Understanding the regulation of plasmid replication and incompatability was an essential prerequisite for developing efficient methods of allele exchange. Three examples of commonly used methods to move mutations from a cloned gene onto the chromosome are described below.
Use of conditional replicons ("suicide plasmids") to contruct genetic duplications and null alleles. Plasmids that are conditional for their replication can be used to create defined duplications within a target genome. In such instances a specific DNA fragment is inserted into a plasmid which is then introduced into a recipient strain and placed under conditions where the plasmid cannot replicate. Since the plasmid cannot replicate, selection for some property of the plasmid, such as an antibiotic resistance marker, results in isolates that have integrated the plasmid into the host chromosome via homology between the cloned fragment and the corresponding region of the recipient chromosome. (This system can also be used to obtain gene disruptions by cloning an internal portion of a gene sequence into such a suicide plasmid, thus generating two incomplete gene copies upon integration of the plasmid into the chromosome.)
Plasmid vectors used as the backbone for such insertional mutagenesis have several properties desirable for this type of gene inactivation. (1) The plasmid must be conditional for replication to allow selection for integration into the chromosome. This can be achieved by using a plasmid that is able to replicate autonomously only in permissive hosts or by using conditional replicons (e.g. a plasmid that is temperature sensitive for replication). (2) The plasmid must carry a selectable marker (e.g. antibiotic resistance). (3) Ideally, the plasmid should be transferable to a variety of other bacteria. Plasmids that can be transferred by conjugation are preferable for situations in which other means of transfer such as transformation or electroporation are not efficient. (4) It is convenient if the plasmid has an array of unique cloning sites.
The plasmid pGP704 is an example of such a system. Plasmid pGP704 is a derivative of pBR322 that is AmpR but has a deletion of the pBR322 origin of replication (oriE1) but that carries, instead, a cloned fragment containing the origin of replication of plasmid R6K. The R6K origin of replication (oriR6K) requires for its function a protein called pi, encoded by the pir gene. In E. coli the pi protein can be supplied in trans by a prophage (lambda pir) that carries a cloned copy of the pir gene. The plasmid also contains a 1.9-kb BamHI fragment encoding the mob region of RP4. Thus, pGP704 can be mobilized into recipient strains by transfer functions provided by a derivative of RP4 integrated in the chromosome of E. coli strain SM10. However, once transferred it is unable to replicate in recipients that lack the pi protein.
Insertion mutations in a chromosomal gene are isolated by first subcloning DNA fragments into pGP704, then mating the plasmid clone into a recipient which lacks the pir gene. Because these plasmids cannot replicate in the recipient, the cloned gene can be incorporated into the recipient genome by homologous recombination involving two crossovers between the the gene present on the plasmid and the corresponding gene in the chromosome. The procedure requires that the mutation has an easily selectable phenotype (for example, if the cloned copy of the gene is disrupted with a KanR insertion). This method can also be used for integrating gene fusions on plasmids into the chromosome to study expression of the gene fusion in single copy and its wild-type context.
Use of suicide vectors for allelic exchange of nonselectable mutations. Because double-crossover events that incorporate a gene from a plasmid into the chromosome are rare, it is not feasible to simply screen for such events if the cloned gene cannot be directly selected. In such cases, a two-step procedure is used instead. First, the entire plasmid is integrated into the chromosome by a single-crossover between the homologous genes, producing a chromosomal duplication. Second, the chromosomal duplication is segregrated by homologous recombination between the flanking direct repeats, ultimately leaving one copy of the gene on the chromosome -- either the wild-type copy or the mutant copy.
Because the direct repeats are often short, the desired recombination event may be very rare. The ApR phenotype of the plasmid provides a direct selection for integration of the plasmid into the chromosome. Once the plasmid is integrated, it is possible to simply screen for segregrants by testing for the loss of the plasmid ApR marker. However, because the segregrants are rare, often a way of selecting for the loss of the integrated plasmid from the chromosome is needed. This can be accomplished by including a counter-selectable marker on the plasmid.
One way of selecting against the integrated plasmid is to use a plasmid that also carries the sacB gene from Bacillus subtilis. Expression of the sacB gene is toxic for gram-negative bacteria when grown in the presence of 5% sucrose, providing a direct selection for loss of the plasmid. The resulting segregrants SucroseR colonies are screened for the simultaneous loss of ApR to ensure that that the sucrose-resistant phenotype is due to loss of the integrated plasmid.
Many modifications on this basic theme have been used in a variety of bacteria. Two other slightly different approaches are worth describing. (1) The inability of plasmids with a colE1 origin to replicate in some bacterial species has been used for allelic exchange in Pseudomonas aeruginosa. (2) Another approach takes advantage of the recessive nature of streptomycin resistance conferred by chromosomal rpsL mutations. Plasmid pRTP1 is a colE1 replicon carrying the wild type rpsL gene in addition to AmpR. The colE1 origin does not support replication in a variety of bacterial species (ie. Bordetella pertussis), and thus acts as a suicide vector when delivered from E. coli into such strains. Integration of the plasmid carrying the wild type rpsL gene into a StrR recipient results in a merodiploid strain that is StrS. Subsequent selection for StrR results in colonies that have lost the vector (like the segregration of sacB vectors described above).
Incompatible plasmids. The incompatibility of certain plasmids has been widely used to select for marker exchange in a variety of gram-negative species. This procedure requires three steps. (1) The mutant DNA fragment is cloned into a broad host range plasmid -- for example, a KanR insertion mutation in a gene may be cloned into a TetR pLAFR derivative maintained in an E. coli host with the necessary mobilization functions. (The pLAFR plasmids require helper functions for their mobilization). (2) The KanR TetR pLAFR plasmid is then mated into the recipient strain by selecting for TetR (with an appropriate counterselection against the donor cells). (3) Recipients carrying this pLAFR plasmid are then mated with a strain carrying a second IncP1 plasmid with a different selectable marker. Plasmid pPH1JI which confers resistance to gentamicin (Gen) is often used for this step (sometimes called "the kickout step"). Since two IncP1 plasmids cannot coexist in the same cell, selection for the GenR from the pPH1JI plasmid results in loss of the KanR TetR pLAFR plasmid. However, if both GenR and KanR are selected, exconjugants arise that have lost the original TetR pLAFR plasmid, and the KanR marker has integrated into the chromosome via the homology surrounding the insertion. Loss of the pLAFR vector can be confirmed by scoring for TcS. Plasmid pPH1JI is somewhat unstable at high temperature, such that it can be cured from the final strain at a frequency of between 5 and 20% by simply growing overnight at 42 degrees in the absence of antibiotic selection.
Confirmation. After putative mutants have been isolated, the presence of the chromosomal insertion should be verified. If the mutation is an insertion or deletion that results in a significant size difference, the allelic exchange event can confirmed by Southern hybridizations or colony PCR. If the mutation is a base substitution, the colony PCR product can be directly sequenced, or confirmed by RPLP analysis if the mutation alters the restriction digest pattern of the amplified fragment.
For a nice review of counterselectable markers, see Reyrat et al. 1998. Infection Immunity 66: 4011-4017.