Vitro Selection Strategies

In the following sections, an overview of the different approaches for in vitro selection is provided. Although this chapter focuses on proteins, we want to briefly explain a nucleic acid selection method that formed the basis of the in vitro evolution of functional proteins. 

In their original publication about SELEX, Tuerk and Gold (1990) already speculated that a similar approach could be adapted to protein selection. They referred to experiments describing the isolation of particular mRNAs from a pool of variants by immunoprecipitation of the nascent polypeptides present in the mRNA-ribosome-polypeptide complexes (Korman et al., 1982 Kraus and Rosenberg, 1982). In fact, soon after the publication of SELEX (Tuerk and Gold, 1990) a patent application was filed (Kawasaki, 1991), proposing a similar approach to enrich peptides from libraries. 

The first experimental demonstration of the ribosome display technology was the selection of short peptides Irom a library using an E. coli S30 in vitro translation system (Mattheakis et al., 1994 Mattheakis et al., 1996). The concept pursued by Mattheakis and coworkers (1994) for peptides was then used for the development of ribosome display of functional proteins by use of the E. coli S30 in vitro translation system (Hanes and Pluckthun, 1997). However, for this purpose it was necessary to significantly modify and optimize the experimental conditions of ribosome display to make this technology efficient enough for the display of correctly folded, functional proteins. 

In the E. coli system, it is important to stop the in vitro translation reaction by rapid cooling on ice. The reaction is usually diluted severalfold in prechilled buffer containing the components for stabilization of the ribosomal complexes. In the E. coli system, the ribosomal complexes can be very efficiently stabilized by low temperature and by high Mg2+ concentrations (50 mM), and then used for affinity selection. It is believed that high Mg2+ condenses the ribosome by binding totherRNA, making it difficult for the peptidyl-tRNA to dissociate or be hydrolyzed. The low temperature probably slows down the hydrolysis of the peptidyl-tRNA ester bond, and perhaps also the thermal motions, which would facilitate dissociation of the peptidyl-tRNA. Such complexes are stable for up to several days. 

The coding region ends with the protein sequence—that is, there is no stop codon present. In the prokaryotic system the presence of a stop codon would result in the binding of the release factors (Grentzmann et al, 1995 Tuite and Stansfield, 1994) and the ribosome recycling factor (Janosi et al., 1994) to the mRNA-ribosome-protein complexes. This would then lead to the release of the protein by hydrolysis of the peptidyl-tRNA (Tate and Brown, 1992), thereby dissociating the ribosomal complexes (Fig. 4A). A similar mechanism exists in eukaryotic systems (Frolova et al., 1994 Zhouravleva et al., 1995). 

---

In vitro selection strategies can be sub-divided into two types direct and indirect selections. These two types of selection experiments directed at the isolation of synthetic catalytic nucleic acids differ mainly by their technical concept, their design and their outcome. 

Tawfik and Griffith (1998) reported an in vitro selection strategy for catalytic activity using compartmentalization. Here, each member of the DNA library is encapsulated in an aqueous compartment in a water in oil emulsion. The compartments are generated from an in vitro transcription-translation system, and contain the components for protein synthesis. The dilution is chosen such that, on average, the water droplets contain less than one DNA molecule. The DNA is transcribed and translated in vitro in the presence of substrate, which is covalently attached to the DNA. Only translated proteins with catalytic activity convert the substrate to the product. Subsequently, all DNA molecules are recovered from the water droplets and the DNA linked to the product is separated from the unmodified DNA linked to the educt, which requires a method to discriminate between both. The modified DNA can then be amplified by PCR and used for a second selection cycle. The principle of this approach is depicted in Figure 6. 

At the same time, several laboratories have used in-vitro selection strategies to identify RNAs and DNAs that catalyze specific reactions [34-40]. The information from these and related studies now opens the possibility of expanding the molecular skeletons capable of storing sequence information into ones that can be read into complementary materials [41]. 

In a subsequent study, a similar in vitro selection strategy was applied towards Mg -, Mn -, and Zn -dependent DNAzymes with RNAse activity [32]. Although the reaction with the resultant DNA enzymes proceeded around 17 times slower than with the Pb -based system, focus remained on the Mg " -dependent systems, as they might eventually be active under intracellular conditions. Next, a transacting catalyst capable of 17 turnovers of a chimeric substrate within 5 h was developed. The most active clone was suggested to contain a three-stem junction, based on the sequence. This makes the structure more complex than that of its Pb " -dependent analog. 

By simulating evolution in vitro it has become possible to isolate artificial ribozymes from synthetic combinatorial RNA libraries [1, 2]. This approach has great potential for many reasons. First, this strategy enables generation of catalysts that accelerate a variety of chemical reactions, e.g. amide bond formation, N-glycosidic bond formation, or Michael reactions. This combinatorial approach is a powerful tool for catalysis research, because neither prior knowledge of structural prerequisites or reaction mechanisms nor laborious trial-and-error syntheses are necessary (also for non-enzymatic reactions, as discussed in Chapter 5.4). The iterative procedure of in-vitro selection enables handling of up to 1016 different compounds... 

In an approach similar to the cell-like compartments, Doi and Yanagawa (1999) used biotinylated DNA to display peptides fused to streptavidin in compartments of water in oil emulsions. The method was named streptavidin-biotin linkage in emulsions, STABLE (Doi and Yanagawa, 1999). Upon in vitro translation each translated peptide is displayed as a fusion to streptavidin that binds to its encoding biotinylated DNA in its compartment. The resulting protein-DNA fusions can then be recovered and used for affinity selection. To avoid cross-contamination, biotin has to be added before recovery because much more streptavidin will be produced in each compartment than biotinylated DNA is present. The selected DNA-protein complexes can then be amplified by PCR. The principle of this selection strategy is shown in Figure 7. 

The in vitro selection/amplification strategy has also been applied to modified ONs, especially where the 2 -ribose position has been changed and where phosphorothioates or other phosphate replacements have been used (Fig. 10.25, top). Several strucmres of modified ON chains that have been synthetically produced to obtain constrained sequences or sequences with higher stability to nucleases have also been reported (Fig. 10.25, bottom). Examples of biosynthetic modified ON libraries are covered in the next section. 

Ribozyme-catalyzed reactions involving C-C bond formations have also been reported. Seelig and Jaschke (233) presented the in vitro selection of ribozyme catalysts for the Diels-Alder reaction between maleimide and anthracene, employing a 2 X lO -member library of 160-mer modified ONs (L28) with 120 randomized positions. The selection strategy used is shown in Fig. 10.40. Library L28 was prepared from the corresponding dsDNA sequences, and transcription initiation was performed in the presence of ternary complexes between guanosine monophosphate (10.57), PEG (10.58), and anthracene (10.59, step a. Fig. 10.40). The library obtained contained a 5 -anthracene-PEG appendage and was incubated with biotin-modified maleimide... 

Figure 9.1. An in-vitro selection experiment comprises various sequential steps, of which the first is the generation of a nucleic acid library of completely random sequences. This library is subjected to an appro-. priatc selection strategy which allows the separation of functional molecules from non-functional ones. The small proportion of nucleic acids with the desired activity is then amplified enzymatically and re-suh-jected to the selection procedure. This is necessary as the complexity of the library, which can contain up to 1016 different oligonucleotide sequences, makes it impossible to enrich for the active sequences in one single selection and amplification cycle. Therefore, a number of cycles are performed sequentially until the functional sequences are the majority species in the library mix, and these can be characterized by cloning and sequencing.

---