Blob-Based Predictions of Protein Folding Times from the Amino Acid-Dependent Conformation of Polypeptides in Solution

A combination of fluorescence blob model and molecular mechanics optimizations was applied to determine the number (N-blob) of amino acids (aa's) located in the volume probed by an excited pyrene for poly(L-lysine) labeled with 1-pyrenebutyric acid (PyBu-PLL), as it adopted an extended conformation in DMSO. The N-blob value of PyBu-PLL was then compared to those of other pyrene-labeled polypeptides adopting a similarly extended conformation over the length scale of a blob to quantify the contribution of the pyrenyl label to N-blob. After subtracting the pyrene contribution, the plot of the corrected N-blob as a function of aa side chain reach (SCR) led to the conclusion that, for a polypeptide in an extended conformation, increasing the SCR of an aa by one bond resulted in an similar to 1.8 aa increase in N-blob. This information could then be applied to predict the intrinsic N-blob value (N(SCR)) of any aa based on its SCR as long as this aa was part of an aa sequence in an extended conformation. The increase in N-blob resulting from the conformational freedom imparted by smaller aa's to a polypeptide backbone was accounted for by multiplying N(SCR) with the bending function f(b)(SCS), which was determined experimentally for the different side chain sizes (SCS) of aa's. By scanning the sequence of a protein, an N-blob value was calculated for each aa from its N(SCR) and f(b)(SCS) values. The N-blob values were then sorted from largest to lowest and averaged to yield < N-blob >. Renormalization group theory was applied to determine the number (Omega) of conformations that would be available to a protein based on its < N-blob > value and the number of blobs found in the protein sequence. Multiplying Omega by the time required by an aa to adopt its possible conformations yielded the protein folding time (tau(F)). A correlation coefficient of 0.73 was obtained from the comparison of the calculated and experimentally determined tF values, demonstrating the ability of this blobbased approach to predict the protein folding times within +/- 1.4 orders of magnitude from the experimentally determined value.

Automated summary

The combination of a fluorescence blob model and molecular mechanics optimizations was used to determine the contribution of pyrene label in polypeptides and predict the protein folding times. This approach showed promising results with a correlation coefficient of 0.73 between calculated and experimentally determined folding times.