Probing the Energy Landscape of Spectrin R15 and R16 and the Effects of Non-native Interactions.

Understanding the details of a protein folding mechanism can be a challenging and complex task. One system with an interesting folding behavior is the α-spectrin domain, where the R15 folds three-orders of magnitude faster than its homologues R16 and R17, despite having similar structures. The molecular origins that explain these folding rate differences remain unclear, but our previous work revealed that a combined effect produced by non-native interactions could be a reasonable cause for these differences. In this study, we explore further the folding process by identifying the molecular paths, metastable states, and the collective motions that lead these unfolded proteins to their native state conformation. Our results uncovered the differences between the folding pathways for the wild-type R15 and R16 and an R16 mutant. The metastable ensembles that speed down the folding were identified using an energy landscape visualization method (ELViM). These ensembles correspond to similar experimentally reported configurations. Our observations indicate that the non-native interactions are also associated with secondary structure misdocking. This computational methodology can be used as a fast, straightforward protocol for shedding light on systems with unclear folding or conformational traps.

Non-Native Cooperative Interactions Modulate Protein Folding Rates.

The energy landscape theory and the funnel description have had remarkable success in describing protein folding mechanisms and function. However, there are experimental results that are not understood using this approach. Among the puzzling examples are the α-spectrin results, in which the R15 domain folds 3 orders of magnitude more rapidly than the homologous R16 and R17, even though they are structurally very similar to each other. Such anomalous observations are usually attributed to the influence of internal friction on protein folding rates, but this is not a satisfactory explanation. In this study, this phenomenon is addressed by focusing on non-native interactions that could account for this effect. We carried out molecular dynamics simulations with structure-based C α models, in which the folding process of α-spectrin domains was investigated. The simulations take into account the hydrophobic and electrostatic contributions separately. The folding time results have shown qualitative agreement with the experimental data. We have also investigated mutations in R16 and R17, and the simulation folding time results correlate with the observed experimental ones. We suggest that the origin of the internal friction, at least in this case, might emerge from a cooperativity effect of these non-native interactions.

Association of the alpha-spectrin R28H mutation with allele alphaLELY and with alphaI/alphaII domain haplotypes in three Brazilian families.

We have studied three Brazilian kindreds presenting spectrin alpha/74 hereditary elliptocytosis (HE) due to a G-->A substitution, responsible for the R28H mutation. The mutant allele was associated with alphaI domain haplotype 1 (XbaI-/MspI-/PvuII+) in all three families and with two different alphaII domain haplotypes (1/RIT, 4/RVR). This result may reflect that this mutation occurs in a "hot spot" and may have arisen more than once or that a crossing over event may have occurred between the two domains studied. We detected one new haplotype in the alphaI domain (haplotype 3 -XbaI(+)/MspI(-)/PvuII(+)). The mutant allele was associated with the lack of the alphaII domain Alu insertion in all three cases. Allele alphaLELY, detected by PCR and restriction enzyme digestion, was present in the heterozygous form in patient 1 (alphaHE/alphaLELY) and in the homozygous form in patients 2 and 3(alphaHE-LEL/alphaLELY). It was found to be associated with domain haplotypes I (RIT) and 4 (RVR) and with the presence and absence of the Alu insertion. This may have arisen through recombination events, since this polymorphism is located in the alphaIV-alphaV domain junction, which is far distant from the alphaII domain.