CA2485203A1 - Error-prone dna polymerase i mutants and methods for targeted random mutagenesis in continuous culture using error-prone dna polymerase i mutants - Google Patents

Abstract

Mutant forms of DNA polymerase I having mutations within motif A and/or motif B in the active domain that increase error rates during replication.
Expression plasmid constructs and cell lines for expressing these low-fidelity polymerase mutants are provided. Methods are also provided for utilizing these low-fidelity DNA polymerase I mutants for generating libraries of randomly-mutagenized genes, which may be prokaryotic or eukaryotic. Random mutagenesis involves the coupling of mutagenesis and selection in continuous culture for convenient iteration, which results in diverse range of base pair substitutions, widely distributed along the sequence. Some advantages include the minimization of deleterious damage to chromosomal DNA, and adaptation to strains that are amenable to complementation, which substantially facilitates the generation and identification of enzymes with altered properties.

Description

ERROR-PRONE DNA POLYMERASE I MUTANTS
AND METHODS FOR TARGETED RANDOM MUTAGENESIS IN
CONTINUOUS CULTURE USING ERROR-PRONE DNA POLYMERASE I
MUTANTS
CROSS-REFERENCE TO RELATED APPLICATION
This application claims the benefit of provisional patent application number 601384,944, filed May 31, 2002, now pending.
TECHNICAL FIELD
This invention relates to mutants of DNA polymerase I and methods for their application, particularly to DNA Pol I mutants engineered to have decreased fidelity of replication and to methods for their application in random mutagenesis for the modification of target sequences in continuous culture.
STATEMENT OF GOVERNMENT INTEREST
This invention has been made with Government support under grant number ROI CA 78885, awarded by the National Institute of Health. The government has certain rights in the invention.
BACKGROUND OF THE INVENTION
Enzymes are routinely used in the pharmaceutical and biotechnology industries, and are finding increasing applications as biocatalysts in chemical synthesis and bioremediation. Native enzymes seldom possess the properties required for direct industrial or medical application, because they have evolved to perform specific functions within specific biochemical environments generally different from the chemical environments of industrial and medical applications. Two general approaches have been used to alter enzyme performance. One involves the replacement of individual amino-acid residues based on detailed structural information, and the other involves the production of large libraries of randomly substituted variants, or mutants. By modifying the structural features already present in the parent enzyme, mutagenesis followed by the identification of mutants with properties of interest allows for fine-tuning enzyme performance with respect to optimal conditions for catalysis and/or substrate specificity. This approach has great potential in improving complex biosynthetic and degradative pathways, and in overcoming the rate-limiting steps for medical and industrial applications.
Unlike site-directed mutagenesis, random mutagenesis requires no prior knowledge of the structure of the targeted enzyme or of the mechanistic aspects of the reaction. The main limitation of random mutagenesis is that only a small fraction of the sequence space can be effectively explored. This is due to the astronomical number of possible combinations of amino acid substitutions in an average enzyme (far exceeding the size of most libraries) and to the frequent generation of functionally deficient mutants. The diversity of random mutant libraries is also constrained by the error-correcting nature of the genetic code. Although a particular amino-acid residue within a protein may be replaced by one of 19 other, naturally occurring amino acids, a given single point mutation within the codon that encodes the particular amino acid results in an average of only 6 different amino-acid substitutions, and these substitutions tend to be conservative, resulting in replacement of the particular amino acid with another amino acid having similar physicochemical properties.
Numerous methods have been developed for introducing random point mutations in vitro, ranging from PCR-mediated chemical mutagenesis [Roufa, D.
J., Methods Mol. Biol. 57: 357-67 (1996)], to mutagenic PCR [Cadwell, R. C. and Joyce, G. F., PCR Methods Appl. 3: 5136-40 (1994)], and ligation of degenerate oligonucleotides [Black, M. E. and Loeb, L. A., Methods Mol. Biol. 57: 335-49 (1996)]. These methods offer some control over the intensity of mutagenesis, and can achieve very high mutation loads. However, even under intense mutagenesis conditions, the probability of generating, by mutagenesis, a specific subsequence of more than 2 or 3 amino acids within an enzyme is minimal. Moreover, high mutation loads increase the proportion of deficient mutants.
One strategy to increase the likelihood of generating mutants with altered properties consists of restricting mutagenesis to areas of the target gene that are directly involved in catalysis [Suzuki et al., Mol. Divers. 2: 111-18 (1996)].
While this strategy has been successful in some cases [Encell et al., Cancer Res.
58: 1013-20 (1998); Glick et al., Embo J. 20: 7303-12 (2001); Patel, P. H. and Loeb, L.
A., Proc.
Natl. Acad. Sci. U S A 97: 5095-100 (2000); Shinkai et al., J. Biol. Chem.
276:
18836-42 (2001 )], it leaves large portions of the protein unexplored, requires substantial knowledge of protein structure and structure-function relationships, and becomes difficult when the sequence of the active site is not contiguous or readily identifiable.
Finding the desired variants is usually a major challenge. This may be done either by true selection or by screening. True selection consists of giving a growth advantage to the clones that show the desired activity, and is therefore performed in vivo. Depending on how strong the growth advantage is, selection-based techniques can be extremely effective for increasing the representation of desired variants in populations. Selection-based techniques have the added advantage of eliminating defective and inactive mutants. There are a number of ways of linking a specific enzymatic property with growth, including genetic complementation, auxotrophy, or drug resistance. Unfortunately, selection-based techniques are not available for all enzymes and desired properties. When selection-based techniques are not available, desired variants can be identified by screening. Screening is more versatile than selection-based techniques, because screening does not require biological activity.
However, screening is more limited in throughput.
When selection-based techniques are available, in vivo mutagenesis offers significant advantages over in vitro methods, because in vivo mutagenesis may be readily coupled to a positive genetic selection. This promotes the accumulation of beneficial mutations, a strategy known as "directed evolution". By the stepwise selection of mutations with a positive effect on the phenotype of interest, we can reach combinations of mutations that had a very low probability at the beginning of the experiment. This iterative process of mutagenesis and selection is greatly facilitated by performing both mutagenesis and selection simultaneously in vivo.
The efficacy of a given in vivo mutagenesis system relies predominantly on the type of cellular hosts or "mutator strains" used, and the currently available mutator strains are inadequate for a number of reasons. For example, mutDS is a mutant of Pol III deficient in proofreading activity [Scheuermann, R. H. and Echols, H., Proc.
Natl. Acad. Sci. U S A 81: 7747-51 (1984)]. Under certain growth conditions, errors introduced by this polymerase, the main replicative polymerase in E. coli, result in the saturation of mismatch repair and a dramatic increase in mutagenesis [Schaaper, R.
M. and Badman, M., Embo. J. 8: 3511-16 (1989)]. Other strains are based on mismatch repair inactivation [Glickman, B. W. and Badman, M., Proc. Natl.
Acad.

Sci. U S A 77: 1063-67 (1980)], or combine mutations affecting mismatch repair (mut S), oxo-dGTP repair activity (mut T), and the 3'-5' exonuclease activity of DNA Pol III (mut D) [Greener et al., Methods Mol. Biol. 57: 375-85 (1996)]. The inactivation of major DNA repair pathways, however, severely limits the performance of these mutator strains [Greener et al., Mol. Biotechnol. 7: 189-95 (1997)]. Since mutagenesis is not directed toward the targeted sequence, growing these mutator strains in culture invariably results in widespread mutagenesis. This leads to substantial loss of fitness of the organism, which, in turn, limits the number of effective iterative cycles that can be performed [Fijalkowska, I. J. and Schaaper, R.
M., Proc. Natl. Acad. Sci. U S A 93: 2856-61 (1996)]. Also, non-specific mutations may obscure phenotypic expression from mutations in the target gene [Long-McGie et al., Biotechnol. Bioeng. 68: 121-25 (2000); Negri et al., Antimicrob.
Agents Chemother. 44: 2485-2491 (2000)] and decrease the efficiency of mutagenesis with prolonged passage in culture [Greener et al., Mol. Biotechnol. 7: 189-95 (1997)].
In summary, although random mutagenesis allows enzyme modification with minimal structural or mechanistic information about the target protein, the number of mutants that need to be analyzed limits the efficacy of this technique.
Although high mutation loads cannot be achieved in vivo, coupling in vivo mutagenesis to positive genetic selection expands the number of mutants that can be effectively analyzed.
Well-characterized mutator strains exist, but have the following limitations with respect to enzyme modification by random mutagenesis: (1) the dependence on specific defects in DNA polymerases and/or in DNA repair pathways of a host;
(2) the unhealthy state of host cells resulting from widespread chromosomal mutagenesis and from defective DNA repair pathways; (3) the increase in phenotypic noise resulting from non-specific chromosomal mutations that makes selection based on phenotypic properties less reliable; and (4) the progressive decrease in mutagenesis rates that occurs with prolonged culture. There is a clear need for improving random mutagenesis in culture. Specifically, healthier host strains and independence from a specific genetic background are desired.
SUMMARY OF THE INVENTION
One embodiment of the present invention relates to the engineering of error-prone DNA polymerases within the Pol I family and methods for their use in implementing random mutagenesis of exogenous target genes of interest in continuous culture at frequencies that exceed those of chromosomal DNA.
Various embodiments of the present invention relate to the engineering of DNA
polymerases by introducing mutations in residues within their active sites that increase the error 5 rates of the polymerases, collectively called "error-prone Pol I mutants".
These mutations involve either motif A, or motif B, or both motifs A and B. The compositions of various embodiments of the present invention comprise the DNA
sequences encoding any portion of the engineered sites of Pol I mutants having low-fidelity, any portion of the Pol I mutants containing the engineered sites conferring low-fidelity, and host strains comprising either the DNA or polypeptide sequences representing these error-prone Pol I mutants.
Other embodiments of the invention relate to a method of expressing these error-prone polymerase mutants under carefully regulated, or "optimized", culture conditions to achieve an elevated frequency of mutagenesis preferentially targeted to an exogenously-introduced plasmid encoding a sequence of interest. The compositions and methods provide for random mutagenesis performed in continuous culture, without plasmid recovery between selection steps, which greatly facilitates the accumulation of beneficial mutations. The compositions and methods can be easily adapted for use in any existing strain suitable for complementation with the target gene because the critical reagents are encoded in transformable vectors, and mutagenesis is not contingent on any specific genetic defect in the host strain.
Suitable strains include cells endogenously expressing wild-type DNA
polymerase and/or wild type DNA repair function. The compositions and methods can introduce a broad spectrum of mutations in any DNA target sequence, prokaryotic or eukaryotic, resulting in the identification of sequences encoding polypeptides with altered properties.
BRIEF DESCRIPTION OF THE DRAWINGS
FIGS. 1 A-B provide a general overview of the 2-plasmid mutagenesis system, involving the co-expression of an error-prone Pol I and an exogenously introduced DNA template.

FIG. 2 illustrates residues within motif A and motif B in the active site of E.
coli DNA polymerase that are critical for the fidelity of replication during nucleotide incorporation.
FIG. 3A illustrates embodiments of template DNA, including prokaryotic and eukaryotic DNA, to be targeted for in vivo mutagenesis by error-prone Pol I.
FIG. 3B illustrates, in particular, several target plasmids having variable distances between the origin of replication (ori) and the ochre stop codon within the (3-lactamase gene used as a reporter for mutagenesis.
FIG. 4 illustrates the chemical structure of three classes of (3-lactamase substrates: penicillin, cephalosporin, and aztreonam. Aztreonam served as a selective agent in Example 7.
FIG. SA illustrates the initial conditions and a method for inducing mutagenesis in culture by the expression of error-prone Pol I in the presence of a target plasmid, as discussed in Example 2.
FIG. SB illustrates the optimized conditions and a method for inducing mutagenesis in culture by the expression of error-prone Pol I in the presence of a target plasmid, as discussed in Example 2.
FIG. 6 illustrates the effect of culture conditions on mutagenesis by the expression of error-prone Pol I mutants, as discussed in Example 3.
FIG. 7 illustrates the copy number of the target plasmid when D424A I709N
A759R error-prone Pol I is expressed under optimized conditions.
FIG. 8A shows the location of mutations identified within a 650 by interval beginning at a distance of 100bp from the ColEl origin of replication (ori), as discussed in Example 5.
FIG. 8B shows how distance affects the frequency of mutations in a target plasmid, as discussed in Example 5.
FIG. 9 illustrates the effect of error-prone Pol I expression on plasmid mutagenesis rates in a host that is wild type for Pol I and proficient in major pathways of DNA repair, as discussed in Example 6.
FIG. 10 shows dose-response curves of aztreonam resistance conferred by a representative subset of mutations identified following selection with aztreonam, as discussed in Example 7.

7 ,.
DETAILED DESCRIPTION OF THE INVENTION
Embodiments of the present invention are directed to a generation of Pol I
mutants, collectively referred to as "error-prone Pol I mutants," that have been altered from the wild-type DNA polymerase at specific residues that form the active domain, resulting in increased misincorporation by these Pol I mutants during catalysis. These low-fidelity Pol I mutants are potentially useful in any industry that employs enzymes to perform a biochemical reaction, where any improvement in the activity of such enzymes is desirable. Embodiments of the present invention are useful for broad applications in the pharmaceutical/medical industry as well as any industry that engages in a process that is enzyme-dependent.
FIG. 1 A provides an overview of the random mutagenesis approach in continuous culture mediated by error-prone Pol I mutants. A prokaryotic host cell containing a first vector encoding a target gene of interest and a second vector encoding an error-prone Pol I mutant is represented. A suitable host cell may or may not express the wild type Pol I endogenously. By using the present mutagenesis system, iterative cycles of mutagenesis and selection result in the directed evolution of a gene of interest.
FIG. 1 B illustrates an example of how the expression of low-fidelity Pol 1 mutants in host cells promote plasmid-specific mutagenesis, without substantially modifying the chromosome of a prokaryotic host when expressed in vivo. As shown, a host cell that has wild type Pol 1I1 (the major DNA polymerase) and temperature-sensitive Pol I activities can be transformed with two plasmids. When mutagenesis occurs at a restrictive temperature, the endogenous temperature-sensitive Pol I can be inactivated. Within a first ("target") plasmid, a target gene of interest can be inserted downstream of a ColEl origin of replication ("ori"), which is specifically recognized by Pol I. A second plasmid ("pSC101 on") can express a mutant DNA polymerase that is engineered within the active site to erroneously replicate substrate DNA, such as the first plasmid containing the target gene. Since DNA synthesis by Pol I
is very limited in the chromosome, mutagenesis is largely restricted to the target plasmid upon expression of the error-prone Pol I mutants in vivo.
The following paragraphs cover two general embodiments of the invention, the compositions relating to the mutagenesis system, and the methods for using the present compositions for implementing directed evolution by continuous culture in vivo.
Components of the Muta~enesis System With respect to the mutagenesis system that includes compositional embodiments directed to error-prone Pol I mutants, the following provides descriptions of: ( 1 ) the development and characterization of motif A and/or B Pol I
mutants, (2) the host strains that can be used to express motif A and/or B Pol I
mutants, and (3) the types of target sequences that can be used by the motif A
and/or B Pol I mutants. Specific Examples relating to these compositions are referenced and provided in the following section.
Development and Characterization of Error-prone Pol I Mutants Type I polymerases consist of a single polypeptide with 3 distinct domains [Kornberg, A. and Baker, T. A., DNA replication, second edition, Vol. 1, p.
931. New York: W.H. Freeman and Company (1992)]. 1n particular, 3 residues (I614, A661, and T664) that are critical for fidelity in one such polymerase, Thermus aquaticus (Taq) polymerase, were previously identified by the present inventors [Suzuki et al., J.
Biol. Chem. 272: 11228-35 (1997); Patel et al., J. Biol. Chem. 276: 5044-51 (2001)].
It has been shown that, upon dNTP binding, motifs A and B create a pocket in the catalytic site that accommodates an incoming nucleotide. Certain amino acid substitutions at specific positions in these motifs favor misincorporations, due to an enlargement of the active site cavity and/or to a more stable "closed"
conformation [Suzuki et al., J. Biol. Chem. 272: 11228-35 (1997); Patel et al., J. Biol.
Chem. 276:
5044-51 (2001)].
One embodiment of the present invention provides compositions relating to variants of E. coli DNA polymerases, i.e., error-prone Pol I mutants, that are engineered within the active site, particularly in motif A and/or motif B, to produce an elevated frequency of replication mistakes within substate DNA. FIG. 2 illustrates the positions within motifs A and B that are identified and characterized by the invention, which are critical for the fidelity of E. coli polymerases. In particular, position I709 of motif A and positions 5756 and A759 of motif B of E. coli DNA Pol I are engineered to increase incorporation of mismatched residues. These mutants are further described in the following sections. Methods for the construction of error-prone Pol I mutants are provided below in Example 1.
One embodiment is directed to E. coli error-prone Pol I mutants bearing amino-acid substitutions in both the motif A and 3' --~ 5' exonuclease domains, which exhibit an elevated frequency of misincorporation during replication of target DNA.
Motif A mutants were altered by substituting wild type aspartic acid (D) at position 424 with alanine (A) to inactivate the exonuclease function essential for proof reading. FIG. 2 provides examples of motif A mutants, lacking exonuclease activity, including the D424A 1709N and D424A I709F mutants, in which the wild type isoleucine at position 709 is substituted with an asparagine (N), or a phenylalanine (F). The biochemical characterization of these two mutants was conducted by the present inventors, and the following publication representing their work is herein incorporated by reference [Shinkai, A. and Loeb, L., J. Biol. Chem. 276: 46759-(2001)]. These error-prone Pol I mutants, bearing one or more mutations in motif A
of the active site, more frequently produce errors during template replication both in vitro and in vivo, and thus may be used in vitro and in vivo.
Another embodiment is directed to E. coli error-prone Pol I mutants, bearing mutations in both the motif B and the 3' ->5' exonuclease domains, that produce an elevated frequency of misincorporation during replication of target DNA. Motif B
mutants were altered by substituting wild type aspartic acid (D) at position 424 with alanine (A) to inactivate the exonuclease function essential for proof reading. F1G. 2 provides examples of motif B mutants, lacking exonuclease activity, that include the D424A S756E and D424A A759R mutants, in which the wild type serine (S) in position 756 is substituted with glutamic acid (E) and the wild type alanine (A) in position 759 is substituted with arginine (R). These error-prone Pol I
mutants, bearing one or more mutations in motif B of the active site, may be used in vitro and in vrvo.
A further embodiment is directed to E. coli error-prone Pol I mutants, bearing amino-acid substitutions in both motif A and motif B, that produce highly elevated frequencies of misincorporation during replication of target DNA. These motifs A
and B mutants were altered to inactivate their 3'~S' exonuclease activity by substituting wild type aspartic acid (D) at position 424 with alanine (A), although this modification doesn't appear to have a significant impact on mutagenesis in these mutants (see Table 2). FIG. 2 provides examples of error-prone Pol I mutants carrying amino-acid substitutions in motif A and motif B. One such mutant is the "D424A I709N A759R," in which the wild type isoleucine (I) at position 709 is 5 substituted with an asparagine (N) and the wild type alanine (A) in position 759 is substituted with arginine (R). Another mutant in both motifs is referred to as the "D424A I709F A759R," in which the wild type isoleucine (I) at position 709 is substituted with a phenylalanine (F) and the wild type alanine (A) in position 759 is substituted with arginine (R). These error-prone Pol I mutants, bearing a single or 10 more amino-acid substitution in each motif A and motif B of the active site, may be used in vitro and in vivo.
Additional embodiments, related to the above-discussed E.coli error-prone Pol I mutants carrying motif A, motif B, and the combination of motif A and motif B
mutations, include homologous mutations that may be engineered into other members of the Pol I family of DNA polymerises, including the DNA polymerise of Staphylococcus aureus, Streptococcus pneumoniae, Bacillus subtillis, or Salmonella enteritidis.
Unrestricted Selection in Host Strains Used for Mutagenesis Mediated by Error-prone Pol I Mutants Suitable hosts for the expression of mutant motif A and/or motif B Pol I genes are not restricted to strains with inactivated DNA repair mechanisms or to strains with defective chromosomal Pol I. The mutagenesis system is therefore uniquely adaptable to a variety of strains imaginable. Various embodiments that employ cells with or without endogenous wild type Pol I activity may be used to express mutators.
Other embodiments are directed to strains with or without inactivated DNA
repair mechanisms. Examples of host strains that may provide the appropriate cellular environment for supporting mutagenesis mediated by error-prone Pol I
expression include E. coli strains such as the JS200 and XL1-Blue. The JS200 strain represents a strain having a temperature-sensitive allele of Pol I (polA72) in the chromosome [Monk, M. and Kinross, J., J. Bacteriol. 109: 971-78 (1972)]. The XL1-Blue strain represents a strain having endogenous wild type Pol I activity and substantial proficiency in DNA repair, and therefore able to support the co-expression of exogenous error-prone Pol I mutants. Another embodiment is directed to a number of existing strains that are deficient in specific enzymes that can be functionally complemented, permitting the exploitation of the system to its fullest extent.
When used in vivo, the error-prone Pol I mutants encoding either motif A
and/or motif B amino-acid substitutions can be delivered into a prokaryotic host as part of a circular vector such as a plasmid or in a linear form, such as in a bacterial phage. In one embodiment, the vehicle encoding the motif A and/or motif B Pol I can remain in an unintegrated form, as an exogenous plasmid for example. In another embodiment, single or multiple copies of genes encoding any portion of motif A
and/or motif B mutator sequences may be integrated into the chromosome of the prokaryotic host, generating stable lines of mutator prokaryotic strains or cell lines.
One example for generating a stable mutator cell line is described as follows.
A double-stranded DNA fragment consisting of 3000 by of the PolAl locus encoding a low-f delity mutant that is flanked on the 5' and 3' sides by approximately 1000 by of the native genomic sequence, can be subcloned into a plasmid with a temperature sensitive origin of replication. Chloramphenicol can be used as a marker for positive selection and sucrose can be used for negative selection. For example, the resulting plasmid can be used to transform JS200 cells and can be selected for chloramphenicol resistance under permissive conditions at 30°C. The culture can be switched to 37°C
in order to block plasmid replication by the temperature-sensitive Pol I
allele. These conditions confer a dramatic growth advantage to cells that have integrated the error-prone Pol I mutants in their chromosome, such that most of the colonies that grow at 37°C are expected to be stable integrants that can be confirmed by PCR.
When this method does not yield desirable results, sucrose can be added in the selection step to actively select against the presence of the plasmid. A number of other methods for generating allelic substitutions that are currently practiced by those skilled in the art are contemplated.
Unrestricted Selection in Target DNA Used for Mutaeenesis Mediated by Error-prone Pol I Mutants The in vivo mutagenesis system contemplates a plurality of suitable templates, as diagrammed in FIG. 3. As one embodiment, the targeted template can be a prokaryotic gene (FIG. 3A) encoding a full-length protein or any fragments thereof.

As another embodiment, the targeted template can be a eukaryotic gene (FIG.
3A) that may be in the form of a cDNA or genomic DNA without introns, or any fragments thereof. For example, an eukaryotic gene in any of these forms may be exogenously introduced and unfaithfully replicated in prokaryotic strains that express endogenously temperature-sensitive or wild type Pol I. Subsequently, a library of mutated eukaryotic DNA may be isolated from their prokaryotic hosts by using, for example, conventional methods for plasmid preparation or for phage recovery.
Once the plasmid library is isolated, it may be subjected to restriction-enzymatic digestion with one or more endonucleases that are commercially available to once skilled in the art of molecular biology to dissociate the targeted gene of interest from the vector.
This can be easily done using gel electrophoresis to separate DNA fragments based on shape and size differences. Upon isolation and repurification, the gene library may be used for subcloning into eukaryotic expression vectors. The mutated versions of the target gene of interest may be transformed into eukaryotic cells or conventional cell-lines to screen for a desired property or function. Other procedures for manipulating DNA that are within the scope of a person skilled in the art are also contemplated. As one example, the x2913recA strain, which is defective in thymidine synthase (dthyA572), may be co-transformed with a vector encoding the human TS cDNA
placed downstream from on and a vector expressing an error-prone Pol I mutant.
Selection may be conducted in increasing concentrations of 5-flourouracyl ("SFU"), a potent inhibitor of thymidine synthase. Clones having increased resistance to SFU
can be isolated and their mutations characterized using materials and methods practiced by persons skilled in the art [Landis et al., Cancer Res. 61: 666-72 (2000)].
The template that is subjected to in vivo mutagenesis by Pol I mutators may be delivered in a circular form, such as a plasmid, or in a linear form. In one embodiment, the template may be delivered by a bacteriophage that can inject phage genomic material into a prokaryotic host expressing one error-prone Pol I
mutant.
When the bacteriophage depends on Pol I for replication, like for example, bacteriophage T4 in a RNAseH-deficient background [Hobbs, L.S. and Nossal, N.G.
J. Bacteriol 178: 6772-6777 (1996)], viral replication in the host results in expansion and mutagenesis of the template, generating a library of mutants.
As another embodiment, the targeted genes) can be exogenously introduced in a Pol I-dependent plasmid, including the plasmids ColEl, pBR322, plSA, pMB, pNT7, pVH5l, RSF1030, CloDFl3, ColE2, pLMV158, pLSI, and their derivatives.
The target plasmid is distinct from the plasmid encoding the error-prone Pol I
mutant, which is Pol I-independent. In its simplest form, the system comprises one plasmid encoding an error-prone Pol I mutant and another plasmid or phage encoding the targeted gene. Other methods of gene delivery that are known and practiced by persons skilled in the art are included in these embodiments.
As an example of the efficacy of error-prone Pol I random mutagenesis for enzyme modification, an isoform of the (3-lactamase gene ("TEM-1 ") representing a hypothetical gene of interest in place of any prokaryotic or eukaryotic gene, was inserted into a plasmid as a target for in vivo mutagenesis, as described in Example 2 provided below. For measuring the frequency of in vivo mutagenesis associated with error-prone Pol I expression, a (3-lactamase reversion assay decribed below, can be used. As illustrated in FIG. 3B, the (3-lactamase gene was inserted into template plasmids at an increasing distance from the position of ori, i.e. from the origin of ColEl plasmid replication ("pLA230," "pLA700," "pLA1400," "pLA2800," and "pLA3700"). In all constructs, an ochre stop codon (TAA) replaces glutamic acid (GAA) 26 codons downstream from the translation start of the TEM-1 [3-lactamase gene. For example, in the reporter plasmid pLA230, the ochre stop codon is positioned approximately 230 by downstream of the ori. Because most mutations within the ochre stop codon result in its reversion to a codon encoding an amino acid and thus allow for (3-lactamase expression, mutations can be scored as carbenicillin-resistant colonies. These may be confirmed by sequencing plasmids isolated from these colonies. To assess the mutagenic effect of error-prone Pol I on a ColEl plasmid sequence, the frequency of ochre reversion in cells expressing error-prone Pol I can be compared to the frequency in cells expressing wild type Pol I. To establish the mutation frequency within the chromosomal DNA, rifampicin resistance, arising from point mutations in the RNA Pol II gene encoded in the chromosome, can be scored [Ovchinnikov et al., Mol. Gen. Genet. 190: 344-348 (1983)].
The (3-lactamase gene can be subjected to in vivo mutagenesis in the presence of increasing concentrations of aztreonam, to select mutations that improve the (3-lactamase substrate recognition of aztreonam. The TEM-1 isoform of (3-lactamase is the most frequently occurring (3-lactamase found in Gram-negative clinical isolates, functioning in the catalytic degradation of (3-lactams by amide-bond hydrolysis [Wiedemann et al., J. Antimicrob. Chemother. 24 Suppl. B: I-22 (1989)]. The catalytic efficiency of [3-lactamase varies greatly depending on the (3-lactam used as a substrate. Whereas (3-lactamase degrades pencillin with high efficiency, hydrolysis of cephalopsorins by [3-lactamase is poor, and it has only residual activity against newer, extended-spectrum antibiotics. Extended-spectrum antibiotics include aztreonam and third-generation cephalosporins, such as cefotaxime and ceftazidime, that carry bulky adducts specifically designed to minimize recognition by (3-lactamase. FIG. 4 illustrates the structure of representatives of each of these (3-lactam substrates including penicillin, cephalosporin, and aztreonam. Aztreonam was used as a selective agent in Examples 4 and 7.
The evolution of the (3-lactamase, which occurs in nature under pressure from widespread use of extended-spectrum antibiotics in the clinics [reviewed in Medeiros, A., Clin. Infect. Dis. 24 Suppl. 1: S19-45 (1997)] can be simulated in the laboratory [Long-McGie et al, Biotechnol. Bioeng. 68: 121-5 (2000); Zaccolo, M. and Gherardi, E., J Mol. Biol. 285: 775-83 (1999); Orencia et al., Nat. Struct. Biol. 8: 238-(2001)]. From clinical isolates and laboratory selections, it is known that at least two specific amino-acid substitutions are required for substantial resistance to extended-spectrum (3-lactams [reviewed in Knox Antimicrob. Agents Chemoth. 39: 2543-(1995)]. Resistance mutations have been identified in a total of 37 different positions, which are listed in http://www.lahey.org~Jstudies/temtable/htm. These mutations fall broadly into two categories: one group of mutations, including the 8164 S/H
(arginine to serine or histidine) and the G238S (glycine to serine), occurs near the active site and increases the catalytic efficiency of the (3-lactamase enzyme.
Mutations belonging to the second group that include the E104K (glutamic acid to lysine) and the M182T (methionine to threonine), are typically positioned distantly from the active site. They are most likely selected in response to mutations in the catalytic site to suppress destabilizing effects resulting from these mutations [Petrosino J. et al., Trends Microbiol. 8: 323-327 (1998)].
Mutagenesis of the [3-lactamase gene coupled with aztreonam selection was chosen to validate the use of error-prone Pol I random mutagenesis for enzyme modification. The following two considerations were critical in this choice of target:

(1) studies of clinical isolates and selections in the laboratory establish that the (3-lactamase gene can be modified through mutation to recognize aztreonam as a substrate; and (2) at least two specific amino-acid substitutions are required to achieve substantial aztreonam resistance. This reduces the potential background contributed 5 by spontaneous mutations, as they are highly unlikely to generate more than one amino acid substitution because of its low frequency (10-~ in JS200 cells).
Method for in vivo Muta~enesis of Target DNA, Mediated by Error-prone Pol I
Mutants, in Combination with Selection or Screenine 10 With respect to the methods embodied for implementing in vivo mutagenesis mediated by error-prone Pol I mutants, the following provides descriptions of:
(1) the advantages and applications of error-prone Pol I-mediated mutagenesis in continuous culture, and (2) the parameters of the optimized culture conditions. Specific Examples relating to these embodiments are referenced and provided in the following 15 section.
Advantages and Applications of Error-prone Pol I-Mediated Muta~enesis in Culture for Enzyme Modification A method of random mutagenesis, followed by selection or screening allows the simultaneous exploration of sequence, structure, and functional space of an enzyme without any detailed mechanistic or structural information [Skandalis et al., Chem. Biol. 4: 889-98 (1997)]. Mutagenesis and selection/screening may be repeated in an iterative fashion so that the sequence space is directed toward changes that improve performance, in an approach referred to as "directed evolution."
Performing mutagenesis and selection simultaneously in continuous culture eliminates the need for recovering target DNA between selection steps.
Directed evolution still explores only a small fraction of the sequence space, so it is unlikely to achieve optimal solutions or to evolve new catalytic functions.
Larger leaps in sequence can be achieved by shuffling sequences within a library of mutants that has been prescreened for activity or within a family of enzymes [Smith, G. P., Nature. 370: 324-5 (1994); Crameri, A. et al., Nature. 391: 288-91 (1998)]. In one embodiment, the invention can also be used in combination with these shuffling technologies. The error-prone DNA Pol I mutants may produce additional mutations during these processes of shuffling in vitro and in vivo.
The error-prone DNA polymerases of the present invention have potential applications in the medical field, both in drug development and in gene therapy. For example, random mutagenesis has been instrumental in generating modified enzymes for the management of cancer [Encell L. P. et al., Nat. Biotechnol. 17: 143-7 (1999)].
These include mutants designed to reduce the side-effects of cancer therapy and mutants with improved pro-drug-activating properties. Examples of the former include: thymidylate synthase resistant to 5-flourouracil; 6-methyguanine-methyl transferase resistant to alkylating agents and to 6-benzoguanine; and dihydrofolate reductase resistant to metrotrexate. Examples of the latter include thymidylate kinase with enhanced specificity for gancyclovir or acyclovir, deoxycytidine-kinase mutants specific for Ara-C, and a cytosine deaminase specific for 5-fluorocytosine.
The error-prone Pol I mutants of the invention also have potential applications in promoting enzyme modification by random mutagenesis in areas as diverse as chemical synthesis, bioremediation, and food processing [reviewed in Chirumamilla et al., Mol. Cell Biochem. 224: 159-68 (2001); Schmidt-Dannert et al., Trends Biotechnol. 17: 135-6 (1999)]. The Pol I mutants of the invention may be used to generate enzymes with altered substrate specificity, including altered esterases, lipases, cytochrome oxidases, and dioxygenases, and enzymes with improved catalysis in non-natural environments, such as alkaline conditions, heat, cold, and/or presence of metal chelators. For example, Pol I mutants may be used to generate enzymes with altered substrate specificity to improve the efficiency and enantiomeric purity of chemical synthesis. Enzymes resistant to heat and to the presence of metal chelators may be obtained by the error-prone Pol I mutagenesis to improve rates of biocatalysis in chemical synthesis. Pol I mutants may be also used to generate enzymes that confer resistance to alkaline conditions for degradation of dyes in laundry, or resistance to cold conditions in food processing applications.
The present invention has foreseeable applications in emerging fields, including biotechnology, gene therapy, and bioremediation. The customization of substrate recognition by DNA-specific reagents, such as recombinases and restriction enzymes [Buchholz et al., Nature Biotechnol. 19: 1047-52 (2001); Santoro et al., Proc. Natl. Acad. Sci. U S A, 99: 4185-90 (2002)], or the improvement of enzymes for alkane degradation [Belhaj et al., Res. Microbiol. 153: 339-344 (2002)], are two examples of emerging areas that can benefit from the use of the embodiments of the present invention.
In vivo Muta~enesis and Selection/Screenin~ Under Optimized Culture Conditions A method for generating enhanced mutagenesis that can be targeted to a sequence of interest is provided, in which error-prone Pol 1 mutants bearing amino-acid substitutions in the motif A and/or B, for example, including amino-acid substitutions at positions I709N/F and A759R, are expressed in host strains under carefully controlled culture conditions, as described below in Example 2. In one embodiment, the method comprises transforming a prokaryotic host strain with an expression vector having any of numerous forms that encodes one or more error-prone Pol I mutants, and transforming a second vector, such as a plasmid that encodes one or more target genes. In another embodiment, the method comprises incorporating one or more copies of genes or fragments that encode error-prone Pol I
within the chromosomes of cell lines or host strains, and subsequently transforming the population of cells expressing the mutants with a vector, such as a plasmid, that encodes one or more target genes. In one embodiment, controlled conditions comprise growing a culture of transformed cells in a rich media and letting the culture reach and maintain a stationary growth phase, such as a growth stage obtained hours after inoculation. As a preferred embodiment, culture conditions by which error-prone Pol I achieves optimal mutagenic performance include: (1) growing the transformed host strain cultures at a cell density indicative of exponential growth (e.g., OD600~0.5); (2) diluting the culture to a low concentration (e.g., 1:105) in nutrient-rich medium (e.g., 2XYT) pre-warmed at 37 °C; (3) incubating the culture for an additional time (e.g., 15 min. at 37 °C); and (4) placing the culture in a 37 °C
shaker to reach a point of saturation (e.g., 15 hours).
As another embodiment, the error-prone Pol I mutants may be used to produce labeled plasmid DNA or fragments thereo, in vivo. For example, one or more labeled nucleotide or nucleoside analogs may be added to the culture medium to be incorporated into the replicated copies of target plasmid DNA. As a related embodiment, the error-prone Pol I mutants may be purified from cell extracts as recombinant proteins, and added to in vitro reactions containing a target sequence for microarray analysis or other high-throughput DNA labelling procedures or for sequencing purposes. Modified deoxynucleoside triphosphates ("dNTPs") or nucleotide analogs, including dATPs, dGTPs, dCTPs, dTTPs, dUTPs, and dITPs, and related analogs may be added in any combination in in vitro DNA polymerization reactions [click et al., Biotechniques 33(5): 1136-42 (2002)]. These nucleotides and nucleosides may be modified by covalent attachment of groups for detection by fluorescence (rhodamine green, fluorescein), chemiluminescence (biotin), or radioactivity (y[32P]). The nucleoside and nucleotide analogs may be synthesized by known methods in the art.
Using the compositions and methods provided, error-prone Pol I mutants bearing mutations in motif A and/or motif B can increase the frequency of mutagenesis, between 4 and 5 orders of magnitude above that of wild type Pol I, with up to a 400-fold differential relative to the chromosome, as shown in Example provided below. The compositions and methods permit mutagenesis targeting nucleotides located distantly from the on including a distance of 3700 bp, far exceeding the 400 by limit reported by previous investigators. Thus, a broad variety of prokaryotic and eukaryotic genes can be targeted by the present invention.
The compositions and methods of the invention permit the targeting of the sequence of interest, resulting in a diverse mutation spectrum and a wide distribution of mutations, as shown in Example 4 provided below. The mutagenesis system of the present invention provides the additional advantage, with respect to existing mutator strains, so that the mutagenesis system representing one embodiment of the current invention may be readily adapted to different strains, as shown in Example 6 provided below. Suitable host strains include: (1) strains endogenously expressing a temperature-sensitive DNA Pol I allele, (2) a wild type DNA polymerase; and (3) both wild type DNA pol and wild type DNA repair phenotypes. To adjust the optimal expression level of error-prone Pol I in various strains for achieving random mutagenesis, inducible promoters may be used. Other embodiments of the in vivo mutagenesis system are directed to a large number of existing strains known to be deficient in specific enzymes and amenable to complementation such as the x2913recA strain defective in thymidine synthase described above.
The embodiments of the compositions and methods permit the accumulation of beneficial mutations within the target gene, in a continuous culture. Two independent selections with increasing concentrations of aztreonam were carried out under optimized mutagenic conditions to obtain 23 sequences representing (3-lactamase variants from each selection, as shown in Example 4 and 7, provided below. Two known amino acid substitutions were identified in these selections:
E104K (glutamic acid changed to lysine at position 104) and R164H (arginine changed to histidine at position 164) in selection 1 and to S (arginine changed to serine at position 164) in selection 2. In addition, a novel mutation, the (glycine changed to arginine at position 267) was identified in selection 1.
The mutations can occur sequentially in this system as exemplified by the introduction of the G267R mutation following the E 104K and R 164H mutations, which can be found in a single clone that encoded a silent mutation also present in all the other plasmids carrying the E104K and R164H mutations that were sequenced. No other silent mutations were found. Table 4 in Example 7 presents the aztreonam phenotypes corresponding to different combinations of these amino acid substitutions after subcloning into vectors that have not been replicated by error-prone Pol I
mutants to ensure that no other mutations were present. FIG. 10 shows dose-response curves of aztreonam resistance conferred by a representative subset of these aztreonam resistance mutations. All non-synonymous mutations contributed to aztreonam resistance. The almost complete absence of irrelevant mutations is consistent with an estimated low mutation load and suggests that each mutation arises sequentially, in a process that simulates natural evolution of resistance to (3-lactamase antibiotics [Negri et al., Antimicrob. Agents Chemother. 44: 2485-2491 (2000); Vuye et al., Antimicrob.
Agents Chemother. 33: 757-761 (1989)]. Relatively mild mutagenesis conditions are more than offset by the large size of the pool (approximately 10' ~ plasmids) and by the discriminating power of iterative functional selection. Finding three relevant mutations in a single (3-lactamse sequence, with probability of 10-~°
[Long-McGie et al., Biotechnol. Bioeng. 68: 121-5 (2000)], is an illustration of the efficient exploration of sequence space allowed by the embodiments of the present invention.
The compositions and methods can be used for prognostic purposes in predicting the evolution of novel resistance mutations, such as the G267R, based on the fact that the selection step can identify the most common mutations found in clinical isolates. The embodiments of the invention are unlike most other mutagenesis approaches that yield a more biased representation relative to naturally occuring mutants [See Orencia et al., Nat. Struct. Biol. 8: 238-242 (2001)].
Mutations at positions E104 and 8164 are the most frequently found in naturally-occurring clinical isolates (http://www.lahey.or~/studies/temtable.htm). This is consistent with 5 a diverse and well-distributed mutation spectrum that approximates the endogenously occuring mutations [Schaaper R., R.M. and Dunn, R. L. Genetics 129, 317-326 (1991)]. No mutation had been previously reported at position 6267, thus the represents a novel determinant that confers aztreonam resistance. This mutation is located in the loop connecting the BS (3-sheet to the N-terminal >-I11 helix, that 10 undergoes a shift in the presence of the third most frequent mutation in clinical isolates, the G238S (glycine changed to serine at position 238).
To provide support for these embodiments, experimental data are provided in Examples 2 through 7, which follow at the end of this section. It should be appreciated that, although specific embodiments of the invention are described herein 15 for purposes of illustration, various modifications may be made without departing from the spirit and scope of the invention. Accordingly, the invention is not limited except as by the appended claims.
EXAMPLES

20 The following Examples demonstrate the embodiments of the invention in various contexts. In Example 1, methods for constructing low-fidelity E. coli Pol I
mutants engineered in motif A and/or B are provided. In Example 2, culture conditions optimized for elevated frequency of plasmid-specific in vivo mutagenesis are provided. In Example 3, the synergistic effect on the efficacy of plasmid targeting in vivo mediated by the D424A I709N A759R Pol I mutant and its sensitivity to culture conditions is provided. In Example 4, the spectrum and frequency of mutations introduced into target sequence/(3-lactamase upon D424A I709N A759R
Pol I mutant expression are provided. In Example 5, the frequency of mutations in a target plasmid as a function of the distance from the origin of replication (ori) is provided. In Example 6, the effect of D424A 1709N A759R Pol expression on the frequency of in vivo mutagenesis in host cells that endogenously produce wild type Pol I and are proficient in DNA repair is provided. In Example 7, an example of directed evolution under selective pressure such as increasing concentration of aztreonam is provided.

Construction of Error-prone Pol I Mutants Construction of plasmids encoding error-prone Pol I mutants of the present invention used in Examples 1-7 were made in the following way. E. coli DHSapolA
was amplified by colony polymerase chain reaction with 59-ATATATATAAGCTTATGGTTCAGATCCCCCAAAATCCACTTATC-39 and 59-ATATATATGAATTCTTAGTGCGCCTGATCCCAGTTTTCGCCACT-39 as primers. To create pECpol I, the 3-kilobase pair amplified fragment was digested with Hindlll and EcoRl, and then cloned under the lactose promoter into pHSG576, a low copy-number plasmid that has a pol 1-independent origin and chloramphenicol as a selectable marker. Site-directed mutagenesis was performed on pECpol I to introduce silent mutations C to A at position 2,067 and G to C at position 2,214 of the polA gene to create Accl and Eagl sites, respectively, that flank the sequence encoding motif A. The resulting plasmid was named pECpol IS.
Plasmids encoding the I709N and I709F motif A mutants of DNA polymerase were isolated from a Pol I library by genetic selection. A random library was constructed by annealing two single-stranded DNA oligonucleotides containing segments with random sequences: Oligo 1 was a 104-mer corresponding to the sense nucleotides 2,053-2,156, and containing an AccI site for cloning (5' GAAGGTCGTCGTATACGCCAGGCGTTTATTGCGCCAGAGGATTAT
[GTGATTGTCTCAGCGGACTACTCGCAGATTGAACTGCGC]ATTATGGCGC
ATCTTTCGCG-3'); Oligo 2 was a 89-mer corresponding to anti-sense strand nucleotides 2,225-2,137 and containing an EagI site (5'-AACACTTC
TGCGGCCGTTGCCCGGTGGATATCTTTTCCTTCCGCGAATGCGGTCAGCAA
GCCTTTGTCACGCGAAAGATGCGCCATAAT-3'). The bracketed nucleotides in Oligo 1 were synthesized to contain 88% wild-type nucleotide and 4% each of the other three nucleotides at every position. The 20-base pair complementary regions of hybridization are underlined. Oligo 1 and Oligo 2 were annealed at their nonrandom complementary regions by mixing 250 pmol of each in 20 pl of H20 and heating to 95 °C for 5 min, followed by cooling for 2 h to room temperature. The partially duplex oligonucleotide was extended by incubation with 50 units of E. coli pol I
Klenow fragment (New England BioLabs, Beverly, MA) for 2 h at 37 °C in a 0.3-ml reaction mixture containing 10 mM Tris-HCI, Ph 7.5, 5 mM MgCl2, 7.5 mM DTT, and 0.5 mM of all four dNTPs. The resulting DNA was digested with Accl and Eagl, purified, and inserted into pECpoldum in place of the stuffer fragment.
Plasmids containing the random library were transformed into E. coli XLl-Blue, and the number of transformed cells was determined by plating an aliquot onto LB agar plates containing 30 pg/ml of chloramphenicol. The remainder of the library was amplified by growing the transformed E. coli XL1-Blue in 3 liters of 2XYT medium for 16 h at 37 °C, and the random library, pECpolLib, was then purified. The pECpolLib was transformed into JS200 cells (with plasmids pHSG576, pECpollS, pECpoldum as controls) and selected for survival on nutrient agar plates at 37 °C
overnight. Only paired samples containing less than 1,500 colonies at 30 °C were analyzed because dense plating of the cells leads to elevated background at 37 °C.
Approximately, 280 colonies that grew at 37 °C were randomly picked and sequenced and found to contain between 1 and 2 amino acid substitutions.
The A759R and S756E amino acid substitutions in motif B were introduced by site-directed mutagenesis into plasmids I709N, 1709F, D424A I709N and D424A
I709F to generate the mutants I709N A759R, D424A I709F A759R, D424A DI709N
A759R, and D424A I709F A759R. For site-directed mutagenesis, QuickchangeT"~
(Stratagene~ La Jolla, CA) was used with the following mutagenic primers: Pol I
S756E-F S'-CCAGCGAGCAA CGCCGTGAGGCGAAAGCGATCAACTTTGG-3 ;
Pol I S756E-R 5'-CCAAAGT
TGATCGCTTTCGCCTCACGGCGTTGCTCGCTGG-3'; Pol I A759R-F 5'-CGCCGTAGCGCGAAACGGATCAACTTTGGTCTGATTTATGGC-3'; Pol I
A759R-R 5'- GCCATAAATCAGACCAAAGTTGATCCGTTT
CGCGCTACGGCG-3'.
To deplete exonuclease function, site-directed mutagenesis on pECpol I to introduce an A-to-C transversion at position 1271, thus changing Asp424 to Ala and inactivating the 3'-5' exonuclease activity was performed [Derbyshire et al., Science 240: 199-201 (1988)]. In the same manner, the 3'-~5' exonuclease domain-inactivating mutation was introduced into the plasmids encoding the I709N and mutations to generate D424A I709N and D424A I709F mutants, respectively.

Optimization of Culture Conditions FIG. SA provides an example of "initial" culture conditions and a method for inducing in vivo mutagenesis by expression of error-prone Pol I that is representative of one conventional method used by those skilled in the art prior to the development of "optimal" culture conditions and methods embodied in the invention. There are two previous reports of a 3-10 fold increase in plasmid mutagenesis relative to chromosomal mutagenesis using culture conditions similar to initial conditions. In one case, the polymerase used was Pol I containing an inactivated proofreading domain [Fabret et al., Nucleic Acids Res. 28: E95 (2000)]. The other example developed by the present inventors used the D424A I709F as an error-prone Pol I
[Shinkai, A., and Loeb, L.A. J. Biol. Chem. 276: 46759-64 (2001)]. The improvement in the frequency and in the targeting of mutagenesis were attributed to the following four variables: (1 ) dilution of the initial inoculum (to at least 1:105), (2) growth of cells in 2XYT, (3) pre-warming cells to 37 °C, and (4) continued growth of the cultures past the point of saturation (typically I S-17h).
RESULTS: Example 2 illustrates the individual culture parameters found to have an effect on mutagenesis by error-prone Pol I expression. Ideally, a system for random in vivo mutagenesis should exhibit frequent mutations throughout a targeted gene, from the 5'end to the 3'end, and as few mutations that are non-target specific.
Culture conditions were varied to achieve optimal mutagenic performance according to these criteria. Cultures should be allowed to reach saturation. Other variables that contributed to the efficacy of random mutagenesis included the type of culture medium, the temperature of the medium at the time of inoculation, and the initial inoculum concentration. Careful optimizations of culture conditions achieved by trial and error are critical for obtaining mutagenesis resulting in a high target specificity and broad distribution of mutations.

The JS200 strain was initially described as SC18-12 [Witkin, E. M. and Roegner-Maniscalco, V., J. Bacteriol. 174: 4166-8 (1992)], and has the following genotype: SC-18 recA718 polAl2 uvrA155 trpE65 lon-11 sulAl. SC-18 carries tetracycline resistance and is insensitive to ~, phage. PolAl2 is a temperature sensitive allele of Pol I [Monk, M. and Kinross, J., J. Bacteriol. 109: 971-8 (1972)].
It appears that polAl2 is misfolded in a way that perturbs the spatial arrangement of polymerase and 5'-~3' exonuclease functional sites, resulting in miscoordination between these two activities (as seen in a nick translation assay). Elevated temperatures further increase misfolding and completely disrupt effective polymerization at a nick [Uyemura et al., J. Biol. Chem. 251: 4085-9 (1976)]. Rec718 allele originated from an unstable recombinant obtained in a conjugation between a recA441 K12 donor and a rec A+ B/r-derived recipient [Witkin, E. M., Mol. Gen. Genet. 185: 43-50 (1982)]. It carries two base substitutions, one that allows the RecA protein to become constitutively activated, and one partial suppressor [(McCall et al., J.
Bacteriol. 169:
728-34 [1987)].
Preparation of reporter plasmids can be described as follows. The reporter plasmids to measure the reversion frequency of the (3-lactamase gene were modifications of plasmid pGPS3 (New England Biolabs Inc., Beverly, MA). This plasmid contains a pUCl9 (ColEl-type) origin of replication and the npt and amp genes for kanamycin and carbenicillin selection. To measure reversion frequencies, a G-to-T transversion at position 76 of the (3-lactamase gene was introduced by site-directed mutagenesis, changing the codon GAA for G1u26 to the ochre stop codon TAA. The reporter plasmids used to measure (3-lactamase reversion frequency were pLA230, and pLA2800, carried 230 by and 2800 by downstream from the pUC 19 origin of replication [Shinkai, A. & Loeb, L.A., J.Biol. Chem. 276, 46759-(2001)]. Plasmid pGPSori corresponds to pLA230 but with the wild-type (3-lactamase gene instead of the interrupted version. It was obtained in the same manner as pLA200, but in this case pGPS3 was used as the template for primers 5'-GCACCCGACATACATGTCCTATTTGTTTATT-3' and 5'AAACTTGGTCGGTACCTTACCAATGCTTAATC-3.' The pLA2800 corresponds to pGPS3 encoding an ochre stop codon at position 76 of the [i-lactamase gene (2748 by from the origin of replication).

Preparation of competent cells can be described as follows. Single colonies growing on LB plates with appropriate antibiotic selection were picked and incubated overnight without shaking at 30 °C in 50 ml of LB plus antibiotic.
After the bacterial culture was shaken for 1 h at 30 °C, it is transferred to a flask containing 450 ml LB
5 antibiotic, and left in the 30 °C shaker for 3-4 hours (to an OD6oo of 0.5-1). Cells were chilled on ice for 20 minutes, pelleted in a Sorval~ RC SB Plus centrifuge and washed twice in 10% glycerol. The pellet was resuspended in ~2m1 10% glycerol, aliquoted in 120 pl aliquots, and quick-frozen on dry ice. To avoid outgrowth of revenants and/or suppressors, competent cells were made using the PoII plasmid first, 10 and subsequently transformed with the reporter plasmids.
Transformations were performed using a A Biorad Gene Pulser TM apparatus (set to 400 S2, 2.20 V and 2.5 pFD). After the electric shock, cells were resuspended in 1 ml LB media, left to shake for lh at 30 °C, and plated in LB
plates with the appropriate antibiotics for plasmid selection. Single colonies were picked into Sml 15 LB with the appropriate antibiotics and grown overnight at 30 °C
with no shaking. In the case of XL1 Blue cells, since temperature or media is not an issue, a more standard protocol involving the resuspension in 2XYT medium, shaking at 37 °C for minutes, and growth of single colonies in LB media in a 37 °C shaker for 16h is followed. Transformation with a reporter gene results most frequently in the loss of 20 the temperature-sensitive phenotype. Replication of reporter plasmids (that are Pol ( dependent and multicopy) likely depletes the limited Pol I functional reserves of JS200 cells even under permissive conditions, favoring the outgrowth of revenants and/or suppressors. Hence, the plasmid carrying Pol I was used for the first transformation to make competent cells, which were used subsequently to transform 25 the reporter plasmids.
The (3-Lactamase reversion assay was performed in the following way. E. coli JS200 transformant strains carrying Pol I (wild type or different low-fidelity mutants) in the pHSG576 plasmid (chloramphenicol-resistant (cm')) were retransformed with a reporter plasmid (kanamycin (kan')). Single colonies exhibiting double resistance 30 (cm', kan') were picked into 5 ml LB (plus tet, kan, cm) and grown overnight without shaking. After 1 hour shaking at 30 °C, the cultures (OD~ 0.5) were diluted 1:105 in 5 ml 2XYT media pre-warmed at 37 °C. These new cultures were maintained at 37 °C

for 15 minutes and incubated for 15 hours in a 37 °C shaker. The (3-lactamase stop codon reversion was detected by plating (in duplicate at least) the appropriate dilution of the saturated cultures onto plates containing SOpg/ml carbenicillin (Island Scientific, Bainbridge Island, WA), in addition to kanamycin and chloramphenicol.
For rifampin resistance, cells were plated in 25 pg/ml rifampin. The results are expressed as frequency relative to viable colonies (grown on kan cm alone) and represent the average of at least two clones that are carried independently.

Synergistic Effect of Motif A and Motif B Mutations in vivo and Significant Targeting of Plasmid Sequences Table 1 below summarizes the results of assays characterizing the frequency and the degree of mutagenesis targeting (results shown in FIG. 6). JS200 cells were transformed with either wild type Pol I or respective Pol I mutators, and reporter plasmid pLA230. The frequency of (3-lactamase reversion (normalized to wild type) is indicative of the frequency of mutagenesis in a given target gene. The ratio of rifampicin resistance relative to carbenicillin resistance reflects the frequency of chromosomal mutagenesis.
Initial conditions represent cultures grown at 30 °C to OD600~0.5 that were diluted 1:50 in nutrient broth and allowed to reach OD=0.7 in the 37 °C
shaker [Shinkai, A. and Loeb, L.A., J. Biol. Chem. 276: 46759-64 (2001)]. Optimized conditions represent cultures grown under permissive conditions to OD600~0.5 that were diluted 1:105 in 2XYT pre-warmed at 37 °C, incubated an additional 15 min. at 37 °C, and placed in the 37 °C shaker for 15 hours (saturation).
Mutagenesis frequencies were determined by (3-lactamase reversion assay (target within plasmid) or by scoring rifampin-resistant colonies (target within chromosome), and the reported values have been normalized against the wild type. The relative frequency of plasmid versus chromosomal mutagenesis for each mutant polymerase was obtained after normalizing to that found for the wild type polymerase. This incorporates variable such as the different nature of the assays (forward assay (Rif ) versus (3-lactamase reversion assay), and the difference in target numbers (single copy (Riir) versus a multicopy [3-lactamase).
Table 1 MutagenesisTargeting PoII Initial OptimizedInitialOptimized wild type 1 1 1 1 D424A I709N 4.4 x 1 2.5 x 12.5 413 A759R l.SxlO~ ND 14.2 ND

D424A A759R 3.0 x102 5.8 x102 4.9 68 I709N A759R 1.8 x103 ND 0.54 ND

D424A1709N A759R3.6 x103 1.5x104 1.0 390 RESULTS: Example 3 characterizes mutagenesis associated with expression of different error-prone Pol I mutants and its sensitivity to culture conditions (FIG. SA
and SB). To establish the effect of low-fidelity mutations in motif A and in motif B, alone and in combination the following Pol I mutants were tested: D424A 1709N, A759R, D424A A759R, I709N A759R and D424A I709N A759R (a subset of which is presented in FIG. 6). Table 1 summarizes these results. The effect of D424A
I709N expression under initial conditions is comparable to published data [Shinkai, A., and Loeb, L.A. J. Biol. Chem. 50: 46759-64 (2001)]. The expression of the D424D A759R mutation results in a frequency of mutagenesis that compares to that of D424A I709N. Similar to motif A mutations, elevated mutagenesis is dependent on the simultaneous inactivation of the proofreading domain. These results confirm that position A759 in motif B is a critical determinant of fidelity in E.
coli. Strikingly, expression of the 1709N A759R D424A triple mutant results in mutation frequencies that are more than additive, 6 to 26 times above those of each double mutant (same applies to I709F A759R D424A). Unlike single motif A or motif B mutants, the mutagenesis by double motif A and motif B Pol I mutants shows little depence on inactivation of proofreading. This suggests that the combination of mutations at positions I709F and A759R leads to the functional inactivation of proofreading.
Under optimized conditions, frequencies of mutagenesis 15,000 times above wild type were achieved (in other experiments, up to a 100,000-fold increase has been observed (FIG. 8B)). This represents a substantial increase over the frequency achieved under "initial" conditions (3,600-fold). Optimized conditions also further restrict mutagenesis to the target plasmid. In the case of the I709N A759R D424A triple mutant, plasmid specificity increases by 400-fold. In sum, I709 (motif A) and (motifB) mutations synergistically enhance target plasmid mutagenesis and both the frequency and targeting of mutations are sensitive to growth conditions.

Determinations of the Spectrum and Frequency of Mutations Introduced in a Target Sequence upon Expression of D424A 1709N A759R Pol I
Table 2 below shows the mutation spectrum of mutations located outside the ochre stop codon in the (3-lactamase gene identified in cells expressing D424A

IS A759R Pol I. The numbers within the heading represent the location of the sequence relative to the origin of replication. Given the nature of the assay, which required functional expression, frameshifts and deletions were not determined.
Table 2 Mutations 100-750 Transitions AT~GC 16 34.7 GOAT 21 44.9 Transversions AT~TA 6 14.3 AT-~CG 1 2.0 GC~TA 2 4.1 GC-~CG 0 0 RESULTS: Example 4 provides an illustration of the mutagenic spectrum resulting from error-prone Pol I expression. To characterize the nucleotide changes associated with I709N A759R D424A Pol I expression, 158 carbenicillin-resistant clones were sequenced. All 155 of these clones showed mutations in the ochre codon.
Of these, only 2 changed to another stop codon (both to opal), which gives an overall specificity of 96.8% for the reversion assay. Wild type sequence was present in 148 of the remaining 153 sequences, in varying amounts, from trace amounts to about 90%, which is consistent with mutations encoded in a multicopy plasmid with diverse degrees of dominance. To establish a complete spectrum of base pair substitutions, good-quality sequence of at least 650 by for all these 158 clones was obtained. A
total of 46 secondary mutations were detected within this 650 by interval. Of these, 40 were located within the open reading frame (ORF), of which 22 were silent.
Only 2 mutations were detected more than once in this analysis, indicating a high degree of diversity. Table 2 shows the mutation spectrum of these 46 secondary mutations.
The expression of D424A I709N A759R Pol I results in a diverse mutation spectrum within the 650 nucleotides of this analysis, although a bias toward transitions (80%) was detected, with a predominance of GC-SAT mutations (56%).
The number of target plasmids present in cells expressing I709N A759R
D424A Pol I was determined in order to estimate the mutation load and found to be of only 10 copies/cell (FIG. 7). This is in contrast with 100 copies/cell in cultures IS expressing the wild type Pol I. This effect of the Pol I point mutations on target plasmid copy number decreases the size of the mutant libraries by 10-fold.
However, this reduction in the number of targets per cell may have improved the selection efficiency, as each new mutation is expected to represent a larger fraction of expressed protein, and is therefore more likely to have a larger impact on the phenotype.
Under optimized mutagenesis conditions, the frequency of stop codon reversion was in the order of one in 500 cells (FIG. 6). Given that 10 target , plasmids/cell were estimated, this translates into one reversion in 5000 codons. Since 2 of the 9 possible base pair substitutions at the ochre codon are not permissible, that is the equivalent of 1 mutation in 3,900 codons. Assuming an even distribution of mutations, i.e., that mutations at the ochre stop codon are representative of those of every other codon in the protein, results an estimated mutation frequency of 1 mutation in I 1,500 by (3 x 3900). The highest mutation frequency reported in vivo is 0.5 mutations/Kbp [Greener, A. et al., 7: 189-95 Mol. Biotechnol. (1997)], but this mutagenesis was unstable upon continued culture and had severe deleterious effects on the host [Greener, A. et al., 7: 189-95 Mol. Biotechnol. (1997)].
JS200 cells expressing the D424A I709N A759R polymerase and with the reporter plasmid pLA230 were grown under optimal conditions for mutagenesis and plated on plates containing 50 pg/ml carbenicillin. Single carbenicillin-resistant colonies were grown. Their plasmids were extracted and the sequence at the ochre codon was determined using oligonucleotide Blac-5. Secondary mutations were confirmed using the following oligonucleotides:
5 Blac-5 5'-TTACGGTTCCTGGCCTTTTGC-3';
Blac-6 5'-GGTTGAGTACTCACCAGTCAC-3';
Blac-7 5'-TCCGATCGTTGTCAGAAGTAA-3'; and Blac-8 5'-CCATTTCCACCCCTCCCAGTT-3'. The sequence was analyzed using SequencherT"~ software.

In vivo Mutagenesis by Error-prone Pol I Mutants as a Function of Distance from the Origin of Replication FIG. 8B demonstrates the frequency of mutagenesis as a function of distance from the origin of replication (ori), determined to assess the level of mutagenic coverage by error-prone Pol 1 mutants containing single and double mutations in the active site. The JS200 cells were transformed with either wild type or with different error-prone Pol 1 mutants and one of the following reporters with increasing distance between the on and the position of the ochre stop codon: pLA230, pLA700, pLA1400, pLA2800, and pLA3700. Cultures grown under permissive conditions to OD600~0.5 were diluted 1:105 in 2XYT, pre-warmed at 37 °C, incubated an additional 15' at 37 °C, and placed in the 37 °C shaker for 15h (saturation). Results show one representative experiment. Each point represents the average of two independent cultures, each plated in duplicate.
RESULTS: Pol I has been reported to synthesize 400 by in the leading strand downstream of ColEl origin [Sakakibara, Y. & Tomizawa, J., Proc. Natl.
Acad. Sci. U S A 71: 1403-1407 (1974)], at which point, replication of both leading and lagging strands is taken over by the Pol III replisome [Staudenbauer, W.L., Mol.
Gen. Genet. 149: 151-158 (1976)]. Thus, most Pol I-dependent mutations were expected to cluster in the 5' end of the target gene. Contrary to this expectation, a similar frequency of hits and a comparable spectrum was detected when mutations further downstream than 500 by were compared to mutations in the sequence corresponding to initiating leader. A decrease in mutagenesis became apparent, however, if the ochre stop codon in the reporter gene was placed further from the origin of replication as shown in FIG. 3B (also Example 5, illustrated in FIG.
8), resulting from the absence of leading-strand synthesis by error-prone Pol I at longer distances. The observed decrease in plasmid mutagenesis at the 700 by point, however, was only moderate and in the case of the triple Pol I mutant mutation frequency remained constant four orders of magnitude above background for another 3 Kb. Significant plasmid mutagenesis by Pol I beyond the sequence corresponding to the initiating leader strand in the plasmid sequence was not expected by a person skilled in the art. Thus, Example 5 demonstrates that target sequences at a distance of at least 3700 by from the on are amenable to mutagenesis by error-prone Pol I
expression.
The preparation of pLA230 and pLA2800 are provided in Example 2, and the 1 S other reporters were derived from plasmids pLA2800 and pGPS~ as follows:
The pLA3800 was generated by amplification of the entire npt (kan~) gene using the synthetic oligonucleotides 5'-CATCGGTACCTTAACCAATTCTGATTAGAAAAAC-3' and 5'-GATGGGTACCCTAGATTTAAATGATATCGGATCC-3' containing flanking Kpnl restriction sites. The 980 by amplified fragment was cloned into the Kpnl site of plasmid pLA2800, adding one extra copy of npt and moving the stop codon to by of the plasmid origin of replication. The pLA2200 resulted from excising of a Sal I 1517 by fragment from plasmid pLA3800 and religating the plasmid backbone.
This brought the stop codon to 2174 by of the plasmid origin of replication.
The pLA1400 was generated by cloning a PCR-amplified npt gene into the Sacl and Aflll sites of pLA2800, bringing the (3-lactamase ochre codon to 1403 by from the ori. The oligonucleotides containing Sacl and AfIIII adapters used for amplification were 5'-CATCGAGCTCTTAACCAATTCTGATTAGAAAAAC-3' and 5'-GATGACATGTCTAGATTTAAATGATATCGGATCC-3'.
The pLA700 resulted from subcloning the PCR-amplified (3-lactamase reporter into Swal and Spel sites of pGPS 0. The oligonucleotides used were 5'-CATCGATATCTTACCAATGCTTAATCAGTG-3' and 5'-GATGACTAGTCCCTATTTGTTTATTTTTCT-3', containing EcoRV and Spel adapters respectively. This places the ochre stop codon in the (3-lactamase gene only 709 nucleotides away from the origin of replication.

The Effect of Error-prone Pol I mutants on in vivo Mutagenesis in Cells Expressing Wild-type Pol I
FIG. 9 (graphical illustration of data in Table 3) demonstrates the plasmid-specific mutation frequency mediated by error-prone Pol I mutants in a Pol I
proficient strain, to determine if expression of chromosomal wild type Pol I
interferes with error-prone Pol I mutagenesis. XLl-Blue cells, a strain that expresses wild type Pol I, were transformed with Pol I mutator (D424 I709N A759R) and reporter plasmid pLA230. They were grown at 30 °C to OD~0.5 and a 10-5 dilution of which was grown for 17 hours in 2XYT medium with or without 100 pg/ml IPTG, and plated in triplicate on carbenicillin plates. The frequencies of (3-lactamase reversion in XL1-Blue cells and in JS200 cells, with or without the same concentration of IPTG, were plotted for comparison.
Table Strain Pol I IPTG Frequency of Carb'Normalized per l Ox cells frequencyb JS200 wild type no 4.20x1012.82 x10 1.00 JS200 D424A I709Nno 3.08 x10510.61 7.4 x103 x105 JS200 wild type Yes 2.70x100.45 x10 0.66 JS200 D424A I709NYes 2.88 x1050.42 x1056.9 x103 XLI-Blue wild type No 1.00 x104 (1.9x1032.4 x102 XL1-Blue D424A I709NNo 1.07 x10514.0 x1042.6 x103 XL1-Blue wild type Yes 2.13 xl OSt9.0 5.1 x102 x103 XL1-Blue D424A 1709NYes 1.36x10612.6 x104 3.3x104 RESULTS: The critical genetic material is provided in a vector and expression of wild-type Pol I and/or intact DNA repair in the host do not significantly interfere with mutagenesis. Thus, random mutagenesis may be performed in strains specifically designed for complementation by a targeted gene. FIG. 9 and Table 3 (in Example 6) illustrates this point demonstrating that error-prone Pol I mutants are co-dominant in a host expressing wild-type Pol I and proficient in DNA repair (XL-Blue). Pol I overexpression of error-prone Pol I increases target plasmid mutagenesis in these cells, by competing with the wild type Pol I for replication of the target plasmid. If overexpressed, error-prone Pol I can be expected to outcompete the wild type. This is illustrated in Fig. 9, which shows that overexpression of of I709N A759R Pol I in XL-1 Blue cells (by inducing expression from the tac promoter with IPTG) further increases plasmid mutagenesis to frequencies comparable to those obtained in JS200. The higher background relative to JS200 that was observed may be due to the GInV mutation (an amber suppressor with some activity against ochre) and therefore should not affect this conclusion. Rifampicin resistance was comparable between the two strains, indicating Pol I mutagenesis is still plasmid-restricted in XL1-Blue cells.
XL1-Blue cells are F'::TnlO proA+B+ laclq d(lacZ) M15/recAl endAl ~rA96(Nalr) thi hsdRl7 (rk mk+) glnv44 relAl lac and are commercially available (can be purchased from Statagene~ for example). Cells were grown under appropriate antibiotic selection: tetracycline (Sigma~ St.Louis MO) at 12.5 pg/ml, chloramphenicol (Sigma~ St.Louis MO) at 30 Pg/ml, and/or kanamycin (Island Scientific, Bainbridge Island, WA) at a concentration of 50 pg/ml.

Phenotypic Analysis of Aztreonam Resistant Mutations Produced by Directed Evolution Table 4 presents the phenotypes of the TEM-I mutations that were identified as a result of the selections performed to establish the applicability of the system to directed evolution of enzymes. JS200 cells that were co-transformed with a plasmid encoding D424A 1709N A759R Pol I and a second plasmid encoding the target TEM-1 (3-lactamase, were subjected to selective pressure under increasing concentrations of aztreonam. Controls included the following transformations: (a) wild type Pol I with wild type (3-lactamase; (b) D424A I709N A759R Pol I with a plasmid devoid of (3-lactamase; and (c) wild type Pol I with a plasmid devoid of (3-lactamase. Two independent selections were carried out under optimized mutagenic conditions.
After one round of mutagenesis without selection, a 1:10 inoculum was incubated with 0.5 pg/ml aztreonam (IC99 =0.2 pg/ml). Viable cells were expanded by growing a 1:10 dilution at the same concentration of drug, after which another 1:10 dilution was selected in 32 pg/ml aztreonam. At this concentration, none of the controls showed significant growth. Surviving cells were plated at 64 pg/ml aztreonam.
Plasmids were obtained from single colonies and the TEM-1 (3-lactamase was directly sequenced using specific primers.
To eliminate the possible effects of mutations elsewhere in the vector and to generate single, double, and triple amino acid substitutions, subcloned the relevant mutations were subcloned into the wild type target plasmid. For phenotypic analysis, the mutants were retransformed into JS200 cells containing the wild type Pol I
(avoiding further mutagenesis). Dose-response curves of selected mutants for aztreonam are shown in FIG. 10 and ICsos (average of two individual transformants) for all mutants are presented in Table 4. Fold aztreonam resistancea represents ICso based on an average of two experiments, normalized against wild type =100.

Table 4 (3-lactamase Fold aztreonam resistancea Deletion 0.57 Wild type 1.00 E 104K 2.70 R164H 2.18 R164S 2.53 G267R 0.84 E104K R164H 43.7 E104K R164S 74.7 E104K G267R 2.87 R164H G267R 2.70 E104K R164H G267R 67.8 E104K R164S G267R 160.0 5 Aztreonam was purchased from the Drug Services of the University of Washington and resuspended in water to a concentration of 50 mg/ml. The aztreonam dose-response curve for JS200 was obtained under the conditions used for Pol I
mutagenesis (1:105 inoculum, pre-warmed 2XYT, 37 °C shaker for 15h).
The inhibitory concentration of Aztreonam that kills >99% of the cells (IC99) was 10 0.2~g/ml. This was true for both JS200 cells transformed with pGPS3ori and cells transformed with pGPS30, confirming that wild-type [3-lactamase does not confer significant aztreonam resistance.
JS200 cells were transformed with a plasmid enconding D424A I709N A759R
Pol I and pGPSori, a plasmid encoding wild type (3-lactamase starting 1 SObp 15 downstream from the ori. Controls included cells expressing the following:
(a) Wild type Pol I and wild type (3-lactamase; (b) D424A I709N A759R Pol I and a target plasmid carrying a large (3-lactamase deletion; and (c) wild type Pol I and the deleted (3-lactamase. Two independent selections were carried out under "optimized"
mutagenic conditions. After one round of mutagenesis without selection, a 1:10 20 dilution of the culture was incubated with 0.5 pg/ml aztreonam (IC99 =0.2 pg/ml).
Viable cells were expanded by growing a 1:10 dilution at the same concentration of drug, after which another 1:10 dilution was selected in 32 pg/ml aztreonam.
None of the controls survived at this point, indicating that the presence of both error-prone Pol I and (3-lactamase was essential for resistance under these conditions.
Surviving cells were plated at 64 pg/ml aztreonam. Plasmids were obtained from single colonies and the TEM (3-lactamase was directly sequenced using specific primers.
Individual colonies were picked from aztreonam plates and grown under permissive conditions (30°C, no shaking) to avoid further mutagenesis in the presence of 10 pg/ml aztreonam. Plasmids were extracted from 3 ml of each of these clones using the Perfectprep~ miniprep kit from Eppendorf AG and resuspended in 50 pl water. They were sequenced by using the primers Blac-5,-6,-7,-8 (Example 4).
The sites HaelII, Scal and Fspl (unique within the target plasmid) were used to subclone the mutations identified in the (3-lactamase ORF by aztreonam selection into the vector encoding (3-lactamase and to generate all possible mutant combinations. These constructs were transformed into JS200 cells carrying the wild-type Pol I plasmid to maintain the same genetic background without inducing further mutagenesis, and grown in the absence of any (3-lactam. Cells were diluted to the standard inoculum of 105 cfu/ml and grown for 16h in the presence of increasing concentrations of aztreonam. The ICSOS were established by plotting OD6oo against drug concentration and finding the concentration that reduces survival by 50%.
To account for variations in expression in individual clones, two independent experiments were performed and averaged. The standard deviation between the two experiments was in the order of 30%.

Claims (45)

We claim:

1. A recombinant mutant DNA polymerase within the Pol I family of polymerases that includes a mutation within an active site having a first motif A and a second motif B, identified as producing a higher rate of misincorporation of nucleotides than the misincorporation rate of a naturally occurring DNA
polymerase.

2. The mutant DNA polymerase of claim 1, wherein the naturally occurring DNA polymerase is a DNA polymerase of Escherichia genus.

3. The mutant DNA polymerase of claim 1, wherein the first motif A is mutated at one or more residues, but the second motif B is not mutated.

4. The mutant DNA polymerase of claim 1, wherein the second motif B is mutated at one or more residues, but the first motif A is not mutated.

5. The mutant DNA polymerase of claim 1, wherein the first motif A and the second motif B are mutated, each at one or more residues.

6. The mutant E. coli DNA polymerase of claim 1, wherein the mutation within the first motif A does not occur at position 709.

7. The mutant E. coli DNA polymerase of claim 1, wherein the mutation within the first motif A comprises an amino-acid residue, substituted at position 709, which is not an isoleucine.

8. The mutant DNA polymerase of claim 7, wherein the amino-acid residue is phenylalanine.

9. The mutant DNA polymerase of claim 7, wherein the amino-acid residue is methionine.

10. The mutant DNA polymerase of claim 7, wherein the amino-acid residue is alanine.

11. The mutant DNA polymerase of claim 7, wherein the amino-acid residue is asparagme.

12. The mutant E. coli DNA polymerase of claim 1, wherein the mutation within the second motif B does not occur at position 756 or 759.

13. The mutant E. coli DNA polymerase of claim 1, wherein the mutation within second motif B comprises an amino-acid residue, substituted at position 756, which is not serine.

14. The mutant DNA polymerise of claim 13, wherein the amino-acid residue is glutamic acid.

15. The mutant E. coli DNA polymerise of claim 1, wherein the mutation within second motif B comprises an amino-acid residue substituted at position 759, which is not an alanine.

16. The mutant DNA polymerise of claim 15, wherein the amino-acid residue is arginine.

17. The mutant DNA polymerise of claim 15, wherein the amino-acid residue is asparagine.

18. The mutant DNA polymerise of claim 15, wherein the amino-acid residue is proline.

19. The mutant DNA polymerise of claim 15, wherein the amino-acid residue is serine.

20. The mutant DNA polymerise of claim 1, wherein the rate of misincorporation is at least 10-fold higher.

21. The mutant DNA polymerise of claim 1, wherein the rate of misincorporation is at least 100-fold higher.

22. The mutant DNA polymerise of claim 1, wherein the rate of misincorporation is at least 1000-fold higher.

23. The mutant DNA polymerise of claim 1, wherein an exonuclease domain within the mutant polymerise has been inactivated.

24. The mutant DNA polymerise of claim 1, which functions in vivo.

25. The mutant DNA polymerise of claim 1, which functions in vitro.

26. A vector encoding the mutant DNA polymerise of claim 1.

27. A host cell comprising the vector of claim 26.

28. A method for generating random mutations within a target DNA sequence by in vivo mutagenesis comprising:
providing a suitable host cell with a first DNA vector expressing a mutant DNA polymerise within the Pol I family of polymerises that includes a mutation within an active site comprising a first motif A and a second motif B, identified as producing a higher rate of misincorporation of nucleotides than the misincorporation rate of a naturally occurring DNA polymerise;

providing the host cell with a second DNA vector encoding a target sequence located downstream of an origin of replication, the second DNA vector subjected to mutagenesis by the mutant DNA polymerase expressed from the first DNA vector, to produce a mutated form of target DNA; and growing the transformed cells under a set of conditions that substantially promotes a rate of plasmid-specific mutagenesis greater than the rate of chromosomal mutagenesis.

29. The method of claim 28 for selecting a mutated target DNA that confers a predetermined biological function, comprising:
contacting the transformed cells with a selective media in which the surviving transformed cells are those that exhibit the predetermined biological function; and isolating the mutated target DNA from surviving transformed cells.

30. A method for generating random mutations within a target DNA
sequence by in vivo mutagenesis comprising:
producing a suitable host cell expressing a mutant DNA polymerase within the Pol I family of polymerases that includes a mutation within an active site comprising a first motif A and a second motif B, identified as producing a higher rate of misincorporation of nucleotides than the misincorporation rate of a naturally occurring DNA polymerase;
providing the host cell with a DNA vector encoding a target sequence located downstream of an origin of replication, the DNA vector subjected to mutagenesis by the expressed mutant DNA polymerase, to produce a mutated form of target DNA;
and growing the transformed cells under a set of conditions that substantially promotes a rate of plasmid-specific mutagenesis greater than the rate of chromosomal mutagenesis.

31. The method of claim 30 for selecting a mutated target DNA that confers a predetermined biological function, comprising:
contacting the transformed cells with a selective media in which the surviving transformed cells are those that exhibit the predetermined biological function; and isolating the mutated target DNA from surviving transformed cells.

32. The method of claim 28 or 30, wherein the set of conditions that are controlled includes:
growing the culture of transformed cells in a rich media; and allowing the culture to reach and to maintain in a stationary growth phase.

33. The method of claim 28 or 30, wherein obtaining the set of conditions that are controlled includes:
starting a culture of transformed cells with a low inoculum of starter cells;
growing the culture of transformed cells in a rich media;
incubating the culture at a restrictive temperature before shaking; and allowing the culture to reach and to maintain in a stationary growth phase.

34. The method of claim 28 or 30, wherein the target DNA represents prokaryotic or eukaryotic sequences.

35. The method of claim 28 or 30, wherein the suitable host cell represents a prokaryotic strain.

36. The method of claim 35, wherein the prokaryotic strain expresses a temperature-sensitive DNA polymerase allele.

37. The method of claim 35, wherein the prokaryotic strain expresses a wild-type DNA polymerase endogenously.

38. The method of claim 35, wherein the prokaryotic strain expresses a wild-type DNA polymerase and exhibits a wild-type mismatch repair function endogenously.

39. The method of claim 35, wherein the prokaryotic strain carries a genetic defect that is amenable to complementation.

40. The method of claim 32 or 33, wherein the rich media comprises a nucleoside analog, wherein the nucleotide analog is labeled with a group selected from a group consisting of fluorophores, chemiluminescers, bioluminescers, or radioisotopes.

41. A kit for producing mutations within a DNA molecule comprising:
a container; and a vector encoding the mutant DNA polymerase of claim 1.

42. The kit of claim 41 comprising:
a second container; and a deoxyribonucleoside triphosphate or related analogs.

43. The kit of claim 42 wherein the deoxyribonucleosides are labeled or unlabeled.

44. A kit for producing mutations within a DNA molecule comprising:
a container; and a vector encoding the mutant DNA polymerase of claim 2.

45. The vector of claim 44 encoding one or more of the following mutant polymerases including:
a mutant having an asparagine at position 709 that substitutes a wild type isoleucine;
a mutant having a phenylalanine at position 709 that substitutes a wild type isoleucine;
a mutant having an alanine at position 709 that substitutes a wild type isoleucine;
a mutant having a methionine at position 709 that substitutes a wild type isoleucine;
a mutant having an arginine in position 759 that substitutes a wild type alanine;
a mutant having an asparagine in position 759 that substitutes a wild type alanine;
a mutant having a proline in position 759 that substitutes a wild type alanine;
a mutant having a serine in position 759 that substitutes a wild type alanine;
a mutant having a glutamic acid in position 756 that substitutes a wild type serine;
a mutant having an asparagine at position 709 that substitutes a wild type isoleucine, and an arginine in position 759 that substitutes a wild type alanine; and a mutant having a phenylalanine at position 709 that substitutes a wild Type isoleucine, and an arginine in position 759 that substitutes a wild type alanine.