**5.2 Generation of gene library and construction of expression system**

The advantages of combinatorial chemistry approach to protein engineering is its ability to associate every protein with its genetic material (Weng & DeLisi, 2002). There are two core parts about applying combinatorial chemistry approach in protein engineering: display techniques and gene expression library. These two core parts are equally important and affect each other. To date, several display techniques have been developed (Table 4) (Daugherty, 2007;Gronwall & Stahl, 2009). Natural properties of peptides with a CSαβ motif are as defenders for their native hosts and most of the peptides express anti-bacterial abilities, virus inhibitory abilities and inhibition of protein translation. How to design a expression library based on a CSαβ motif scaffold is a tough task.


Table 4. Comparison of three display techniques.

Peptides with a CSαβ motif are tightly held by disulphide bridges and may be not easily folded into appropriate structures inside bacterial cells (Villemagne et al., 2006). Ribosomal display technique can overcome the problems (Figure 7), but it requirs an ultra clean

Fig. 7. Concept of ribosome display techniques. (a) expression system, (b) & (c) transcription, translation and folding while translation, (d) binding and selection, (e) isolation of mRNA from the complex, (f) reverse transcription and construction of expression system for next cycle of selection. A specific holder gene usually is constructed to the expression system. The holder can arrest protein translation and the nascent peptide can form a stable complex with ribomsome.

Fig. 7. Concept of ribosome display techniques. (a) expression system, (b) & (c) transcription, translation and folding while translation, (d) binding and selection, (e) isolation of mRNA from the complex, (f) reverse transcription and construction of expression system for next cycle of selection. A specific holder gene usually is constructed to the expression system. The holder can arrest protein translation and the nascent peptide can form a stable complex

with ribomsome.

environment to protect mRNA from degradation (Fennell et al., 2010;Villemagne et al., 2006). The gene library of a CSαβ motif has to be artificially synthesized then constructed to the expression system. There are two key issues of the synthesized gene library have to be considered, one is the size of library the other one is how many random residue positions it should have (Gronwall & Stahl, 2009;Michel-Reydellet et al., 2005). These two questions are hard to answer. A large library with more randomized residue positions could have more genotypes/phenotypes but it also has a high possibility to generate nonsense genes for the unvoided stop codon on unexpected positions. A library with a randomized loop 3 of a CSαβ motif scaffold has been constructed and successfully used in a noval binder to human TNF-α (Zhao et al., 2004). It also has been reported that all three loops of peptides with a CSαβ motif can equip unique functions and the structural regions also fine regulate the functions of the peptides (Assadi-Porter et al., 2010;Liu et al., 2006;Yang et al., 2009;Zhao et al., 2002). For a library of long length artificial genes, several cycles of PCR overlapping extensions are required and may reduce quality of the library. A circular or a linear expression systems can be recruited (Fennell et al., 2010;Shimizu et al., 2001;Villemagne et al., 2006). To construct the linear library, all required DNA elements, as promotor, ribosomal binding site and poly T tail, are synthesized or copied and constructed with the desired gene library through PCR (Katzen et al., 2005;Shimizu et al., 2001) . In the linear expression system, the library can have a maximal size up to 1013 (Table 4) (Gronwall & Stahl, 2009).When a linear expression system is recruited, the qualities of final PCR products should be well controlled. It is better for the library to be freshly constructed everytime the protein display is performed. The size of a library constructed with a circular expression system is much smaller for the genotype lost during construction, however, the library can be stably stored in a good condition after construction.

### **5.3 Mass production and folding of proteins**

To screening possible funtions of a protein, it requires a large amount of the protein. The amount of proteins with a CSαβ motif in their native host is rare and to obtain eounght of the native proteins for assays requires a lot of raw materials. To produce recombinant proteins of a CSαβ motif in bacteria cells, the proteins have to be fused with a large tag to reduce their toxicity to bacteria. Protein folding is also another problem for apporpriately forming the disulphide bonds of a CSαβ motif. To overcome the problems, peptides with CSαβ motif can be fused with a thioredoxin tag and a cleavagable site is placed between the peptide and the tag. The recombinant proteins are expressed in *E. Coli* and purified. After the thioredoxin tag is cut off, the desired proteins are further purified. A simple method to cut off the thioredoxin tag is to place a acidohydrolysis site, -Asp-Pro-, between the peptide and the tag (Liu et al., 2006). A large amount of reconbinant proteins can be obtained by following the procedures and the native properties and functions of the proteins are preserved. The acidohydrolysis process is not suitable for proteins with a CSαβ motif containing proline for the proteins would be unstable in a heat and acidic environment (Cunningham & O'Connor, 1997;LeBlanc & London, 1997;Wilce et al., 1998). To resolve the problem, an alternative enzyme cleavage site can be introduced to replace the acidohydrolysis site.

### **5.4 Unpredictable biochemical function**

Enzyme inhibitory function and microbial killing ability are the two greatest advantages of peptide with a CSαβ motif. The conventional bacterial/fungal killing assay is time consuming and it is possible for a performer to be exposed to pathogens during the operation. The peptides with a CSαβ motif have been shown to form pores on an artificial membrane, but the artificial membrane does not represent the membrane of different organisms. A new high throughput screening recruiting membrane of different pathogens will be a great help in resolving the problem (Yang & Lyu, 2008).

Not simple as binding assay, enzymes act on specific substrates and it is extremely difficult to unify experimental conditions for a broad spectrum of enzymes. *In vitro* enzyme inhibitory assays are costly and unable to represent the actually physiological conditions. Computer aided drug screening methods have been established to screen drugs with small molecule weights that can fit into active sites of target proteins (Barrons, 2004;Weideman et al., 1999). Drug dynamics also can be *in silico* simulated (Sinek et al., 2009;Zunino et al., 2009). An antibody can be as a simple binder and its binding targets usually are short fragments exposed to the surface of the proteins. The binding fragments are predicable for a huge amount of knowledge accumulated in the last several decades (Blythe & Flower, 2005;Kulkarni-Kale et al., ;Odorico & Pellequer, 2003). Currently, we do not well understand the interface between peptides with a CSαβ motif and their targets. How functional loops access active sites of the targets is needed to be decrypted. We know too little about the scaffold and it is to difficult to engineer it. It is hard to develop an algorithm to predict the biochemical functions of the peptides with a CSαβ motif. For the complexity of proteinprotein interaction, it is a long way to screen peptide drugs *in silico*.
