Posttranslational modifications of brain proteins

The brain and neural tissues harbour an especially rich collection of unusual structural features of protein-linked glycans, which in yet only vaguely understood ways contribute to neural functions and dysfunctions. Our hypothesis is that comprehensive data on the glycosylation of brain proteins and the ability to measure glycomic changes during neuronal development or disease progression contribute to the understanding of brain function. This project aims at the measurement of N- and O-glycans of mouse brain glycoproteins in an innovative, isomer- specific way that delivers quantitative results from minimal sample quantities. Technically, this will be realized by the graphitic carbon-LC-ESI MS/MS methodology recently developed in our laboratory. The object of the study will be the mouse brain glycoproteins CD24 and L1, which play essential roles in neurite outgrowth. Recent results indicate that the molecular mechanisms underlying their interaction and function include protein-carbohydrate interactions. Samples of these proteins extracted at various stages of brain development will be obtained from the group of Prof. Melitta Schachner of the University of Hamburg. Phase one of the project shall serve to find optimal conditions for the quantitative LC-ESI-MS analysis of N- as well as O-glycans from mouse brain cell surface glycoproteins. A byproduct will be information about which glycans should be chosen for closer investigation. Phase two shall deal with the structural elucidation of the relevant peaks. This shall be performed by the classical approaches of glycosidase digestions and MS/MS. An additional, innovative approach for a true structural assignment will involve reference glycans synthetized with the help of specific glycosyltransferase, e.g. Lewis- fucosyltransferases and ?2,3- or ?2,6-sialyltransferases. This strategy has already been proven useful for biantennary N-glycans and is expected to be applicable to larger N-glycans and to O-glycans, for which the art of structural analysis lags behind considerably. This activity shall yield an LC-MS/MS data set of mouse brain N- and O-glycans, where the major compounds (i.e. those exhibiting developmental changes plus the most abundant peaks) will enjoy a complete structural assignment. In phase three, the samples obtained from the collaborators lab will be processed and evaluated for qualitative and quantitative differences with the data from phase two as an important resource. Possibly, time may allow to define the ligand specificity of L1 by analyzing the glycans binding to recombinant L1.

Recent publication pointed at the significance of sugar determinants on glycoproteins in brain development and learning in the mouse. These determinants appeared to be defined by the presence of fucose in a particular linkage. The reliability of the applied lectin methods is limited and therefore the need for instrumental analysis by a combination of liquid chromatography (LC) with mass spectrometry (MS) is felt. Here, the glycans are cut from the protein part and separated according to shape. Then, the mass and also the fragment mass pattern are determined. Now, the situation is such that just one fucose residue on an ordinary N-glycan with 5 hexose and 4 hexosamine units can give rise to at least 40 different isomeric structures, i.e. glycans with the same mass but different structure. In this project, we could show that LC-MS is a promising way towards a reliable and yet fast and sensitive structural analysis. Retention times and fragment patterns of possible isomers must be recorded. A major body of work in this project consisted in the generation of the various isomers with the help of 11 self-produced, recombinant enzymes. At first we prepared the about 25 structures that seemed most probable but we had to learn that mouse brain mainly presented other isomers, which by fragment analysis alone could not be identified. So, while the complexity of the mouse protein glycans made our task difficult, it confirmed the validity of our experimental strategy.The multitude of possible isomers makes the assignment of peaks in an LC chromatogram difficult. Hence, we developed two strategies for the incorporation of heavy isotopes into glycans. As isotope-labelled activated sugars are not commercially available, we set out to generate 13C labelled, activated galactose again using recombinant enzyme and then isotope-labelled glycans. These will be useful for implementing glycan analysis of a hitherto unknown level of structural specificity and also quantitative reliability. Not at least, the developed methodology is highly sensitive, which enables to analyse the small quantities of sample obtained from mouse brain fractions, e.g. synapse preparations.