Final Report Summary - GLYCAN HETEROGENEITY (Influence of the protein sequence and structure on the glycan micro-heterogeneity)
Presently many pharmaceutical industries rely on recombinant proteins to treat various diseases. Among these recombinant proteins a large part are glycoproteins. Glycoproteins are proteins modified post-translationally by a sugar moiety. N-glycosylation is a major post-translational modification involved in protein folding, protein maturation, trafficking and secretion, and can modulate protein functions. N-glycosylation consists in the addition of a glycan precursor onto a nascent protein in the endoplasmic reticulum. The glycan is then trimmed and matured in the Golgi by addition of various sugars motifs, leading to a large heterogeneity of glycans structures. Heterogeneity of glycosylation is observed from organism to organism, cell to cell, protein to protein, but also at the glycosylation site level. Until now the underlying causes of this microheterogeneity have not clearly been identified and the influence of the sugar-nucleotides pool and the glycosyltransferases present in the cell was proposed. Although heterogeneity has biological relevance, when producing pharmaceutical compounds homogenous glycosylation, or specific and reproducible glycan biodiversity has a wide importance. A better understanding of the glycan structures roles and synthesis, would help to reduce the glycan heterogeneity of recombinant proteins without multiplying high-cost purification steps.
In this project, we aimed to prove that heterogeneity of glycan structures is highly influenced by the three-dimensional protein structure and the interaction between the protein and the glycan. To do so, we used classic biochemical and molecular biology approaches in combination with high-performance mass spectrometry and to modify the glycan and/ or the structure of a model glycoprotein, containing 5 sites of glycosylation. The model protein was expressed and purified from mammalian cells. The site-specific glycosylation of the 5 sites was analyzed by mass spectrometry and the enzymatic conversion rates and fluxes were calculated. We observed on the wildtype protein that each of the 5 sites had a specific glycan profile. In particular site 4 presented the least processed glycan structures such as oligomannose and hybrid glycans, while other sites presented mainly highly processed fucosylated complex glycans. Since the maturation process is a non-template driven modification, the flux analysis allowed us to determine the most likely path followed for the processing of each sites. In parallel, the enzymatic conversion rate calculation in combination with molecular dynamic studies allowed to pinpoint the bottleneck steps. For example, we determined that the site 4 glycan profile was due to a lack of accessibility of the enzyme ER-mannosidase I and Mannosidase II to the glycan. Molecular dynamic studies showed that the glycan on site 4 was interacting strongly with the protein surface, preventing the trimming enzymes to access their substrate. These simulation indicated specific aminoacids potentially creating this interaction. Thanks to a site-directed mutagenesis approach, it was proven that a tyrosine in the vicinity of the site 4 was tightly interacting with the glycan and upon mutation of this tyrosine to a non-aromatic aminoacid the glycan profile was drastically shifted to highly processed glycans. Interestingly mutation of the tyrosine to a phenylalanine was to the contrary leading to an increase of the oligomannose and hybrid structures. These results demonstrate the possibility to modify the glycosylation profile of a specific site of glycosylation thanks to a single point mutation of the protein. We created another set of mutants where this time, a tyrosine was introduced in the vicinity of a glycosylation site which is normally carrying fucosylated complex glycans and were able to decrease the abundance of complex glycans in favor of oligomannose.
Altogether these results show that the protein tridimensional structure influence on the glycan trimming is not only due to a steric hindrance provided by the protein structure surrounding the glycan and burying it to reduce the accessibility to the processing enzyme, but is also due to direct interaction of specific aminoacids with the glycan structure. This project show also the possibility to glycoengineer in a site-specific manner a glycoprotein by modifying the protein sequence, in opposition to most of the present glycoengineering techniques which are changing the glycan machinery at a cellular level by knocking down or overexpressing glycosylation enzymes and thus modify simultaneously all the glycan carried by a protein. Combining site-directed glycoprotein mutagenesis and classical glycoengineering techniques could allow to obtain specific targeted glycan profile on a single site, acquisition of such define glycan profile would be a great plus for the industry of recombinant glycoproteins.
In this project, we aimed to prove that heterogeneity of glycan structures is highly influenced by the three-dimensional protein structure and the interaction between the protein and the glycan. To do so, we used classic biochemical and molecular biology approaches in combination with high-performance mass spectrometry and to modify the glycan and/ or the structure of a model glycoprotein, containing 5 sites of glycosylation. The model protein was expressed and purified from mammalian cells. The site-specific glycosylation of the 5 sites was analyzed by mass spectrometry and the enzymatic conversion rates and fluxes were calculated. We observed on the wildtype protein that each of the 5 sites had a specific glycan profile. In particular site 4 presented the least processed glycan structures such as oligomannose and hybrid glycans, while other sites presented mainly highly processed fucosylated complex glycans. Since the maturation process is a non-template driven modification, the flux analysis allowed us to determine the most likely path followed for the processing of each sites. In parallel, the enzymatic conversion rate calculation in combination with molecular dynamic studies allowed to pinpoint the bottleneck steps. For example, we determined that the site 4 glycan profile was due to a lack of accessibility of the enzyme ER-mannosidase I and Mannosidase II to the glycan. Molecular dynamic studies showed that the glycan on site 4 was interacting strongly with the protein surface, preventing the trimming enzymes to access their substrate. These simulation indicated specific aminoacids potentially creating this interaction. Thanks to a site-directed mutagenesis approach, it was proven that a tyrosine in the vicinity of the site 4 was tightly interacting with the glycan and upon mutation of this tyrosine to a non-aromatic aminoacid the glycan profile was drastically shifted to highly processed glycans. Interestingly mutation of the tyrosine to a phenylalanine was to the contrary leading to an increase of the oligomannose and hybrid structures. These results demonstrate the possibility to modify the glycosylation profile of a specific site of glycosylation thanks to a single point mutation of the protein. We created another set of mutants where this time, a tyrosine was introduced in the vicinity of a glycosylation site which is normally carrying fucosylated complex glycans and were able to decrease the abundance of complex glycans in favor of oligomannose.
Altogether these results show that the protein tridimensional structure influence on the glycan trimming is not only due to a steric hindrance provided by the protein structure surrounding the glycan and burying it to reduce the accessibility to the processing enzyme, but is also due to direct interaction of specific aminoacids with the glycan structure. This project show also the possibility to glycoengineer in a site-specific manner a glycoprotein by modifying the protein sequence, in opposition to most of the present glycoengineering techniques which are changing the glycan machinery at a cellular level by knocking down or overexpressing glycosylation enzymes and thus modify simultaneously all the glycan carried by a protein. Combining site-directed glycoprotein mutagenesis and classical glycoengineering techniques could allow to obtain specific targeted glycan profile on a single site, acquisition of such define glycan profile would be a great plus for the industry of recombinant glycoproteins.