Downloaded from www.microbiologyresearch.org by IP: 54.70.40.11 On: Sat, 15 Dec 2018 23:40:34 Adenovirus type 5 fiber knob domain has a critical role in fiber protein synthesis and encapsidation Petra Henning, 1,2 3 Emma Lundgren, 1 3 Mattias Carlsson, 1 Karolin Frykholm, 1 Jenny Johannisson, 1 Maria K. Magnusson, 1,2 Erika Ta ˚ ng, 1 Laure Franqueville, 3 Saw See Hong, 3 Leif Lindholm 1,2 and Pierre Boulanger 3,4 Correspondence Leif Lindholm leif.lindholm@gotagene.se Pierre Boulanger Pierre.Boulanger@ sante.univ-lyon1.fr 1 Got-A-Gene AB, O ¨ stra Kyviksva ¨ gen 18, SE 42930 Kullavik, Sweden 2 Institute for Biomedicine, Department of Microbiology and Immunology, University of Go ¨ teborg, PO Box 435, SE 40530 Go ¨ teborg, Sweden 3 Laboratoire de Virologie et Pathoge ´ne ` se Virale, Universite ´ Claude Bernard de Lyon and CNRS UMR-5537, Faculte ´ de Me ´ decine RTH Laennec, 7 Rue Guillaume Paradin, 69372 Lyon Cedex 08, France 4 Laboratoire de Virologie Me ´ dicale, Domaine Rockefeller, Hospices Civils de Lyon, 8 Avenue Rockefeller, 69373 Lyon Cedex 08, France Received 1 March 2006 Accepted 3 July 2006 Adenovirus serotype 5 (Ad5) vectors carrying knobless fibers designed to remove their natural tropism were found to have a lower fiber content than recombinant Ad5 with wild-type (WT) capsid, implying a role for the knob-coding sequence or/and the knob domain in fiber encapsidation. Experimental data using a variety of fiber gene constructs showed that the defect did not occur at the fiber mRNA level, but at the protein level. Knobless fiber proteins were found to be synthesized at a significant slower rate compared with knob-carrying fibers, and the trimerization process of knobless fibers paralleled their slow rate of synthesis. A recombinant Ad5 diploid for the fiber gene (referred to as Ad5/R7-ZZ wt /E1 : WT-fiber) was constructed to analyse the possible rescue of the knobless low-fiber-content phenotype by co-expression of WT fiber. Ad5/R7-ZZ wt /E1 : WT-fiber contained a knobless fiber gene in its natural location (L5) in the viral genome and an additional WT fiber gene in an ectopic position in E1. Knobless fiber was still synthesized at low levels compared with the co-expressed E1 : WT fiber and the recovery of the two fiber species in virus progeny reflected their respective amounts in the infected cells. Our results suggested that deletion of the fiber knob domain had a negative effect on the translation of the fiber mRNA and on the intracellular concentration of fiber protein. They also suggested that the knob control of fiber protein synthesis and encapsidation occurred as a cis effect, which was not modified by WT fiber protein provided in trans by the same Ad5 genome. INTRODUCTION One of the possible strategies for redirecting adenovirus serotype 5 (Ad5) vectors to desired cell targets consists of (i) removing their natural tropism and (ii) inserting a new cell- binding ligand in one of the capsid proteins. Step (i) can be achieved by deletion of the entire fiber knob domain and its replacement by a non-viral trimerization motif. In earlier studies, we and others have described the construction of viable, knobless adenoviruses in which the deletion of the knob was compensated for by the incorporation of sequences encoding a non-viral trimerization signal (Krasnykh et al., 2001; Magnusson et al., 2001; van Beusechem et al., 2000). However, such recombinant viruses showed some defective functions compared with wild-type (WT) virus: growth and infectivity were impaired, virus entry was hampered in some cases and the fiber content of virus particles carrying knobless fibers was lower than the theoretical number of 12 copies per virion. In contrast, recombinant Ad5 with fibers carrying their knob domain in addition to new cell ligands have nearly normal fiber numbers, suggesting that the knob domain plays a role, direct or indirect, in the fiber content of the virion (Henning et al., 2002, 2005; Magnusson et al., 2001, 2002; van Beusechem et al., 2000). The lower growth rate and infectivity index of our Ad5 vectors could be due to negative effects on multiple steps of the virus life cycle, including their binding to cellular receptors, endocytosis, intracellular trafficking, assembly, 3These authors contributed equally to this paper. 0008-1992 G 2006 SGM Printed in Great Britain 3151 Journal of General Virology (2006), 87, 3151–3160 DOI 10.1099/vir.0.81992-0