M00192 865.87

Discovery of Superior Enzymes by Directed directed evolution ´ enzymes ´ gene technology ´ molecular evolution ´ mutagenesis 1. IntroductionNatural selection has created optimal catalysts that exhibit their identify individuals showing the desired properties, either by convincing performance even with a number of sometimes selection or by screening (Figure 1). After each round, the genes counteracting constraints. Optimal performance of enzyme of improved variants are deciphered and subsequently serve as catalysis does not refer necessarily to maximum reaction rate.
parents for another round of optimization. This review covers the Rather, it may involve a compromise between specificity, rate, most important aspects of directed evolution and summarizes stability, and other chemical constraints; in some cases, it may key solutions to problems of optimizing and understanding involve ªintelligentº control of rate and specificity.[1] Because enzymes are capable of catalyzing reactions under mildconditions and with high substrate specificity that often is accompanied by high regio- and enantioselectivity, it is notsurprising that a continually increasing number of industrial and The route of evolutionary optimization of a certain enzyme can academic reports concern the use of enzyme catalysts in be described as an adaptive walk in a ªfitness landscapeº that chemical synthesis as well as in biochemical and biomedical consists of peaks (sequences with high fitness) which are connected by ridges and separated by saddles, valleys (sequen- However, the demands of modern synthesis and their ces with low fitness), or planes. The fitness landscape is commercial application were obviously not targeted during associated to a ªsequence spaceºÐa network-like arrangement the natural evolution of enzymes. Considering a specific, non- of all possible amino acid sequences of a given length.[13, 14] natural application, any property (or combination of properties) Under the influence of mutation and selection, the enzyme may of an enzyme may therefore need to be improved. Of course, walk along the ridges toward peaks, that is, sequences of higher scientists desired to mimick nature's powerful concepts for fitness.[15] However, exhaustive searching of all possible protein tailoring specific enzymatic properties:[2] Following pioneering sequences for the individual variant with maximum fitness experiments for evolving molecules in the test tube,[3±6] evolu- seems like a daunting undertaking because the sequence space tionary engineering of biomolecules was successfully realized even for a protein of moderate sequence length of 100 amino with first selections of functional nucleic acids (ribozymes) by acids is extraordinarily large (about 10130 sequences). Taking using the SELEX (systematic evolution of ligands by exponential again nature as an example, optimal solutions can be found by enrichment with integrated optimization by non-linear analysis) exploring only small fractions of the sequence space: A series of procedure,[7, 8] and with the development of high-affinity ligands experimental strategies have been developed for generating (aptamers) by using similar techniques. Meanwhile, evolutionary mutant libraries in the laboratory which differ in diversity, that is, engineering, also termed ªdirected evolutionº, has emerged as a in the extent of covered sequence space, and in approaching key technology for biomolecular engineering and generated intelligent solutions for dealing with complexity.
impressive results in the functional adaptation of enzymes toartificial environments.[9±11] Certainly, evolution in the laboratorydoes not come to a halt at the optimization of single genes and proteins. Recent results excitingly demonstrate the success of Random mutagenesis at the nucleotide level is a widespread ªmolecular breedingº of metabolic pathways, and even of strategy which targets whole genes. This may be achieved by complete genomes,[12] and the end is not in sight yet.
passing cloned genes through mutator strains,[16, 17] by treating Directed evolution in the laboratory is highly attractive because its principles are simple and do not require detailedknowledge of structure, function, or mechanism. Essentially like natural evolution, directed evolution comprises the iterative Max Planck Institute for Biophysical Chemistry implementation of (1) the generation of a ªlibraryº of mutated genes, (2) its functional expression, and (3) a sensitive assay to CHEMBIOCHEM 2001, 2, 865 ± 871  WILEY-VCH-Verlag GmbH, D-69451 Weinheim, 2001 1439-4227/01/02/12 $ 17.50+.50/0 extremely rare. Codon-level random mutagenesis ofcomplete genes therefore would be desirable, but hasnot been realized yet.
Much effort, however, has been devoted to targeting single amino acids or selected regions of a proteinwhich have been identified in previous experiments tobe important for a certain function. By focusing only onthe positions of interest and/or their close environment,also the size of a mutant library can be drasticallyreduced (so-called ªdoped librariesº).[23] Methods forrandomizing small gene fragments are among theearliest techniques applied to in vitro evolution.
Typically, they employ the substitution of wild-typegene fragments by synthetic oligonucleotides whichcontain random positions or stretches (random cas-settes),[5, 24] or semi-random ranges (spiked oligonu-cleotides).[25] Randomization of defined positions orregions is achieved with automatic solid-phase DNAsynthesis by programming the desired InternationalUnion of Biochemistry (IUB) mix codes. The introduc-tion of stop codons can be excluded by allowing only Gand C (mix code: S) at the third position of each codon.
Complete permutation of a single amino acid position(saturation mutagenesis) may thus enable the findingof nonconservative replacements which are inaccessi-ble by random point mutagenesis.[26] Meanwhile,automatic solid-phase DNA synthesis also allows forthe selective introduction of codon mixtures by usingtrinucleotide b-cyanoethyl phosphoramidites,[27] evenby combining conventional dimethoxytrityl (DMT)protection with 9-fluorenylmethoxycarbonyl (Fmoc)chemistry.[28] Recombination of DNA represents an alternative or Figure 1. A) Schematic representation of directed enzyme evolution by in vivo selection.
additional approach for generating genetic diversity Target genes are mutated and inserted into a plasmid vector to yield a mutant gene library.
that is based on the mixing and concatenation of After transformation of a suitable host (which is, for example, auxotrophic with respect to the genetic material from a number of parent sequences.
target function), selective conditions are applied to the culture. Those cells which express active target gene variants that perform their activity under the applied constraints can As compared to random mutagenesis, recombination survive, whereas other cells harboring inactive gene products die out. B) Schematic may be advantageous in concentrating beneficial representation of directed evolution by screening. Bacteria harboring mutant target genes mutations which have arisen independently and may are plated and subsequently individualized, for example, in a microwell array. The target be additive, and likewise, in concentrating deleterious reaction is started with the addition of such substrates that facilitate the detection of successfully performed reactions, for example, chromogenic substrates. Desired variants are mutations which subsequently might be more effi- ciently purged from the population by selection.[29, 30]DNA shuffling was the first technique introduced forrandom in vitro recombination of gene variants created single-stranded DNA with various chemical mutagens,[18±20] or by by random mutagenesis.[31] This method employs the PCR error-prone PCR.[21, 22] Due to its simplicity and versatility, random reassembly of whole genes from a pool of short overlapping PCR mutagenesis emerged as the most common technique DNA sequences (typical length: 100 ± 300 bp) which are gen- which can result in mutation frequencies as high as 2% per erated by random enzymatic fragmentation of different parental nucleotide position. With alterations of some PCR conditions, the genes. Alternative protocols include StEP (staggered extension mutation rate may also be adjusted to lower values. However, process),[32] and random-priming recombination.[33] In vitro the number of amino acid substitutions accessible by using recombination by StEP is forced in a PCR-like reaction with very error-prone PCR is limited because this reaction biases the short annealing and extension steps that promote the formation distribution of mutation type in favor of transitions (A > G, T> of premature extension products. In following cycles, the C) and because multiple substitutions within a single codon are truncated strands may anneal randomly to a parent strand, thus combining the information of different parent strands. As an Irrespective of whether a protein library is expressed in a alternative to DNA shuffling, random-priming recombination recombinant host or displayed on bacteriophage, the available produces random fragments for reassembly by annealing of protein diversity is limited by the transformation or transfection short, random primers to a certain template gene and extension efficiencies of bacterial, or eukaryotic cells. Furthermore, the expression of nonhomologous or even toxic proteins may There is no reason that the concept of in vitro recombination severely interfere with some host environments. Thus, selection should be limited to pools of gene variants generated by may enrich only those cells which survive by reducing or random mutagenesis. In an extension of the idea, naturally circumventing the expression of the specific protein, or by occuring genes showing high similarity in sequence and preventing the correct protein folding. Cell-free transcription ± function can serve as an enormous pool of ªinformationº for translation systems that were recently developed may provide the creation of new, chimeric enzymes. Recombination of the basis for protein evolution in the absence of cells: They homologous parent genes, also called ªfamily shufflingº,[34] establish a physical genotype ± phenotype linkage in vitro either could access yet unexplored regions of the sequence space by stabilizing the mutual attachment of correctly folded because it combines genetic variability that has already been complete protein and its encoding mRNA to the ribosome selected in nature to be functional.
(called ribosome display),[39] by generating covalent fusions Homologous recombination may also be achieved in vivo.
between a peptide or protein and its mRNA,[40] or by distribution Most common are methods based on the transformation of of a library-based transcription ± translation system within Saccharomyces cerevisiae with a linearized plasmid and target aqueous droplets in a water-in-oil emulsion.[41] gene variants. Intermolecular homology-dependent recombina-tion may occur, which yields a circular plasmid that can be detected by using a selection marker.[35] Another concept for exploiting natural sources is called Functional protein libraries can be created rather easily by using modular protein design. This idea emphasizes the predominance one of the strategies described above. Therefore, the most of a limited number of elementary secondary structure units challenging step in directed evolution experiments is to develop which could be adopted by protein sequences having a low a screening or selection scheme that is sensitive to the degree of similarity. The permutation of protein modules properties of interest. Selection can be used in vivo if a requires nonhomologous recombination for generating func- substantial growth advantage is conferred to those clones that tional diversity rather than sequence diversity.[36] harbor a protein variant with the desired improvement. Mostoften, this is achieved by genetic complementation of hosts thatare deficient in a certain pathway or activity. In other cases, the positive feedback coupling between a property of interest and cell survival may be achieved with alteration of genetic contexts, for example, employing enzyme-specific control elements liketranscriptional promoters.[42] Selective enrichment of only those The classic evolution experiments with RNA and DNA were clones that express the particular enzyme function can be very successful in vitro because nucleic acids represent both function efficient. It should be noticed, however, that in vivo selection (phenotype) and genetic information (genotype). The directed systems usually represent highly specific solutions and often are evolution of enzymes, however, differs insomuch as diversity is difficult to implement because microbial hosts are extremely created on the DNA level, but selection or screening act on the flexible in circumventing the applied constraints and in invent- level of encoded protein. Therefore, functional expression of the ing solutions that are not necessarily related to the targeted information-carrying DNA libraries, whether generated by ran- dom mutagenesis or by recombination, is a necessary prereq- In vitro enrichment procedures that are detached from cell uisite for the detection of improved enzyme variants. The most survival may also be termed selection. Originally, these tech- common approaches for recombinant protein expression em- niques have been developed for the ªbiopanningº of phage- ploy the cellular trancription and translation machineries of well- displayed peptide libraries by binding to a ligand that is established organisms such as Escherichia coli, Saccharomyces immobilized on an appropriate column matrix. Recently, the cerevisiae, or Bacillus subtilis. These cellular expression systems approach has also been applied to the selective enrichment of also guarantee the association of a specific protein variant and phage-displayed functional enzyme libraries. Therefore, the idea its encoding gene. This is essential for the identification and of panning optimal binding partners needs to be extended to a amplification of desired mutants after selection or screening, as panning of optimal catalysts by using either transition state well as for further cycles of evolution. Alternatively, a physical analogues,[43] immobilized suicide substrates,[44] or reaction link between genotype and phenotype may be established by substrates that are covalently linked to the same or another generating fusions between the protein of interest and a phage via a second fusion.[45±47] However, the enrichment of bacteriophage coat protein. Following intracellular assembly, improved enzymes by biopanning remains challenging because recombinant phages express the protein variants on their the assessment of phage-displayed enzymes on the basis of their surface while enclosing the appertaining genetic information catalytic activity, that is, on the basis of their kinetic parameters, 4. Successful application of directed evolution Screening is an important alternative to selection which requires During the past few years, many enzymes have been improved that the specific property is directly observable by using physical by directed evolution (Table 2). Some of these new biocatalysts or biochemical analysis. As compared with in vivo selection, the are tuned for use in organic synthesis, and commercial screening approach enables a better control of the applied applications of some other enzymes are already in sight.
constraints, and also is more versatile, predominantly in Enzymes that exhibit increased activity in aqueous-organic unnatural environments, or with unnatural substrates. The solvents were among the first products of directed-evolution number of individual mutants that can be tested in a certain experiments.[52, 53] These biocatalysts enable the performance of period of time (throughput), however, may be lowerÐdepend- reactions at increased substrate solubility and stability, and thus ing on the enzymatic reaction and the sensitivity of the applied effect altered reaction equilibria, higher reaction rates, and higher product yields. The directed evolution of a large number The comparative assessment of individual mutants usually of enzymes that exhibit increased performance at elevated requires that the mutant libraries are diluted and distributed.
temperatures has been driven by a similar motiva- This can either be achieved by conventional plating of trans- tion.[26, 32, 54±58, 60±62] Furthermore, increased thermostability may formants on agar plates or filter membranes, or by distribution of be beneficial regarding the long-term stability of proteins at the mutant pool in a microtiter format (usually 96- or 384-well plates, but also formats with higher sample density, including Directed evolution has also been employed to improve the silicon wafers). This time-consuming step is sometimes accel- expression and folding of recombinant eukaryotic enzymes erated by using robotic systems. Common assays are based on which fail to adopt their active conformation in a heterologous visual or spectroscopic detection, for example formation, alter- host, or which are expressed in an artificial context, for example, ation, or destruction of colors or fluorescence characteristics. The in the form of a fusion protein.[63±66] Likewise, the secretion of determination of the optical parameters can also be accom- correctly folded enzymes has been facilitated by using evolu- plished by using automatic plate-reading systems, which enable tionary optimization.[67] In most of these cases, the increased a normalization of measured intensity values to the respective levels of expression and native folding have been attributed to cell densities and, furthermore, may be used for monitoring few point mutations within the structural genes.
reaction kinetics. There is an increasing tendency toward The narrow range of substrates that are accepted by natural automation and parallel processing by using decreasing sample enzymes often retards or prevents their use in new synthetic and volumes and concentrations, for example by applying the highly commercial applications. By far most results therefore reflect the sensitive fluorescence correlation spectroscopy (FCS) technique, efficient tuning of catalytic efficiency toward nonnatural sub- which requires concentrations in the femtomolar range.[49] strates.[34, 42, 43, 65, 68±88] Similarly, the enantioselectivity of specific Alternative approaches like confocal fluorescence coincidence bioconversions has been significantly improved by using evolu- analysis (CFCS)[50] or a fluorescence-activated cell sorter (FACS) tionary approaches.[89±91] Enzymes with altered substrate speci- can directly analyze single cells or proteins and, thus, gain ficity that yield yet inavailable products have also recently been (ultra)high throughput by avoiding the transfer of individuals.
generated by using the molecular breeding of new biosynthetic Disregarding whether a specific directed-evolution experi- pathways.[92] Together with similar attempts at mixing subunits ment employs a selection or screening approach (for a of multi-enzyme complexes,[93] this approach opens up the comparison, see Table 1), it should finally be emphasized that horizon toward new biologically active compounds.
it is important to choose selective constraints that precisely The conversion of a specific enzymatic activity into another reflect the desired property. Otherwise, ªyou get what you screen has very recently been achieved by using the methods of directed evolution, and by using a combination of rational and Table 1. Comparison of strategies for searching improved biocatalysts.
selection linkage between desired activity and cell survival false positives (viable but undesired mutants) screening individualization of mutant clones; in some cases, direct testing of each single clone for the multiple pipetting/washing/transfer steps isolation of mutant proteins from competitive cellular activitiesoften: need for fluorogenic or chromogenic sub- assays in nonnatural environments (artificial low throughput (ca. 105 individual mutants in qualitative and quantitative assay of one or Table 2. Examples of enzymes that were successfully optimized by using directed evolution.[a] suppressor tRNAs with diverse ribosomally (1) increased activity toward b-branched amino activity toward polychlorinated biphenyls (PCBs) degradation of various PCBs, polychlorinated benzene, thermostability at 658C without decrease in b-glucosidase CelB (Pyrococcus furiosus) increased catalytic activity at 208C retention of function after glutaraldehyde increased stability toward glutaraldehyde and inactivation of 20000-fold higher concentration of cefo- sixfold higher PRAI activity than wild-type enzyme PRAI activity while retaining ProFARI activity PRAI activity 3 ± 11  104 lower than wild-type; ProFARI increase in intrinsic peroxidase activity activity of monomeric and hexameric enzyme increased activity toward naphthalene in the enantioselectivity and increased activity conversion into L-hydantoinase, 5-fold activity 316-fold decrease in LD100 of transformed E. coli in the presence of AZT,[b] 11-fold decrease with ddC[b] hydrolysis of sterically hindered 3-hydroxy esters activity with an increase in enantiomeric excess from increased thermal and oxidative stability 174-fold thermostability, 100-fold oxidative stability 1000-fold increase in specificity towards p-nitrophenyl furanoside, 66-fold increase in specific activity activity toward a range of new substrates DNA-dependent DNA polymerase activity and resistance [81] increased activity toward ABTS[b] and guaiacol total, 40-fold; ABTS, 5.4-fold; guaiacol, 2.3-fold (B. subtilis)kanamycin nucleotidyl transferase (1) 200-fold increase in t1/2 at 60 ± 658C increased activity in the absence of cofactor increase in enantiomeric excess from 2% to 81% lipases (Staphylococcus hyicus, S. aureus) substrate specificity (phospholipids vs. short- 3-fold increase in activity toward long-chain pNB[b] esters [82] substrate specificity (activity on phospholipids) 11.6-fold increase in abolute phopholipase activity, 11.5-fold increase in phospholipase/lipase ratio 6-fold increase in yield of D-amino acids fusion proteinO6-alkylguanine alkyltransferase reduction of BG inhibitory concentration to 50% 2.7 ± 5.5-fold decrease in mutation frequency phytoene desaturase, lycopene cyclase new carotenoid pathway synthesis of 3,4,3',4'-tetradehydrolycopene and torulene 148C increase in Tm without any decrease in activity at increased activity in aqueous/organic solvents 50 ± 150-fold activity toward different pNB esters in 6-fold activity, while retaining 3.3-fold lower activity at 708C [106] altered specificity and increased activity 20 ± 45-fold increase in rate and a greater selectivity increased activity at decreased temperature 178C increase in Topt, increased activity at all temperatures [62] functional complementation of E. coli DNA retention of activity of active-site mutants 43-fold toward ganciclovir, 20-fold toward aciclovir [a] See also ref. [96]. [b] Abbreviations: ABTS ˆ 2,2'-azinobis-(3-ethylbenzothiazoline-6-sulfonic) acid; AZTˆ 3'-azido-3'-deoxythymidine; BG ˆ O6-benzylgua- nine; ddC ˆ dideoxycytidine; IGPS ˆ indoleglycerol phosphate synthase; pNB ˆ para-nitrobenzoate; PRAI ˆ phosphoribosyl anthranilate isomerase; ProFARI ˆ N'-[(5'-phosphoribosyl)formimino]-5-aminoimidazole-4-carboxamide ribonucleotide isomerase.
evolutionary design principles.[94, 95] These approaches essential- [10] A. Plückthun, C. Schaffitzel, J. Hanes, L. Jermutus, Adv. Prot. Chem. 2000, ly copy natural evolution by recruiting existing functional protein scaffolds and refitting them to new enzymes. Experi- [11] P. L. Wintrode, F. H. Arnold, Adv. Prot. Chem. 2000, 55, 161 ± 225.
[12] J. E. Ness, S. B. Del Cardayre, J. Minshull, W. P. C. Stemmer, Adv. Prot.
ments of this type may also unravel the evolutionary relations between enzymes that share some common properties.
[13] ªThe roles of mutation, inbreeding, crossbreeding, and selection in In conclusion, many impressive examples have demonstrated evolutionº: S. Wright in International Proceedings of the 6th International that directed evolution represents a powerful and reliable tool Congress in Genetics, Vol. 1 (Ed.: D. F. Jones), Ithaca, New York, 1932, for improving biocatalysts in reasonably short periods of time.
[14] P. K. Schuster, Biophys. Chem. 1996, 66, 75 ± 110.
The technique itself will continue to evolve and address many of [15] M. Eigen, R. Winkler-Oswatitsch, A. Dress, Proc. Natl. Acad. Sci. USA 1988, the present questions and limitations. Certainly, the future of evolutionary biotechnology will be exciting! [16] E. C. Cox, Annu. Rev. Genet. 1976, 10, 135 ± 156.
[17] A. Greener, M. Callaghan, B. Jerpseth, Methods Mol. Biol. 1996, 57, 375 ± [18] D. Shortle, D. Nathans, Proc. Natl. Acad. Sci. USA 1978, 75, 2170 ± 2174.
[1] M. Eigen, C. K. Biebricher, M. Gebinoga, W. C. Gardiner, Biochemistry [19] J. T. Kadonaja, J. R. Knowles, Nucleic Acids Res. 1985, 13, 1733 ± [2] M. Eigen, W. Gardiner, Pure Appl. Chem. 1984, 56, 967 ± 978.
[20] J. O. Deshler, Gen. Anal. Techn. Appl. 1992, 9, 103 ± 106.
[3] S. Spiegelman, I. Haruna, I. B. Holland, G. Beaudreau, D. Mills, Proc. Natl.
[21] D. W. Leung, E. Chen, D. V. Goeddel, Technique 1989, 1, 11 ± 15.
Acad. Sci. USA 1965, 54, 919 ± 927.
[22] R. C. Cadwell, G. F. Joyce, PCR Methods Appl. 1992, 2, 28 ± 33.
[4] C. K. Biebricher, L. E. Orgel, Proc. Natl. Acad. Sci. USA 1973, 70, 934 ± 938.
[23] D. Tomandl, A. Schober, A. Schwienhorst, J. Comput. Aided Mol. Des.
[5] M. S. Horwitz, L. A. Loeb, Proc. Natl. Acad. Sci. USA 1986, 83, 7405 ± 7409.
[6] A. R. Oliphant, K. Struhl, Methods Enzymol. 1987, 155, 558 ± 568.
[24] L. A. Loeb, Adv. Pharmacol. 1996, 35, 321 ± 347.
[7] D. Irvine, C. Tuerk, L. Gold, J. Mol. Biol. 1991, 222, 739 ± 761.
[25] L. J. Jensen, K. Andersen, A. Svendsen, T. Kretzschmar, Nucleic Acids Res.
[8] D. S. Wilson, J. W. Szostak, Annu. Rev. Biochem. 1999, 68, 611 ± 647.
[9] M. T. Reetz, K. E. Jaeger, Top. Curr. Chem. 1999, 200, 32 ± 57.
[26] K. Miyazaki, F. H. Arnold, J. Mol. Evol. 1999, 49, 716 ± 720.
[27] M. H. Lyttle, E. W. Napolitano, B. L. Calio, L. M. Kauvar, BioTechniques [73] M. Zaccolo, E. Gherardi, J. Mol. Biol. 1999, 285, 775 ± 783.
[74] Q. S. Li, U. Schwaneberg, P. Fischer, R. D. Schmid, Chem. Eur. J. 2000, 6, [28] P. Gaytan, J. Yanez, F. Sanchez, H. Mackie, X. Soberon, Chem. Biol. 1998, 5, [75] H. Joo, Z. Lin, F. H. Arnold, Nature 1999, 399, 670 ± 673.
[29] C. Zeyl, G. Bell, Nature 1997, 388, 465 ± 468.
[76] A. Iffland, P. Tafelmeyer, C. Saudan, K. Johnsson, Biochemistry 2000, 39, [30] G. L. Moore, C. D. Maranas, J. Theor. Biol. 2000, 205, 483 ± 503.
[31] W. P. C. Stemmer, Proc. Natl. Acad. Sci. USA 1994, 91, 10747 ± 10751.
[77] S. L. Washington, M. S. Yoon, A. M. Chagovetz, S. X. Li, C. A. Clairmont, [32] H. Zhao, L. Giver, Z. Shao, J. A. Affholter, F. H. Arnold, Nat. Biotechnol.
B. D. Preston, K. A. Eckert, J. B. Sweasy, Proc. Natl. Acad. Sci. USA 1997, 94, [33] Z. Shao, H. Zhao, L. Giver, F. H. Arnold, Nucleic Acids Res. 1998, 26, 681 ± [78] W. Knecht, B. Munch-Petersen, J. Piskur, J. Mol. Biol. 2000, 301, 827 ± 837.
[79] U. Bornscheuer, J. Altenbuchner, H. Meyer, Biotechnol. Bioeng. 1998, 58, [34] A. Crameri, S. A. Raillard, E. Bermudez, W. P. C. Stemmer, Nature 1998, [80] J. H. Zhang, G. Dawes, W. P. C. Stemmer, Proc. Natl. Acad. Sci. USA 1997, [35] T. L. Orr-Weaver, J. W. Szostak, Methods Enzymol. 1983, 101, 228 ± 245.
[36] M. A. Fuchs, C. Buta, Biophys. Chem. 1997, 66, 203 ± 210.
[81] B. Kim, T. R. Hathaway, L. A. Loeb, J. Biol. Chem. 1996, 271, 4872 ± 4878.
[37] K. Johnsson, L. Ge, Curr. Top. Microbiol. Immunol. 1999, 243, 87 ± 105.
[82] M. D. van Kampen, N. Dekker, M. R. Egmond, H. M. Verheij, Biochemistry [38] G. P. Smith, V. A. Petrenko, Chem. Rev. 1997, 97, 391 ± 410.
[39] J. Hanes, A. Plückthun, Proc. Natl. Acad. Sci. USA 1997, 94, 4937 ± 4942.
[83] T. Lanio, A. Jeltsch, A. Pingoud, J. Mol. Biol. 1998, 283, 59 ± 69.
[40] R. W. Roberts, J. W. Szostak, Proc. Natl. Acad. Sci. USA 1997, 94, 12297 ± [84] L. D. Graham, K. D. Haggett, P. A. Jennings, D. S. Le Brocque, R. G.
Whittaker, P. A. Schober, Biochemistry 1993, 32, 6250 ± 6258.
[41] D. S. Tawfik, A. D. Griffith, Nat. Biotechnol. 1998, 16, 652 ± 656.
[85] M. Suzuki, A. K. Avicola, L. Hood, L. A. Loeb, J. Biol. Chem. 1997, 272, [42] S. Brakmann, S. Grzeszik, ChemBioChem 2001, 2, 212 ± 219.
[43] L. O. Hansson, M. Widersten, B. Mannervik, Biochem. J. 1999, 344, 93 ± [86] M. E. Black, T. G. Newcomb, H. M. Wilson, L. A. Loeb, Proc. Natl. Acad. Sci.
[44] S. Vanwetswinkel, B. Avalle, J. Fastrez, J. Mol. Biol. 2000, 295, 527 ± 540.
[87] F. C. Christians, L. Scapozza, A. Crameri, G. Folkers, W. P. Stemmer, Nat.
[45] J. L. Jestin, P. Kristensen, G. Winter, Angew. Chem. 1999, 111, 1196 ± 1200; Angew. Chem. Int. Ed. 1999, 38, 1124 ± 1127.
[88] M. D. van Kampen, M. R. Egmond, Eur. J. Lipid Sci. Technol. 2000, 102, [46] H. Pedersen, S. Hölder, D. P. Sutherlin, U. Schwitter, D. S. King, P. G.
Schultz, Proc. Natl. Acad. Sci. USA 1998, 95, 10523 ± 10528.
[89] O. May, P. T. Nguyen, F. H. Arnold, Nat. Biotechnol. 2000, 18, 317 ± 320.
[47] S. Demartis, A. Huber, F. Viti, L. Lozzi, L. Giovannoni, P. Neri, G. Winter, D.
[90] E. Henke, U. T. Bornscheuer, Biol. Chem. 1999, 380, 1029 ± 1033.
Neri, J. Mol. Biol. 1999, 286, 617 ± 633.
[91] M. T. Reetz, A. Zonta, K. Schimossek, K. Liebeton, K.-E. Jäger, Angew.
[48] M. Olsen, B. Iverson, G. Georgiou, Curr. Opin. Biotechnol. 2000, 11, 331 ± Chem. 1997, 109, 2961 ± 2963; Angew. Chem. Int. Ed. Engl. 1997, 36, [49] S. Sterrer, K. Henco, J. Recept. Signal Transduction Res. 1997, 17, 511 ± 520.
[92] C. Schmidt-Dannert, D. Umeno, F. H. Arnold, Nat. Biotechnol. 2000, 18, [50] T. Winkler, U. Kettling, A. Koltermann, M. Eigen, Proc. Natl. Acad. Sci. USA [93] C. Khosla, R. S. Gokhale, J. R. Jacobsen, D. E. Cane, Annu. Rev. Biochem.
[51] C. Schmidt-Dannert, F. H. Arnold, Trends Biotechnol. 1999, 17, 135 ± 136.
[52] J. C. Moore, F. H. Arnold, Nat. Biotechnol. 1996, 14, 458 ± 467.
[94] M. M. Altamirano, J. M. Blackburn, C. Aguayo, A. R. Fersht, Nature 2000, [53] K. Chen, F. H. Arnold, Proc. Natl. Acad. Sci. USA 1993, 90, 5618 ± 5622.
[54] G. Gonzalez-Blasco, J. Sanz-Aparicio, B. Gonzalez, J. A. Hermoso, J.
[95] C. Jürgens, A. Strom, D. Wegener, S. Hettwer, M. Wilmanns, R. Sterner, Polaina, J. Biol. Chem. 2000, 275, 13 708 ± 13 712.
Proc. Natl. Acad. Sci. USA 2000, 97, 9925 ± 9930.
[55] J. H. Lebbink, T. Kaper, P. Bron, J. van der Oost, W. M. de Vos, Biochemistry [96] http://www.che.caltech.edu:80/groups/fha/Enzyme/directed.html.
[97] A. Crameri, G. Dawes, E. Rodriguez, S. Silver, W. P. Stemmer, Nat.
[56] M. Kikuchi, M. Delarue, S. Harayama, Gene 1999, 236, 159 ± 167.
[57] F. Buchholz, P. Angrand, A. Stewart, Nat. Biotechnol. 1998, 16, 657 ± 662.
[98] I. Matsumura, J. B. Wallingford, N. K. Surana, P. D. Vize, A. D. Ellington, [58] J. R. Cherry, M. H. Lamsa, P. Schneider, J. Vind, A. Svendsen, A. Jones, A. H.
Nat. Biotechnol. 1999, 17, 696 ± 701.
Pedersen, Nat. Biotechnol. 1999, 17, 379 ± 384.
[99] Y. S. Kim, H. C. Jung, J. G. Pan, Appl. Environm. Microbiol. 2000, 66, 788 ± [59] S. Akanuma, A. Yamagishi, N. Tanaka, T. Oshima, Prot. Sci. 1998, 7, 698 ± [100] T. Matsuura, T. Yomo, S. Trakulnaleamsai, Y. Ohashi, K. Yamamoto, I.
[60] H. Liao, T. McKenzie, R. Hageman, Proc. Natl. Acad. Sci. USA 1986, 83, Urabe, Protein Eng. 1998, 11, 789 ± 795.
[101] L. Wan, M. B. Twitchett, L. D. Eltis, A. G. Mauk, M. Smith, Proc. Natl. Acad.
[61] L. Giver, A. Gershenson, P. O. Freskgard, F. H. Arnold, Proc. Natl. Acad. Sci.
[102] J. Hoseki, T. Yano, Y. Koyama, S. Kuramitsu, H. Kagamiyama, J. Biochem.
[62] H. Zhao, F. H. Arnold, Protein Eng. 1999, 12, 47 ± 53.
[63] T. K. Cheong, P. J. Oriel, Enzyme Microb. Technol. 2000, 26, 152 ± 158.
[103] S. J. Allen, J. J. Holbrook, Protein Eng. 2000, 13, 5 ± 7.
[64] G. McBeath, P. Kast, D. Hilvert, Science 1998, 279, 1958 ± 1961.
[104] F. C. Christians, B. J. Dawson, M. M. Coates, L. A. Loeb, Cancer Res. 1997, [65] B. Morawski, Z. L. Lin, P. Cirino, H. Joo, G. Bandara, F. H. Arnold, Protein [105] F. C. Christians, L. A. Loeb, Proc. Natl. Acad. Sci. USA 1996, 93, 6124 ± 6128.
[66] G. J. Kim, Y. H. Cheon, H. S. Kim, Biotechnol. Bioeng. 2000, 68, 211 ± 217.
[106] P. L. Wintrode, K. Miyazaki, F. H. Arnold, J. Biol. Chem. 2000, 275, 31635 ± [67] D. Naki, C. Paech, G. Ganshaw, V. Schellenberger, Appl. Microbiol.
[107] J. E. Ness, M. Welch, L. Giver, M. Bueno, J. R. Cherry, T. V. Borchert, W. P. C.
[68] D. R. Liu, P. G. Schultz, Proc. Natl. Acad. Sci. USA 1999, 96, 4780 ± 4785.
Stemmer, J. Minshull, Nat. Biotechnol. 1999, 17, 893 ± 896.
[69] T. Yano, S. Oue, H. Kagamiyama, Proc. Natl. Acad. Sci. USA 1998, 95, [108] S. Taguchi, A. Ozaki, T. Nonaka, Y. Mitsui, H. Momose, J. Biochem. 1999, [70] S. Oue, A. Okamoto, T. Yano, H. Kagamiyama, J. Biol. Chem. 1999, 274, [109] M. Suzuki, D. Baskin, L. Hood, L. A. Loeb, Proc. Natl. Acad. Sci. USA 1996, [71] T. Kumamaru, H. Suenaga, M. Mitsuoka, T. Watanabe, K. Furukawa, Nat.
[72] W. P. C. Stemmer, Nature 1994, 370, 389 ± 391.

Source: http://www.biocomputing.it/cutruzzola/lezioni/BM2-2008/ART4-evol.pdf