SUMMARY Glycoproteins represent a class of mammalian proteins that presents challenges for structural and functional characterization, particularly if the native glycosylation, which affects structure, stability and interaction with other molecules, is to be preserved. The best route to proteins with native glycosylation is expression in mammalian cell cultures. Unfortunately, for X-ray crystallography, this produces proteins with heterogenous glycosylation, often preventing formation of suitable crystals. For traditional nuclear magnetic resonance (NMR) methods, this forces use of expensive substrates for isotopic labeling and prohibits the perdeuteration often required to maintain resolution for larger proteins. The investigators involved in this proposal have worked together to develop an efficient mammalian cell expression system that produces proteins sparsely labeled using a restricted set of less expensive isotope enriched amino acids and maintains resolution without the aid of perdeuteration. This has been accompanied by the development of resonance assignment programs and data analysis protocols that allow structural and functional characterization from basic, high sensitivity, two-(and three-)dimensional NMR experiments. This project is designed to turn those developments into an integrated protocol that can be adopted by an expanded community of users. It centers on the refinement of a software package that accomplishes NMR resonance assignment of sparsely labeled proteins. It will be bolstered by extensive validation of program output, introduction of new data types and data analysis methods, and extension of program capabilities to the refinement of computer-generated models for protein structure. The potential impact will be a new route to structure and function studies of a class of proteins intimately involved with human physiology and disease.

PROJECT NARRATIVE Glycoproteins are intimately involved with human health and disease, both as targets for pathogen attachment and sources of altered function when improperly glycosylated. Yet, the understanding of these processes at a structural level is limited by the challenges these proteins pose for one of the primary approaches to structural characterization, nuclear magnetic resonance (NMR). This proposal would produce a protocol based on expression of sparsely labeled proteins in mammalian cell culture, software facilitated assignment of NMR resonances from those proteins and computer generated structural models validated by NMR data.