A project in the framework of the 'Action for the promotion of and co-operation with the Belgian Coordinated Collection of Micro-organisms, BCCM'.

Partners

• BCCM/LMBP and VIB Department for Molecular Biomedical Research, Ghent University
• BCCM/MUCL and Unité de Microbiology (MBLA), Université Catholique de Louvain
• VIB Department for Transgene Technology and Gene Therapy, Katholieke Universiteit Leuven

Context

Over a period of about 20 years, bio-molecules have taken up an important position within the spectrum of pharmaceuticals. The advance of these compounds was significantly stimulated by the successful application of recombinant DNA technology and the development of cell lines and microorganisms that can serve as hosts for the heterologous production of proteins. Several therapeutic proteins can be produced in prokaryotic organisms. However, they lack several mechanisms needed to perform post-translational modifications, which are in many cases necessary for the correct folding and full biological function of the produced protein. A well-documented example of such a modification is N-glycosylation. Therefore, CHO cells – a mammalian cell line capable of producing human-like N-glycans – are currently a commonly used expression host for human therapeutic glycoproteins. However, there are several drawbacks to this system such as high expense, limited up-scaling of production capacities and complicated downstream processing.

Figure 4: N-glycan analysis of a GlcNAcMan₅GlcNAc₂ producing Pichia strain, transformed with different construct aiming at the conversion to GalGlcNAcMan₅GlcNAc₂.

Because of these drawbacks, the application of lowereukaryotic organisms such as yeast and fungi for the production of recombinant proteins has been under investigation for several years. On several levels they seem to be ideal expression hosts as they combine specific advantages of prokaryotes, such as ease of handling and manipulation with the ability of higher eukaryotic organisms to co- and post-translationally modify their proteins in the secretion pathway. However, unlike mammalian cell lines, the fungal organisms that have been evaluated up to now for recombinant protein production are unable to synthesize complex human-like N-glycans. Rather, their oligosaccharides are of the high-mannose type and as such do not carry terminal sialic acid residues. This will result in a rapid clearance of the recombinant glycoprotein when injected into the blood stream, which renders it inefficient for therapeutic applications. Furthermore, the high-mannose Nglycans can be modified via the addition of fungal-specific glycosyl structures such as galactofuranose and mannose phosphate. These structures are immunogenic for humans, which renders the glycoprotein even more unsuitable for therapeutic purposes. Hence, the generation of fungal organisms that are able to produce proteins with a human- like N-glycosylation could mean a significant breakthrough in the availability and application of biopharmaceuticals.

At the VIB Department for Molecular Biomedical Research (DMBR), much research has been done on modifying the glycosylation pathway of lower eukaryotes for the production of therapeutic glycoproteins. In this context, several plasmids carrying genes important for genetic engineering of the N-glycosylation pathway have been constructed. Some of these authentic constructs have already been deposited in the BCCM/LMBP plasmid collection to assure their long-term preservation and their distribution to third parties. During this project, these plasmids will be further evaluated into a suitable fungal organism like Pichia pastoris. Furthermore, reshuffling several deposited genes will generate new plasmids that can improve the results and/or simplify the lengthy process of the N-glycan pathway engineering. Examples of these reshufflings are the generation of new combinations of genes, important for glycan engineering, with other promoter sequences to drive their expression or with different selection markers for fungal transformation as well as the creation of fusion constructs between different genes that play an important role in the synthesis of human-like N-glycan structures. The applicability and functionality of each of the newly generated plasmids is evaluated via transformation to Pichia pastoris and following phenotypic screening of several transformants via N-glycan analysis. The latter is, in first instance, done on a rather crude extract of yeast cell wall or secreted medium proteins using DSA-FACE technology (currently applicable to both slabgel- based as well as capillary-based DNA sequencers) (Laroy et al., 2006). This technology is now also offered by BCCM/LMBP as a new service to the scientific community (‘ProfileThoseSugars’ – announcement in this issue of the BCCM Newsletter). Figure 4 shows a clear application within this project. Several plasmids were constructed in which a fusion between the S. pombe GAL10 gene and the human beta-1,4- galactosyltransferase was combined with different yeast localization signals and promoter sequences. A Pichia strain modified to produce GlcNAcMan5GlcNAc2 structures, an intermediate in the synthesis of human-like complex N-glycans, was transformed with these constructs and their efficiency for the conversion to GalGlcNAcMan5GlcNAc2 was evaluated via N-glycan analysis.

Despite of the promising results in for example Pichia pastoris, N-glycan engineering from hypermannosyl to complex-like structures remains a lengthy process. However, since more than a thousand different yeast species have been identified, nature might have already provided us with interesting organisms with the potential of being an expression host for foreign proteins and synthesizing simplified N-glycans that are easily convertible to human-like oligosaccharides. Data on N-glycosylation are limited to only a few species. Therefore, this project also envisages a high-throughput N-glycan screening of the fungal collection at BCCM/MUCL which has more than 3000 different strains of about 400 different species. N-glycan analysis is currently in process on a selection of species that should represent all of the phylogenetically distinct groups. These were determined by means of phylogenetic information, derived from ribosomal DNA sequences. The strains are taken from the collection under standard quality controls and grown on a small scale in defined minimal medium of light acidic pH (5 to 7). N-glycans are isolated from both medium proteins as well as cell-wall proteins and analyzed via the above-mentioned DSA-FACE technique. Currently, over more than 300 strains are under analysis. Since many yeast species show a lack of morphological differentiation, expansion of our knowledge on phenotypic properties such as N-glycosylation, might improve future characterizations and even classifications. Hence, in combination with already existing taxonomic and (phylo)genetic data, the project aims at the generation of a better-characterized subcollection within the BCCM/MUCL library and is exploring the possibility to using other technologies for taxonomic purposes and phenotypic evaluation of micro-organisms. Not only will this increase the quality of the identification/characterization service offered at BCCM/MUCL, it can also result in a new understanding of the evolution of the N-glycosylation machinery within fungal species. Today, the proven feasibility of screening large numbers of biological samples within this project has already resulted into the set-up of the new ‘ProfileThoseSugars’ service at BCCM/LMBP.

Several fungi have the intrinsic ability to secrete a large amount of proteins into their environment. However, some recombinant proteins are difficult to produce in any system. Cell-surface display in combination with FACS (fluorophore-activated cell sorting) can provide an elegant way to evaluate whether a given yeast organism has the potential to secrete a certain foreign protein. A yeast-surface-display (YSD) system, based on a-agglutinin has been developed in the past for S. cerevisiae (Boder & Wittrup, 1997). The a-agglutinin consists of a heterodimer of Aga1p, anchored into the cell wall via covalent binding to the beta-glucan layer, and Aga2p whose N-terminus interacts with Aga1p via two disulphide bridges. Surface expression of peptides and proteins can easily be obtained by in-frame cloning of their coding sequence to the 3’ end of the AGA2 sequence. Because of its higher intrinsic capacity to express recombinant proteins, we have recently implemented the a-agglutinin based YSD system into Pichia pastoris via the construction of new plasmids. Since a-agglutinin is also well expressed on the Pichia surface, the level of surface display after fusing a protein to the C-terminus of Aga2p will be a measure for the efficiency with which the fusion partner can be secreted by the Pichia cells. As such, YSD in combination with high-throughput FACS analysis and mutagenesis techniques (mutagenesis of either the expression strain or the coding sequence of the fusion partner) provides an interesting set-up in the search for better expression hosts of a given protein. The project aims at validating the newly constructed plasmids as well as the proposed exploitation of generating better producing strains. As a test case, two protein domains of great interest for the VIB Department for Transgene Technology and Gene Therapy, are under evaluation.

The combination of this technology with glyco-engineering should provide fungal organisms able to produce significant amounts of a given protein carrying a homogeneous human-like N-glycan structure.

Literature:

• Boder ET & Wittrup KD. Yeast surface display for screening combinatorial polypeptide libraries. Nature Biotechnology 15, 553-557, 1997.
  
  
• Laroy W, Contreras R, Callewaert N. Glycome mapping on DNA sequencing equipment. Nature Protocols 1, 397-407, 2006.