Acid-induced assembly of a reconstituted silk protein system

Silk cocoons are reconstituted into an aqueous suspension, and protein stability is investigated by comparing the protein's response to hydrochloric acid and sodium chloride. Aggregation occurs for systems mixed with hydrochloric acid, while sodium chloride over the same range of concentrations does not cause aggregation. We measure the structures present on the protein and aggregate length scales in these solutions using both optical and small-angle neutron scattering, while mass spectrometry techniques shed light on a possible mechanism for aggregate formation. We find that the introduction of acid modulates the aggregate size and pervaded volume of the protein, an effect that is not observed with salt.

Figure 1

$37\mu \mathrm{M}$ silk protein solutions with (a) $c_H=0\to 7.4$ mM (left to right) ($0\le \mathrm{ME}\le 200$) and (b) NaCl over the same range $0\le \mathrm{ME}\le 200$. (c) Data corresponding to samples from (a): the normalized transmittance for each sample at 488 nm ($•$), the measured pH ($▴$), and the expected pH value ($-$). Transmittance for the samples used in the neutron scattering experiment ($\circ$).

Figure 2

Scattering intensity $I\left(q\right)$ of deuterated silk samples at $c_P=37\mu \mathrm{M}$: 0.5 mM NaCl ($▵$), 1.0 mM NaCl ($▴$), 0.5 mM HCl ($\circ$), and 1.0 mM HCl ($•$).

Figure 3

(a) Lithium ($\circ$) and bromine ($•$) ions per protein during twelve extra dialysis iterations beyond the normal reconstitution protocol. Iteration number one represents the ion concentrations present after the standard reconstitution protocol. (b) Lithium ions in the elution with the addition of either HCl ($\circ$) or NaCl ($▵$). (c) A common segment of the protein sequence (G-A-G-A) with lithium ions drawn to associate with carbonyl oxygens.

Figure 4

$I\left(q\right)$ changes monotonically as ME increases along the arrow from $0\to 58$ except for the two lowest values of ME: $\mathrm{ME}=0$ ($•$) and $\mathrm{ME}=1.9$ ($•$). $I\left(q\right)$ for each value of ME are fit with a double Guinier-Porod model and are plotted as the solid lines.

Figure 5

(a) Aggregate $R_H$ as determined from DLS as a function of ME. The measured samples for $\mathrm{ME}>9$ ($\circ$) are used to approximate the hydrodynamic radii of the two unmeasured samples ($*$) by linear interpolation, and the error bars correspond to the propagation of uncertainty in the interpolation. (a-inset) $d_{\mathrm{aggregate}}$ determined by the double Guinier-Porod fit of $\mathrm{I}\left(q\right)$. (b) Protein $R_G$ as determined through the double Guinier-Porod fit. (b-inset) $d_{\mathrm{protein}}$ determined by the single Guinier-Porod fit.