Temporal and spatial dynamics of C. obsidiansis biofilm formation
To visualize the process of biofilm formation by C. obsidiansis on a model cellulose substrate, cells were grown in serum bottles containing a regenerated cellulose chad as the sole carbon source. Based on imaging data, the dynamic process of biofilm formation and growth can be differentiated into multiple steps, including: i) initial cell attachment to the substrate; ii) cell growth and division and iii) inverted colony formation; iv) crater-like depression formation due to degradation of the cellulose substrate by the microbial colony; v) fusion of the depressions due to continued growth and substrate degradation, leading to vi) a biofilm of uniform thickness.
Initial microbial attachment and growth
Initial attachment by C. obsidiansis to the cellulose substrate occurred during the first 16 h of incubation in the serum bottles. By 8 h after inoculation, single cells were observed randomly attaching to the cellulose surface (Figure 1b). These cells appeared to grow by cell division on the surface, forming small clusters of cells (Figure 1c). A three-dimensional reconstruction of one representative cluster is shown in Figure 2a. These data suggest that the cells are likely distributed as a monolayer on the cellulose surface. This observation is supported by a cross-sectional view of the cluster (Figure 3a). Interestingly, it appears that many of the cells are positioned vertically on the cellulose surface (Figures 2a and 3a). Whether this positioning is due to physical crowding of the cells on the surface or is the result of a specific attachment mechanism is the focus of ongoing studies.
Inverted colony formation
By 24 h after inoculation, the formation of C. obsidiansis colonies was observed on the cellulose substrate (Figure 1d). The diameters of the colonies varied in size. Three-dimensional reconstructions of colony morphology revealed that these colonies were inverted; that is, the colonies were growing into the cellulose substrate rather than on the surface (Figure 2b). This inverted colony morphology can be seen clearly in the cross-sectional view (Figure 3b, Additional file 1). Measurements taken from this perspective indicate that the radius of the colony is larger than its height, with the width at 35 μm but the maximum depth at 10 μm. The formation of inverted colonies is likely due to cellulose hydrolysis by C. obsidiansis.
Additional file 1:
C. obsidiansis biofilm formation at 24 h. Visualization of the three-dimensional structure of an inverted colony of C. obsidiansis growing into regenerated cellulose substrate at 24 h (MPG 462 KB)
Formation and fusion of crater-like depressions
As the experiment continued, the dimensions of the colonies continued to grow. By 44 h after inoculation, large depressions about 50 μm in width were observed on the cellulose substrate with adjacent depressions beginning to fuse (Figure 1e). Smaller depressions were also seen at this stage (Figure 1e). Three-dimensional reconstructions indicated depressions in the cellulose substrate were lined by C. obsidiansis cells (Figure 2c, Additional file 2). Measurements from a cross-sectional view indicate that the maximum biofilm thickness in the depression was about 10 μm (Figure 3c). By 48 h, multiple individual depressions had fused (Figure 1f) and by 56 h, the cellulose substrate was dominated by large, irregular (approximately 200 μm) depressions into the substratum (Figure 1g). From this point on, individual depressions could not be distinguished and the surface of the cellulose substrate was covered with a thin biofilm (Figure 1h). A three dimensional reconstruction of the cellulose substrate after 68 h incubation shows a rather uniform surface without any prominent cavities or depressions as seen in earlier time points (Figure 2d). The cross-sectional view shows that the biofilm thickness remains constant at approximately 10 μm after 68 h growth on the substrate (Figure 3d). At this point, it appears that a dynamic equilibrium was reached between biofilm growth and detachment, stabilizing the biofilm thickness at a constant value. Moreover, the cell density measured in this mature biofilm is about 1.69 × 1011 cells cm-3, which is much greater than the cell density typically found in a biofilm grown on a soluble substrate (Zhang and Bishop 1994; Ito et al. 2002).
Additional file 2:
C. obsidiansis biofilm formation at 44 h. Visualization showing the three-dimensional structure of crater-like depressions formed by C. obsidiansis on regenerated cellulose at 44 h (MPG 514 KB)
Cellulose hydrolysis
It should be emphasized that the regenerated cellulose chad provides the sole carbon source for C. obsidiansis growth in this study. Hence, the hydrolysis of the cellulose chad occurs concurrently with biofilm formation. The change in chad thickness can be used as an indicator of cellulose hydrolysis and was measured throughout the experiment. The first measurable reduction in chad thickness was observed after 24 h incubation, which corresponds to the formation of inverted colonies (Figure 4a). From this point on, the cellulose chad thickness decreased at a nearly constant rate (Figure 4a). After 72 h incubation, the cellulose chad displayed significant degradation with irregular holes being visible (Figure 4c) in comparison with the new chad at the 0 h time point (Figure 4b). Our previous work indicated that a C. obsidiansis biofilm growing on cellulose generates more hydrolysate than it can utilize in order to establish an intra-biofilm substrate concentration high enough to support growth (Wang et al. 2011). The excess hydrolysate diffuses through the biofilm and is released into the bulk liquid where it can support planktonic cell growth (Wang et al. 2011).
Biofilm formation on linter cellulose
Although regenerated cellulose chads provide an ideal platform to image the process of biofilm formation and cellulose utilization (Figures 1, 2 and 3), it was unknown whether biofilm formation and degradation on natural cellulose occurred in the same manner. To address this question, a similar experiment was performed using linter cellulose, which is a natural cotton fiber containing higher crystallinity than regenerated cellulose (Gümüskaya et al. 2003). In order to create a flat surface for microscopy, linter cellulose chads were fabricated through a high-pressure homogenization method and used as the sole carbon source to culture C. obsidiansis. As with regenerated cellulose, biofilm growth on linter cellulose was characterized by the formation and fusion of depressions on the surface (Figure 5). C. obsidiansis biofilm formation on linter cellulose, however, was much slower than on regenerated cellulose, requiring four days to reach a growth stage comparable to 24 h growth on regenerated cellulose (compare Figure 3b with Figure 5). The higher crystallinity of linter cellulose likely accounts for this slower biofilm formation and cellulose degradation.
Biofilm formation by C. thermocellum
In C. thermocellum, the cellulosome is thought to play important roles in promoting bacterial attachment to cellulose and in cellulose hydrolysis (Adams et al. 2006). C. thermocellum was used as a model cellulosome-producing organism to compare whether the presence of cellulosomes altered the dynamics of biofilm formation on cellulose compared to non-cellulosome producing bacteria. In this study, C. thermocellum was grown with regenerated cellulose chads as the sole carbon source. Results showed a very similar biofilm formation process to that of the C. obsidiansis, characterized by the formation of depressions in the cellulose substrate (Figure 6).