Helical ultrastructure of the L-ENA spore aggregation factor of a *Bacillus paranthracis* foodborne outbreak strain

Endospores of Bacillota are highly resilient, multi-layered structures that enable long-term survival and dissemination of pathogenic Bacillus and Clostridium species. The outermost exosporium mediates interactions with the environment and host, with the collagen-like glycoprotein BclA known to modulate complement. Filamentous endospore appendages (ENAs) have been observed for decades on spores of pathogenic species but have resisted molecular identification due to extreme robustness. Recent cryoID work identified S-type ENAs (S-ENA) formed by DUF3992 proteins (Ena1A/B) in Bacillus cereus s.l., suggesting roles in self-adherence and spore clustering. Spores of the food-poisoning outbreak strain Bacillus paranthracis NVH 0075-95 display a second class, ladder-like ENAs (L-ENA), which are shorter, exosporium-tethered fibers with a single distal ruffle implicated in spore-spore aggregation. The study aims to identify the structural subunit of L-ENA, resolve its ultrastructure, define the associated gene cluster and accessory components responsible for exosporium anchoring and ruffle formation, and assess distribution and expression patterns relevant to pathogenesis.

Historical EM studies documented ENAs on spores of pathogenic Clostridia and Bacillus but not typically on saprophytes (e.g., B. subtilis). Pili on vegetative Gram-positive cells are often sortase-assembled and function in adhesion and biofilms. Recent cryoID resolved S-ENA as DUF3992-based helical fibers (Ena1A/B) with lateral β-sheet augmentation and disulfide stabilization; S-ENA are widespread in pathogenic B. cereus s.l. (>90%) and may contribute to virulence and spore clustering. BclA forms the hairy nap of the exosporium and recruits complement factor H, attenuating complement activation. Collagen-like proteins (CLPs), including BclA, are common on Firmicute spores, though many functions remain unclear. The exosporium is a disulfide cross-linked 2D lattice mainly of ExsY hexamers. These findings frame hypotheses that ENAs, like pili, mediate interactions important for dissemination and biofilms, and that specialized tip adhesins (ruffles) may diversify function across ENA subclasses.

- Sample preparation and imaging: ENAs were enriched from B. paranthracis NVH 0075-95 spores by shear-induced detachment and purification. Negative-stain TEM (nsTEM) used uranyl acetate staining on glow-discharged formvar/carbon-coated grids, imaged on a 120 kV JEOL 1400. - Cryo-EM: For ex vivo ENAs, GO-coated Quantifoil grids were prepared and plunge frozen. Data collected on a JEOL CRYO ARM 300 with omega energy filter and K2/K3 detectors. Ex vivo dataset: 1568 movies (0.766–0.82 Å/pixel), yielding a 5.8 Å C7-symmetry helical reconstruction (twist 17.04°, rise 43.82 Å). Recombinant Ena3A (recEna3A) dataset: 10886 movies (0.782 Å/pixel, 64.66 e-/Ų over 60 frames), refined to 3.3 Å global resolution with C7 symmetry (twist 18.5°, rise 44.97 Å) in CryoSPARC v4.0.3. 2D/3D classification and 3D variability analysis (ChimeraX) probed flexibility. - Gene/protein identification: HMMsearch 3.0 with a DUF3992 HMM (from Ena1/2 sequences) against NVH 0075-95 genome (GCA_027945115.1) identified a DUF3992 protein (WP_017562367), designated ena3A. Synthetic gene cloning of ena3A into pET28a enabled expression in E. coli C43(DE3). Insoluble fractions after lysozyme and 1% SDS extraction were examined by nsTEM to confirm fiber formation. - Recombinant fiber stability assays: recEna3A fibers subjected to autoclaving, 8 M urea, SDS, proteinase K, and formic acid. Post-treatment nsTEM and 2D class averages assessed integrity and depolymerization. - Gene cluster discovery and distribution: cblaster (CAGECAT v1.0) remote search of NCBI RefSeq NR using Ena3A, ExsL, BcIA, and Tn3 transposase queries identified ena3 clusters. Quality-checked assemblies (QUAST) from B. cereus s.l. (n=656) and B. subtilis (n=136) were analyzed by tBLASTn (coverage >70%, identity >30%, conserved synteny) to score presence/absence, paralogs, and genomic location. Mashtree and Microreact visualized phylogeny and metadata. - Expression analysis: Sporulation timeline established by phase-contrast/TRITC fluorescence microscopy (FM4-64 staining). RNA extracted at 4, 8, 12, 16 h from sporulation medium cultures (3 biological replicates). cDNA synthesized for qRT-PCR (SYBR Green, Pfaffl method) targeting exsL, l-bclA, ena3A relative to rpoB. Standard PCR across exsL→l-bclA and l-bclA→ena3A junctions tested operonic structure. - Genetics: Markerless deletion mutants ΔexsL, Δl-bclA, Δena3A constructed with PMAD-I-Scel-based allelic exchange using ~700–800 bp homology arms. Complementation in pHT304 (low-copy) under native upstream region restored phenotypes. nsTEM of spores assessed presence/absence of L-ENA and ruffles. - Structural modeling: AlphaFold2/Multimer predicted complexes: Ena3A–ExsL (pLDDT 82.6, ptmscore 0.73), ExsL–ExsY (pLDDT 80.2, ptmscore 0.71), ExsY hexamer (pLDDT 81.3, ptmscore 0.86), and L-BcIA94–336 homotrimer (pLDDT 94, ptmscore 0.9). Domain architectures annotated with InterPro. Comparisons to known structures (e.g., BclA-C PDB:1WCK; S-ENA Ena1B PDB:7A02). - Genome sequencing: Hybrid PacBio RS II and Illumina NovaSeq reads assembled with Unicycler; quality assessed by QUAST, Bandage, BUSCO; corrected by Pilon; annotated by NCBI PGAP; plasmid prediction by PlasFlow; species validation via TYGS. Data deposited (GCA_027945115.1). - Data deposition: EMDB EMD-17579 (ex vivo L-ENA), EMD-17627 (recEna3A); PDB 8PDZ (recEna3A atomic model).

- Identification and structure of L-ENA: L-ENA in B. paranthracis NVH 0075-95 are ~7–8 nm diameter fibers with a ladder-like pattern (4.6 nm repeat), tethered to the exosporium and terminating in a single 2 nm tip fibrillum (ruffle) comprising a ~45 nm stalk and globular head. - Subunit and architecture: The major L-ENA subunit is Ena3A (WP_017562367), a DUF3992 protein. Cryo-EM of recombinant Ena3A fibers resolved a 3.3 Å structure of axially stacked heptameric rings with C7 symmetry (twist 18.5°, rise 44.97 Å). Rings interlock via N-terminal connectors (first 14 residues, Ntc) inserting into the lumen of the ring below. - Covalent cross-linking: Three inter-molecular disulfide bridges stabilize the fiber: lateral intra-ring Cys22–Cys82 and Cys14–Cys15; longitudinal inter-ring Cys8–Cys21 between the Ntc and ring lumen. Each ring contains 21 disulfide bonds arranged in three concentric rings. - Flexibility and stability: Despite extensive disulfide cross-links, L-ENA fibers exhibit notable flexibility via rocking of rings around Ntc hinges. Fibers withstand extreme treatments (autoclaving, 8 M urea, SDS, proteinase K) with preserved morphology; 100% formic acid likely unfolds monomers without depolymerization, indicating non-reduction of disulfides. - Gene cluster and components: ena3A resides in a rare three-gene cluster (ena3) on plasmid CP116205 (flanked by Tn3 transposase and topoisomerase fragments) with exsL (N-terminal Spore coat protein Z/Y-like, C-terminal Ena-core) and l-bclA (N-terminal collagen-like domain and C1q/TNF-like BclA C-terminal domain). ExsL serves as an exosporium anchor; L-BcIA forms the ruffle. - Genetics/phenotypes: ΔexsL spores lack exosporium-tethered L-ENA but show detached L-ENA with ruffles in supernatants, indicating ExsL is required for anchoring. Δl-bclA spores retain L-ENA fibers but lack ruffles, implicating L-BcIA as the L-ENA tip fibrillum. Δena3A spores lack L-ENA entirely. Complementation restores wild-type phenotypes. - Expression: PCR across gene boundaries indicates l-bclA and ena3A form a bicistronic operon; exsL is separate. qRT-PCR shows sporulation-specific upregulation: at 12 h, ena3A ~13,000-fold, l-bclA ~1,200-fold, exsL ~40-fold over 4 h vegetative baseline; transcripts absent/low during early vegetative growth (≤8 h). - Distribution: The ena3 cluster is rare in B. cereus s.l.: 9.5% (62/656) genomes, including B. cereus (51), B. thuringiensis (4), B. anthracis (1), B. paranthracis (1), B. toyonensis (1), B. mobilis (2), Bacillus spp. (2). Some strains carry multiple paralogous copies (up to five). Not found in B. subtilis or other saprophytes. Genomic context suggests mobility via transposons/integrases (mobilome). - Structural modeling of anchoring and tip: AF2 supports Ena3A–ExsL heterodimerization via ExsL C-terminal Ena-core mimicking Ena3A lateral contacts (predicted disulfide ExsL Cys224–Ena3A Cys22). AF2 also supports ExsL–ExsY interaction via ExsL N-terminus, consistent with ExsY hexameric lattice integration. L-BcIA is a trimeric C1q-like CTD fused to a collagen-like stalk; predicted lengths (~52 nm) and head size (~4 nm) match measured ruffle dimensions (~50 nm) and docking into the Ena3A ring lumen at the fiber apex.

The study resolves the molecular identity and architecture of a second ENA class (L-ENA) in B. paranthracis spores, addressing longstanding questions about ENA composition and function. L-ENA fibers are heptameric ring stacks of DUF3992 subunits (Ena3A) stabilized by a dense network of disulfide bonds, conferring exceptional mechanical and chemical robustness while allowing flexibility through N-terminal connector hinges. Genetic and structural analyses reveal a minimal three-gene system responsible for L-ENA biogenesis, surface display, and functional decoration: ExsL anchors fibers into the exosporium lattice (likely through ExsY-mimicking interactions and disulfide continuity), Ena3A forms the fiber body, and L-BcIA, a C1q-like collagen trimer, constitutes the distal ruffle that mediates spore-spore interactions. Expression is tightly linked to sporulation, ensuring assembly at the relevant developmental stage. The rarity and mobilome association of ena3 across B. cereus s.l. suggest horizontal acquisition and potential selective advantages. Functionally, ENA-mediated spore aggregation likely promotes biofilm fortification, environmental persistence, immune evasion (by clumping), and increased infectious dose, positioning L-ENA as a secondary effect virulence factor. Distinctions between S-ENA and L-ENA (e.g., different helical architectures, tip fibrillae stoichiometry, and possibly distinct CLP adhesins) hint at functional diversification akin to specialized adhesins at pilus tips in other pathogens. The proposed ExsY→ExsL→Ena3A disulfide-linked anchoring chain and ruffle docking into the terminal ring lumen provide a mechanistic model for assembly and display that can be experimentally tested.

This work identifies and structurally characterizes L-ENA fibers on B. paranthracis spores, demonstrating that a compact ena3 gene cluster encodes all components needed for assembly, anchoring, and functional decoration: Ena3A forms stacked heptameric, disulfide-cross-linked rings; ExsL anchors fibers into the exosporium; and L-BcIA, a C1q-like collagen trimer, forms the distal ruffle. The 3.3 Å cryo-EM structure reveals inter- and intra-ring disulfide networks and a lumen-docking mechanism for fiber elongation and ruffle attachment. Expression is sporulation-specific, and deletion phenotypes validate roles of each gene. The ena3 cluster is a rare, mobile element within B. cereus s.l. Given demonstrated roles of ENAs in spore aggregation, L-ENA likely enhance survival, dissemination, and virulence. Future research should: (i) experimentally define ExsL stoichiometry and binding epitopes within the exosporium lattice; (ii) biochemically confirm predicted disulfide linkages and complex assemblies; (iii) identify and characterize the S-ENA ruffle protein(s); (iv) determine molecular targets of L-BcIA and assess adhesion specificities; and (v) quantify contributions of L-ENA to biofilm formation, environmental persistence, and pathogenicity in vivo.

- The ex vivo L-ENA cryo-EM map was limited to 5.8 Å due to sparse particle images and intrinsic fiber flexibility, precluding de novo atomic modeling from native material; high-resolution modeling relied on recombinant Ena3A fibers. - AF2-based predictions of ExsL–Ena3A and ExsL–ExsY interactions, and disulfide linkages, await experimental validation (biochemistry/structural data in complex). - The specific ligand(s) or binding targets for the L-BcIA ruffle remain unidentified, and the mechanistic basis of spore-spore adhesion is unresolved. - The S-ENA ruffle protein(s) were not identified here. - Distribution analyses depend on available genome assembly quality and may underestimate presence in poorly assembled genomes. - Functional impacts on virulence and biofilms are inferred from structural/aggregation data and require in vivo validation.