying interaction strengths within the IDRs, particularly in the C30 region. Both force fields predict attractive interactions within residues the stretch defined by residues 191-316 (this region is represented as a gray line in panel B).

C30 forms disulfide cross-linked oligomers.

A: Diagram of C30 showing the position of Cys (green dots) and Lys (white dots) residues (numbering is from the N-terminus of SafAFL; see also S1_Fig). B: C30WT, C30K to A (the two Lys residues replaced by Ala), C30C to S (the four Cys residues replaced by Ser) and C30K to A/C to S (the two groups of substitutions combined) were resolved by SDS-PAGE (12.5% gels) under reducing (left panel) or non-reducing (right panel) conditions. The blue dashed lines shows the dye front. C: C30WT and C30C to S were resolved by BN-PAGE (5-15% gradient) in the absence or in the presence of DTT (-/+ 2 mM). In panels B and C, the position of the relevant proteins is indicated by arrowheads. The gels were stained with Coomassie. D: SEC analysis of C30WT and C30C to S on a Sephacryl S-300 column. The estimated sizes of C30WT and C30C to S are shown. The elution positions of blue dextran (void volume, V0) and of molecular weight standards (in kDa) are indicated.

The N- and C-terminal pairs of Cys residues have different functions.

A and B: C30WT, C30N-ter (Cys214 and Cys228 replaced by Ser) and C30C-ter (Cys323 and Cys325 replaced by Ser) and C30C to S (all Cys residues replaced by Ser) were resolved by SDS-PAGE (12.5% gels) under reducing (A) or non-reducing (B) conditions. C30WT was loaded in the absence (WT-DTT) and after treatment with DTT (2 mM; WT +DTT) C: The same proteins as in A were resolved by BN-PAGE (5-15% gradient) and the gels stained with Coomassie. C30WT was also analyzed after incubation with DTT (as above). In A and B, the various oligomeric forms of C30 are indicated by arrowheads. The gels were stained with Coomassie. D: C30 has a parallel orientation within dimers (C302) and in higher order complexes (C306). Disulfide bonds involving the C-terminal Cys residues (red crosses) orient and stabilize the dimer, while the N-terminal Cys residues cross-link adjacent dimers (black). The sizes of the various species detected by different electrophoretic techniques is also show at the bottom of the panel. E: Mass spectrometry analysis of the C302 dimer identifies disulfide bonds between Cys214 and Cys228 (a). and Cys323-Cys325 (b; see also S7_Fig).

Disulfide bonds are required for the formation of large C30 oligomers.

A: C30WT, pre-incubated in the absence or in the presence of DTT (2 mM) or C30C to S were incubated for 120 min in the absence (“-“) or in the presence (“+”) of Tgl. The reaction products were resolved by non-reducing (upper panel) or reducing (lower panel) SDS-PAGE (10%) and the gels stained with Coomassie. B: C30WT or C30C to S were incubated alone (“-“) or with Tgl (“+”) for 120 min and then the proteins were resolved by BN-PAGE (5-15% gradient). In A or B, the position of C30 and C302 are indicated by a blue arrowhead and the position of higher order oligomers is indicated by the red arrowheads. The position of Tgl is indicated by a green arrowhead and the brown arrowhead indicates a faster-migrating form of C30. C: SEC analysis of C30WT (top panel) and C30C to S (bottom), before and after incubation with Tgl, using a Sephacryl S-300 column. The estimated sizes of C30WT and C30C to S before and after cross-linking are shown. The elution positions of blue dextran (void volume, V0) and of molecular weight standards (in kDa) are indicated.

Disulfide bonds enforce the proper kinetics of C30 cross-linking by Tgl.

C30WT (A), C30C to S (B), C30K to A (C), and C30K to A/C to S (D) were purified and incubated in the absence (-) or in the presence of Tgl (+). Samples were taken at the indicated time points (in min) and resolved by SDS-PAGE (10%) under reducing conditions. In A-D, the blue arrowheads show the position of C30 or C302 whereas the red and pink arrowheads show the position of (C306)n or [(C306)n]n. The green arrowhead shows the position of Tgl, and the brown arrowhead shows the position of a C30C to S species that migrates faster than the monomer. Gels were stained with Coomassie. E: kinetics of cross-linking of C30WT, C30C to S, C30K to A, and C30K to A/C to S by Tgl as measured by the rate of disappearance of C30. F: kinetics of formation of the [(C306)n]n species from C30WT (red) and C30C to S (blue) in the presence of Tgl. G: interaction of Tgl with C30WT or C30C to S measured by microscale thermophoresis. The estimated KD for the interaction is indicated. The data in E to G are from three independent experiments.

Importance of the lysine residues in the cross-linking of C30.

C30WT (A) and variants K177A (B), K318A (C), and K177A/K318A (K to A for simplicity) (D) were purified and incubated alone (“-“) or in the presence of Tgl (“+”). Samples were taken at the indicated times (in min) and resolved by SDS-PAGE (10%) under reducing conditions. The gels were stained with Coomassie. In A, B, C and D panels, the blue arrowheads show the position of C30 and C302; the red and the faint red arrowheads show the position of HMW forms. The green arrowhead shows the position of Tgl. E: kinetics of cross-linking of each variant was measured by the rate of disappearance of the C30 monomer. The graph shows the data obtained with three independently purified C30 and cross-linking assays. F: model for the role of K177 and K318 in formation of the C30 cross-linked oligomers. In C30WT, K177 is involved in intramolecular (intra-dimer) cross-links, whereas K318 is thought to be involved in intermolecular (between pairs of dimers) cross-links. The panel also represents the altered pattern of C30C to S cross-linking.

Models of C30WT oligomers.

A: SEC-SAXS chromatograms of C30 (top) and C30 after incubation with Tgl (C30 + Tgl; bottom). The UV absorbance at 280 nm is shown across the elution frames; the corresponding radius of gyration (Rg) are shown as red dots. The inset shows violin plots of Rg distributions across each peak; for the C30 + Tgl sample, a dotted line links peak p to the Rg distribution plot. The position of Tgl is indicated by a black arrow. B: Kratky plots of SAXS profiles for C30 (blue) and C30 + Tgl (green), averaged from frames with constant Rg values, shown alongside the distribution of their P(r) versus (same color). C: Normalized pair-distance distribution functions, P(r), derived from the experimental SEC-SAXS data for C30 (blue) and C30 + Tgl (green). D: Low-resolution ab initio reconstructions (DAMFILT of 20 independent runs) of C30 (light blue) and C30 + Tgl (green) show elliptical, globular particles with a dense core, suggesting a compact nature of the assemblies. The estimated Mw of the complexes is indicated.

Tgl controls the extractability and dynamics of SafAFL/C30 complexes.

A: Spores of the following strains were purified, the coat proteins extracted and resolved by SDS-PAGE (15%): the WT, a safA in-frame deletion mutant (ΔsafA) and a tgl deletion mutant (Δtgl); other strains have a ΔsafA deletion and are complemented in trans, at the non-essential amyE locus with the WT, safAC to S, safAK to A and safAK to A/C to S alleles. The gels were stained with Coomassie (top panel) or immunoblotted with anti-SafA (middle) or anti-Tgl antibodies (bottom). The arrowheads show the position of the following proteins: SafAFL, bands 1 and 2, light blue; C30, dark blue; GerQ and Yeek, black; CotG, red; Tgl, green. Note that the K to A substitutions retard the migration of SafAFL or C30. B: quantification of the levels of the various proteins extracted from spores of the ΔsafA mutant complemented with the indicated safA alleles at amyE. The data are from three experiments using independently prepared spore suspensions and shown normalized to the WT. Note that the quantification of bands 1 and 2, not seen in the WT, is relative to SafAFL. For the statistical analysis we used a one-way ANOVA followed by Dunnett’s multiple comparison test (⍺ = 0.05). Asterisks indicate statistically significant differences (p < 0.05), with * denoting values <0.05 relative to the WT. C: Represents cells at the time when SafA begins to be synthesized under αE control and localizes to the spore surface, and cells after engulfment completion, when Tgl is produced under the control of αK. At this stage SafA forms caps at the MCP and MCD forespore poles. For both poles, the edge of the cap was photobleached and the signal recorded over time. The recovery period is only illustrated for the MCP pole; the protein is dynamic, if the signal recovers, or otherwise static. D: Cells were collected 5 hours after the onset of sporulation and imaged by confocal microscopy. The signal from SafA-YFP was photobleached at the MCP forespore pole in the WT and in Δtgl mutant. E: The signal in the photobleached areas was then monitored for up to 25 seconds and plotted with the standard deviation for the mother cell proximal (top) and distal (bottom) forespore poles.

Assembly of SafA is biphasic and relies on a hierarchical cross-linking cascade.

A. Prior to engulfment completion, SafAFL (not shown for simplicity) and C30 are produced and self-assemble into oligomers that undergo disulfide cross-linking. Following engulfment completion, activation of σK leads to production of Tgl, which is recruited to the spore surface via pre-assembled SafAFL/C30 complexes. Tgl then “spotwelds” these complexes resulting in their immobilization and reduced extractability. Tgl itself may also become cross-linked to the SafAFL/C30 complexes. B. C30 relies on the interaction with SafAFL for localization to the cortex/inner coat interface. Each C30 hexamer may incorporate one or more SafAFL molecules, anchoring the complex to the cortex. SafAFL/C30 complexes may recruit different inner coat proteins (a, b) before or after Tgl-mediated cross-linking (asterisks). The complexes are not drawn to scale and the CAM6 motif is omitted for simplicity.

Primary structure and organization of SafA.

A: The domains of SafAFL are indicated: LysM, light blue; region A, light orange; the linker, light mint green; and C30, pale yellow. The 6 amino acid-long C-terminal acidic peptide (CAM6: PEEENE) is highlighted in mustard yellow. Relevant amino acids are highlighted as follows: Cys residues, green; Lys residues, reddish-pink; Gln residues, magenta. Potential cleavage sites for Trypsin and Asp-N are indicated by green or black dots, respectively and verified cleavage sites are shown by arrowheads with numbers indicating the order of cleavage. B. C30 is resistant to proteolysis. SafAFL was incubated with either Trypsin or AspN (left and middle panels). Control experiments for trypsin digestion were performed using BSA or ⍺-sinuclein as substrates (right panels). The reactions were conducted for 180 min, samples were collected at the indicated time points and resolved by SDS-PAGE (15%) under reducing conditions and the gels stained with Coomassie solution. Gels run in parallel were transferred to PVDF membranes for N-terminal sequencing analysis. Red arrowheads mark the position of SafAFL in the left and the middle panels and the forms of BSA and ⍺-sinuclein in the right panels. The black arrowhead marks the position of a stable proteolytic fragment of SafAFL that contains part of C30. Mw markers (KDa) are shown on the left side of the panels.

Similarity between C30 and human amelogenin.

The figure shows the alignment of the amino acid sequences of human amelogenin (UniProt code: Q99217) and C30. The alignment was generated using Clustal W (www.ebi.ac.uk) and edited using JalView (https://www.jalview.org). The two possible starts for C30, M161 or M164, are indicated. The Lys (K177 and K318) and Cys (C214, C228, C323 and C325) residues in C30 are indicated, as well as the CAM6 acidic motif at its C-terminus (see also Fig 2A). The domain organization of amelogenin is indicated (lines above the sequence): the N-terminal hydrophilic, tyrosine-rich region (TRAP), the central hydrophobic region, rich in histidine, glutamine and proline residues (HR) (grey lines above the sequence) and the C-terminal hydrophilic region (CTHR). The brown line below the C30 sequence represents the stretch of residues thought to include self-interacting motifs (positions 191-316 of SafAFL or P29 to V155 of C30) as defined in Fig 2B and 2E (see the main text for details).

Tgl activity decreases with the concentration of SafAFL.

A: The indicated concentrations of SafAFL were incubated with Tgl (4 μM) at 37°C; samples were collected at the indicated time points and resolved by SDS-PAGE (10%) under reducing conditions. Red and blue arrowheads mark the position of SafAFL and Tgl, respectively. The asterisk denotes likely degradation products of SafAFL. The red line shows the position of cross-linked products of SafAFL, [(SafAFL)n]. Mw markers (kDa) are shown on the left side of the panels. Gels were stained with Coomassie. B: Tgl activity was estimated by measuring the loss of SafAFL over time using Image J. C: Decrease in the level of Tgl-His6, at a fixed concentration of 4 μM, at the indicated times after the onset of the reaction (time 0), as a function of the concentration of SafAFL.

Cross-linking of SafAFL increases with the concentration of Tgl.

A: SafAFL (12 μM) was incubated with different concentrations of Tgl, as indicated, at 37°C. Samples were collected at the indicated time points and resolved by SDS-PAGE (10%) under reducing conditions and the gels were stained with Coomassie. Red and light blue arrowheads mark the position of SafAFL and Tgl, respectively. The asterisk denotes likely degradation products of SafAFL. The red line shows the position of cross-linked products of SafAFL, denoted as [(SafAFL)n]. Mw markers (in kDa) are shown on the left side of the panels. B: Tgl activity was measured and normalized as a function of the loss of SafAFL over time, using Image J. C: Decrease in the level of Tgl at the indicated time points, as a function of the concentration of Tgl, for a fixed concentration of SafAFL (12 μM).

SafAN21 is not efficiently cross-linked by Tgl.

A: SafAN21 was incubated, at four different concentrations with a fixed concentration of Tgl (4 μM) at 37°C. Samples were collected at the indicated time points, resolved by SDS-PAGE (10%) under reducing conditions and the gels stained with Coomassie. Red and light blue arrowheads mark the position of SafAN21 and Tgl, respectively. The asterisk mark the position of a possible intramolecular cross-linked species. Mw markers (kDa) are shown on the left side of the panels. B: Domain organization of SafAFL. At least in vitro, cross-linking of the protein by Tgl seems restricted to the C30 domain.

Effect of DDM on the complexes formed by C30.

C30WT (A and B) and C30C to S (C and D) were loaded at a concentration of 0.3 mg/ml on a Sephacryl S-300 SEC column in the absence (A and C) or in the presence (B and D) of 0.05% DDM. The protein and DDM concentrations mimic those used for the Blue Native PAGE analysis. The position of the peaks is identified for each chromatogram and an estimation of the MW is shown. The position of MW markers is shown by the light red arrowheads.

Identification of the Cys214-Cys228 and Cys323-Cys325 cross-links.

A and B: LC-MSMS analysis of disulfide bond Cys214-Cys228 and Cys323-Cys325 (numbering is from the N-terminal of SafAFL). The MSMS fragmentation pattern is show. The highlighted ions identify the Cys214-Cys228 and Cys323-Cys325 disulfide bonds. See also Fig 4D.

Cross-linking within the C30 oligomers.

Purified C30 was incubated for up to 120 min in the absence (“-“) or in the presence TglWT or its variant TglR185A (“+”). Samples withdrawn at the indicated time points (in min) were resolved by reducing SDS-PAGE (10%). In A and B, the blue arrowheads show the position of C30 and multimeric forms of the protein; the red arrowheads show the positions of (C306)n and [(C306)n]n. The green arrowhead shows the position of Tgl. Gels were stained with Coomassie. C: Comparison of the activity of TglWT and TglR185A, measured by the rate of disappearance of C30. D: a model for the cross-linking of C30 units within the oligomer, taken as an example. The model considers the possibility that K177 may only be involved in intermolecular cross-linking (between dimers) whereas K318 may only form intramolecular cross-links (within a dimer). The first cross-linking event involves K177 (1); then, a cross-link forms via K318 (2) followed by another cross-link involving K177 (3), until all the C30 units in the oligomer are cross-linked. The model accounts for the detection of a C30 trimer by SDS-PAGE.

Spore properties.

A: Heat and Lysozyme resistance of safA and C30 spores. Spores were density gradient purified from 24h sporulating cultures of a ΔsafA mutant or strains bearing either the WT safA allele or point mutations in C30 inserted at the non-essential amyE locus in a safA in-frame deletion mutant. Aliquots of the spore suspensions were serially diluted and plated on LB agar plates before and after incubation with lysozyme or at 80°C (for 20 min). After overnight incubation, the survivors were enumerated. The data shown are the mean and standard deviation (SD) of the results from three independent experiments and are normalized to the reference strain (MsafA with safAWT at amyE). B: Kinetics and extent of spore germination: spores of the indicated strains were density gradient purified and resuspended in a buffer. Germination was induced by adding a mixture of asparagine, glucose, fructose and KCl (AGFK) and followed by monitoring the drop in the OD580 for up to 1 hour. Control cultures were incubated in buffer, with no addition of AGFK, for the same duration. C: shows the extent of germination for the indicated strains, measured before adding the germinant mixture and 60 min thereafter, normalized to the reference strain. In A, B and C: the data shown are the mean and standard deviation (SD) of the results from three independent experiments. The data in A and C are normalized to the reference strain. Statistical analysis used a one-way ANOVA followed by Dunnett’s multiple comparison test (⍺ = 0.05). Statistical significance(P < 0.05) is indicated by asterisks **, P < 0.05; ****, P < 0.0001, and “ns” for non-significant differences.

SafA properties.

MSMS data of peptide ADHDDC323GC325DGDHQPY carrying disulfide bond Cys323-Cys325.

Bacterial strains used in this study.

Oligonucleotide primers.

Plasmids used in this study.