Goldenberg and Creighton (1984) demonstrate that polyacrylamide gel electrophoresis, normally used to suppress conformational effects, can be inverted into a probe of protein conformation. Their key innovation is the transverse gradient gel: a slab cast with a linear gradient of denaturant (0β8 M urea) or acrylamide perpendicular to migration, so a single electrophoretic run generates a continuous profile of mobility versus denaturant concentration . This is powerful because folding transitions are cooperative and often two-state, and heterogeneity that broadens bulk unfolding measurements can be resolved as separate bands on the gel.
Using five bovine proteins (BPTI, RNase A, Ξ±-lactalbumin, Ξ²-lactoglobulin, serum albumin) in native, reduced/carboxamidomethylated (RCAM), and reduced/carboxymethylated (RCM) forms, the authors dissect mobility into charge and sieving components via Ferguson plots. Notably, they find retardation coefficients did not correlate obviously with molecular weight across their small, narrow-range protein setβcontradicting prior literatureβand that RNase's unfolded KR depended on the thiol-blocking group, honestly exposing the limits of interpreting KR as a compactness measure .
Reported values (averaged where duplicated). Unfolding slows migration for basic proteins (BPTI, RNase), but charge reversal (RCM Ξ±-lactalbumin) increases mobility β illustrating the chargeβconformation confound the authors emphasize.
Where NβU interconversion is fast on the electrophoresis timescale, the observed mobility is an equilibrium-weighted average, and extrapolation of ΞG from the transition region gave β10.0 Β± 1.5 kcal/mol for ferricytochrome c, reasonably consistent with calorimetric (β8.0) and guanidinium-extrapolation (β7.3) values . When interconversion is slow, low-temperature rapid electrophoresis resolves kinetic species: two refolding populations of RNase A (attributed to proline isomerization), four conformational states of penicillinase (N, U, and intermediates H and I), and an in vivo protrimer folding intermediate of phage P22 tail spike detected by pulse-labeling β a result the authors note would have been impossible without conformation-sensitive separation .
Strengths: intellectually honest (the authors repeatedly stress that only changes in mobility are interpretable, and that KR shape-dependence is uncertain); methods are fully described (casting protocols, catalysts, temperatures); multi-protein validation with orthogonal thermodynamic comparisons. Limitations the authors themselves flag: poor sensitivity to small conformational changes; charge changes on unfolding can exactly compensate mobility changes (e.g., partially succinylated RNase showed no apparent transition); radius of gyration of a ~60-residue random coil equals that of a fully Ξ±-helical state, so hydrodynamic volume is not a universal folding probe; zero-urea position on the gel is imprecise. Remaining blind spots: extrapolation of ΞG assumes linear denaturant dependence, whose physical basis was itself contested; only five small proteins were tested for sieving behavior, limiting generality of KR conclusions; the Ferguson-plot gradient method trades absolute KR accuracy for comparative sensitivity, which the authors acknowledge. No conflicts of interest were declared; funding was an NIH postdoctoral fellowship.
The paper's claims would be falsified if mobility patterns failed to reproduce unfolding transitions measured by calorimetry/CD across proteins and conditions β the reasonable cytochrome c agreement and RNase proline-isomer kinetics (corroborated by independent kinetic analyses) support the core validity, within its stated qualitative scope.
Know what changed, what holds up, and what remains uncertain. Every Friday. No ads.