Structural insights into Arginine-Serine rich proteins and N-H spin-spin coupling constants

This thesis is constituted of two parts. In the first part of this thesis, I studied the structural phenomena of SR protein phosphorylation. The serine/arginine-rich (SR) protein family plays multiple functions in the whole process of RNA metabolism such as transcription, RNA splicing, RNA exporting, translation, and nonsense decay of RNA. SR proteins contain at least one essential arginine/serine-rich (RS) domain, from which the protein name is derived. The RS domain mediates both protein-protein and protein-RNA interactions. A number of studies have shown that the serine residues within the RS domains of SR proteins are extensively phosphorylated. The majority of phosphorylation occurs on serine residues in the RS domain. This phosphorylation appears to influence interactions and the subcellular localization of SR proteins, thereby modulating their functions. In spite of the versatile and vital functions of SR proteins, the structure of the whole protein remains unknown so far. In this study, we use the combined approaches of NMR and MD simulation to study the structures of wild type and phosphorylated RS dipeptide repeat peptides which are derived from the prototypical SR protein ASF/SF2 (200-219). In wild type form, the RS repeat peptide is disordered, which is revealed by its negative heteronuclear NOE values and highly degenerate spectra. Upon phosphorylation, the RS repeat peptide gets structured but is still not as fully rigid as folded proteins. After unbiased MD simulations of both, a sub-ensemble selection procedure was carried out to obtain a representative structural ensemble of the native and phosphorylated ASF/SF2 (200-219) peptides, which constrains the simulated ensembles to reproduce experimental NMR data. The phosphorylated peptide adopts an “arch-like” structure, which get structured at both backbone and side chain levels. The structural changes also show pH dependency, which is related to the charges which require -2 charges on the side chain of serine residues. Hence RE or RD repeats cannot reproduce the same phenomena as phosphorylated RS repeats. NMR data of the intact full length RS domains in ASF/SF2 or hPrp28 showed that the phosphorylated RS repeats in domain also undergoes a disorder-to-order transition upon phosphorylation, which is similar to the isolated RS peptide. This similarity suggests the phosphorylated RS peptide is a general model of structure transitions upon phosphorylation for all SR and SR-related proteins. In the second part of this thesis, I studied the relationship between N-H spin-spin couplings and hydrogen bonds. Hydrogen bonds are essential for the structure of many biochemical compounds. Protein folding, the formation of amyloid aggregates, enzymatic catalysis, drug-receptor interactions, and many other phenomena are intrinsically connected to hydrogen bonding. It has been theoretically predicted that 1JNH becomes more negative upon hydrogen bond formation. In spite of the high accuracy of 1JNH measurement, 1JNH values have not been used for the detection of hydrogen bonds. In this study, I first measured large numbers of 1JNH values in disordered proteins to serve as a reference. The residue-specific mean values of 1JNH were used as random coil values. The influence of pH and temperature were also checked by systematically measuring the couplings in various conditions. Comparing the random coil values to 1JNH in folded proteins, it was demonstrated that the magnitude of 1JNH was increased by up to 1.6Hz due to hydrogen bond formation. Thus the H-N spin coupling can be a sensitive tool for the detection of hydrogen bonds. The results also provide a basis for further investigating the relation between hydrogen bonds and 1JNH values using theoretical calculations