During meiosis, homologous chromosomes—maternal and paternal chromosomes that have almost identical DNA sequences—pair up and exchange sequences by recombination. To achieve this, both DNA strands in the participating chromosomes must be cleaved, and then hybrid DNA sequences are formed by joining fragments from the homologous chromosomes. But homologous recombination is not just a special characteristic of germ cells: it is also deployed in somatic cells as a way of repairing double-strand DNA breaks, and to rescue stalled or collapsed replication forks.
Unlike in meiosis, there is no mechanism for pairing of homologous chromosomes in somatic cells. However, after the DNA of a single chromosome has replicated, the two double-stranded DNA molecules are held tightly together, forming sister chromatids. Sister chromatid exchange—a type of homologous “recombination” —involves formation of double-strand breaks at analogous locations on the paired DNA molecules of sister chromatids, followed by joining of fragments originally located on different DNA molecules. an invading DNA strand from a chromatid can be used as a template to direct repair of the break on the opposing sister chromatid, a form of homologous recombination that allows flawless repair of spontaneous double strand breaks in cellular DNA. Homologous recombination can occasionally also take place between sequences on homologous chromosomes in somatic cells.
The hallmark of homologous recombination is that it takes place only between DNA duplexes having extensive regions of sequence similarity (sequence homology). Stable base pairing is a necessary requirement because recombination occurs only after invading single strands from each DNA duplex form heteroduplexes with a sufficiently high degree of base matching.
Homologous recombination is not just confined to naturally occurring DNA molecules in a cell. By introducing a transgene containing sequences that are identical to the sequence of a specific portion of a chromosomal DNA molecule, artificial homologous recombination may take place between the introduced transgene and the homologous target sequence on a chromosomal DNA molecule. That approach enabled the first attempts to alter the genome of intact cells, in a pre-determined way, and at a predetermined specific target sequence. Often the target sequence is a gene, and this use of genome editing is sometimes known as gene targeting.
Genome editing using homologous recombination: general strategies and the need for selection systems
Using artificial homologous recombination, a specific target gene of interest can be modified within an intact cell in any way that we want. Usually, a suitable DNA clone containing a faithful copy of the target sequence is modified in some way so that a central segment has a modified sequence, or an entirely new sequence. The transgene is transfected into suitable cultured cells, and then the cells are screened to identify rare cases where homologous recombination between the transgene and the chromosomal target sequence has resulted in transfer of the altered sequence into the target sequence in chromosomal DNA (Figure 1A).
Fig1. Homologous recombination can be used to replace a specific chromosomal DNA sequence with a similar but modified transgene sequence. (A) The general principle. A transgene, designed to contain sequences identical to those at a desired target site on a chromosomal DNA molecule apart from a central region that has some modified or new sequence (MOD), can recombine with the endogenous target sequence, thereby introducing the desired modified sequence into the target site. (B) The modified sequence has a deletion. (C) A reporter gene may be introduced into a genomic site and expressed under the control of an endogenous promoter (P). Expression can be tracked in all cells that descend from the transfected cell, which can be important when transfected cells are introduced into the germ line to make transgenic animals (see main text). (D) Editing a codon in genomic DNA so that it makes a different amino acid (ATG, specifying methionine, is replaced by AGG, which specifies arginine).
The strategy might be to delete an entire gene, a specific exon, or a specific regulatory sequence, such as an enhancer sequence. Very precise changes can also be made, including substituting, deleting, or inserting a single nucleotide at a unique, pre-determined position. Or it may be desirable instead to insert some advantageous sequence. Figure 1B–D shows some possibilities.
The gene targeting may be designed to produce different outcomes. We may wish to inactivate the gene, for example, the first stage in an exercise where the endpoint is homozygous gene inactivation (gene knockout) as a way of trying to establish what the gene normally does in a cell. That may be achieved by deleting the whole gene, but because genes are often large it may be simpler to delete key sequence elements. In the case of a protein-coding gene, a general approach might be to delete one or more of the first few exons so as to produce a frameshift in the translational reading frame. Alternatively, the motivation may be to test whether a candidate pathogenic mutation really is pathogenic. A requirement would be that there is a functional assay for the gene in question.
There is a problem: homologous recombination normally occurs at extremely low frequencies in mammalian cells—about 104–105 times less frequently than random integration (when the transgene becomes inserted at sites of temporary chromo some breaks). Because the frequency depends on both the length of the homologous regions and the degree of base matching, it is usual for the transgene to have a couple of homologous sequences that are several kilobases long and show 100% identity to sequences at the target site. Even then, genuine homologous recombination events are so infrequent (compared to random integrations) that transgenes also need to carry marker genes that allow selection for cells in which homologous recombination has occurred. A widely used approach, positive-negative selection, uses a marker gene that is intended to be inserted into the target sequence (positive selection) plus a different marker gene that is located near the end of the transgene and outside of the region of homology; Figure 2 demonstrates the principle.
Fig2. Using positive-negative selection to select cells containing a desired gene-targeting event. Here, the linearized plasmid used for gene targeting contains two marker genes: the neomycin phosphotransferase gene (neo), which confers resistance to neomycin and its analog G418, and the Herpes simplex thymidine kinase gene (tk). (A) In some gene-targeting events that occur by homologous recombination, a double crossover leads to incorporation of the neo gene but not of the tk gene. (B) Random integration of the gene-targeting construct at a chromosome break leads to integration of both the neo and tk genes. Suitably modified cells can be identified by selecting for neo+tk− cells.
Site-specific recombination allows conditional gene inactivation and chromosome engineering
Site-specific recombination systems are naturally used by several bacteriophages and by bacteria and yeast. In such systems, the recombinase recognizes a specific recognition sequence and induces recombination between two copies of the recognition sequence. Recognition sites for recombinases can be engineered easily into transgenes or targeting vectors. The recombinase enzymes can be provided conditionally by supplying genes that are expressed under the control of regulated or inducible promoters.
Two of these site-specific recombination systems have been widely employed in genetic manipulation of mammalian cells: the Cre–loxP system derived from bacteriophage P1, and the FLP-FRT system derived from the 2 μm plasmid of S. cerevisiae. Both the Cre (causes recombination) recombinase and the FLP (flippase) recombinase recognize specific 34 bp target sequences: respectively, the loxP sequence and the FRT sequence (FLP recombinase target).
Although their sequences are different, loxP and FRT have essentially the same structure: inverted 13 bp repeats separated by a central, asymmetric 8 bp spacer (Figure 3A). If two copies of the recombinase target site are located on the same DNA molecule and in the same orientation, recombination results in excision of the intervening sequence; if the loxP sites are in opposite orientations, recombination produces an inversion that can be very large. Site-specific recombination between two target sequences on different DNA molecules is also possible and can produce chromosome translocations (Figure 3B).
Fig3. Site-specific recombination using Cre–loxP and FLP-FRT. (A) The 34 bp recognition sequences for loxP and FRT. Both sequences contain an asymmetric 8 bp core sequence (shown in bold) flanked by two 13 bp inverted repeats (arrows). The phage P1 Cre recombinase recognizes the loxP (locus of X-over, P1) sequence and the S. cerevisiae FLP (flippase) recombinase recognizes the FRT (flippase recognition target) sequence. In each case, a recombinase monomer binds to each inverted repeat, and the two monomers form an active dimer that asymmetrically cleaves the central core sequence; breakpoints following cleavage by Cre recombinase (top) or flippase (bottom) are shown by yellow triangles. The asymmetric central sequence confers orientation: GCATACAT for loxP in the orientation shown here, and ATGTATGC for the opposite orientation. (B) Standard homologous recombination is first carried out to insert transgenes containing a loxP sequence into desired sites in the genome. (i) Two transgenes containing loxP sequences can be inserted in the same orientation flanking a sequence of interest (A), whereupon introduction of a Cre cDNA transgene to express Cre recombinase results in recombination between the two loxP sequences and deletion of sequence A. (ii) Alternatively, loxP sequences can be inserted in opposing orientations flanking a sequence of interest (a-b-c-d-e-f). Expressing an introduced Cre cDNA transgene then results in an inversion. (iii) Another possibility is to engineer a specific translocation by expressing a Cre cDNA transgene in cells after two loxP sequences have been inserted into desired target sites on different chromosomes.
The Cre–loxP system has been applied in many different ways. It can allow site- specific integration of transgenes, conditional activation and inactivation of transgenes, and the deletion of unwanted marker genes. Perhaps the most important applications, however, are conditional gene inactivation (described in Section 8.6) and conditional recombination.
