UGA researcher receives NIH award for high-risk, high-reward carbohydrate research

July 28, 2017

Figure 1: Antibodies are often used to diagnose and develop treatments for disease; their development is the most rapidly growing area of the pharmaceutical industry. Shown is a computer-simulated 3D structure for an antibody that can recognize specific types of cancers on the basis of the glycans (complex carbohydrates, green structure) present on the cancer cell surface. Figure 2: In a combination of methods called ???carbohydrate threading,??? data obtained from two high-throughput technologies ??“ virtual glycan array screening and experimental glycan array screening ??“ are combined to produce the 3D structure of a protein, such as an antibody, in combination with a large carbohydrate. Producing a structure of this complexity would be difficult using either X-ray crystallography or NMR spectroscopy.

The problem, according to Woods, has been that pertinent data from traditional analytical methods are extremely difficult to generate. Established ways of generating 3D molecular structures are not necessarily applicable to complex carbohydrates, he explained, as the large carbohydrate molecules, unlike proteins, are floppy and often ???too big??? or too heterogeneous for these methods. Woods said that with current methods of defining the 3D structures of carbohydrate-protein complexes, ???determining the 3D structure for a potential diagnostic agent becomes a Ph.D. thesis.???

During the last 10 years, he said, the technology to identify disease-related glycans has evolved rapidly, principally due to advances in mass spectrometry and glycan array screening, two technologies that speedily analyze millions of samples. However, Woods said that as libraries of carbohydrate structures have grown, ???among the things we??™ve discovered is that antibodies against carbohydrates are not as specific as we had thought. That can lead to false positives or unanticipated cross-reactions??”results that diagnostically and therapeutically are very bad. We can predict 3D structures, but without some corroborating data from the laboratory, people should be cautious before putting a lot of trust in it.???

Now, said Woods, combining two disparate high-throughput technologies??”glycan array screening and computer modeling??”provides a new approach to generating detailed atomic-level structures that provide insight into the origin of protein-glycan specificity.

Woods said the field of computational glycoscience holds promise for making structural information about such interactions widely accessible to the biotechnology industry. ???This insight can guide the engineering of enhanced antibody specificity or assist in the development of antibodies as therapeutic agents,??? he said. ???We should be able to re-engineer the antibodies to be more specific and more effective as therapeutics and diagnostics.???