Chitosan Flocculation

Chitosan is a linear polymer composed of randomly distributed β-(1→4)-linked D-glucosamine and N-acetyl-D-glucosamine subunits. The amine group in chitosan has a pKa of 6.5 leading to a protonation at slightly acidic pH, making it a bioadhesive which readily binds to negatively-charged surfaces. Since gas vesicle surfaces are known to be negatively-charged, we speculates that the use of chitosan might help flocculate these vesicles.

Structure of chitosan (left) derived by deacetylation of chitin (right). Taken from [1]

Biotin-Streptavidin Interaction

NHS-biotin is a biotinylating reagent that reacts with primary amines (e.g. lysines) in a peptide chain and attaches a small spacer-separated biotin moiety. Since we know that two lysine residues in the GvpA monomer are exposed to the solvent, we expect a huge number of lysine residues to be available for biotinylation on the gas vesicle surface. By adding NHS-biotin to a suspension of gas vesicles, we can biotinylate their surface!

Chemical structure of NHS-biotin

Streptavidin is a tetravalent biotin-binding molecule (each streptavidin molecule binds 4 biotin moieties) and this made us propose a biotin-streptavidin mediated strategy for increasing the effective hydrodynamic radius of these vesicles. The streptavidin-biotin binding is one of the strongest non-covalent interactions known to biology. Avidin — another notable biotin-binding protein — has an even higher affinity for biotin, but streptavidin is a better biotin-conjugate binder, i.e. streptavidin has higher affinity to biotin when the biotin is conjugated to another molecule. Also, significant non-specific binding can be prevented by using streptavidin.

Tetrameric structure of streptavidin with two bound biotin molecules

By incubating biotinylated gas vesicles with free tetravalent streptavidin, we should be able to cluster gas vesicles, increase their hydrodynamic radius and cause them to float up faster!

SpyCatcher-SpyTag Binding

SpyCatcher and SpyTag are two parts of a split protein generated by splitting a fibronectin-binding protein from Streptococcus pyogenes (Spy) into SpyCatcher, a 15.2 kDa protein and SpyTag, a short (13 aa) peptide. These two pieces, when mixed in solution, spontaneously form an isopeptide bond over a wide range of temperatures, pH values, buffers and even with non-ionic detergents!

By having two distinct populations of gas vesicles — one expressing SpyCatcher (fused to GvpC) and another expressing SpyTag on the surface (fused to GvpC) — and mixing them when desired, these heterodimerizing proteins should cause the vesicles to aggregate!

SEM

Electron microscopy

Multiple dilutions of pure gas vesicles suspended in PBS were imaged using a scanning electron microscope after applying a 10 nm gold sputter. In the images, gas vesicles can be seen as translucent polygon-shaped particles. Note that some lysed gas vesicle membranes are also seen in the image owing to the drying step during the sample preparation that precedes electron microscopy. Air-drying can be carried out over a longer period of time to reduce the frequency of such lysis events. Three dilutions were prepared for microscopy; out of these the 0.01 µg/µl samples gave the best results.

Dynamic Light Scattering

Gas vesicle suspensions prepared as in the spectrophotometry assay were used to perform dynamic light scattering. Three replicates of each concentration were run through the machine thrice. It was noted that the average particle size decreased after every run — indicating that the particles were either sedimenting or floating up.

To assay biotin-streptavidin aggregation, the gas vesicles were centrifuged at 500 rpm overnight (9 hours) and the top layer was resuspended in 1 mL PBS. This was done to remove the excess biotin that might bind with the streptavidin. Different concentrations of streptavidin were then added to these suspensions and they were analysed using a DLS machine.

The theory behind dynamic light scattering becomes quite simple if the implications of Einstein's Brownian motion hypothesis are well known. Smaller particles tend to get a stronger "kick" when a solvent particle hits them. What the system actually detects are the correlations that persist in the scattered intensities at consequent time intervals. A large correlation implies that the particle hasn't moved much in the interval and is hence larger.

The actual values obtained from the system are those of the translation diffusion coefficient, to which the software applies the famous Einstein relation (see Mathematical model) giving the hydrodynamic diameter,

\[ d_{H}=\frac{kT}{3 \pi \eta D} \]

where d_H is the hydrodynamic diameter and D the translation diffusion coefficient.

Verification of presence of Gas Vesicles

The easiest way to assay presence of gas vesicles is their disappearance under high pressure under a microscope. This was observed even during normal experiments. Fully filled micro-centrifuge tubes containing dilute gas vesicle suspensions lost their faint opalescence when the tube was closed, leading to a loss of samples. A stricter assay was performed using DLS and SEM Imaging to pinpoint the exact size of the nano-particles. It was found that these gas vesicles have an effective hydrodynamic radius of around 230 nm. This estimate was particularly valuable in the development of our model.

Typical H. salinarum gas vesicles measure 300 nm in length and around 200 nm in diameter. A. flos-aquae vesicles are slightly larger in size but the culturing conditions required for the algae make them harder to extract. All the gas vesicles used in our flocculation experiments were extracted from H. salinarum and were stripped of GvpC by using the 6M urea lysis method.

References