Chromosome Engineering In The Mouse

The conventional gene knockout technology can generate identifiable genomic deletions up to 19 kb (24). However, this is not sufficient to model large chromosome rearrangements found in human genomic disorders. Chromosome rearrangements can be induced by either X-ray or chemicals in the mouse (25-27). However, the randomness of the endpoints in the irradiation-induced chromosome rearrangements has hampered broader applications for such genetic approaches to produce research resources. Chromosome engineering technology was developed to enable the construction of rearrangements of defined sizes in specific genomic locations. It combines gene targeting that defines chromosome rearrangement endpoints, and a site-specific recombination system from bacteriaphage P1, Cre-loxP, that can mediate recombination between two loxP sites. Cre is a 38-kDa recombinase that alone can catalyse recombination between two 34-bp loxP sites in mammalian cells (28). Prior to its application in chromosome engineering, Cre-loxP system had been used to make small genomic deletions in conditional knockout mice (29,30).

In 1993, while a graduate student in the laboratory of Professor Allan Bradley, together with Ramiro Ramirez-Solis, a postdoctoral fellow in the lab, I tested the Cre-loxP system for its ability to induce large chromosomal rearrangements. It was known that the Cre-loxP system is capable of efficiently inducing smaller deletions of a few thousands basepairs in mammalian cells. We anticipated that the intrachromosomal recombination efficiencies between two loxP sites that are millions of basepairs away, or that are on two different chromosomes, could be very low and a positive selection might, therefore, be required to detect and to recover the large recombination products. In an ideal positive selection system, a selection marker would be activated on Cre-loxP recombination and the recombinants could, thus, be selected out in appropriate medium. We chose the human hypoxanthine phosphoribosyl transferase (HPRT) minigene as the positive selection marker for two reasons. First, an Hprt deficient ES cell line, AB2.2, was available in the lab and it had been shown to be capable of highly efficient germline transmission. The second reason is the unique structure of the HPRT minigene which contains intron 2 of the human HPRT gene that is required for efficient expression in mouse ES cells. A loxP site can therefore be conveniently inserted into the middle of this intron (XbaI site) without affecting the HPRT minigene activity. We subsequently divided the HPRT minigene into two non-functional parts, the 5' and the 3' cassettes with the intronic loxP site embedded in both cassettes (Fig. 1). On recombination mediated by loxP sites and catalyzed by Cre, a full-length HPRT minigene is reconstituted and the recombinant can be selected out using hypox-anthine amniopterin thymidine (HAT) medium (Fig. 1).

To generate a defined chromosome rearrangement, the recombination cassettes need to be targeted, on a sequential manner, to two endpoints. We tested the ability of the HPRT selection system to generate chromosome rearrangements between the Hsd17$ and the Gastrin loci separated by approx 1 Mb on the mouse chromosome 11 (31). As illustrated in Fig. 1, the 5' HPRT cassette was targeted to the Hsd17$ locus in ES cells. The 3' HPRT cassette was then targeted to the Gastrin locus in ES cells with the 5' HPRT cassette already targeted into the Hsd17$ locus. To induce the recombination, double-targeted ES cells were transfected with a plasmid that transiently expressed Cre. We recovered all of the expected recombinants in

Fig. 1. Chromosome engineering between Hsd17$ and Gastrin loci on the mouse chromosome 11. (A) Loci on the chromosome 11 are represented by a, b, c, and d. (B) The two recombination cassettes (5' and 3') are targeted consecutively to the two desired rearrangement endpoints (Hsd17^> and Gastrin) in embryonic stem (ES) cells. (C) Recombination between the two loxP sites catalyzed by Cre recombinase regenerates a full-length HPRT minigene that enables the deletion to be selected for in HAT medium. (D) When the two cassettes are targeted to the two chromosome 11 homologous, both the deletion and the duplication can be recovered in a single ES cell. Filled arrow, loxP site; P, puromycin resistance cassette; N, neomycin resistant cassette.

Fig. 1. Chromosome engineering between Hsd17$ and Gastrin loci on the mouse chromosome 11. (A) Loci on the chromosome 11 are represented by a, b, c, and d. (B) The two recombination cassettes (5' and 3') are targeted consecutively to the two desired rearrangement endpoints (Hsd17^> and Gastrin) in embryonic stem (ES) cells. (C) Recombination between the two loxP sites catalyzed by Cre recombinase regenerates a full-length HPRT minigene that enables the deletion to be selected for in HAT medium. (D) When the two cassettes are targeted to the two chromosome 11 homologous, both the deletion and the duplication can be recovered in a single ES cell. Filled arrow, loxP site; P, puromycin resistance cassette; N, neomycin resistant cassette.

HAT medium and confirmed the anticipated rearranged genomic structures by Southern hybridization (31).

In order to generate a desired chromosome rearrangement, it is critical to know the orientation of two endpoints relative to each other on the chromosome and to the centromere because of the nature of the selection. As we described in the initial chromosome engineering experi ments (31), to obtain a deletion without prior knowledge of the endpoint orientation, four types of double-targeted ES cells had to be generated and tested. For example, generation of a deletion requires the two recombination cassettes in the correct orientation so that after recombination the HPRT minigene is reconstituted on the deletion chromosome in order to survive HAT selection. The finished mouse genome sequence (9) now enables one to obtain the information regarding the position and orientation of the two endpoints for a specific rearrangement. As a result, the recombination substrates can be targeted to the endpoints and the desired rearrangement can be obtained reliably and diagnosed readily.

Cre-loxP-mediated recombination can occur either in cis (two loxP sites are on the same chromosome) or in trans (two loxP sites are on different chromosomes). Figure 1 shows that when two recombination cassettes are targeted to two homologous chromosomes, both the deletion and the duplication can potentially be generated and recovered in a single ES cell. When these genetically balanced ES cells contribute to the mouse germline, from one chimaera, offspring with either the deletion or the duplication chromosomes can be recovered (32).

Following the initial success with 1-Mb chromosomal rearrangements, we found that even larger chromosomal rearrangements could be successfully induced and recovered with the HPRT minigene selection system, with sizes ranging from 3 to 22 cM (32). Most importantly, after multiple rounds of manipulation, ES cells harboring the engineered chromosomal rearrangements still retained the competence for germline transmission. This finding is often overlooked but represents a key step in chromosome engineering because, after multiple rounds of manipulation, ES cells were originally thought to lose germline competence. Further extension of the HPRT minigene selection system has facilitated the development of induced mitotic recombination in mouse ES cells (33).

The ability to generate designed chromosomal rearrangements opens a new research field and offers new genetic resources. For example, in addition to modeling human disease, chromosome deletions can be used to uncover recessive mutations in genetic screens in vitro and in vivo (34). Chromosome inversions can be engineered to function as balancers to facilitate N-ethyl-N-nitrosourea mutagenesis screens and to maintain mutant mouse lines (35).

In addition to Cre-loxP-mediated chromosome engineering, other approaches have been developed to generate chromosome rearrangements. One of these approaches is to irradiate F1 hybrid mouse ES cells (36). The irradiated ES cells have a negative selection marker, Herpes simplex virus thymidine kinase (HSV-TK), targeted by homologous recombination to a defined locus. If deletions caused by irradiation encompass the TK-tagged locus, the cells will survive in 2'-fluoro-2'-deoxy-5-iodo-1-P-D-arabinofuranosyluracil medium. The advantage ofthis approach is that from a single experiment, multiple deletions of various sizes centered on the tagged locus can potentially be recovered. These deletions are useful for estimating haploinsufficiency tolerance of various chromosome regions in ES cells and may be used directly for in vitro genetic screens (37). Importantly, ES cells carrying these irradiation-induced deletions still retain their ability to contribute to germline development in chimaeras. Thus, large genomic deletions from these experiments are useful for estimating haploinsufficiency in the mouse prior to making precise chromosomal rearrangements in this genomic region using Cre-loxP and for modeling certain genomic disorders. For instance, several deletions on the mouse chromosome 5 syntenic to human 4p16.3 were produced in one experiment (38). The mice heterozygous for these deletions have phenotype similar to Wolf-Hirschhorn syndrome caused by monosomy 4p16.3. Thus, these deletion mice can directly serve as the mouse models for human disease syndromes (38).

Compared to the TK-anchored deletion strategy, chromosome engineering using Cre-loxP has advantages. First, defined chromosome rearrangements can only be obtained with the precisely engineered endpoints (preselected loci). Second, other types of rearrangements, such as duplications, inversions, and translocations, can also be generated with chromosome engineering. Third, when a large chromosome rearrangement needs to be induced only in somatic cells, site-specific recombination is much more efficient.

Was this article helpful?

0 0

Post a comment