Structural Profiling Tools
In all these engineering campaigns, the availability of robust, high-resolution and high-sensitivity N-glycan structural profiling tools is of paramount importance. It also is important in monitoring production process variations for their influence on product N-glycosylation and in ensuring batch-to-batch reproducibility. Moreover, a high-throughput, low-cost analytics platform that can be operated by nonspecialist technicians is a desirable feature.
In our laboratory, the workhorse analytical technique is N-glycan profiling on the Applied Biosystems(www.appliedbiosystems.com) gel-based or capillary array-based DNA sequencers.
We provide examples here on how this method has allowed N-glycan profiling to become a technique as routine as SDS-PAGE for proteins and sequencing of DNA, effectively being executed on over 10,000 samples per year on a single four-capillary instrument.
Samples of glycoproteins are worked up through an optimized multistep 96-well plate-based protocol, and the released N-glycans are labeled with the fluorophore 8-aminopyrene-1,3,6-trisulfonic acid, followed by removal of the excess label. The workflow for sample preparation is provided in Figure 1.
In a typical N-glycan engineering campaign in the yeast Pichia pastoris, a series of vectors for the inactivation of unwanted host cell genes, for the overexpression of glycosyltransferases, etc., needs to be transformed to a wild type strain, which produces maximal levels of a target biopharmaceutical under an optimized fermentation regime. Upon each transformation round, only a fraction of the selected clones will contain the transgene in a desired genomic configuration, and among these clones, there still is significant clonal variation in the efficiency of the N-glycan engineering, which is achieved. Therefore, phenotypic screening of a relatively large number of clones is warranted to identify the clone with the most complete engineering efficiency, and to minimize the heterogeneity of the N-glycosylation pattern that is finally obtained.
To perform this screening, we grew small cultures of the clones in 24 deep-well plates covered with a gas-permeable membrane and incubated in a shaking incubator.
As a first screen, we typically prepare a rather crude extract of the cell wall glycoproteins of the clones and profile the N-glycans of these yeast-endogenous glycoproteins, which are produced at high levels.
Figure 2 shows how such analysis allows the selection of those clones with the highest engineering efficiency in a rapid fashion.
If structural information on the observed N-glycans is desired, we perform exoglycosidase array sequencing, either on the total mixture if the major compounds are of interest, or on HPLC-purified minor compounds from the mixture. The integration of a normal phase ion-pairing HPLC option in our standard workflow allows for the complete analysis of complex mixtures (Figure 3).
Along similar lines, we have employed N-glycan profiling on DNA sequencers in analysis service to academic scientists and to the biopharmaceutical industry. The high sensitivity of the integrated analytical protocols is a significant advantage of the method, with 1–5 micrograms of protein being sufficient to obtain results of exoglycosidase array sequencing. This enables the implemention of N-glycan profiling at a much earlier stage of cell-line development than has been possible in the past.
Expertise was gained for the analysis of mAb N-glycosylation at an early stage of clone selection (24-well plate culture scale) for the analysis of mAbs and total glycoprotein extracts produced in plants (proteins extracted from parts of single Arabidopsis plants are sufficient), for the analysis of N-glycans derived from Western-blotted protein bands, for recombinant Factor IX secreted by muscular tissue implants, for EPO produced in the novel expression system Leishmania tarentolae, for N-glycosylation analysis of antigens derived from parasitic trematodes, and many more.
Customized contract analytical services (specializing in miniaturized comparative analysis of larger series of samples) built on this technology and performed by the developers, are available through the ProfileThoseSugars service at BCCM/LMBP (bccm.belspo.be/about/lmbp.php); contact: Kristien Neyts (kristien.neyts @dmbr.ugent.be).