Predicting crossover generation in DNA shuffling

A patent-based consideration of latest platforms in the art of directed evolution: a decade long untold story

Directed (or in vitro) evolution of proteins and metabolic pathways requires tools for creating genetic diversity and identifying protein variants with new or improved functional properties. Besides simplicity, reliability, speed, versatility, universal applicability and economy of the technique, the new science of synthetic biology requires improved means for construction of smart and high-quality mutant libraries to better navigate the sequence diversity. In vitro CRISPR/Cas9-mediated mutagenic (ICM) system and machine-learning (ML)-assisted approaches to directed evolution are now in the field to achieve the goal. This review describes the gene diversification strategies, screening and selection methods, in silico (computer-aided), Cas9-mediated and ML-based approaches to mutagenesis, developed especially in the last decade, and their patent position. The objective behind is to emphasize researchers the need for noting which mutagenesis, screening or selection method is patented and then selecting a suitable restriction-free approach to sequence diversity. Techniques and evolved products subject to patent rights need commercial license if their use is for purposes other than private or experimental research.

Forty Years of Directed Evolution and its Continuously Evolving Technology Toolbox - A Review of the Patent Landscape

Generating functional protein variants with novel or improved characteristics has been a goal of the biotechnology industry and life sciences, for decades. Rational design and directed evolution are two major pathways to achieve the desired ends. Whilst rational protein design approach has made substantial progress, the idea of using a method based on cycles of mutagenesis and natural selection to develop novel binding proteins, enzymes and structures has attracted great attention. Laboratory evolution of proteins/enzymes requires new tools and analytical approaches to create genetic diversity and identifying variants with desired traits. In this pursuit, construction of sufficiently large libraries of target molecules to search for improved variants and the need for new protocols to alter the properties of target molecules has been a continuing challenge in the directed evolution experiments. This review will discuss the in vivo and in vitro gene diversification tools, library screening or selection approaches, and artificial intelligence/machine-learning-based strategies to mutagenesis developed in the last forty years to accelerate the natural process of evolution in creating new functional protein variants, optimization of microbial strains and transformation of enzymes into industrial machines. Analyzing patent position over these techniques and mechanisms also constitutes an integral and distinctive part of this review. The aim is to provide an up-to-date resource/technology toolbox for research-based and pharmaceutical companies to discover the boundaries of competitor's intellectual property (IP) portfolio, their freedom-to-operate in the relevant IP landscape, and the need for patent due diligence analysis to rule out whether use of a particular patented mutagenesis method, library screening/selection technique falls outside the safe harbor of experimental use exemption. While so doing, we have referred to some recent cases that emphasize the significance of selecting a suitable gene diversification strategy in directed evolution experiments. This article is protected by copyright. All rights reserved.

Codon-Optimization of Wild-Type Adeno-Associated Virus Capsid Sequences Enhances DNA Family Shuffling while Conserving Functionality

Adeno-associated virus (AAV) vectors have become one of the most widely used gene transfer tools in human gene therapy. Considerable effort is currently being focused on AAV capsid engineering strategies with the aim of developing novel variants with enhanced tropism for specific human cell types, decreased human seroreactivity, and increased manufacturability. Selection strategies based on directed evolution rely on the generation of highly variable AAV capsid libraries using methods such as DNA-family shuffling, a technique reliant on stretches of high DNA sequence identity between input parental capsid sequences. This identity dependence for reassembly of shuffled capsids is inherently limiting and results in decreased shuffling efficiency as the phylogenetic distance between parental AAV capsids increases. To overcome this limitation, we have developed a novel codon-optimization algorithm that exploits evolutionarily defined codon usage at each amino acid residue in the parental sequences. This method increases average sequence identity between capsids, while enhancing the probability of retaining capsid functionality, and facilitates incorporation of phylogenetically distant serotypes into the DNA-shuffled libraries. This technology will help accelerate the discovery of an increasingly powerful repertoire of AAV capsid variants for cell-type and disease-specific applications.

Shuffle Optimizer: A Program to Optimize DNA Shuffling for Protein Engineering

DNA shuffling is a powerful tool to develop libraries of variants for protein engineering. Here, we present a protocol to use our freely available and easy-to-use computer program, Shuffle Optimizer. Shuffle Optimizer is written in the Python computer language and increases the nucleotide homology between two pieces of DNA desired to be shuffled together without changing the amino acid sequence. In addition we also include sections on optimal primer design for DNA shuffling and library construction, a small-volume ultrasonicator method to create sheared DNA, and finally a method to reassemble the sheared fragments and recover and clone the library. The Shuffle Optimizer program and these protocols will be useful to anyone desiring to perform any of the nucleotide homology-dependent shuffling methods.

DNA secondary structure formation by DNA shuffling of the conserved domains of the Cry protein of Bacillus thuringiensis

Background

The Cry toxins, or δ-endotoxins, are a diverse group of proteins produced by Bacillus thuringiensis. While DNA secondary structures are biologically relevant, it is unknown if such structures are formed in regions encoding conserved domains of Cry toxins under shuffling conditions. We analyzed 5 holotypes that encode Cry toxins and that grouped into 4 clusters according to their phylogenetic closeness. The mean number of DNA secondary structures that formed and the mean Gibbs free energy \documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$ \left(\overline{\varDelta G}\right) $$\end{document}ΔG¯ were determined by an in silico analysis using different experimental DNA shuffling scenarios. In terms of spontaneity, shuffling efficiency was directly proportional to the formation of secondary structures but inversely proportional to ∆G.

Results

The results showed a shared thermodynamic pattern for each cluster and relationships among sequences that are phylogenetically close at the protein level. The regions of the cry11Aa, Ba and Bb genes that encode domain I showed more spontaneity and thus a greater tendency to form secondary structures (<∆G). In the region of domain III; this tendency was lower (>∆G) in the cry11Ba and Bb genes. Proteins that are phylogenetically closer to Cry11Ba and Cry11Bb, such as Cry2Aa and Cry18Aa, maintained the same thermodynamic pattern. More distant proteins, such as Cry1Aa, Cry1Ab, Cry30Aa and Cry30Ca, featured different thermodynamic patterns in their DNA.

Conclusion

These results suggest the presence of thermodynamic variations associated to the formation of secondary structures and an evolutionary relationship with regions that encode highly conserved domains in Cry proteins. The findings of this study may have a role in the in silico design of cry gene assembly by DNA shuffling techniques.

Assembly of Designed Oligonucleotides: a useful tool in synthetic biology for creating high-quality combinatorial DNA libraries

The method dubbed Assembly of Designed Oligonucleotides (ADO) is a powerful tool in synthetic biology to create combinatorial DNA libraries for gene, protein, metabolic, and genome engineering. In directed evolution of proteins, ADO benefits from using reduced amino acid alphabets for saturation mutagenesis and/or DNA shuffling, but all 20 canonical amino acids can be also used as building blocks. ADO is performed in a two-step reaction. The first involves a primer-free, polymerase cycling assembly or overlap extension PCR step using carefully designed overlapping oligonucleotides. The second step is a PCR amplification using the outer primers, resulting in a high-quality and bias-free double-stranded DNA library that can be assembled with other gene fragments and/or cloned into a suitable plasmid subsequently. The protocol can be performed in a few hours. In theory, neither the length of the DNA library nor the number of DNA changes has any limits. Furthermore, with the costs of synthetic DNA dropping every year, after an initial investment is made in the oligonucleotides, these can be exchanged for alternative ones with different sequences at any point in the process, fully exploiting the potential of creating highly diverse combinatorial libraries. In the example chosen here, we show the construction of a high-quality combinatorial ADO library targeting sixteen different codons simultaneously with nonredundant degenerate codons encoding various reduced alphabets of four amino acids along the heme region of the monooxygenase P450-BM3.

Restriction enzyme-mediated DNA family shuffling

DNA shuffling is an established recombinatorial method that was originally developed to increase the speed of directed evolution experiments beyond what could be accomplished using error-prone PCR alone. To achieve this, mutated copies of a protein-coding sequence are fragmented with DNase I and the fragments are then reassembled in a PCR without primers. The fragments anneal where there is sufficient sequence identity, resulting in full-length variants of the original gene that have inherited mutations from multiple templates. Subsequent studies demonstrated that directed evolution could be further accelerated by shuffling similar native protein-coding sequences from the same gene family, rather than mutated variants of a single gene. Generally at least 65-75 % global identity between parental sequences is required in DNA family shuffling, with recombination mostly occurring at sites with at least five consecutive nucleotides of local identity. Since DNA shuffling was originally developed, many variations on the method have been published. In particular, the use of restriction enzymes in the fragmentation step allows for greater customization of fragment lengths than DNase I digestion and avoids the risk that parental sequences may be over-digested into unusable very small fragments. Restriction enzyme-mediated fragmentation also reduces the occurrence of undigested parental sequences that would otherwise reduce the number of unique variants in the resulting library. In the current chapter, we provide a brief overview of the alternative methods currently available for DNA shuffling as well as a protocol presented here that improves on several previous implementations of restriction enzyme-mediated DNA family shuffling, in particular with regard to purification of DNA fragments for reassembly.

Engineering cellulosic bioreactors by template assisted DNA shuffling and in vitro recombination (TADSir)

The current study focuses on development of a bioreactor engineering strategy based on exploitation of the Arabidopsis thaliana genome. Chimeric A. thaliana glycosyl hydrolase (GH) gene libraries were assembled using a novel directed evolution strategy (TADSir: template assisted DNA shuffling and in vitro recombination) that promotes DNA recombination by reassembly of DNA fragments on unique gene templates. TADSir was modeled using a set of algorithms designed to simulate DNA interactions based on nearest neighbor base stacking interactions and Gibb's free energy differences between helical coil and folded DNA states. The algorithms allow for target gene prediction and for in silica analysis of chimeric gene library composition. Further, the study investigated utilization of A. thaliana GH sequence space for bioreactor design by evolving 20 A. thaliana genes representing the GH1, GH3, GH5, GH9 and GH10 gene families. Notably, TADSir achieved streamlined engineering of Saccharomyces cerevisiae and spinach mesophyll protoplast bioreactors capable of processing CM cellulose, Avicel and xylan.

Probabilistic methods in directed evolution: library size, mutation rate, and diversity

Directed evolution has emerged as an important tool for engineering proteins with improved or novel properties. Because of their inherent reliance on randomness, directed evolution protocols are amenable to probabilistic modeling and analysis. This chapter summarizes and reviews in a nonmathematical way some of the probabilistic works related to directed evolution, with particular focus on three of the most widely used methods: saturation mutagenesis, error-prone PCR, and in vitro recombination. The ultimate aim is to provide the reader with practical information to guide the planning and design of directed evolution studies. Importantly, the applications and locations of freely available computational resources to assist with this process are described in detail.

Identification of amino acids important for substrate specificity in sucrose transporters using gene shuffling

Plant sucrose transporters (SUTs) are H(+)-coupled uptake transporters. Type I and II (SUTs) are phylogenetically related but have different substrate specificities. Type I SUTs transport sucrose, maltose, and a wide range of natural and synthetic α- and β-glucosides. Type II SUTs are more selective for sucrose and maltose. Here, we investigated the structural basis for this difference in substrate specificity. We used a novel gene shuffling method called synthetic template shuffling to introduce 62 differentially conserved amino acid residues from type I SUTs into OsSUT1, a type II SUT from rice. The OsSUT1 variants were tested for their ability to transport the fluorescent coumarin β-glucoside esculin when expressed in yeast. Fluorescent yeast cells were selected using fluorescence-activated cell sorting (FACS). Substitution of five amino acids present in type I SUTs in OsSUT1 was found to be sufficient to confer esculin uptake activity. The changes clustered in two areas of the OsSUT1 protein: in the first loop and the top of TMS2 (T80L and A86K) and in TMS5 (S220A, S221A, and T224Y). The substrate specificity of this OsSUT1 variant was almost identical to that of type I SUTs. Corresponding changes in the sugarcane type II transporter ShSUT1 also changed substrate specificity, indicating that these residues contribute to substrate specificity in type II SUTs in general.

Development of chimeric laccases by directed evolution

DNA recombination methods are useful tools to generate diversity in directed evolution protein engineering studies. We have designed an array of chimeric laccases with high-redox potential by in vitro and in vivo DNA recombination of two fungal laccases (from Pycnoporus cinnabarinus and PM1 basidiomycete), which were previously tailored by laboratory evolution for functional expression in Saccharomyces cerevisiae. The laccase fusion genes (including the evolved α-factor prepro-leaders for secretion in yeast) were subjected to a round of family shuffling to construct chimeric libraries and the best laccase hybrids were identified in dual high-throughput screening (HTS) assays. Using this approach, we identified chimeras with up to six crossover events in the whole sequence, and we obtained active hybrid laccases with combined characteristics in terms of pH activity and thermostability.

Algorithms for optimizing cross-overs in DNA shuffling

BACKGROUND

DNA shuffling generates combinatorial libraries of chimeric genes by stochastically recombining parent genes. The resulting libraries are subjected to large-scale genetic selection or screening to identify those chimeras with favorable properties (e.g., enhanced stability or enzymatic activity). While DNA shuffling has been applied quite successfully, it is limited by its homology-dependent, stochastic nature. Consequently, it is used only with parents of sufficient overall sequence identity, and provides no control over the resulting chimeric library.

RESULTS

This paper presents efficient methods to extend the scope of DNA shuffling to handle significantly more diverse parents and to generate more predictable, optimized libraries. Our CODNS (cross-over optimization for DNA shuffling) approach employs polynomial-time dynamic programming algorithms to select codons for the parental amino acids, allowing for zero or a fixed number of conservative substitutions. We first present efficient algorithms to optimize the local sequence identity or the nearest-neighbor approximation of the change in free energy upon annealing, objectives that were previously optimized by computationally-expensive integer programming methods. We then present efficient algorithms for more powerful objectives that seek to localize and enhance the frequency of recombination by producing "runs" of common nucleotides either overall or according to the sequence diversity of the resulting chimeras. We demonstrate the effectiveness of CODNS in choosing codons and allocating substitutions to promote recombination between parents targeted in earlier studies: two GAR transformylases (41% amino acid sequence identity), two very distantly related DNA polymerases, Pol X and β (15%), and beta-lactamases of varying identity (26-47%).

CONCLUSIONS

Our methods provide the protein engineer with a new approach to DNA shuffling that supports substantially more diverse parents, is more deterministic, and generates more predictable and more diverse chimeric libraries.

Replacing Mn(2+) with Co(2+) in human arginase i enhances cytotoxicity toward l-arginine auxotrophic cancer cell lines

Replacing the two Mn(2+) ions normally present in human Arginase I with Co(2+) resulted in a significantly lowered K(M) value without a concomitant reduction in k(cat). In addition, the pH dependence of the reaction was shifted from a pK(a) of 8.5 to a pK(a) of 7.5. The combination of these effects led to a 10-fold increase in overall catalytic activity (k(cat)/K(M)) at pH 7.4, close to the pH of human serum. Just as important for therapeutic applications, Co(2+) substitution lead to significantly increased serum stability of the enzyme. Our data can be explained by direct coordination of l-Arg to one of the Co(2+) ions during reaction, consistent with previously reported model studies. In vitro cytotoxicity experiments verified that the Co(2+)-substituted human Arg I displays an approximately 12- to 15-fold lower IC(50) value for the killing of human hepatocellular carcinoma and melanoma cell lines and thus constitutes a promising new candidate for the treatment of l-Arg auxotrophic tumors.

In vitro DNA recombination by L-Shuffling during ribosome display affinity maturation of an anti-Fas antibody increases the population of improved variants

The use of random mutagenesis in concert with protein display technologies to rapidly select high affinity antibody variants is an established methodology. In some cases, DNA recombination has been included in the strategy to enable selection of mutations which act cooperatively to improve antibody function. In this study, the impact of L-Shuffling DNA recombination on the eventual outcome of an in vitro affinity maturation has been experimentally determined. Parallel evolution strategies, with and without a recombination step, were carried out and both methods improved the affinity of an anti-Fas single chain variable fragment (scFv). The recombination step resulted in an increased population of affinity-improved variants. Moreover, the most improved variant, with a 22-fold affinity gain, emerged only from the recombination-based approach. An analysis of mutations preferentially selected in the recombined population demonstrated strong cooperative effects when tested in combination with other mutations but small, or even negative, effects on affinity when tested in isolation. These results underline the ability of combinatorial library approaches to explore very large regions of sequence space to find optimal solutions in antibody evolution studies.

Revealing biases inherent in recombination protocols

Background

The recombination of homologous genes is an effective protein engineering tool to evolve proteins. DNA shuffling by gene fragmentation and reassembly has dominated the literature since its first publication, but this fragmentation-based method is labor intensive. Recently, a fragmentation-free PCR based protocol has been published, termed recombination-dependent PCR, which is easy to perform. However, a detailed comparison of both methods is still missing.

Results

We developed different test systems to compare and reveal biases from DNA shuffling and recombination-dependent PCR (RD-PCR), a StEP-like recombination protocol. An assay based on the reactivation of β-lactamase was developed to simulate the recombination of point mutations. Both protocols performed similarly here, with slight advantages for RD-PCR. However, clear differences in the performance of the recombination protocols were observed when applied to homologous genes of varying DNA identities. Most importantly, the recombination-dependent PCR showed a less pronounced bias of the crossovers in regions with high sequence identity. We discovered that template variations, including engineered terminal truncations, have significant influence on the position of the crossovers in the recombination-dependent PCR. In comparison, DNA shuffling can produce higher crossover numbers, while the recombination-dependent PCR frequently results in one crossover. Lastly, DNA shuffling and recombination-dependent PCR both produce counter-productive variants such as parental sequences and have chimeras that are over-represented in a library, respectively. Lastly, only RD-PCR yielded chimeras in the low homology situation of GFP/mRFP (45% DNA identity level).

Conclusion

By comparing different recombination scenarios, this study expands on existing recombination knowledge and sheds new light on known biases, which should improve library-creation efforts. It could be shown that the recombination-dependent PCR is an easy to perform alternative to DNA shuffling.

A comparative study of metabolic engineering anti-metabolite tolerance in Escherichia coli

A problem in strain engineering is that mutations that benefit the expression of a phenotype in one environment may impose a cost to biological fitness in a new environment. The overall objective of this study was to improve understanding of this phenomenon within the context of a classic anti-metabolite selection strategy. We have engineered Escherichia coli using three mutagenesis techniques (chemical mutagenesis, insertional mutagenesis, and plasmid-based overexpression) and assessed the relative costs and benefits to biological fitness of mutants selected for tolerance to five amino acid analogs whose target amino acids (glutamatic acid, aspartic acid, tryptophan, glycine, and serine) differ in metabolic connectivity and biosynthetic energy requirements. Our major findings include (i) the fold increase in anti-metabolite tolerance, independent of mutagenesis strategy, was much greater for aspartic acid beta-hydroxamate (AAH) compared to all other tested hydroxamates, (ii) increased tolerance to glutamic acid gamma-hydroxamate (GAH) was not achieved using any of the mutagenesis strategies, and (iii) characteristics of the anti-metabolite, rather than those of the corresponding metabolite, were more important in determining the ability to increase tolerance.

IPRO: an iterative computational protein library redesign and optimization procedure

A number of computational approaches have been developed to reengineer promising chimeric proteins one at a time through targeted point mutations. In this article, we introduce the computational procedure IPRO (iterative protein redesign and optimization procedure) for the redesign of an entire combinatorial protein library in one step using energy-based scoring functions. IPRO relies on identifying mutations in the parental sequences, which when propagated downstream in the combinatorial library, improve the average quality of the library (e.g., stability, binding affinity, specific activity, etc.). Residue and rotamer design choices are driven by a globally convergent mixed-integer linear programming formulation. Unlike many of the available computational approaches, the procedure allows for backbone movement as well as redocking of the associated ligands after a prespecified number of design iterations. IPRO can also be used, as a limiting case, for the redesign of a single or handful of individual sequences. The application of IPRO is highlighted through the redesign of a 16-member library of Escherichia coli/Bacillus subtilis dihydrofolate reductase hybrids, both individually and through upstream parental sequence redesign, for improving the average binding energy. Computational results demonstrate that it is indeed feasible to improve the overall library quality as exemplified by binding energy scores through targeted mutations in the parental sequences.

Towards the preparative and large-scale precision manufacture of virus-like particles

Virus-like particles (VLPs) are of interest in vaccination, gene therapy and drug delivery, but their potential has yet to be fully realized. This is because existing laboratory processes, when scaled, do not easily give a compositionally and architecturally consistent product. Research suggests that new process routes might ultimately be based on chemical processing by self-assembly, involving the precision manufacture of precursor capsomeres followed by in vitro VLP self-assembly and scale-up to required levels. A synergistic interaction of biomolecular design and bioprocess engineering (i.e. biomolecular engineering) is required if these alternative process routes and, thus, the promise of new VLP products, are to be realized.

Strategies and computational tools for improving randomized protein libraries

In the last decade, directed evolution has become a routine approach for engineering proteins with novel or altered properties. Concurrently, a trend away from purely 'blind' randomization strategies and towards more 'semi-rational' approaches has also become apparent. In this review, we discuss ways in which structural information and predictive computational tools are playing an increasingly important role in guiding the design of randomized libraries: web servers such as ConSurf-HSSP and SCHEMA allow the prediction of sites to target for producing functional variants, while algorithms such as GLUE, PEDEL and DRIVeR are useful for estimating library completeness and diversity. In addition, we review recent methodological developments that facilitate the construction of unbiased libraries, which are inherently more diverse than biased libraries and therefore more likely to yield improved variants.

Evolution of highly active enzymes by homology-independent recombination

The theta-class GST enzymes hGSTT1-1 (human GSTTheta-1-1) and rGSTT2-2 (rat GSTTheta-2-2) share 54.3% amino acid identity and exhibit different substrate specificities. Homology-independent techniques [incremental truncation for the creation of hybrid enzymes (ITCHY) and SCRATCHY] and low-homology techniques (recombination-dependent exponential amplification PCR) were used to create libraries of chimeric enzymes containing crossovers (C/Os) at positions not accessible by DNA family shuffling. High-throughput flow cytometric screening using the fluorogenic rGSTT2-2-specific substrate 7-amino-4-chloromethyl coumarin led to the isolation of active variants with either one or two C/Os. One of these enzymes, SCR23 (83% identity to hGSTT1-1), was encoded by a gene that exchanged helices 4 and 5 of hGSTT1-1 with the corresponding sequence from rGSTT2-2. Compared with either parent, this variant was found to have an improved k(cat) with the selection substrate and also exhibited activity for the conjugation of glutathione to ethacrynic acid, a compound that is not recognized by either parental enzyme. These results highlight the power of combinatorial homology-independent and low-homology recombination methods for the generation of unique, highly active enzymes and also suggest a possible means of enzyme "humanization."

Predicting enzyme function from protein sequence

There are two main reasons to try to predict an enzyme's function from its sequence. The first is to identify the components and thus the functional capabilities of an organism, the second is to create enzymes with specific properties. Genomics, expression analysis, proteomics and metabonomics are largely directed towards understanding how information flows from DNA sequence to protein functions within an organism. This review focuses on information flow in the opposite direction: the applicability of what is being learned from natural enzymes to improve methods for catalyst design.

Recombination during in vitro evolution

Recombination, the swapping of large portions of genetic information between and among parental genotypes, can be applied to in vitro evolution experiments on functional nucleic acids. Both homologous and heterologous recombination can be achieved using standard laboratory techniques. In many cases, recombination can allow for the discovery of a ribozyme or DNAzyme phenotype that would not likely be encountered by reliance on point mutations alone. In addition, recombination can often aid in the discovery of global optima in sequence space and/or lessen the number of generations it would take to reach optima. Recombination is most efficiently used in combination with point mutations and applied after the first couple of rounds of selection but before high-fitness genotypes dominate the selection. The "recombination zone" describes that region of sequence space-defined by the residues that will ultimately participate in the function of the winning nucleic acid(s)-where recombination is expected to be the most beneficial in the search for high-fitness genotypes.

DNA shuffling and screening strategies for improving vaccine efficacy

The efficacy of vaccines can be improved by increasing their immunogenicity, broadening their crossprotective range, as well as by developing immunomodulators that can be coadministered with the vaccine antigen. One technology that can be applied to each of these aspects of vaccine development is MolecularBreeding directed molecular evolution. Essentially, this technology is used to evolve genes in vitro through an iterative process consisting of recombinant generation followed by selection of the desired recombinants. We have used DNA shuffling and screening strategies to develop and improve vaccine candidates against several infectious pathogens including Plasmodium falciparum (a common cause of severe and fatal human malaria), dengue virus, encephalitic alphaviruses such as Venezuelan, western and eastern equine encephalitis viruses (VEEV, WEEV, and EEEV, respectively), human immunodeficiency virus-1 (HIV-1), and hepatitis B virus (HBV). By recombining antigen-encoding genes from different serovar isolates, new chimeras are selected for crossreactivity; these vaccine candidates are expected to provide broader crossprotection than vaccines based on a single serovar. Furthermore, the vaccine candidates can be selected for improved immunogenicity, which would also improve their efficacy. In addition to vaccine candidates, we have applied the technology to evolve several immunomodulators that when coadministered with vaccines can improve vaccine efficacy by fine-tuning the T cell response. Thus, DNA shuffling and screening technology is a promising strategy to facilitate vaccine efficacy.

Chemical and biochemical strategies for the randomization of protein encoding DNA sequences: library construction methods for directed evolution

Directed molecular evolution and combinatorial methodologies are playing an increasingly important role in the field of protein engineering. The general approach of generating a library of partially randomized genes, expressing the gene library to generate the proteins the library encodes and then screening the proteins for improved or modified characteristics has successfully been applied in the areas of protein-ligand binding, improving protein stability and modifying enzyme selectivity. A wide range of techniques are now available for generating gene libraries with different characteristics. This review will discuss these different methodologies, their accessibility and applicability to non-expert laboratories and the characteristics of the libraries they produce. The aim is to provide an up to date resource to allow groups interested in using directed evolution to identify the most appropriate methods for their purposes and to guide those moving on from initial experiments to more ambitious targets in the selection of library construction techniques. References are provided to original methodology papers and other recent examples from the primary literature that provide details of experimental methods.

DNA shuffling: induced molecular breeding to produce new generation long-lasting vaccines

The paradigm for classic vaccines has been to mimic natural infection, and their success relies mostly on the induction of neutralizing antibodies followed by long-lasting immunity. The outcome of aggressive chronic infections such as HIV and HCV, the reappearance of fastidious diseases such as tuberculosis and the progression of cancer growth suggest that natural immune responses are definitely insufficient in many cases. A new paradigm is needed to design and develop a new high-efficiency generation of vaccines ideally able to surpass the capabilities of natural immune responses. In vitro evolution is a new, important laboratory method to evolve molecules with desired properties, which appears as an appealing alternative to achieve this goal. In its battle against disease, the vertebrate immune system triggers a series of well-known molecular events in order to produce protective neutralizing antibodies. This natural in vivo response shares remarkable similarities with the in vitro technique known as molecular breeding or "DNA shuffling." This method exploits the recombination between genes to dramatically accelerate the rate at which genes can be evolved under selection pressure in the laboratory, producing optimized high-efficiency mutant proteins. Since new generation vaccines are aimed to overcome natural selection and environmental pressures to fully inactivate rapidly developing pathogen variants, they could be engineered, developed and selected through the application of directed DNA shuffling procedures. This review highlights the potential of the procedure in the complex context of natural immune responses and the equilibrium and interaction existing in nature between hosts and pathogens.

Enhanced crossover SCRATCHY: construction and high-throughput screening of a combinatorial library containing multiple non-homologous crossovers

SCRATCHY is a methodology for the construction of libraries of chimeras between genes that display low sequence homology. We have developed a strategy for library creation termed enhanced crossover SCRATCHY, that significantly increases the number of clones containing multiple crossovers. Complementary chimeric gene libraries generated by incremental truncation (ITCHY) of two distinct parental sequences are created, and are then divided into arbitrarily defined sections. The respective sections are amplified by skewed sets of primers (i.e. a combination of gene A specific forward primer and gene B specific reverse primer, etc.) allowing DNA fragments containing non-homologous crossover points to be amplified. The amplified chimeric sections are then subjected to a DNA shuffling process generating an enhanced crossover SCRATCHY library. We have constructed such a library using the rat theta 2 glutathione transferase (rGSTT2) and the human theta 1 glutathione transferase (hGSTT1) genes (63% DNA sequence identity). DNA sequencing analysis of unselected library members revealed a greater diversity than that obtained by canonical family shuffling or with conventional SCRATCHY. Expression and high-throughput flow cytometric screening of the chimeric GST library identified several chimeric progeny that retained rat-like parental substrate specificity.

Synthetic gene libraries: in search of the optimal diversity

Directed evolution has proven to be an effective method for evolving proteins with desired properties. A key step is the creation of suitably diverse gene libraries. Two new methods for creating such libraries make sole use of synthesized oligonucleotides and allow researchers to tailor the diversity of a library with greater precision and create libraries with greater diversity than was previously possible. Such increased diversity appears to accelerate directed evolution.

Identifying residue-residue clashes in protein hybrids by using a second-order mean-field approach

In this article, a second-order mean-field-based approach is introduced for characterizing the complete set of residue-residue couplings consistent with a given protein structure. This information is subsequently used to classify protein hybrids with respect to their potential to be functional based on the presenceabsence and severity of clashing residue-residue interactions. First, atomistic representations of both the native and denatured states are used to calculate rotamer-backbone, rotamer-intrinsic, and rotamer-rotamer conformational energies. Next, this complete conformational energy table is coupled with a second-order mean-field description to elucidate the probabilities of all possible rotamer-rotamer combinations in a minimum Helmholtz free-energy ensemble. Computational results for the dihydrofolate reductase family reveal correlation in substitution patterns between not only contacting but also distal second-order structural elements. Residue-residue clashes in hybrid proteins are quantified by contrasting the ensemble probabilities of protein hybrids against the ones of the original parental sequences. Good agreement with experimental data is demonstrated by superimposing these clashes against the functional crossover profiles of bidirectional incremental truncation libraries for Escherichia coli and human glycinamide ribonucleotide transformylases.

Computational and experimental analysis of DNA shuffling

We describe a computational model of DNA shuffling based on the thermodynamics and kinetics of this process. The model independently tracks a representative ensemble of DNA molecules and records their states at every stage of a shuffling reaction. These data can subsequently be analyzed to yield information on any relevant metric, including reassembly efficiency, crossover number, type and distribution, and DNA sequence length distributions. The predictive ability of the model was validated by comparison to three independent sets of experimental data, and analysis of the simulation results led to several unique insights into the DNA shuffling process. We examine a tradeoff between crossover frequency and reassembly efficiency and illustrate the effects of experimental parameters on this relationship. Furthermore, we discuss conditions that promote the formation of useless "junk" DNA sequences or multimeric sequences containing multiple copies of the reassembled product. This model will therefore aid in the design of optimal shuffling reaction conditions.

Synthetic shuffling expands functional protein diversity by allowing amino acids to recombine independently

We describe synthetic shuffling, an evolutionary protein engineering technology in which every amino acid from a set of parents is allowed to recombine independently of every other amino acid. With the use of degenerate oligonucleotides, synthetic shuffling provides a direct route from database sequence information to functional libraries. Physical starting genes are unnecessary, and additional design criteria such as optimal codon usage or known beneficial mutations can also be incorporated. We performed synthetic shuffling of 15 subtilisin genes and obtained active and highly chimeric enzymes with desirable combinations of properties that we did not obtain by other directed-evolution methods.

Growth factor engineering by degenerate homoduplex gene family recombination

There is great interest in engineering human growth factors as potential therapeutic agonists and antagonists. We approached this goal with a synthetic DNA recombination method. We aligned a pool of "top-strand" oligonucleotides incorporating polymorphisms from mammalian genes encoding epidermal growth factor (EGF) using multiple polymorphic "scaffold" oligonucleotides. Top strands were then linked by gap filling and ligation. This approach avoided heteroduplex annealing in the linkage of highly degenerate oligonucleotides and thus achieved completely random recombination. Cloned genes from a human-mouse chimeric library captured every possible permutation of the parental polymorphisms, creating an apparently complete recombined gene-family library, which has not been previously described. This library yielded a chimeric protein whose agonist activity was enhanced 123-fold. A second library from five mammalian EGF homologs possessed the highest reported recombination density (1 crossover per 12.4 bp). The five-homolog library yielded the strongest-binding hEGF variant yet reported. In addition, it contained strongly binding EGF variants with antagonist properties. Our less biased approach to DNA shuffling should be useful for the engineering of a wide variety of proteins.

Structure-based combinatorial protein engineering (SCOPE)

Presented here is the development a semi-rational protein engineering approach that uses information from protein structure coupled with established DNA manipulation techniques to design and create multiple crossover libraries from non-homologous genes. The utility of structure-based combinatorial protein engineering (SCOPE) was demonstrated by its application to two distantly related members of the X-family of DNA polymerases: rat DNA polymerase beta (Pol beta) and African swine fever virus DNA polymerase X (Pol X). These proteins share similar folds but have low sequence identity, and differ greatly in both size and activity. "Equivalent" subdomain elements of structure were designed on the basis of the tertiary structure of Pol beta and the corresponding regions of Pol X were inferred from homology modeling and sequence alignment analysis. Libraries of chimeric genes with up to five crossovers were synthesized in a series of PCR reactions by employing hybrid oligonucleotides that code for variable connections between structural elements. Genetic complementation in Escherichia coli enabled identification of several novel DNA polymerases with enhanced phenotypes. Both the composition of structural elements and the manner in which they were linked were shown to be essential for this property, indicating the importance of these aspects of design.

eCodonOpt: a systematic computational framework for optimizing codon usage in directed evolution experiments

We present a systematic computational framework, eCodonOpt, for designing parental DNA sequences for directed evolution experiments through codon usage optimization. Given a set of homologous parental proteins to be recombined at the DNA level, the optimal DNA sequences encoding these proteins are sought for a given diversity objective. We find that the free energy of annealing between the recombining DNA sequences is a much better descriptor of the extent of crossover formation than sequence identity. Three different diversity targets are investigated for the DNA shuffling protocol to showcase the utility of the eCodonOpt framework: (i) maximizing the average number of crossovers per recombined sequence; (ii) minimizing bias in family DNA shuffling so that each of the parental sequence pair contributes a similar number of crossovers to the library; and (iii) maximizing the relative frequency of crossovers in specific structural regions. Each one of these design challenges is formulated as a constrained optimization problem that utilizes 0-1 binary variables as on/off switches to model the selection of different codon choices for each residue position. Computational results suggest that many-fold improvements in the crossover frequency, location and specificity are possible, providing valuable insights for the engineering of directed evolution protocols.

Analysis of shuffled gene libraries

In vitro recombination of homologous genes (family shuffling) has been proposed as an effective search strategy for laboratory evolution of genes and proteins. Few data are available, however, on the composition of shuffled gene libraries, from which one could assess the efficiency of recombination and optimize protocols. Here, probe hybridization is used in a macroarray format to analyze chimeric DNA libraries created by DNA shuffling. Characterization of hundreds of shuffled genes encoding dioxygenases has elucidated important biases in the shuffling reaction. As expected, crossovers are favored in regions of high sequence identity. A sequence-based model of homologous recombination that captures this observed bias was formulated using the experimental results. The chimeric genes were found to show biases in the incorporation of sequences from certain parents, even before selection. Statistically different patterns of parental incorporation in genes expressing functional proteins can help to identify key sequence-function relationships.

Creating multiple-crossover DNA libraries independent of sequence identity

We have developed, experimentally implemented, and modeled in silico a methodology named SCRATCHY that enables the combinatorial engineering of target proteins, independent of sequence identity. The approach combines two methods for recombining genes: incremental truncation for the creation of hybrid enzymes and DNA shuffling. First, incremental truncation for the creation of hybrid enzymes is used to create a comprehensive set of fusions between fragments of genes in a DNA homology-independent fashion. This artificial family is then subjected to a DNA-shuffling step to augment the number of crossovers. SCRATCHY libraries were created from the glycinamide-ribonucleotide formyltransferase (GART) genes from Escherichia coli (purN) and human (hGART). The developed modeling framework eSCRATCHY provides insight into the effect of sequence identity and fragmentation length on crossover statistics and draws contrast with DNA shuffling. Sequence analysis of the naive shuffled library identified members with up to three crossovers, and modeling predictions are in good agreement with the experimental findings. Subsequent in vivo selection in an auxotrophic E. coli host yielded functional hybrid enzymes containing multiple crossovers.