The C2 fusion protein is composed of the silica-binding peptide (Si4, in red), the cellulose-binding domain of Clostridium cellulovorans cellulose-binding protein A (CBPa, in green), the B domain of staphylococcal protein A (BpA, in blue.

Once we received the sequence encoding for this protein (named construction C1 or C2, size 921bp, with a C-ter or N-ter His-Tag respectively), we amplified it by PCR by using specific primers (For, Rev, see PCR protocol) (Fig. 2) and a Taq polymerase without exonuclease activity. In lanes 1 and 4 we see that a PCR product was amplified with the expected size. As negative controls, neither amplification was possible with a single primer (lanes 2, 3, 5, 6), nor in the absence of primers (lane 7) or DNA template (lane 8).  </br></br>

In order to have more DNA, we cloned it into TOPO vector (Fig. 3A), and transformed competent bacteria Escherichia coli TOP10, resulting in white clones (Fig. 3B). After bacteria culture and plasmid DNA extraction, we verified the presence of an insert by using Xba I and Hind III restriction enzymes (data not shown). After that, insert was extracted from the gel, and ligated into digested and dephosphorylated pET43.1a, the expression vector (Fig. 4A). We repeated the procedure, and we proved that our vector contained the insert by electrophoresis (Fig. 4B). Sequencing confirmed that it was the correct sequence. </br></br>

Once checked, we cloned our construct into the Escherichia coli BL21(DE3) strain, a specific dedicated strain to produce high amounts of desired proteins under a T7 promoter. Bacteria were grown on large scale (4 l), and we made a growth curve (Fig. 5). Protein expression was induced with IPTG overnight at 15°C. Protein purification was achieved using the HisTag. Owing to the intrinsic affinity of C2 for cellulose, we had to revert to a polystyrene column for purification to work. We eluted our protein using a gradient of imidazole-containing buffer, and two peaks were detected (Fig. 6). We checked the presence of proteins in the fractions by SDS-PAGE. We clearly noted the appearance of bands at about 25 kDa, the expected size of our fusion protein (24 967 Da), but also at about 50 kDa (Fig. 7). We hypothesized that it could be monomers (25 kDa) and dimers (50 kDa). Indeed, since Si4 of C2 is able to condense silicic acid, it could potentially form dimers via Si-O bonds. </br></br>

Transformed E. coli BL21(DE3) were grown in LB supplemented with carbenicillin (50 µg/mL). Time points were taken and OD600nm was measured every 20 minutes. When OD600nm was about 0.7, the culture was induced with IPTG at 0.3 mM (red arrow). </br></br>

After determining the concentration of each fraction by Bradford assay, we tested the efficiency of our protein to bind to cellulose and to catalyze the biosilification reaction. Unfortunately, the described methods to the last component, i.e. to test the ability of our protein to bind to antibodies, require high amounts of antibodies (hundreds of micrograms), making the experiment too expensive to perform.  

We first tested the ability of our protein to bind to cellulose. To do that, we used several types of cellulose: Avicell, Sigmacell, and carboxymethyl-cellulose. As described by Goldstein et al1, we mixed cellulose with an excess of competitor non-specific protein (BSA), and with or without our protein of interest (Fig. 8A). After washing and centrifugation, we harvested proteins from supernatant and the cellulose-based pellet in order to analyze them by SDS-PAGE. We clearly noted that the 25 kDa and 50 kDa proteins were retained by cellulose, instead of the non-specific BSA (Fig. 8B). Indeed, data showed that almost no monomer remained in the supernatant after the second wash, but the pellet contained most of the protein. However, some dimers seem to remain in the second washing supernatant because the initial protein concentration was too high: the binding sites were saturated. The pellet also contains a lot of the dimers. As control, we observed that BSA remained in the supernatant and didn’t bind to cellulose.  Therefore, we can conclude that our protein binds to cellulose.  

Then, we investigated whether our protein was able to catalyze the biosilification reaction. To do that, we drew inspiration for the 2011 Minnesota iGEM team and their work about Si4 to evaluate the silification process. First, we used a source of silicic acid, the tetraethyl orthosilicate (TEOS), which is an inactive form of silicic acid. By activating it in acidic conditions, we released the free silicic acid (Fig. 9A). After incubation with or without our fusion protein, we determined the quantity of free silicic acid by a spectrophotometric method, since biosilification process consumes silicic acid to form silica (Fig. 9B). We clearly observed a precipitation into the test tube, instead of the negative control (Fig. 10A). By quantifying it by molybdate assay using a standard curve (Fig. 10B), we deduced the corresponding mass of silicic acid left after silification: 33 µg. Before silification, the concentration was 208 µg/ml. The fusion protein led to the production of 175 µg of silica after 2 hours. Therefore, the silification yield after two hours is up to 84% with the protein whereas the yield without the protein is 0% (Fig. 10C). We concluded that our protein worked.