Conditional Alleles in Mice Practical Considerations for Tissue

REVIEW Conditional Alleles in Mice:Practical Considerations for Tissue-Specific KnockoutsKin-Ming Kwan*Department of Molecular Genetics,University of Texas,M.D.Anderson Cancer Center,Houston,TexasReceived7January2002;Accepted8January2002Key words:conditional genetics;Cre-loxP;Flp-FRT;tissue-specific knockout;mosaicINTRODUCTIONRecent advances using site-specific recombinase systems to manipulate the mouse genome have opened new doors,revealing the complexity of gene function in mammals and for designing better models of human diseases(Torres and Kuhn,1997;Dymecki,2000).The application of these site-specific recombinase systems along with gene targeting techniques in embryonic stem (ES)cells have now made it possible to modify the mouse genome in almost any desired manner,from cre-ating specific point mutations to achieving large site-specific chromosomal rearrangements(Nagy and Mar, 2001;Yu and Bradley,2001).Furthermore,these tech-niques have proven to be very useful tools for activating and inactivating specific genes in a“conditional”manner in spatially and temporally restricted patterns(Lobe and Nagy,1998;Nagy,2000).This helps to circumvent the limitations of conventional gene inactivation,in particu-lar,that a given gene has multiple roles at different times and in different tissues.These methodologies are partic-ularly useful for elucidating a complete picture of gene function in which:1)the conventional knockout leads to an early lethal phenotype preventing the study of its later roles,and/or2)the phenotype affects multiple tissues, preventing the detailed study of its function in a partic-ular cell lineage.Basic Properties of Site-Specific Recombinase SystemsSite-specific recombinase systems consist of two basic elements:the recombinase enzyme and a DNA sequence that is specifically recognized by the particular recombi-nase.Recognition is through a consensus sequence and the recombinase catalyzes recombination between two of its recognition sites to bring about the modification of the associated DNA such as deletion,insertion,inver-sion,or translocation,depending on the orientation and location of the recognition sites(Fig.1B;Torres and Kuhn,1997).Currently,there are two such recombinase systems that have been applied in mouse gene targeting experiments:the Cre-loxP system from the bacterio-phage P1(Lakso et al.,1992;Orban et al.,1992)and the Flp-FRT system from the budding yeast Saccharomyces cerevisiae(Dymecki,1996).Each of these recombinases (Cre or Flp)recognize a34-bp consensus sequence con-sisting of two13-bp inverted repeatsflanking an8-bp nonpalindromic core that define the orientation of the overall sequence of the recognition site(loxP or FRT) (Fig.1A).This minimal34-bp target site is very unlikely to occur at random in the mouse genome,yet is small enough to be considered a“neutral”sequence when integrated into chromosomal DNA.Another major ad-vantage of these recombination systems is that the re-combinase functions autonomously with highfidelity,so no cofactors or additional sequence elements are re-quired for efficient recombination(Stark et al.,1992).In addition,recombination can occur over large distances (Mb)and in a wide range of cell types in vitro and in vivo,including undifferentiated as well as postmitotic cells(Sauer and Henderson,1988;O’Gorman et al., 1991;Agah et al.,1997).Together these properties make the Cre-loxP and Flp-FRT systems very useful for site-specific DNA modifications,especially for generating conditional alleles in mice.Basic Design of a Recombinase-Dependent Conditional Genetic ExperimentTwo basic reagents,usually in the form of two trans-genic mouse lines,are required for a recombinase-de-pendent conditional genetic experiment.One mouse line expresses the recombinase(Cre or Flp)in selected cell lineages or tissues at a specific stage of embryonic or postnatal development.Another line carries the condi-tional allele,a gene segment(or the entire gene)flanked by the two recognition target sites(loxP or FRT).After intercrossing,cells expressing the recombinase in the double transgenic offspring delete/modify the condi-*Correspondence to:Kin-Ming Kwan,Department of Molecular Genet-ics,University of Texas,M.D.Anderson Cancer Center,1515Holcombe Blvd.,Houston,TX77030.E-mail:kkwan@©2002Wiley-Liss,Inc. genesis32:49–62(2002)DOI10.1002/gene.10068tional allele,while the target gene remains intact and functional in the cells of all other tissues where the recombinase is not expressed.Thus,the function of the target gene can be studied in the cell/tissue type of interest at a speci fic time.One major advantage of this binary design is that it offers great versatility for gene function studies.For example,one floxed mouse line can be crossed to different Cre lines to study gene function in different cell types or tissues or,conversely,the one Cre line can be used to study the functions of different floxed genes in one lineage or tissue.The generation of mice equipped with site-speci fic recombination systems (Cre-loxP and/or Flp-FRT )is a powerful system for studying tissue-speci fic gene func-tion and for creating better human disease models.Suc-cess in achieving such goals depends both on reliable and consistent recombinase expression in vivo and the careful design of the conditional alleles.Currently,the most popular/powerful use of conditional genetics is its application for cell/lineage or tissue-speci fic knockouts by conditional gene inactivation.In a previous specialissue of genesis ,Nagy (2000)comprehensively reviewed the properties and applications of Cre recombinase for mouse genome manipulations.This current review will focus on the experimental design and practical consid-erations for constructing a conditional allele,the focus of this special issue of genesis .A summary of the literature of conditional knockout alleles in mice is also provided and discussed.DESIGNS FOR GENERATING A CONDITIONAL NULL ALLELETargeting Vector Construction for Conditional Gene InactivationThe basic concept underlying the design of a targeting construct for generating a conditional allele is to use gene targeting methods to insert two recombinase target sites flanking the entire gene of interest (or essential functional part of the gene).Basically,this type of tar-geting construct is similar to those used for the conven-tional gene knockout experiments (see Fig.2A;CheahFIG. 1.A :Structure and se-quence of the recombinase rec-ognition sites (loxP and FRT ).Each site consists of two 13-bp inverted repeats flanking an asymmetric 8-bp core region which de fine the orientation of the recognition site indicated by the red or blue arrow in the loxP or FRT sequence,respectively.B :Products of intra-and inter-molecular recombinase-medi-ated directional recombination between two recognition sites.The reactions are basically re-versible.The orientation of the recognition site is represented as a filled red triangle.50KWANand Behringer,2000)that includes the following essen-tial components and properties:1)two arms of isogenic DNA sequence homologous to that of the ES cells to be targeted;2)the length of homologous DNA sequence should be maximized and avoid repetitive sequences;3)a positive selection gene cassette (neomycin,hygromy-cin,or similar resistance gene)is placed between the two homology arms for selection for integration;and4)FIG.2.Strategies of conditional gene targeting for generating a loxP -flanked conditional null allele.The gray boxes represent exons of the gene of interest;the pink box with a ϩ,positive selectable marker such as neo r ;orange box with a Ϫ;negative selectable marker such as HSVtk;red triangles,loxP sites;blue ovals,FRT sites.A :2-loxP strategy.The selection marker is still in the final targeted conditional null allele.Some targeted ES cell clones will lack the 5ЈloxP site because of crossover events within the loxP -flanked genomic sequence (marked by X).The frequency of this event declines with the distance between the loxP site and the selection marker.Y marks restriction sites for Southern blot analysis.An engineered restriction site is introduced to mark the loxP site for diagnostic purposes as shown by the double-headed arrows.B,C :3-loxP strategy.After the gene replacement step,transient expression of Cre will generate three types of ES cell subclones due to partial recombination:one harboring the desired floxed allele (type II deletion),the other two are null alleles (type I and III deletions).The three different deletion events can be distinguished by PCR and Southern analyses.B :The negative selection maker is placed at one end of the targeting vector.Both positive and negative selection are used in the gene replacement step for more effective selection of the targeted clones with homologous recombination.C :The negative selection marker is floxed together with the positive selection marker.Only positive selection is used for enriching the targeted allele,while negative selection can be used for enriching the selection for the desired Cre-mediated type II deletion.D :2-loxP-2-FRT strategy:a two-step strategy using both the Cre-loxP and Flp-FRT recombination systems.The flrted positive selection marker is removed by Flp-mediated deletion either in vitro or in vivo after the gene replacement step.51CONDITIONAL ALLELES IN MICEa negative selection gene(usually thymidine kinase/TK or diphtheria toxin-A/DT-A)is placed at the end of a homology arm for selection against random integration. In addition to these basic components,the conditional gene inactivation construct should have at least two recombinase target sites placed in the same orientation flanking the entire gene of interest(or essential func-tional domain of the gene)(Fig.2A).Although the basic strategy is very simple and straightforward,there are a number of important considerations that have to be taken into account to generate a successful design for a conditional null allele.Experimental Design Considerations1)The placement of the recombinase target sites and the selectable marker(s)is critical.The strategy must avoid any potential interference with wild-type gene function and yet keep the conditional allele fully func-tional for recombination.Basically,the target sites and the selectable markers should be placed outside coding regions or within introns and avoiding known or poten-tial regulatory regions.On the other hand,the target sites should be placed in a manner whereupon excision should result in a null allele.The most obvious strategy would beflanking the entire coding region with the recombinase target sites.However,this may be difficult if the gene is large,the gene structure is undefined, and/or there are regulatory elements lying in the5Јor3Јflanking regions.In most cases,the target sites are placed toflank the exon(s)(flanked by loxP sites,floxed;flanked by FRT sites,flrted)that(a)encodes a function-ally essential part of the gene(including the one with the translation start site)and/or(b)results in a frameshift after excision by the recombinase.It is usually better to flank the exon(s)in the5Јgene segment to reduce the chances of making a partial protein product.Thus,it is important to have a defined gene structure and to be aware of any alternative splicing of the gene of interest. Also,information from previous convention knockout experiments(if any)should help to determine which exon(s)should beflanked with target sites.If you are flanking internally located exons,then you should deter-mine by sequence analysis whether the mRNA encoded by the remaining exons after excision by the recombi-nase could still potentially generate a protein.Thus,it is best toflank exons that upon excision results in frame-shifts.2)It is important to design an optimal screening strat-egy to identify targeted ES cell clones.One should note that the efficiency of obtaining the correct targeted clone containing all the recombinase target sites declines with the distance between the target site and the select-able marker(Fig.2A).It is very useful to mark the recombinase target sites by introducing an appropriate restriction enzyme site(s)that remains even after Cre-mediated excision.Thus,one can screen directly for the homologous recombination event as well as the integra-tion of all of the target sites in the correct manner by Southern blot analysis(Fig.2A).Moreover,it is essential to design a strategy(PCR or Southern)to distinguish between wild-type,targeted,and recombined alleles in ES cells and mouse tissue.3)As mentioned above,conditional gene inactivation requires the gene of interest to function normally until excision by the specific recombinase.The presence of a selectable marker(for example a neo expression cas-sette)has been shown to influence the expression of the targeted gene when it is placed in the region5Јto the transcription start site(Rucker et al.,2000)or within an intron(Meyers et al.,1998;Nagy et al.,1998).These are probably due to the interference of promoter activity of the targeted gene or cryptic splice sites in the neo cassette,respectively.To eliminate this potential influ-ence on the expression of the target gene and maintain two recombinase target sites intact,the“three-loxP”strategy developed by Rajewsky and colleagues has been shown to be very successful(Gu et al.,1994).The targeting vector contains three loxP sites all in the same orientation,twoflanking the selectable marker and the third one positioned in one of the homology arms in a way that will cause the desired excision after recombi-nation by Cre(Fig.2B).After confirming successful tar-geting of the3-loxP construct into the gene locus,the selectable marker can be removed by transient expres-sion of Cre in correctly targeted ES cell clones transiently transfected with a Cre expression plasmid(Sauer and Henderson,1988;Gu et al.,1994).As all three loxP sites are in the same orientation in the targeted allele,there will be three possible recombination products due to partial excision by Cre.Isolation of the desired clone(s) (type II deletion)is not always straightforward because of the removal of the positive selection marker(Fig.2B). In most cases,the type I deletion is predominant(Fig. 2B)that can serve as a null allele control because it will be thefinal excised allele after conditional excision in vivo.One way to facilitate the selection of the type II deletion is toflox the positive and negative selection markers together(or as a fused gene)and use negative selection to enrich for the type II deletion(Fig.2C).In this strategy,it is still possible to perform the initial gene targeting using negative selection.However,the initial negative selection used for gene replacement screening must be different from the one used to select for the type II deletion(for example,first DT-A,then TK).Alterna-tively,afloxed hprt cassette can be used as both a positive and negative selection marker when hprt-defi-cient ES cells are used for the targeting experiment (Giovannini et al.,2000).The extra manipulation of the ES cells during this additional selection procedure can be a problem,as time and effort would be wasted if germline competency is compromised in thefinal correctly modified ES cell clones,as mentioned by Wiebel et al.(2002)in this issue of genesis.However,several approaches have been de-veloped to overcome this problem:a)a Cre-expression plasmid can be microinjected to the pronuclei of fertil-ized eggs from mice with the3-loxP allele to excise the floxed selection marker from the targeted allele(Xu et52KWANal.,2001);b)recombinant Cre adenovirus can be used to infect morulae stage embryos to remove the floxed marker from the targeted allele (Kaartinen and Nagy,2001);or c)mice containing the floxed marker can be crossed with a mild Cre deleter mouse line such as EIIa-Cre (Lakso et al.,1996;Xu et al.,2001).All of these approaches produce mosaic mice with partial and com-plete excision.The extent of mosaicism is probably floxed gene-and Cre transgene-dependent.The mosaic mice are then bred with wild-type mice and the offspring containing the desired conditional null allele can be identi fied.In the case of (c),this outcross also segregates the Cre transgene from the desired conditional and null alleles.Su et al.(this issue)have targeted a CMV-Cre transgene to the hprt locus on a 129inbred genetic background.Thus,this mild deleter mouse line can be used to modify a 3-loxP allele while maintaining an inbred genetic background.These approaches provide different options for removing the selection marker from the targeted allele in vivo and minimize the manipulation of ES cells in vitro.Another effective solution is to make use of the Flp-FRT system together with the Cre-loxP system.The pos-itive selectable marker is flanked with FRT sites instead of loxP sites while the target gene segment is floxed (Fig.2D).Flp-mediated excision of the positive selectable marker can be achieved either i)in vitro by transient expression of Flp in targeted ES cell clones (Buchholz,et al.,1998;Schaft et al.,2001)or ii)in vivo by breeding mice bearing the targeted allele with a Flp deleter trans-genic line (Farley et al.,2000;Rodriguez et al.,2000;Sun et al.,2000).The advantage of this approach is that it eliminates the stochastic aspects of the 3-loxP strategy and if option (ii)is used,minimizes the manipulation of ES cells to preserve germline competency of the targeted clones.4)Although the removal of the selection marker(s)from the conditional allele will prevent the undesired effect of those marker genes on the endogenous func-tion of the gene of interest,those conditional alleles harboring the selection marker may still be valuable.In some cases,those alleles containing the marker may be hypomorphic,in which their expression is reduced in comparison to wild-type (see Kulessa and Hogan,this issue).Such hypomorphic alleles have proven to be very useful tools for exploring the functions of the gene of interest (Meyers et al.,1998;Nagy et al.,1998;Rucker et al.,2000).Thus,in some cases it may be anadvantageFIG.3.Strategy for linking the con-ditional gene inactivation to re-porter gene activation.Gray rectan-gle,the exon of the gene of interest;SA,splice acceptor.53CONDITIONAL ALLELES IN MICEnot to remove the selection marker in the targeted ES cell clones but to use them to generate mice harboring the conditional allele with the selection marker.The selection marker can then be removed if desired by the in vivo approaches discussed above.5)In a typical conditional gene inactivation experi-ment,the design of the construct is aimed at generating a conditional null allele in the mice.However,one should consider that this same strategy can also be very useful to delete a speci fic functional domain or isoform of a protein in a tissue-speci fic manner without total knockout of gene function (Hajihosseini et al.,2001;Waltz et al.,2001).This provides an additional in vivo approach to dissect the functions of a protein in more detail.In this case,it is essential to document the speci fic deletion of the speci fic domain or isoform after excision by the recombinase using different expression assays.6)It is always important to verify the orientation and functionality of the loxP and/or FRT sites in the final gene targeting construct prior to ES cell transfection.This can be accomplished by DNA sequencing and trans-formation of the targeting construct into E.coli express-ing Cre or Flp for in vitro recombination (Buchholz et al.,1996).MOSAICISM IN CONDITIONAL GENETIC EXPERIMENTSA consistent and appropriately expressing Cre mouse strain is the other critical factor for a successful tissue-speci fic knockout experiment.Bear in mind that many of the tissue-speci fic Cre mouse lines will not be 100%ef ficient (Nagy,2000);partial deletion of the conditional allele is very likely to happen when these lines are used for conditional gene inactivation.This will create mosa-icism in which gene inactivation is induced in only a variable fraction of the desired cells,while the wild-type population of cells may potentially obscure and/or even mask the inactivation phenotype.Most likely,variable phenotypes will be observed which make the analysis and interpretation of the phenotypes potentially dif fi-cult.This mosaicism could become a prominent prob-lem if the target gene acts in a noncell-autonomous manner,in which a small fraction of wild-type cells expressing the gene product would fully rescue the mutant phenotype.This is especially true for secreted signaling molecules.Furthermore,the stability of the gene product (RNA and/or protein)may be another problem.The more stable the gene product meanstheFIG.4.Basic breeding scheme for conditional gene inactivation.Mice harboring the tissue-speci fic Cre transgene and heterozygous for the floxed and null alleles (Cre /ϩ;floxed/null )can be generated after two generations at a frequency of 25%,which will be used for phenotypic analysis.The floxed/floxed mice are usually maintained as homozygous female stock.ϩrepresents the wild-type allele.54KWANdelay between the onset of genetic ablation and the actual biological effect of the inactivation is longer.This may initially mask the inactivation phenotype.Thus,one should consider the nature of the gene of interest before applying conditional gene inactivation for the study of its functions.In order to deal with this problem of mosaicism,there is a need to trace and distinguish the two cell popula-tions (wild-type and mutant)at a cellular level after conditional inactivation.One way of achieving this is by using in situ hybridization and immunostaining to detect the speci fic mRNA transcript and protein,respectively,in those wild-type cells without gene inactivation.But this depends very much on the availability and speci fic-ity of RNA probes and antibodies.Indeed,if only a small portion of the coding region is conditionally deleted,it may not be suf ficiently large enough to provide a good in situ hybridization probe to detect the loss of mRNA expression.Thus,conditionally deleting a larger portion of the coding region provides a better opportunity for identifying a useful in situ hybridization probe,espe-cially if no antibody is available to the protein product.A more direct and reliable way is to modify the con-ditional targeting vector in which a reporter gene (lacZ,alkaline phosphatase,eGFP ,etc.)is inserted in such a way that activation of the reporter is linked to simulta-neous inactivation of the targeted gene (Brakebusch et al.,2000;Moon and Capecchi,2000;Moon et al.,2000).The reporter gene (usually with a splice acceptor at its 5Јend)is placed behind the most 3Јrecombinase target site (Fig.3).Thus,recombinase-mediated recombination of the reporter-containing conditional allele not only inactivates the gene in a tissue-speci fic manner,but also allows reporter activity to label those cells in which recombinase-mediated inactivation has occurred.An-other successful example using this approach is re-ported in this issue by Kulessa and Hogan (2002).Bythese means,one can assay the degree of gene inactiva-tion mosaicism (if any)in the tissue/lineage of interest that should help to ease the analysis and interpretation of mutant phenotype(s).On the other hand,there are still potential advantages of this cell/tissue-type restricted mosaicism due to in-complete conditional gene inactivation.It may serve as an informative tool to identify cell lineages for whose development a given targeted gene is critical (Betz et al.,1996).Moreover,mosaicism resulting from sporadic re-combination of conditional oncogenes and tumor sup-pressor genes by Cre is especially useful for mimicking certain human disease conditions with sporadic (in)acti-vation of those genes to establish better disease models (Shibata et al.,1997;Meuwissen et al.,2001).BREEDING SCHEMES FOR TISSUE-SPECIFIC KNOCKOUTSThe basic breeding scheme to achieve conditional gene inactivation in mice is shown in Figure 4.Basically,mice harboring the tissue-speci fically expressed Cre transgene and heterozygous for the floxed and null alleles (Cre/ϩ;floxed/null )can be generated after two generations at a 25%frequency.This frequency assumes that the floxed locus and the Cre transgene are not linked.The null allele can come from a previous knockout mouse or the deleted conditional allele (type I deletion in Fig.2B)as a by-product of the excision of the selection marker.In many cases,pregnant female mice from the second gen-eration (F2)are sacri ficed for developmental studies of the F3embryos having conditional gene inactivation.Thus,the (Cre/ϩ;null/ϩ)mice are usually male,whereas the homozygous (floxed/floxed )mice are usu-ally female.In this way,effort is saved for genotyping the homozygous (floxed/floxed )line that is going to be sac-ri ficed very frequently during the course of theexperi-FIG. 5.The number of condi-tional knockouts published each year since 1994.55CONDITIONAL ALLELES IN MICETable1Summary of the Current Literature on Conditional Alleles in MiceGene Genetic mapinformationChr(cM)Strategies forgenerating theconditionalallele ReferenceAP-2␥(or Tcfap2c)2(unknown)F Werling and Schorle,2002*Apc18(15.0)ŒShibata et al.,1997Arnt(aryl hydrocarbon receptor uncleartranslocator)3(47.9)F Tomita et al.,2000Bcl-x2(87.5)F Rucker et al.,2000;Wagner et al.,2000 Bdnf(brain-derived neurotrophic factor2(62.0)F Rios et al.,2001Bmp414(15)✚Kulessa and Hogan,2002*Bmpr1a(BMP receptor type IA)14(13.0)F Ahn et al.,2001;Mishina et al.,2002* Brca111(60.5)F Xu,et al.,1999Brca25(84.0)ŒLudwig et al.,2001F Jonkers et al.,2001Calb1(calbindin D-28K)4(10.5)F Barski et al.,2002*CamK4(Caϩϩ/calmodulin-dependent kinase4)18(12.0)✚Casanova et al.,2002*␤-catenin9(72.0)F Brault et al.,2001ŒHarada et al.,1999F Huelsken et al.,2001c/ebp␣(liver-enriched transcriptional regulator)7(12.0)F Lee et al.,1997Cx43(connexin43)10(29.0)F Liao et al.,2001Dnmt1(DNA methyltransferase1)9(5.0)F Lee et al.,2001;Jackson-Grusby et al.,2001;Fan et al.,2001ErbB2(erythroblastosis oncogene B2)11(57.0)F Garratt et al.,2000Foxa2(or Hnf3␤)2(84.0)F Sund et al.,2000and2001Fgf4(fibroblast growth factor4)7(72.4)✚Sun et al.,2000ŒMoon et al.,2000Fgf8(fibroblast growth factor8)19(45.0)✚Lewandoski et al.,2000✚Moon and Capecchi,2000Fgfr1(FGF receptor1)8(10.0)F Xu et al.,2000*Fgfr2-IIIc(FGF receptor2)7(63.0)F Hajihosseini et al.,2001Frda(Friedreich ataxia gene)unknown F Puccio et al.,2001Galnt6(polypeptide N-acetylgalactosaminyltransferase6)unknownŒHennet et al.,1995Gk(glucokinase)11(0.7)F Postic et al.,1999Glut4(glucose transporter4)11(40.0)ŒAbel et al.,1999;Zisman et al.,2000 Gp13013(unknown)F Betz et al.,1998;Hirota et al.,1999Gr11(glucocorticoid receptor)18(20.0)F Tronche et al.,1999H2(MHC-II beta chain gene)17(23.0)F Hashimoto et al.,2002*Hdh(Huntington disease gene)5(20.0)ŒDragatsis et al.,2000Hoxa26(26.29)✚Ren et al.,2002*Hnf4␣(hepatocyte nuclear factor4␣)2(94.0)F Hayhurst et al.,2001F Parviz et al.,2002*IL-2r␥(interlukin-2receptor␥)unknown F Disanto et al.,1995;Betz et al.,1996␤1-integrin unknownŒBrakebusch et al.,2000F Raghavan et al.,2000Igf1(insulin-like growth factor1)10(48.0)ŒLiu et al.,1998;Sjogren et al.,1999;Yakar et al.,1999Ir(insulin receptor)8(1.0)F Bruning et al.,1998;Kulkarni et al.,1999;Burning et al.,2000Kif3a(kinesin family member3a)11(29.5)F Marszalek et al.,2000Krox20(or Egr2)10(35.0)F Taillebourg et al.,2002*Lim1(or Lhx1)11(48.0)F Kwan and Behringer,2002*Lrp(low density lipoprotein receptor-related protein)15(unknown)Rohlmann et al.,1996and1998Mb-1(or H2-M1)117(20.23)Pelanda et al.,2002*Mdm210(66.0)ŒSteinman and Jones,2002*F Griver et al.,2002*Mecp2(methyl CpG binding protein2)X(29.6)F Chen et al.,2001Men1(multiple endocrine neoplasias type1)19(unknown)F Biondi et al.,2002*Mre119(unknown)F Xiao and Weaver,1997Mst1r(macrophage stimulating1receptor)9(60.0)F Waltz et al.,2001Mttp(microsomal triglyceride transfer protein)3(66.2)ŒChang et al.,1999F Raabe et al.,1999c-Myc15(32.0)F de Alboran et al.,2001Nf1(neurofibromatosis type1)11(46.1)ŒZhu et al.,2001Nf2(neurofibromatosis type2)11(0.25)F Giovannini et al.,2000Nmdar-1(N-methyl-D-aspartate receptor1)2(12.0)ŒHuerta et al.,2000;Iwasato et al.,2000 56KWAN。