In the SMD simulations, water molecules were pulled through GerAB from the outside of the spore IM to the inside, as indicated in Figure 2A. For each individual SMD run, the cumulative work as a function of the z coordinate (perpendicular to the plane of the membrane) was calculated. For all runs, the exerted work increases at z = 5 nm, until z = 7.5 nm. This means that the water molecule does not spontaneously go through the protein at the pulling speed. At z > 7.5 nm, the cumulative works stays approximately constant, which means that no additional work is required to pull the water molecule through bulk water after it has exited the protein (Figure 2B). This observation shows that the SMD does not suffer from artifacts caused by pulling the water too fast, as pulling the water molecule through bulk water happens at a velocity typical for water diffusion at 298 K (Maginn, 2007). This indicates that the pulling speed is sufficiently slow.

The pulling work profiles all have a similar shape but differ in when the plateau is reached. To quantify this variation, differences in the cumulative work are defined as Δw = w_end – w_start. Δw ranged from Δw_min = 37 kJ/mol to Δw_max = 293 kJ/mol (Figure 2B). The Δw values show a large variation and are reported here to indicate that there are different ways in which the water molecule can go through the protein when pulled. In all simulations, a group of three amino acids with bulky side chains termed the triad, consisting of Y97 (located on TM3), L199 (located on TM6), and F342 (located on TM10) (Figure 2C), blocked the movement of the water molecule along the z-axis. To move through GerAB across the blockade formed by the triad, the water molecule passed on the side of L199 and F342, located between TM6 and TM10, and further away from Y97 in the run with Δw_min (Figure 2D). In contrast, in all other SMD simulation trajectories, the water molecule went through the center of the three bulky side chains forming the triad, as exemplified by run Δw_max (Figure 2E). A previous study (Blinker et al., 2021) using MD simulations and HOLE analysis (Smart et al., 1996) predicted a water channel formed by TM1, 2, 3, 6, and 8. In this study, we identified Y97 and L199 located on TM3 and TM6, respectively. We also identified F342 on TM10. Since the bulky side chain triad interfered with water molecule passage in all the SMD simulations trajectories, we checked if they are conserved in other Bacillus spore-forming species. With CLUSTAL Omega (Madeira et al., 2024), we aligned the amino acid sequences of GerAB with those of GerBB, GerKB from B. subtilis, GerVB from Bacillus megatarium QM B1551 and GerRB from B. cereus ATCC14579. This sequence alignment indicated that Y97 is fully conserved, while F342 and L199 are not conserved in the three germinant receptor subunits (Supplementary Figure 1). Therefore, Y97 might have a functional role in GRs of spore formers beyond B. subtilis. To verify the function of the triad in vivo, the triad residues have been mutated to alanine to test the function of the bulky side chains in the germination behavior of B. subtilis spores.

Four mutant strains of B. subtilis PY79 were successfully constructed, Y97A, L199A, F342A, and triA (Y97A, L199A and F342A triple mutant). Differences in germination efficiency of each strain with L-alanine or AGFK shows that each triad residue has an important role in both L-alanine mediated germination and AGFK initiated germination due to the reduced ability of mutant spores to respond to both types of germinants compared to wt (wild-type) spores. In the presence of L-alanine, only Y97A mutant spores germinated, with a germination efficiency of 61%, while the other mutants did not respond to L-alanine (Figures 3A, C). A similar pattern occurred when measuring germination by the OD₆₀₀ drop where Y97A spores showed a partial OD₆₀₀ drop (∼85%) compared to that of wt spores (∼60%), while L199A, F342A and triA mutants exhibited no OD₆₀₀ drop under same experimental conditions (Figure 4A). All mutants responded to AGFK with slightly slower kinetics compared to wt spores. Note that a lower OD₆₀₀ drop indicates a lower germination efficiency, albeit with different levels of the different mutant spores (Figure 4B). This is in agreement with single-spore germination efficiency assays, which revealed a drop around 10% of germination efficiency in L199A, F342A and triA mutant spores (Figure 4A). Notably, the OD₆₀₀ drop, and the germination efficiency drop with Y97A spores are both less than that of the other mutant spores.

To determine whether the mutant GerAB proteins were incorporated in the spores, we relied on the finding that GR stability depends on a strict 1:1:1 stoichiometry of the A, B and C subunits, and that the existence of GerAA and GerAC in spores depend on the stability of GerAB. This means that the number of GerAA proteins in spores is directly related to the number of GerAB proteins and the GerA complexes. Absence or reduction of GerAA levels in the mutants is a reliable marker for failure of assembling the complex (Cooper and Moir, 2011; Mongkolthanaruk et al., 2011). We therefore performed western blot analysis on wt and mutant spore lysis using a purified GerAA antibody. Only wt and Y97A mutant spores exhibited GerAA signals, although the level in the Y97A mutant is lower than that in wt spores (Figure 5). Since assembly of the GerA GR requires all three subunits, this means that the lack of response to L-alanine of L199A, F342A and triA mutant spores was due to the absence of whole GerA GR (Cooper and Moir, 2011). That is to say GerAB Y97A is produced and assembled into a complex with GerAA and GerAC, albeit the lower level may indicate its lower stability. At the same time, GerAB L199A, GerAB F342A and GerAB triA are either unstable or unable to form a complex with GerAA and GerAC proteins. These findings indicate that residues Y97, L199, and F342 are all critical for maintaining the structural stability of GerAB. Among them, L199 and F342 appear to be even more essential than Y97 for GerAB structure, as their mutation completely abolished the formation of the GerA germinant receptor complex.

To explore the structural aspects of the triad residues, we performed protein contact analyses on the ten starting structures of the SMD simulations, listed in Supplementary Table 1 in the Supporting Information. In particular, F342 has at least one π–π stacking interactions in all starting structures with F55, Y97, F198 or F346. While Y97 interacts with only F342 in five starting structures via π–π stacking. Counting the number of contacts, with a contact defined as two residues within 7Å. L199 showed the highest total number of contacts in seven of the starting structures (Supplementary Table 1). These results suggest that L199 and F342 engage in more interactions within GerAB, and their mutation has a stronger destabilizing effect on the protein compared to Y97. However, with respect to the original hypothesis of this study, the inability of the L199A, F342A, and triA mutants to form the GerA complex prevents us from directly testing whether L199 and F342 function as a barrier to water crossing. In contrast, the Y97A mutant was still able to form the GerA complex and exhibited reduced germination efficiency. Yet, it remains unclear whether this reduction is due to lower numbers of GerA complexes or a dysfunctional water channel.

The GerAB mutant spores constructed in this study without residual GerA showed slower germination kinetics when triggered by AGFK (Figures 3B, D, 4B). This observation was unexpected since AGFK is sensed collaboratively by GerB and GerK, which are not known to need involvement of GerA (Moir et al., 2002). Ultimately, the deletion of the gerA operon also led to a compromised germination in response to AGFK compared to a wt strain (Figure 6). This result was different when gerA: spe was used (Gao et al., 2022), as there was no delayed germination observed when spores were triggered with AGFK. This result indicated a crosstalk between GerA GR and GerB/GerK GR, since the absence of the GerA GR altered the function of GerB and GerK GR. These observations are in agreement with the current understanding of germinosome function where all three GRs in B. subtilis spores form protein complexes in germinosomes (Griffiths et al., 2011; Yi et al., 2011) and may exchange GR subunits (Amon et al., 2021). Also, the GerA GR is stimulated via D-glucose binding to GerK, although D-glucose does not trigger spore germination alone (Yi et al., 2011), indicating structural, and even functional interdependency. Taking this even further, the structural integrity of germinosome may require the presence of all three GRs. Our observations provide evidence that the triad is structurally important in the assembly of GerAB and suggest that GerAB structure has an influence on the function of the entire germinosome.

One 100 ns Molecular Dynamics (MD) simulation of a protein structure model based on Methanocaldococcus jannaschii ApcT (PDB: 3GI8) (Shaffer et al., 2009) generated by the structure prediction tool raptorX (Källberg et al., 2012) was conducted using the same starting structure and settings as those described by Blinker et al. (2021). The simulation was performed with GROMACS 2020.4 (Abraham et al., 2015), using the CHARMM36 (Aguayo et al., 2012) forcefield. In a cubic periodic box with starting dimensions of 10 by 10 by 11 nm, the system was solvated by TIP3P water molecules (Mark and Nilsson, 2001). The simulation was maintained at a temperature of 298 K using the Bussi velocity rescaling thermostat (Bussi et al., 2007) and at constant pressure of 1.0 bar with the Parrinello-Rahman barostat (Parrinello and Rahman, 1981; Nosé and Klein, 1983). Long-range electrostatic interactions were treated using the Particle Mesh Edward (Darden et al., 1993) method with a maximum grid spacing at 1.6 Å. For short-range electrostatic and Van-der-Waals interactions, the cutoff radius was set to 12 Å. The production run lasted 100 ns with a timestep of 2 fs, with coordinates saved every 2 ps. The term “(free) MD simulation” refers to the MD simulation production run conducted for this research.

Steered Molecular Dynamics (SMD) simulations were performed to pull a water molecule through the GerAB protein. To extract a suitable starting structure from the MD simulation trajectory, the protein was first positioned at the center of the simulation box with its center of mass between 4.76 nm and 4.85 nm along the x- and y-coordinates, and membrane boundaries at 3.0 nm < z < 7.0 nm. The entrance of the water was defined within the region 3.5 nm < x < 6.5 nm, 3.5 nm < y < 6.5 nm, and 4.6 nm < z < 5.0 nm. Ten frames, each containing a single water molecule within the defined entrance region, were extracted as the starting configurations for Steered Molecular Dynamics (SMD) simulations. The selected water molecule was then subjected to the pulling force during the simulation. The steered MD simulations were conducted using PLUMED 2.7.0 (Bonomi et al., 2009, 2019; Tribello et al., 2014) embedded with GROMACS 2020.4, where the steering collective variable (CV) was the position in the z-direction of the oxygen atom of the pulled water molecule. In all SMD simulations, the selected water molecule was pulled 4 nm over 100 ns along the z-coordinate toward the inside of the inner membrane (IM), following the direction of water intake during germination. Over the 100 ns simulation, a moving harmonic potential kept the water oxygen constrained to z with a force constant κ = 10000 kJ mol^–1 nm^–2. The harmonic restraint moved at a constant velocity of 0.04 nm/ns. In addition, the water oxygen atom was constrained in the x and y direction with a harmonic potential with a force constant κ = 10 kJ mol^–1 nm^–2, to keep the water molecule inside the protein. The SMD simulations were performed with the same settings as the MD simulation. The work performed to constrain the water molecule along the z coordinate in the SMD simulations was computed as the cumulative force applied to the water molecule over the simulation time. In total, 10 SMD runs were performed. Visual inspection of the simulation trajectories was conducted using Visual Molecular Dynamics (VMD) 1.9.4a53 version (Humphrey et al., 1996). For the 10 starting frames of the SMD simulations, protein contact analysis was carried out via Contact Map Explorer¹ built on the MDTraj library (McGibbon et al., 2015). First, the minimum distances between all residue pairs in the protein were determined. Two residues are counted as a contact if their minimal distance was 7Å or less. For the three residues in the triad, contacts were counted. Then, a more detailed analysis of the contacts formed by the triad residues was performed by RING, which identifies hydrogen bonds, van der Waals interactions and π – π stacking between residues (Conte et al., 2023).

All strains were derived from B. subtilis PY79. The gerA operon was cloned by PCR, inserted in plasmid vector pUC19 with the Gibson Assembly Master Mix kit (New England BioLabs, NEB # E2611S). Mutagenesis was carried out with the QuikChange Lightning Site-Directed Mutagenesis Kit (Agilent Technologies, Cat # 210518-5). Mutant gerAB sequences were integrated into the original gerAB locus by double crossover along with an erythromycin (erm) resistance cassette. The precision of the mutagenesis in the gerAB locus was confirmed by Sanger sequencing of all mutants. Strains involved in this study are listed in Table 1.

Bacterial strains were streaked onto LB agar plates with appropriate antibiotics and incubated overnight to avoid stationary phase; a single colony was inoculated into liquid LB with antibiotics. Once OD₆₀₀ reached 1.0-2.0, 200 μL liquid culture was spread onto 2xSG agar plates (Difco Nutrient Broth 16 g/L, KCl 26 mM, MgSO₄ 2 mM, MnCl₂ 0.1 mM, FeSO₄ 1.08 μM, Ca (NO₃)₂ 1 mM, Glucose 5.5 mM, Agar 15 g/L) without antibiotics. Plates were incubated upside down in plastic bags at 37 °C for 2–5 days to allow sporulation, monitored by phase-contrast microscopy. After sporulation, the plates were dried on the benchtop for approximately 2 days to allow cell lysis. Spores were scraped from the plates and transferred to 50 mL centrifuge tubes with cold MQ water. The spore suspensions were then sonicated for 1 min at full power, cooled on ice, and centrifuged at ∼8,000 rpm for 20 min to remove debris. The purity of spore preparation was checked with phase contrast microscopy.

Purified phase-bright spores at OD₆₀₀ of 5 in Milli-Q water (Milli-Q ultrapure water, resistivity ≥18.2 MΩ⋅cm, Millipore, Billerica, MA, USA) were heat-activated at 70 °C for 30 min followed by 20 min incubation on ice. Before the addition of the germinants, the germinant solutions were kept on ice. A final concentration of 1 mM L-alanine or 10 mM AGFK solution (10 mM each of L-asparagine, D-glucose, D-fructose and KCl) (Cabrera-Martinez et al., 2003; Artzi et al., 2021) was added prior to slide preparation and imaging according to Wen et al. (2019). In essence, to prepare a slide for wide field imaging, slides and two types of coverslips (size 22 mm round and size 22 mm by 30 mm rectangular) were cleaned with 70% EtOH and air dried in a vertical position minimizing dust collection. Two rectangular coverslips were prewarmed for several seconds on a 70 °C heating block, 65 μl 2% agar was deposited on top of one coverslip, while the other coverslip was placed on top of it, spreading the agar in between. The agar patch was dried for approximately 10 min, after which the coverslips were slid off each other. The agar patch was cut into a 1 × 1 cm section to which 0.4 μl of purified spore suspension with germinant added. The patch was transferred onto a round coverslip by placing it onto the patch and sliding it off. A G-frame was stuck onto the air-dried slide, onto which the coverslip was placed, closing all corners of the frame, and completing the slide for microscopy. For wide field time-lapse experiments, we used a Nikon Eclipse Ti equipped with a Nikon Ti Ph3 phase contrast condenser. Connected to it were a Nikon Plan Apo Plan Apo λ Oil Ph3 DM lens (100×, NA = 1.49, T = 23 °C), a Lumencor Spectra Multilaser (470 nm, 555 nm) using Lambda 10-B filter blocks, a NIDAQ Lumencor shutter, a Ti XY-and Z drive, and a Hamamatsu C11440-22C camera. All hardware was connected to a computer running NIS-Elements AR 4.50.00 (Build 1117) Patch 03. Before every use the condenser was adjusted to the optimal setting bringing the pixel size to 0.07 μm. A 100-min time lapse video was recorded for every slide with a 30 s interval between frames. The slide chamber was kept at 37 °C, and single-cell germination was analyzed with SporeTrackerX (Omardien et al., 2018). The percentages of germinated spores at the end of the time-lapsed session were defined as germination efficiency.

Purified phase-bright spores, normalized to an OD₆₀₀ of 1.2 in 25 mM HEPES buffer (pH 7.4), were heat-activated at 70 °C for 30 min, followed by incubation on ice for 20 min. 100 μL of the heat-activated spores were added, and germinant solutions were dispensed into a 96-well plate to a final concentration of 1 mM for L-alanine and 10 mM for AGFK. The OD₆₀₀ was monitored every 2.5 min for 100 min using a plate reader, with the plate maintained at 37 °C and agitated between measurements. The percentage of OD drop is measured against the initial OD. Every experiment was repeated 3 times and the average OD drop for each condition is shown.

CaDPA release was measured using a fluorescence plate reader following the method of Yi and Setlow (2010). Spores, at an OD₆₀₀ of 0.5, were germinated with 1 mM L-alanine in 25 mM K-HEPES buffer (pH 7.4) at 45 °C without prior heat activation. At various time points, 190 μL of the sample was mixed with 10 μL of 1 mM TbCl₃, and relative fluorescent units (RFU) were recorded using a plate reader.

To prepare samples for western blotting, the spores were first decoated and then lysed as follows. 50 ODs of spores were pelleted by centrifugation. The pelleted spores were resuspended in 1 mL TUDSE buffer (8 M Urea, 50 mM Tris-HCl, pH 8.0, 1% SDS, 50 mM DTT, 10 mM EDTA) and incubated for 45 min at 37 °C, centrifuged (3 min, max rpm, room temperature) and resuspended in 1 mL TUDS (8 M Urea, 50 mM Tris-HCl, pH 8.0, 1% SDS) buffer. The suspension was then incubated for another 45 min at 37 °C, centrifuged (3 min, max rpm, room temperature) and then washed six times by centrifugation (3 min, max rpm, room temperature) and then all suspended in 1 mL TEN buffer (10 mM Tris-HCl pH 8, 10 mM EDTA, 150 mM NaCl). The final decoated spores were resuspended in 1mL water if samples were not immediately given further treatment. 50 OD of the decoated spore preparation was treated with 1 mg lysozyme in 0.5 mL TEP buffer (50 mM Tris-HCl pH 7.4, 5 mM EDTA) containing 1 mM phenylmethylsulfonyl fluoride (PMSF), 1 μg RNase, 1 μg DNase I, and 20 μg of MgCl₂ at 37 °C for 6–8 min, and then put on ice for 20 min. Glass disruptor beads (0.10–0.18 mm, 100 mg) were added to each sample and spores were disrupted with three bursts of sonication (micro probe, medium power for 10 s, and placed on ice for 30 s between bursts). Following the final sonication, samples were allowed to settle for 15 s and 100 μl of the upper liquid was withdrawn and added to 100 μl 2× Laemmli sample buffer (BIO-RAD cat. #1610737) containing 5% (v/v) 2-mercaptoethanol and 1 mM MgCl₂ and incubated at 90 °C–95 °C for 3 min. This was saved as the total lysate and ready to run on SDS-PAGE. 20 μg of total protein was loaded onto each lane of SDS-PAGE on a 10% acrylamide gel (BIO-RAD cat. #4561034) and run 45 min at 60 V followed by 45 min at 110 V. After SDS-PAGE, proteins on the gel were transferred to a 0.22 μm PVDF membrane. The membrane was then blocked for 30 min by 2.5% low fat milk in TBST buffer (20 mM Tris, 150 mM NaCl, 0.1% Tween 20) and incubated with 1:3000 anti-GerAA antibody overnight (Ramirez-Peralta et al., 2012). After washing three times with TBST, the membrane was incubated with 1:2500 diluted goat anti rabbit HRP antibody (BIO-RAD cat. #1706515) for an hour before visualization.