Abstract
Breakthrough in genome sequencing projects has revealed large quantity of potentially new genes. Ever since microbial genome sequences became accessible to most research laboratories, reverse genetic analysis has become a standard experimental approach to study bacterial gene function. Point mutations can be introduced into bacterial chromosomes which is important for progress in further studies in functional genomics. Chromosomal systems are more favorable than artificial plasmid based systems for many investigations. Analysis of these genes for their functions needs gene manipulation methods which are simple and efficient allowing targeted modification of particular sequences in their chromosomal position. Gene replacement is a technique of modifying endogenous gene at its original locus in the chromosome. The strategies which have been developed for replacing segments of the chromosomal genes are based on recombination between the target locus and a cloned DNA fragment containing the desired modification. Gene replacement in E. coli chromosome can be done by variety of techniques. In this work we present different methods by which mutations can be introduced into E. coli chromosome by gene replacement.
Gene replacement is an effective technique that takes advantage of the natural ability of an organism to recombine homologous regions of DNA. It enables the specific replacement of targeted genome sequences with copies of those carrying defined mutations. Therefore gene replacement can promote the assignment of function to cloned genes. Obviously there are many potential uses of this method in research on microorganisms. Exploitation of large quantity of potentially new genes from recent microorganism genome sequence projects will, in many cases, require the ability to verify in an efficient manner the gene assignments based on sequence homologies and has opened many new experimental avenues.
The ability to make precise genetic modifications to the bacterial chromosome and then to study the resulting phenotypic behavior is very important for functional studies. Construction of mutant alleles can be done in vitro by recombinant DNA techniques available. To study the effect of mutation by introducing the mutant allele into the cell where it will replace the wild-type gene by homologous recombination then the effect of mutation can be tested by expressing the gene at its native location. Gene replacement strategies basically involve three stages. First, a cloned region of the genome is disrupted; for example, by the insertion of an antibiotic-resistance marker. Second, the mutated sequence carried on a suitable vector is introduced into the host organism by transformation. Finally, transformants are screened to identify those in which the desired gene replacement events have occurred. Recombination events between the homologous sequences present on the chromosome and the DNA vector results in the insertion of the transforming DNA into the host chromosome. A double crossover, that is two homologous recombination events, one on either side of the mutation in the cloned sequence, results in replacement of the wild-type gene with the mutated copy. Recombination at only one end of the wild-type sequence results in a single crossover and the insertion of the transforming DNA into the host chromosome.
Techniques exploiting homologous recombination to achieve specific gene replacement have been developed for use in many micro-organisms. Gene replacement on the E. coli chromosome can be done by a variety of techniques. One of the ways is called linear DNA-transformation. An appropriate linear DNA, containing a mutated or deleted gene flanked by homologous regions of the chromosome, is transferred into recombination-proficient strains, such as recBC, sbcBC, or recD. Another general method involves cointegration of a plasmid containing the mutated gene of interest into the E. coli chromosome by single-crossover recombination, followed by resolution of the cointegrate. The key to this procedure is the use of appropriate vectors that cannot replicate under the conditions used for selection of the cointegrate. This review will describe different approaches for obtaining gene replacement in E. coli.
2. Recombination as a genetic tool
Genetic recombination is a process by which a molecule of nucleic acid is broken and then joined to a different DNA molecule. Recombination in E. coli and other bacteria is mediated by RecA which its activity is ATP depended and requires a minimal length of DNA homology between the donor sequence and a double strand recipient DNA. Homologous recombination (HR) requires sequence similarity between the incoming donor DNA and DNA sequence of the recipient. In E. coli a minimum of approximately 20 base pairs is acquired for successful recombination with circular DNA (Watt et al. 1985).
In most cases, DNA is integrated into the recipient genome through HR. HR is an invaluable aid to the molecular geneticist in both prokaryote and eukaryote systems, providing the means by which DNA can be accurately excised, replaced or inserted into the chromosome. Targeted disruption of genes is commonly achieved using HR in a variety of bacteria (Winans et al. 1985; Miller and Mekalanos, 1988; Stibitz et al. 1989). In principle, a simple approach utilizes a plasmid vector carrying a single homologous region of DNA which is complementary to part of a specific gene and entirely internal of the open reading frame. Introduction of this into the cell and resultant Campbell-type recombination (Campbell 1962) would integrate the vector into the open reading frame of the target gene on the chromosome thus creating two truncated genes. A more reliable approach involves using a vector with two regions of homologous DNA flanking an antibiotic marker. Recombination between both homologous vector sequences and the chromosome would result in excision of any chromosomal sequence between the homologous regions and integration of the antibiotic marker: allelic exchange. Mutant phenotypes can be confirmed by reintroduction of the wild type gene on a plasmid. A chromosomal fragment is cloned into a plasmid not replicating in laboratory strain, and the construction is transferred into laboratory strain. The homologous fragments in the plasmid and in the chromosome recombine, and results in merging of the plasmid into the chromosome (Fig. 1), and the integrants can be found with selection. Depending on the arrangement of the homologous fragments, Single crossover (SCO) or double crossover (DCO) resulting in insertions or deletions in the chromosome will occur (Biswas et al. 1993; Mills 2001).
Illegitimate recombination (IR) (nonhomologous) is a type of recombination that occurs between DNA molecules sharing no homology or with very short regions of homology, typically 4 to 10 nucleotides. It occurs widely in both prokaryotes and eukaryotes and is responsible for major genome rearrangements including deletions, duplications, translocations and insertions. The mechanisms thought to be involved in IR are varied and in E. coli the mechanisms of short-homology-dependent and short-homology independent IR is mediated by subunit exchange of DNA gyrase whereas short-homology-dependent IR involves the formation of double-strand DNA breaks and then processing, annealing and ligation of DNA ends (Shimizu et al. 1997). Heterologous recombination events requiring between 150-200bp DNA similarity, have been named homology-facilitated illegitimate recombination (De Vries and Wackernagel 2002). IR events are associated with very small sequences ("anchors") that join DNA molecules at sites with no or a few identical nucleotide sequences. These short sequences can be as small as 3 or 8 base pair and do not require RecA protein to be initiated. Such events are rare and happen in low frequency among bacteria (De Vries and Wackernagel 2002).
Figure 1: Gene replacement strategy
3. Gene replacement methods
Recent technical advances have produced several methods to introduce a mutant sequence synthesized in vitro into the E. coli chromosome. The two main previously established methods are the 'in-out' method and the linear fragment method.
3.1. 'In-out' method
In the 'in-out' method, (Hamilton et al. 1989; Martinez-Morales et al. 1999; Posfai et al. 1999) a mutant sequence is introduced into the cell on a multicopy circular plasmid. A Rec mediated single homologous crossover results in cointegration of the whole circle into the genome at the target site with the plasmid vector between a wild type and mutant copy of the target sequence ('in'). The cointegrate is resolved by a second single crossover ('out'). When in and out crossovers span the mutant site, the desired mutation is transferred to the genome. The mutant allele can be delivered on a suicide plasmid into the cell. For insertion of the circular molecule into the chromosome a single crossover between the mutant and the wild type allele is required. This cointegrate generated can be resolved by spontaneous recombination of the allele pair resulting in cells with either a wild type or a mutant allele in the chromosome (Blomfield et al. 1991; Link et al. 1997). Because the rare resolution event, to eliminate cells which retain the cointegrate structure and carry a counterselectable gene located on the inserted plasmid effective counterselection is needed (Dean 1981; Gay et al. 1985; Russell and Dahlquist 1989). The most commonly used method is the sacB/sucrose counterselection system (Gay et al. 1985), but the use of the method is limited by its strain-medium and temperature-dependence (Blomfield et al. 1991; Link et al. 1997). The key to this procedure is the use of appropriate vectors that are not capable to replicate under the situation used for selection of the cointegrate. The best used vectors include ColE1-drived plasmids that do not replicate in polA mutants, (Gutterson and Koshland 1983; Saarilhti and Pavla 1985) a temperature-sensitive pSC101 replication (Hamilton et al. 1989; Kato et al. 1989; Link et al. 1997) and a phage-based vector (Slater and Maurer 1993). Although such a cointegration scheme has been successfully and widely used, a problem is that resolution of the cointegrate occurs at a relatively low frequency and may not always give the replacement desired. However, if there is a positive-selection device for monitoring resolution of the excised vector, this problem would be overcome. Recently, Link et al. (1997) reported such an integration vector that carries a positive-selection marker, the sacB gene encoding levansucrase.
3.1.1. Gene replacement with one-sided homologous recombination
Vectors for replacing endogenous chromosomal sequences with transferred DNA are traditionally constructed so that the replacement segment is flanked on both sides by DNA sequences which are identical to sequences in the chromosomal target gene. The amount of homology between targeting vector and target is important for maximizing the frequency of homologous recombination. In mammalian cells a positive relationship has been found between length of homology and targeting frequency. (Hasty et al. 1991, Schulman et al. 1990). In mammalian cells larger amounts of homology are used to obtain gene replacements. It appears that quality of homology is also accounted during gene replacement such as small sequence differences between DNAs leading to decreased frequency of gene replacement. To test the importance of bilateral regions of homology, Berinstein et al. (1992) measured recombination between transferred and chromosomal immunoglobulin genes when the transferred segment was homologous to the chromosomal gene only on the 3' side indicating that targeting with fragments bearing one-sided homology can be as efficient as with fragments with bilateral homology, provided that the overall length of homology is comparable. The frequency of gene targeting increases monotonically with the length of the region of homology. Vectors designed for one-sided homologous recombination might be advantageous for some applications in genetic engineering. Berinstein et al. (1992) used vectors with a conventional (bilateral homology regions) and a nonconventional (unilateral homology region) design to modify the chromosomal, µ heavy-chain gene and found that specific gene replacement occurs when the transferred DNA is homologous on only one side suggesting simplified methods for modifying chromosomal genes. For example, it might be possible to use this approach to generate a set of nested deletions to test the importance of adjoining DNA in gene expression. The experiments involving a unilateral homology region had several intriguing aspects. Berinstein et al. (1992) expected that the absence of a 5' homology region would substantially reduce the frequency of recombination in the µ gene. Hence, they consider it significant that the frequency of recombination for vectors bearing one-sided homology was as high as for vectors bearing bilateral homology. Their results contrast with those of Smith and Kalogerakis (1990), who found that homologous recombination was significantly reduced when only one-sided homology was provided. The explanation for this may be related to the short (110bp) region available for nonhomologous crossovers in their system. Thus, under circumstances in which there is a long region in which nonhomologous crossovers can occur, the frequency of recombination is determined by the total length of the homology region rather than the distribution of homology.
3.1.2. DSB (double-stranded break) - stimulated gene replacement
DSB stimulated gene replacement offers an efficient and simple way to manipulate chromosomal sequences. This method is used in recombination-proficient wild type E. coli as it produces markerless replacements at high efficiency and does not require specific growth conditions, and is potentially applicable in other microorganisms as well. This method is based on the recombination and repair activities of the cell and permits the targeted construction of markerless insertions, point mutations, as well as large deletions can be created in the E. coli genome. In this method the integration of the mutant gene which is carried on a circular plasmid (suicide plasmid) is inserted at a homologous locus into the genome by homologous recombination between the mutant and the wild-type (wt) alleles resulting in a direct duplication (Pósfai et al. 1999). The cointegrate formed via intramolecular recombination and the resolution of the allele pair results in either a mutant or a wild type chromosome which can be distinguished by allele-specific PCR screening. The cointegrate is resolved by introducing a unique DSB by the meganuclease I-SceI into the chromosome. The enzyme recognizes an 18-bp sequence and generates a DSB with a 4-base 3' hydroxyl overhang (Montelheit et al. 1990). Cleavage by the nuclease not only enhances the frequency of resolution by two to three orders of magnitude, but also selects for the resolved products. Use of the method was demonstrated by Pósfai et al. 1999 where they constructed a 17bp and a 62kb deletion in the MG1655 chromosome. Cleavage of the chromosome induced the SOS response but did not lead to an increased mutation rate.
Figure 2: General procedure of the DSB-stimulated gene replacement.
For inserting mutant alleles into the chromosome two sets of suicide plasmids, equipped with an I-SceI recognition site, are available. (Pósfai et al. 1997). Insertion creates a duplication of the target sequence. Cells which resolve this cointegrate can be selected by various counterselection methods. The novelty of the method presented by Pósfai et al. (1999) is that I-SceI cleavage serves not only as a selection tool, but also as a stimulator of the resolution process, increasing the efficiency of resolution by two to three orders of magnitude. This method has been used in the laboratory in various E. coli strains, including pathogens, to introduce mutations into the genome at six different loci. It can be assumed that modified versions of this system can be applied to various microorganisms since DSBs are known to stimulate homologous recombination in a wide range of systems.
3.1.3. Gene Replacement by Xer Recombination
Gene Replacement by Xer Recombination is an efficient method for gene replacement for unlabelled and stable insertion of gene into bacterial chromosomes and for selectable marker gene excision by Xer recombination. This technique make use of native Xer recombinases that normally function to restore the plasmid as well as chromosomal dimers produced by RecA back to monomers. Xer recombinases are present in bacteria naturally to excise gene for antibiotic resistance followed by chromosomal integration, by that means eliminate the requirement for an exogenous SSR system. Xer recombinases present in E. coli are XerC and XerD (Leslie and Sherratt 1995), and have homologues in Bacillus subtilis such as RipX and CodV (Sciochetti et al. 2001) and in many other species and have been used for gene insertions and deletions successfully by Xer-cise technology. These Xer recombinases are present in the majority of bacterial species.
Xer-cise technology can be applied in gene deletion which is helpful in accurate excision of target alleles preventing reversion in mutant strains and in gene insertion by inserting new genes for protein expression or for altering the phenotype. The labelling of integration cassette with gene for antibiotic resistance for selection of mutants has been traditionally accomplished in bacterial chromosomes for deletion and insertion of genes. The benefits of this method are; chromosomal insertion was followed by excision of the antibiotic resistance gene by native Xer recombinases, no requirement of exogenous recombinases such as Cre, Flp, Xer-cise technology works in a wide species range of bacteria and it makes able multiple gene integration events in the identical strain. This approach has some disadvantages such as: expression of surrounding genes altered by insertion of foreign gene permanently; the presence of antibiotic resistance gene on the chromosome can affect the use of the same gene for plasmid maintenance and the difficulty of multiple gene integration events due to the deficiency of suitable genes for antibiotic resistance. To address this, genes for antibiotic resistance have been flanked with sites for site-specific recombinase enzymes supplied in trans on a plasmid. Additional transformation is required as an extra step followed by further culturing to remove the helper plasmid. Additionally, this technique has only been optimized for a small range of bacteria. This technology should be applicable to all prokaryotes with the ubiquitous Xer dimer resolution system (Recchia and Sherratt 1999).
Bloor and Cranenburgh 2006 proposed the term "Xer-cise" to describe technique. They constructed cassette consisting of gene for antibiotic resistance flanked by dif sites and regions of homology to the chromosomal target. This gene was integrated into the chromosome after amplified or cloned into a plasmid. Cells that have undergone intramolecular Xer recombination at dif sites during further culture were identified by antibiotic sensitivity and verified by PCR. The inclusion of a counter-selectable gene is not necessary as the Xer recombination frequency is high enough for recombinant clone detection without the antibiotic selection. This is beneficial, since certain counter-selectable genes can be mildly toxic such as sacB even when counter-selection is not present, leading to an accumulation of cells carrying mutations in sacB and for that reason generating false-positive results during clone selection (Bloor and Cranenburgh 2006). Xer-cise technique will simplify and enhance the production of unlabeled E. coli mutants and other bacteria for which native dif site has been elucidated (Chalker et al. 2000).
3.1.4. Gene replacement without selection: 'gene gorging' method
The term 'gene gorging' comes from forcibly incorporating the desired allele into the genome by imposing copious quantities of it into the cell. This method greatly simplifies the process of mutagenesis by combining the efficient steps and eliminating the multiple selections needed for the earlier methods. As in the 'in-out' method, the mutant allele is introduced on a stable plasmid, and as in the linear fragment method, Red recombination is used to integrate the cloned allele into the genome. The key difference in gene gorging is the linearization of the plasmid in vivo. Since virtually every cell receives linear donor, and Red is extremely efficient, the yield of the mutant allele incorporated into the genome is at least 1%. Thus the desired alteration can be identified by individual examination of around one hundred colonies. Gene Gorging enables investigaters to introduce precise sequence changes into genome of E. coli in a direct way without leaving undesirable drug markers or other 'scars'. This technique can be used to make variety of mutations, like deletions of changes to the amino acid sequence of a gene. The frequency at which the mutations are introduced is extremely high; no selection is required but by just simple screening.
In a similar vein, Poteete and Fenton (2000) used lambda phage as the vehicle to efficiently introduce recombinogenic fragments into E. coli for Red recombination. The Pae R7 class II restriction modification system based on a six base cut site was established in the recipient E. coli to cut the incoming unmethylated donor DNA creating linear substrate fragments for Red recombination. They reported conversion of 1.5% of cells to the introduced chloramphenicol resistant phenotype in wild type and 5.6% in a recG background. Ellis et al. (2001) have reported the introduction of unselected mutations in up to 7% of viable cells using electroporated single stranded DNA. Previous methods of directed mutagenesis in E. coli rely on the use of positive and negative selections for recombination intermediates because the desired events occur at very low frequency. Herring et al. (2003) presented gene gorging, a new method in which the efficiency of gene replacement is high enough to make selection of recombinants unnecessary. Replacement occurred in 1-15% of the cell population, making it feasible to identify mutants by PCR of individual colonies or other means of direct screening. The approach by Herring et al. (2003) resulted in generally higher levels of replacement for wild type cells and is greatly simplified by using plasmid rather than lambda techniques and meganuclease I-SceI, which does not cut in the genome of E. coli or most other organisms. In their study, they changed alanine codons to amber stop codons, but virtually any mutation can be introduced in a straightforward way without complicated manipulations.
Figure 3: Gene gorging strategy. The desired genomic modification is produced as a PCR fragment with the recognition sequence for I-SceI on one or both of the primers, then cloned into a standard cloning vector (donor plasmid). After electroporation of both plasmids into E. coli the Red and I-SceI genes are induced, leading to linearization of the donor plasmid and a double recombination with the chromosomal target.
3.2. Linear fragment method
The linear fragment method (Murphy 1998; Datsenko and Wanner 2000; Murphy et al. 2000; Yu et al. 2000), utilizes the Red recombination system encoded by bacteriophage lambda genes gam, bet and exo which operates on linear DNA. Electroporation is used to introduce a linear DNA fragment carrying the synthesized mutation directly into the cell where Red favors double crossover events in the progeny since a single crossover would result in a chromosome break. Incorporation of the mutation into the chromosome occurs where the double crossover spans the mutant site.
Linear fragment method is an alternative procedure to deliver the mutant allele on a linear DNA-fragment into the cell (Jasin and Schimmel 1984; Dabert and Smith 1997; Zhang et al. 1998). An appropriate linear DNA, containing a deleted or mutated gene flanked by homologous regions of the chromosome, is transferred into recombination-proficient strains, such as recBC, sbcBC, or recD. Double cross-over recombination between the E. coli chromosome and both ends of the linear DNA fragment results in gene replacement in these particular genetic backgrounds at a high frequency. A disadvantage of this method is that it is restricted to specific nuclease-deficient recombination-proficient genetic backgrounds. Variations of Linear fragment method depend on requirement of extensive DNA engineering using long PCR primers, introduction of a marker along with the mutant gene into the genome and specifically altered host cells. This introduction of a marker along with the mutant gene can have polar effects or can prevent multiple manipulations of the genome and can be eliminated only in a second round of allele replacement (Zhang et al. 1998). But this step requires the use of a counterselection system with its intrinsic limitations.
Both the 'in-out' and linear fragment methods include inefficient steps and require strong genetic selections to achieve useful frequencies. The 'in-out' method uses a positive selection such as a drug resistance marker to select for the 'in' step and a negatively selectable marker such as sucrose resistance to drive the 'out' step. The linear fragment method is limited by the very low efficiency of electroporation. Even with high DNA concentrations and cells of the highest competency, it is impossible to introduce donor DNA into more than a tiny fraction of target cells. To overcome this limitation, a two step process may be used to first select for the insertion of a cassette with both negative and positive selectable markers at the chromosomal target site and then to perform a second (negative) selection to create the desired mutation. The second step can either employ a second electroporation with Red, site specific recombinases (in which a 'scar' is typically left behind) (Datsenko and Wanner 2000) or Rec mediated reduction of the genetic intermediate with a duplication incorporated in the original fragment (Kolisnychenko et al. 2002). The latter is especially useful in the production of deletions.
3.2.1. Gene replacement by electrotransformation
Due to exonucleolytic degradation of incoming DNA gene targeting using linear dsDNA fragments in wild-type E. coli transformation is generally inefficient. To overcome transformation inefficiency strains having high recombination-proficiency in which the exonucleolytic activity of RecBCD is inactivated have been used as transformation recipients (such as recB recC sbcA, recB recC sbcB sbcCD, and recD mutants or strains expressing bacteriophage recombination functions).
Recently, an approach was developed to achieve gene replacement in wild-type cells, in which the transformation of linear DNA containing Chi sequences (5'-GCTGGTGG-3') at both ends flanking the homologies (Dabert and Smith 1997). These sequences are known to decrease the activity of RecBCD exonuclease and stimulate its activity of recombination (Dixon and Kowalczykowski 1993; Myers and Stahl 1994; Karoui et al. 1999). If electrocompetent cells are used, gene replacements with the use of linear DNA without Chi sequences can be achieved in wild-type E. coli, on a plasmid as well as a chromosomal target. The use of electrocompetent cells and electrotransformation technique appears to lessen the exonucleolytic activity of RecBCD in E. coli, in this way allowing gene replacement to occur. This method provides a simple way to perform gene replacement in many E. coli strains. The system of gene replacement by electrotransformation requires a particular E. coli strain although extremely efficient, and in this way limiting its range of use. In contrast, the method described by Karoui et al. (1999) to achieve gene replacement can be used in many different E. coli strains and does not necessitate special DNA constructions. The frequencies of gene replacement events obtained (with a chromosomal target) are comparable to those obtained in the Chi-stimulated recombination method (Dabert and Smith 1997). In this way electrotransformation may constitute a direct method to get gene replacements with linear DNA in wild-type E. coli on plasmid and chromosomal targets. The linear DNA containing the chi sites had no effect on gene replacement efficiency although the chi sites are known to block DNA degradation and stimulate recombination in E. coli. In electroporated cells the RecBCD-mediated exonucleolytic activity was found to be diminished. Thus elctrotransformation provides a easy way to carry out gene replacements in several E. coli strains (Karoui et al. 1999). It may also be used to make gene disruptions on plasmid-carried targets which can then be transferred to the organism of interest.
3.2.2. Chi sites enhanced Gene Replacement
Chi sites are octameric nucleotide sequences in DNA that stimulate the RecBCD pathway of homologous recombination in E. coli. Stimulation is maximal at the chi site, decreases approximately factor of two for each 2-3 kb to one side, but is insignificant to the other side of chi. Chi stimulates recombination by interaction with RecBCD enzyme, which has multiple enzymatic activities and multiple physiological roles in recombination, repair, and replication. Chi appears to be active throughout the enteric bacteria; other nucleotide sequences may similarly interact with RecBCD-like enzymes in other bacteria. The nucleotide sequence of chi was shown to be 5´GCTGGTGG3´ (or its complement or the duplex) by a comparison of the sequences of active and inactive Chi sites and their flanking sequences (Smith et al. 1984).
Transformation with short (<10 kb) linear fragments (gene replacement) occurs at a very low or undetectable frequency in wild-type (rec+ sbc+) E. coli but does occur in E. coli mutants expressing the RecF (recBC sbcBC ) or ++ (recD) pathway (e.g. Jasin and schimmel 1984; Winans et al. 1985; Shevell et al. 1988; Russel et al. 1989). Inclusion of two properly oriented and positioned chi sites on the linear fragment enhances gene replacement during transformation about 40-fold in wild-type cells; dual Chi sites also enhance a specialized transduction in which intracellular EcoRI restriction enzyme cuts the linear Chi-containing fragment out of infecting λ DNA (Dabert and Smith 1997). Transformation or transduction in this way offers a useful method of "gene targeting" with cloned DNA fragments in wild-type E. coli (as opposed to recBC sbc or recD mutants used previously). These observations indicate that Chi is an important element in E. coli homologous recombination. In its absence recombination following conjugation or transduction in wild-type cells would presumably occur at very low or undetectable frequency. Testing this proposal is thwarted by the ≈1000 Chi sites in the E. coli genome.
This technique uses the characteristic of Chi sites to govern RecBCD exonuclease activity and stimulate recombination. When wild-type E. coli cells are made competent by treatment with CaCl2 Chi sites occurring near the ends of linear DNA fragments stimulate the frequency of gene replacement events. (Dabert and Smith 1997). One disadvantage of this technique is that it needs DNA constructions that add Chi sites at the fragment extremities.
3.2.3. PCR-mediated gene replacement
The hyper-recombinogenic properties of an E. coli strain in which the recBCD genes have been replaced by λ red recombination functions were exploited in the development of a general PCR-mediated gene replacement scheme for E. coli. An experimental system that allows for PCR-mediated disruption of genes in yeast has greatly aided studies of this organism (Baudin et al. 1993; Wach et al. 1994; Lorenz et al. 1995). In this scheme, PCR- generated DNA fragments containing a selectable marker flanked by ~50 bases of sequences upstream and downstream of the gene of interest are introduced into yeast cells by electroporation. Given their high rate of homologous recombination, nearly 95% of the transformed yeast cells carry the designed gene disruption. E. coli does not recombine as readily Saccharomyces cerevisiae, requiring researchers to rely on recombination-proficient mutant strains (e.g. recBCsbcBC or recD) to perform gene replacement (Marinus et al. 1983; Jasin and Schimmel 1984; Russell et al. 1989). Transformation and/or electroporation of these strains with linear DNA substrates containing a drug resistance-conferring marker near or in place of the gene of interest results in integration of the mutant or deleted gene by homologous recombination. However, successful gene replacement with these strains usually requires extensive regions of homology and/or occurs at a low frequency, limitations that have precluded the development of an efficient PCR-mediated gene disruption method for E. coli.
This method of gene replacement does not require prior cloning of the gene, and can be used to easily construct precise gene disruptions. This technique is well suited for the generation of precise deletions of unknown open reading frames predicted from the genome sequence of E. coli (and possibly those of other bacteria) in an effort to identify essential functions that might serve as suitable targets for drug design.
The gene replacement technology described by Murphy et al. (2000) allows for the easy one-step replacement of almost any gene in E. coli. The system takes advantage of PCR-promoted recombination to generate marked gene deletions and the hyper-rec environment of an E. coli strain containing the red for recBCD allele. This method of gene replacement offers an advantage over existing gene replacement techniques presently employed for E. coli: that is, the gene of interest does not have to be cloned and plasmid-chromosome co-integrants do not have to be formed and resolved. The cloning-free gene replacement scheme is suitable for cases where secondary stage products generating marked gene deletions are difficult to clone.
Figure 4: Schematic of λ Red-promoted PCR-mediated recombination. PCR products, containing a drug marker flanked by 40-60 bp of target DNA, are generated by primers designated (UP) and (UP). Chromosomal replacements were verified by PCR using primers upstream (U), downstream (D) and within the drug marker (M).
3.2.5. λ red-promoted gene replacement
It is known that bacteriophage λ recombination system Red, have the ability to act on linear DNA substrates promoting homologous recombination in the absence of RecBCD activity of the E. coli. The absence of any hotspot requirements for λ Red-mediated recombination and the break-join mechanism (Stahl et al. 1990), it seems that for the promotion of E. coli gene replacement λ Red would be best suited. Red system offers a number of advantages over other recombination systems in the bacteria. First, it is extremely efficient. Secondly, there is no noticeable increase of spontaneous mutation rates when RED proteins are transiently expressed. Thirdly, very short homology sequence is required (45bp) so a targeting vector for DNA modification can be easily obtained from PCR. Fourthly, single strand DNA oligos are better substrates for recombination with RED system. Therefore generating point mutations can be achieved with DNA oligos. Moreover when single strand oligos are used, the only known recombination protein required is the beta-protein (Lee et al. 2001). One requirement for using RED system is that one needs to have the genomic sequence of a gene, at least the region that needs to be manipulated (Liu et al. 2003).
Red/ET recombination first published under the name ET cloning (Zhang et al. 1998) and also known as recombineering, is an easy to use modification system for prokaryotic functional genomics. Zhang and coworkers demonstrated in 1998 for the first time that a pair of phage coded proteins (RecE and RecT) only need 42bp long homology arms to mediate the homologous recombination between a linear DNA molecule (e.g. a PCR product) and circular DNA (plasmid, BAC or E. coli chromosome). This method was used to disrupt the endogenous lacZ gene of E. coli strain JC9604 (Zhang et al. 1998). One year later the system was extended by the same group in replacing recE and recT by their respective functional counterparts of phage lambda redα and redβ (Muyrers et al. 1999). Since the year 2000 the system, which is protected by several international patents from Gene Bridges, has been used by other academic groups to disrupt several chromosomal genes in E. coli (e.g. Datsenko and Wanner 2000, Yu et al. 2000).
Murphy et al. 1998 developed a new system that can produce gene replacement in almost any E. coli strain at high frequency. This method does not dependent on a cointegrate, and does not require cloning of the gene in advance. This method uses the bacteriophage λ recombination functions to stimulate gene replacement. This method is extremely efficient but needs use of a particular E. coli strain limiting its range of use. The recombination functions such as exo, bet, and gam of bacteriophage λ are expressed from a multicopy plasmid and transformation carried out with linear DNA substrates to promote gene replacement in a wild-type E. coli host. Gene replacement was promoted by Lambda recombination functions within the lacZ gene at a higher rate i.e. about 15 to 130 times than recBC sbcBC or recD strains. λ red-promoted gene replacement is not only applicable widely to any E. coli but also possibly for other bacteria as well. The reason for this is since homologous recombination is raised to the higher level by the addition of functions to the host, rather than induced by alteration of host functions. The value of this method is emphasized by gene replacement with DNA fragments generated by PCR.
Conclusion
In the emerging post-genome sequencing era, high-throughput evaluation of uncharacterized ORFs becomes a necessity. Gene replacement techniques would be the ultimate tool to help investigators to systematically assign functions to the vast number of these new ORFs. With these new techniques and findings in E. coli gene replacement genomic modifications can be created with enhanced efficiency and speed and have reduced the time and work load compared to previous strategies. These approaches provide powerful tools to clone and modify genes precisely for functional analysis and thus perform biochemical or behavioral experiments to clarify functions. Application of these techniques is not problem-free but further expansion in this research area is likely to continue with the several excellent and technically demanding protocols. However, we are hopeful that in future we could see new gene replacement strategies making the more traditional gene disruption strategies obsolete.
