Strategies and computational tools for improving randomized protein libraries

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.

Iterative saturation mutagenesis: a powerful approach to engineer proteins by systematically simulating Darwinian evolution

Iterative saturation mutagenesis (ISM) is a widely applicable and powerful strategy for the efficient directed evolution of enzymes. First, one or more amino acid positions from the chosen enzyme are assigned to multi-residue sites (i.e., groups of amino acids or "multisites"). Then, the residues in each multisite are mutated with a user-defined randomization scheme to all canonical amino acids or a reduced amino acid alphabet. Subsequently, the genes of chosen variants (usually the best but not necessarily) are used as templates for saturation mutagenesis at other multisites, and the process is repeated until the desired degree of biocatalyst improvement has been achieved. Addressing multisites iteratively results in a so-called ISM scheme or tree with various upward branches or pathways. The systematic character of ISM simulates in vitro the natural process of Darwinian evolution: variation (library creation), selection (library screening), and amplification (template chosen for the next round of randomization). However, the main feature of ISM that distinguishes it from other directed evolution methods is the systematic probing of a defined segment of the protein sequence space, as it has been shown that ISM is much more efficient in terms of biocatalyst optimization than random methods such as error-prone PCR. In addition, ISM trees have also shed light on the emergence of epistasis, thereby rationally improving the strategies for evolving better enzymes. ISM was developed to improve catalytic properties such as rate, substrate scope, stereo- and regioselectivity using the Combinatorial Active-site Saturation Test (CAST), as well as chemical and thermal stability employing the B-Factor Iterative Test (B-FIT). However, ISM can also be invoked to manipulate such protein properties as binding affinity among other possibilities, including protein-protein interactions. Herein, we provide general guidelines for ISM, using CAST as the case study in the quest to enhance the activity and regioselectivity of the monooxygenase P450BM3 toward testosterone.

Site-saturation mutagenesis by overlap extension PCR

Site-saturation mutagenesis is a proven strategy for generating high-quality variant gene libraries of a defined size. Variation is introduced via incorporation of degenerate base combinations at specific codon locations, giving rise to a precise series of amino acid substitutions in the encoded protein. Here we describe a simple and efficient overlap PCR protocol for the introduction of degenerate bases at either single or multiple codon locations. The resulting libraries can then be directly screened for improved protein function as either an independent directed evolution study or an adjunct to random mutagenesis strategies (such as error-prone PCR) that are, in isolation, unlikely to access the full repertoire of possible amino acid substitutions at any given position.

Biocatalysis in organic chemistry and biotechnology: past, present, and future

Enzymes as catalysts in synthetic organic chemistry gained importance in the latter half of the 20th century, but nevertheless suffered from two major limitations. First, many enzymes were not accessible in large enough quantities for practical applications. The advent of recombinant DNA technology changed this dramatically in the late 1970s. Second, many enzymes showed a narrow substrate scope, often poor stereo- and/or regioselectivity and/or insufficient stability under operating conditions. With the development of directed evolution beginning in the 1990s and continuing to the present day, all of these problems can be addressed and generally solved. The present Perspective focuses on these and other developments which have popularized enzymes as part of the toolkit of synthetic organic chemists and biotechnologists. Included is a discussion of the scope and limitation of cascade reactions using enzyme mixtures in vitro and of metabolic engineering of pathways in cells as factories for the production of simple compounds such as biofuels and complex natural products. Future trends and problems are also highlighted, as is the discussion concerning biocatalysis versus nonbiological catalysis in synthetic organic chemistry. This Perspective does not constitute a comprehensive review, and therefore the author apologizes to those researchers whose work is not specifically treated here.

Fitness loss and library size determination in saturation mutagenesis

Saturation mutagenesis is a widely used directed evolution technique, in which a large number of protein variants, each having random amino acids in certain predetermined positions, are screened in order to discover high-fitness variants among them. Several metrics for determining the library size (the number of variants screened) have been suggested in the literature, but none of them incorporates the actual fitness of the variants discovered in the experiment. We present the results of an extensive simulation study, which is based on probabilistic models for protein fitness landscape, and which investigates how the result of a saturation mutagenesis experiment – the fitness of the best variant discovered – varies as a function of the library size. In particular, we study the loss of fitness in the experiment: the difference between the fitness of the best variant discovered, and the fitness of the best variant in variant space. Our results are that the existing criteria for determining the library size are conservative, so smaller libraries are often satisfactory. Reducing the library size can save labor, time, and expenses in the laboratory.

Optimal codon randomization via mathematical programming

Codon randomization via degenerate oligonucleotides is a widely used approach for generating protein libraries. We use integer programming methodology to model and solve the problem of computing the minimal mixture of oligonucleotides required to induce an arbitrary target probability over the 20 standard amino acids. We consider both randomization via conventional degenerate oligonucleotides, which incorporate at each position of the randomized codon certain nucleotides in equal probabilities, and randomization via spiked oligonucleotides, which admit arbitrary nucleotide distribution at each of the codon's positions. Existing methods for computing such mixtures rely on various heuristics.

Library construction and evaluation for site saturation mutagenesis

We developed a method for creating and evaluating site-saturation libraries that consistently yields an average of 27.4±3.0 codons of the 32 possible within a pool of 95 transformants. This was verified by sequencing 95 members from 11 independent libraries within the gene encoding alkene reductase OYE 2.6 from Pichia stipitis. Correct PCR primer design as well as a variety of factors that increase transformation efficiency were critical contributors to the method's overall success. We also developed a quantitative analysis of library quality (Q-values) that defines library degeneracy. Q-values can be calculated from standard fluorescence sequencing data (capillary electropherograms) and the degeneracy predicted from an early stage of library construction (pooled plasmids from the initial transformation) closely matched that observed after ca. 1000 library members were sequenced. Based on this experience, we suggest that this analysis can be a useful guide when applying our optimized protocol to new systems, allowing one to focus only on good-quality libraries and reject substandard libraries at an early stage. This advantage is particularly important when lower-throughput screening techniques such as chiral-phase GC must be employed to identify protein variants with desirable properties, e.g., altered stereoselectivities or when multiple codons are targeted for simultaneous randomization.

Reducing codon redundancy and screening effort of combinatorial protein libraries created by saturation mutagenesis

Saturation mutagenesis probes define sections of the vast protein sequence space. However, even if randomization is limited this way, the combinatorial numbers problem is severe. Because diversity is created at the codon level, codon redundancy is a crucial factor determining the necessary effort for library screening. Additionally, due to the probabilistic nature of the sampling process, oversampling is required to ensure library completeness as well as a high probability to encounter all unique variants. Our trick employs a special mixture of three primers, creating a degeneracy of 22 unique codons coding for the 20 canonical amino acids. Therefore, codon redundancy and subsequent screening effort is significantly reduced, and a balanced distribution of codon per amino acid is achieved, as demonstrated exemplarily for a library of cyclohexanone monooxygenase. We show that this strategy is suitable for any saturation mutagenesis methodology to generate less-redundant libraries.

When second best is good enough: another probabilistic look at saturation mutagenesis

We developed new criteria for determining the library size in a saturation mutagenesis experiment. When the number of all possible distinct variants is large, any of the top-performing variants (e.g., any of the top three) is likely to meet the design requirements, so the probability that the library contains at least one of them is a sensible criterion for determining the library size. By using a criterion of this type, one may significantly reduce the library size and thus save costs and labor while minimally compromising the quality of the best variant discovered. We present the probabilistic tools underlying these criteria and use them to compare the efficiencies of four randomization schemes: NNN, which uses all 64 codons; NNB, which uses 48 codons; NNK, which uses 32 codons; and MAX, which assigns equal probabilities to each of the 20 amino acids. MAX was found to be the most efficient randomization scheme and NNN the least efficient. TopLib, a computer program for carrying out the related calculations, is available through a user-friendly Web server.

Site-saturation mutagenesis: a powerful tool for structure-based design of combinatorial mutation libraries

This unit describes a method for site-saturation mutagenesis (SSM) using PCR amplification with degenerate synthetic oligonucleotides as primers. SSM allows the substitution of predetermined protein sites against all twenty possible amino acids at once. Therefore, SSM is a powerful approach in protein engineering to characterize structure-function relationships, as well as to create improved protein variants. The procedure accepts double-stranded plasmid isolated from the dam(+) E. coli strain. The procedure is simple, fast, efficient, and eliminates time-consuming subcloning and ligation steps.

Iterative saturation mutagenesis accelerates laboratory evolution of enzyme stereoselectivity: rigorous comparison with traditional methods

Efficacy in laboratory evolution of enzymes is currently a pressing issue, making comparative studies of different methods and strategies mandatory. Recent reports indicate that iterative saturation mutagenesis (ISM) provides a means to accelerate directed evolution of stereoselectivity and thermostability, but statistically meaningful comparisons with other methods have not been documented to date. In the present study, the efficacy of ISM has been rigorously tested by applying it to the previously most systematically studied enzyme in directed evolution, the lipase from Pseudomonas aeruginosa as a catalyst in the stereoselective hydrolytic kinetic resolution of a chiral ester. Upon screening only 10,000 transformants, unprecedented enantioselectivity was achieved (E = 594). ISM proves to be considerably more efficient than all previous systematic efforts utilizing error-prone polymerase chain reaction at different mutation rates, saturation mutagenesis at hot spots, and/or DNA shuffling, pronounced positive epistatic effects being the underlying reason.

Increased ethanol productivity in xylose-utilizing Saccharomyces cerevisiae via a randomly mutagenized xylose reductase

Baker's yeast (Saccharomyces cerevisiae) has been genetically engineered to ferment the pentose sugar xylose present in lignocellulose biomass. One of the reactions controlling the rate of xylose utilization is catalyzed by xylose reductase (XR). In particular, the cofactor specificity of XR is not optimized with respect to the downstream pathway, and the reaction rate is insufficient for high xylose utilization in S. cerevisiae. The current study describes a novel approach to improve XR for ethanol production in S. cerevisiae. The cofactor binding region of XR was mutated by error-prone PCR, and the resulting library was expressed in S. cerevisiae. The S. cerevisiae library expressing the mutant XR was selected in sequential anaerobic batch cultivation. At the end of the selection process, a strain (TMB 3420) harboring the XR mutations N272D and P275Q was enriched from the library. The V(max) of the mutated enzyme was increased by an order of magnitude compared to that of the native enzyme, and the NADH/NADPH utilization ratio was increased significantly. The ethanol productivity from xylose in TMB 3420 was increased ∼40 times compared to that of the parent strain (0.32 g/g [dry weight {DW}] × h versus 0.007 g/g [DW] × h), and the anaerobic growth rate was increased from ∼0 h(-1) to 0.08 h(-1). The improved traits of TMB 3420 were readily transferred to the parent strain by reverse engineering of the mutated XR gene. Since integrative vectors were employed in the construction of the library, transfer of the improved phenotype does not require multicopy expression from episomal plasmids.

High-throughput metabolic engineering: advances in small-molecule screening and selection

Metabolic engineering for the overproduction of high-value small molecules is dependent upon techniques in directed evolution to improve production titers. The majority of small molecules targeted for overproduction are inconspicuous and cannot be readily obtained by screening. We provide a review on the development of high-throughput colorimetric, fluorescent, and growth-coupled screening techniques, enabling inconspicuous small-molecule detection. We first outline constraints on throughput imposed during the standard directed evolution workflow (library construction, transformation, and screening) and establish a screening and selection ladder on the basis of small-molecule assay throughput and sensitivity. An in-depth analysis of demonstrated screening and selection approaches for small-molecule detection is provided. Particular focus is placed on in vivo biosensor-based detection methods that reduce or eliminate in vitro assay manipulations and increase throughput. We conclude by providing our prospectus for the future, focusing on transcription factor-based detection systems as a natural microbial mode of small-molecule detection.

Ultrahigh-throughput FACS-based screening for directed enzyme evolution

Directed enzyme evolution has proven to be a powerful tool for improving a range of properties of enzymes through consecutive rounds of diversification and selection. However, its success depends heavily on the efficiency of the screening strategy employed. Fluorescence-activated cell sorting (FACS) has recently emerged as a powerful tool for screening enzyme libraries due to its high sensitivity and its ability to analyze as many as 10(8) mutants per day. Applications of FACS screening have allowed the isolation of enzyme variants with significantly improved activities, altered substrate specificities, or even novel functions. This review discusses FACS-based screening for enzymatic activity and its potential application for the directed evolution of enzymes, ribozymes, and catalytic antibodies.

"Fuzzy oil drop" model applied to individual small proteins built of 70 amino acids

The proteins composed of short polypeptides (about 70 amino acid residues) representing the following functional groups (according to PDB notation): growth hormones, serine protease inhibitors, antifreeze proteins, chaperones and proteins of unknown function, were selected for structural and functional analysis. Classification based on the distribution of hydrophobicity in terms of deficiency/excess as the measure of structural and functional specificity is presented. The experimentally observed distribution of hydrophobicity in the protein body is compared to the idealized one expressed by a three-dimensional Gauss function. The differences between these two distributions reveal the specificity of structural/functional characteristics of the protein. The residues of hydrophobicity deficiency versus the idealized distribution are assumed to indicate cavities with the potential to bind ligands, while the residues of hydrophobicity excess are interpreted as potentially participating in protein-protein complexation. The distribution of hydrophobicity irregularity seems to be specific for particular structures and functions of proteins. A comparative analysis of such profiles is carried out to identify the potential biological activity of proteins of unknown function.

Synonymous genes explore different evolutionary landscapes

The evolutionary potential of a gene is constrained not only by the amino acid sequence of its product, but by its DNA sequence as well. The topology of the genetic code is such that half of the amino acids exhibit synonymous codons that can reach different subsets of amino acids from each other through single mutation. Thus, synonymous DNA sequences should access different regions of the protein sequence space through a limited number of mutations, and this may deeply influence the evolution of natural proteins. Here, we demonstrate that this feature can be of value for manipulating protein evolvability. We designed an algorithm that, starting from an input gene, constructs a synonymous sequence that systematically includes the codons with the most different evolutionary perspectives; i.e., codons that maximize accessibility to amino acids previously unreachable from the template by point mutation. A synonymous version of a bacterial antibiotic resistance gene was computed and synthesized. When concurrently submitted to identical directed evolution protocols, both the wild type and the recoded sequence led to the isolation of specific, advantageous phenotypic variants. Simulations based on a mutation isolated only from the synthetic gene libraries were conducted to assess the impact of sub-functional selective constraints, such as codon usage, on natural adaptation. Our data demonstrate that rational design of synonymous synthetic genes stands as an affordable improvement to any directed evolution protocol. We show that using two synonymous DNA sequences improves the overall yield of the procedure by increasing the diversity of mutants generated. These results provide conclusive evidence that synonymous coding sequences do experience different areas of the corresponding protein adaptive landscape, and that a sequence's codon usage effectively constrains the evolution of the encoded protein.

Engineering a circularly permuted GFP scaffold for peptide presentation

The use of peptides as in vivo and in vitro ligand binding agents is hampered by the high flexibility, low stability and lack of intrinsic detection signal of peptide aptamers. Recent attempts to overcome these limitations included the integration of the binding peptide into a stable protein scaffold. In this paper, we present the optimization and testing of a circularly permuted variant of the green fluorescent protein (GFP). We examined the ability of the optimized scaffold to accept peptide insertions at three different regions. The three regions chosen are localized in close spatial proximity to each other and support different conformations of the inserted peptides. In all the three regions peptides with a biased, but still comprehensive, amino acid repertoire could be presented without disturbing the function of the optimized GFP-scaffold.

GLUE-IT and PEDEL-AA: new programmes for analyzing protein diversity in randomized libraries

There are many methods for introducing random mutations into nucleic acid sequences. Previously, we described a suite of programmes for estimating the completeness and diversity of randomized DNA libraries generated by a number of these protocols. Our programmes suggested some empirical guidelines for library design; however, no information was provided regarding library diversity at the protein (rather than DNA) level. We have now updated our web server, enabling analysis of translated libraries constructed by site-saturation mutagenesis and error-prone PCR (epPCR). We introduce GLUE-Including Translation (GLUE-IT), which finds the expected amino acid completeness of libraries in which up to six codons have been independently varied (according to any user-specified randomization scheme). We provide two tools for assisting with experimental design: CodonCalculator, for assessing amino acids corresponding to given randomized codons; and AA-Calculator, for finding degenerate codons that encode user-specified sets of amino acids. We also present PEDEL-AA, which calculates amino acid statistics for libraries generated by epPCR. Input includes the parent sequence, overall mutation rate, library size, indel rates and a nucleotide mutation matrix. Output includes amino acid completeness and diversity statistics, and the number and length distribution of sequences truncated by premature termination codons. The web interfaces are available at http://guinevere.otago.ac.nz/stats.html.

PCR-based strategy for construction of multi-site-saturation mutagenic expression library

There is an increasing demand for efficient and effective methods to engineer protein variants for industrial applications, structural biology and drug development. We describe a PCR-based strategy that produces multi-site-saturation mutagenic expression library using a circular plasmid carrying the wild-type gene. This restriction digestion- and ligation-independent method involves three steps: 1) synthesis of the degenerate oligonucleotide primers, 2) incorporation of the mutations through PCR, 3) transformation into the expression host. Our strategy is demonstrated through successful construction of an E. coli K12 malic enzyme expression library that contains members with simultaneous mutations on amino acid residues G311, D345 and G397. This method is in principle compatible with any circular vector that can be propagated with a dam(+)E. coli host to generate protein variant library with multiple changes, including mutation, short sequence deletion and insertion, or any mix of them.

Molecular diversity in engineered protein libraries

Engineered protein libraries, defined here as a collection of different mutant variants of a single specific protein, are intentionally designed to be rich in molecular diversity and can span ranges from as little as 400 different variants to greater than 10(12) members per library. The goal of engineering libraries is to generate new protein variants, identified upon screening, that possess desired novel properties. Exploitation of the natural organization of the genetic code has led to 'focused' libraries that are lower in overall complexity yet biased towards variants with preferred biophysical properties. An emerging trend, in which computational algorithms are blended with in vivo screens, is also leading towards greater and more rapid success in the field of protein design.

Incorporating Synthetic Oligonucleotides via Gene Reassembly (ISOR): a versatile tool for generating targeted libraries

The directed evolution of proteins has benefited greatly from site-specific methods of diversification such as saturation mutagenesis. These techniques target diversity to a number of chosen positions that are usually non-contiguous in the protein's primary structure. However, the number of targeted positions can be large, thus leading to impractically large library size, wherein almost all library variants are inactive and the likelihood of selecting desirable properties is extremely small. We describe a versatile combinatorial method for the partial diversification of large sets of residues. Our library oligonucleotides comprise randomized codons that are flanked by wild-type sequences. Adding these oligonucleotides to an assembly PCR of wild-type gene fragments incorporates the randomized cassettes, at their target sites, into the reassembled gene. Varying the oligonucleotides concentration resulted in library variants that carry a different average number of mutated positions that comprise a random subset of the entire set of diversified codons. This method, dubbed Incorporating Synthetic Oligos via Gene Reassembly (ISOR), was used to create libraries of a cytosine-C5 methyltransferase wherein 45 individual positions were randomized. One library, containing an average of 5.6 mutated residues per gene, was selected, and mutants with wild-type-like activities isolated. We also created libraries of serum paraoxonase PON1 harboring insertions and deletions (indels) in various areas surrounding the active site. Screening these libraries yielded a range of mutants with altered substrate specificities and indicated that certain regions of this enzyme have a surprisingly high tolerance to indels.

Iterative saturation mutagenesis (ISM) for rapid directed evolution of functional enzymes

Iterative saturation mutagenesis (ISM) is a new and efficient method for the directed evolution of functional enzymes. It reduces the necessary molecular biological work and the screening effort drastically. It is based on a Cartesian view of the protein structure, performing iterative cycles of saturation mutagenesis at rationally chosen sites in an enzyme, a given site being composed of one, two or three amino acid positions. The basis for choosing these sites depends on the nature of the catalytic property to be improved, e.g., enantioselectivity, substrate acceptance or thermostability. In the case of thermostability, sites showing highest B-factors (available from X-ray data) are chosen. The pronounced increase in thermostability of the lipase from Bacillus subtilis (Lip A) as a result of applying ISM is illustrated here.