Peptidoglycan degradation machinery in Clostridium difficile forespore engulfment

Summary Clostridium difficile remains the leading cause of antibiotic‐associated diarrhoea in hospitals worldwide, linked to significant morbidity and mortality. As a strict anaerobe, it produces dormant cell forms – spores – which allow it to survive in the aerobic environment. Importantly, spores are the transmission agent of C. difficile infections. A key aspect of sporulation is the engulfment of the future spore by the mother cell and several proteins have been proposed to be involved. Here, we investigated the role of the SpoIID, SpoIIM and SpoIIP (DMP) machinery and its interplay with the SpoIIQ:SpoIIIAH (Q:AH) complex in C. difficile. We show that, surprisingly, SpoIIM, the proposed machinery anchor, is not required for efficient engulfment and sporulation. We demonstrate the requirement of DP for engulfment due to their sequential peptidoglycan degradation activity, both in vitro and in vivo. Finally, new interactions within DMP and between DMP and Q:AH suggest that both systems form a single engulfment machinery to keep the mother cell and forespore membranes together throughout engulfment. This work sheds new light upon the engulfment process and on how different sporeformers might use the same components in different ways to drive spore formation.


Circular dichroism
Affinity purification of polyclonal antibodies from rabbit serum Detection of proteins using immunoblotting

Protein stability
Wild type and point mutant versions of  and  recombinantly expressed were obtained at a high degree of purity, as shown in immunoblotting ( Figure S6A) and SDS-PAGE gels ( Figure S6B). Moreover, all proteins seem stable, as no degradation was detected ( Figure S6A, B). Analysis of CD spectra from 190 to 240 nm ( Figure S6C) indicates all proteins are folded. However, point mutations of the zinc-binding residues of SpoIID -H134, C140, H145 and C146 -lead to reduced protein stability as reflected by reduced melting temperatures (T m ), with C140A and H145A proteins exhibiting the lowest values. This effect appears to be linked to some reduction in helical content, as calculated by CDSSTR ( Figure S6C). Conversely, mutating the catalytic aspartate 101 to alanine seems to lead to increased protein stability, with Tm increasing by 3 °C. A similar stabilizing effect seems to occur when mutating catalytic residues in SpoIIP, where both mutants have a higher Tm than the wild type protein. The mutations seem to lead to decreased helical and strand content, with accompanying increase of the coiled/turn content.
Overall, it seems that, although all proteins are folded in solution, the zinc-binding mutations in SpoIID affect their relative stability which could impact both in vitro catalysis and in vivo activity.
Indeed, when analyzing in vitro PG degradation activity, C140A and H145A proteins appear mostly inactive (Figure 7), whilst H134A and C146A retain some activity.
The varying stability could also explain the observed differences in sporulation efficiency difficile where isogenic mutants were complemented with the mutation-containing isoforms. Analysis of fractionated protein extracts by probing with affinity-purified antibodies revealed that both SpoIID and SpoIIP were present in a pool of membrane associated proteins, extracted from sporulating cells ( Figure S10A, B). Signal arising from SpoIIP was also detected at similar levels in a soluble fraction of cell extract (Figure S10B, D); SpoIID however, presented differential detection pattern (Figure S10A, C). Noticeably lower amounts of SpoIID was observed for Zn-binding mutants compared to WT in the membrane fraction of lysates, highlighting a potential role of the metal in protein stability in vivo, reflecting the in vitro observations ( Figure S10C). A loss of signal in the soluble fraction of corresponding lysates ( Figure S10C) further supports the hypothesis that lower stability of these isoforms leads to increased protein degradation and resulting reduced availability. Conversely, increased amount of protein was detected for the catalytic mutant of SpoIID (ΔspoIID complemented with spoIIDE101A) when compared to WT. This indicates accumulation of a more stable ( Figure S10C) but inactive (Figure 7) form of the protein on the membrane of cells that are unable to complete engulfment and progress in sporulation ( Figure 8).
Together, these results indicate that zinc and the residues involved in metal coordination are likely to play a structural, stabilizing role in SpoIID, although not all residues seem to be essential.

Protein expression in BACTH strains
To confirm protein expression and stability in the BACTH strains, immunoblotting with the relevant rabbit-raised antibodies was carried out for selected fusion combinations. A monoclonal anti-CyaA (adenylate cyclase, see main text for details) antibody recognising T18 fragment was used to cross-validate this immunoblot analysis. Combinations where only one of the fusion orientations (T18:25 or T25:T18) resulted in a detectable interaction were tested, as well as those described in the engulfment mechanism proposed in Figure 1A. Despite cross-reaction of the polyclonal antibodies with native E. coli proteins, particularly for anti-SpoIID and anti-SpoIIQ, as well as observed degradation issues, the combined information for the protein-specific and anti-CyaA antibodies allowed us to confirm the presence of all proteins in all investigated orientations, but at varying levels of expression and stability. The fact that some interactions are only detected in one combination suggests that protein orientation might also play an important role in establishing the correct engulfment machinery.
Immunoblotting revealed clear differences in protein stability and expression levels under analyzed conditions. Notably, fusion forms of SpoIIPH142R and SpoIIIAH seem to degrade, with the full length SpoIIPH142R or SpoIIIAH protein being detected when probed with the respective antibody ( Figure S7C, purple arrowheads; G, orange arrowheads, respectively). Importantly, probing with the anti-CyaA antibody confirms that the complete fusion is still present (D, purple arrowheads; H, orange arrowheads, respectively). This partial degradation could account for the weaker interactions detected for SpoIIPH142R and SpoIIIAH ( Figure 5). SpoIIM is an integral membrane protein and no antibodies could be obtained but probing with anti-CyaA reveals the presence of the T18-SpoIIM fusions (light pink arrowheads, B and D) albeit at low levels, which could lead to the relatively low interactions observed in -galactosidase assay. T18-SpoIID fusions were clearly visible when using anti-CyaA antibodies, despite some degradation (B, pink arrowheads), confirming our tentative identification of SpoIID when probing with anti-SpoIID antibodies (A, pink arrowheads). Finally, presence of SpoIIQ in the BACTH strains was confirmed combining information from immunoblots using anti-SpoIIQ (C, blue arrowheads) and anti-CyaA (D, blue arrowheads).

Circular dichroism
Circular dichroism (CD) data were collected for soluble SpoIID and SpoIIP recombinant proteins using a JASCO J-810 spectrophotometer and a 1 mm path length quartz cuvette (Hellma), where the temperature was maintained at 20 o C, by a PTC-4235 Peltier temperature controller.
Proteins were buffer exchanged into 50 mM Na2HPO4 pH 8.0, 50 mM NaF, using Satorious Vivaspin 500 centrifugal concentrators (10,000 MWCO), as per manufacturer's instructions. Measurements were acquired for proteins at a final concentration of approximately 0.01-0.02mg/ml as confirmed by Bradford assay under the following conditions: 2 nm bandwidth, 4 second response, over 260-185nm wavelengths at 0.5 nm pitch with a scanning speed of 50nm/min. Presented sSpoIID data are the result of 4 accumulations, whereas sSpoIIP scans are the average of 10 accumulations. Spectrum scan data were corrected to a buffer-only reference scan and 10-neighbour Savitzky-Golay smoothing (Savitzky and Golay, 1964) was applied to final curves. Estimation of secondary structure composition was performed using the CDSSTR program, reference data set 4 (Sreerama and Woody, 2000) available via DichroWeb (Whitmore and Wallace, 2004). Thermostability experiments were performed by monitoring a change of CD signal of proteins (approximately 0.2 mg/ml concentration in 50 mM Na2HPO4 pH 8.0, 50 mM NaF buffer) at 222 nm between 4 and 95 o C, at 1 o C /s, with 2 nm bandwidth, 4 second response time using the aforementioned cuvette. Mean residue ellipticity was calculated and the apparent melting temperature, Tm (°C), was determined from unfolding curves using the midpoint of the sigmoidal fit of the calculated MRE values using GraphPad Prism 7 software (La Jolla California USA, www.graphpad.com). The unfolding curves were normalized to give 0 for fully folded signal at 4 °C and 1 for fully unfolded signal at 90 °C.

Affinity purification of polyclonal antibodies from rabbit serum
Purified SpoIID26-354 and SpoIIP27-339 were used to immunize rabbits for polyclonal antibody production (Moravian Biotechnology). Carbonyldiimidazole (CDI)-activated crosslinked 6% beaded agarose resin (Pierce) was used for purification of antibodies from final bleed rabbit sera via affinity to column-immobilized immunogen. A sample of 0.5 ml of beads was washed with 100 mM borate pH 8.5 buffer to remove acetone storage solution and equilibrate the resin. An aliquot of immunogen protein at 1-5 mg/ml concentration in 100 mM borate pH 8.5 buffer was added to agarose and incubated for 24h at 4°C. Unbound protein was removed and resin was incubated at 4°C for 5h with 1 ml of 1M Tris pH 8.8 and washed with PBS 0.1% (v/v) Tween20. Agarose-immobilised protein was added to 15 ml of final bleed rabbit sera and incubated with mixing at 4°C for 18h. Serum was recovered, and agarose was washed with PBS 0.1% (v/v) Tween20. IgG elution was carried out with 0.1M glycine pH 2.5 in 5 consecutive steps of 30min incubation at room temperature, centrifugation 1 min at 3,000 x g, and neutralization of recovered supernatant with 0.1ml of 1M Tris pH 8.8.
Concentration of purified antibodies was tested using NanoDrop 2000.
Recombinantly expressed and purified proteins were prepared as detailed in the main text.
Whole cell extracts of BACTH strains were prepared from 2 ml samples of cultures grown in 10 ml LB broth, supplemented with carbenicillin (100 µg/ml), kanamycin (50 µg/ml) and 0.5 mM IPTG to an OD600 of 0.5. Cells were harvested by centrifugation (5 minutes at 17,000 x g), resuspended in 100 µl 20 mM Tris pH7.5 and 150 mM NaCl with the addition of 30 µl 4x Laemmli sample buffer and boiled for 10 minutes.
For detection of SpoIID and SpoIIP proteins in C. difficile cell extracts were prepared from sporulating cultures in 5ml SM broth, supplemented where necessary with 250 ng/ml of ATc (Ptet-spoIIP variants), and grown statically for 10h at 37 °C in a DG250 workstation (Don Whitley Scientific) under anaerobic conditions (10 % H2, 10 % CO2, 80 % N2) as previously described (Fagan and Fairweather, 2011). All mutant and complemented strains were created in the 630ΔermΔpyrE background via allele-coupled exchange as described in the main text. Cells were harvested by centrifugation (10 min at 4,000 x g), pellets were washed with 1 ml of PBS and frozen. Thawed cell pellets were resuspended in PBS containing protease inhibitors, 1.4 mg/ml lysozyme and 0.12 μg/ml DNaseI to an OD600 nm of 20 and incubated at 37°C for 1 h. Membranes were harvested by centrifugation at 21,100 g for 10 min at 4°C. The supernatant containing the soluble proteins was removed and mixed with an equal volume of 2 x Laemmli sample buffer. The harvested membranes were washed twice with 500 μl PBS, resuspended in PBS and solubilized with 1% SDS to a final equivalent OD600 nm of 20 and mixed with an equal volume of 2 x Laemmli sample buffer. B A Figure S1. Conservation of SpoIIDMP sequences between B. subtilis and C. difficile (A) Pairwise alignments of SpoIID, SpoIIM and SpoIIP sequences. Sequence identity is provided above each alignment. Conserved residues deemed essential for enzymatic activity are highlighted in blue.

Figure S5. SIM analysis of SpoIIQ-SNAP and SpoIIIAH-SNAP localization in spoIIDMP mutants
In wild type and spoIIM mutants, SpoIIQ and SpoIIIAH follow the engulfing membrane as sporulation progresses. Both proteins localize at the flat septa or partially curved membrane seen in spoIID and spoIIP mutants as cells are arrested at early stages of engulfment. Although lack of D/P has no effect in the initial recruitment of SpoIIQ and SpoIIAH, a potential role in maintaining Q:AH localization later in the process cannot be excluded. SIM images of wild type and spoIIDMP mutant sporangia cells expressing either SpoIIQ (left) or SpoIIIAH (right) SNAP fusions, harvested after 14h of growth in SM broth and stained with MitoTracker Green (membrane) and TMR-Star (SNAP substrate). Scale bars corresponds to 2 μm. Images are representative of at least 3 biological replicates.

Figure S7. Immunoblot analysis of E. coli cell extracts used in BACTH
Presence of the different fusion proteins in the combination strains used in the BACTH assays was investigated using a combination of anti-SpoIID (A), anti-SpoIIP (C), anti-SpoIIQ (E) or anti-SpoIIIAH (G).
The same strain combinations were tested using a monoclonal antibody against adenylate cyclase CyaA, which recognises the T18 fragment (B, D, F, H). All proteins tested were present in these strains, despite clear differences in protein amount and stability.
BACTH strains were grown to an OD600 of 0.5, the cells harvested and extracts resolved on 12% SDS-PAGE gels which were then subject to immunoblot analysis using rabbit anti-SpoIID (A), anti-SpoIIP (B), anti-SpoIIQ (C) or anti-SpoIIIAH (D) at a 1:5000 dilution, and an anti-rabbit secondary antibody conjugated to HRP was used at dilution of 1:2500. Secondary immunoblot analyses for the T18 fragments were conducted using mouse anti-CyaA conjugated to HRP (Santa Cruz) at a 1:200 dilution.

Figure S8. Metal content analysis of SpoIID26-354 in the presence of EDTA and of SpoIIP27-339
SpoIID retained zinc at an occupancy of 1:1 after incubation with 5mM EDTA, indicating that metal coordination is relatively strong, with 5 mM chelating agent not enough to remove it. Attempts to remove zinc by increasing EDTA concentration to 20mM resulted in protein precipitation, which indicates a possible structural role for the metal ion.
SpoIIP does not contain any of the metal divalent cations tested, frequently associated with enzymatic activity in metal binding proteins. Anhydro-disaccharide and anhydro-tetrasaccharide fragments detected when digesting PG with an active SpoIIP ( Figure 7A, B, peaks 2 and 5), were no longer detectable when digesting PG with SpoIIP alone (A), indicating that they are a result of the amidase activity of SpoIIP acting on the naturally occurring anhydro-MurNAc glycan termini followed by cellosyl treatment.
H143A mutation does not completely abolish activity as a significant amount of anhydro-disaccharide (B, peak 2) is detected when treating PG with SpoIIDH143A and SpoIIP. This suggest that not all zincbinding residues are essential for activity, and that H143 together with C146, seem to have a less significant role.     This study 630ΔermΔpyrE carrying a 900 bp deletion (aa 21-320) in spoIID (CD630_01240) 630ΔermΔspoIID This study 630ΔermΔpyrEΔspoIID with pyrE restored through homologous recombination with pMTL-YN1 630ΔermΔspoIIM Pnat-spoIID This study 630ΔermΔpyrEΔspoIID complemented with spoIID under the control of its native promoter and pyrE restored through homologous recombination with pMLD101 630ΔermΔspoIIM Pnat-spoIID E101A This study 630ΔermΔpyrEΔspoIID complemented with spoIID E101A under the control of its native promoter and pyrE restored through homologous recombination with pAXK003 630ΔermΔspoIIM Pnat-spoIID H134A This study 630ΔermΔpyrEΔspoIID complemented with spoIID H134A under the control of its native promoter and pyrE restored through homologous recombination with pAXK007

Relevant details
630ΔermΔspoIIM Pnat-spoIID C140A This study 630ΔermΔpyrEΔspoIID complemented with spoIID C140A under the control of its native promoter and pyrE restored through homologous recombination with pAXK004 630ΔermΔspoIIM Pnat-spoIID H145A This study 630ΔermΔpyrEΔspoIID complemented with spoIID H145A under the control of its native promoter and pyrE restored through homologous recombination with pAXK005 630ΔermΔspoIIM