Segmental isotope labeling of proteins for NMR structural study using a protein S tag for higher expression and solubility

A common obstacle to NMR studies of proteins is sample preparation. In many cases, proteins targeted for NMR studies are poorly expressed and/or expressed in insoluble forms. Here, we describe a novel approach to overcome these problems. In the protein S tag-intein (PSTI) technology, two tandem 92-residue N-terminal domains of protein S (PrS₂) from Myxococcus xanthus is fused at the N-terminal end of a protein to enhance its expression and solubility. Using intein technology, the isotope-labeled PrS₂-tag is replaced with non-isotope labeled PrS₂-tag, silencing the NMR signals from PrS₂-tag in isotope-filtered ¹H-detected NMR experiments. This method was applied to the E. coli ribosome binding factor A (RbfA), which aggregates and precipitates in the absence of a solubilization tag unless the C-terminal 25-residue segment is deleted (RbfAΔ25). Using the PrS₂-tag, full-length well-behaved RbfA samples could be successfully prepared for NMR studies. PrS₂ (non-labeled)-tagged RbfA (isotope-labeled) was produced with the use of the intein approach. The well-resolved TROSY-HSQC spectrum of full-length PrS₂-tagged RbfA superimposes with the TROSY-HSQC spectrum of RbfAΔ25, indicating that PrS₂-tag does not affect the structure of the protein to which it is fused. Using a smaller PrS-tag, consisting of a single N-terminal domain of protein S, triple resonance experiments were performed, and most of the backbone ¹H, ¹⁵N and ¹³C resonance assignments for full-length E. coli RbfA were determined. Analysis of these chemical shift data with the Chemical Shift Index and heteronuclear ¹H–¹⁵N NOE measurements reveal the dynamic nature of the C-terminal segment of the full-length RbfA protein, which could not be inferred using the truncated RbfAΔ25 construct. CS-Rosetta calculations also demonstrate that the core structure of full-length RbfA is similar to that of the RbfAΔ25 construct.