Monoclonal antibodies are used in a wide range of applications including as tools for research and for therapeutic purposes (1, 2). The range of possible applications is broadening rapidly with single-chain variable fragment (scFv) and antigen-binding fragments (Fab), parts of complete antibodies that are small, expressible in Escherichia coli, alluding for screening as well as selection of high-affinity binding molecules (3). However, one of the most important limitations of antibody fragment production is the time needed for generating and analyzing binding efficiency and the effect on target molecules (1, 4, 5). In recent years, a considerable acceleration—as well as a more economical generation—of antibodies was achieved through expression in E. coli cells (6). The synthesis of one Fab has been shown in insect cells in vivo (7). Cell-based methods of antibody production, however, require time and labor intensive cell cultivation. Moreover, for bacterial expression, additional work has to be performed to clone antibody-encoding genes into expression vectors (8).

In contrast, cell-free expression is now regarded as a promising alternative for overcoming limitations of cell-based methods. Reasons for that are most notably the clear improvement of productivity and economy, as well as the activity of the synthesized proteins in prokaryotic (9-11) and eukaryotic (12-16) cell-free systems in recent years. Modifications of some of these systems allowed the production of disulfide containing proteins including scFv (17, 18). However, because of their better binding properties, Fabs are a more attractive than scFv (19).

Recently, cell-free expression of one Fab in an E. coli-based system has been shown (20, 21). However, disadvantages of E. coli-based cell-free systems include lower specific activity of produced Fab and the requirement for the removal of toxins present in the expression system before the synthesized Fab can be applied to cell-based assays. By now, cell-free E. coli systems in their simple format (batch) yield up to 1 mg de novo synthesized protein per milliliter reaction. Nevertheless, the highest yield reported for cell-free synthesis of functional Fab is only 30 µg/mL (21). Despite the high productivity of E. coli-based cell-free systems, only a small percentage of the synthesized protein is usable. Moreover, in order to obtain accurate data, the active and inactive forms must be separated.

Cell-free E. coli-based systems exhibit a high-intrinsic activity, which changes the redox potential during disulfide protein synthesis remarkably. This change reduces the activity of synthesized Fab. An inhibition of this change was achieved by chemical pre-treatment of the cell lysate prior to protein synthesis reaction (22, 23). However, this procedure is detrimental for the handling of the system since it requires an additional work step that is difficult to automate.

For the synthesis of noteworthy amounts of soluble and active Fab antibody fragments, E. coli-based systems are supplemented with protein disulfide isomerase (PDI) and chaperones like GroE and DnaK (17, 18, 20, 21). The cell-free synthesis of disulfide proteins using an insect cell-based system has been shown. However, this system also requires supplementation of PDI and the synthesis of Fab has not been demonstrated so far (24).

Hence, for the synthesis of Fab antibody fragments, there is a need for an easy to handle and automatable cell-free expression system without the requirements for chemical pre-treatment and supplementation with chaperons and PDI, allowing the synthesis of Fab with high-specific activity. Ideally, such a system is free of endotoxins and compatible with rapidly generated linear templates produced by PCR (25), thus avoiding time consuming cloning of expression vectors.

Materials and methods

Generation of DNA templates

Primers were purchased from IBA (Göttingen, Germany). Molecular biology enzymes were from NEB (Frankfurt am Main, Germany). All Fab constructs, except controls, include the sequence for the honey bee melittin signal peptide. Genes encoding light and heavy chains (L- and H-chains) of HyHEL-10 anti-lysozyme Fab (26) were chemically synthesized by GeneArt (Regensburg, Germany) without their native signal peptide and with codon optimization for E. coli. Chemical synthesis also included an upstream sequence encoding honey bee melittin signal peptide and a sequence for the restriction enzyme recognition site BspHI: 5'-GAGCTCATGAAATTCTTAGTC AACGTTGCCCTGGTTTTTATGGTGGTGTATATTAGCTATATTTATGCCGAT-3′. Following the translational stop codons, a sequence for XhoI was also included. The genes were subcloned into vector pIX5.0 (RiNA, Berlin, Germany), resulting in plasmids pIX5.0-Mel-LaLys and pIX5.0-Mel-HaLys by using restriction enzymes NcoI and XhoI for the vector and BspHI and XhoI for the genes.