Production of human cytomegalovirus glycoprotein B in Nicotiana tabacum BY-2 suspension cells

The production of pharmaceutical recombinant proteins in plant has benefited from a growing interest in recent years. However, whole plant-based production faces several hurdles such as opposition to genetically modified crops and difficulties to apply Good Manufacturing Practices. We suggest here the use of BY-2 suspension cells as an efficient alternative expression platform for vaccinal glycoprotein production. These cells combine micro-organisms advantages with those of the higher eukaryotes, while reducing N-glycosylations heterogeneity compared to mammalian cells. However, plant-typical β-1,2-xylose (Xyl) and α-1,3-fucose (Fuc) could trigger allergic response, accelerate drug clearance and thus reduce their efficiency. In the frame of this work, we firstly characterised a BY-2 cell line in which the genes coding for β-1,2-xylosyltransferases (XylT) and α-1,3-fucosyl-transferases (FucT) have been targeted by a multiplex Cas9 strategy. No Xyl or Fuc residues were detected in total soluble proteins. Furthermore, Cas9-induced indels were observed in each PCR-amplified XylT/FucT sequence, suggesting that all alleles are indeed knocked out. Then, we compared non-glycoengineered cell lines previously transformed with three genetic constructs allowing the production of a vaccinal human cytomegalovirus glycoprotein B (gB). Each construct differs in the number of genetic insulators (SARs) flanking the gB expression cassettes. Recent studies suggest that SARs increase transgene expression. We determined their impact on the expression of gB. However, we did not observe any positive SARs effect on gB production. The highest production rate was even obtained for the construct which do not contain SAR. Therefore, the XylT/FucT KO cell line was transformed with this construct as well as two other constructs which differ in the type of gB isoform used. Elite lines were selected for each construct and expression of both isoforms were compared. One of them seems to allow higher production yield. Finally, we discuss the CRISPR-Cas9 cleavage efficiency, off-target concerns and strategies that could be considered to avoid them. We also propose alternative constructs which could improve the SARs influence on gB expression.