Classifications

• • G—PHYSICS
  • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
  • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
  • G16B30/00—ICT specially adapted for sequence analysis involving nucleotides or amino acids
• • G—PHYSICS
  • G16—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR SPECIFIC APPLICATION FIELDS
  • G16B—BIOINFORMATICS, i.e. INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR GENETIC OR PROTEIN-RELATED DATA PROCESSING IN COMPUTATIONAL MOLECULAR BIOLOGY
  • G16B30/00—ICT specially adapted for sequence analysis involving nucleotides or amino acids
  • G16B30/10—Sequence alignment; Homology search

Definitions

• This invention is in the field of genetic algorithms and the application of genetic algorithms to nucleic acid shuffling methods.
• Recursive nucleic acid recombination provides for the rapid evolution of nucleic acids, in vitro and in vivo. This rapid evolution provides for the generation of encoded molecules (e.g., nucleic acids and proteins) with new and/or improved properties. Proteins and nucleic acids of industrial, agricultural and therapeutic importance can be created or improved through DNA shuffling procedures.
• a number of publications by the inventors and their co-workers describe DNA shuffling. For example, Stemmer et al. (1994) "Rapid Evolution of a Protein” Nature 370:389-391; Stemmer (1994) “DNA Shuffling by Random Fragmentation and Reassembly: in vitro Recombination for Molecular Evolution," Proc. Natl.
• DNA shuffling can also be found in various published applications, such as WO95/22625, WO97/ 20078, WO96/33207, WO97/33957, WO98/27230, WO97/35966, WO98/ 31837, WO98/13487, WO98/13485 and WO989/42832.
• the present invention relates, e.g., to methods of defining, selecting and making polypeptides and nucleic acids. For example, a variety of methods of selecting crossovers in sequence strings corresponding to either nucleic acids or polypeptides are provided. These methods include, e.g., consideration of structural, functional and/or statistical data for nucleic acids or polypeptides which correspond to the sequence strings.
• the invention provides methods of making recombinant nucleic acids and/or polypeptides.
• a plurality of parental character strings corresponding to a plurality of nucleic acids or polypeptides are provided. These character strings, when aligned for maximum similarity/ identity, include at least one region of heterology.
• the character strings are aligned and a plurality of cross-over sites are selected in the character strings.
• a set of character string subsequences is defined, which includes subsequences of at least two of the plurality of parental character strings.
• a set of oligonucleotides corresponding to the set of character string subsequences is provided which includes a plurality of bridging oligonucleotides which correspond to the plurality of cross-over sites.
• the set of oligonucleotides is annealed, and one or more members of the set of oligonucleotides is elongated with a polymerase, or by ligating at least two members of the set of oligonucleotides with a ligase, thereby producing one or more recombinant nucleic acid, which, optionally, encodes one or more polypeptide.
• the two or more parental sequences can display very low sequence similarity, at either the nucleic acid or polypeptide level. However, the two or more parental sequences can, of course, display higher levels of sequence similarity.
• the methods can also include further steps such as selection of sequences in silico (or by standard methods), determination of sequences in silico (or by standard sequencing methods, or both), or the like.
• the method optionally includes determining one or more sequence for one or more putative recombinant nucleic acid resulting from in silico recombination of the two or more parental sequences at the cross- over sites, and performing one or more in silico simulation of activity for the one or more putative recombinant nucleic acid.
• nucleic acids used for gene reassembly/ recombination can be synthetic oligonucleotides.
• fragments of nucleic acids corresponding to the sequences can be used.
• recombinant nucleic acids can be made by providing fragments of the two or more parental nucleic acids (or synthetic oligonucleotides, or both fragments and ' oligonucleotides) and a plurality of bridge oligonucleotides, hybridizing the fragments and the bridge oligonucleotides and elongating the hybridized fragments with a polymerase or a ligase.
• parameters can be used to select the cross-over site.
• such parameters include consideration of one or more features encoded by a polypeptide/polypeptide corresponding to the character string.
• Such features include one or more of, e.g., structural stability, an allosteric effect, a 3-D energetic constraint, structural symmetry, hydrophobicity, distribution of hydrophobic residues in 1, 2 or 3 dimensions, periodicity of amino acid residues, residue properties in 1, 2 or 3- dimensions, distribution of charged residues, co-variation of residues, evolutionary rates of change for local regions or entire polypeptides, codon usage, codon distribution, GC bias, synonymous substitution rates, non-synonymous substitution rates, motif distribution, orientation of primary structural units, orientation of secondary structural units, orientation of tertiary structural units, orientation of quaternary structural units, one or more 3-D motif, one or more structural pocket, one or more active site, one or more binding site, and/or one or more interaction strength between two or more secondary structural elements.
• Analysis of amino acid composition of an encoded polypeptide corresponding to the character string can also be performed to determine appropriate cross-over sites, including, e.g., analysis of hydrophobicity as determined by a ⁇ G of transfer from one or more solvents, an analysis of hydrophilicity as determined by column retention on one or more column, analysis of one or more charges on one or more amino acids of the encoded polypeptide, polarity of one or more amino acids on the encoded polypeptide, pka of on one or more amino acids of the encoded polypeptide, bulkiness of one or more amino acids of the encoded polypeptide, side chain entropy of one or more amino acids of the encoded polypeptide, one or more alpha helix/beta sheet propensities of the encoded polypeptide, or one or more hydration potentials of the encoded polypeptides.
• methods of producing recombinant nucleic acids or polypeptides include providing two or more parental sequences, selecting a plurality of cross-over sites for recombination between the two or more parental sequences (thereby defining one or more recombinant sequences that result from a cross-over between at least two of the two or more parental sequences), selecting a plurality of ' bridging oligonucleotides which correspond to the cross-over sites, predicting a recombinant sequence for at least one of the one or more recombinant sequences, selecting the at least one recombinant sequence in silico for one or more expected activity, and synthesizing the at least one recombinant sequence.
• Synthesizing the at least one recombinant sequence can include, e.g., providing nucleic acid fragments which at least partly correspond to the two or more parental sequences and at least one bridge oligonucleotide, hybridizing the fragments and the bridge oligonucleotide and elongating the hybridized fragments with a polymerase or a ligase.
• the two or more parental sequences can display low sequence similarity, given that bridging sequences are used in the elongation step.
• Selecting the at least one recombinant sequence in silico can include, e.g., performing an energy minimization analysis of the at least one recombinant sequence, performing a stability analysis of the at least one recombinant sequence, comparing an energy minimized model of the at least one recombinant sequence to an energy minimized model of one or more of the two or more parental sequences, performing polypeptide threading on one or more polypeptide corresponding to the parental or recombinant sequences, or selecting the cross-over sites for recombination between the two or more parental sequences to occur within regions of structural overlap, thereby determining the sequence of the at least one recombinant sequence.
• selection can include performing PDA, a branch-and-terminate, a combinatorial optimization analysis, a dead end elimination, a genetic or, e.g., mean-field analysis, or analysis of polypeptide folding by threading of the at least one recombinant sequence.
• PDA can be performed on at least one of the two or more parental sequences.
• a PDA of the at least one recombinant sequence can be compared to a PDA of at least one of the two or more parental sequences.
• any of the structural parameters for the embodiment noted above structural stability, allosteric effects, etc.
• the step of selecting cross-over sites for recombination between the two or more parental sequences and the step of selecting the at least one recombinant sequence in silico are optionally performed simultaneously.
• the invention provides methods of producing one or more recombinant nucleic acids or encoded polypeptides by providing a plurality of first nucleic acid or first polypeptide sequences, selecting a plurality of cross-over sequences between the plurality of first nucleic acid or first polypeptide sequences by defining structural, statistical, or logical criteria for the cross-over sequences in silico, and artificially synthesizing a plurality of recombinant nucleic acids comprising or encoding the cross-over sequences.
• the first nucleic acid or polypeptide sequences can comprise homologous or non-homologous sequences.
• structurally or functionally related sequences can be recombined, as can structurally or functionally complementary sequences.
• the sequences can be natural or artificial.
• polymerase or ligase-mediated assembly methods can be used to generate recombinant nucleic acids or polypeptides.
• artificially synthesizing a plurality of recombinant nucleic acids comprising or encoding the cross-over sequences can include synthesizing a plurality of oligonucleotides, one or more of which encodes part or all of one or more of the cross-over sequences and incubating the plurality of oligonucleotides with a polymerase or a ligase, or both a polymerase and a ligase.
• Any nucleic acid can be expressed by any available method.
• This approach can also use any of the other structural or statistical criteria above or herein for selection of cross-over sites, including, e.g., consideration of one or more features of at least one polypeptide corresponding to at least one of the first polypeptide sequences. These include, e.g., structural stability, allosteric effects, 3-D energetic constraints, etc.
• Fig. 1 is a flow chart describing a portion of directed evolution by
• Fig. 2 is a flow chart describing a portion of directed evolution by GAGGS.
• the flow chart of Fig. 2 is optionally contiguous from Fig. 1.
• Fig. 3 is a flow chart describing a portion of directed evolution by GAGGS.
• the flow chart of Fig. 3 is optionally contiguous from Fig. 2.
• Fig. 4 is a flow chart describing a portion of directed evolution by GAGGS.
• the flow chart of Fig. 4 is optionally contiguous from Fig. 3.
• Fig. 5 is a chart and relational tree showing percent similarity for different subtilisins (an exemplar shuffling target).
• Fig. 6 is a pairwise dot-plot alignment showing homology areas for different subtilisins.
• Fig. 7 is a pairwise dot-plot alignment showing homology areas for 7 different parental subtilisins.
• Panels A-C are pairwise histograms showing conditions determining probability of crossover point selection can be independently controlled for any region over a selected gene length, as well as independently for the pairs of parents.
• Fig. 9 is a chart showing introducing indexed crossover points marker into the sequence of each parent.
• Fig 10 shows a procedure for oligonucleotide assembly to make nucleic acids.
• Fig. 11 is a continuation of figure 13 showing an oligonucleotide assembly scheme.
• Fig. 12 is a difference plot and relatedness tree for shuffling Naphthalene deoxygenase.
• Fig. 13 is a schematic of a digital system of the invention.
• Fig. 14 is a schematic showing a geometric relationship between nucleotides.
• Fig. 15 is a schematic of an HMM matrix.
• sequence strings which can be converted into physical molecules, shuffled and tested for a desired property.
• character strings as "virtual substrates" for shuffling protocols, when coupled with gene reconstruction methods, eliminates the need to obtain parental physical molecules encoding genes.
• the invention provides a variety of methods for selecting cross over sites for recombination between nucleic acids or proteins (or character or other data strings corresponding to such molecules) based upon structural, logical or statistical criteria.
• bridge oligonucleotide(s) can be designed in silico, synthesized and used to achieve any desired physical recombination of any nucleic acid or encoded polypeptide.
• GAs Genetic algorithms
• GAs are used in a wide variety of fields to solve problems which are not fully characterized or which are too complex to allow for full characterization, but for which some analytical evaluation is available. That is, GAs are used to solve problems which can be evaluated by some quantifiable measure for the relative value of a solution (or at least the relative value of one potential solution in comparison to another).
• the basic concept of a genetic algorithm is to encode a potential solution to a problem as a series of parameters. A single set of parameter values is treated as the "genome" or genetic material of an individual solution. A large population of candidate solutions are created.
• a genetic algorithm is used to provide a character string-based representation of the process of generating biopolymer diversity (computational evolution of character strings by application of one or more genetic operators to a provided population (e.g., a parent library) of character strings, e.g. gene sequences).
• a representation of a GA-generated character string population (or "derivative library”) is used as a sequence instruction set in a form suitable to control polynucleotide synthesis (e.g. via non-error-prone synthesis, error-prone synthesis, parallel synthesis, pooled synthesis, chemical synthesis, chemoenzymatic synthesis, (including assembly PCR of synthetic oligonucleotides), and the like). Synthesis of polynucleotides is conducted with sequences encoded by a character string in the, derivative library. This creates a physical representation (a library of polynucleotides) of the computation-generated "gene" (or any other character string) diversity.
• Physical selection of the polynucleotides having desired characteristics is also optionally (and typically) conducted. Such selection is based on results of physical assays of properties of polynucleotides, or polypeptides, whether translated in-vitro, or expressed in-vivo.
• Sequences of those polynucleotides found to have desired characteristics are deconvoluted (e.g., sequenced, or, when positional information is available, by noting the position of the polynucleotide). This is performed by DNA sequencing, by reading a position on an array, real time PCR (e.g., TaqMan), restriction enzyme digestion, or any other method noted herein, or currently available.
• deconvoluted e.g., sequenced, or, when positional information is available, by noting the position of the polynucleotide. This is performed by DNA sequencing, by reading a position on an array, real time PCR (e.g., TaqMan), restriction enzyme digestion, or any other method noted herein, or currently available.
• any nucleic acid which is generated in silico can be synthesized and shuffled by any known DNA shuffling method, including those taught in the references by the inventors and their co- workers cited herein. Such synthesized DNAs can also be mutagenized or otherwise modified according to existing techniques.
• the present invention provides new "in silico" DNA shuffling techniques, in which part, or all, of a DNA shuffling procedure is performed or modeled in a computer system, avoiding (partly or entirely) the need for physical manipulation of nucleic acids. These approaches are collectively termed Genetic Algorithm Guided Gene Synthesis or "GAGGS.”
• the invention provides methods for obtaining a "chimeric" or “recombinant" polynucleotide or polypeptide (or other bio-polymer) having a desired characteristic.
• at least two parental character strings encoding sequence information for one or more polypeptides and/or for one or more single-stranded or double-stranded polynucleotides are provided. All or a part of the sequences (i.e., one or more subsequence regions) contain areas of identity and areas of heterology.
• a set of character strings of a pre-defined or selected length is provided that encodes single-stranded oligonucleotide sequences which include overlapping sequence fragments of at least a part of each of the parental character strings, and/or at least a part of polynucleotide strands complementary to the parental character strings.
• the invention provides methods of generating libraries of biological polymers.
• the method include generating a diverse population of character strings in a computer, where the character strings are generated by alteration (recombination, mutagenesis, etc.) of pre-existing character strings.
• the diverse population of character strings is then synthesized to comprise the library of biological polymers (nucleic acids, polypeptides, peptide nucleic acids, etc.).
• the members of the library of biological polymers are selected for one or more activity.
• an additional library or an additional set of character strings is filtered by subtracting the additional library or the additional set of character strings with members of the library of biological polymers which display activity below a desired threshold.
• the additional library or additional set of character strings is filtered by biasing the additional library, or the additional set of character strings, with members of the library of biological polymers which display activity above a desired threshold.
• a set of single-stranded oligonucleotides made according to the set of sequences defined in the character strings is provided. Part or all of the single stranded nucleotides produced are pooled under denaturing or annealing conditions, where at least two single-stranded oligonucleotides represent parts of two different parental sequences.
• the resultant population of the single-stranded oligonucleotides is incubated with a polymerase under conditions which result in annealing of the single-stranded fragments at areas of identity to form pairs of annealed fragments. These areas of identity are sufficient for one member of the pair to prime replication of the other, resulting in an increase in the length of the oligonucleotides.
• the resulting mixture of double- and single-stranded oligonucleotides are denatured into single-stranded fragments. These steps are repeated, such that at least a part of the resultant mixture of single-stranded chimeric and mutagenized polynucleotides are used in the steps of subsequent cycles. Recombinant polynucleotides having evolved toward a desired property are selected or screened for.
• sequence strings corresponding to the oligonucleotides noted above are selected by the computer from sequence strings corresponding to one or more of the following sets of single-stranded oligonucleotides: a) oligonucleotides synthesized to contain randomly or non-randomly preselected mutations of the parental sequences according to modified sequences including replacement of one or more characters with another character, or deletion or insertion of one or more characters; b) oligonucleotide sequences synthesized to contain degenerate, mixed or unnatural nucleotides, at one or more randomly or non-randomly pre-selected positions; and, c) chimeric oligonucleotides synthesized according to artificial sequences of character substrings designed to contain joined partial sequences of at least two parental sequences.
• oligonucleotides of set (c) contain one or more mutated or degenerate positions defined in sets (a) and (b).
• the oligonucleotides of set (c) are optionally chimeric, nucleotides with crossover points selected according to a method allowing identification of a plurality of character substrings displaying pairwise identity (homology) between any or all of the string pairs comprising sequences of different parental character strings.
• Crossover points for making chimeric oligonucleotide sequences are optionally selected randomly, or approximately in the middle of each or a part of the identified pairwise identity (homology) areas, or by any other set of selection criteria as noted herein.
• at least one crossover point for at least one chimeric oligonucleotide sequence is selected from those not within detected identity areas.
• the mixtures of single stranded oligonucleotides described above are pooled at least once with an additional set of polynucleotides comprising one or more double-stranded or single-stranded polynucleotide encoded by a part and/or by an entire character string of any of the parental sequences provided, and/or by another character string(s) which contains areas of identity and areas of heterology with any of the parental character strings provided.
• the polynucleotides from the additional set of polynucleotides can be obtained by oligonucleotide synthesis of oligonucleotides corresponding to any parental character string (or homolog thereof), or by random fragmentation (e.g., by enzymatic cleavage e.g., by a DNAse, or by chemical cleavage of the polynucleotide) and/or by a restriction-enzyme fragmentation of polynucleotides encoded by character strings defined above, and/or by another character string(s) which contains areas of identity and areas of heterology with any of the parental character strings provided. That is, any nucleic acid generated by GAGGS can be further modified by any available method to produce additionally diversified nucleic acids. Furthermore, any diversified nucleic acid can serve as a substrate for further rounds of GAGGS.
• the above methods are suitably adapted to a wide range of lengths for synthetic oligos (e.g., 10-20 nucleotides or more, 20-40 nucleotides or more, 40-60 nucleotides or more, 60-100 nucleotides or more, 100-150 nucleotides or more, etc.), a wide variety of types of parental sequences (e.g., for therapeutic proteins such as EPO, insulin, growth hormones, antibodies or the like; agricultural proteins such as plant hormones, disease resistance factors, herbicide resistance factors (e.g., p450s,) industrial proteins (e.g., those involved in bacterial oil desulfurization, synthesis of polymers, detoxification proteins and complexes, fermentation or the like)) and for a wide variety in the number of selection/screening cycles (e.g., 1 or more cycle, 2 or more cycle, 3-4 or more cycles, 10 or more cycles, 10-50 or more cycles, 50-100 or more cycles, or more than 100 cycles).
• therapeutic proteins such as EPO
• Rounds of GAGGS evolution can be alternated with rounds of physical nucleic acid shuffling and/or selection assays under various formats (in vivo, or in vitro).
• Selected nucleic acids i.e., those with desirable properties
• Selected nucleic acids can be deconvoluted by sequencing or other procedures such as restriction enzyme analysis, real-time PCR analysis or the like, so that the processes can be started over using the sequence information to guide gene synthesis, e.g., without any physical manipulation of DNA obtained from previous GAGGS rounds.
• the sets of character strings, encoding single- strand oligonucleotides comprising fragments of parental strings, including chimeric and mutated/degenerate fragments of a pre-defined length are generated using a device comprising a processing element, such as a computer with software for sequence string manipulation.
• a processing element such as a computer with software for sequence string manipulation.
• the invention provides for single parent GAGGS.
• the invention also provides methods of producing recombinant nucleic acids using bridging oligonucleotide selection strategies.
• two or more parental nucleic acid sequences are provided.
• Cross -over sites are selected for recombination between the two or more parental nucleic acid sequences, thereby defining one or more recombinant nucleic acids that result from a cross-over between at least two of the two or more parental nucleic acids.
• Corresponding bridging oligonucleotides are defined.
• a recombinant sequence for at least one of the one or more recombinant nucleic acids is determined.
• the at least one recombinant sequence is selected in silico for one or more expected activity and the at least one recombinant sequence is synthesized.
• the synthetic step is typically performed by providing fragments of the two or more parental nucleic acids and at least one of corresponding bridge oligonucleotides, hybridizing the fragments and the bridge oligonucleotides and elongating the hybridized fragments with a polymerase or a ligase. This dramatically simplifies the overall synthesis strategy for creating recombinant nucleic acids between two or more parental sequences, including sequences which display low levels of sequence similarity.
• the present invention further provides methods of producing one or more recombinant nucleic acids or encoded polypeptides.
• first nucleic acid or first polypeptide sequences are provided.
• the first nucleic acid or polypeptide sequences optionally comprise homologous or non-homologous sequences, and the sequences can comprise artificial or natural sequences.
• Cross-over sequences are selected between the plurality of first nucleic acid or first polypeptide sequences by defining structural, statistical, or logical criteria for the cross-over sequences in silico.
• Defining the structural logical or statistical criteria can include any of a variety of methods set for the herein, including performing structural modeling of at least one of the first polypeptide sequences to define one or more region of structural interest in the at least one first polypeptide sequence and selecting one or more cross-over sequence to preserve or disrupt the region of structural interest; defining a structural or sequence- based motif in at least one of the first polynucleotide or polypeptide sequences to define one or more conserved region in the at least one first polynucleotide or polypeptide sequence and selecting one or more cross-over sequence to preserve or disrupt the motif; identifying one or more nucleotides or amino acids within at least one of the first polynucleotide or polypeptide sequences which shows activity or structural co-variance for one or more desired activities or structural features of the first polynucleotide or polypeptide sequence and selecting one or more cross-over sequence to preserve or disrupt the co-variance; performing an energy minimization analysis of the first polynucleotide
• a plurality of recombinant nucleic acids comprising or encoding the cross-over sequences are synthesized. This can include, for example, artificially synthesizing a plurality of recombinant nucleic acids comprising or encoding the cross-over sequences comprises synthesizing a plurality or oligonucleotides, one or more of which encodes part or all of one or more of the cross-over sequences.
• the plurality of oligonucleotides are typically incubated with a polymerase or a ligase (or both a polymerase and a ligase).
• GAGGS is an evolutionary process which includes an information manipulation step (application of a genetic algorithm to a character string representing a biopolymer such as a nucleic acid or protein), to create a set of defined information elements (e.g., character strings) which serve as templates for synthesizing physical nucleic acids.
• the information elements can be placed into a database or otherwise manipulated in silico, e.g., by the recursive application of a GA to the sequences which are produced.
• Corresponding physical nucleic acids can be subjected to recombination/selection or other diversity generating procedures, with the nucleic acids being deconvoluted (e.g., sequenced or otherwise analyzed) and the overall process repeated, as appropriate, to achieve a desired nucleic acid.
• deconvoluted e.g., sequenced or otherwise analyzed
• sequence information is used for oligo design and selection.
• a variety of public databases provide extensive sequence information, including, e.g., GenbankTM and those noted supra. Additional sequence databases are available on a contract basis from a variety of companies specializing in genomic information generation and storage.
• sequences from inaccessible, non-cultivable organisms can be used for GAGGS.
• sequences from pathogenic organisms can be used without actual handling of the pathogens.
• All of the sequence types suitable for physical DNA shuffling, including damaged and incomplete genes (e.g., pseudo genes), are amenable to GAGGS.
• GAGGS has applicability to the self-learning capability of artificial intelligence (optimization algorithm output parameter profiles based on feedback entry of yields, success rates and failures of physical screens, etc.).
• sequences with frame-shift mutations which are generally undesirable
• are eliminated or fixed discarded from the character set, or repaired, in silico.
• entries with premature terminations are discarded or repaired and entries with loss of sequence features known to be important for display of a desired property (e.g. conservative ligands for metal binding) are discarded or repaired.
• wild-type parents do not contaminate derivative libraries with multiple redundant parental molecules, as, in one preferred embodiment, only a priori modified genes are subjected to physical shuffling and/or screening (which, in some cases can be expensive, or low throughput, or otherwise less than ideal, depending on the assay available).
• protein sequences can be shuffled in the same way in silico as nucleic acid sequences, and retrotranslation of the resulting shuffled sequences can be used to alleviate codon usage problems and to minimize the number of oligos needed to build one or more library of coding nucleic acids.
• protein sequences can be shuffled in silico using genetic operators that are based on recognition of structural domains and folding motifs, rather than being bound by annealing-based homology criteria of DNA sequences, or simple homology of AA sequences. Furthermore, rational structure-based biases are easily incorporated in library construction, when such information is available.
• CHARACTER STRINGS in general, a character string can be any representation of an array of characters (e.g., a linear array of characters provides
• character strings are preferably those which encode polynucleotide or polypeptide strings, directly or indirectly, including any encrypted strings, or images, or arrangements of objects which can be transformed unambiguously to character strings representing sequences of monomers or multimers in polynucleotides, polypeptides or the like (whether made of natural or artificial monomers).
• GENETIC ALGORITHM Genetic algorithms generally are processes which mimic evolutionary processes. Genetic algorithms (GAs) are used in a wide variety of fields to solve problems which are not fully characterized or too complex to allow full characterization, but for which some analytical evaluation is available. That is, GAs are used to solve problems which can be evaluated by some quantifiable measure for the relative value of a solution (or at least the relative value of one potential solution in comparison to another).
• GAs are used to solve problems which can be evaluated by some quantifiable measure for the relative value of a solution (or at least the relative value of one potential solution in comparison to another).
• a genetic algorithm is a process for selecting or manipulating character strings in a computer, typically where the character string can be corresponded to one or more biological polymer (e.g., a nucleic acid, protein, PNA, or the like).
• a biological polymer is any polymer which shares some structural features with naturally occurring polymers such as an RNAs, DNAs and polypeptides, including, e.g., RNAs, RNA analogues, DNAs, DNA analogues, polypeptides, polypeptide analogues, peptide nucleic acids, etc.
• DIRECTED EVOLUTION OF CHARACTER STRINGS OR OBJECTS A process of artificially changing a character string by artificial selection, i.e., which occurs in a reproductive population in which there are (1) varieties of individuals, with some varieties being (2) heritable, of which some varieties (3) differ in fitness (reproductive success determined by outcome of selection for a predetermined property (desired characteristic).
• the reproductive population can be, e.g., a physical population or a virtual population in a computer system.
• GENETIC OPERATORS user-defined operations, or sets of operations, each comprising a set of logical instructions for manipulations of character strings.
• Genetic operators are applied to cause changes in populations of individuals in order to find interesting (useful) regions of the search space (populations of individuals with predetermined desired properties) by predetermined means of selection.
• Predetermined (or partially predetermined) means of selection include computational tools (operators comprising logical steps guided by analysis of information describing libraries of character strings), and physical tools for analysis of physical properties of physical objects, which can be built (synthesized) from matter with the purpose of physically creating a representation of information describing libraries of character strings.
• some or all of the logical operations are performed in a computer. Genetic Operators
• MULTIPLICATION is a form of reproduction of character strings, producing additional copies of character strings comprising parental population/library of strings.
• Multiplication operators can have many variations. They can be applied to individual strings or to groups of identical or non-identical strings. Selecting groups of strings for multiplication can be random or biased.
• all mutation types in each member of a set of strings can be described by several simple operations which can be reduced to elements comprising replacement of one set of the characters with another set of characters.
• One or more characters can be mutated in a single operation. When more than one character is mutated, the set of characters may or may not be continuous over an entire string length (a feature useful to simulate closely clustered mutations by certain chemical mutagens).
• a Single point mutation operator replaces a single character with another single character.
• the nature of the new characters can vary, and they can be from the same set of characters making up parental strings, or from different, (e.g., to represent degenerate nucleobases, unnatural nucleobases or amino acids, etc).
• a Deletion mutation is a more complex operator which removes one or more characters from strings.
• Individual single point deletions in nucleic acid-encoding strings may be not desirable for manipulating strings representing polynucleotide sequences; however, 3x clustered (continuous or dispersed) deletions may be acceptable ("triple deletion frameshifts").
• Single point deletions, though, are useful and acceptable for evolutionary computation of strings encoding polypeptides.
• Insertion mutations are operationally similar to deletions, except that one or more new characters are inserted. The nature of the added characters optionally vary, and they can be from the same set of characters making up parental strings, or from different, (e.g., to represent degenerate nucleobases, unnatural nucleobases or amino acids, etc).
• Death can simply be defined as a variation of the deletion operator. It takes place when the result of an application of a genetic operator (or combinations thereof) yields a deletion of an entire individual character string, or entire (sub) population of character strings. Death can also be defined as a variation of an elitism-prone multiplication operator (multiplication of values defining abundance level of one or more strings by zero). Death can also be defined as a default non- selection action in operators, effecting selections of sub-populations of string and transfer manipulations with various sorting and indexing operations of indexed libraries of strings (all non-transferred strings can be considered as dead or non-existent for subsequent computations).
• FRAGMENTATION OF STRINGS are an important class of non- elemental (complex) optional operators which can have advantages for simulating evolution of strings in various formats of DNA shuffling.
• fragmentation can be described as a formal variation of a combination of a deletion operator and a multiplication operator.
• Fragmentation operations may be random or biased. Different ranges of fragment sizes can be predetermined. String fragments may be left in the same population with parental stings, or they may be transferred to different population. Strings fragments from various population strings can be pooled to from new populations.
• CROSSOVER (RECOMBINATION)- This operator formally comprises joining a continuous part of one string with a continuous part of another string in such a way that one or two hybrid strings are formed (chimeras), where each of the chimeras contain at least two connected continuous string areas each comprising partial sequence of two different recombining strings.
• Crossover operations can be combined with mutation operations affecting one or more characters of the recombined strings in a proximity of the crossover area/point of joining.
• LIGATION is a variation of an insertion mutation operator where essentially the entire content of one string is combined with the entire content of another string in a way that the last character of one string is followed by the first character of another string.
• Ligation operation can be combined with mutation operation affecting one or more characters of the ligated strings in a proximity of the point of joining. Ligation can also be viewed as a means of forming chimeras.
• ELITISM is a concept that provides a useful form of bias which imposes discriminating criteria for use of any of the genetic operators, and various types of positive and negative biases can be designed and implemented.
• the rational for the design of elitist operators is based on the concept of fitness. Fitness can be determined using string analysis tools which recognize various sequence-specific features (GC- content, frameshifts, terminations, sequence length, specific substrings, homology properties, ligand-binding and folding motifs, etc) and/or indexed correlated parameters acquired from physical selection of physical representations of character strings (enzyme activity, stability, ligand binding, etc.). It is understood that different elitism criteria can be applied separately to any of the above described genetic operators, or combinations of operators.
• elitism in the same evolutionary computation process, with several operators of the same type, where input/output parameters of each of the similar operators can be controlled independently (or interdependently). Different elitism criteria can be used to control changes in the string character populations caused by action of each of the individual operators.
• SEQUENCE HOMOLOGY or SEQUENCE SIMILARITY is an especially important form of sequence-specific elitism useful for controlling changes in populations of character strings caused by crossover/ recombination operators in those genetic algorithms used to evolve character strings encoding polynucleotide and polypeptide sequences.
• Optimal alignment of sequences for comparison can be conducted, e.g., by the local homology algorithm of Smith & Waterman, Adv. Appl. Math. 2:482 (1981), by the homology alignment algorithm of Needleman & Wunsch, J. Mol. Biol. 48:443 (1970), by the search for similarity method of Pearson & Lipman, Proc. Nat'l. Acad. Sci. USA 85:2444 (1988), by computerized implementations of these algorithms (GAP, BESTFIT, FASTA, and
• HSPs high scoring sequence pairs
• initial neighborhood word hits act as seeds for initiating searches to find longer HSPs containing them.
• the word hits are then extended in both directions along each sequence for as far as the cumulative alignment score can be increased. Cumulative scores are calculated using, for nucleotide sequences, the parameters M (reward score for a pair of matching residues; always > 0) and N (penalty score for mismatching residues; always ⁇ 0). For amino acid sequences, a scoring matrix is used to calculate the cumulative score. Extension of the word hits in each direction are halted when: the cumulative alignment score falls off by the quantity X from its maximum achieved value; the cumulative score goes to zero or below, due to the accumulation of one or more negative-scoring residue alignments; or the end of either sequence is reached.
• the BLAST algorithm parameters W, T, and X determine the sensitivity and speed of the alignment.
• W wordlength
• E expectation
• BLASTP program uses as defaults a wordlength (W) of 3, an expectation (E) of 10, and the BLOSUM62 scoring matrix (see Henikoff & Henikoff (1989) Proc. Natl. Acad. Sci. USA 89:10915).
• W wordlength
• E expectation
• BLOSUM62 scoring matrix see Henikoff & Henikoff (1989) Proc. Natl. Acad. Sci. USA 89:10915.
• the BLAST algorithm also performs a statistical analysis of the similarity between two sequences (see, e.g., Karlin & Altschul (1993) Proc. Nat'l. Acad. Sci. USA 90:5873-5787).
• P(N) the smallest sum probability
• a nucleic acid is considered similar to a reference sequence (and, therefore, homologous) if the smallest sum probability in a comparison of the test nucleic acid to the reference nucleic acid is less than about 0.1, or less than about 0.01, and or even less than about 0.001.
• An additional example of a useful sequence alignment algorithm is
• PDMEUP creates a multiple sequence alignment from a group of related sequences using progressive, pairwise alignments. It can also plot a tree showing the clustering relationships used to create the alignment.
• PDMEUP uses a simplification of the progressive alignment method of Feng & Doolittle, J. Mol. Evol. 35:351-360 (1987). The method used is similar to the method described by Higgins & Sharp, CABIOS5: 151- 153 (1989).
• the program can align, e.g., up to 300 sequences of a maximum length of 5,000 letters.
• the multiple alignment procedure begins with the pairwise alignment of the two most similar sequences, producing a cluster of two aligned sequences.
• This cluster can then be aligned to the next most related sequence or cluster of aligned sequences.
• Two clusters of sequences can be aligned by a simple extension of the pairwise alignment of two individual sequences. The final alignment is achieved by a series of progressive, pairwise alignments.
• the program can also be used to plot a dendogram or tree representation of clustering relationships. The program is run by designating specific sequences and their amino acid or nucleotide coordinates for regions of sequence comparison.
• Homology-based elitism of crossover operators can thus be used (a) to find suitable recombination pairs of strings in a population of strings, and/or (b) to find/predetermine particularly suitable/desired areas/points of recombination over lengths of character strings selected for recombination.
• Setting predetermined types and stringency of similarity/homology as a condition for crossover to occur is a form of elitism for control of formation of chimeras between representative parental character strings of various degree of homology.
• Each genetic operator can be applied to randomly selected strings and/or to randomly selected positions over one or more string's length, with occurrence frequencies randomly selected within a range.
• Order determining applications of individual GOs to product derivative libraries of character strings may be different and may depend on the composition of a particular set of individual GOs selected for practicing various formats of GAGGS.
• the order may be linear, cyclic, parallel, or a combination of the three and can typically be represented by a graph.
• Many GO arrangements can be used to simulate natural sexual and mutagenic processes for generating genetic diversity, or artificial protocols, such as single-parent or family DNA shuffling.
• the purpose of GA is not in limited to simulation of some known physical DNA manipulation methods. Its main aim is in the provision of a formal and intelligent tool, based on understanding natural and artificial evolution processes, for creation and optimization of evolutionary protocols of practical utility which may provide effective advantages over currently practiced methods.
• genes encoded by derivative libraries of character strings obtained by operation of genetic algorithms, is the primary means to create a physical representation of matter that is amenable to a physical assay for a desired property or to produce substrates that are further evolved in physical diversity generation procedures.
• one aspect of the present invention relates to the synthesis of genes with sequences selected following one or more computer shuffling procedure as set forth herein.
• GAGGS For GAGGS to be a time and resource effective technology, gene synthesis technology is used, typically to construct libraries of genes in a consistent manner, and in close adherence to the sequence representations produced by GA manipulations. GAGGS typically uses gene synthesis methods which allow for rapid construction of libraries of 10 4 -10 9 "gene" variations. This is typically adequate for screening/selection protocols, as larger libraries are more difficult to make and maintain and sometimes cannot be as completely sampled by a physical assay or selection method. For example, existing physical assay methods in the art (including, e.g., "life-and-death” selection methods) generally allow sampling of about 10 9 variations or less by a particular screen of a particular library, and many assays are effectively limited to sampling of 10 4 -10 5 members. Thus, building several smaller libraries is a preferred method, as large libraries cannot easily be completely sampled. However, larger libraries can also be made and sampled, e.g., using high-throughput screening methods.
• oligonucleotides e.g., for use in in vitro amplification/ gene reconstruction methods, for use as gene probes, or as shuffling targets (e.g., synthetic genes or gene segments) are typically synthesized chemically according to the solid phase phosphoramidite triester method described by Beaucage and Caruthers (1981), Tetrahedron Letts., 22(20): 1859-1862, e.g., using an automated synthesizer, as described in Needham-VanDevanter et al. (1984) Nucleic Acids Res., 12:6159-6168. Oligonucleotides can also be custom made and ordered from a variety of commercial sources known to persons of skill.
• nucleic acid can be custom ordered from any of a variety of commercial sources, such as The Midland Certified Reagent Company (mcrc@oligos.com), The Great American Gene Company (http://www.genco.com), ExpressGen Inc. (www.expressgen.com), Operon Technologies Inc. (Alameda, CA) and many others.
• peptides and antibodies can be custom ordered from any of a variety of sources, such as PeptidoGenic (pkim@ccnet.com), HTI Bio-products, inc. (http://www.htibio.com), BMA Biomedicals Ltd (U.K.), Bio-Synthesis, Inc., and many others.
• nucleic acid ligases or polymerases allows great freedom of oligo design and generation of relevant mixtures. Further, synthetic and assembly parameters permits considerable control over library design. Whether polymerase mediated or ligase mediated assembly methods (or a combination thereof) are used, oligos for libraries assembled by DNA ligase are synthesized, e.g., by conventional chemistry, by split-pool synthesis, or use of trinucleotide phosphoramidites, as described herein.
• oligos are then assembled into full-length sequences of interest. Oligo mixtures can be spiked with partial or full-length homologous sequences (e.g., single or double-stranded sequences) e.g., from naturally occurring, synthetic or cloned sequences, to facilitate gene reassembly methods.
• partial or full-length homologous sequences e.g., single or double-stranded sequences
• polymerase mediated, ligation-mediated and combination ligation/ polymerase mediated assembly methods are suitable for construction of individual sequences and/or synthetic libraries (see also, "OLIGONUCLEOTIDE MEDIATED NUCLEIC ACID RECOMBINATION" by Crameri et al., Filed January 18, 2000, USSN PCT/USOO/01203).
• top and bottom strand oligos can be designed to be overlapping, but with the oligos for each strand being abutting, rather than overlapping, as in typical polymerase-mediated assembly reactions.
• oligos are optionally phosphorylated, e.g., with a phosphorylase or a kinase enzyme, or by chemical addition of a phosphate during or following oligonucleotide synthesis.
• Phosphorylated oligos are assembled with a DNA ligase, e.g., T DNA ligase or another available DNA ligase. Either thermostable or thermolabile ligases can be used.
• a ligation chain reaction can be performed to achieve assembly, e.g., where a thermostable ligase is used for assembly.
• LCR ligation chain reaction
• An example of an LCR-mediated gene synthesis approach is described by Au et al. (1998) "Gene Synthesis by a LCR-Based Approach: High level Production of Leptin- L54 Using Synthetic Gene in Escheria coli” Biochemical and Biophysical Research Communications 248:200-203. The gene synthetic strategies described supra and in "OLIGONUCLEOTIDE MEDIATED NUCLEIC ACID RECOMBINATION" by
• a number of the publications of the inventors and their co-workers, as well as other investigators in the art also describe techniques which facilitate DNA shuffling, e.g., by providing for reassembly of genes from small fragments, or even oligonucleotides.
• One aspect of the present invention is the ability to use family shuffling oligonucleotides and cross over oligonucleotides as recombination templates/intermediates in various DNA shuffling methods.
• a number of the publications by the inventors and their co- workers, as well as other investigators in the art also describe techniques which facilitate reassembly of genes from small fragments, including oligonucleotides.
• synthetic recombination methods are used, in which oligonucleotides corresponding to different homologues are synthesized and reassembled in PCR or ligation reactions which include oligonucleotides which correspond to more than one parental nucleic acid, thereby generating new recombined nucleic acids.
• oligonucleotide-mediated recombination is the ability to recombine homologous nucleic acids with low sequence similarity, or even to recombine non-homologous nucleic acids.
• these low-homology oligonucleotide shuffling methods one or more set of fragmented nucleic acids is recombined, e.g., with a set of crossover family diversity oligonucleotides.
• Each of these crossover oligonucleotides have a plurality of sequence diversity domains corresponding to a plurality of sequence diversity domains from homologous or non-homologous nucleic acids with low sequence similarity.
• the fragmented oligonucleotides which are derived by comparison to one or more homologous or non-homologous nucleic acids, can hybridize to one or more region of the crossover oligos, facilitating recombination.
• Such oligonucleotide sets are selected in silico according to the methods herein.
• sets of overlapping family gene shuffling oligonucleotides (which are derived by comparison of homologous nucleic acids and synthesis of oligonucleotide fragment sets, which correspond to regions of similarity and regions of diversity derived from the comparison) are hybridized and elongated (e.g., by reassembly PCR), providing a population of recombined nucleic acids, which can be selected for a desired trait or property.
• the set of overlapping family shuffling gene oligonucleotides include a plurality of oligonucleotide member types which have consensus region subsequences derived from a plurality of homologous target nucleic acids.
• family gene shuffling oligonucleotide are provided by aligning homologous nucleic acid sequences to select conserved regions of sequence identity and regions of sequence diversity.
• a plurality of family gene shuffling oligonucleotides are synthesized (serially or in parallel) which correspond to at least one region of sequence diversity. Further details regarding family shuffling is found in USSN 09/408,392, cited above.
• Sets of fragments, or subsets of fragments used in oligonucleotide shuffling approaches can be provided by cleaving one or more homologous nucleic acids (e.g., with a DNase), or, more commonly, by synthesizing a set of oligonucleotides corresponding to a plurality of regions of at least one nucleic acid (typically oligonucleotides corresponding to a full-length nucleic acid are provided as members of a set of nucleic acid fragments).
• homologous nucleic acids e.g., with a DNase
• synthesizing a set of oligonucleotides corresponding to a plurality of regions of at least one nucleic acid typically oligonucleotides corresponding to a full-length nucleic acid are provided as members of a set of nucleic acid fragments.
• these cleavage fragments can be used in conjunction with family gene shuffling oligonucleotides, e.g., in one or more recombination reaction to produce recombinant nucleic acids.
• Gene assembly by PCR from single-strand complementary overlapping synthetic oligos is a method of choice for practicing in GAGGS. Optimization of this method can be performed e.g., including varying oligo length, the number of oligos in the recombination reaction, the degree of oligonucleotide overlap, levels and nature of sequence degeneracy, specific reaction conditions and particular polymerase enzymes used in the reassembly, and in controlling the stringency of gene assembly to decrease or increase the number of sequence deviations during gene synthesis.
• the method can also be practiced in a parallel mode where each of the individual library members, including a plurality of the genes intended for subsequent physical screening, are synthesized in spatially separated vessels, or arrays of vessels, or in a poolwise fashion, where all, or part, of the desired plurality of genes are synthesized in a single vessel.
• Many other synthesis methods for making synthetic nucleotides are also known, and specific advantages of use of one vs. another for practicing GAGGS may be readily determined by one skilled in the art.
• Sequence deconvolution is performed on those variants of polynucleotides which are found to have desired properties, in order to confirm changes in corresponding character strings (i.e., corresponding to physical sequences for biopolymers) yielding desired changes in the relevant composition of matter (e.g., a polynucleotide, polypeptide, or the like).
• In vitro amplification methods can also be used to amplify and/or sequence GAGGS generated nucleic acids, e.g., for cloning, and selection.
• Examples of techniques sufficient to direct persons of skill through typical in vitro amplification and sequencing methods including the polymerase chain reaction (PCR) the ligase chain reaction (LCR), Q ⁇ -replicase amplification and other RNA polymerase mediated techniques (e.g., NASBA) are found in Berger, Sambrook, and Ausubel, id., as well as in Mullis et al, (1987) U.S. Patent No. 4,683,202; PCR Protocols A Guide to Methods and Applications (Innis et al. eds) Academic Press Inc.
• PCR polymerase chain reaction
• LCR ligase chain reaction
• Q ⁇ -replicase amplification e.g., NASBA
• directed evolution by GAGGS, one or more deconvoluted character string(s), encoding sequences of those variants displaying certain changes in level of desired properties (where the level is arbitrarily defined by increase/decrease/ratios between measures of several properties), can be used to comprise a new library of character strings for a new round of GAGGS.
• Recursive GAGGS unlike typical DNA shuffling, does not use physical manipulation of the polynucleotides in order to produce subsequent generations of gene diversity.
• GAGGS simply uses sequence information describing acquired beneficial changes as a foundation for generating additional changes leading to subsequent changes (improvements) in the desired properties of molecules encoded by the character strings.
• Recursive GAGGS can be performed until strings of characters are evolved to the point when the encoded polynucleotides and polypeptides attain arbitrarily set levels of desired characteristics or until further changes in characteristics cannot be obtained (e.g., enzyme turnover reached theoretical diffusion rate limit under conditions of a physical assay). Genetic algorithm parameters, gene synthesis methods and schemes, as well as physical assays and sequence deconvolution methods can vary in each of the different rounds/cycles of directed evolution by recursive GAGGS.).
• GAGGS constitutes a self-sufficient and independent technology which can be practiced separately from or in conjunction with DNA shuffling or any other available directed evolution methods.
• one or more rounds of GAGGS can be, and often is, practiced in combination with physical shuffling of nucleic acids, and/or in combination with any other mutagenic technique, including site directed mutagenesis, error-prone PCR (e.g., as alternating cycles of a directed evolution process) or any other diversity generation methods.
• GAGGS -generated libraries of polynucleotides can be subjected to nucleic acid modification, whether by shuffling or any other method, and polynucleotides found to have desired characteristics following rounds of in silico and/or physical shuffling can be selected and sequenced to provide character strings to evaluate GAGGS processes or to form character strings for further GAGGS operations.
• GAGGS can be performed as a stand-alone technology, or can be preceded or followed by shuffling, mutagenesis, random priming PCR, or any other method.
• the methods of the invention entail performing physical recombination (including, e.g., shuffling) and screening or selection to evolve individual genes, whole plasmids, viruses, multigene clusters, or even whole genomes
• physical recombination including, e.g., shuffling
• screening or selection to evolve individual genes, whole plasmids, viruses, multigene clusters, or even whole genomes
• the techniques of the inventors and their co-workers are particularly useful. For example, reiterative cycles of recombination and screening/selection can be performed to further evolve the nucleic acids of interest which are generated by performing a GO on a character string (e.g., followed by synthesis of corresponding oligonucleotides, and gene generation/regeneration, e.g., by assembly PCR).
• shuffling or "recursive recombination" of nucleic acids to provide new nucleic acids with desired properties
• shuffling or "recursive recombination" of nucleic acids to provide new nucleic acids with desired properties
• Any of these methods are integrated with those of the present invention by incorporating nucleic acids corresponding to character strings produced by performing one or more GO on one or more selected parental character string, e.g., in combination with selection of crossover site(s) according to the present invention. Any of these methods can be adapted to the present invention to evolve GAGGS produced nucleic acids as discussed herein to produce new nucleic acids with improved properties.
• nucleic acids such as those produced by synthesis of sets of nucleic acids corresponding to character strings produced by GO manipulation of character strings, or available homologues of such sets, or both, can be recombined in vitro by any of a variety of techniques discussed in the references above, including e.g., DNAse digestion of nucleic acids to be recombined followed by ligation and/or PCR reassembly of the nucleic acids, with or without a template (e.g., a single-stranded nucleic acid sequence) to help direct reassembly.
• a template e.g., a single-stranded nucleic acid sequence
• sets of nucleic acids corresponding to character strings produced by GO manipulation of character strings, and/or available homologues of such sets can be recursively recombined in vivo, e.g., by allowing recombination to occur between the nucleic acids while in cells.
• whole cell genome recombination methods can be used in which whole genomes of cells are recombined, optionally including spiking of the genomic recombination mixtures with desired library components such as with sets of nucleic acids corresponding to character strings produced by GO manipulation of character strings, or available homologues of such sets.
• oligonucleotides corresponding to different homologues are synthesized and reassembled in PCR or ligation reactions which include oligonucleotides which correspond to more than one parental nucleic acid, thereby generating new recombined nucleic acids.
• Oligonucleotides can be made by standard nucleotide addition methods, or can be made by tri-nucleotide synthetic approaches.
• Fifth, purely in silico methods of recombination can be effected in which GOs are used in a computer to recombine sequence strings which correspond to nucleic acid or proteins homologues.
• the resulting recombined sequence strings are optionally converted into nucleic acids by synthesis of nucleic acids which correspond to the recombined sequences, e.g., in concert with oligonucleotide synthesis/ gene reassembly techniques. Any of the preceding general recombination formats, separately or together, can be practiced in a reiterative fashion to generate a diverse set of recombinant nucleic acids.
• nucleic acids of the invention can be recombined (with each other or with related (or even unrelated) nucleic acids to produce a diverse set of recombinant nucleic acids, including homologous nucleic acids.
• Other diversity generating approaches can also be used to modify character strings or nucleic acids in combination with the methods of the present invention. Additional diversity can be introduced into input or output nucleic acids by methods which result in the alteration of individual nucleotides or groups of contiguous or non-contiguous nucleotides, i.e., mutagenesis methods.
• Mutagenesis methods include, for example, recombination (PCT/US 98/05223; Publ. No. WO98/42727); oligonucleotide-directed mutagenesis (for review see, Smith, Ann. Rev .Genet. 19: 423- 462 (1985); Botstein and Shortle, Science 229: 1193-1201 (1985); Carter, Biochem. J. 237: 1-7 (1986); Kunkel, "The efficiency of oligonucleotide directed mutagenesis" in Nucleic acids & Molecular Biology, Eckstein and Lilley, eds., Springer Verlag, Berlin
• kits for mutagenesis, library construction and other diversity generation methods are also commercially available.
• kits are available from, e.g., Stratagene (e.g., QuickChangeTM site-directed mutagenesis kit; and ChameleonTM double-stranded, site-directed mutagenesis kit), Bio/Can Scientific, Bio-Rad (e.g., using the Kunkel method described above), Boehringer Mannheim Corp., Clonetech Laboratories, DNA Technologies, Epicentre Technologies (e.g., 5 prime 3 prime kit); Genpak Inc, Lemargo Inc, Life Technologies (Gibco BRL), New England Biolabs, Pharmacia Biotech, Promega Corp., Quantum Biotechnologies, Amersham International pic (e.g., using the Eckstein method), and Boothn Biotechnology Ltd (e.g., using the Carter/Winter method).
• Stratagene e.g., QuickChangeTM site-directed mutagenesis kit
• Non-Stochastic methods of generating nucleic acids and polypeptides are proposed in Short “Non-Stochastic Generation of Genetic Vaccines and Enzymes” WO 00/46344. These methods, including proposed non-stochastic polynucleotide reassembly and site-saturation mutagenesis methods can be applied in conjunction with the methods of the present invention as well.
• Random or semi-random mutagenesis using doped or degenerate oligonucleotides is also described in, e.g., Arkin and Youvan (1992) "Optimizing nucleotide mixtures to encode specific subsets of amino acids for semi- random mutagenesis" Biotechnology 10:297-300; Reidhaar-Olson et al. (1991) "Random mutagenesis of protein sequences using oligonucleotide cassettes" Methods Enzymol 208:564-86; Lim and Sauer (1991) "The role of internal packing interactions in determining the structure and stability of a protein” J Mol. Biol.
• this can include testing for and identifying any detectable or assayable activity, by any relevant assay in the art.
• a variety of related (or even unrelated) properties can be assayed for, using any available assay.
• a recombinant nucleic acid produced by recursively recombining one or more polynucleotide of the invention (produced by GAGGS methods) with one or more additional nucleic acid (e.g., via physical methods) forms a part of the invention.
• the one or more additional nucleic acid may include another polynucleotide of the invention; optionally, alternatively, or in addition, the one or more additional nucleic acid can include, e.g., a nucleic acid encoding a naturally-occurring sequence or a subsequence, or any homologous sequence or subsequence.
• recombining steps can be performed in vivo, in vitro, or in silico as described in more detail in the references above and herein.
• a cell containing any resulting recombinant nucleic acid, nucleic acid libraries produced by recursive recombination of the nucleic acids set forth herein, and populations of cells, vectors, viruses, plasmids or the like comprising the library or comprising any recombinant nucleic acid resulting from recombination (or recursive recombination) of a nucleic acid as set forth herein with another such nucleic acid, or an additional nucleic acid.
• Corresponding sequence strings in a database present in a computer system or computer readable medium are a feature of the invention.
• a typical physical recombination procedure starts with at least two substrates that generally show at least some identity to each other (i.e., at least about 30%, 50%, 70%, 80% or 90% or more sequence identity), but differ from each other at certain positions (however, in purely in silico or cross-over oligonucleotide mediated formats, nucleic acids can show little or no homology).
• two or more nucleic acids can be recombined herein.
• the differences between the nucleic acids can be any type of mutation, for example, substitutions, insertions and deletions. Often, different segments differ from each other in about 1-20 positions.
• the starting materials differ from each other in at least two nucleotide positions. That is, if there are only two substrates, there should be at least two divergent positions. If there are three substrates, for example, one substrate can differ from the second at a single position, and the second can differ from the third at a different single position.
• any GO herein can be used to modify the nucleic acid to produce a diverse array of nucleic acids that can be screened for an activity of interest.
• starting DNA segments can be natural variants of each other, for example, allelic or species variants.
• the segments are derived from one or more homologous nucleic acid sequence.
• the segments can also be from nonallelic genes showing some degree of structural and usually functional relatedness.
• the starting DNA segments can also be induced variants of each other. For example, one DNA segment can be produced by error-prone PCR replication of the other, or by substitution of a mutagenic cassette. Induced mutants can also be prepared by propagating one (or both) of the segments in a mutagenic strain. In these situations, strictly speaking, the second DNA segment is not a single segment but a large family of related segments.
• the different segments forming the starting materials are often the same length or substantially the same length. However, this need not be the case; for example, one segment can be a subsequence of another.
• the segments can be present as part of larger molecules, such as vectors, or can be in isolated form.
• the nucleic acids of interest are derived from DE by GAGGS. CODQN-VAR ⁇ ED OLIGONUCLEOTIDE METHODS
• Codon-varied oligonucleotides are oligonucleotides, similar in sequence but with one or more base variations, where the variations correspond to at least one encoded amino acid difference. They can be synthesized utilizing tri-nucleotide, i.e., codon-based phosphoramidite coupling chemistry, in which tri-nucleotide phosphoramidites representing codons for all 20 amino acids are used to introduce entire codons into oligonucleotide sequences synthesized by this solid-phase technique.
• all of the oligonucleotides of a selected length e.g., about 20, 30, 40, 50, 60, 70, 80, 90, or 100 or more nucleotides
• codon-varied oligonucleotide sequences can be based upon sequences from a selected set of nucleic acids, generated by any of the approaches noted herein. Further details regarding tri-nucleotide synthesis are found in USSN 09/408,393 "USE OF CODON VARIED OLIGONUCLEOTIDE SYNTHESIS FOR SYNTHETIC SHUFFLING" by Welch, et al., filed 09/28/1999.
• Oligonucleotides can be made by standard nucleotide addition methods, or can be made by tri-nucleotide synthetic approaches.
• An advantage of selecting changes which correspond to encoded amino acid differences is that the modification of triplets of codons results in fewer frameshifts (and, therefore, likely fewer inactive library members). Also, synthesis which focuses on codon modification, rather than simply on base variation, reduces the total number of oligos which are needed for a synthesis protocol.
• sets of oligos can be combined for assembly in many different formats and different combinations schemes to effect correlation with genetic events and operators at the physical level.
• overlapping sets of oligonucleotides can be synthesized and then hybridized and elongated to form full-length nucleic acids.
• a full length nucleic acid is any nucleic acid desired by an investigator which is longer than the oligos which are used in the gene reconstruction methods. This can correspond to any percentage of a naturally occurring full length sequence, e.g., 10%, 20%, 30%, 40%, 50%, 60%, 70%, 80%, or 90% or more of the corresponding natural sequence.
• Oligo sets often have at least about 5, sometimes about 10, often about 15, generally about 20, or more, nucleotide overlap sequences to facilitate gene reconstruction.
• Oligo sets are optionally simplified for gene reconstruction purposes where regions of fortuitous overlap are present, i.e., where repetitive sequence elements are present or designed into a gene sequence to be synthesized. Lengths of oligos in a set can be the same or different, as can the regions of sequence overlap. To facilitate hybridization and elongation (e.g., during cycles of PCR), overlap regions are optionally designed with similar melting temperatures.
• Parental sequences can be gridded (conceptually or physically) and the common sequences used to select common sequence oligos, thereby combining oligo members into one or more sets to reduce the number of oligos required for making full- length nucleic acids.
• oligonucleotides with some sequence similarity can be generated by pooled and/or split synthesis where pools of oligos under synthesis are split into different pools during the addition of heterologous bases, optionally followed by rejoined synthesis steps (pooling) at subsequent stages where the same additions to the oligos are required.
• Oligo shuffling formats heterologous oligos corresponding to many different parents can be split and rejoined during synthesis.
• nucleobase can be added during single synthetic steps to produce two or more variations in sequence in two or more resulting oligonucleotides.
• the relative percentage of nucleobase addition can be controlled to bias synthesis towards one or more parental sequence.
• partial generacy can be practiced to prevent the insertion of stop codons during degenerated oligonucleotide synthesis.
• Oligos which correspond to similar subsequences from different parents can be the same length or different, depending on the subsequences. Thus, in split and pooled formats, some oligos are optionally not elongated during every synthetic step (to avoid frame-shifting, some oligos are not elongated for the steps corresponding to one or more codon).
• crossover oligos can be constructed at one or more point of difference between two or more parental sequences (a base change or other difference is a genetic locus which can be treated as a point for a crossover event).
• the crossover oligos have a region of sequence identity to a first parental sequence, followed by a region of identity to a second parental sequence, with crossover point occurring at the locus. For example, every natural mutation can be a cross over point.
• Another way of biasing sequence recombination is to spike a mixture of oligonucleotides with fragments of one or more parental nucleic acid (if more than one parental nucleic acid is fragmented, the resulting segments can be spiked into a recombination mixture at different frequencies to bias recombination outcomes towards one or more parent).
• Recombination events can also be engineered simply by omitting one or more oligonucleotide corresponding to one or more parent from a recombination mixture.
• diversity can be modulated by the addition of selected, pseudo-random or random oligos to elongation mixture, which can be used to bias the resulting full-length sequences.
• mutagenic or non-mutagenic conditions can be selected for PCR elongation, resulting in more or less diverse libraries of full length nucleic acids.
• oligo sets which correspond to different parents in the elongation mixture oligo sets which correspond to just one parent can be elongated to reconstruct that parent. In either case, any resulting full-length sequence can be fragmented and recombined, as in the DNA shuffling methods noted in the references cited herein.
• oligonucleotide sets and synthetic variations which can be correlated to genetic events and operators at the physical level are found in "OLIGONUCLEOTIDE MEDIATED NUCLEIC ACID RECOMBINATION" by Crameri et al., filed February 5, 1999 (USSN 60/118,813) and filed June 24, 1999 (USSN 60/141,049) and filed September 28, 1999 (USSN 09/408,392) and "USE OF CODON-BASED OLIGONUCLEOTIDE SYNTHESIS FOR SYNTHETIC SHUFFLING" by Welch et al, filed September 28, 1999 (USSN 09/408,393).
• any nucleic acid can be shuffled using the GAGGS methods herein. No attempt is made herein to identify the hundreds of thousands of known nucleic acids.
• Common sequence repositories for known proteins include GenBank EMBL, DDB J and the NCBI. Other repositories can easily be identified by searching the internet.
• One class of preferred targets for GAGGS methods includes nucleic acids encoding therapeutic proteins such as erythropoietin (EPO), insulin, peptide hormones such as human growth hormone; growth factors and cytokines such as epithelial Neutrophil Activating Peptide-78, GRO ⁇ /MGSA, GRO ⁇ , GRO ⁇ , MlP-l ⁇ , MIP-16, MCP-1, epidermal growth factor, fibroblast growth factor, hepatocyte growth factor, insulin-like growth factor, the interferons, the interleukins, keratinocyte growth factor, leukemia inhibitory factor, oncostatin M, PD-ECSF, PDGF, pleiotropin, SCF, c-kit ligand, VEGEF, G-CSF etc.
• therapeutic proteins such as erythropoietin (EPO), insulin, peptide hormones such as human growth hormone
• growth factors and cytokines such as epithelial Neutrophil Activating Peptide-
• GAGGS GAGGS
• transcription and expression activators include genes and proteins that modulate cell growth, differentiation, regulation, or the like. Expression and transcriptional activators are found in prokaryotes, viruses, and eukaryotes, including fungi, plants, and animals, including mammals, providing a wide range of therapeutic targets.
• expression and transcriptional activators regulate transcription by many mechanisms, e.g., by binding to receptors, stimulating a signal transduction cascade, regulating expression of transcription factors, binding to promoters and enhancers, binding to proteins that bind to promoters and enhancers, unwinding DNA, splicing pre-mRNA, polyadenylating RNA, and degrading RNA.
• Expression activators include cytokines, inflammatory molecules, growth factors, their receptors, and oncogene products, e.g., interleukins (e.g., DM-1, IL-2, IL-8, etc.), interferons, FGF, IGF-I, IGF-II, FGF, PDGF, TNF, TGF- ⁇ , TGF- ⁇ , EGF, KGF, SCF/c- Kit, CD40L/CD40, VLA-4/VCAM-1, ICAM-l/LFA-1, and hyalurin/CD44; signal transduction molecules and corresponding oncogene products, e.g., Mos, Ras, Raf, and Met; and transcriptional activators and suppressors, e.g., p53, Tat, Fos, Myc, Jun, Myb, Rel, and steroid hormone receptors such as those for estrogen, progesterone, testosterone, aldosterone, the LDL receptor ligand and corticosterone.
• proteins from infectious organisms for possible vaccine applications including infectious fungi, e.g., Aspergillus, Candida species; bacteria, particularly E. coli, which serves a model for pathogenic bacteria, as well as medically important bacteria such as Staphylococci (e.g., aureus), Streptococci (e.g., pneumoniae), Clostridia (e.g., perfringens), Neisseria (e.g., gonorrhoea), Enterobacteriaceae (e.g., coli), Helicobacter (e.g., pylori), Vibrio (e.g., cholerae), Capylobacter (e.g.
• Pseudomonas e.g., aeruginosa
• Hemeophilus e.g., influenzae
• Bordetella e.g., pertussis
• Mycoplasma e.g., pneumoniae
• Ureaplasma e.g., urealyticum
• Legionella e.g., pneumophila
• Spirochetes e.g., Treponema, Leptospira, and Borrelia
• Mycobacteria e.g., tuberculosis, smegmatis
• Actinomyces e.g., israelii
• Nocardia e.g., asteroides
• Chlamydia e.g., trachomatis
• Rickettsia e.g., Coxiella, Ehrilichia, Rochalimaea, Brucella, Yersinia, Fracisella, and Pasteurella
• RNA viruses examples include Rhabdo viruses, e.g., VSV; Paramyxo viruses, e.g., RSV; Orthomyxovimses, e.g., influenza; Bunyaviruses; and Arenaviruses), dsDNA viruses (Reoviruses, for example), RNA to DNA viruses, i.e., Retroviruses, e.g., especially HIV and HTLV, and certain DNA to RNA viruses such as Hepatitis B virus.
• Rhabdo viruses e.g., VSV
• Paramyxo viruses e.g., RSV
• Orthomyxovimses e.g., influenza
• Bunyaviruses Bunyaviruses
• Arenaviruses Arenaviruses
• RNA to DNA viruses i.e., Retroviruses, e.g., especially HIV and HTLV
• certain DNA to RNA viruses such as Hepatitis B virus.
• nucleic acids encoding proteins relevant to non-medical uses such as inhibitors of transcription or toxins of crop pests e.g., insects, fungi, weed plants, and the like, are also preferred targets for GAGGS.
• Industrially important enzymes such as monooxygenases, proteases, nucleases, and lipases are also preferred targets.
• subtilisin can be evolved by shuffling selected forms of the gene for subtilisin (von der Osten et al., J. Biotechnol. 28:55-68 (1993) provide a subtilisin coding nucleic acid). Proteins which aid in folding such as the chaperonins are also preferred.
• Preferred known genes suitable for codon alteration and shuffling also include the following: AIpha-1 antitrypsin, Angiostatin, Antihemolytic factor, Apolipoprotein, Apoprotein, Atrial natriuretic factor, Atrial natriuretic polypeptide,
• Atrial peptides C-X-C chemokines (e.g., T39765, NAP-2, ENA-78, Gro-a, Gro-b, Gro- c, rP-10, GCP-2, NAP-4, SDF-1, PF4, MIG), Calcitonin, CC chemokines (e.g., Monocyte chemoattractant protein- 1, Monocyte chemoattractant protein-2, Monocyte chemoattractant protein-3, Monocyte inflammatory protein- 1 alpha, Monocyte inflammatory protein-1 beta, RANTES, 1309, R83915, R91733, HCC1, T58847, D31065, T64262), CD40 ligand, Collagen, Colony stimulating factor (CSF), Complement factor 5 a, Complement inhibitor, Complement receptor 1, Factor IX, Factor VII, Factor VIII, Factor X, Fibrinogen, Fibr nectin, Glucocerebrosidase, Gonadotropin, Hedgehog proteins
• TNF beta Tumor necrosis factor beta
• TNFR Tumor necrosis factor receptor
• TNF alpha Tumor necrosis factor-alpha
• Urokinase Urokinase
• the present invention overcomes this difficulty by providing for in silico design of a "diplomat" sequence which has an intermediate level of homology to each of the sequences to be recombined, thereby facilitating cross-over events between the sequences and facilitating chimera formation.
• This diplomat sequence can be a character string produced by any of a variety of GO to establish intermediate sequence similarity in the diplomat sequence as compared to the sequences to be recombined, including by alignment of the sequences to select a consensus sequence, codon modification to optimize similarity between diverse nucleic acids, or the like.
• the consensus sequence is generated by comparison and lining-up/pile-up of a family of genes (DNA consensus), or of amino acid sequence line-up/pile-up (aa consensus).
• DNA consensus DNA consensus
• amino acid sequence line-up/pile-up aa consensus
• the amino acid consensus sequence are optionally back-translated using a desirable codon bias to further enhance homology, or to enhance host organism for expression, or to select for alternate codon usages in order to enable access to alternative sets of amino acid codons.
• consensus sequence itself may encode an improved enzyme. This has been observed elsewhere (e.g. presentation at International Conference “Enzyme Opportunities on the Next Millenium", Chicago, DM, May 5-7, by Dr. Luis Pasamontes, Roche Vitamins, Inc., on "Development of Heat Stable Phytase"- a consensus phytase had an increase of 16 degrees C in thermostability).
• consensus Interferon IFN-conl
• diplomat sequences can be designed using selected GO criteria and, optionally, physically synthesized and shuffled using any of the techniques herein.
• Automatic processing steps e.g., performed in a digital system as described herein
• Automatic processing steps that perform the following functions facilitate selection of oligonucleotides in synthetic shuffling techniques herein.
• the system can include an instruction set which permits inputting of amino acid sequences of a family of proteins of interest of interest.
• codon usage parameters e.g., optimal usage parameters for one or more organism to be used for expression, or to optimize sequence alignments to facilitate recombination, or both.
• codon usage can be selected for multiple expression hosts, e.g. E. coli and S. cerevisiae. In some cases, simply optimizing codon usage for expression in a host cell will result in making homologous sequences more similar, as they will lose their natural species codon bias. Sequences are aligned, and a consensus sequence is produced, optionally showing degenerate codons.
• Oligonucleotides are designed for synthetic construction of one or more corresponding synthetic nucleic acid for shuffling.
• Input parameters on oligonucleotide design include minimum and maximum lengths, minimum length of identical sequence at the ends, maximum degeneracy per oligonucleotide, length of oligo overlap, etc.
• an alternative to back-translation to achieve optimal codon usage for expression in a particular organism is to back-translate sequences to optimize nucleotide homology between family members. For example, amino acid sequences are aligned. All possible codons for each amino acid are determined and codons that minimize differences between the family of aligned sequences are chosen at each position.
• sequences of naturally occurring enzymes that catalyze similar or even identical reactions can vary widely: sequences may be only 50% identical or less.
• the method described here can also be applied to non-catalytic proteins (i.e. ligands such as cytokines) and even nucleic acid sequences (such as promoters that may be inducible by a number of different ligands), wherever multiple functional dimensions are encoded by a family of homologous sequences.
• libraries of variants can be prepared from a family of homologous natural sequences by DNA family shuffling. These libraries contain the diversity of the original set of sequences, in a large number of different combinations. If individuals from the library are then tested under a specific set of conditions for a particular property, the optimal combinations of sequences from the parental set for those conditions can be determined.
• sequence motifs have been identified, proteins are manipulated in, e.g., any of a number of ways. For example, identified changes are optionally deliberately introduced into other sequence backgrounds. Sequences conferring different specific properties can be combined. Identified sequence regions of importance for a specific function can be targeted for more thorough investigation, for example by complete randomization using degenerate oligonucleotides, e.g., selected by in silico processes. IDENTIFICATION OF PARENTAL CONTRIBUTORS TO CHIMERAS PRODUCED BY FAMILY SHUFFLING
• This example provides a method for the identification of parental contributors to chimeras produced by family shuffling.
• the method takes as an input the sequences of parental genes, and the sequences of chimeras, and compares each chimera with each parent. It then builds sequence and graphical maps of each chimera, indicating the parental source of each chimeric fragment. Correlation of this with functional data permits identification of parents that contribute to specific properties and thereby facilitates the selection of parents for new more focused libraries which can be made by any of the methods noted herein, and screened for any desired functional property.
• family 3 and 4 genes contribute to an activity at, e.g., pH 5.5, while family 1 and 2 genes are better at pH 10.
• the parental composition to create a library would be biased towards 3 and 4, while for high pH a predominantly 1 and 2-based library would be appropriate.
• a GO can be implemented which selects oligonucleotides for gene reconstruction predominantly from families 3 and 4.
• Principal component analysis is a mathematical procedure that transforms a number of (possibly) correlated variables into a (smaller) number of uncorrelated variables called "principal components.”
• the first principal component accounts for as much of the variability in the data as possible, and each succeeding component accounts for as much of the remaining variability as possible.
• principal component analysis is performed on a square symmetric SSCP (pure sums of squares and cross products) matrix, covariance (scaled sums of squares and cross products) matrix, or correlation (sums of squares and cross products from standardized data) matrix.
• the analysis results for objects of type SSCP and covariance are similar.
• a correlation object is used when the variances of individual variants differ substantially or the units of measurement of the individual variants differ.
• Objectives of principal component analysis include, e.g., to discover or to reduce the dimensionality of a data set, to identify new meaningful underlying variables, and the like.
• Partek (PCA) software discussed above has an "experimental design" component, that identifies variables that appear to have an effect on a specific function. As applied to the present example, this is useful in an iterative process in which a family library is constructed and screened and the resulting chimeras analyzed for functional correlations with sequence variations. This is used to predict sequence regions for a particular function, and a library is selected in silico by any desired GO directed changes of the region which correlates to functional activity. A focused library which has diversity in those regions is constructed, e.g., by oligonucleotide synthetic methods as described herein. The resulting library members (chimeras) are analyzed for functional correlations with sequence variations.
• This approach focuses the search of variation in sequence space on the most relevant regions of a protein or other relevant molecule. After sequences which are active are deconvoluted, the resulting sequence information is used to refine further predictions for in silico operations, e.g., in a neural net training approach.
• neural net approaches can be coupled to genetic algorithm- type programming, for example, NNUGA (Neural Network Using Genetic Algorithms) is an available program (http://www.cs.bgu.ac.il/ ⁇ omri/NNUGA/) which couples neural networks and genetic algorithms.
• NNUGA Neuron Using Genetic Algorithms
• An introduction to neural networks can be found, e.g., in Kevin Gurney (1999) An Introduction to Neural Networks, UCL Press, 1 Gunpowder Square, London EC4A 3DE, UK. and at http://www.shef.ac.uk/psychology/gurnev/notes/index.html.
• Additional useful neural network references include those noted above in regard to genetic algorithms and, e.g., Christopher M.
• One aspect of the present invention is the coupling of logical filtering mechanisms to nucleic acid or polypeptide sequences in silico and, e.g., random physical shuffling of logically "filtered" nucleic acids or polypeptides.
• logical filtering mechanisms to nucleic acid or polypeptide sequences in silico and, e.g., random physical shuffling of logically "filtered" nucleic acids or polypeptides.
• in silico approaches can be used to apply any desired criteria to selection of recombination events, which are optionally coupled to physical shuffling processes to generate selected, random or pseudo-random recombined physical sequences for subsequent activity selection.
• Three basic logical GO filters are discussed herein in detail, though others will be apparent to one of skill.
• structural considerations can be used to design logical filters which preserve or eliminate any structural feature of interest in a nucleic acid or encoded protein (as discussed herein, direct design and synthesis of recombinant proteins without nucleic acid intermediates can be performed, but, for simplicity of illustration, the following discussion generally describes the use of nucleic acids to generate proteins).
• structural criteria include e.g., energy minimization calculations, combinatorial automated protein design algorithms, sequence motifs, application of GAs or GOs to structural and/or sequence criteria, and, e.g., structural information based upon any available structural or modeling data.
• structural data is provided, e.g., as derived from physical protein (or nucleic acid) analysis (e.g., crystal structure, nmr, epr, circular dichrosim, intrinsic fluorescence, mass spectrometry, and any of the myriad other available structural analysis methods) or by any structural modeling method.
• physical protein (or nucleic acid) analysis e.g., crystal structure, nmr, epr, circular dichrosim, intrinsic fluorescence, mass spectrometry, and any of the myriad other available structural analysis methods
• modeling and physical structure information are optionally used in conjunction and that cycles of physical information analysis, modeling and application of this information in one or more GO in silico can be performed in conjunction.
• certain combinatorial protein design algorithms themselves rely on cycles of experimentation and design (such approaches are discussed in more detail below).
• Structural considerations can include, e.g., a logical filter (e.g., as part of any GO) which maintains or disrupts any interaction of two or more amino acids in a protein sequence.
• a logical filter e.g., as part of any GO
• the two or more amino acids will be selected to maintain their physical relationship and will, therefore, show co-variance in subsequent recombinants. That is, cross-overs or other GOs are selected by the filter such that the physical-structural relationship of the two or more amino acids is linked or maintained (or disrupted, if desired).
• physical linkages can include any sort of structural element maintenance or disruption, including maintaining or disrupting physical distance relationships, energy minimization relationships, etc. It is, of course, useful to maintain or disrupt particular structures to derive similar or fundamentally different nucleic acids or proteins from a set of parental nucleic acids or proteins.
• statistical analysis of empirical results can suggest that certain structural events or features are relevant, and can be used to refine modeling prediction and to elucidate structural analysis, and this information can be applied to the next cycle of GO application in silico to select additional cross-over points or to apply any other GO of interest.
• one aspect of the present invention is the use of any form of structural information in applying a logical filter to design cross-over points, or, for that matter, to design or modify any other GO as noted herein.
• These GOs include, without limitation, mutation of one or more parental character strings or one or more character string subsequences, multiplication of one or more parental character strings or one or more character string subsequences, fragmentation of one or more parental character strings or one or more character string subsequences, crossover between any of the one or more parental character strings or one or more character string subsequences or an additional character string, ligation of the one or more parental character strings or one or more character string subsequences, an elitism calculation, a calculation of sequence homology or sequence similarity of aligned strings, a recursive use of one or more genetic operator for evolution of character strings, application of a randomness operator to the one or more parental character strings or the one or more character string subsequences, a deletion mutation of
• any GO can include a component which maintains, removes, modifies, or in any way modulates one or more structural relationship during application of the GO in a GA.
• Protein design cycles involving cycling between theory and experiment, has led to recent advances in rational protein design (reviewed,, e.g., in Street and Mayo (1999) "Computational Protein Design” Structure with Folding and Design 7(5):R105- R109).
• Protein design programs can be used to build or modify proteins with any selected set of design criteria and these design criteria can be used as filters for any GO as noted herein, and/or in recursive cycles of design (by modeling or structural analysis), in silico GO or GA application, and/or physical recombination of nucleic acids of interest.
• linkages between amino acids or nucleotides in secondary and tertiary structures of proteins linkages can be assigned e.g., at the dipeptide level (or any other level) to provide structure/design criteria. Again, these linkages can be selected for or disrupted, depending on the desired outcome, in selection of cross-over events or other GOs, e.g., in silico, according to the present invention.
• Penalization for exposed hydrophobic surface structures also improved design performance. More recently, pairwise expressions with one scalable parameter have been used that reproduce true buried and true solvent-accessible areas (Street and Mayo (1998) "Pairwise Calculation of Protein Solvent- Accessible Surface Areas” Folding & Design 3: 253-258).
• steric and other constraints can be used to consider structural linkages between amino acids (and in protein structures generally) and these linkages can be used to set filtering criteria for any GO or GA of interest, e.g., to select appropriate cross-over points between sequences of interest.
• a feature of the invention optionally uses recursive cycles of design and experimentation, with observed results being used to improve both the activity of any protein of interest and the modeling tools used to predict cross-overs and to perform other GOs.
• thermophillic protein was increased by selecting residues for mutation based upon design algorithms.
• such combinatorial algorithms can be used to select desired improvements, e.g., based upon the functions of the algorithms and these desired improvements can be incorporated into one or more in silico sequence string to be recombined with any other sequence string, physically or by in silico methods.
• desired improvements e.g., based upon the functions of the algorithms and these desired improvements can be incorporated into one or more in silico sequence string to be recombined with any other sequence string, physically or by in silico methods.
• residues which the algorithm indicates as targets for mutation are modified in silico, and cross-overs between the resulting modified sequence(s) and any other sequence(s) are designed to maintain or disrupt the modifications in any subsequent recombination steps.
• these methods can be used to bias or modulate any GO (e.g., cross-over selection) to maintain or alter any such force field or, more simply, to maintain or alter residue any sequence which the algorithms indicate to be of interest. That is, residues which the algorithm(s) indicates as targets for maintenance or mutation are modified or maintained in silico, as noted herein.
• Cross-overs between the resulting sequence(s) and any other sequence(s), or any other GO of interest are designed to maintain or to disrupt the sequences of interest in any subsequent recombination step(s).
• designed or modified proteins or character strings corresponding to proteins can be directly shuffled in silico, or, e.g., reverse translated and shuffled in silico and/or by physical shuffling (that is, many design algorithms apply primarily to proteins, while recombination is conveniently performed between coding nucleic acids — though as noted herein, direct recombination between proteins, particularly in silico, can also be performed).
• one aspect of the invention is the coupling of high-throughput rational design and in silico or physical shuffling and screening of genes to produce activities of interest. It is possible, using the present methods, to couple high-throughput rational design and random or semi-random recombination methods.
• molecular dynamic simulations such as those above and, e.g., Ornstein et al. (http://www.emsl.pnl.gov:2080/homes/tms/bms.html; Curr Opin Struct Biol (1999) 9(4):509-13) provide for "rational" enzyme redesign by biomolecular modeling & simulation to find new enzymatic forms that would otherwise have a low probability of evolving biologically. For example, rational redesign of p450 cytochromes and alkane dehalogenase enzymes are a target of current rational design efforts.
• Any rationally designed protein e.g., new p450 homologues or new alkaline dehydrogenase proteins
• the dynamic simulations can be used as predictors of residues of interest and coupling or linkages between residues of interest and cross-over selection (or other GOs) can be performed to specifically maintain or eliminate such linkages in silico.
• homology modeling can also be used to provide structural predictors and to identify which residues are relevant to activity. See, Haney et al. (1997) "Structural basis for thermostability and identification of potential active site residues for adenylate kinases from the archaeal genus Methanococcus" Proteins 28(1): 117-30. As applied to the present invention, this combination of structure and sequence analysis can be used to identify putative relationships between amino acids in a structure and can, accordingly, be used in cycles of design, in silico GO application, recombination, etc. Here again, cross-over selection or other GOs of interest can be maintained or eliminated in silico.
• any structural information can be used to select cross-over sequences between nucleic acids (or to perform any other GO of interest). For example, comparison of protein crystal structures to predict crossover points based on structural rather than sequence homology considerations can be conducted and crossover can be effected by oligos to direct chimerization as discussed herein. This can be performed with or without the use of protein design algorithms, i.e., even simple inspection of crystal structure can provide a basis for selecting or eliminating particular residues. Thus, with knowledge about structure, either complex design algorithms or simple structural analysis can be used to select residues, secondary structures, tertiary structures, or the like, to be maintained or disrupted in any recombined coding nucleic acids.
• Multivariate data analysis and experimental design is widely applied in industry, government and research centers. It is typically used for things like formulating gasoline, or optimizing a chemical process. In the classic example of gasoline formulation, there may be more than 25 different additives that can be added in different amounts and in different combinations. The output of the final product is also multifactorial (energy level, degree of pollution, stability etc. etc.).
• experimental design a limited number of test formulations can be made where the presence and amounts of all additives are altered in a non-random fashion in order to maximally explore the relevant "formulation space.” The appropriate measurements of the different formulations are subsequently analyzed. By plotting the datapoints in a multivariate (multidimensional) fashion, the formulation space can be graphically envisioned and the ideal combination of additives can be extracted.
• PCA Principal Component Analysis
• this type of matrix is used to correlate each multidimensional datapoint with a specific output vector in order to identify the relationship between a matrix of dependant variables Y and a matrix of predictor variables X.
• a common analytical tool for this type of analysis is Partial Least Square Projections to Latent Structures (PLS). This, for instance, is often used in investment banker's analysis of fluctuating stock prices, or in material science predictions of properties of novel compounds.
• PLS Partial Least Square Projections to Latent Structures
• Nucleotide or amino acid sequence analysis has traditionally been concentrated on qualitative pattern recognition (e.g., sequence classification). This mainly involves identifying sequences based on similarity. This works well for predictions or identifications of classification, but does not always correlate with quantitative values. For instance, a consensus transcriptional promoter may not be a good promoter in a particular application, but is, instead, the average promoter among an aligned group of related sequences.
• Relative transcriptional level was measured by dot-blot using the vector derived ⁇ -lactamase gene as internal standard.
• the 28 promoters each with 68 nucleotides
• the result is a 28x204 matrix (28 promoters x 68 nucleotides x 3 parameters (steric bulk, hydrophobicity and polarizability)).
• the unique sequence of each promoter can be represented as a single point in a 204-dimensional hyperspace. This compilation of 28 promoters thus formed a cluster of 28 points in this space.
• nucleotides and amino acids are identical to nucleotides.
• quantitative descriptors have to be used to parameterize amino acids.
• the relevant features of the amino acids have been determined and can be extracted from the literature (Hellberg et al. (1986) Acta Chem. Scand. B40:135- 140; Jonsson et al, (1989) Quant. Struct. -Act. Relat. 8:204-209).
• the mutated proteins are analyzed, both those that are “better” and those that are “worse” than the initial sequences.
• Statistical tools such as PLS
• Cross-overs can be selected (or other GOs performed) that provide for generation of the extrapolated sequences.
• any protein encoded by a DNA sequence can be plotted as a distinct point in multidimensional space using statistical tools.
• a "normal" lkb gene can constitute, e.g., about 330 amino acids. Each amino acid can be described by, e.g., three major physico-chemical quantitative descriptors (steric bulk, hydrophobicity and polarizability) for each amino acid (other descriptors for proteins are largely dependent on these three major descriptors). See also, Jonsson et al. (1989) Quant. Struct.- Act. Relat. 8:204-209.
• a 1 kb gene is modeled in 330 (number of amino acids)
• X 20 possible amino acids at each position
• X 3 the three main descriptors noted above, for each amino acid
• a number of shuffled sequences are used to validate sequence activity related predictions. The closer the surrounding sequences are in space (percent similarity), the higher the likelihood that predictive value can be extracted. Alternatively, the more sequence space which is analyzed, the more accurate predictions become.
• This modeling strategy can be applied to any available sequence. As described above, cycles of design and experimentation can be used to refine the model. Alternatively, neural networks can be used to learn a type of pattern and predict the generated outcome of given variations.
• neural networks examples include Schneider and Wrede (1998) "Artificial neural networks for computer-based molecular design” Prog Biophvs Mol Biol 1998;70(3):175-222; Schneider et al. (1998) “Peptide design by artificial neural networks and computer-based evolutionary search” Proc Natl Acad Sci U S A 95(21): 12179-84; and Wrede et al. (1998) “Peptide design aided by neural networks: biological activity of artificial signal peptidase I cleavage sites” Biochemistry 37(11):3588-93.
• An overall issue for the above strategy is the availability of enough related sequences generated through shuffling to provide useful information.
• An alternative to shuffling sequences is to apply the modeling tools to all available sequences, e.g., the GenBank database and other " public sources. Although this entails massive computational power, current technologies make the approach feasible. Mapping all available sequences provides an indication of sequence space regions of interest.
• the information can be used as a filter which is applied to in silico shuffling events to determine which virtual progeny are preferred candidates for physical implementation (e.g., synthesis and/or recombination as noted herein).
• cross-overs and other GOs are selected to provide preferred recombinants and/or substrates for shuffling.
• Synthetic shuffling optionally uses bridging oligos to force crossovers at defined location(s) in order to generate chimeric progeny.
• bridging oligos to force crossovers at defined location(s) in order to generate chimeric progeny.
• the general concept of defining the location for forced crossovers on certain structural (e.g., primary, secondary or tertiary sequence based) or statistical considerations is discussed herein.
• a protein's need to fold into a native structure is a universal constraint on all proteins irrespective of its function.
• the sequence-function landscape of proteins is therefore overlaid on sequence-structure foldadility landscape.
• Genes that share very low sequence identity cannot be shuffled using conventional shuffling methods.
• Crossover recombination can be achieved using crossover oligos or other means described herein. This approach permits generation of a library of recombined sequences that are likely to be folded and therefore of better functional quality.
• structural super-families span sequences displaying low sequence similarity/ identity and can span a variety of functions.
• the present invention provides an effective way to engineer diversity at sequence, structure and function levels.
• Parameters that can be used for determining the appropriate crossover locations include, but are not limited to: structural stability, allosteric effects, 3-D energetic constraints, structural symmetry, regions of hydrophobicity and patterns in the distribution of hydrophobic residues in 1-D, 2-D and 3-D, periodicity of residues and residue properties in 1-D, 2-D and 3-D, distribution of charged residues in all dimensions and electrostatic fields, complexity of sequences and structure, information and mutual information content in sequences and co-variation of residues, evolutionary rates over local regions and entire proteins, codon usage, distribution and GC biases, synonymous and non-synonymous substitution rates, motifs distribution along the 1-D sequence and 3-D structure, orientations of primary, secondary, tertiary and quaternary structural units, analysis of structural, sequence and evolutionary domains, 3-D motifs, structural pockets,
• relevant considerations include: hydrophobicity as determined by ⁇ G of transfer from solvents, hydrophilicity from column retention, charges on the amino acids, polarity, pka of the amino-acids, bulkiness, side chain entropy, alpha helix/beta sheet propensities, hydration potentials, codon degeneracies and the like.
• crossover points For example, the above can be considered to select for structural integrity and functional conservation in recombined proteins.
• One advantage to this approach over standard "wet" laboratory approaches is that in-silico libraries can be generated and analyzed on a large scale. Recombined libraries based on designed crossover points can be screened and the information used to provide feedback to generate 2 nd generation crossovers and recombined libraries.
• One alternate to using purely structure-function design-based criteria for cross-over selection is the use of statistical matrixes such as Markov chains to pinpoint ideal crossover locations.
• the following example presents two alternative criteria for defining crossover locations based on statistical considerations.
• Co-variation of amino acids during evolution allows proteins to retain a given fold or function while altering other traits, such as specificity.
• the co-variation identified in a large data set can be useful in addressing possible crossover locations as it identifies co-evolving amino acids in a given family and allows bridging oligos to be engineered so that such functional constraints are retained.
• MI Mutual Information
• MIx Y ⁇ ⁇ P(Xi, Y j )log n PJXi.Yi) , i j P(Xi)P(Yj)
• P(Xj) is the probability of I at site X
• P(Y j ) is the probability of j at site Y
• P(X ⁇ ,Y j ) is the joint probability of i at site X and j at site Y (X ⁇ Y).
• the maximum MI value occurs when the variation at two sites is perfectly correlated.
• the data presented by Wollenberg and Atchley uses the sequence elements as symbol variables with no underlying metric. The same statistical correlation can be done using the principal components for each amino acid. Replacing the symbols with true physicochemical properties can improve the significance of the MI relationship.
• a statistical acceptability threshold permits the identification, within a quantifiable error, of those intersite associations most probably arising from structural/functional causes.
• the reminder of the intersite associations can be deemed originating from phylogenetic resemblance.
• Cross-overs or other GOs are selected to produce the intersite associations of interest.
• the present invention provides for the shuffling of "evolutionary intermediates."
• evolutionary intermediates are artificial constructs which are intermediate in character between two or more homologous sequences, e.g., when the sequences are grouped into an evolutionary dendogram.
• Nucleic acids are often classified into evolutionary dendograms (or "trees") showing evolutionary branch points and, optionally, relatedness.
• cladistic analysis is a classification method in which organisms or traits (including nucleic acid or polypeptide sequences) are ordered and ranked on a basis that reflects origin from a postulated common ancestor (an intermediate form of the divergent traits or organisms).
• Cladistic analysis is primarily concerned with the branching of relatedness trees (or "dendograms") which shows relatedness, although the degree of difference can also be assessed (a distinction is sometimes made between evolutionary taxomomists who consider degrees of difference and those who simply determine branch points in an evolutionary dendogram (classical cladistic analysis); for purposes of the present invention, however, relatedness trees produced by either method can produce evolutionary intermediates).
• Cladistic or other evolutionary intermediates can be determined by selecting nucleic acids which are intermediate in sequence between two or more extant nucleic acids.
• sequence may not exist in nature, it still represents a sequence which is similar to a sequence in nature which had been selected for, i.e., an intermediate of two or more sequences represents a sequence similar to the postulated common ancestor of the two or more extant nucleic acids.
• evolutionary intermediates are one preferred shuffling substrate, as they represent "pseudo selected" sequences, which are more likely than randomly selected sequences to have activity.
• evolutionary intermediates as substrates for shuffling (or of using oligonucleotides which correspond to such sequences) is that considerable sequence diversity can be represented in fewer starting substrates (i.e., if starting with parents A and B, a single intermediate "C" has at least a partial representation of both A and B). This simplifies the oligonucleotide synthesis scheme for gene reconstruction/ recombination methods, improving the efficiency of the procedure. Further, searching sequence databases with evolutionary intermediates increases the chances of identifying related nucleic acids using standard search programs such as BLAST.
• Intermediate sequences can also be selected between two or more synthetic sequences which are not represented in nature, simply by starting from two synthetic sequences.
• Such synthetic sequences can include evolutionary intermediates, proposed gene sequences, or other sequences of interest that are related by sequence. These "artificial intermediates" are also useful in reducing the complexity of gene reconstruction methods and for improving the ability to search evolutionary databases.
• character strings representing evolutionary or artificial intermediates are first determined using alignment and sequence relationship software and then synthesized using oligonucleotide reconstruction methods.
• the intermediates can form the basis for selection of oligonucleotides used in the gene reconstruction methods herein.
• HMM Hidden Markov Model
• the HMM matrix shown in Fig. 15 exemplifies a family of 8 amino acid peptides.
• the peptide can be a specific amino acid (one of the 20 present in the boxes), an insertion (diamonds), or a deletion (circles).
• the probability for each to occur is dependent on how often it occurs among the compiled parents. Any given parent can subsequently be 'threaded' through the profile in such way that all allowed paths are given a probability factor.
• HMM can be used in other ways as well.
• the HMM profile can be used as a template to generate de novo family members (e.g., intermediate members of a cladistic tree of nucleic acids).
• the program, HMMER is available (http://hmmer.wustl.edu/). This program builds a HMM profile on a defined set of family members.
• a sub-program, HMMEMtT reads the profile and constructs de novo sequences based on that.
• the original purpose of HMMEMIT is to generate positive controls for the search pattern, but the program can be adapted to the present invention by using the output as in silico generated progeny of a HMM profile defined shuffling.
• oligonucleotides corresponding to these nucleic acids are generated for recombination, gene reconstruction and screening.
• sequence context of each position is accounted for in a probabilistic fashion, the number of non-active progeny is significantly lower than a shuffling reaction that simply randomly selects such progeny.
• Crossover between genetic modules occur where they occur in nature (i.e. among the parents) and co-evolution of point mutations or structural elements is retained throughout the shuffling process.
• the following is an outline for a program for generating sequence intermediates from alignments of related parental nucleic acids.
• One aspect of the present invention is to use positive or negative data in sequence design and selection methods, either in silico, or in physical processing steps, or both.
• the use of positive or negative data can be in the context of a learning heuristic, a neural network or by simply using positive or negative data to provide logical or physical filters in design and library synthesis processes. Learning networks are described supra, and provide one convenient way of using positive or negative data to increase the chances that additional sequences which are subsequently generated will have a desired activity.
• the ability to use negative data to reduce the size of libraries to be screened provides a considerable advantage, as screening is often a limiting step in generating improved genes and proteins by forced evolution methods.
• the use of positive data to bias libraries towards sequences of interest is another way of focusing libraries.
• positive or negative data can be used to provide a physical or logical "filter" for any system of interest. That is, sequences which are shown to be inactive provide useful information about the likelihood that closely related sequences will also prove to be inactive, particularly where active sequences are also identified. Similarly, sequences which are active provide useful information about the likelihood that closely related sequences will also prove to be active, particularly where inactive sequences are also identified. These active or inactive sequences can be used to provide a virtual or physical filter to bias libraries (physical or virtual) toward production of more active members.
• hybridization rules or other parameters can be used to select against members that are likely to be similar to inactive sequences.
• oligonucleotides used in gene reconstruction methods can be biased against sequences which have been shown to be inactive.
• libraries or character strings are filtered by subtracting the library or set of character strings with members of an initial library of biological polymers which display activity below a desired threshold.
• hybridization rules or other parameters can be used to select for members that are likely to be similar to active sequences. For example, oligonucleotides used in gene reconstruction methods can be biased towards sequences which have been shown to be active.
• libraries or character strings are filtered by biasing the library or set of character strings with members of an initial library of biological polymers which display activity above a desired threshold.
• in silico approaches can be used to produce libraries of inactive sequences, rather than active sequences. That is, inactive sequences can be shuffled in silico to produce libraries of clones that are less likely to be active. These inactive sequences can be physically generated and used to subtract libraries (typically through hybridization to library members) generated by other methods. This subtraction reduces the size of the library to be screened, primarily through the elimination of members that are likely to be inactive.
• each contiguous selected region or motif e.g., the selected window can be, e.g., a 20 base region
• the selected window can be, e.g., a 20 base region
• Second generation libraries are synthesized in which the library is selected to be enriched for good motifs and depopulated with bad motifs, using any filtering or learning process as set forth herein. A variety of methods for measuring frequencies of motifs in populations of genes are available.
• hybridization to membranes containing spatially addressable motifs and measuring relative signal intensities for probe before and after selection can also be performed in an essentially similar fashion, e.g., using standard Southern or northern blot methods. Relative ratios of identified desirable/undesirable features on the chips also provides an indication of overall library quality.
• phage display or other expression libraries can be used to assess library features, i.e., by evaluating expression products.
• real time quantitative PCR can be performed where PCR oligos are highly discriminating for the feature of interest. This can be done by, for example, having a polymorphism unique to the motif present at or near the 3' end of an oligo such that it will only prime the PCR efficiently if there is a perfect match.
• Real time PCR product analysis by, e.g., FRET or TaqMan (and related real time reverse-transcription PCR) is a family of known techniques for real time PCR monitoring that has been used in a variety of contexts (see, Laurendeau et al.
• sequencing primers up and down the gene are used, then one can look at the sequences in parallel on a sequencing gel.
• the sequence polymorphisms near the primer can be read out to see the relative ratio of bases at any given site. For example, if the population starts out with 50%T and 50% C at a given position, but 90% T and 10% after selection, one could easily quantitate this base ratio from a sequencing run that originates near the polymorphism. This method is limited, because as one gets further from the primer and reads through more polymorphisms, the mobilities of the various sequences gets increasingly variable and the traces begin to run together. However, as the cost of sequencing continues to decline and the cost of oligos continues to decrease, one solution is simply to sequence with many different oligos up and down the gene.
• Typical sequence spaces are very large compared to the number of sequences that can be physically cloned and characterized.
• Methods for fractionating sequence space such that it is enriched for molecules which are predicted by a given model to have greater or lesser fitness with respect to a phenotype of interest would be useful for testing such predictive models.
• shuffled TEN genes typically on the order of 10 - 10
• convert them to ssDNA pass them over an affinity column consisting of an oligonucleotide complementary human IFN alphal over these residues, wash under appropriate stringency, elute the bound molecules, PCR amplify the eluted genes, clone the material, and perform functional tests on the expressed clones.
• This protocol allows one to physically bias a library of shuffled genes strongly in favor of containing this motif which is predicted by this very simple model to confer an improvement in the desired activity.
• affinity electing oligonucleotides that encode motifs of interest prior to gene recombination/resynthesis reduces the diversity of populations of nucleic acids that are produced in gene recombination/ resynthesis methods as noted herein.
• oligonucleotides encoding motifs can be selected by enzymatically degrading molecules that are not perfectly matched with the oligos, e.g., again prior to gene recombination/ resynthesis methods.
• genes that match imperfectly with the oligonucleotides can be selected for, e.g., by binding to mutS or other DNA mismatch repair proteins.
• Polymerization events during recombination/gene synthesis protocols can be primed using one or more oligos encoding the motif(s) of interest. That is, mismatches at or near the 3' end of hybridized nucleic acids reduce or block elongation. In this variation, only newly polymerized molecules are allowed to survive (used in subsequent library construction/ selection steps). This can be done, for example, by priming reverse transcription of RNA and then degrading the RNA.
• DNA with a high frequency of uracil incorporation can be synthesized.
• Polymerase-based synthesis is primed with oligos and extended with dNTPs containing no uracil.
• the resulting products are treated with uracil glycosylase and a nuclease that cleaves at apurinic sites, and the degraded template removed.
• RNA nucleotides can be incorporated into DNA chains (synthetically or via enzymatic incorporation); these nucleotides then serve as targets for cleavage via RNA endonucleases.
• cleavable residues are known, including certain residues which are targets for enzymes or other residues and which serve as cleavage points in response to light, heat or the like.
• residues which are targets for enzymes or other residues and which serve as cleavage points in response to light, heat or the like.
• polymerases are currently not available with activity permitting incorporation of a desired cleavage target, such polymerases can be produced using shuffling methods to modify the activity of existing polymerases, or to acquire new polymerase activities.
• Localized motifs can easily be translated into affinity selection procedures. However, one sometimes wants to impose a rule that molecules have multiple sequence features that are separated in space in the gene (e.g., 2, 3, 4, 5, 6, etc. sequence features). This can be timed into a selection by making a nucleic acid template that contains all motifs of interest separated by a flexible linker. The T m for molecules having all motifs is greater than for molecules having only one or two of the motifs. It is, therefore, possible to enrich for molecules having all motifs by selecting for molecules with high T m s for the selecting oligo(s). A "gene" of many such motifs strung together can be synthesized separated by flexible linkers or by bases such as inosine that can base pair promiscuously.
• Careful design of the selecting nucleic acid template allows one to enrich for genes having a large number of sequence motifs that are predicted to bias genes containing them toward having a phenotype of interest.
• a set of motifs is defined, e.g., based upon sequence conservation between different homologs, or the motifs can even be randomly selected motifs. As long as the sequence space is not isotropic (equally dense with good members in all directions), then one can simply fractionate the sequence space based on a designed or on a random set of motifs, measure the average fitness of clones in the region of sequence space of interest, and then prospect more heavily in the regions that give the highest fitness.
• Rational design can be used to produce desired motifs in sequences or sequence spaces of interest. However, it is often difficult to predict whether a given designed motif will be expressed in a functional form, or whether its presence will affect another property of interest.
• An example of this is the process of designing glycosylation sites into proteins such that they are accessible to cellular glycosylation machinery and such that they do not negatively affect other properties of the protein such as blocking binding to another protein by virtue of steric hindrance by the attached polysaccaride groups.
• One way of addressing these issues is to design motifs or multiple variations of motifs into multiple candidate sites within the target gene.
• the sequence space is then screened or selected for the phenotype(s) of interest.
• Molecules that meet the specified design criterion threshold are shuffled together, recursively, to optimize properties of interest.
• Motifs can be built into any gene.
• Exemplary protein motifs include: N- linked glycosylation sites (i.e. Asn-X-Ser), 0-linked glycosylation sites (i.e. Ser or Thr), protease sensitive sites (i.e. cleavage by collagenase after X in P-X-G-P) Rho-dependent transcriptional termination sites for bacteria, RNA secondary structure elements that affect the efficiency of translation, transcriptional enhancer elements, transcriptional promoter elements, transcriptional silencing motifs, etc.
• high throughput rational design methods are also useful.
• high throughput rational design methods can be used to modify any given sequence in silico, e.g., before recombination/ synthesis.
• PDA Protein Design Automation
• PDA starts with a protein backbone structure and designs the amino acid sequence to modify the protein's properties, while maintaining it's three dimensional folding properties.
• Large numbers of sequences can be manipulated using PDA, allowing for the design of protein structures (sequences, subsequences, etc.).
• PDA is described in a number of publications, including, e.g., Malakauskas and Mayo (1998) "Design, Structure and Stability of a Hyperthermophilic Protein Variant” Nature Struc. Biol. 5:470; Dahiyat and Mayo (1997) “De Novo Protein Design: Fully Automated Sequence Selection” Science, 278, 82-87. DeGrado, (1997) “Proteins from Scratch” Science.
• PDA and other design methods can be used to modify sequences in silico, which can be synthesized/ recombined in shuffling protocols as set forth herein.
• PDA and other design methods can be used to manipulate nucleic acid sequences derived following selection methods.
• design methods can be used recursively in recursive shuffling processes.
• one way to increase the number of crossover sites is to build "bridging" or "cross-over" oligonucleotides that have some number of bases (e.g., sufficient to mediate hybridization, typically, e.g., about 20 or more, though fewer can also be used) identical to one parent, then some number (typically, e.g., also about 20 or more bases) identical to the second parent.
• some number of bases e.g., sufficient to mediate hybridization, typically, e.g., about 20 or more, though fewer can also be used
• bridging oligos there are at least four basic ways of providing bridging oligos, i.e., 1) arbitrarily choosing bridging oligos; 2) making all possible bridging oligos; 3) making a subset of possible bridging oligonucleotides based upon known features of the sequences to be shuffled (e.g., to permit particular domains or subdomains to be recombined or disrupted, e.g., to account for 3-D or empirically derived considerations), and 4) using computational methods to optimize bridging oligonucleotide selection and design.
• This last option is especially useful in the context of the present invention, although the first three approaches can also be applicable.
• this section uses shuffling of p450 super family members to illustrate certain methods of performing relevant steps of the invention; however, the steps are universally applicable to other systems.
• the methods are applicable to any set or sets of proteins for which more than one sequence is available and for which at least one structure is either known or can be calculated/ estimated.
• desirable crossover points can be selected between two or more sequences, e.g., following an approximate sequence alignment, performing Markov chain modeling, or any other desired selection method. This desirable subset of all possible crossover points is smaller than all possible crossover points, and all possible recombinants between two or more low homology sequences using these reasonable crossover points can be computationally generated.
• Steps 2-5 are optionally omitted and only bridging oligonucleotides fulfilling given criteria (retaining certain substructures, statistical considerations, etc.) are used for subsequent library generation).
• PDA is one particularly useful computationally driven system for the design and optimization of proteins and peptides, as well as for the design of proteins and peptides.
• Many other energy minimization and other protein/nucleic acid design algorithms are known and available. These include computer search algorithms in protein modification and design (Desjarlais and Clarke (1999) Structure Fold Des ;7(9): 1089-98), Theoretical and algorithmical optimization of the dead-end elimination theorem (Desmet et al.
• Genetic algorithms can be used to solve problems in protein threading, both in this context and generally in relation to the present invention (see also, Yadgari et al. (1998) Ismb 6:193-202). In protein threading approaches, a sequence is aligned and identified to the fold with which it is most compatible (this process is often referred to in the literature as protein "threading"). Genetic algorithms can be used to solve such problems, which are not expected to have simple polynomial solutions. See also, Yadgari et al, id. 4. Use protein design algorithms to assess the ability of each in silico recombinant to fold into structures resembling the template structures.
• a filtering criteria e.g., stability, structural similarity of one or more region of the putative cross-over construct to one or more parental, desired regions of hydrophobicity or hydrophilicity, the ability of selected regions to form particular secondary, tertiary or even quatranery structures, or any other useful criteria
• a filtering criteria e.g., stability, structural similarity of one or more region of the putative cross-over construct to one or more parental, desired regions of hydrophobicity or hydrophilicity, the ability of selected regions to form particular secondary, tertiary or even quatranery structures, or any other useful criteria
• Oligonucleotide design is optionally further refined, for example, to ensure that a disallowed crossover is not facilitated, or that all crossover oligos have identical melting temperatures to each parent (and to all other crossover oligos generated).
• desirable primers optionally incorporate any of several useful properties.
• primers include, inter alia, that the hybridization of the primers to their complementary sequences is uniform; that individual primers hybridize only to their complementary regions in the system, and do not significantly cross hybridize with primers complementary to other parental sequence regions; that if there are selected regions associated with the primers that are not complementary to a target (e.g., cloning sites, secondary PCR primer binding sites, etc.) that the selected regions do not hybridize to a corresponding probe set, etc.
• a target e.g., cloning sites, secondary PCR primer binding sites, etc.
• One available computer program for primer selection is the Mac VectorTM program from Kodak.
• An alternate program is the MFOLD program (Genetics Computer Group, Madison WI) which predicts secondary structure of, e.g., single-stranded nucleic acids.
• MFOLD program Genetics Computer Group, Madison WI
• sequence information from selected recombinants can be used to provide information for bridging oligonucleotide design for subsequent steps.
• any selected recombinant can be used in any diversity generation reaction according to any other protocol herein (e.g., shuffling, mutagenesis, or any in silico procedure herein).
• the types of enzymatic activity that arise as a result of this overall process depend upon the recombinants chosen and the template structure screened against. For example, there is putative structural homology between Bacillus megaterium P450BM-3 and murine nitric oxide synthase protein (see Degtyarenko, K.N. & Archakov, A.I.
• Novel P450s can be computationally constructed by making all possible recombinants between the nitric oxide synthase N-terminus and the entire P450, following comparison with a P450 template structure to eliminate undesired sequences.
• Libraries of synthesized variants are constructed that can be tested for alterations in various P450 properties, e.g., substrate range, acceptable electron donors, etc.
• novel nitric oxide synthases are computationally constructed by making all possible recombinants between the N-terminal domain of the P450 and the entire nitric oxide synthase and comparing putative recombinants resulting from the cross-overs with a nitric oxide template structure for one or more property (energy minimization, stability, etc., as noted above). Desirable sequences are constructed and functionally screened. For example, in this particular example, a functional screen on these enzymes can screen, e.g., for a different property rate of electron transfer and/or nitric oxide synthesis.
• step 2 other ways of computationally generating new protein sequences can be used. For example incremental truncation for the creation of hybrid enzymes (ITCHY), circular permutations (with or without additional protein sequence insertions), sequence duplications, complementation, and many other techniques are applicable. Any approach that takes naturally occurring or artificial sequences and recombines them in silico can be used. For example, Ostermeier et al. "A combinatorial approach to hybrid enzymes independent of DNA homology" (1999) Nat Biotechnol 17(12): 1205-9 describe incremental truncation for the creation of hybrid enzymes (ITCHY), that creates combinatorial fusion libraries between genes in a manner that is independent of DNA homology.
• ITCHY and DNA shuffling to create interspecies or intraspecies fusion libraries between fragments of genes with little or no homology can identify a more diverse set of active fusion points including those in regions of nonhomology and those with crossover points that diverge from a sequence alignment.
• GFPs For example, several rearrangements of GFPs, in which the amino and carboxyl portions are interchanged and rejoined with a short spacer connecting the original termini, still become fluorescent. These circular permutations have altered pKa values and orientations of the chromophore with respect to a fusion partner. Furthermore, certain locations within GFP tolerate insertion of entire proteins, and conformational changes in the insert can have profound effects on the fluorescence. For example, insertions of calmodulin or a zinc finger domain in place of Tyr-145 of a yellow mutant (enhanced yellow fluorescent protein) of GFP result in indicator proteins whose fluorescence can be enhanced upon metal binding.
• the calmodulin graft into enhanced yellow fluorescent protein can monitor cytosolic Ca(2+) in single mammalian cells.
• the tolerance of proteins for circular permutations and insertions shows that the folding process is robust and offers a general strategy for creating new diversified sequences, including completely non-homologous sequences which are joined by bridging oligonucleotides according to the present invention.
• active sequences are produced by complementation.
• Yang and Schachman (1993) “In vivo formation of active aspartate transcarbamoylase from complementing fragments of the catalytic polypeptide chains” Protein Sci 2(6): 1013-23 and, e.g., Yang and Schachman (1996) "A bifunctional fusion protein containing the maltose-binding polypeptide and the catalytic chain of aspartate transcarbamoylase: assembly, oligomers, and domains" Biophys Chem 59(3):289-97 describe formation of an active stable enzyme in vivo, even with fragmented catalytic chains.
• domains or other subsequences can be separately designed and synthesized and tested for complementary effects with one another. Because the domains/sequences are on separate chains, complementation can be used to assess combinatorial effects of the separate domains. This approach can increase the number of sequence combinations that are assessed for activity, without increasing the number of sequences that are actually synthesized.
• Incomplete protein fragments or putative protein domains can also be included in any recombination reaction. Domains can be approximately identified, or even arbitrarily designated, e.g., as described supra. Additional details regarding protein domain designation/ identification are found, e.g., in Crameri et al. "OLIGONUCLEOTIDE MEDIATED NUCLEIC ACID RECOMBINATION" Filed January 18, 2000; USSN: PCT/USOO/01203.
• the six different forms of p450s include prokaryotic and eukaryotic 3 part systems with separate FAD-containing, Fe-S and P450 (heme-containing) subunits; prokaryotic and eukaryotic 2 part systems with a combined FAD- and FMN-containing subunit and a P450 (heme-containing) subunit; at least one prokaryotic one-component system (Bacillus megaterium P450BM-3) with one polypeptide containing FAD- FMN- and heme- group, and at least one eukaryotic one-component system (murine nitric oxide synthase) with one polypeptide containing FAD- FMN-and heme- group.
• prokaryotic and eukaryotic 3 part systems with separate FAD-containing, Fe-S and P450 (heme-containing) subunits prokaryotic and eukaryotic 2 part systems with a combined FAD- and FMN-containing subunit and a P450 (heme-containing)
• the combined FAD- and FMN-containing subunit from either prokaryotic or eukaryotic 2-component systems are optionally computationally recombined with the FAD- FMN-containing domain of either of the one-component systems, and then tested against a one-component template structure.
• the P450 subunits from either the 2 or 3 component systems can be computationally recombined with the heme-containing domain of the one-component system and then tested against a one-component template structure.
• a domain from the one-component system can computationally recombined with a subunit from the multi-component systems, and screened in silico for one or more computed property, e.g., against a structure of one of the subunits as a template.
• Potential sequences for recombination can also be identified by computational methods other than by direct homology or structural information, e.g.
• Pegg and Babbitt provide an example of using the Shotgun program to identify both new superfamily members and to reconstruct known enzyme superfamilies, using BLAST database searches.
• An analysis of the false-positive rates generated in the analysis and other control experiments show that high Shotgun scores indicate evolutionary relationships. Shotgun is also a useful tool for identifying subgroup relationships within superfamilies and for testing hypotheses about related protein families.
• PRINTS is a compendium of protein motif fingerprints derived from the OWL composite sequence database. Fingerprints are groups of motifs within sequence alignments whose conserved nature allows them to be used as signatures of family membership. Fingerprints can provide improved diagnostic reliability over single motif methods by virtue of the mutual context provided by motif neighbors.
• the database is now accessible via the UCL Bioinformatics Server on http: @ www.biochem.ucl.ac.uk/bsm/dbbrowser/. Atwood et al. describe the database, its compilation and interrogation software, and its Web interface. See also, Attwood et al. (1997) "Novel developments with the PRINTS protein fingerprint database" Nucleic Acids Res 25(l):212-7.
• Crossover points can also be "calculated” by simply comparing the structures (either from crystals, nmr, dynamic simulations, or any other available method) of proteins corresponding to nucleic acids to be recombined. All possible pairwise combinations of structures can be overlaid. Amino acids can be identified as possible crossover points when they overlap with each other on the parental structures
• Variations on the basic approach noted above can also be performed by varying calculation of structural information. Variations include the following.
• the entire process can be conducted at the level of sequence domains, e.g., by identifying domains in a structure and then identifying homologous domains for each target domain. Calculations can be made for the ability of in silico recombinants to form a complete folded protein, or for any given domain to form a structural domain.
• the FAD- and FMN-containing domain and the P450 (heme containing) domain can be treated as structurally separable.
• the FAD- and FMN-containing domain can be computationally recombined with the FAD- and FMN-containing domain of murine nitric oxide synthase, the FAD- and FMN-containing subunit of 2-component systems, and the FAD-containing subunit of 3-component systems.
• These can be fit either to a complete P450BM-3 template structure, or to a template structure that only includes the FAD- and FMN-containing domain and is not constrained by the presence of the heme- containing domain.
• this invention provides, inter alia, a process by which variants produced computationally can be synthesized physically, without the need for massive parallel synthesis of each individual calculated gene.
• Parental genes can be cloned or synthesized and then fragmented (e.g., by the uracil cleavage methods noted supra, e.g., to allow for multi-library formats as noted below), or fragments of the parents can simply be synthesized. The fragments are mixed with the relevant crossover oligos, assembled, expressed and assayed.
• each oligonucleotide is sufficient for each crossover oligonucleotide.
• Many different progeny can then be synthesized from gene fragments and crossover oligonucleotides simply by assembling different combinations of parents and oligonucleotides. The actual synthesis is performed either as each individual variant synthesized independently, or in increasing pool sizes up to a single library of all variants, e.g., by using a 96 -well parallel format (or other common screening format). For example, the following steps can be used in one basic format of this method. 1.
• Each variant is synthesized, e.g., in a single microtitre well, e.g., by including two parents and one bridging oligonucleotide.
• crossover oligonucleotides are synthesized, representing different pairs of parents. For example, for 5 parents, there are 10 different pairwise combinations of crossover oligonucleotides.
• Combinations of the above steps can be performed, separately or individually and in different orders.
• permutated or truncated parental sequences it is useful to synthesize, clone or otherwise construct the truncated or permutated parental sequences, as well as the bridging oligonucleotides.
• the method can also incorporate a variety of variations to produce an iterative process.
• the following are example variations.
• Improved variants from the first round of recombination can be shuffled by any of the processes noted herein, including use of bridging or cross-over oligonucleotides.
• first round improved variants can be computationally fitted to the parental structures used, and the improved structures can then be calculated, e.g., by energy minimization.
• New structures of the improved variants are, e.g., used as template structures for second round calculated recombinants. In this way it is also possible to gain structural understanding of functional changes which are obtained. Neural net and/or other statistical approaches can be used to further refine second round recombinants.
• many of the methods of the invention involve generating diversity in sequence strings in silico, followed by oligonucleotide gene recombination/ synthesis methods.
• non-oligonucleotide based recombination methods are also appropriate.
• entire genes can be made which correspond to any diversity created in silico, without the use of oligonucleotide intermediates. This is particularly feasible when genes are sufficiently short that direct synthesis is possible.
• solid phase polypeptide synthesis can be performed.
• solid phase peptide arrays can be constructed by standard solid phase peptide synthesis methods, with the members of the arrays being selected to correspond to the in silico generated sequence strings.
• solid phase synthesis of biological polymers, including peptides has been performed at least since the early "Merrifield" solid phase peptide synthesis methods, described, e.g., in Merrifield (1963) J. Am. Chem. Soc. 85:2149-2154 (1963).
• Solid-phase synthesis techniques are available for the synthesis of several peptide sequences on, for example, a number of "pins.” See e.g., Geysen et al. (1987) L Immun. Meth. 102:259-274, incorporated herein by reference for all purposes.
• Other solid-phase techniques involve, for example, synthesis of various peptide sequences on different cellulose disks supported in a column. See, Frank and Doling (1988) Tetrahedron 44:6031-6040.
• Still other solid-phase techniques are described in U.S. Patent No. 4,728,502 issued to Hamill and WO 90/00626. Methods of forming large arrays of peptides are also available. For example, Pirrung et al., U.S.
• Patent No. 5,143,854 and Fodor et al. PCT Publication No. WO 92/10092, disclose methods of forming arrays of peptides and other polymer sequences using, for example, light- directed synthesis techniques. See also, Stewart and Young, Solid Phase Peptide Synthesis, 2d. ed., Pierce Chemical Co. (1984); Atherton et al. (1989) Solid Phase Peptide Synthesis, 1P.L Press, Greene, et al. (1991) Protective Groups In Organic Chemistry, 2nd Ed., John Wiley & Sons, New York, NY and Bodanzszyky (1993) Principles of Peptide Synthesis second edition Springer Verlag, Inc. NY. Other useful information regarding proteins is found in R.
• character string diversity generated in silico can be corresponded to other biopolymers.
• the character strings can be corresponded to peptide nucleic acids (P ⁇ As) which can be synthesized according to available techniques and screened for activity in any appropriate assay. See, e.g., Peter E. Nielsen and Michael Egholm (eds) (1999) Peptide Nucleic Acids: Protocols and Applications ISBN 1-898486-16-6 Horizon Scientific Press, Wymondham, Norfolk, U.K for an introduction to PNA synthesis and activity screening.
• P ⁇ As peptide nucleic acids
• Directed Evolution by GAGGS can use any physical assays known in the art for detecting polynucleotides encoding desired phenotypes.
• Synthetic genes are amenable to conventional cloning and expression approaches; thus, properties of the genes and proteins they encode can readily be examined after their expression in a host cell. Synthetic genes can also be used to generate polypeptide products by in-vitro (cell-free) transcription and translation. Polynucleotides and polypeptides can thus be examined for their ability to bind a variety of predetermined ligands, small molecules and ions, or polymeric and heteropolymeric substances, including other proteins and polypeptide epitopes, as well as microbial cell walls, viral particles, surfaces and membranes.
• many physical methods can be used for detecting polynucleotides encoding phenotypes associated with catalysis of chemical reactions by either polynucleotides directly, or by encoded polypeptides. Solely for the purpose of illustration, and depending on specifics of particular pre-determined chemical reactions of interest, these methods may include a multitude of techniques well known in the art which account for a physical difference between substrate(s) and product(s), or for changes in the reaction media associated with chemical reaction (e.g. changes in electromagnetic emissions, adsorption, dissipation, and fluorescence, whether UV, visible or infrared (heat).
• These methods can be selected from any combination of the following: mass-spectrometry; nuclear magnetic resonance; isotopically labeled materials, partitioning and spectral methods accounting for isotope distribution or labeled product formation; spectral and chemical methods to detect accompanying changes in ion or elemental compositions of reaction product(s) (including changes in pH, inorganic and organic ions and the like).
• Other methods of physical assays, suitable for use in GAGGS can be based on the use of biosensors specific for reaction product(s), including those comprising antibodies with reporter properties, or those based on in vivo affinity recognition coupled with expression and activity of a reporter gene.
• Enzyme-coupled assays for reaction product detection and cell life-death-growth selections in vivo can also be used where appropriate. Regardless of the specific nature of the physical assays, they all are used to select a desired property, or combination of desired properties, encoded by the GAGGS -generated polynucleotides. Polynucleotides found to have desired properties are thus selected from the library.
• the methods of the invention optionally include selection and/or screening steps to select nucleic acids having desirable characteristics.
• the relevant assay used for the selection will depend on the application. Many assays for proteins, receptors, ligands and the like are known. Formats include binding to immobilized components, cell or organismal viability, production of reporter compositions, and the like.
• each well of a microtiter plate can be used to run a separate assay, or, if concentration or incubation time effects are to be observed, every 5-10 wells can test a single variant (e.g., at different concentrations).
• a single standard microtiter plate can assay about 100 (e.g., 96) reactions. If 1536 well plates are used, then a single plate can easily assay from about 100- about 1500 different reactions.
• cells, viral plaques, spores or the like, comprising GAGGS shuffled nucleic acids are separated on solid media to produce individual colonies (or plaques).
• an automated colony picker e.g., the Q-bot, Genetix, U.K.
• colonies or plaques are identified, picked, and up to 10,000 different mutants inoculated into 96 well microtiter dishes containing two 3 mm glass balls/well.
• the Q-bot does not pick an entire colony but rather inserts a pin through the center of the colony and exits with a small sampling of cells, (or mycelia) and spores (or viruses in plaque applications).
• the time the pin is in the colony the number of dips to inoculate the culture medium, and the time the pin is in that medium each effect inoculum size, and each parameter can be controlled and optimized.
• the uniform process of automated colony picking decreases human handling error and increases the rate of establishing cultures (roughly 10,000/4 hours). These cultures are optionally shaken in a temperature and humidity controlled incubator.
• Optional glass balls in the microtiter plates act to promote uniform aeration of cells and the dispersal of cellular (e.g., mycelial) fragments similar to the blades of a fermenter.
• Clones from cultures of interest can be isolated by limiting dilution.
• plaques or cells constituting libraries can also be screened directly for the production of proteins, either by detecting hybridization, protein activity, protein binding to antibodies, or the like.
• a prescreen that increases the number of mutants processed by 10-fold can be used.
• the goal of the primary screen is to quickly identify mutants having equal or better product titers than the parent strain(s) and to move only these mutants forward to liquid cell culture for subsequent analysis.
• One approach to screening diverse libraries is to use a massively parallel solid-phase procedure to screen cells expressing shuffled nucleic acids, e.g., which encode enzymes for enhanced activity.
• Massively parallel solid-phase screening apparatus using absorption, fluorescence, or FRET are available. See, e.g., United States Patent 5,914,245 to Bylina, et al. (1999); see also, http://www.kairos- scientific.com/; Youvan et al. (1999) "Fluorescence Imaging Micro-Spectrophotometer (FIMS)" Biotechnology et alia ⁇ www.et-al.com> 1:1-16; Yang et al.
• FIMS Fluorescence Imaging Micro-Spectrophotometer
• High throughput screening systems are commercially available (see, e.g., Zymark Corp., Hopldnton, MA; Air Technical Industries, Mentor, OH; Beckman Instruments, Inc. Fullerton, CA; Precision Systems, Inc., Natick, MA, etc.). These systems typically automate entire procedures including all sample and reagent pipetting, liquid dispensing, timed incubations, and final readings of the microplate in detector(s) appropriate for the assay. These configurable systems provide high throughput and rapid start up as well as a high degree of flexibility and customization.
• Zymark Corp. provides technical bulletins describing screening systems for detecting the modulation of gene transcription, ligand binding, and the like.
• a variety of commercially available peripheral equipment and software is available for digitizing, storing and analyzing a digitized video or digitized optical or other assay images, e.g., using PC (Intel x86 or pentium chip- compatible DOSTM, OS2TM WINDOWSTM, WINDOWS NTTM or WINDOWS95TM based machines), MACINTOSHTM, or UNIX based (e.g., SUNTM work station) computers.
• PC Intel x86 or pentium chip- compatible DOSTM, OS2TM WINDOWSTM, WINDOWS NTTM or WINDOWS95TM based machines
• MACINTOSHTM e.g., UNIX based (e.g., SUNTM work station) computers.
• Integrated systems for analysis typically include a digital computer with GO software for GAGGS, and, optionally, high-throughput liquid control software, image analysis software, data interpretation software, a robotic liquid control armature for transferring solutions from a source to a destination operably linked to the digital computer, an input device (e.g., a computer keyboard) for entering data to the digital computer to control GAGGS operations or high throughput liquid transfer by the robotic liquid control armature and, optionally, an image scanner for digitizing label signals from labeled assay components.
• the image scanner can interface with image analysis software to provide a measurement of probe label intensity.
• the probe label intensity measurement is interpreted by the data interpretation software to show whether the labeled probe hybridizes to the DNA on the solid support.
• GOs genetic algorithms
• digital or analog systems such as digital or analog computer systems can control a variety of other functions such as the display and/or control of output files.
• standard desktop applications such as word processing software (e.g., Microsoft WordTM or Corel WordPerfectTM) and database software (e.g., spreadsheet software such as Microsoft ExcelTM, Corel Quattro ProTM, or database programs such as Microsoft AccessTM or ParadoxTM) can be adapted to the present invention by inputting one or more character string into the software which is loaded into the memory of a digital system, and performing a GO as noted herein on the character string.
• word processing software e.g., Microsoft WordTM or Corel WordPerfectTM
• database software e.g., spreadsheet software such as Microsoft ExcelTM, Corel Quattro ProTM, or database programs such as Microsoft AccessTM or ParadoxTM
• systems can include the foregoing software having the appropriate character string information, e.g., used in conjunction with a user interface (e.g., a GUI in a standard operating system such as a Windows, Macintosh or LINUX system) to manipulate strings of characters, with GOs being programmed into the applications, or with the GOs being performed manually by the user (or both).
• a user interface e.g., a GUI in a standard operating system such as a Windows, Macintosh or LINUX system
• specialized alignment programs such as PDMEUP and BLAST can also be incorporated into the systems of the invention, e.g., for alignment of nucleic acids or proteins (or corresponding character strings) as a preparatory step to performing an additional GO on the resulting aligned sequences.
• Software for performing PCA can also be included in the digital system.
• Systems for GO manipulation typically include, e.g., a digital computer with GO software for aligning and manipulating sequences according to the GOs noted herein, or for performing PCA, or the like, as well as data sets entered into the software system comprising sequences to be manipulated.
• the computer can be, e.g., a PC (Intel x86 or Pentium chip- compatible DOS,TM OS2,TM WINDOWS,TM WINDOWS NT,TM WTNDOWS95,TM WINDOWS98,TM LINUX, Apple-compatible, MACINTOSHTM compatible, Power PC compatible, or a UNIX compatible (e.g., SUNTM work station) machine) or other commercially common computer which is known to one of skill.
• Software for aligning or otherwise manipulating sequences can be constructed by one of skill using a standard programming language such as Visualbasic, Fortran, Basic, Java, or the like, according to the methods herein.
• Any controller or computer optionally includes a monitor which can include, e.g., a cathode ray tube ("CRT") display, a flat panel display (e.g., active matrix liquid crystal display, liquid crystal display), or others.
• Computer circuitry is often placed in a box which includes numerous integrated circuit chips, such as a microprocessor, memory, interface circuits, and others.
• the box also optionally includes a hard disk drive, a floppy disk drive, a high capacity removable drive such as a writeable CD-ROM, and other common peripheral elements.
• Inputting devices such as a keyboard or mouse optionally provide for input from a user and for user selection of sequences to be compared or otherwise manipulated in the relevant computer system.
• the computer typically includes appropriate software for receiving user instructions, either in the form of user input into a set parameter fields, e.g., in a GUI, or in the form of preprogrammed instructions, e.g. , preprogrammed for a variety of different specific operations.
• the software then converts these instructions to appropriate language for instructing the system to carry out any desired operation.
• a digital system can instruct an oligonucleotide synthesizer to synthesize oligonucleotides for gene reconstruction, or even to order oligonucleotides from commercial sources (e.g., by printing appropriate order forms or by linking to an order form on the internet).
• the digital system can also include output elements for controlling nucleic acid synthesis (e.g., based upon a sequence or an alignment of a sequences herein), i.e., an integrated system of the invention optionally includes an oligonucleotide synthesizer or an oligonucleotide synthesis controller.
• the system can include other operations which occur downstream from an alignment or other operation performed using a character string corresponding to a sequence herein, e.g., as noted above with reference to assays.
• GOs of the invention are embodied in a fixed media or transmissible program component containing logic instructions and/or data that when loaded into an appropriately configured computing device causes the device to perform a GO on one or more character string.
• Figure 13 shows example digital device 700 that should be understood to be a logical apparatus that can read instructions from media 717, network port 719, user input keyboard 709, user input 711 or other inputting means. Apparatus 700 can thereafter use those instructions to direct GO modification of one or more character string, e.g., to construct one or more data set (e.g., comprising a plurality of GO modified sequences corresponding to nucleic acids or proteins).
• One type of logical apparatus that can embody the invention is a computer system as in computer system 700 comprising CPU 707, optional user input devices keyboard 709, and GUI pointing device 711, as well as peripheral components such as disk drives 715 and monitor 705 (which displays GO modified character strings and provides for simplified selection of subsets of such character strings by a user.
• Fixed media 717 is optionally used to program the overall system and can include, e.g., a disk-type optical or magnetic media or other electronic memory storage element.
• Communication port 719 can be used to program the system and can represent any type of communication connection.
• the invention can also be embodied within the circuitry of an application specific integrated circuit (ASIC) or programmable logic device (PLD).
• ASIC application specific integrated circuit
• PLD programmable logic device
• the invention is embodied in a computer readable descriptor language that can be used to create an ASIC or PLD.
• the invention can also be embodied within the circuitry or logic processors of a variety of other digital apparatus, such as PDAs, laptop computer systems, displays, image editing equipment, etc.
• the digital system comprises a learning component where the outcomes of physical oligonucleotide assembly schemes
• compositions, abundance of products, different processes are monitored in conjunction with physical assays, and correlations are established. Successful and unsuccessful combinations are documented in a database to provide justification/preferences for user- base or digital system based selection of sets of parameters for subsequent GAGGS processes involving the same set of parental character strings/nucleic acids/proteins (or even unrelated sequences, where the information provides process improvement information).
• the correlations are used to modify subsequent GAGGS processes to optimize the process. This cycle of physical synthesis, selection and correlation is optionally repeated to optimize the system. For example, a learning neural network can be used to optimize outcomes.
• the methods of this invention can be implemented in a localized or distributed computing environment.
• the methods may implemented on a single computer comprising multiple processors or on a multiplicity of computers.
• the computers can be linked, e.g. through a common bus, but more preferably the computer(s) are nodes on a network.
• the network can be a generalized or a dedicated local or wide-area network and, in certain preferred embodiments, the computers may be components of an intra-net or an internet.
• a client system typically executes a Web browser and is coupled to a server computer executing a Web server.
• the Web browser is typically a program such as IJBM's Web Explorer, Internet explorer, NetScape or Mosaic.
• the Web server is typically, but not necessarily, a program such as IBM's HTTP Daemon or other WWW daemon (e.g., LTNUX-based forms of the program).
• the client computer is bi-directionally coupled with the server computer over a line or via a wireless system.
• the server computer is bi-directionally coupled with a website (server hosting the website) providing access to software implementing the methods of this invention.
• a user of a client connected to the Intranet or Internet may cause the client to request resources that are part of the web site(s) hosting the application(s) providing an implementation of the methods of this invention.
• Server program(s) then process the request to return the specified resources (assuming they are currently available).
• a standard naming convention has been adopted, known as a Uniform Resource Locator ("URL"). This convention encompasses several types of location names, presently including subclasses such as Hypertext Transport Protocol (“http"), File Transport Protocol (“ftp”), gopher, and Wide Area Information Service (“WAIS").
• http Hypertext Transport Protocol
• ftp File Transport Protocol
• WAIS Wide Area Information Service
• the software implementing the method(s) of this invention can run locally on the server hosting the website in a true client-server architecture.
• the client computer posts requests to the host server which runs the requested process(es) locally and then downloads the results back to the client.
• the methods of this invention can be implemented in a "multi-tier" format wherein a component of the method(s) are performed locally by the client. This can be implemented by software downloaded from the server on request by the client (e.g. a Java application) or it can be implemented by software "permanently" installed on the client.
• the'application(s) implementing the methods of this invention are divided into frames.
• a typical application generally includes a set of menu items, each of with invokes a particular frame—that is, a form which manifest certain functionality of the application.
• an application is viewed not as a monolithic body of code but as a collection of applets, or bundles of functionality. In this manner from within a browser, a user would select a Web page link which would, in turn, invoke a particular frame of the application (i.e., subapplication).
• one or more frames may provide functionality for inputting and/or encoding biological molecule(s) into one or more character strings, while another frame provides tools for generating and/or increasing diversity of the encoded character string(s).
• the methods of this invention are implemented as one or more frames providing, e.g., the following functionalit(ies).
• the functions to encode two or more biological molecules can provide one or more windows wherein the user can insert representation(s) of biological molecules.
• the encoding function also, optionally, provides access to private and/or public databases accessible through a local network and/or the intranet whereby one or more sequences contained in the databases can be input into the methods of this invention.
• the end user inputs a nucleic acid sequenced into the encoding function
• the user can, optionally, have the ability to request a search of GenBank and input one or more of the sequences returned by such a search into the encoding and/or diversity generating function.
• EXAMPLE 1 DECISION TREE FOR EXAMPLE GAGGS PROCESS
• FIG. 1 provides an example decision making process from an idea of a desired property to selection of a genetic algorithm.
• Figure 2 provides a directed evolution decision tree from selection of the genetic algorithm to a refined library of parental character strings.
• Figure 3 provides example processing steps from the refined parental library to a raw derivative library of character strings.
• Figure 4 processes the raw character strings to strings with a desired property.
• the charts are schematics of arrangements for components, and of process decision tree structures. It is apparent that many modifications of this particular arrangement for DEGAGGS, e.g., as set forth herein, can be developed and practiced. Certain quality control modules and links, as well as most of the generic artificial neural network learning components are omitted for clarity, but will be apparent to one of skill.
• the charts are in a continuous arrangement, each connectable head-to tail. Additional material and implementation of individual GO modules, and many arrangements of GOs in working sequences and trees, as used in GAGGS, are available in various software packages.
• non-chip parallel devices for oligo synthesis provide an effective capacity to complete simultaneous (single-load) synthesis of 196 (2x96) individual 60-mer oligos in less than 5 hours, with the cost of hardware under $100K, and the cost of reagents under $0.07 per base. Therefore, with an understanding of these costs, the cost estimates made in the examples below can be reduced by at least 8 fold.
• EXAMPLE 3 GAGGS OF A SINGLE PARENT LOW MUTAGENICITY LIBRARY.
• This example describes GAGGS of a single parent low mutagenicity library derived from an average gene (-1.6 kb), given the sequence information of a single 1.6 kb gene (encoding 500 aa + "convenience" start/end oligos). The goal is to build a library of gene variants with all possible single amino acid changes, one aa change per each gene copy in the library.
• Relevant parameters include the number of oligos and cost to build 1 parental 1.6 kb gene, e.g., from 40 mer oligos, with complete 20+20 base overlaps e.g.,, by non-error prone assembly PCR, the number of all possible single aa replacement mutations, the number of distinct non-degenerate 40-mer oligos used to "build-in" all possible single aa mutations, the minimal number of all distinct fixed-position single- codon-degenerate oligos used to incorporate all possible single aa mutations, but not terminations, and the minimal number of all distinct fixed-position single-codon-fully degenerate oligos used to incorporate all possible single aa mutations.
• variable codons two of which are degenerate: NNT, VAA, TGG
• the same physical oligo inventory used for the first round GAGGS is used in the second round of GAGGS to synthesize a library which contains -95% of all possible combinations of any of two single aa changes. To have 100% coverage (to include for combinations of mutations within +/- 20 bp proximity, additional oligos are used.
• Tails for chimerizing each area of homology.
• Relevant parameters include: the number of oligos and resulting cost to build 6 parental 1.6 kb genes (from 40 mer oligos, complete 20+20 overlaps, by non- error-prone assembly PCR), the number of distinct pairwise crossovers between all matching homology areas, assuming 1 crossover event per pairwise homology region), the number of all possible chimeras using the crossovers, the theoretical library size, and the number of distinct oligos and cost to build all possible chimeras.
• X ⁇ 5.315 ⁇ 10 9 , calculated according to the formula:
• the cost of running multiple rounds of GAGGS is not additive, as most of the excess oligos from previous rounds can be reused in synthesis of the later generation libraries.
• This example provides a GAGGS family model stepwise protocol.
• a family of genes/proteins (DNA or AA sequence) is selected. All possible pairwise alignments are made to identify pairwise homology regions satisfying crossover operator conditions (length, % identity, stringency). Crossover points are selected, one per each of the pairwise homology substrings, in the middle of each substring, or randomly, or according to an annealing-based probability model built on histograms of crossover probability ranks for every pair of parents.
• Oligos are selected for assembly PCR and synthesized. Genes/libraries are assembled from synthesized oligos. The libraries are screened/ selected as set forth above.
• Amino acid sequences were aligned (Codon usage can be optimized on retrotranslation for a preferred expression system, and number of oligos for synthesis can be minimized).
• a Dot plot pairwise alignment of all possible pairs of 7 parents was made (Figs. 5, 6, 7).
• Figure 5 is a percent similarity alignment for 7 parents. Amino acid sequences are aligned, with the leader peptide excluded.
• Figure 6 is a dot-plot alignment of the sequences to identify regions of similarity.
• Figure 7 is a dot plot showing pairwise crossover points in the alignment.
• Pair 6 and 7 show 95% percent identity per each window of >7aa, while all other pairs show 80% percent identity per each window of >7aa.
• stringency of alignment and subsequent representation of crossover between parents
• the stringency of alignment can be manipulated individually for each pair, so that low homology crossovers can be represented at the expense of highly homologous parents.
• No structural biases or active site biases were incorporated in this model.
• oligos From 40 mers, with full overlap assembly (20 +20 bp overlaps), about 300 oligos are used. For pairwise crossover oligos to build chimeras, based on alignment results, with one crossover per each homologous substring, there are about 180 homologous substrings, with 170 in the coding region and 10 in the leader region. With 2 60 mers per each crossover point, and 2 head-tail sets for each pair of parents, about 360 additional oligos dedicated to build crossovers can be used. The total number of oligos is about 660 (300 40 mers and 360 60 mers). At a total cost of oligos of $0.70 per base, the oligos would be about $23,520. The cost of reagents would run about $0.07 per base, for a total cost of about $2,252 dollars.
• Napthalene dioxygenase is a non-heme reductive dioxygenase. There are at least three closely related but catalytically distinct types of Napthalene dioxygenases.
• Figure 12 provides a schematic of a percent similarity plot for the three different Napthalene dioxygenase types, with the amino acid sequence for the ISP large subunit (which is responsible for substrate specificity) being provided.
• one aspect of the invention provides for single parent GAGGS.
• polynucleotides having desired characteristics are provided. This is accomplished by: (a) providing a parental sequence character string encoding a polynucleotide or polypeptide; (b) providing a set of character strings of a pre-defined length that encode single-stranded oligonucleotide sequences comprising overlapping sequence fragments of an entire parental character string, and an entire polynucleotide strand complementary to the parental character string (splitting the sequence of a parent into oligos suitable for assembly PCR); (c) creating a set of derivatives of parental sequence comprising variants with all possible single point mutations, with, e.g., one mutation per variant string (defining all possible single point mutations); (d) providing a set of overlapping character strings of a pre-defined length that encode both strands of the parental oligonucleotide sequence, and a set of overlapping character strings of a predefined length that encode
• wild type oligos are excluded at mutations; and (g) selecting or screening for recombinant polynucleotides having evolved toward a desired property.
• the method includes deconvoluting sequence of the mutated polynuceotides (i.e., determining which library member has a sequence of interest, and what that sequence is) having evolved toward a desired property to determine beneficial mutations (when assembly PCR is one per container format, this is done by positional sequence deconvolution, rather than actual sequencing, i.e., the physical location of the components are adequate to provide knowledge of the sequence).
• the method includes assembling a library of recombinant variants which combine some or all possible beneficial mutations in some or all possible combinations, from single-stranded oligos by assembly PCR. This is performed from the same set of oligos; if some of the mutations are positionally close (within any one oligo), then additional single strand oligos are made which incorporate combinations of mutations.
• An optional step (j) includes selecting or screening for recombinant polynucleotides having evolved further toward a desired property.
• Genome length 1000 bp.
• First round mutation rate 1 amino acid/ gene.
• non degenerate oligos 13320. Partially degenerate 40 mers, one pg position per oligo: 1920. Fully degenerate oligos, 40 mers, one fg position per oligo: 666.
• the number of additional oligos to allow for construction of all possible recombinations with beneficial mutations would use about 10% of the preceding number of oligos. However, about 95% of all possible recombinants having beneficial mutations can be made from the initial set produced above.
• EXAMPLE 9 A PROCESS FOR DESIGN OF CROSSOVER OLIGONUCLEOTIDES FOR SYNTHESIS OF CHIMERICAL POLYNUCLEOTIDES:
• substrings are identified and selected in parental strings for applying a crossover operator to from chimeric junctions. This is performed by: a) identifying all or part of the pairwise homology regions between all parental character strings, b) selecting all or part of the identified pairwise homology regions for indexing at least one crossover point within each of the selected pairwise homology regions, c) selecting one or more of the pairwise non-homology regions for indexing at least one crossover point within each of the selected pairwise non-homology regions ("c" is an optional step which can be omitted, and is also a step where structure-activity based elitism can be applied), thereby providing a description of a set of positionally and parent-indexed regions/areas (substrings) of parental character strings suitable for further selection of crossover points.
• step 1 further selection of crossover points within each of the substrings of the set of the substrings selected in step 1 above is performed.
• the steps include: a) randomly selecting at least one of the crossover points in each of the selected substrings, and/or b) selecting at least one of the crossover points in each of the selected substrings, using one or more of annealing-simulation-based models for determining probability of the crossover point selection within each of the selected substrings and/or c) selecting one crossover point approximately in the middle of each of the selected substrings, thereby creating a set of pairwise crossover points, where each point is indexed to corresponding character positions in each of the parental strings desired to from a chimeric junction at that point.
• codon usage adjustments are performed.
• the process can be varied. For example, if a DNA sequences was used: a) adjustment of codons for the selected expression system is performed for every parental string, and b) adjustment of codons among parents can be performed to standardize codon usage for every given aa at every corresponding position.
• This process can significantly decrease total number of distinct oligos for gene library synthesis, and may be particularly beneficial for cases where AA homology is higher than DNA homology, or with families of highly homologous genes (e.g. 80%+ identical).
• AA sequences are used: a) retrotranslate sequence to degenerate DNA; b) define degenerate nucleotides using position-by-position referencing to codon usage in original DNA (of majority of parents or of corresponding parent), and/or - exercise codon adjustments suitable for the selected expression system where a physical assay will be performed.
• This step can also be used to introduce any restriction sites within coding parts of the genes, if any, for subsequent identification/QA/deconvolution/manipulations of library entries. All crossover points identified in step 2 above (indexed to pairs of parents) are correspondingly indexed to the adjusted DNA sequences.
• oligo arrangements are selected for a gene assembly scheme. This step includes several decision steps: Uniform 40-60 mer oligos are typically used (using longer oligos will result in decrease of # of oligos to build parents, but uses additional dedicated oligos for providing representation of closely positioned crossovers/mutations.
• a "Yes” decision cuts the total number of oligos for high homology genes of different lengths with gaps (deletion/insertion), esp for l-2aa.
• convenience sequences are designed in front and in the back of the parent strings. Ideally, it is the same set which will be built in every library entry at the end. These include any restriction sites, primer sequences for assembled product identifications, RBS, leader peptides and other special or desirable features.
• the convenience sequences can be defined at a later stage, and at this stage, a "dummy" set of appropriate length can be used, e.g. a substring from an easily recognizable forbidden letters.
• An indexed matrix of oligo strings for building every parent is created, according to the selected scheme.
• An index of every oligo includes: a parent identifier (parenfTD), indication of coding or complementary chain, and position numbers.
• Crossover-points are determined for indexed coding string of every parent with head and tail convenience substrings.
• a complementary chain of every string is generated. Every coding string is selected according to the selected assembly PCR scheme in step 4 above (e.g. in increments of 40 bp). Every complement string is split according to the same scheme (e.g. 40 bp with 20 bp shift)
• an indexed matrix of oligos is created for every pairwise crossover operation.
• Third, every set of 4 oligo strings are taken which have been labeled with the same crossover marker, and another derivative set of 4 chimeric oligo strings comprising of characters encoding 2 coding and 2 complement chains (e.g. with 20 bp shift in 40 20+20 scheme) are made.
• Two coding strings are possible, having a forward end sequence substring of one parent followed by the backward end of the second parent after crossover point.
• Complement strings are also designed in the same fashion, thereby obtaining an indexed complete inventory of strings encoding oligos suitable for gene library assembly by PCR.
• EXAMPLE 10 PROGRAM ALGORITHM FOR DESIGNING OLIGONUCLEOTIDES FOR SYNTHESIS
• the following is a program outline for designing oligonucleotides for use in synthetic/ recombination protocols.
• oligos given bounds which contains a minimum amount of sequence from current sequence 5' of the gap and a minimum amount of sequence from current sequence 3' of the gap
• Oligos for reverse bounds add Oligos: given a list of sequences (DNA) and bounds for each position in the bounds get all unique bases at this position from DNA sequences in the list generate a base (or degenerate base symbol) for this position if total degenerate positions is greater than user defined number split sequence list in two, add Oligos given sequence list one, add Oligos given sequence list two (recursive) else add this oligo (set of bases for each position) to oligo list display all oligo in oligo list
• Figs. 8-11 are schematics of various processes and process criteria for oligonucleotide selection for recombination between parental nucleic acids.
• panel A shows a typical dot-plot alignment of two parents and the increase in crossover probability that results in regions of similarity.
• Panel B shows that crossovers can be selected based upon a simple logical/physical filter, i.e., the physical or virtual annealing temperature of oligonucleotides, e.g., using a linear annealing temperature.
• Panel C shows various more complex filters which vary annealing temperature to achieve specific crossovers, i.e., by appropriately controlling the physical or virtual annealing temperature.
• Figure 9 schematically represents the introduction of indexed crossover points into the sequence of each of the aligned parents.
• sequences are aligned and the positional index of each crossover point (marker field) is represented schematically by a vertical identifier mark.
• the crossover point, for parents m and n, as represented in Figure 8, is represented by an identifier, a position number for parent m (a head) and a position number for parent n (a tail). This process is repeated for every parent in a data set, applying the oligonucleotide gridding operator (grid of positional indexes which indicate the start and end of every oligonucleotide in a PCR assembly operation) to each of the parents.
• the oligonucleotide gridding operator grid of positional indexes which indicate the start and end of every oligonucleotide in a PCR assembly operation
• Figure 10 schematically represents the complete inventory of oligonucleotide sequences to assemble all of the parents.
• the data set is simplified by identifying all pairs of oligonucleotide sequences with matching pairwise crossover indexes, providing a sub inventory of oligos with crossover markers.
• Figure 11 provides a schematic for obtaining an inventory of sequences for chimeric oligonucleotides for each of the selected crossover points.
• two pairs of oligo sequences with matching pairwise crossover indexes are selected (down arrow 1).
• Sequences of chimeric oligonucleotides are generated around the pairwise crossover point (head-tail, tail-head, s or a chain) (down arrow 2).
• EXAMPLE 12 SYNTHETIC SHUFFLING BY REPEATED CYCLES OF MELTING. ANNEALING AND POLYMERIZATION:
• the methods herein provide, inter alia, synthetic shuffling using, e.g., a primerless PCR reaction to assemble a series of overlapping oligos designed to capture the diversity represented in a family of related sequences.
• synthetic shuffling provides a direct route from an in silico database to shuffled library.
• Synthetic shuffling projects are not limited to natural diversity, and need not result in equal representation of diversity, and diversity can occur with any desired granularity of diversity (e.g., at the level of single amino acids, or in blocks, depending on the sequences that are selected). Freedom of oligo sequence, length, and number; as well as freedom to mix oligos in any desired combination and quantities under any desired assembly parameters allows great control over library design.
• subtilisins also described supra.
• 16 overlapping 60-mers (8 top strand and 8 bottom strand) with 20 bp of homology between ends of top and bottom strand oligos formed a backbone were designed to capture the majority of natural diversity represented in a 660 bp gene segment (also known as the diversified region) of a family of 15 subtilisin sequences.
• the diversity was largely captured by introducing degeneracies into the oligonucleotide backbone. Diversity that was not captured in the backbone oligos was encoded in 11 additional 39- to 45-mers.
• the backbone and spiking oligos optimized codon usage for Bacillus subtilis and maximized recombination. Oligos were mixed at equimolar concentrations and assembled in a primerless PCR reaction. A full-length library was rescued from the assembly by conventional PCR with primers annealing to construct ends. The library was designed such that it recombines at the level of single amino acids and results in equal representation of all possible amino acids in a particular position (assuming unbiased assembly and equal representation of nucleotides at degenerate positions).
• Kits will optionally additionally comprise instructions for performing methods or assays, packaging materials, one or more containers which contain assay, device or system components, or the like.
• kits embodying the methods and apparatus herein optionally comprise one or more of the following: (1) a shuffled component as described herein; (2) instructions for practicing the methods described herein, and/or for operating the selection procedure herein; (3) one or more assay component; (4) a container for holding nucleic acids or enzymes, other nucleic acids, transgneic plants, animals, cells, or the like, (5) packaging materials, and (6) software for performing any of the decision steps noted herein related to GAGGS.
• the present invention provides for the use of any component or kit herein, for the practice of any method or assay herein, and/or for the use of any apparatus or kit to practice any assay or method herein.

Landscapes

• Physics & Mathematics (AREA)
• Life Sciences & Earth Sciences (AREA)
• Bioinformatics & Cheminformatics (AREA)
• Bioinformatics & Computational Biology (AREA)
• Biophysics (AREA)
• Proteomics, Peptides & Aminoacids (AREA)
• Health & Medical Sciences (AREA)
• Engineering & Computer Science (AREA)
• Chemical & Material Sciences (AREA)
• Analytical Chemistry (AREA)
• Biotechnology (AREA)
• Evolutionary Biology (AREA)
• General Health & Medical Sciences (AREA)
• Medical Informatics (AREA)
• Spectroscopy & Molecular Physics (AREA)
• Theoretical Computer Science (AREA)
• Enzymes And Modification Thereof (AREA)
• Preparation Of Compounds By Using Micro-Organisms (AREA)

Applications Claiming Priority (5)

Publications (1)

Family

ID=27066107

Family Applications (1)

Country Status (3)

Families Citing this family (116)

Family Cites Families (4)

• 2001
  • 2001-03-30 AU AU4964101A patent/AU4964101A/xx active Pending
  • 2001-03-30 EP EP01922889A patent/EP1272967A2/de not_active Withdrawn
  • 2001-03-30 WO PCT/US2001/010231 patent/WO2001075767A2/en not_active Application Discontinuation

Non-Patent Citations (1)

Also Published As

Similar Documents

Legal Events