INTOXICACION ALIMENTARIA POR SALMONELLA ENTERITIDIS

Untreated: WTA and surface proteins present

Pronase treated: WTA present

HF treated: Surface proteins present

137 Figure 4.13 Lectin labelling in S.aureus ΔclfAΔgtfAB

ΔclfA ΔgtfAB cells were treated so that surface proteins un-glycosylated and WTA (A), only WTA (B),

only surface proteins un-glycosylated (C) and neither (D) remained present on the cell surface. Brightfield light and fluorescence images show that whilst all the cells are strongly labelled with WGA, the labelling of ConA is absent (C) and (D) when compared to ‘untreated’ (A) and (B) where cells are still exhibiting WTA. This labelling pattern confirms that ConA binds to proteins

glycosylated by GtfA and/or GtfB. Each row of images represents a different sample; Scale bars represent 2µm.

Enlarged cells (inset boxes) show the same hazy punctate pattern associated with WTA in (A) and (B). Scale bar represents 1µm.

Hazy punctate pattern

Hazy punctate pattern

No ConA binding

138 4.3 Discussion

Although the localisation of WTA hasn’t been extensively studied it is fairly well established that it is present across the entire peptidoglycan moiety (Swoboda et al., 2010; Umeda et al., 1992; Wheeler R., 2012; Xia et al., 2010a). The recent suggestion that it is not present at the septum or that if present it is not yet fully polymerised (Schlag et al., 2010) is

controversial. My work has suggested that WTA biosynthesis machinery is associated with the divisome (Chapter 3). Upon AFM microscopy it was seen that the WTA appear to form a ‘furry’ layer across the entire cell surface including the ‘piecrust’ ribs that dictate the

septum (Turner et al., 2010). Indeed Umeda et al ( 1987) have described a ‘fuzzy coat that consists of fine fibres or an electron dense mass’ on the S.aureus surface that they

identified as being made of teichoic acids and proteins. This layer was also been seen in

L.lactis, E.faecalis and S.pneumoniae and had to be removed by weak HF treatment to more

clearly observe the peptidoglycan structure and annular features (septa and equatorial rings) (Wheeler et al., 2011). Furthermore AFM images and interaction maps using ConA functionalised tips, showed that the cylinder of Lactobacillus plantarum was abundant in WTAs and ‘rough’, whilst cell poles were much poorer in WTA and had a smooth

architecture (Andre et al., 2011).

As ConA has been shown to bind the WTA backbone it was possible that a ConA- fluorophore conjugate could be used to detect WTA on the surface of S.aureus cells. Microscopy of SH1000 indicated that ConA binds both wall teichoic acids and surface proteins, with HF stripped and pronase treated cells acting as suitable controls. This observation was confirmed by using ΔsrtA and ΔtarO which respectively exhibit no covalently bound surface proteins and WTA. Schlag et al. (2010) showed that ConA binds only to cells expressing tarO, whereas this study only observed abrogation of binding where both teichoic acids and proteins were removed; this may be due to a higher level of

sensitivity in my study.

The absence of binding at the septum seen in several cells within this study and reported in

S.aureus by Schlag et al (2010), could be explained either by a lack of sensitivity of the

fluorescent probe, the presence of immature WTA or because it has no access to its binding partner. My study has shown that enzymes involved with the final step (in which fully

139 polymerised WTA are linked to the peptidoglycan), interact with the divisome suggesting that mature WTA are present at the septum. The latter offers a likely explanation because peptidoglycan is more cross linked at the septum, reducing ConA access. Peptidoglycan then undergoes processing and remodelling after cell division (Turner et al., 2010). Cells were broken, allowing the fluorophore access to the septum, and then stained for WTA. Binding was seen at the cross-wall which confirms the hypothesis that absence of binding was due to access. Similar binding aberration is seen with WGA, both within this study and previously (Endl et al., 1983; Pinho and Errington, 2003), and again most likely explained by nascent peptidoglycan having tight cross linking and this decreases permeability.

The binding pattern to ConA of both surface proteins and WTA was analysed. A hazy punctate pattern was seen with the binding of WTA, this pattern was seen across all conditions where only WTA was present on the cell surface (SH1000, pronase; Δsrt, ‘untreated’; Δsrt pronase). The patterning was seen to extend around the entire cell surface. In ΔsrtA ,the ‘untreated’ sample binding was slightly hazier, than seen with either SH1000 pronase or Δsrt pronase. This could be explained by unglycosylated surface proteins or surface proteins which although aren’t covalently attached to the surface are still associated with the cell wall and not be fully extracted by SDS interfering with clear binding. Once treated with pronase, all proteinacious material is degraded and therefore only WTA binds. Surface protein binding was in a different pattern to WTA; it was seen as a distinct punctate pattern across the entire cell surface. The number of dots did not appear to be uniform although it was not seen to exceed 6, when the cells were in focus. Once again where ΔtarO, displaying only surface protein, was left ‘untreated’ the binding pattern seen was slightly hazier when compared to ‘treated’ samples.

The data gathered from the microscopy of treated SH1000 cells and mutant cells allowed the development of a schematic representation of the localisation of surface proteins and wall teichoic acids (Figure 4.11). This schematic shows that when compared with each other, WTA forms a complete layer across the entire cell with denser patches whilst surface proteins (proven to be glycoprotein) produce distinct foci.

A ‘line and dot’ pattern was described for DivIB localisation and used to help suggest its localisation to piecrust and rib features (Bottomley, 2011) (Chapter 1.6; Figure 1.8B for

140

Figure 4.15 Peptidoglycan patterning models