TECHNOLOGY REPORT

Highly Efficient Transgene-Independent RecombinationDirected by a Maternally Derived Sox2Cre TransgeneStephane D. Vincent* and Elizabeth J. RobertsonDepartment of Molecular and Cellular Biology, Harvard University, Cambridge Massachusetts

Received 20 June 2003; Accepted 14 July 2003

Summary: The Cre/loxP site-specific recombination sys-tem is a powerful tool that allows gene inactivation in atissue- and time-specific manner. Several reports haveshown that the Sox2Cre transgenic strain provides avery efficient means to delete gene function from theearly epiblast (Hayashi et al.: Gene Expr Patterns 2:93–97, 2002; Vincent et al.: Genes Dev 17:1646–1662, 2003).Routinely, male studs carrying one null allele of the geneof interest and the Cre transgene are crossed to femaleshomozygous for the conditional allele. Normally, exci-sion is observed only in the progeny inheriting both theCre transgene and the conditional allele. Here we reportthat when the Sox2Cre transgene is inherited maternally,excision occurs in all offspring irrespective of whetherthey carry the Cre transgene. Thus, Sox2Cre femalesprovide a generally useful tool for rapid and efficientremoval of loxP flanked sequences in vivo. genesis 37:54–56, 2003. © 2003 Wiley-Liss, Inc.

Key words: mouse, Cre recombination, gene deletion, female

The Cre protein is a recombinase that specifically recog-nizes a 34-bp sequence called the loxP site. Expressionof Cre allows recombination between two loxP sites,leading to deletion of the intervening DNA sequences.This system has been adapted for use in the mouse andhas proven very useful for bypassing early requirementsfor gene function to allow study of later tissue-specificphenotypes (reviewed in Nagy, 2000).

The Sox2Cre transgenic strain was generated by in-jecting a 12.5-kb fragment containing the promoter andenhancer sequences of the Sox2 locus linked to the Crecoding sequence (Hayashi et al., 2002). These sequencesdirect Cre expression in the early epiblast, thereforeallowing efficient conditional deletion of loxP flankedalleles in the embryo proper (Hayashi et al., 2002; Vin-cent et al., 2003). This strain therefore provides a usefultool not only for studying gene function in the epiblast,but also for eliminating drug selection cassettes intro-duced during the process of generating conditional al-leles.

We used the Sox2Cre strain in order to remove aloxP-flanked hygromycin selection cassette from an en-gineered allele of the Blimp locus (Fig. 1a; unpubl. obs.).We observed differences in the efficiency of recombina-

tion, depending on the parental origin of the Sox2Cretransgene (see Table 1). As expected, when the Cretransgene is contributed by the male, excision is de-tected in progeny that carry both the mutant allele andthe Cre transgene. Surprisingly, when Sox2Cre femaleswere bred with males carrying a loxP flanked cassette(BAH/�), no hygro cassette was detected by PCR in anyof the resulting progeny (Fig. 1b). Interestingly, asshown in Figure 1b, PCR analysis shows that offspring 1,5, and 12 carry the recombined test allele as assessed bythe absence of hygro-specific product and the presenceof the diagnostic band corresponding to the rearrangedBA locus, but lack the Cre transgene. This finding washighly consistent, since the same result was observedafter testing in excess of 14 litters from this cross. Toconfirm that this was a generalized phenomenon, wealso tested a second independent allele containing a loxPflanked selection cassette and observed similar results(data not shown). Therefore, we conclude that when theSox2Cre transgene is of maternal origin, recombinationoccurs regardless of whether the resulting progeny in-herit the transgene.

There are two likely explanations for the mechanismof these parent-of-origin differences. The first is that thetransgene is expressed in the egg but its expression isinfluenced by parental imprinting. A more likely expla-nation is that Sox2Cre is expressed during differentiationof the female germline. Cre protein synthesized duringfolliculogenesis may be stored in the oocyte nucleus,thereby allowing recombination to occur after the fusionof the pronuclei in the fertilized egg. A delay betweenCre transgene expression and Cre-mediated recombina-tion has been observed in the male germline when DNAis too compacted, although Cre recombination can oc-cur in the male pronucleus (M. Rassoulzadegan and F.Cuzin, pers. commun.). In order to test this hypothesis,

* Correspondence to: Stephane D. Vincent, Department of Molecular andCellular Biology, Harvard University, 16 Divinity Avenue, Cambridge MA01238.

E-mail: vincent@fas.harvard.eduContract grant sponsors: the NIH, the HFSP (to SDV).

DOI: 10.1002/gene.10226

© 2003 Wiley-Liss, Inc. genesis 37:54–56 (2003)

we performed a Cre whole-mount in situ on adult ovariesfrom Sox2Cre females. As shown in Figure 2c–f, Cretranscripts are detected during all stages of oocyte dif-ferentiation, including in the most mature follicles (as-terisk in Fig. 2f). This result clearly indicates that thetransgene is expressed during the differentiation of thefemale germ cells, therefore permitting nuclear accumu-lation of Cre protein.

We note that a second Cre transgenic line, the CAG-cre line, (Sakai and Miyazaki, 1997) has similar charac-teristics to the Sox2Cre transgene and, in this case, Sakaiand Miyazaki have shown that recombination occursprior to the two-cell embryo stage, suggesting stronglythat the transgene is expressed in the developing oo-cytes, allowing accumulation of Cre protein in the ma-ture oocytes.

Our findings show that the Sox2Cre line provides arapid and efficient genetic tool to remove selection cas-settes or generate null configurations of conditional al-leles in vivo. It is thus very straightforward to obtainanimals carrying the excised allele but lacking the Cretransgene in one generation. This is especially usefulsince founder germline chimeric males can be bred di-

rectly to Sox2Cre female partners. Normally, using othermale germline deleter strains, such as the Prm-1 Cre line(O’Gorman et al., 1997), two generations of breedingwould be needed to generate the desired progeny.

MATERIALS AND METHODS

Mouse StrainsThe Sox2Cre mice were maintained on an outbred

CD1 background (Hayashi et al., 2002). The BAH allelewas generated using gene targeting (unpubl. data) toreplace 3 kb of genomic sequences by a loxP flanked pgkhygro cassette. The heterozygous BAH/� mice weremaintained on a (129 � C57BL/6) hybrid background.

Genotyping Procedures

Mice were genotyped by PCR screening of genomictail DNA. The Cre transgene was genotyped as describedpreviously (Vidal et al., 1998). The presence of the pgkhygro cassette was detected using the following prim-ers: hygro 5� 5�-CAGCGAGAGCCTGACCTATTGC-3� andhygro 3� 5�-CGATCCTGCAAGCTCCGGATG-3� that am-plify a 530-bp product.

FIG. 1. Transgene-independentCre recombination from the ma-ternally derived Sox2Cre trans-gene. a: Map of the test alleleused. The test allele (BAH) is theresult of a targeting event thatremoves 3 kb (EcoRI–PstI) ofgenomic sequences and inte-grates a loxP flanked pgk hygro-mycin selection cassette. The as-terisk indicates that the PstI siteis not the only one on the locus.b: PCR and Southern analysis ofone litter obtained from a crossbetween a BAH/� male and aSox2Cre/� female. As shown bythe BA PCR analysis and con-firmed by Southern analysis,progeny 1, 5, 7, 8, and 12 carriedthe recombined test allele; how-ever, only animals 7 and 8 addi-tionally carry the Cre transgene.

Table 1Maternal and Paternal Differences in Sox2Cre Transgene-mediated Recombination

Male Sox2Cre/� � female BAH/�

�/�;BA/� �/�;BAH/� Sox2Cre/�;BA/� Sox2Cre/�;BAH/� Sox2Cre/�;�/� �/�0 (0%) 9 (32.25%) 4 (14.25) 0 (0%) 7 (25%) 8 (28.5%)

2 litters

Male BAH/� � female Sox2Cre/�

�/�;BA/� �/�;BAH/� Sox2Cre/�;BA/� Sox2Cre/�;BAH/� Sox2Cre/�;�/� �/�32 (22.25%) 0 (0%) 38 (26.5%) 0 (0%) 34 (23.5%) 40 (27.75%)

14 litters

55EFFICIENT TRANSGENE-INDEPENDENT RECOMBINATION

The BAH and BA alleles were detected by the sameprimer pairs: mut�: 5�-TGTATGCCCCGTGTGTTTAG-3�and mut–: 5�-CGAAGTTATGGATCATCAAGG-3� amplify-ing a 260-bp fragment. The wildtype allele (WT) wasdetected using primers corresponding to sequences con-tained in the 3 kb deleted fragment: WT�: 5�-TTGGTA-AAGAAGTCTGCGGC-3� and WT–: 5�-TCGTACCCACAC-GTTTTTCC-3� amplifying a 310-bp fragment.

Southern Analysis

Five-�g samples of genomic tail DNAs were digestedwith 100 U of EcoRI overnight at 37°C. Fragments wereseparated and transferred according to standard proce-dures (Sambrook and Russell, 2001). Hybridization wasperformed using a 5� external probe that recognized a 9kb WT and 6.5 kb BAH or BA fragments.

Whole-Mount In Situ Hybridization (WISH)

Whole-mount in situ hybridization (WISH) was per-formed according to standard procedures (Hogan et al.,1994) using a full-length Cre cDNA probe (kindly pro-vided by Cindy Lu). To improve probe penetration, ova-ries were cut in halves before hybridization.

LITERATURE CITED

Hayashi S, Lewis P, Pevny L, McMahon AP. 2002. Efficient gene mod-ulation in mouse epiblast using a Sox2Cre transgenic mouse strain.Gene Expr Patterns 2:93–97.

Hogan BL, Beddington RS, Costantini F, Lacy E. 1994. Manipulating themouse embryo, a laboratory manual. Cold Spring Harbor, NY:Cold Spring Harbor Laboratory Press.

Nagy A. 2000. Cre recombinase: the universal reagent for genometailoring. genesis 26:99–109.

O’Gorman S, Dagenais NA, Qian M, Marchuk Y. 1997. Protamine-Crerecombinase transgenes efficiently recombine target sequences inthe male germ line of mice, but not in embryonic stem cells. ProcNatl Acad Sci USA 94:14602–14607.

Sakai K, Miyazaki J. 1997. A transgenic mouse line that retains Crerecombinase activity in mature oocytes irrespective of the cretransgene transmission. Biochem Biophys Res Commun 237:318–324.

Sambrook J, Russell DW. 2001. Molecular cloning: a laboratory manual,3rd ed. Cold Spring Harbor, NY: Cold Spring Harbor LaboratoryPress.

Vidal F, Sage J, Cuzin F, Rassoulzadegan M. 1998. Cre expression inprimary spermatocytes: a tool for genetic engineering of the germline. Mol Reprod Dev 51:274–280.

Vincent SD, Dunn NR, Hayashi S, Norris DP, Robertson EJ. 2003. Cellfate decisions within the mouse organizer are governed by gradedNodal signals. Genes Dev 17:1646–1662.

FIG. 2. Cre expression during fol-liculogenesis in Sox2Cre females.Whole-mount in situ hybridizationof adult ovaries from wildtype (a)and Sox2Cre (c,e) females usinga Cre riboprobe. a: No signal isdetected in the oocyte nuclei (ar-rows) from the control ovary. b:Higher-power view of ovary nuclei(a). c,e: The nuclei of oocytes atdifferent stages of folliculogen-esis express Cre. As shown athigher magnification (d,f), Cre ex-pression can be detected fromthe early stage (primary follicle) tothe latest stage (antral follicle, in-dicated by asterisk in f).

56 VINCENT AND ROBERTSON