Structure-functional relationship study of glycosyltransferases

Abstract

Gangliosides are a group of functional molecules and are synthesized by glycosyltransferases in a stepwise manner. Mechanism of ganglioside profile change in normal conditions or in diseases is not well understood. Gene regulations, protein structures and catalytic mechanisms of related glycosyltransferases may provide clues to elucidate this phenomenon. In our project, GD3-synthase is used as a model of glycosyltransferases for protein structure-functional relationship study. GD3-synthase is a key enzyme in the ganglioside biosynthetic network. It catalyzes the biosynthesis of ganglioside GD3, which is the entry substrate of biosynthesis of the b- and c- series gangliosides. GD3 is a minor ganglioside in adult vertebrate tissues, while it is highly expressed during embryonic development and in pathological conditions, such as cancer. GD3 is also involved in aging, cell proliferation, and cell differentiation. Aspects that are still poorly understood include: 1) the regulatory mechanisms for change of GD3 levels in normal condition and in diseases, 2) the structure of GD3-synthase, 3) how the structure is related to function, and 4) how the synthesis and degradation of the enzyme are regulated. To carry out the structure-functional relationship study on GD3-synthase, we developed a strategy to obtain a large amount of pure human GD3-synthase for crystallographic analysis. E. coli, yeast, and baculovirus systems were screened for selection of the expression system. Transmembrane domain"truncated human GD3- synthase was expressed in E. coli, but it aggregated into inclusion bodies. The refolded protein did not have enzyme activity. Human GD3-synthase could not be expressed in yeast cells, due to the presence of two yeast-specific stop codons. Although codon optimization was performed, the protein still could not be expressed in yeast cells. The soluble human GD3-synthase with enzymatic activity was expressed in Trichoplusia ni (T. ni) insect cells and secreted into the culture medium. The recombinant protein was purified from the culture medium with a yield of 1.45 mg/L. This is the first report of a procedure for expression and purification of GD3-synthase with a high yield, since the eDNA sequence was determined in 1994. As an alternative strategy, protein modeling was performed to study the structurefunctional relationship of GD3-synthase. Cstii, a bacterial sialyltransferase, was identified as a remote homologue of human GD3-synthase with a low sequence similarity. Cstii and vertebrate sialyltransferases share a similar topology of protein structure, which makes it possible to build the structure of human GD3-synthase by using homology modeling. Sequence comparison (including primary sequence and secondary structure alignments) between Cstii and human GD3-synthase was performed to identify the possible functional sites. Between sialylmotifs L and S, four highly conserved amino acid residues (Asnl88, Prol89, Serl90, and Arg272) were identified. Protein sequence alignment of human sialyltransferases suggests that all conserved residues identified in our study are ST8Sia :subfamily-specific. Functional analysis of these residues in human GD3 -synthase was performed by using site-directed mutational analysis. Mutational analysis of these conserved residues suggests that Asnl88, Serl90, and Arg272 are necessary for enzyme activity. The predicted 3-D structures of human GD3-synthase with docking of substrates support the data of mutational analysis and elucidate the contributions of these residues to the enzymatic activity, which suggests: 1) Asn!88 is acceptor binding-related, 2) Ser190 functions to lock the acceptor substrate, and 3) Arg272 is acceptor binding-related. We also suggest that the protein modeling approach can be applied to structure-functional relationship studies only for those regions, which are highly conserved between vertebrate sialyltransferases and Csti!Cstii, due to the low sequence similarity.