Abstract
Surface layers (S-layers) are protective protein coats which form around all archaea and most bacterial cells. Clostridium difficile is a gram-positive bacterium with an S-layer covering its peptidoglycan cell wall. The S-layer in C. difficile is constructed mainly of S-layer protein A (SlpA), which is a key virulence factor for these bacteria and an absolute requirement for disease. Studying the formation and maintenance of the C. difficile S-layer will allow the discovery of specific therapeutic interventions in the future. Here we use microscopy to examine the subcellular localization of S-layer growth and SlpA secretion. We observed formation of S-layer at specific sites that coincide with cell wall synthesis, while conversely the secretion of SlpA from the cell is delocalized. We conclude that this delocalized secretion of SlpA leads to a pool of precursor in the cell wall which is available to repair openings in the S-layer formed during cell growth or following damage.
Introduction
Clostridium difficile infection (CDI) of the gut is the major cause of antibiotic associated diarrhoea (Hull & Beck, 2004) and can lead severe inflammatory complications (Napolitano & Edmiston, 2017). This gram-positive bacterium has its cell wall encapsulated with a surface-layer (S-layer). The S-layer is a proteinaceous, paracrystalline array which acts as a protective semipermeable shell which can interact with the host (Merrigan et al., 2013) and is essential for virulence (Kirk et al., 2017). In C. difficile the S-layer consists mainly of SlpA, the most abundant surface protein (Wright et al., 2005). SlpA is produced as a pre-protein (Figure 1A) that is secreted and processed by Cwp84 into low molecular weight (LMW) and high molecular weight (HMW) SLP subunits (Kirby et al., 2009),(Figure 1B). These two subunits form a heterodimeric complex which can then be incorporated into the crystalline lattice of the S-layer, that is anchored to cell wall polysaccharide PS-II via three cell wall binding (CWB2) motifs within the HMW region (Willing et al., 2015) (Figure 1A).
The production and secretion of S-layer components are energetically expensive for the cell, suggesting that the process will display evolved efficiency. However, it is not yet clear how S-layer formation is spatially regulated and whether SlpA is targeted to areas of cellular growth before or after secretion (Figure 1C). C. difficile express two homologs of the E. coli cytosolic protein export ATPase, SecA: SecA1 and SecA2 (Fagan & Fairweather, 2011). These two SecAs are thought to promote post-translational secretion through the general secretory (Sec) pathway. SecA2 is required for efficient SlpA secretion (Fagan & Fairweather, 2011) and is encoded adjacent to slpA on the chromosome (Monot et al., 2011). It has been shown that some SecA2 systems secret specific substrates (reviewed by (Bensing, Seepersaud, Yen, & Sullam, 2014)) which may ease the burden on the general Sec system and help to spatially or temporally regulate secretion.
As an obligate anaerobe (Ransom, Ellermeier, & Weiss, 2015), C. difficile has been notoriously difficult to visualize using standard microscopy techniques with commonly used oxygen-dependent fluorescent proteins and this is further complicated by intrinsic autofluorescence in the green spectrum. To circumvent these problems, we have used a variety of labeling techniques to avoid the requirement for oxygen maturation and any overlap with autofluorescence. Using fluorescence microscopy, we identified areas of S-layer biogenesis and SlpA secretion to determine if this S-layer component is specifically targeted to growing parts of the cell. Firstly, we probed the localization of newly synthesized S-layer which was detected at discrete regions which coincided with areas of new cell wall biosynthesis. We continued by studying the internal localization of SecA2 and SlpA, discovering that SlpA is secreted all over the cytoplasmic membrane. Having observed delocalized secretion of SlpA yet localized new surface S-layer we conclude that there is a pool of SlpA that resides within the cell wall which is available to seep out and construct regions of the developing S-layer.
Results
Newly synthesized S-layer co-localizes with areas of new cell wall
During exponential growth, C. difficile cells are constantly growing and dividing, requiring the production of new peptidoglycan cell wall and the accompanying S-layer. Peptidoglycan can be labelled by growing C. difficile cells in the presence of the fluorescent D-amino acid, HCC-amino-D-alanine (HADA), (Kuru et al., 2012). Chasing cell wall staining with unlabeled media reveals sites of newly synthesized peptidoglycan that appear less intense for HADA at dividing septum (Figure 2S1). Combining this with the inducible expression of the immunologically distinct SlpAR20291 in C. difficile strain 630 (Figure 2S2A), allowed areas of newly assembled S-layer to be visualized by immunofluorescence (Figure 2A and 2S2B). Tracing the intensity of cell wall staining with the signal from newly synthesized surface SlpA reveals a crude anti-correlation (Figure 2B and 2S3) and suggests that new S-layer is formed at areas of newly formed underlying cell wall. The areas of newly synthesized cell wall that are void of SlpAR20291 signal are likely to be filled with endogenous SlpA630 that is expressed at much higher levels, observed in extracellular extracts (Figure 2S2A).
During cell division, at the septum, a large amount of new surface SlpA can be detected (Figure 3). This staining pattern suggests that S-layer is actively being formed on the mother cell over the newly synthesized cell wall, preparing each daughter cell with a new S-layer cap before cell division is complete. Numerous cells display areas of new cell wall at one of their poles that co-insides with new S-layer staining (Figure 3), we interpret these as new daughter cells that have completed cell division during the HADA stain chase. New polar S-layer can be sorted into three categories: staining distributed over the whole cell cap, on the tip of the cap or at the sides of the new cap close to the older cell wall (Figure 3). Daughter cells with their poles completely covered in new S-layer have most likely expressed SlpAR20291 throughout cell division and have SlpAR20291 distributed all over the new S-layer cap. The apex of the cell cap marks the final place of new daughter cell formation and those caps stained just at the tip have probably expressed SlpAR20291 towards the final stages of division as the two daughter cells separate and the cap is completed. Areas stained at the connecting edge between the pole and the main body of the cell must represent areas of growth once the S-layer cap was fully formed when cell division was completed. Together, these staining patterns support the hypothesis that S-layer is assembled on the mother cell at the septum to form polar caps for the daughter cells to maintain a continuous protective barrier following cell separation.
SecA2 localization in C. difficile
As newly synthesized S-layer is formed at specific points on the cell surface (Figure 3) we wanted to identify if these areas correlate with concentrated points of SlpA secretion from the cytosol. Having designated sites of secretion would allow the efficient targeting of S-layer precursors to where they are needed. As it has been shown that SecA2 is essential for cell survival (Dembek et al., 2015) and performs a critical role in SlpA secretion (Fagan & Fairweather, 2011), we assumed that intracellular positioning of SecA2 will reveal where SlpA is secreted. To confirm this, we set out to create a functional, fluorescently tagged SecA2 for monitoring SecA2 localization by microscopy. A C. difficile strain 630 mutant was generated that encodes a C-terminal SNAP-tagged SecA2 (SecA2-SNAP) on the genome in the original locus and under the control of the native promoter. SecA2-SNAP was the only SecA2 protein detected in membrane fractions by western immunoblot analysis (Figure 4S1A) and, when stained with TMR-Star, this protein species was the only one visualized by in-gel fluorescence (Figure 4S1B). Cells expressing SecA2-SNAP displayed similar growth dynamics to the wild-type parental strain (Figure 4S1C), suggesting that the fusion protein is fully functional as SecA2 is essential for growth. Imaging cells by widefield microscopy revealed that SecA2 is distributed throughout the cell and not localized to specific areas (Figure 4A). Higher resolution, Airyscan confocal images displayed the same widespread distribution but with pockets of higher intensity signal (Figure 4B). By combining SecA2 localization with HADA chase staining and new S-layer labelling, no correlation between SecA2 within the cell and areas of newly synthesized S-layer on the cell periphery could be determined (Figure 4B). Together these data suggest that SecA2 is not the determining factor in localization of new S-layer growth.
SlpA secretion from C. difficile
Although SecA2 was visualized throughout the cell, SecA2 has additional secretory substrates so it is possible that secretion of SlpA itself may be localized. Although immunofluorescence has been used to detect surface SlpA, S-layer pore size is thought to be too small to allow access of antibodies to protein located in the cell wall or indeed within the cell (Fagan & Fairweather, 2014). To discover where SlpA is being secreted from the cell, two different SlpA fusion proteins were constructed, an SlpA-SNAP fusion that can be secreted and is found in the extracellular fraction and a SNAP tagged SlpA-dihydrofolate reductase (SlpA-DHFR) fusion that associates with the membrane cell fraction (Figure 5A & 5S1A). DHFR is a fast folding protein that has been used to block and probe protein translocation mechanisms for many years (Arkowitz, Joly, & Wickner, 1993; Bonardi et al., 2011; Eilers & Schatz, 1986; Rassow et al., 1989). Expressing SlpA-DHFR decreases the secretion of native C. difficile extracellular proteins (Figure 5S1B) and leads to the build-up of precursor SlpA within the cell (Figure 5S2A). The decrease in extracellular proteins is also dependent on the SlpA-DHFR signal sequence (Figure 5S1B). Together these findings show that SlpA-DHFR fusion used in this investigation is specifically targeted to and occupies the same secretory channel required for wild-type SlpA secretion. To prevent protein secretion, the DHFR domain must first fold correctly (Arkowitz et al., 1993) and will therefore only prevent secretion via the post-translational pathway. The lack of SlpA-DHFR signal in the extracellular fraction (Figure 5S1A) shows for the first time that SlpA is exclusively post-translationally translocated. Using these two SNAP fusion proteins we can now probe the intercellular localization of SlpA secretion (SlpA-DHFR-SNAP) and localization once secreted (SlpA-SNAP).
Widefield images display a diffuse SlpA-SNAP signal on the cell surface with a halo of TMR-Star signal surrounding most of the cells (Figure 5B and C). Surface intensity plots reveal a broad cross-section of TMR-Star intensity across the cell width (Phase vs TMR-Star signal, Figure 5D). Within these cross-sections there are flat peaks of TMR-Star signal with smaller peaks towards the cell periphery (Figure 5D), this correlates with the large in-gel florescence signal observed for SlpA-SNAP-TMR-Star in the extracellular fraction of cells (Figure 5S1A). Cells expressing SlpA-DHFR-SNAP show some heterogeneity of expression (Figure 5E and F), perhaps caused by leaky expression leading to a negative selective pressure for the plasmid and the drastic effects this DHFR fusion protein has on secretion (Figure 5S1C). Again, the signal from SlpA-DHFR-SNAP appears diffuse throughout the cell and not located at specific sites (Figure 5E and F). Signal intensity traces reveal a narrow TMR-Star signal peak towards the interior of the cell (Figure 5G), suggesting a more intracellular location for SlpA-DHFR-SNAP than SlpA-SNAP which fits with SlpA-DHFR-SNAP being trapped in the cell at the membrane (Figure 5S1A). The distribution of signal in these images suggest that SlpA secretion occurs all over the cell and not just at sites where new S-layer is being formed.
Discussion
For an S-layer to function correctly it must completely encapsulate the cell (de la Riva, Willing, Tate, & Fairweather, 2011; Kirk et al., 2017). We propose here that S-layer is assembled at areas of newly synthesized peptidoglycan to maintain a stable S-layer that continually protects the C. difficile cell. Although newly synthesized SlpA is secreted from all regions of the cell, only a relatively small proportion of this was detected at the surface. This irregularity suggests that C. difficile possess reserves of SlpA beneath the S-layer in the cell wall (Figure 6). Although excess SlpA production and storage will be quite energetically expensive for the cell, this reservoir of SlpA could provide a positive fitness advantage by allowing cells to respond quickly to repair gaps in this important barrier (Figure 6). Examples of self-repair mechanisms are present thorough all forms of life from intracellular vesicular mediated membrane healing (McNeil & Baker, 2001; Tang & Marshall, 2017) up to a tissue level such as wound healing (Greaves, Ashcroft, Baguneid, & Bayat, 2013). In addition to allowing rapid repair, having a stockpile of SlpA in the cell wall, C. difficile may also create a buffer so less de novo SlpA translation and translocation is required to safely complete cell division. The amount of SlpA that lays below the S-layer should be studied further to determine what proportion of extracellular SlpA is part of the S-layer and how much is in reserve.
Although the S-layer is a rigid structure (Mescher & Strominger, 1976), fractures in the S-layer must form to allow the cells to grow. Our data suggests that new SlpA emerges through these gaps to be incorporated into the crystalline lattice. Higher resolution imaging techniques may allow direct observation these gaps in the S-layer and how the separate S-layer sections assemble. Future work could also examine the localization of intracellular cytoskeletal and motor proteins that power cell growth (Colavin, Shi, & Huang, 2018) and how the localization of these relate to areas of newly forming cell wall and S-layer along the lateral wall.
We have also demonstrated for the first time that SlpA is secreted post-translationally. Proteins transported in this way usually interact with cytosolic chaperones that prevent folding prior to translocation (Kim & Kendall, 2000). The identity of these chaperones and the exact role SecA2 plays in SlpA secretion has yet to be determined. Since SlpA must undergo a post-secretion protease modification (Figure 1) (Kirby et al., 2009), having a dwell time in the cell wall will allow time for correct processing. However this also poses the question of how S-layer components located there are prevented from oligomerizing (Takumi, Koga, Oka, & Endo, 1991) prior to assembly at the surface. It is tempting to speculate that the S-layer assembly pathway may also involve extracellular chaperones to facilitate processing and targeting while preventing premature self-assembly. The revelation that there is a pool of SlpA in the cell wall and the accessibility of the cell wall to drugs may provide opportunities for the identification of novel narrow spectrum targets that affect the assembly of this essential virulence factor.
In summary, we have found that S-layer is formed at sites of cell wall synthesis and there is an underlying supply of the S-layer precursor, SlpA, located throughout the cell wall.
Methods
Media and Growth Conditions
All strains, plasmids and oligonucleotides used in this investigation are displayed in Table 1. CA434 and NEB5α E. coli were grown in LB broth or on LB agar supplemented with 15 µg/ml chloramphenicol for plasmid selection. C. difficile were grown in reduced TY (3% Bacto tryptose, 2% yeast extract) broth or on Brain Heart Infusion agar under strict anaerobic conditions. Cultures were supplemented with 15 µg/ml thiamphenicol when selecting for plasmids.
For SlpA-DHFR-Strep expression, bacteria were subcultured from overnight cultures to an OD600 of 0.05, grown for 30 minutes then supplemented with 200 µM methotrexate. Bacteria were then grown for a further 30 minutes before expression of SlpA-DHFR-Strep was induced with 20 ng/ml anhydrotetracycline (Atc). Bacteria were grown for a further 3 hours before harvesting at 4,000 xg, for 10 min at 4°C.
Molecular Biology
Chemically competent E. coli were transformed by heat shock using standard methods and plasmids were transferred to C. difficile strain 630 by conjugation using the E. coli donor strain CA434 (Kirk & Fagan, 2016). Standard molecular biology techniques were used for restriction digest and ligations. DNA modifications were performed using Phusion High-Fidelity DNA Polymerase (Thermo Fisher) and Q5 Site-Directed Mutagenesis Kit (NEB) as per manufacturers’ instructions.
Gene editing of C. difficile 630 was achieved through allele exchange using a construct based on pMTLsc7315 as described by Cartman et al (Cartman, Kelly, Heeg, Heap, & Minton, 2012).
Microscopy
TMR-Star Staining: cells were grown from an OD 0.05 to 0.4 and treated with 250 µM TMR-Star for at least 30 minutes. Transient expression of SlpA-SNAP or SlpA-DHFR-SNAP was induced for 10 minutes with 10 ng/ml Atc or 20 ng/ml for 1 hour for in-gel fluorescence experiments.
HADA Staining: Cells were grown to an OD600 of approximately 0.1 before the addition of 0.5 mM HADA and continued growth for at least 2 hours to an OD600 of 0.5-0.6. To chase the HADA staining, cells were subsequently harvested at 4,000 x g for 5 minutes, with care being taken not to expose the cells to air, washed once by resuspension in 8 ml reduced TY and finally resuspended in 2x the original volume of reduced TY media. The cells were then grown for 25 minutes before inducing the expression of SlpAR20291 with 100 ng/ml Atc for the final 5 minutes.
Fixation: Cells were harvested at 4,000 x g for 5 minutes at 4°C, washed two times in 1 ml ice cold PBS with spins at 8,000 x g for 2 min at 4°C before being fixed with 4% paraformaldehyde in PBS for 30 minutes at room temperature. After fixation, cells were washed three times in PBS. For immunofluorescence the fixed cells were blocked overnight with 3% BSA in PBS at 4°C. Cells were harvested at 8,000 x g for 2 min at 4°C, resuspended in 1:500 Primary antibody (Mouse Anti-027 SlpAR20291 LMW SLP) and incubated at room temperature for 1 hour. Cells were then washed three times in 1 ml 3% BSA in PBS before being resuspended in 1:500 secondary antibody (Goat anti-mouse-Cy5, Thermo Fisher). Cells were incubated for 1 hour at room temperature then washed three times in 3% BSA in PBS before being resuspended in PBS. Washed cells were dried down to glass cover slips and mounted with SlowFade Diamond (Thermo Fisher).
Images were taken on a Nikon Ti eclipse widefield imaging microscope using NIS elements software or a ZEISS LSM 880 with Airyscan using ZEN imaging software. Image J based FiJi was used for image analysis.
Cell Fractionation
Extracellular glycine preparation: Cells were harvested at 4,000 x g for 5 minutes at 4°C. In all the following wash steps bacterial cells we spun at 8,000 xg for 2 minutes. Pellets were washed twice by resuspension in 1 ml ice cold PBS. Cells were treated with 10 µl per OD600U of extraction buffer (0.2 M Glycine, pH 2.2) and incubated at room temperature for 30 minutes to strip extracellular proteins. Stripped cells were harvested and the supernatant, containing extracellular protein, was taken and neutralized with 0.15 µl 1.5 M Tris pH 8.8 per 1 µl extract. The stripped cells were washed twice in 1 ml ice cold PBS before being frozen at −80C. Cells were thawed and then resuspended in 11.5 µl per OD600U cell lysis buffer (PBS, 1x protease inhibitor cocktail, 5 mM EDTA, 20 ng/ml DNase, 120 mg/ml purified CD27L endolysin (Mayer, Garefalaki, Spoerl, Narbad, & Meijers, 2011)). Lysis was induced by incubating at 37°C shaking for 30 minutes. Cell membranes were harvested by centrifugation at 20,000 xg for 20 minutes and the soluble intracellular protein fraction retained before the pellet was washed twice with 1 ml PBS. Membranes were solubilized using 11.5 µl per D600U solubilization buffer (1xPBS, 1x Protease Arrest, 5 mM EDTA, 20 ng/ml DNase, 1.5% sarkosyl) and agitated by rotating for 1 hour at room temperature. Insoluble material was harvested at 20,000 xg for 5 minutes and the solubilized membrane fraction was taken. Alternatively, for the SNAP tagged SlpA constructs; cells lysates were supplemented with 1.5% sarkosyl, incubated for 1 hour and harvested at 20,000 xg for 5 minutes to create a total cellular extract.
Protein Gels
Proteins were separated using standard SDS-PAGE techniques on a mini-protein III system (Bio-Rad) before being either; analyzed for in-gel florescence on a ChemiDoc imaging system (Bio-Rad), stained with Coomassie or transferred to nitrocellulose membranes using a semi-dry blotter (Bio-Rad) for western blot analysis. Band intensities were measured using Image Lab Software (Bio-Rad).
Statistics
Statistics were performed in Origin using one-way analysis of variance (ANOVA), a difference with p ≤ 0.05 was considered significant.
Acknowledgements
We would like to thank Darren Robinson and Christa Walther at The Wolfson Light Microscopy Facility at the University of Sheffield for their help with microscopy.
This work was supported by the Medical Research Council (grant number MR/N000900/1) and the Wellcome Trust (grant number 204877/Z/16/Z).