Замещение Последовательностей Человек → Мышь

Manipulating the Mouse Genome to Engineer Precise Functional Syntenic Replacements with Human Sequence Helen A.C. Wallace, Fatima Marques-Kranc, MeMlle Richardson, Francisco Luna-Crespo, Jackie A. Sharpe, Jim Hughes, William G. Wood, Douglas R. Higgs, and Andrew J.H. Smith Cell V.128, No 1, P.197-209, 2007
Разработана стратегия замены крупных сегментов (более 100 т.п.н.) генома мыши эквивалентной синтеничной областью человека. Технология связана с модификацией хромосом ES клеток мыши и ВАС человека путем введения гетеротипических lox сайтов, для фланкирования предполагаемых замен, и затем с использованием Cre recombinase для получения замен сегмента. Продемонстрирована эффективность подхода на примере замещения регуляторного домена α globin мыши синтеничной областью и последующей генерацией мышей, гомозиготных только по цепочке α globina человека. Кроме того модифицированные ES клетки можно использовать для функциональных исследований. Например, был использован RMGR для продукции точной мышиной модели человеческой a таллассемии.	Studies of how human genes are normally regulated and the effects of mutations that cause human genetic disease are limited by the availability of appropriate primary cells and tissues and the necessary constraints on experimental interventions. Consequently, over the past 20 years, transgenic mice have become Important experimental models to understand how human genes are regulated. Small transgenes, however, often do not contain all of the cis-acting elements required for fully regulated expression since such sequences may be found tens or even hundreds of kilobases from the gene in question (Kleinjan and van Heyningen, 2005). Even when large transgenes derived from PACs, BACs or YACs are used, they are frequently rearranged; it is often very difficult to fully analyze their structural integrity and copy number (Peterson et al., 1998). Furthermore, their expression is often influenced by their position of integration in the genome (Alami et al., 2000; Kaufman et al., 1999). Making and characterizing directed mutations for structure/function studies is also difficult when using such large molecules. Even when successful, the interpretation of these transgenic experiments is complicated by the fact that the endogenous mouse genes are still present unless bred against mice in which they have been deleted or inactivated by homologous recombination. A complementary approach is to Identify the mouse ortholog and alter this gene or its regulatory elements by homologous recombination using conventional gene-targeting methods (Nagy, 2003). Clearly, mutating a single gene at its normal chromosomal locus avoids many of the problems associated with transgenes (copy number differences, complex mapping problems, position effects), and, in this process, the normal endogenous gene does not remain intact to complicate the analysis. However, there is increasing evidence demonstrating considerable differences in the ways In which orthologous human and mouse genes are normally regulated. In many instances, mutations known to cause disease in human do not mimic the phenotype when introduced into the mouse ortholog (Cdledge et al., 1995; Engle et al., 1996; Garrick et al., 2006). These include the α globin locus, where deletion of the major regulatory element causes a much milder condition in mouse than In human (Angurta et al., 2002). Therefore, although mouse models have provided many Insights into human genetic disease, It is becoming Increasingly clear that they are also beset by inherent problems and difficulties In Interpretation resulting from differences in basic biological processes that have evolved over the 70 million years of evolution that separate humans and mice. Clearly, the next step in developing mouse models of human disease is to replace farge-segments (100-1-OOOs kb) of the mouse genome with the wild-type or mutated syntenic region of the human sequence, thereby including all remote regulatory elements and inserting the human segment as a single copy at a natural chromosomal position for the genes contained within it A range of chromosome-engineering technologies are now available that, combined in a novel way, could make this feasible (Yu and Bradley, 2001; Copeland et al., 2001; Muyrers et al., 2001; Testa et al., 2003; Valenzuela et al., 2003; Yang and Seed, 2003), including application of various heterotypic site-specific recombination systems (Cre, FLP, >C31) to achieve efficient exchange of sequence cassettes at targeted chromosomal loci, a process known as recombinase-mediated cassette exchange (RMCE) (Baer and Bode, 2001). In this study, we have combined various aspects of these technologies to develop what we believe to be a new, general strategy for replacing large segments of the mouse genome with wild-type or mutated syntenic regions of the human genome. We refer to this process as recombinase-mediated genomic replacement (RMGR). This allows precise, reproducible, genetically selectable genomic replacement on a scale not previously attained and should be generally applicable to large-scale modification of the mouse genome. We have established a "proof of principle" for its application in generating "humanized" mice by replacing the syntenic region encompassing the mouse α globin regulatory domain (the region containing all cis-acting sequences required for fully regulated expression of the α globin genes) with human BAC-derived sequence in ES cells and subsequently deriving viable mice homozygous for this syntenic substitution producing only human a-globin chains. In addition, we have demonstrated the potential of this approach for making authentic mouse models of human disease by replacing the mouse region with a recombineered human BAC in which the α globin major regulatory element has been deleted from the α globin regulatory domain. This created a model of a thalassemia that faithfully mimics the phenotype of patients with this common human genetic disease. DISCUSSION Here we have shown that it is possible to precisely replace a large segment of the mouse genome containing multiple loci with the corresponding, syntenic segment of the human genome. Combining a variety of developments in chromosome engineering. RMGR has significantly extended the principles underpinning RMCE to enable extensive tracts (100s to 1000s kb) of mouse genomic DNA to be replaced by the orthologous sequences of other species cloned in BAC libraries. RMGR will therefore be widely applicable and offers several advantages over other approaches to large-scale genome modification. In particular, correct genomic replacement using RMGR is genetically selectable, precise, and reproducible. This reduces the effort of screening and greatly simplifies the characterization of the chromosomal breakpoints. Minimal extraneous DNA remains to affect the functional integrity of the engineered chromosomal region, provided the locations of the recombination sites are chosen carefully. Once correctly targeted, germline transmitting ES cells are established, RMGR allows iterative replacement with mutated BACs for detailed functional dissection of genomic sequences. Although the frequency of replacement Is low, it is reproducible. General application may reveal locus-specific and/or BAC-specific variations in efficiency requiring adjustments to the electroporation parameters. The RMGR strategy will be adaptable for application with other (e.g.,φ C31) (Berteki et al., 2003) site-specific recombination systems to improve the overall efficiency and allow the Cre/lox system to be used to introduce conditional deletions within the human sequence. It is present form, the extent of sequence replacement is limited only by the size of the BAC Inserts. Recent approaches to construct mega-BACs (Shen et a!., 2005) in E. coli will therefore extend the size range of replacements possible. By analyzing the well-characterized α globin regulatory domain by RMGR, we could be confident that the structurally intact segment contains all known cis-acting elements, at a single copy in a chromosomal position that is appropriate for globin gene expression. Furthermore, by carefully examining gene expression in erythropoiesis and evaluating the resulting, sensitive red cell phenotype, we could accurately assess the degree to which regulation of the human globin genes mimic their normal pattern and levels of expression in the mouse model. We found that under these conditions the human genes are expressed in an appropriate developmental stage- and tissue-specific manner, although the level of expression of the a genes is only about 40% of that of the endogenous mouse α globin genes. This leads to abnormal red blood cells, and homozygotes for the human synteny replacement may die prematurely even though they are not frankly anemic. These differences between expression of the human gene in man and mouse could result from changes in the chromosomal environment, although the relationship between transcription and nuclear sublocalization remains unclear (Brown et al., 2006). Perhaps more likely, changes in the structure or recognition sequences of the key transcription factors have altered during the evolution of the two species such that the binding or stability of mouse transcription factors on human sequences are suboptimal. Knowing the phenotype resulting from expression of the normal human α globin cluster in mouse, it is now possible to compare and evaluate the effects of specific mutations introduced into the human sequence. Using RMGR, the only variable is the mutation itself. Such mutations could either be introduced by further rounds of homologous recombination in the humanized ES cells or, as demonstrated here, by recombineering the appropriate mutation into the modified BAC followed by a new round of RMGR at the target region. Again using the α globin cluster as the model we could show that removal of just 1.1 kb containing the major regulatory element (HS -40) virtually abolishes human α globin expression consistent with previous extensive observations (Higgs et al., 1998). In homozygotes, this deletion mimics the lethal clinical phenotype of hydrops fetalis seen in human patients with this degree of et globin deficiency. We have previously shown that removal of the mouse sequence (HS -26}, which is the equivalent of the human major regulatory element (HS -40), from the endogenous locus only reduces mouse α globin expression to ~50% of normal, and homozygotes for this allele survive with only mild anemia rather than developing hydrops fetalis (Angurta et al., 2002). This clearly shows that the detailed regulation of human and mouse genes may be quite different and emphasizes the point -that extrapolating conclusions from studying regulation of an orthologous gene may be misleading when one wishes to determine how a particular human gene is regulated. It seems likely that the observations set out here represent general recurrent issues that arise when making and analyzing mouse models of human gene regulation and genetic disease. RMGR will exclude most of the variables associated with current transgenic experiments, but it is still likely that human genes will not be completely faithfully ' regulated In mouse systems. However, comparing wild-type and mutant chromosomal domains iteratively inserted into an appropriate chromosomal environment will allow one to determine more precisely the effect of specific mutations. This will not only be of value when studying previously well-characterized loci but will also enable those studying complex traits to assess the functional role of mutations and SNPs localized within less well characterized large segments of the human genome.

Studies of how human genes are normally regulated and the effects of mutations that cause human genetic disease are limited by the availability of appropriate primary cells and tissues and the necessary constraints on experimental interventions. Consequently, over the past 20 years, transgenic mice have become Important experimental models to understand how human genes are regulated.

Small transgenes, however, often do not contain all of the cis-acting elements required for fully regulated expression since such sequences may be found tens or even hundreds of kilobases from the gene in question (Kleinjan and van Heyningen, 2005). Even when large transgenes derived from PACs, BACs or YACs are used, they are frequently rearranged; it is often very difficult to fully analyze their structural integrity and copy number (Peterson et al., 1998). Furthermore, their expression is often influenced by their position of integration in the genome (Alami et al., 2000; Kaufman et al., 1999). Making and characterizing directed mutations for structure/function studies is also difficult when using such large molecules. Even when successful, the interpretation of these transgenic experiments is complicated by the fact that the endogenous mouse genes are still present unless bred against mice in which they have been deleted or inactivated by homologous recombination.

A complementary approach is to Identify the mouse ortholog and alter this gene or its regulatory elements by homologous recombination using conventional gene-targeting methods (Nagy, 2003). Clearly, mutating a single gene at its normal chromosomal locus avoids many of the problems associated with transgenes (copy number differences, complex mapping problems, position effects), and, in this process, the normal endogenous gene does not remain intact to complicate the analysis. However, there is increasing evidence demonstrating considerable differences in the ways In which orthologous human and mouse genes are normally regulated. In many instances, mutations known to cause disease in human do not mimic the phenotype when introduced into the mouse ortholog (Cdledge et al., 1995; Engle et al., 1996; Garrick et al., 2006). These include the α globin locus, where deletion of the major regulatory element causes a much milder condition in mouse than In human (Angurta et al., 2002). Therefore, although mouse models have provided many Insights into human genetic disease, It is becoming Increasingly clear that they are also beset by inherent problems and difficulties In Interpretation resulting from differences in basic biological processes that have evolved over the 70 million years of evolution that separate humans and mice.

Clearly, the next step in developing mouse models of human disease is to replace farge-segments (100-1-OOOs kb) of the mouse genome with the wild-type or mutated syntenic region of the human sequence, thereby including all remote regulatory elements and inserting the human segment as a single copy at a natural chromosomal position for the genes contained within it A range of chromosome-engineering technologies are now available that, combined in a novel way, could make this feasible (Yu and Bradley, 2001; Copeland et al., 2001; Muyrers et al., 2001; Testa et al., 2003; Valenzuela et al., 2003; Yang and Seed, 2003), including application of various heterotypic site-specific recombination systems (Cre, FLP, >C31) to achieve efficient exchange of sequence cassettes at targeted chromosomal loci, a process known as recombinase-mediated cassette exchange (RMCE) (Baer and Bode, 2001).

In this study, we have combined various aspects of these technologies to develop what we believe to be a new, general strategy for replacing large segments of the mouse genome with wild-type or mutated syntenic regions of the human genome. We refer to this process as recombinase-mediated genomic replacement (RMGR). This allows precise, reproducible, genetically selectable genomic replacement on a scale not previously attained and should be generally applicable to large-scale modification of the mouse genome. We have established a "proof of principle" for its application in generating "humanized" mice by replacing the syntenic region encompassing the mouse α globin regulatory domain (the region containing all cis-acting sequences required for fully regulated expression of the α globin genes) with human BAC-derived sequence in ES cells and subsequently deriving viable mice homozygous for this syntenic substitution producing only human a-globin chains. In addition, we have demonstrated the potential of this approach for making authentic mouse models of human disease by replacing the mouse region with a recombineered human BAC in which the α globin major regulatory element has been deleted from the α globin regulatory domain. This created a model of a thalassemia that faithfully mimics the phenotype of patients with this common human genetic disease.

DISCUSSION

Here we have shown that it is possible to precisely replace a large segment of the mouse genome containing multiple loci with the corresponding, syntenic segment of the human genome. Combining a variety of developments in chromosome engineering. RMGR has significantly extended the principles underpinning RMCE to enable extensive tracts (100s to 1000s kb) of mouse genomic DNA to be replaced by the orthologous sequences of other species cloned in BAC libraries. RMGR will therefore be widely applicable and offers several advantages over other approaches to large-scale genome modification. In particular, correct genomic replacement using RMGR is genetically selectable, precise, and reproducible. This reduces the effort of screening and greatly simplifies the characterization of the chromosomal breakpoints. Minimal extraneous DNA remains to affect the functional integrity of the engineered chromosomal region, provided the locations of the recombination sites are chosen carefully. Once correctly targeted, germline transmitting ES cells are established, RMGR allows iterative replacement with mutated BACs for detailed functional dissection of genomic sequences. Although the frequency of replacement Is low, it is reproducible. General application may reveal locus-specific and/or BAC-specific variations in efficiency requiring adjustments to the electroporation parameters. The RMGR strategy will be adaptable for application with other (e.g.,φ C31) (Berteki et al., 2003) site-specific recombination systems to improve the overall efficiency and allow the Cre/lox system to be used to introduce conditional deletions within the human sequence. It is present form, the extent of sequence replacement is limited only by the size of the BAC Inserts. Recent approaches to construct mega-BACs (Shen et a!., 2005) in E. coli will therefore extend the size range of replacements possible.

By analyzing the well-characterized α globin regulatory domain by RMGR, we could be confident that the structurally intact segment contains all known cis-acting elements, at a single copy in a chromosomal position that is appropriate for globin gene expression. Furthermore, by carefully examining gene expression in erythropoiesis and evaluating the resulting, sensitive red cell phenotype, we could accurately assess the degree to which regulation of the human globin genes mimic their normal pattern and levels of expression in the mouse model. We found that under these conditions the human genes are expressed in an appropriate developmental stage- and tissue-specific manner, although the level of expression of the a genes is only about 40% of that of the endogenous mouse α globin genes. This leads to abnormal red blood cells, and homozygotes for the human synteny replacement may die prematurely even though they are not frankly anemic. These differences between expression of the human gene in man and mouse could result from changes in the chromosomal environment, although the relationship between transcription and nuclear sublocalization remains unclear (Brown et al., 2006). Perhaps more likely, changes in the structure or recognition sequences of the key transcription factors have altered during the evolution of the two species such that the binding or stability of mouse transcription factors on human sequences are suboptimal.

Knowing the phenotype resulting from expression of the normal human α globin cluster in mouse, it is now possible to compare and evaluate the effects of specific mutations introduced into the human sequence. Using RMGR, the only variable is the mutation itself. Such mutations could either be introduced by further rounds of homologous recombination in the humanized ES cells or, as demonstrated here, by recombineering the appropriate mutation into the modified BAC followed by a new round of RMGR at the target region. Again using the α globin cluster as the model we could show that removal of just 1.1 kb containing the major regulatory element (HS -40) virtually abolishes human α globin expression consistent with previous extensive observations (Higgs et al., 1998). In homozygotes, this deletion mimics the lethal clinical phenotype of hydrops fetalis seen in human patients with this degree of et globin deficiency. We have previously shown that removal of the mouse sequence (HS -26}, which is the equivalent of the human major regulatory element (HS -40), from the endogenous locus only reduces mouse α globin expression to ~50% of normal, and homozygotes for this allele survive with only mild anemia rather than developing hydrops fetalis (Angurta et al., 2002). This clearly shows that the detailed regulation of human and mouse genes may be quite different and emphasizes the point -that extrapolating conclusions from studying regulation of an orthologous gene may be misleading when one wishes to determine how a particular human gene is regulated.

It seems likely that the observations set out here represent general recurrent issues that arise when making and analyzing mouse models of human gene regulation and genetic disease. RMGR will exclude most of the variables associated with current transgenic experiments, but it is still likely that human genes will not be completely faithfully ' regulated In mouse systems. However, comparing wild-type and mutant chromosomal domains iteratively inserted into an appropriate chromosomal environment will allow one to determine more precisely the effect of specific mutations. This will not only be of value when studying previously well-characterized loci but will also enable those studying complex traits to assess the functional role of mutations and SNPs localized within less well characterized large segments of the human genome.