I. General Introduction
5. Directed Evolution of DNA Polymerases
The process of in vitro evolution, especially directed enzyme evolution,[126, 127] has proven to be a powerful method to generate these enzyme variants with improved or new properties tailored for specific applications. In contrast to a rational design of mutants, the directed evolution of proteins requires no structural information of the protein, as mutations are introduced randomly. The method comprises an iterative process of three different steps:
Random mutagenesis in order to generate a library of enzyme variants, the expression of the enzymes and a subsequent screening or selection step. During this process mutations accumulate until a desired level of improvement is achieved, with the gene of the most promising variant selected after every round and employed as template in the next cycle.[128]
Mutations can be either introduced on the entire target gene coding for the respective protein or on selected amino acid positions. The introduction can be facilitated by various techniques such as saturation mutagenesis, DNA shuffling, StEP (Staggered Extension Process) or error-prone PCR.[121, 129-131] Subsequent transformation into a host organism, e.g. E. coli, generates the library. However, it is crucial that the phenotype and genotype are ‘connected’ in the library due to the following selection or screening step. High-throughput screening strategies achieve separation from other variants by conventional compartmentalization based on multi-well plates. One method for high-throughput screening was established in our lab which employs the fluorescent dye SYBRGreen I to identify active polymerase mutants.[132] The dye exhibits an increased fluorescence signal upon binding to the minor groove of double-stranded DNA (emission at 520 nm). Therefore, the amplification of DNA in PCR by active DNA polymerase mutants can be visualized either in real time or through end-point determination.
Connecting the phenotype to the genotype in selection based strategies can be facilitated e.g.
in phage display, ribosome display, mRNA display or water-oil emulsions.[133-135] The method of compartmentalized self-replication (CSR)[136] also relies on the formation of water and oil emulsions and provides a powerful tool for the evolution of DNA polymerases. It is based on a simple feedback loop with active polymerases replicating their own gene. Thus, adaptive gain is directly translated into genetic amplification of the encoding gene.
5.2 DNA Shuffling
The next chapter will focus on DNA shuffling, as it was the method of choice in this work. DNA shuffling is defined as the in vitro recombination of selected genes by random fragmentation and PCR reassembly.[130, 137] This method is based on four different steps consisting of gene preparation, DNA fragmentation, reassembly of these fragments in a self-priming polymerase reaction and an amplification of the recombined fragments in PCR (Figure 10).
DNA fragmentation can be achieved via DNase I digestion of the parental DNA[130, 137] or via short randomly designed primers which anneal to the parental DNA and are extended by a DNA polymerase at or below room temperature.[138] In the following reassembly step, the fragmented genes are reassembled in a ‘reverse’ PCR without using primers. The fragments replace the primer, as homologous stretches anneal and form a primer/template complex elongated by a thermostable DNA polymerase. Consequently, the number of DNA molecules decreases during DNA reassembly, whereas in standard PCR the number of DNA molecules exponentially increases.[137] This step also offers the possibility to either introduce mutations
DNA shuffling was first reported to be successfully applied in a -lactamase model system resulting in enzyme mutants with an increased antibiotic resistance against cefotaxime.[130] A high point mutation rate of 0.7 % was observed which is comparable to the rate in error-prone PCR. Whereas a high-error rate is desired for gaining diversity in in vitro evolution applications, the opposite is true for studies focusing on the structure-function relationship between homologous genes or, studies in which beneficial mutations were already identified and the respective mutants are to be recombined without gaining new mutations.
Consequently, protocols were developed in which each step was optimized to yield a low error-rate.[139] Thus, one protocol reported an error-rate as low as 0.05 %, which was mainly achieved by including high-fidelity DNA polymerases during gene preparation, reassembly and in the post-amplification PCR step.[139]
Figure 10. Principle of DNA shuffling.
Depicted are the fragmentation of the parental DNA (homologous genes), the recombination, reassembly and amplifi-cation in PCR.
5.3 Thermostable DNA Polymerases with Reverse Transcriptase Activity
The evolution of DNA polymerases towards the acceptance of non-natural substrates facilitates a variety of applications in both molecular biology and diagnostics, as described before (chapter I 5.1). The acceptance of non-cognate substrates by DNA-directed DNA polymerases also includes the usage of RNA as template for DNA synthesis in a process called reverse transcription. How DNA- or RNA- dependent DNA polymerases discriminate between the natural templates (RNA vs DNA) and maintain their substrate specificity is still a subject of ongoing investigations.[140, 141] Structural studies yielding insights into this process are lacking and thus designing DNA polymerases in a rational fashion to accept both, DNA and RNA, as template remains a challenge. However, thermostable DNA polymerases accepting both substrates would provide a crucial tool for the so-called reverse transcription PCR (RT-PCR), a fundamental technique utilized in many applications in molecular biology and clinical diagnostics such as transcriptome analysis, pathogen detection as well as disease-specific marker recognition.[14, 142]The detection and quantification of RNA in RT-PCR is generally based on the enzyme-mediated reverse transcription of RNA to its complementary DNA (cDNA) by a reverse transcriptase and a subsequent amplification of the resulting DNA by a DNA-dependent DNA polymerase in PCR. The detection can be even monitored in real time. The reverse transcriptase and the DNA-dependent DNA polymerase can be applied either in separate (two-enzymes/two tubes) or single (two-enzymes/one tube) reactions. One tube reactions having the reverse transcription prior to PCR amplification, termed one-step RT-PCR, are time- and work-saving. Additionally, the risk of contamination is reduced as, in general, an RNA digestion step or the addition of different buffers can be omitted.[14, 142]
Although two enzyme mixtures are state of the art, several drawbacks arise from the heat-instability of commonly used retroviral mesophilic reverse-transcriptases[143] such as MoMLV and AMV. Performing the reverse transcription step within a one-step RT-PCR set-up requires low temperatures (i.e. 45 °C) to allow activity of the reverse transcriptase, which facilitates unspecific priming, low yield on complex targets e.g. from secondary structure formation of the mRNA template and premature reaction termination.[142] Furthermore, the reverse transcription step results in a time addition to the PCR protocol, a disadvantage especially in the field of point of care testing or outbreak situations when hundreds of swabs need to be analysed in a short period of time. Therefore, the development or discovery of heat-stabile reverse transcriptases would be desirable but was shown to have its limitations.[143-146] So far an increase in thermostability was gained by eliminating the RNase H activity,[143] by site- directed mutagenesis[145, 146] or random mutations,[144] but the achieved thermostability was insufficient for the use of these enzymes in PCR.
Consequently, strong efforts have been undertaken to evolve thermostable DNA-dependent DNA polymerases with reverse transcriptase activity applicable in RT-PCR.[147-151] These enzymes offer the possibility to perform one step RT-PCR at high temperatures minimizing
shown to be applicable in RT-PCR[147, 148, 150, 151] and to the best of my knowledge only two of these enzymes are currently commercially available, one belonging to sequence family A and one isolated from a viral metagenomic library.[147, 151] Thus, the demand for DNA polymerases with increased reverse transcriptase activity persists.