3.1. Microstructure and phase analysis
The XRD pattern in Fig.1 (a) shows a broad hump and absence of any crystalline peaks confirming that the powders calcined at 800°C (58S-C800) are completely amorphous. XRD patterns of BG58S-C900 (Fig.1 (a)) and BG58S-C1000 (Fig.1 (a)) powders calcined at 900°C and 1000°C respectively show a broad hump with emergence of peaks corresponding to pseudowollastonite (Ca3Si3O9), wollastonite (CaSiO3), quartz (SiO2) and calcite phase (CaCO3) indicating the beginning of crystallisation. In this work, we have explored the effects of this phase change on initial adhesion and antibacterial behaviour of bioactive glass.
The powder size (Fig.1b) is obtained in the size range of 50-80 nm. The presence of calcium, silicon and phosphorous is confirmed via elemental quantification (Fig. 1c). The Si:Ca:(P2) molar ratio of the powder was calculated using EDX elemental analysis. The ratio 60.68 mol% Si 34.86 mol% Ca and 4.46 mol% (P2) is observed to be consistent with bioactive glass 58S composition, thereby confirming successful synthesis of BG58S bioglass [57].
3.2. Antibacterial tests
The results of disc diffusion test show bactericidal activity of BG58S-C800, BG58S-C900 and BG58S-C1000 (Fig. 2). There is a distinct zone of inhibition around the BG58S-C800 and BG58S-C900 samples, whereas the BG58S-C1000 disc shows no zone of inhibition. The dissolution is higher for more amorphous samples (for both E. coli and S. aureus) and, consequently, the zone of inhibition is also larger.
In Fig. 3 (a), a significant decrease in E. coli viability is observed in the presence of BG58S-C800 and BG58S-C900 as when compared to the control sample. S. aureus’ growth is also greatly affected due to the presence of BG58S-C800 and BG58S-C900 (Fig. 3 (b)) but the effect is less pronounced when compared to E. coli in Fig. 3(a). This shows that BG58S-C800 and BG58S-C900 inhibit bacterial growth and they are more potent against gram-negative bacteria. Semi-crystalline BG58S-C1000, however, does not restrain the growth of either bacteria.
A higher antibacterial activity and bacterial resistance is observed for more amorphous phases calcined at higher temperatures. This can be understood by observing the change in pH with time in Fig. 3(d). Increase in pH is due to dissolution of bioactive glass in fluid which leads to leaching out of ions, demonstrated schematically in Fig. 3(c). Therefore, completely amorphous bioactive glass is more antibacterial than semi-crystalline phases calcined at higher temperatures. Also, the antibacterial effect is stronger against gram negative E. coli, this can be due to a thinner peptidoglycan layer as compared to gram positive S. aureus [58].
3.3. AFM images and biomechanical mapping
AFM images of live planktonic cells captured the structural details and biomechanical properties of the cell surface. E. coli cell (Fig. 4 A1,A2) has concentric oblong ring like patterns and S. aureus cell (Fig.4 B1, B2) has an uneven morphology with nano-ridge like patterns. These features correspond to accumulation of macromolecules on the bacterial surface which are characteristic of a bacterial strain. High resolution images of such macromolecules (proteins, lipids) have been reported in literature [59]. These macromolecules include surface proteins which are responsible for initial adhesion of bacteria.
E. coli adhesion map (Fig. 4 A3) highlights adhesive and repulsive nano-domains on the cell’s surface. Adhesion values up to 750 pN are measured in adhesive parts of the surface and repulsion up to 970 pN observed in other areas. Note that these adhesion maps are measured using silicon nitride tip in AFM tapping mode and these values are w.r.t. the AFM tips’ interaction with glass surface. This information helps us understand the findings of adhesion measurements (F-D curves) better. Similar observations are made in case of S. aureus cell surface (Fig. 4 B3), the distribution of adhesive and non-adhesive domains is not as symmetric as in case of E. coli. The highly adhesive domains show adhesion values as high as 3.8 nN and other areas have shown repulsive forces as high as 640 pN. The LogDMTModulus (logarithmic of elastic modulus based on DMT model [60] values are also mapped (Fig. 4 A4, Fig. 4 B4)) giving us a deeper insight on the effect of macromolecule placement on cell stiffness. E. coli cells are relatively larger in size providing more surface area for adhesive bonds to form. Further, the uneven distribution of adhesive domains on S. aureus cell envelop further limits the area available for adhesion [61].
3.4. Adhesion forces between bioactive glass and planktonic bacteria
Representative force-distance curves (retraction) and adhesion force histograms are summarized in Figure 5. The major and minor peaks of a retraction curve may correspond to any of the following events: breaking of receptor-ligand bond, protein unfolding, mechanical protein stretching or de-adhesion of single cell-biointerface. A closer look at the retraction curve shape and force values gives us a lot of information about molecular events at play [62].
In Fig. 5, the adhesion force values increase for all bacteria-bio mineral system with every 250 ms increment in contact time. A decrease in adhesion force observed with increase in crystallinity. The representative curves (Fig. 5 a,c,e,g) show an increase in number of minor peaks and Fig. 6 d shows an increase in number of total peaks which indicate an increase in number of bonds formed between bacteria and probe. Each peak corresponds to either of the following: a protein unbinding event, ligand-receptor bond breakage or protein unzipping depending on the shape of the retraction curve [63]. In Fig. 5a, multiple binding events are observed between E. coli and BG58S-C800 and the number of binding events increase with time. The area under the curve is also much larger (Fig. 6c) and increases with contact time. In Fig. 5e, maximum number of ligand-receptor bonds are observed indicating a very strong bonding in a small contact time, indicating a higher likelihood of transition of E. coli BG58S-C900 adhesion from reversible to irreversible. The area under the curve is higher in E. coli (Fig. 6d) indicating a higher work of adhesion required for detachment as compared to S. aureus (Fig. 5 c,g). In Fig. 5(c), protein unbinding are unzipping events are largely responsible for bond strengthening between S. aureus planktonic cell and BG58S-C800. More number of binding events are observed (Fig.5 c-inset) which indicates that more ligand receptor bonds are forming between biomineral tip and bacteria in short period of time. In Fig. 5(g), protein unzipping and unbinding are primarily responsible for adhesion strength and the number of unbinding events is the minimum of all cases indicating a relatively weak adhesion.
Fadh values are the adhesion force values corresponding to the biggest de-adhesion peak. The average Fadh and standard deviation is plotted in Fig. 6. Figure 6 (a,b) shows an increase in average adhesion force between both bacteria and BG58S-C800 by ~0.25 nN with each 250 ms increment of contact time. The spread of adhesion force values is higher for E. coli than in case of S. aureus, with adhesion force values as high as 6 nN as compared to 3 nN for S. aureus within 1 s of contact. This can be understood by revisiting the adhesion maps of both cells in Fig. 4 and observing the contrast in adhesion values over both cell surfaces. E. coli has more adhesion sites distributed throughout the cell. In general, there is a decrease in adhesion between bacteria and bio-mineral with increase in crystallinity.
A low initial adhesion indicates a lower chance of bacterial colonisation. However, a high initial adhesion does not mean a high chance of bacterial infection in case of bioactive glass, given their antibacterial mechanism (Fig. 3). The bioactive glass and glass ceramics form a much stronger bond with E. coli indicated both by the higher de-adhesion force values and larger area under the curve. Bond strengthening between bacteria and bio-mineral is comprised of short range and long range force components. Poisson analysis is used to decouple these forces and calculate the change in their contributions with contact time.
3.5. Poisson analysis
The values calculated from adhesion force analysis of E. coli and S. aureus with BG58S-C800 are listed in table 1. A negative force value indicates attraction and a positive force value indicates repulsion. The high value of R2 (~ 0.9 and higher) in most cases indicate a good fit of our values to the model in general. However, low values R2 (less than 0.5) are observed in some cases (S. aureus at 0 s and 250 ms) can be due to very low contact times at which the measurements were carried out and improve as we increase the contact times to values over 500 ms.
Short range interactions between bacteria and surface are responsible for permanent adhesion of bacteria to a surface [64]. We observe that the bond-strengthening between E. coli and bioactive glass is mainly driven by short range forces (fSR= -5.05±0.25 nN at 1 s), the long range contributions are much lower (FLR=-1.07±0.31 nN at 1 s). For S. aureus, the short range force values increase with time but not as sharply as E. coli. An effect of this was observed in adhesion measurements (Fig.5 b) where we saw the adhesion force histograms show much higher force values for E. coli adhesion. This also explains the larger work of adhesion i.e.area under the retraction curve observed for E. coli (3.17 10-15 J at 1 s) compared to that for S. aureus (3.78 10-16 J at 1 s). We can conclude, based on short range force contributions and number of peaks, that E. coli is likely to adhere much sooner than S. aureus and therefore is more likely get arrested and form a strong irreversible bond with BG58S-C800. This is due to the stronger short range interactions made possible by the many adhesive nano-domains spread over E. coli surface.
It is important to note that force values calculated using Poisson analysis comprise of forces associated with major and minor deadhesion peaks and not just the Fadh reported in Fig. 5 i.e. it takes into account all the adhesive events [51]. Therefore, the force values reported in Table 1 and Figure 7, which were calculated using Poisson analysis, may be much higher than the average values reported in figure 6, which only considers the adhesion force value of the largest de-adhesion peak.
Figure 7 schematically summarises the key aspects of this AFM study. Fig. 7 (a) illustrates the preparation of bio-mineral probe and force spectroscopy measurements. Fig. 7 (b) conveys the key findings of this study. As observed in Fig. 4, the adhesive nano-domains cover a greater area on E. coli surface than on S. aureus surface. This implies a greater number of binding sites (Fig. 6d) and more binding proteins on the E. coli surface as compared to that of S. aureus surface. The results is more number of adhesive interactions and binding events, which are indicted by major and minor peaks in representative graphs reported in Fig. 5 and total number of peaks and energy reported in Fig. 6, for E. coli planktonic cells as compared to S. aureus planktonic cells. The outcome of these interactions is the different adhesive force values reported in Table 1.