2Alacris Theranostics GmbH, Berlin, Germany
3Department of Vertebrate Genomics, Max Planck Institute for Molecular Genetics, Berlin, Germany
Directed evolution of nucleotide libraries using recombination or mutagenesis is an important technique for customizing catalytic or biophysical traits of proteins. Conventional directed evolution methods, however, suffer from cumbersome digestion and ligation steps. Here, we describe a simple method to increase nucleotide diversity using single-stranded DNA (ssDNA) as a starting template. An initial PCR amplification using phosphorylated primers with overlapping regions followed by treatment with lambda exonuclease generates ssDNA templates that can then be annealed via the overlap regions. Double-stranded DNA (dsDNA) is then generated through extension with Klenow fragment. To demonstrate the applicability of this methodology for directed evolution of nucleotide libraries, we generated both gene shuffled and regional mutagenesis synthetic antibody libraries with titers of 2×108 and 6×107, respectively. We conclude that our method is an efficient and convenient approach to generate diversity in nucleic acid based libraries, especially recombinant antibody libraries.
Directed evolution allows an in vitro mimicry of the natural in vivo evolution process to generate new or improved traits. The directed evolution process involves repetitive cycles of genotype diversification accompanied by a physical selection process to sieve out the best phenotype. The approach is synonymous with display techniques, such as phage, yeast, and ribosome display, that promote a physical linkage between genotype and phenotype. A basic requirement for any display technique is a diverse nucleotide library (1). Numerous methods have been developed to introduce diversity into these libraries, including random mutagenesis (2), in vitro or in vivo DNA shuffling (3-5) and site-specific recombination (6-8). Random mutagenesis can be carried out using several different methods, although most commonly through the use of a low fidelity DNA polymerase with MnCl2 (9) and Escherichia coli mutator strains (10). Other methods include mutagenic polymerases (11), mixtures of triphosphate derivatives of nucleoside analogs (12), site directed mutagenesis (13), mutational hotspots (14), and parsimonious mutagenesis (15). Although random mutagenesis can generate randomization at any position in the sequence, the technique is mostly limited to short stretches of DNA. This prevents the alteration of entire binding sites, which are generally longer DNA regions. In addition, as some mutations may not be directed, errors such as frameshifts or stop codons could also occur during mutagenesis that may result in the disruption of the phenotype.
Antibody libraries are the mainstay in display approaches due to the nature of diversification present in antibody genes. Here, gene shuffling can be accomplished by numerous methods including chain shuffling (16), DNA shuffling (4) or staggered extension process (StEP) (17). Chain shuffling makes use of shuffling either the heavy or light chain variable regions of the antibody genes to generate new variants by conventional restriction digestion and ligation. In DNA shuffling, the antibody gene is digested with DNase I, randomly reassembled and amplified by PCR. StEP, which is also PCR-based, allows template switching by shortening extension times and, hence, will shuffle various portions of several parental antibody genes. DNA shuffling has been widely used for library generation, and a few recent modifications have been described that improve the efficiency of the technique including the use of single-stranded DNA (ssDNA) instead of double-stranded DNA (dsDNA) as template and use of restriction enzymes or endonuclease V instead of DNase I during DNAs fragmentation (18, 19). However, this method still presents limitations, including low frequency of chimeric genes due to preferred homoduplex formation, a limited distribution of restriction site, and a lack of high-resolution crossover (20, 21).
Lambda exonuclease is an enzyme that assists in the repair of dsDNA breaks in viral DNA (22). Lambda exonuclease is a highly processive 5′→3′ dsDNA exonuclease that selectively degrades a phosphorylated chain of the duplex to yield mononucleotides and ssDNA (23-25). ssDNA template used in directed evolution experiments is usually generated as the substrate for pairing enzymes that promote homologous recombination. The main characteristic of the enzyme is the requirement of a phosphate group at the 5′ dsDNA end. The use of lambda exonuclease to generate ssDNA has assisted in other method developments and technologies such as next-generation sequencing platforms (26), DNA-chips (27), SELEX (28), subtractive hybridization techniques (29), sample preparation for electrospray ionization mass spectrometry (30), and recombination methods (31).
DNA polymerase I is a single polypeptide chain with three separate functional domains: a polymerization domain, a 3′ exonuclease (or proofreading) domain, and a 5′ nuclease domain (32). Klenow fragment (KF) of DNA polymerase I is obtained by removing the 5′ nuclease domain from the protein by limited protease digestion. KF is a 68 kDa protein with the polymerase and 3′→5′ proof reading exonuclease activity, but lacking 5′→3′ exonuclease activity (33). Thus, KF is commonly used for fill-in reactions during molecular cloning as the enzyme has the ability to add nucleotides without degrading any overhangs. In addition, KF has been used to generate dsDNA from highly complexed templates, such as degenerate oligos for peptide-displaying phage library cloning (34). Another application of KFs is to conduct low temperature PCR cycling with the aid of proline (35).