|Home | About | Journals | Submit | Contact Us | Français|
We have developed an easy, reliable two-step method for the insertion of large DNA fragments into any desired location in the E. coli chromosome. The method is based on the recombineering of a small (~1.3 kbp) “Landing Pad” into the chromosome at the insertion site, to which the large construct is subsequently delivered via I-SceI endonuclease excision from a donor plasmid. To demonstrate the power of this method, we here show the insertion of a fragment containing the entire lac operon (~9 kbp) into four predefined novel locations in the E. coli chromosome, a feat not possible with existing technologies. In addition, the chromosomal breaks induced by landing pad excision provide sufficient selective pressure that positive selection by antibiotics is unnecessary, making precise, exact insertion without extraneous sequence possible.
One of the basic methodologies of modern biology is the introduction of exogenous genes and gene networks into cells to measure their properties, for example the qualitative effects of gene expression on phenotype or quantitative characteristics of regulation of a gene network. One of the standard methods for delivery and expression of exogenous genes in bacteria is molecular cloning of the genes of interest into a multi-copy plasmid or bacterial artificial chromosome (BAC), followed by transformation of this construct into the cell. After transformation, the plasmid is maintained by inclusion of a selection marker, usually genes encoding proteins conferring antibiotic resistance.
In many cases gene delivery via plasmid is sufficient for the purpose at hand. However, for some applications expression from multi-copy plasmids is not ideal, and it would be far more desirable to incorporate the construct directly into the host cell's chromosome. Currently, there are two primary technologies for chromosomal insertion in E. coli: recombineering1–5 and phage-derived methods.6,7 However, both of these methods have significant drawbacks.
Recombineering is very easy to perform and can target constructs to any desired site in the chromosome. However, the efficiency of insertion at the targeted location drops precipitously as the size of the construct increases.8 Several labs report routinely performing recombineering of fragments up to 3–4 kbp.9–12 We have found recombineering of constructs this size or larger to be prohibitively difficult due to nonspecific recombination of fragments into sites other than the target,8 making identification and isolation of correct integrants a time consuming and tedious task. Nonspecific integration can be mitigated through the use of counterselection methods using e.g., galK2 or tetA,13,14 however the low efficiency of integration still makes this a non-trivial endeavor and success is by no means guaranteed.
The remaining possibility for insertion is methods which apply phage-derived mechanisms.6 These methods have the benefit of being able to reliably deliver much larger DNA fragments into the chromosome than recombineering. The downside is reduced flexibility. It is only possible to incorporate DNA into endogenous phage attachment sites already present in the chromosome. Furthermore, targeting different attachment sites requires the tedious construction of entirely new constructs, one for each insertion location specific to that particular phage attachment site.
To circumvent these limitations, we have developed a new standardized method for the integration of large constructs into the E. coli chromosome.8 Our system combines the advantages of recombineering and phage-based methods: the ability to insert very large constructs, and the ability to target the insertion specifically to any location in the chromosome. The system is very easy to use and effective, while also being extremely flexible.
The system uses three plasmids, shown in Figure 1A, and the strategy (modified to illustrate exact integration—see below) is shown in Figure 2B. The helper plasmid, pTKRED, carries a gene encoding the homing endonuclease I-SceI, as well as genes encoding the recombinogenic λ-Red proteins3 and RecA. The Landing Pad plasmid pTKS/CS carries a tetracycline resistance gene, tetA, flanked by I-SceI restriction sites and novel 25 bp sequences we refer to as “Landing Pad Regions”, or LPs. These LPs were designed to be unique to the E. coli genome, and ultimately serve as the recombination target for insertion of the large construct. This construct, encompassing the LPs, I-SceI recognitions sites, and tetA gene, is amplified by PCR and recombineered using standard methods5 into the desired insertion site.
Next, a donor plasmid, pTKIP, is transformed into the cell. This plasmid carries the insertion fragment, which is flanked on either side by LPs and I-SceI restriction sites. Upon induction of I-SceI and λ-Red expression from the helper plasmid, both the insertion fragment and the Landing Pad are excised from their host molecules, and λ-Red mediates homologous recombination between the fragment and chromosomal LP regions. The fragment is therefore integrated into the chromosome, repairing the chromosomal breaks induced by excision of the landing pad. The helper plasmid pTKRED, which contains a temperature sensitive pSC101 replication origin, can then be cured from the cell by growth at nonpermissive temperature. The end result is a bacterial strain in which the construct has been successfully inserted into the desired location in the chromosome.
As a demonstration, we have used this method to integrate the entire lac operon, including the genes lacI, lacZ, lacY and lacA, as well as their promoters and terminators, into the four novel locations in the E. coli chromosome indicated in Figure 1B. The insertion was performed once in a given strain, generating a set of bacterial strains with the lac operon integrated into one of the four different chromosomal positions. An agarose gel of PCR products verifying insertion into the indicated sites is shown in Figure 1C. To simplify selection of successful integrants, the fragment also contained the kanamycin resistance gene neo (see Fig. 1D), yielding a fragment size of ~9 kbp. As far as we are aware, site specific insertion of such large constructs is impossible with any other existing technology.
One of the main strengths of our method is the ability to deliver constructs into any desired locus in the E. coli chromosome. However, in many instances, a researcher attempting to integrate a large construct into the chromosome may not be concerned about the precise location of the insertion, as long as the construct is stably integrated and reliably expressed. In cases such as this, integration via phage based mechanisms has been the method of choice, and it appears at first glance that our method would provide no additional benefit in such cases. However, our method has advantages that we believe may prove useful.
Both recombineering and phage-based methods require the inclusion of genes, generally encoding proteins conferring antibiotic resistance, along with the insertion fragment that allow for positive selection. Once the construct is integrated into the genome, the included selection marker gene is expressed and allows for the preferential growth of successful integrants. In contrast, with our method we have shown that the repair of induced chromosomal breaks at the Landing Pad site via integration of the construct provides sufficient selective pressure that additional selection for integrants is unnecessary. Consequently, our method allows for the incorporation of DNA sequence into the chromosome without the need for any extraneous sequence encoding the positive selection marker; the only additional sequences required for integration and repair (and therefore selection) are the Landing Pad region sequences.
Our method therefore allows for the exact integration of only the desired sequence into the chromosome. This ability allows for possibilities such as the integration of a gene within a polycistronic operon to add to or replace existing cistrons without additional disruption to the reading frame of downstream genes. An example would be the exact replacement of the lacY gene in the lac operon without introducing extraneous sequence or disrupting the reading frame of either lacZ or lacA.
For ease of use and rapid deployment, the plasmids that comprise the system have been engineered to include standardized LP sequences which are not already present in the E. coli genome. The integration of these standardized LP regions will introduce extraneous sequence along with the insertion fragment (i.e., the red squares in Fig. 1D). In many cases this can be tolerated. However, to achieve exact integration, it is necessary to modify the landing pad and donor plasmid to use the existing sequence of the desired insertion site to serve as the LP regions. The necessary modifications, shown in Figure 2A, are minor and extremely quick and easy to perform.
Modification of the Landing Pad is indirect; the plasmid itself remains unaltered. All that is required is judicious design of the primers used to amplify the landing pad for insertion via recombineering. The 50 bp homology arms [designated Flanking Region 1 (FR1) and Flanking Region 2 (FR2) in Fig. 2A] that upon PCR amplification are used to target the Landing Pad to the insertion location will serve as the LP regions. The only direct modification necessary is to the LP regions on the donor plasmid which will carry the insertion fragment. This modification can be made quickly by amplifying the insertion fragment with primers containing the new FR regions and I-SceI restriction sites, digesting the resulting product and donor plasmid with I-SceI, and ligating the fragment back into the donor plasmid backbone. After these modifications are made, the insertion protocol proceeds as described above, resulting in the exact, markerless integration of the construct into the desired site.
We additionally note that the lack of a requirement for positive selection using antibiotics makes this method particularly attractive for use in human-pathogenic bacteria, where limited use of antibiotics is advantageous. However, due to the relatively low frequency of Landing Pad integration via recombineering, the use of antibiotics is still required for the selection of successful landing pad integrants. We are currently attempting to adapt the method to use non-antibiotic positive selection mechanisms to circumvent this requirement to make the method even more attractive.
The authors would like to acknowledge support by the National Institutes of Health [GM078591, GM071508]; and the Howard Hughes Medical Institute .
Previously published online: www.landesbioscience.com/journals/biobugs/article/12386