The first step of reconstituting silk is to boil the cocoons in aqueous sodium carbonate. Until Ben Partlow and I looked more closely at this step, the range of boil times varied haphazardly from 5 minutes to 2 hours. Using gel electrophoresis (a) of reconstituted silks with varying degrees of boil time, and some image analysis, we show that the molecular weight distribution broadens and its mean shifts to lower molecular weights as boiling/extraction time increases (b).
The molecular weight distributions in (b) are summarized by the red curves in (c), with the maximum and the middle 50% covered by the error bars. This change may not seem so significant, but when you use this range to calculate the overlap concentration (black lines in (c)) for the denatured protein, the changes become enormous. From a practical point of view, a sample at a concentration of 40mg/mL is either dilute or semidilute if you are using silk that was boiled for 60 or 10 minutes respectively. As a consequence, these two samples will behave very differently.
We wanted to know if we could predict this polydispersity by using Monte Carlo simulations where we computationally “cleaved” the protein at random or particular amino acids. And depending on the number of cuts per protein, we could create a molecular weight distribution since we know the amino acid sequence of the protein. An example of this is shown in (d) for randomly cut protein, and the symbols represent ensembles with a different number of cuts per protein. Its not perfect, but it matches qualitatively well with the changes in distribution shape in (b).
Now, you can choose the boil time you want depending on either the desired molecular weight distribution or the rheological behavior for your working concentration. Our publication can be found here.