Abstract 2220

The commercial production of recombinant factor VIII (rFVIII) is 2 to 3 orders of magnitude lower than that of comparably sized proteins in heterologous mammalian cell expression systems. One of the major contributing factors to its very low expression levels is protein misfolding and chaperone mediated retention in the endoplasmic reticulum (ER). Glucose regulated protein 78, also known as BiP, a central player in ER homeostasis, plays an important role in retention of FVIII in the ER. The secretion of Factor V (FV), a coagulation factor homologue of FVIII, is much more efficient (∼10- to 15-fold higher) in mammalian cells. Although the proteins share an identical domain structure, with close to 40% homology in their A and C domains, unlike with FVIII, BiP does not bind to and retain FV in the ER. Successful bioengineering strategies to improve the secretion of FVIII have been rationally informed by detailed study of ER chaperone interactions. A previous report indicated that the presence of multiple asparagine (N)-linked glycans within the first 50 amino acid residues of the NH2-terminus of some glycoproteins steers the nascent polypeptide away from BiP and towards the calnexin-calreticulin lectin protein folding machinery, resulting in more efficient folding and secretion of the protein (Molinari & Helenius, Science 2000). We hypothesized that the lack of BiP binding and the more efficient secretion of FV over FVIII could be the result of the presence of two N-linked glycans at positions 23 and 27 in FV, whereas FVIII has a single N-linked glycan at position 41. We investigated the role of N-linked glycans at the NH2-termini in the secretion and function of FV and FVIII. Single and double glycosylation mutants of FV were created by replacing Asn residues with Gln residues at positions 23 and 27. Similarly, Asn at position 41 of FVIII was mutated to Gln. The mutants were transiently transfected in COS-1 and CHO cells and their secretion and function were analyzed and compared to that of the respective wild type (WT) proteins. Antigen and activity assays revealed that the secretion and function of single and double glycosylation mutants of FV were no different from the WT protein indicating that the two N-linked glycans at the NH2–terminus did not contribute to the secretion or function of FV. Similarly, the FVIII glycosylation mutant did not display a significant reduction in secretion or function. We next tested if increasing the density of N-linked glycosylation at the NH2–terminus of FVIII might help improve folding and secretion. Two additional N-linked glycosylation sequons were introduced at positions 17 and 47 of FVIII by mutating M17N and K47N. Single and double glycosylation mutants were analyzed by transfections in COS-1 and CHO cells and compared to the WT protein. While the M17N glycosylation mutant was quite similar to the WT protein, the K47N mutant displayed a significant reduction in secretion. However, K47 is located adjacent to K48, a residue which has been shown to be the site of known hemophilia A missense mutations. The secretion of the M17N/K47N double mutant was reduced even further to almost half of the WT protein in cell media. We conclude that, unlike some other glycoprotein-ER chaperone interactions, early N-linked oligosaccharides at the NH2–termini of FV and FVIII do not contribute significantly to the secretion and function of either protein. Previous structural comparisons between FV and FVIII have revealed that a stretch of hydrophobic amino acids within the FVIII A1 domain may contribute to BiP interaction and a targeted point mutation (F309S) improved secretion of FVIII up to 3-fold (Swaroop et al, J Biol Chem 1997). Additional structure and function comparative studies between FV and FVIII may identify new targets for bioengineering strategies with an aim to further enhance the expression of rFVIII in mammalian cell lines.


No relevant conflicts of interest to declare.

Author notes


Asterisk with author names denotes non-ASH members.