DISCUSSION

In this study, the slightly modified ribosome display using eukaryotic translation system was established. Anti-DNA antibody, 3D8 was specifically selected against ssDNA as a model system. The selected 3D8 RNAs sequences from translation complexes were recovered by RT-PCR. By applying this model system, TP-specific scFvs were selected and enriched from the immunized mouse library.

Four TP-specific scFv clones were selected and their sequences were analyzed.

These four scFvs showed intact scFv structure and specific binding activity to TP-peptide; three of them, the G6, G2, and D10 clones, also bound functional human HBV DNA polymerase in an ELISA. These selected anti- TP scFvs could be used in the future to further understand HBV replication.

In vitro display techniques have some potential advantages over in vivo display. These include ease of generation of larger libraries; fewer biases caused by cellular expression; and more facile application of round-by-round mutagenesis.

Ribosome display, one of the in vitro display methods, was achieved here using an in vitro transcription and translation process. These processes, which were based on Escherichia coli or rabbit reticulocyte system, have been described previously.^{17, 20} In these reports, functional scFvs were expressed and selected by ribosome display.

Since TP-peptide was poorly immunogenic, I needed a large library to obtain specific scFvs using the display method. Because the library size of a ribosome

display is not limited by transformation, antigen-specific scFvs would be more easily obtained from DNA libraries using ribosome display than by in vivo display methods.

A eukaryotic method was slightly modified from that described by He et al. and Makeyev et al.^{20, 33} In my work, the transcription and translation steps were carried out separately and the latter being performed using a rabbit reticulocyte system. By uncoupling transcription and translation, translation without a reducing agent was possible, and this could potentially result in greater functional expression of a single-chain antibody.⁴⁸ Translation using a rabbit reticulocyte system has some potential advantages over an E. coli ribosome display system in that the selection conditions are less complex due to lower RNase activity.¹⁵ Moreover, eukaryotic conditions might improve the translation and/or folding efficiency of some proteins.

Internal primers were used for the RT-PCR in this system. In principle, there should be no advantage to use an internal primer in ribosome display since mRNA is released from the complex before RT-PCR. However, the results are much better with an internal primer than with a 3' end primer (Fig. 12). If some of the ribosomes remained associated with mRNA after the dissociation procedure, this might inhibit the RT-PCR reaction. However, I treated the samples with EDTA and an RNA isolation kit to dissociate the mRNA from the ribosomes. Furthermore, little mRNA should remain associated with the ribosome after processing, as an RNA isolation kit was used originally for isolation from the cell. However, using nested PCR with internal primers seems to increase the sensitivity and specificity of repeated PCR of a library.

One advantage of ribosome display is the ease of generating larger libraries.

In this study, even after one round of selection, some portion of library was able to bind to the antigen, indicating that the library was of sufficient size. The selected anti-TP scFvs had higher immunogenecity than anti- TP clone s selected by phage display in direct comparison with periplasmic extracts³⁸ (Fig. 19). Notably, however, the selected antibody library was still diversified after four rounds of selection.

Although I observed antigen-specific gene recovery by RT-PCR and monitored the enrichment rate during the selecting procedure, our results suggest that four rounds may not be sufficient for complete selection. One reason may be that low temperature and RNase-free conditions are required to maintain the ARM complex, which hampers efficient selection in ribosome display. A modified ribosome display method, the ribosome- inactivation display system (RIDS), has been reported to have enhanced stability at room temperature, but has not yet been applied to selection from an antibody library.⁵⁹ Another reason might be some translated dead end antibody fragment. As antibody library is composed of a lot of variable antibody clones and be made by PCR reaction, a stop codon could be contained within a library. A stop codon within a library can induce a incomplete antibody fragment that does not connected with ribosome. A small sized protein shown in Fig. 11 suggests this possibility. If there are free antibody fragments in translation mixture during selection process, these could bind competitively with ARM complexes, and result in decreased efficiency of selection. The factors affecting the stable maintenance of ARM complexes and specific selection will be further studied in order to develop

more efficient selection process in ribosome display using eukaryotic translation system.

Although ribosome display has many advantages over phage display, to the best of my knowledge, only 4 reports regarding antibody selection from a library using ribosome display have been published. With only one using eukaryotic translation system.¹⁹ Therefore, my work gives a solid example to select antibody from antibody library by ribosome display using eukaryotic translation system.

Furthermore, the selected anti-TP scFvs in this study could be used to inhibit HBV replication. As HBV DNA polymerase is expressed at much lower levels than other gene products during HBV infectio n, an intrabody to the TP domain would be quite likely to efficiently inhibit HBV replication.

0 0.1 0.2 0.3 0.4 0.5 0.6 0.7

RD1 RD2 PD1 PD2

A405

TP-peptide no antigen

Fig. 19. Direct comparison of immunoreactivity of selected anti-TP scFvs by ribosome display and phage display. TP peptide coated as antigen to ELISA plate and the periplasmic extracts containing anti-TP scFvs were analyzed. The clone of anti-TP scFvs selected by ribosome display and phage display was denoted as RD and PD respectively.