My research interest is regarding protein folding and misfolding behaviors in order to answer how proteins can fold into its native structure and how certain proteins can misfold and cause disease. Studies about protein folding are mainly the development of new methodologies to explore the folding process in order to unravel the intrinsic folding properties. Studies about protein misfolding are focus on two diseases: prion disease and Alzheimer’s disease. We aim to tackle the biological problem from a chemist’s point of view. We are interested in studying the mechanism of amyloid fibril formation and the factors influencing molecular assembly, designing inhibitor for amyloid formation, and evaluating the species barrier in prion disease transmission.

I. Protein folding:
(1) Early stage protein folding
        Due to the lack of new technology, the progress of protein folding studies slows down apparently in recent years. We have long term interest in developing new strategy to study early stage of protein folding which were ignored before due to the limitation of dead time in initiating refolding reaction and the speed limitation of the detection method. Recently, we have developed a novel strategy to follow the early kinetic events in the refolding of small structural motifs in proteins based on laser flash photolysis of photo-labile “caged” peptides. Because the photolabile cage is cleaved within one nanosecond, it is possible to follow the re-formation of secondary and tertiary structures in peptides and proteins on a nanosecond timescale. Photoacoustic calorimetry was used in detection of the folding process. This technique can detect folding reaction as early as 20 ns. By making several caged peptides with different turn sequences, we are able to explore how the turn sequence affects the folding kinetics. Application of similar strategy to a small protein, RD1, is undergoing successfully. Using this method, the kinetic behavior of proteins, which fold fast and can not be studied before, can be explored now.
　

Photolabile caging strategy

(2) Single molecule FRET study
        In our previous studies, multiple phases were found in the refolding kinetic traces of ubiquitin, implying that possibility of the existence of multiple pathways. We aim to explore the folding of ubiquitin on a microsecond (or even nanosecond) timescale by using the Fluorescence Resonance Energy Transfer (FRET) method and single molecule technique → spFRET. FRET experiments have been widely used in protein-protein interactions that donor and acceptor dyes are attached to different protein molecules. To study the conformational change within one molecule, two dyes have to be added to the same molecule. How to label the protein with two dyes remains a big challenge. Recently, we developed a new method to efficiently make doubly labeled protein. Donor and acceptor dyes can be site-specifically labeled on ubiquitin with high coupling efficiency for single molecule studies. This method also allows us to swap the labeling position easily. We have found that swapping the labeling sites will create doubly-labeled protein with different FRET efficiencies. The data imply that dyes might interact with protein surface, hence specifically labeled protein is necessary for single molecule study in order to avoid signal averaging problem due to the existence of reversely-labeled proteins. Single molecule spectroscopy can unmask the heterogeneity problem. FRET method provides high sensitivity in signal measurement. Combing these two techniques, we should be able to tackle the nature of protein folding.
　

Design of labeling strategy of donor and acceptor dyes

Conformational heterogeneity of protein

II. Protein misfolding:
(3) Evaluation of species barrier of prion transmission
        It has been known that protein sequence is related to the transmission efficiency of prion diseases among different species. However, how sequence determines the efficiency is not clear yet. Previous studies have often reported different results regarding seeding efficiency, the efficiency of initiating amyloid propagation by adding pre-existing amyloid fibrils as seed. We successfully developed a new method called “seed titration method” to quantify the seeding efficiency between hamster and mouse Using time-resolved circular dichroism spectroscopy, we found out the relationship between prion sequence and seeding efficiency. We used synthetic peptides as a simple system to determine the sequence-dependent transmission barrier between hamster and mouse. We found that the heterologous seeding efficiency of hamster and mouse prion peptides was four times less than that of homologous seeding. Moreover, residue 139 was not the only residue in determining seeding efficiency. When the seed had Ile at this position, the homology at this position between seed and monomer determined the seeding efficiency. When the seed had Met at this position, homology at residues 109 and 112 determined the seeding efficiency. We are the first to clearly describe how the sequence difference determines the transmission efficiency between hamster and mouse using an in vitro model. Our results can well explain the origin of the contradiction in the previously reported in vivo data. Our study is important in estimating the transmission risk of prion diseases among different species without sacrificing any animals.

Illustration of “seed titration method”.

(4) Conformational tuning of polypeptide chain by mutations and post-translational modifications
        We explored the amyloidogenesis of the prion peptide and a de novo designed helical peptide. We found the amyloidogenesis of the prion peptide is dramatically affected by adding a sugar on the Ser-135 of the prion peptide. This effect is strongly dependent on the nature of the attached sugar unit. The glycosylation with the glycosidic bond in the α-configuration has much greater effect than that in the β-configuration. We also discovered that the effect of glycosylation in the structural conversion of the glycosylated peptide is due to the interaction between the sugar and the side-chain of Arg-136. Mutating Arg-136 to Gly dramatically accelerates the amyloidogenesis of the mutant peptide. Our results imply that Arg-136 is the kink region of the structural conversion and that it might be the turn region of the final amyloid structure. Compared with glycosylation, phosphorylation has less inhibitory effect on amyloidogenesis of the prion peptide and the de novo designed helical peptide (αtα). We concluded that the conformational properties of a polypeptide chain can be tuned by mutations and post-translational modifications.

The nucleation-dependent polymerization model for amyloidogenesis.

The prion peptides glycosylated with α-GalNAc and α-GlcNAc at Ser-135 can not form amyloid fibrils.

(5) Inhibition of amyloid formation and amyloid cleaning
        There are about one hundred thousand people die because of Alzheimer’s disease every year. With the increasing human lifespan, there will be more and more people suffering from this malady. Recently, the amyloid structures formed of Aβ40 and Aβ42 were disclosed. The difference in the amyloid core between these two structures raised our idea that different parts of Aβ might all have the potential to form amyloid structure. In order to test our postulation, a ^DP-G sequence was introduced at different positions of the Aβ40 peptide to work as a “structural clip”. Surprisingly, we found that the mutated peptides containing the ^DP-G sequence at the C-terminal region will not form amyloid fibrils any more. Instead, they show random coil structure at low peptide concentration and reversibly convert into β-structure at high peptide concentration. Interestingly, this β-aggregate has most of the characteristics of amyloid except fibril morphology. Thioflavin T and Congo red, the dyes usually employed in amyloid detection and quantification, were able to bind to this structure. We concluded that these Aβ40 mutants form a new amyloid-like aggregate. It is a new finding showing that spectroscopic evidence is not necessarily correlated to amyloid formation. Moreover, the mutant peptide V24P, when mixed with Aβ40, can attenuate the cytotoxicity of Aβ40. We surmise that V24P, with the same hydrophobic segment as Aβ40, might associate with Aβ40 and prohibit the formation of toxic oligomer. We proposed that designing a peptide which is able to “trap” Aβ peptides or divert the misfolding pathway of the Aβ peptides might be another strategy in the therapy of Alzheimer’s disease.

A proposed working model to explain how V24P can decrease Aβ toxicity.

Download slides of courses