Abstract

Imprinted genes are commonly clustered in domains across the mammalian genome, suggesting a degree of coregulation via long-range coordination of their monoallelic transcription. The distal end of mouse chromosome 7 (Chr 7) contains two clusters of imprinted genes within a ∼1 Mb domain. This region is conserved on human 11p15.5 where it is implicated in the Beckwith–Wiedemann syndrome. In both species, imprinted regulation requires two critical cis-acting imprinting centres, carrying different germline epigenetic marks and mediating imprinted expression in the proximal and distal sub-domains. The clusters are separated by a region containing the gene for tyrosine hydroxylase (Th) as well as a high density of short repeats and retrotransposons in the mouse. We have used the Cre-loxP recombination system in vivo to engineer an interstitial deletion of this ∼280-kb intervening region previously proposed to participate in the imprinting mechanism or to act as a boundary between the two sub-domains. The deletion allele, Del7AI, is silent with respect to epigenetic marking at the two flanking imprinting centres. Reciprocal inheritance of Del7AI demonstrates that the deleted region, which represents more than a quarter of the previously defined imprinted domain, is associated with intrauterine growth restriction in maternal heterozygotes. In homozygotes, the deficiency behaves as a Th null allele and can be rescued pharmacologically by bypassing the metabolic requirement for TH in utero. Our results show that the deleted interval is not required for normal imprinting on distal Chr 7 and uncover a new imprinted growth phenotype.

INTRODUCTION

The distal end of mouse chromosome 7 (Chr 7) contains a large cluster of imprinted genes, located in cytogenetic band 7F5. This domain defines a ∼1 Mb region rich in imprinted transcripts (1). The region shares syntenic homology with human 11p15, where abnormal regulation of imprinting is associated with the Beckwith–Wiedemann syndrome (BWS, OMIM 130650), a congenital overgrowth syndrome (2,3). Comparative analyses between mouse and human suggest a high degree of conservation, not only of the genomic organization and sequences of the orthologous regions, but also of the imprinted expression of several key genes within these clusters (4–6). Genetic studies on distal Chr 7 have therefore provided important models for BWS, while contributing to our knowledge of the general mechanism and function of imprinting in mammals.

Our current understanding of the regulation of imprinting on distal Chr 7 suggests that two imprinting centres (ICs) are responsible for the reciprocal imprinting of two sub-domains within this cluster. In the proximal sub-domain, a CpG-rich sequence located ∼2 kb upstream of the H19 gene acquires a DNA methylation imprint in sperm but not in eggs (7–9). This sequence termed H19 DMR or IC1 acts as an IC (10) and regulates monoallelic expression of nearby genes via two different mechanisms. DNA methylation of the paternal IC1 leads to silencing of H19 in cis as well as paternal allele-specific expression of at least two protein-coding genes, Igf2 and Ins2, via a CTCF-dependent insulator or boundary element acting exclusively on the unmethylated maternal allele (11–14).

The telomeric (distal) sub-domain is also marked by a differentially methylated CpG-rich region, but here the germline imprint is exclusively inherited from the oocyte. Located within intron 10 of Kcnq1, this sequence termed KvDMR1 or IC2 acts as a promoter for the large non-coding RNA Kcnq1ot1 (15,16). Expression of Kcnq1ot1 exclusively from the paternal homolog has been linked to the silencing of several protein-coding genes in cis, such as Cdkn1c, Phlda2, and Ascl2 (15–17). These genes are consequently imprinted and expressed preferentially from the maternal homolog. Imprinted monoallelic expression in this distal sub-domain is therefore thought to be under the regulation of Kcnq1ot1 transcription and/or the transcript.

Whereas several studies have focused on the function of IC1 and IC2, their independent imprinting at ectopic genomic locations, and the transcripts they each regulate, very little is known about possible interactions between the two imprinted sub-domains in the context of distal Chr 7 (10,17–19). The linkage between these two ICs is conserved in eutherians and provides a model to study long-range interactions between centres as well as to address the possible presence of boundaries or insulators defining the range of action of each centre. In the mouse, the interval between Ins2 (20), located at the distal end of the known IC1-regulated domain, and Ascl2, the most proximal transcript known to be regulated by IC2 (17,21), represents close to a third of the entire distal Chr 7 domain. This region contains a single known protein-coding gene, Th, which codes for tyrosine hydroxylase, the rate-limiting enzyme in the synthesis of catecholamines.

We have used an in vivo Cre-loxP approach to engineer a ∼280-kb deletion of the IC1–IC2 intervening region. Proximal and distal breakpoints of this deletion were defined by targeted loxP site insertions next to the Ins2 and Ascl2 loci. From Cre-mediated trans-allelic recombination in spermatocytes we recovered the reciprocal deletion (Del7AI) and duplication (Dp7AI) alleles. We show here that the initiation and maintenance of epigenetic marking at IC1 and IC2 by DNA methylation are not affected by this deletion. Both maternal and paternal deletion heterozygotes are viable but maternal heterozygotes exhibit an intrauterine growth retardation (IUGR) phenotype. We also report the post-implantation lethality of Del7AI/Del7AI homozygotes and the developmental rescue of this phenotype by provision of exogenous L-DOPA in utero during gestation. These results provide genetic evidence for the absence of new developmentally required genes in this large region of Chr 7 and suggest that deletion homozygotes phenocopy a Th deficiency with respect to developmental survival.

RESULTS

Unique features of the Ins2–Ascl2 interval

The paternally expressed imprinted genes Igf2 and Ins2 belong to an ancient chromosomal paralogy group linking genes for insulin-related proteins (IGF) to members of the aromatic amino acid hydroxylase family (AAAH) in vertebrates (22). This group extends to the genes for the achaete-scute family of transcription factors (ASCL) in several vertebrate genomes. In the mouse genome, Igf2-Ins2-Th-Ascl2 maps to the distal end of Chr 7, whereas the paralogs Igf1-Pah-Ascl1 are present on Chr 10. Members of the first group are regulated by genomic imprinting, but this is not the case for the paralogs on Chr 10. The same regions are present on human chromosomes 11 and 12, respectively (23). Although both IGF2 and INS are imprinted in human (24–26), there is still no definitive evidence for the imprinting of TH or ASCL2, which are both maternally expressed in the mouse placenta (21,27).

We have defined two intervening regions or gaps in these paralogy regions, gap 1 (IGF-AAAH) and gap 2 (AAAH-ASCL) and analyzed their structure and sequence composition in seven vertebrate genomes (Supplementary Material, Fig. S1A). Our analysis shows that the greatest variation in gap sizes is observed in gap 1 of the mouse Igf2-Ins2 to Th region (Supplementary Material, Fig. S1B). Most of this size increment is attributable to a significant increase in the content of repetitive elements, notably LINEs and LTR elements (Supplementary Material, Fig. S1C). Compared with the gap 1 of the IGF1-PAH paralogy region, there is also more variation in the GC content of the IGF2-INS-TH region between non-imprinting (chicken, zebrafish) and imprinting mammalian species (mouse, human, dog, cow) (Supplementary Material, Fig. S1C). Other unique features of the mouse Igf2-Ins2-Th-Ascl2 region which have been previously described include the presence of a large block of tandem repeats of a 30-bp sequence (28), as well as the identification by circular chromosome conformation capture of five sequences in gap 1 physically interacting with the unmethylated maternal IC1 (asterisks in Fig. 1A) (29). To address the function of these features in the regulation of imprinting on distal Chr 7, we have engineered a deletion of the Ins2 to Ascl2 interval.

Figure 1.

Targeted lox P site insertions on distal Chr7. (A) Gene arrangement on distal Chr 7 shown to scale (coordinates from NCBI m37 build), oriented from the telomere (Tel7q, left) to the centromere (Cen, right). Genes showing biallelic, paternal and maternal expressions are represented by shaded, black and white rectangles, respectively, with arrows indicating their transcriptional orientation. The map also presents the positions of the following features: (above) the imprinting centres IC1 and IC2, the sites physically interacting with the maternal IC1 (asterisks) and the breakpoints of the deletion generated in this study (Del7AI); (below) three genomic PAC clones (Phlda2, Th, and H19) used in DNA FISH experiments. (B) Targeting of a loxP site at Ins2: the I2 and I2loxP alleles. The generation of these alleles has been described previously (30). White and black rectangles show the Ascl2 and Ins2 exons, respectively. The black circle on the right represents the centromere. In the I2 allele, the PGK promoter can be deleted by Cre. The remaining promoter-less loxP-neo-pA of the I2loxP allele can be activated by recombination with the PGK-loxP of the M2loxP allele, shown in C. The genomic probe In1 was used to analyze the duplication allele (Supplementary Material, Fig. S3); In2 was used to confirm the structure of the deletion allele (Fig. 3A). (C) Targeting of a loxP site at Ascl2: the M2 and M2loxP alleles. Diagrams centred on the Ascl2-Ins2 interval and showing, from the top, the wild-type locus (wt), the targeting vector, the targeted M2 allele and the M2loxP insertion. The selectable marker cassettes consist of a PGK promoter (white box), driving neo expression from a neo-pA construct (black box and open oval). In the M2 allele the neo-pA is flanked by loxP sites (black arrowheads). A1: external 5′ probe; A2, A3: 3′ and 5′ internal genomic probes. E, EcoRI; N, NcoI. Southern blot analysis of genomic DNA from ES cell clones, digested with EcoRI and hybridized with A1 and A2 (bottom left and middle). The M2loxP allele was obtained by transient Cre expression in Ascl2+/M2 cells and analyzed with probe A2 (bottom right). Samples include the parental R1 ES cell line (+/+), 4 positive Ascl2+/M2loxP and one un-recombined Ascl2+/M2 clones. PGK2 and MR1 on M2 allele: PCR primers used to genotype the M2 and M2loxP alleles (Supplementary Material, Fig. S2B). The A3 probe was used to monitor rearrangements at the M2 allele in vivo (Supplementary Material, Fig. S2A), and to confirm the structure of the deletion allele (Fig. 3A).

Figure 1.

Targeted lox P site insertions on distal Chr7. (A) Gene arrangement on distal Chr 7 shown to scale (coordinates from NCBI m37 build), oriented from the telomere (Tel7q, left) to the centromere (Cen, right). Genes showing biallelic, paternal and maternal expressions are represented by shaded, black and white rectangles, respectively, with arrows indicating their transcriptional orientation. The map also presents the positions of the following features: (above) the imprinting centres IC1 and IC2, the sites physically interacting with the maternal IC1 (asterisks) and the breakpoints of the deletion generated in this study (Del7AI); (below) three genomic PAC clones (Phlda2, Th, and H19) used in DNA FISH experiments. (B) Targeting of a loxP site at Ins2: the I2 and I2loxP alleles. The generation of these alleles has been described previously (30). White and black rectangles show the Ascl2 and Ins2 exons, respectively. The black circle on the right represents the centromere. In the I2 allele, the PGK promoter can be deleted by Cre. The remaining promoter-less loxP-neo-pA of the I2loxP allele can be activated by recombination with the PGK-loxP of the M2loxP allele, shown in C. The genomic probe In1 was used to analyze the duplication allele (Supplementary Material, Fig. S3); In2 was used to confirm the structure of the deletion allele (Fig. 3A). (C) Targeting of a loxP site at Ascl2: the M2 and M2loxP alleles. Diagrams centred on the Ascl2-Ins2 interval and showing, from the top, the wild-type locus (wt), the targeting vector, the targeted M2 allele and the M2loxP insertion. The selectable marker cassettes consist of a PGK promoter (white box), driving neo expression from a neo-pA construct (black box and open oval). In the M2 allele the neo-pA is flanked by loxP sites (black arrowheads). A1: external 5′ probe; A2, A3: 3′ and 5′ internal genomic probes. E, EcoRI; N, NcoI. Southern blot analysis of genomic DNA from ES cell clones, digested with EcoRI and hybridized with A1 and A2 (bottom left and middle). The M2loxP allele was obtained by transient Cre expression in Ascl2+/M2 cells and analyzed with probe A2 (bottom right). Samples include the parental R1 ES cell line (+/+), 4 positive Ascl2+/M2loxP and one un-recombined Ascl2+/M2 clones. PGK2 and MR1 on M2 allele: PCR primers used to genotype the M2 and M2loxP alleles (Supplementary Material, Fig. S2B). The A3 probe was used to monitor rearrangements at the M2 allele in vivo (Supplementary Material, Fig. S2A), and to confirm the structure of the deletion allele (Fig. 3A).

New deletion and duplication alleles on distal Chr 7

We have used the Cre-loxP site-specific recombination system to introduce a chromosomal deletion removing the entire region located in between the sub-domains regulated by the ICs IC1 and IC2. The deleted interval of ∼280-kb is flanked by the Ascl2 and Ins2 loci, respectively (Fig. 1A). For this purpose we designed bipartite PGK-neo-pA cassettes in which loxP sites flank only the PGK promoter or only the neo-pA element, and generated targeting vectors carrying these complementary selectable markers (30). Gene targeting in R1 ES cells (31) was used to introduce loxP-carrying insertional mutations at the Ascl2 and Ins2 loci (Fig. 1B and C). The proximal breakpoint is defined by a loxP site inserted 2.56 kb telomeric of Ins2 (Ins2tm1Nagy allele, referred to below as the I2 allele; Fig. 1B). We previously described the construction and targeting of this allele at the Ins2 locus (30). For the distal breakpoint of the deletion we targeted a loxP site 4.34 kb centromeric to Ascl2 (Ascl2tm2Nagy allele, referred to below as the M2 allele; Fig. 1C). The conditionality of the neo cassette was demonstrated by transient expression of Cre recombinase in these cell lines and the characterization of G418-sensitive Ascl2+/M2loxP cell line variants (Ascl2tm2.1Nagy allele) with a simple PGK-loxP insertion (Fig. 1C). We generated chimeras with ES cell clones carrying the individual M2 and I2 insertion alleles and obtained germline transmission for both of these alleles. In each case, the heterozygous and homozygous mice are viable and fertile, with no obvious phenotype, suggesting that the M2 and I2 alleles constitute phenotypically silent insertions.

The availability of Ascl2M2/M2 and Ins2I2/I2 homozygous mouse lines offered the possibility of generating the desired deletion in vivo, using a Cre trans-recombination in spermatocytes. This strategy, referred to as TArgeted MEiotic REcombination (TAMERE), was used previously to generate reciprocal deletion and duplication alleles within the Hoxd locus (32,33). It is based on the use of the Sycp1-Cre transgenic mouse line expressing the Cre recombinase in primary spermatocytes, driven by the promoter for the synaptonemal complexe protein 1 gene, Sycp1 (34,35). First, we generated Ascl2M2/+Sycp1-Cre double transgenic males by crossing Ascl2M2/M2 females with males hemizygous for the Sycp1-Cre transgene. As previously reported (35), we observed ubiquitous somatic excision of the M2 allele to form M2loxP in some of the progeny (1 in 11 males analyzed, Supplementary Material, Fig. S2A). To confirm that the double transgenic males selected for the next cross carry a functionally active Cre transgene, we only bred the Sycp1-Cre positive males showing no sign of somatic excision of the M2 allele.

Second, we crossed Ins2I2/I2 homozygous females to one of the Ascl2M2/+Sycp1-Cre double transgenic male. For this cross, we observed Cre-mediated conversion of M2 to M2loxP in 10/11 progeny heterozygous for the loxP site insertion at Ascl2, confirming efficient germline activity of the Cre transgene in this particular male (Supplementary Material, Fig. S2B). Three of these mice were confirmed triple transgenic males with the Ascl2+/M2loxPIns2I2/+Sycp1-Cre genotype. In these males, the loxP sites are therefore present in a trans-configuration, one on each parental Chr 7 homolog.

In the final crosses, wild-type outbred females were mated with males trans-heterozygous for the loxP site insertions and carrying the Sycp1-Cre transgene (Fig. 2A). Recombination between these non-allelic loxP sites, located ∼280 kb apart, is expected to generate two reciprocal recombination products: a deletion (Del7AI) and duplication (Dp7AI) of the intervening sequences (Fig. 2B). From these crosses, viable mice heterozygous for the deletion or duplication alleles were recovered at weaning, as initially confirmed by allele-specific PCR assays (Fig. 2C). In total, we analyzed six litters from two transgenic males. From a total of 84 progeny genotyped at weaning, we obtained 13+/Del7AI and 16+/Dp7AI paternal heterozygous pups (Supplementary Material, Table S1, cross 1).

Figure 2.

TAMERE from a triple transgenic male: the Del7AI and Dp7AI alleles. (A) Breeding scheme to produce novel recombined alleles from wild-type females crossed with triple transgenic males, carrying the Sycp1-Cre transgene as well as loxP site insertions at the maternal Ins2 and paternal Ascl2 loci. (B) Diagrams of the Ascl2–Ins2 interval presented as in Figure 1 and showing the positions of the genotyping PCR primers F1, F2, R1 and R2. From the top, these show the structures of the loxP insertion alleles at Ascl2 (M2loxP) and Ins2 (I2loxP) present in the male, as well as the two predicted reciprocal recombination products from a trans-allelic Cre reaction: the deletion allele Del7AI and the duplication allele Dp7AI. (C) Representative PCR genotyping of 14 progeny, including 5+/Del7AI heterozygotes (#3, 4, 11–13) and 6+/Dp7AI heterozygotes (#2, 5–8, 14). The new alleles were detected using specific PCR assays against novel unique junctions. Controls are: A, wild-type DNA; B, ES cells heterozygous for the deletion allele.

Figure 2.

TAMERE from a triple transgenic male: the Del7AI and Dp7AI alleles. (A) Breeding scheme to produce novel recombined alleles from wild-type females crossed with triple transgenic males, carrying the Sycp1-Cre transgene as well as loxP site insertions at the maternal Ins2 and paternal Ascl2 loci. (B) Diagrams of the Ascl2–Ins2 interval presented as in Figure 1 and showing the positions of the genotyping PCR primers F1, F2, R1 and R2. From the top, these show the structures of the loxP insertion alleles at Ascl2 (M2loxP) and Ins2 (I2loxP) present in the male, as well as the two predicted reciprocal recombination products from a trans-allelic Cre reaction: the deletion allele Del7AI and the duplication allele Dp7AI. (C) Representative PCR genotyping of 14 progeny, including 5+/Del7AI heterozygotes (#3, 4, 11–13) and 6+/Dp7AI heterozygotes (#2, 5–8, 14). The new alleles were detected using specific PCR assays against novel unique junctions. Controls are: A, wild-type DNA; B, ES cells heterozygous for the deletion allele.

The structure of the Dp7AI duplication allele was confirmed by Southern blot analysis of genomic DNA samples isolated from heterozygous progeny and wild-type litter mates. In all heterozygous +/Dp7AI samples analyzed, Ascl2 and Ins2 genomic probes internal to the loxP-flanked region hybridized to a novel, common, mutant restriction fragment (Supplementary Material, Fig. S3A and B). This is consistent with the repetition of Ascl2 3′-flanking sequences next to the Ins2 loxP site. We also amplified and sequenced this newly formed junction between the 5′ end of Ins2 and the 3′ end of Ascl2. The DNA sequence obtained confirmed the presence of a single junctional loxP site, flanked by the expected linker sequences from the targeting vectors and the genomic sequences of the appropriate arms of homology (Supplementary Material, Fig. S3C).

All the +/Dp7AI mice obtained from TAMERE are phenotypically indistinguishable from their wild-type litter mates and the same is true upon maternal and paternal transmission of the allele from +/Dp7AI mice (L.L., unpublished data). The duplication allele has not been characterized further for the present study.

The Del7AI allele brings the Ascl2 and Ins2 loci in close proximity

We analyzed the structure of the Del7AI deletion allele by Southern blot analysis, using genomic probes from the Ascl2 and Ins2 loci, flanking and external to the deletion breakpoints. All the restriction enzyme digests tested present a new fragment, unique to the Del7AI allele and shared by the two probes (Fig. 3A). Further molecular evidence for the deletion of Chr 7 sequences was provided by confirming the hemizygosity of a polymorphic marker internal to the deleted region in +/Del7AI heterozygous embryos at E13.5 (Supplementary Material, Fig. S4). Some of these E13.5 embryos were also used to culture primary embryonic fibroblasts, which were grown in absence or presence of G418. All the G418-resistant fibroblast cultures were from +/Del7AI embryos, demonstrating that a functional PGK-loxP-neo-pA marker had been re-formed at the recombination junction (Supplementary Material, Fig. S5). These fibroblasts also provided a source of metaphase spreads for DNA-FISH experiments using as probes genomic PAC clones from regions proximal (H19), internal (Th) or distal (Phlda2) to the deletion (Fig. 1A for probe positions). Whereas the internal Th probe gave a signal on both Chr 7 homologs in wild-type cells (Fig. 3B, right panel), only one of the two H19-positive Chr 7 homologs also gave a signal for the internal Th probe in Del7AI heterozygous cells (Fig. 3B, left panel). The integrity of the Chr 7 region telomeric to the Ascl2 loxP site in the Del7AI allele was also confirmed with the distal Phlda2 probe. Here again, the flanking probe gave a signal on both Chr 7 homologs, only one of which also carries the Th locus, identifying it as the wild-type Chr 7 (Fig. 3B, middle panel). Together our results confirm the structure of the deletion allele which brings Ins2 and Ascl2 in close proximity.

Figure 3.

Molecular and cytogenetic characterization of the deletion allele Del7AI. (A) Southern blot analysis of tail genomic DNA, isolated from a +/Del7AI heterozygote (#1) and a wild-type littermate (#2), digested with different restriction enzymes, and probed with the Ascl2 and Ins2 genomic probes A3 and In2. The diagrams of the wild-type (wt) and mutant (Del7AI) alleles (left) present the positions of the probes A3 and In2 relative to the loxP site insertion sites. Representative restriction enzyme recognition sites (e) illustrate the presence of a new fragment shared by the two probes in the DNA sample from the +/Del7AI heterozygote (#1). (B) DNA FISH on interphase nuclei and metaphase chromosome spreads from wild-type or +/Del7AI primary embryonic fibroblasts, using genomic PAC clones from Chr 7 regions proximal (H19, green), internal (Th, red), or distal (Phlda2, green) to the Del7AI breakpoints as probes (Fig. 1A). Arrowheads point to Chr 7 variants negative for the Th probe and thus carrying the Del7AI allele.

Figure 3.

Molecular and cytogenetic characterization of the deletion allele Del7AI. (A) Southern blot analysis of tail genomic DNA, isolated from a +/Del7AI heterozygote (#1) and a wild-type littermate (#2), digested with different restriction enzymes, and probed with the Ascl2 and Ins2 genomic probes A3 and In2. The diagrams of the wild-type (wt) and mutant (Del7AI) alleles (left) present the positions of the probes A3 and In2 relative to the loxP site insertion sites. Representative restriction enzyme recognition sites (e) illustrate the presence of a new fragment shared by the two probes in the DNA sample from the +/Del7AI heterozygote (#1). (B) DNA FISH on interphase nuclei and metaphase chromosome spreads from wild-type or +/Del7AI primary embryonic fibroblasts, using genomic PAC clones from Chr 7 regions proximal (H19, green), internal (Th, red), or distal (Phlda2, green) to the Del7AI breakpoints as probes (Fig. 1A). Arrowheads point to Chr 7 variants negative for the Th probe and thus carrying the Del7AI allele.

Growth retardation in maternal Del7AI heterozygotes

Although the heterozygous mutant progeny obtained by TAMERE from triple-transgenic males were viable and fertile, the inherited deletion allele was generated de novo in primary spermatocytes and did not go through reprogramming in the male germline. Therefore, we first assessed the phenotype associated with the deletion Del7AI through reciprocal crosses between wild-type mice and the +/Del7AI progeny obtained by TAMERE. The distal Chr 7 imprinted domain contains several developmentally important genes, including Igf2 (36,37), Cdkn1c (38–40), Phlda2 (41) and Ascl2 (21,42). Consequently, disruption of epigenetics regulation at IC1 and/or IC2 is expected to yield distinct and characteristic embryonic or early postnatal phenotypes. Paternal +/Del7AI and maternal Del7AI/+ heterozygotes were recovered at the expected frequency at weaning from the carrier parents (Supplementary Material, Table S1, crosses 2 and 3). No abnormalities or growth differences were observed in the paternal heterozygotes (Fig. 4A and B), but maternal transmission of the deletion allele caused intrauterine growth restriction in mutant progeny, which were almost 20% smaller than their wild-type littermate at birth on average (Fig. 4A). This growth difference is maintained to weaning, with no sign of catch up by the mutants (Fig. 4C). Both Del7AI/+ and +/Del7AI heterozygotes grow to become mature and fertile adults.

Figure 4.

Intrauterine growth restriction in maternal Del7AI heterozygotes. Newborns from five litters obtained by crossing +/Del7AI heterozygous females to wild-type CD-1 outbred males and from six reciprocal crosses were weighted the morning of their delivery (p0) and genotyped. (A) Scatter plot showing the distribution of the weights at birth. The average weight of maternal heterozygous pups (1.42 g) is 17.4% smaller than that of their wild-type litter mates (1.72 g). Between the maternal heterozygotes and wild-type pups, P = 1.82E-8 (t-test). Number of animals analyzed: n = 58 wild-type, n = 18+/Del7AI, n = 34 Del7AI/+ and n = 18+/Del7AI treated with L-DOPA. (BD) Body weight of mutant and wild-type litter mates at p0, p7 and p21 (time of weaning). Unlike the paternal heterozygotes (+/Del7AI), maternal heterozygous pups (Del7AI/+) do not catch up to their wild-type litter mates before weaning. This growth retardation is not rescue by addition of L-DOPA to the drinking water of pregnant females. Bars show the average weight ± standard deviation. At each time point, P < 0.001 (t-test) between maternal heterozygotes and wild-types in C. All values and numbers of animals analyzed are given in Supplementary Material, Table S2.

Figure 4.

Intrauterine growth restriction in maternal Del7AI heterozygotes. Newborns from five litters obtained by crossing +/Del7AI heterozygous females to wild-type CD-1 outbred males and from six reciprocal crosses were weighted the morning of their delivery (p0) and genotyped. (A) Scatter plot showing the distribution of the weights at birth. The average weight of maternal heterozygous pups (1.42 g) is 17.4% smaller than that of their wild-type litter mates (1.72 g). Between the maternal heterozygotes and wild-type pups, P = 1.82E-8 (t-test). Number of animals analyzed: n = 58 wild-type, n = 18+/Del7AI, n = 34 Del7AI/+ and n = 18+/Del7AI treated with L-DOPA. (BD) Body weight of mutant and wild-type litter mates at p0, p7 and p21 (time of weaning). Unlike the paternal heterozygotes (+/Del7AI), maternal heterozygous pups (Del7AI/+) do not catch up to their wild-type litter mates before weaning. This growth retardation is not rescue by addition of L-DOPA to the drinking water of pregnant females. Bars show the average weight ± standard deviation. At each time point, P < 0.001 (t-test) between maternal heterozygotes and wild-types in C. All values and numbers of animals analyzed are given in Supplementary Material, Table S2.

We then assessed the compatibility of the deletion allele with erasure and re-establishment of parental imprints by taking the mutation thought the opposite germline. This was achieved by crossing paternal heterozygous +/Del7AI females or maternal heterozygous Del7AI/+ males with wild-type mice. Here again, deletion heterozygotes were recovered at the expected Mendelian ratio with no progeny loss in utero (Supplementary Material, Table S1, crosses 4 and 5).

Normal imprinting at IC1 and IC2 in deletion heterozygotes

Our breeding data suggested that the deletion of the large region flanked by the IC1- and IC2-regulated imprinted sub-domains is silent with regard to epigenetic regulation on distal Chr 7. To assess the effect of Del7AI on the imprinting at IC1 and IC2 we looked at imprinted expression of the non-coding RNAs and DNA methylation marks at each imprinting centres in wild-type and reciprocal heterozygous Del7AI embryos (Figs 5 and 6). For allele-specific studies, we crossed Del7AI heterozygotes with mice from a congenic mouse line, derived in our laboratory, and carrying the Mus musculus castaneus variant (denoted C) of distal Chr 7 on the 129S1 background. From these crosses, we obtained reciprocal wild-type (C/+ and +/C) and Del7AI heterozygous (C/Δ and Δ/C) embryos with expressed polymorphisms at H19 and Kcnq1ot1. The allele-specific RT–PCR analyzes confirmed exclusive maternal expression of H19 (Fig. 5B) and paternal expression of Kcnq1ot1 (Fig. 6B). Partial DNA methylation at the ICs was first demonstrated by Southern blot analysis of embryonic genomic DNA (Figs 5C and 6C) and the parent-of-origin imprints at IC1 and IC2 were confirmed in reciprocal Del7AI heterozygotes by bisulfite conversion and sequencing, using SNPs between castaneus and domesticus variants to assign parental origin of each sequenced strand (Figs 5D and 6D). Our results therefore show that both paternal and maternal primary epigenotypes are unperturbed by the deletion allele and that the deleted region is dispensable for imprinting at IC1 and IC2.

Figure 5.

Maintenance of imprinting at IC1 in reciprocal Del7AI heterozygotes. Expression and methylation imprinting at IC1 were analyzed in reciprocal heterozygotes for the Del7AI allele. E14.5 embryos were recovered from reciprocal crosses between +/Del7AI heterozygotes and M. m. castaneus (CAST/EiJ) adult mice. (A) Diagram of the IC1-H19 region showing the region amplified by RT–PCR to analyze H19 expression in (B) (RT), the position of the genomic probe used in (C) (probe) and the part of IC1 analyzed in (D) (BS). (B) Allele-specific RT–PCR analysis of embryonic H19 expression. Maternal (mat) and paternal (pat) alleles are as follows: C, wild-type castaneus; +, wild-type domesticus; and Δ, Del7AI. A Cac8I RFLP was used to distinguish the castaneus (cast) from the domesticus (dom) variants. (C) DNA methylation analysis by Southern blotting of genomic DNA digested with EcoRI (E) and the methylation sensitive restriction enzyme HpaII (Hp). Maternal (mat) and paternal (pat) alleles are identified as wild-type (+) or Del7AI (Δ). Also shown are the positions of methylated EcoRI fragment (me+) and the main unmethylated (me−) bands. (D) Allele-specific DNA methylation analysis by sodium bisulfite sequencing. The upstream region of the IC1 DMR was analyzed in reciprocal Del7AI heterozygotes carrying SNPs allowing the identification of the parental alleles. Methylated CpG sites are shown in black, unmethylated sites in white.

Figure 5.

Maintenance of imprinting at IC1 in reciprocal Del7AI heterozygotes. Expression and methylation imprinting at IC1 were analyzed in reciprocal heterozygotes for the Del7AI allele. E14.5 embryos were recovered from reciprocal crosses between +/Del7AI heterozygotes and M. m. castaneus (CAST/EiJ) adult mice. (A) Diagram of the IC1-H19 region showing the region amplified by RT–PCR to analyze H19 expression in (B) (RT), the position of the genomic probe used in (C) (probe) and the part of IC1 analyzed in (D) (BS). (B) Allele-specific RT–PCR analysis of embryonic H19 expression. Maternal (mat) and paternal (pat) alleles are as follows: C, wild-type castaneus; +, wild-type domesticus; and Δ, Del7AI. A Cac8I RFLP was used to distinguish the castaneus (cast) from the domesticus (dom) variants. (C) DNA methylation analysis by Southern blotting of genomic DNA digested with EcoRI (E) and the methylation sensitive restriction enzyme HpaII (Hp). Maternal (mat) and paternal (pat) alleles are identified as wild-type (+) or Del7AI (Δ). Also shown are the positions of methylated EcoRI fragment (me+) and the main unmethylated (me−) bands. (D) Allele-specific DNA methylation analysis by sodium bisulfite sequencing. The upstream region of the IC1 DMR was analyzed in reciprocal Del7AI heterozygotes carrying SNPs allowing the identification of the parental alleles. Methylated CpG sites are shown in black, unmethylated sites in white.

Figure 6.

Imprinting at IC2 in reciprocal Del7AI heterozygotes. Expression and methylation imprinting at IC2 were analyzed in reciprocal heterozygotes for the Del7AI allele. (A) Diagram of the IC2 region, located within the intron 10 of Kcnq1, showing the region amplified by RT–PCR to analyze Kcnq1ot1 expression in (B) (RT), the position of the genomic probe used in (C) (probe) and the 335-bp sequence containing the 31 CpG sites analyzed in (D) (BS). (B) Allele-specific RT–PCR analysis of embryonic Kcnq1ot1 expression. Maternal (mat) and paternal (pat) alleles are as follows: C, wild-type castaneus; +, wild-type domesticus; and Δ, Del7AI. An expressed StuI RFLP was used to distinguish the castaneus (cast) from the domesticus (dom) Kcnq1ot1 variants. (C) DNA methylation analysis by Southern blotting of genomic DNA digested with SpeI and the methylation sensitive restriction enzyme EagI. Maternal (mat) and paternal (pat) alleles are identified as wild-type (+) or Del7AI (Δ). Also shown are the positions of methylated (me+) and unmethylated (me−) SpeI fragment. (D) Allele-specific DNA methylation analysis by sodium bisulfite sequencing. The central region of the IC2 DMR was analyzed in reciprocal Del7AI heterozygotes carrying SNPs allowing the identification of the parental alleles. Methylated CpG sites are shown in black, unmethylated sites in white, and omitted sites denote sequencing ambiguities.

Figure 6.

Imprinting at IC2 in reciprocal Del7AI heterozygotes. Expression and methylation imprinting at IC2 were analyzed in reciprocal heterozygotes for the Del7AI allele. (A) Diagram of the IC2 region, located within the intron 10 of Kcnq1, showing the region amplified by RT–PCR to analyze Kcnq1ot1 expression in (B) (RT), the position of the genomic probe used in (C) (probe) and the 335-bp sequence containing the 31 CpG sites analyzed in (D) (BS). (B) Allele-specific RT–PCR analysis of embryonic Kcnq1ot1 expression. Maternal (mat) and paternal (pat) alleles are as follows: C, wild-type castaneus; +, wild-type domesticus; and Δ, Del7AI. An expressed StuI RFLP was used to distinguish the castaneus (cast) from the domesticus (dom) Kcnq1ot1 variants. (C) DNA methylation analysis by Southern blotting of genomic DNA digested with SpeI and the methylation sensitive restriction enzyme EagI. Maternal (mat) and paternal (pat) alleles are identified as wild-type (+) or Del7AI (Δ). Also shown are the positions of methylated (me+) and unmethylated (me−) SpeI fragment. (D) Allele-specific DNA methylation analysis by sodium bisulfite sequencing. The central region of the IC2 DMR was analyzed in reciprocal Del7AI heterozygotes carrying SNPs allowing the identification of the parental alleles. Methylated CpG sites are shown in black, unmethylated sites in white, and omitted sites denote sequencing ambiguities.

Th deficiency phenotype in Del7AI homozygotes

In order to assess the phenotypic consequence of homozygous loss of the deleted region, we performed intercrosses between Del7AI heterozygotes. Live Del7AI/Del7AI pups were not recovered from such crosses, suggesting an embryonic lethality phenotype (Fig. 7A). Analysis of litters at E10.5 and E14.5 showed that the homozygous mutant embryos are viable at those stages, although abnormalities suggestive of cardiovascular defects were often observed in older embryos (Fig. 7B). These observations, together with the late developmental or perinatal lethality, were reminiscent of the phenotype described by several groups for the Th-null embryos (43,44). This developmental requirement for tyrosine hydroxylase can be shunted pharmacologically by administration of L-DOPA, the product of the reaction, in the diet of the mother during gestation. This phenotypic rescue was attempted for the Del7AI allele by repeating heterozygous crosses and adding L-DOPA to the drinking water of the pregnant females throughout gestation. Using this protocol, we were able to recover viable Del7AI/Del7AI pups (Fig. 7A and C). These TH-deficient mice are indistinguishable from heterozygous and wild-type litter mates at birth, but failed to thrive postnatally and died before weaning (Fig. 7D). The phenotypic characteristics and metabolic rescue of the homozygous Del7AI/Del7AI mutants generated here suggest that this large deletion is a phenocopy of a Th null allele with regard to the developmental phenotype and that the deleted region is thus devoid of additional genes, other than Th, required for normal development.

Figure 7.

The embryonic lethality of Del7AI homozygotes is pharmacologically rescued by provision of L-DOPA in utero. (A) Live progeny from heterozygous crosses expressed in percentage (%) of wild-type (+/+), Del7AI heterozygotes (het) and Del7AI homozygous mutants (Δ/Δ) recovered at birth (p0) or at post-implantation embryonic stages (E14.5 and E10.5). Homozygous mutants were rescued to term when the pregnant +/Del7AI heterozygous female was provided with L-DOPA in her drinking water during gestation (p0+). (B) Wild-type (+/+) and homozygous mutant (Δ/Δ) E10.5 embryos from a heterozygous cross. (C) Southern blot analysis of tail genomic DNA from live progeny of heterozygous crosses. Samples 1–4 are from a mother treated with L-DOPA, sample 1 is from a rescued homozygote. The expected EcoRV bands are 5.2 kb for wild-type and 9.7 kb for Del7AI. (D) Picture of 12-day-old pups from a heterozygous cross with L-DOPA treatment.

Figure 7.

The embryonic lethality of Del7AI homozygotes is pharmacologically rescued by provision of L-DOPA in utero. (A) Live progeny from heterozygous crosses expressed in percentage (%) of wild-type (+/+), Del7AI heterozygotes (het) and Del7AI homozygous mutants (Δ/Δ) recovered at birth (p0) or at post-implantation embryonic stages (E14.5 and E10.5). Homozygous mutants were rescued to term when the pregnant +/Del7AI heterozygous female was provided with L-DOPA in her drinking water during gestation (p0+). (B) Wild-type (+/+) and homozygous mutant (Δ/Δ) E10.5 embryos from a heterozygous cross. (C) Southern blot analysis of tail genomic DNA from live progeny of heterozygous crosses. Samples 1–4 are from a mother treated with L-DOPA, sample 1 is from a rescued homozygote. The expected EcoRV bands are 5.2 kb for wild-type and 9.7 kb for Del7AI. (D) Picture of 12-day-old pups from a heterozygous cross with L-DOPA treatment.

Recent results suggested that Th is imprinted in the developing placenta, and preferentially expressed from the maternal allele (27). This observation raises the possibility that the imprinted growth retardation phenotype noted only in Del7AI/+ maternal heterozygotes is caused by loss of TH function during development. We directly tested this hypothesis by asking whether L-DOPA provision to pregnant Del7AI heterozygous females is able to rescue to growth defects of their mutant progeny. Our results showed that L-DOPA was not able to rescue the IUGR phenotype since Del7AI/+ progeny are indistinguishable from those obtained from non-treated females at birth (Fig. 4A) and up to weaning (Fig. 4D). Our results suggest that loss of imprinted Th is not responsible for IUGR. Further studies will address the roles of IC2-regulated genes in this novel imprinted phenotype.

DISCUSSION

Several studies have provided annotation and comparative sequence analysis data for the distal Chr 7 imprinted region. Motivated by the importance of the region in BWS and as a model for imprinted gene regulation in mammals, these have addressed the conservation of transcripts and intergenic sequences between mouse, human, chicken, zebrafish and marsupials (5,6,23,45–47). Other analyses have focused on the evolution of specific gene families in the region, notably the insulin-like coding genes and the aromatic amino acid hydroxylase families (22). In the context of our study, two different aspects of this work are of particular relevance: (i) the function of the IC1–IC2 intervening region and (ii) the evolution of imprinting at IC1 and IC2.

We present here the generation and description of the largest interstitial deletion within the mouse ortholog of the BWS region. Deletion of the gap 1-Th-gap 2 intervening region in the Igf2-Ins2-Th-Ascl2 paralogy region is silent for primary germline imprinting at the flanking ICs IC1 and IC2. We show that Del7AI is associated with an imprinted IUGR phenotype and that homozygous mutants fail to develop to term. Furthermore, we show that this lethality but not growth defects can be pharmacologically rescued by bypassing the developmental requirement for the deleted Th gene. Our work has important implications for the regulation and role of imprinting on distal Chr 7.

We have uncovered an imprinted IUGR phenotype observed only in maternal heterozygotes. Since a key regulator of embryonic growth, Igf2, is expressed from the paternal chromosome, it is unlikely that the maternal Del7AI allele exerts its effect via Igf2. Trans effects have been previously described in specific allele combinations at the Ins2 locus (48), but because of functional redundancy with Ins1, Ins2-deficient mice have no reported abnormal phenotype (49,50). Another possibility was that Del7AI/+ embryos are growth retarded due to absence of the maternal Th allele. The genomic organization of human TH relative to the IC1 and IC2 sub-domains is conserved in humans, although it is located closer to the insulin locus INS due to deletion–insertion rearrangements. In human, mutations in the TH gene have been identified in patients with recessive juvenile distonia, a parkinsonian-like condition referred to as Segawa syndrome (OMIM 605407) (51). The genetic evidence therefore suggests that the Th/TH locus is not regulated by genomic imprinting, despite its genomic location. However, recent evidence suggests that Th is preferentially expressed from the maternal allele in the mouse placenta (27). However, our inability to rescue the IUGR phenotype by provision of L-DOPA during development and the results of previous studies on Th-deficient mice which have not reported any abnormalities in the heterozygotes (27,43,44) all argue against a role for loss of imprinted Th in this phenotype.

Finally, another possibility for the phenotype of Del7AI/+ embryos could be the partial disruption of expression of any one of the maternally-expressed genes of the IC2 sub-domain. Those which have been implicated in the regulation of embryonic growth and placental function include Ascl2 (42), Cdkn1c (52) and Phlda2 (41). Of these, downregulation of Ascl2 or upregulation of Cdkn1c could be expected to lead to the observed IUGR phenotype, for instance via the deletion of specific enhancer or silencer elements regulating these genes. A similar although perhaps less pronounced growth deficiency was previously described in mice expressing IC2-regulated genes biallelically following deletion of the paternal Kcnq1ot1 promoter (17).

Most known genes in the region, including Ascl2, Th, Ins2, Igf2, and H19, are all transcribed from the (−) strand of the mouse genome sequence. Other than the Th mRNA, few transcripts originating from either strand have also been described or annotated in the Ascl2–Ins2 interval. The current mouse build (Ensembl release 54) annotates a transcript from the (+) strand (cDNA clone AK139090), located 40 kb proximal of Th (gene EG624121). This predicted transcript is made of 68% sequences from repetitive elements, LINE 1 and LTR elements, and codes for a predicted protein of unknown function and without homology to known proteins. The developmental rescue of Del7AI homozygotes by L-DOPA confirms that the product of this gene does not perform an essential developmental function.

Analysis of the Del7AI allele narrows down the candidate region of cis-acting elements that are required for the imprinting of IC2 itself. This sequence is methylated during oogenesis but is unmethylated in the male gamete. These differential gametic epigenotypes are thought to be directly responsible for the imprinted, paternal-allele specific transcription of the Kcnq1ot1 ncRNA, and indirectly required for the monoallelic paternal silencing of all protein-coding genes in the region. With this current model, two critical events remain to be explained namely the mechanisms for the oocyte-specific acquisition of a DNA methylation imprint at IC2, and the establishment of a silent chromatin domain from the unmethylated paternal IC2 in cis. Two previous studies have tested the ability of large transgenic constructs from the IC2 domain to recapitulate the imprinted regulation seen at the endogenous locus. In the first study, a 260-kb BAC transgene containing the Cdkn1c gene and the 3′ end of the Kcnq1 gene, extending to IC2, failed to exhibit allele-specific imprinted expression of Cdkn1c (53). In the second study, a 800-kb YAC covering the entire IC2 domain with its centromeric end overlapping with roughly half of our Del7AI deletion, showed faithful imprinting of Phlda2, Slc22a18, Kcnq1 and Cdkn1c, but failed to exhibit allele-specific imprinted expression of Tssc4 and Ascl2 (19). IC2 function itself was not perturbed in our study of Del7AI. However, together with the results of these transgenic studies, our results raise the possibility that additional control elements, perhaps towards the telomeric end of the IC1–IC2 interval, may be critical to maintain imprinting of at least a few of the IC2-regulated genes, a possibility which will require future investigation.

MATERIALS AND METHODS

Complementary recyclable PGK-neo-pA cassettes

Two complementary PGK-neo-pA cassettes were used to insert loxP sites defining the breakpoint of the deletion on Chr 7, while providing a positive, dominant selection for a Cre-mediated recombination between these sites. The construction of these two partially recyclable cassettes, loxP-PGK-loxP-neo-pA and PGK-loxP-neo-pA-loxP has been described previously (30).

Targeting vectors, alleles and Southern blot analyses

The targeting vector to generate the I2 allele (pI2TV) as well as the description of G418r R1 ES cell lines carrying the I2 allele (Ins2I2, or Ins2tm1Nagy) and the G418sI2loxP variants obtained by transient Cre production (Ins2I2loxP, or Ins2tm1.1Nagy) have been described previously (30). To construct the targeting vector for the Ascl2M2 allele, a 4.1-kb EcoRV–EcoRI fragment of Ascl2/Mash2 3′ flanking sequences was subcloned from the 129/Sv genomic clone 19ER1 (42), into pBluescript II KS+ cut with EcoRV and EcoRI. The first NcoI site 3′ of exon 3 of Ascl2 was destroyed by partial digestion and blunt ended with T4 DNA polymerase. In pM2TV, the PGK-loxP-neo-pA-loxP cassette was inserted as a blunt EcoRI–HindIII fragment into the blunted 3′ NcoI site, located 4.3 kb downstream of the end of exon 3 of Ascl2, such that both genes were in the same transcriptional orientation. In this vector, the EcoRI site 5′ of the PGK promoter is regenerated at the junction with the 3.06-kb 5′ arm of homology from Ascl2 genomic sequences. In the M2loxP allele obtained by transient Cre expression in Ascl2+/M2 cells, the neo-pA of M2 is excised, leaving behind a PGK-loxP insertion. Positive targeted M2 clones and excised M2loxP clones were screened by Southern blotting of EcoRI-digested ES cell genomic DNA and hybridization with flanking and internal probes (Fig. 1B, probe A1: 1.4-kb EcoRI–XhoI fragment; probe A2: 1.05-kb NcoI–EcoRI fragment, the 3′ arm of the M2 targeting vector pM2TV). Excision of M2 to M2loxP in vivo in Sycp1-Cre transgenic males was monitored by KpnI digestion of genomic DNA samples and hybridization with probe A3 (0.6-kb KpnI–NcoI fragment). KpnI sites flank the M2 insertion site and yield bands of 0.9, 1.4 and 2.5-kb for the wild-type Ascl2 locus, the M2loxP and the M2 alleles, respectively.

For the analysis of progeny from heterozygous crosses (Fig. 7C), genomic DNA was digested with EcoRV and the Southern blot membrane was hybridized with the 1.6-kb EcoRV–HindIII fragment from the 3′ flanking region of Ascl2. Hybridizing bands are 5.2 kb for the wild-type allele and 9.7 kb for Del7AI.

ES cell culture and electroporation

The mouse embryonic stem cell line R1 (31) was maintained, expanded and electroporated as described (54). For selection the ES medium was supplemented with G418 (Sigma G9516) at 150 µg/ml. The PGK-neo-pA cassettes were first tested by electroporation of R1 cells with 20 µg of pPGK-loxP-neo-pA-loxP and ploxP-PGK-loxP-neo-pA, digested with EcoRI, followed by G418 selection. For the two rounds of targeting and transient Cre expression, cells were electroporated with the following constructs: (i) R1 ES cells with 20 µg of pM2TV linearized with NotI, followed by G418 selection to obtain the M2 allele; (ii) M2/+ cells with 15 µg of circular pBS165, a PGK-Cre expression vector (55) to obtain the M2loxP allele; (iii) M2loxP/+ cells with 20 µg of pI2TV linearized with NotI (30), followed by G418 selection to obtain the I2 allele; (iv) M2loxP/+ I2/+ or simply I2/+ cells with 15 µg of circular pCAGGS-nlsCre, a subclone of nlsCre into pCAGGS [(56); A.N.] to obtain cells carrying the deletion allele or I2loxP.

Mice and ES cell lines

R1 ES cell clones targeted at Ascl2, Ins2, or carrying the deletion allele, were aggregated with diploid host embryos as described (54). Mutant mice carrying the loxP site insertion alleles on Chr 7 were maintained on an outbred background (ICR, Harlan). The Sycp1-Cre transgenic mice (35) were obtained from Dr F. Cuzin (Nice). Mus musculus castaneus mice were from the Jackson Laboratories (CAST/EiJ strain). The congenic mouse line with distal Chr 7 M. m. castaneus SNPs on the 129S1 background was derived in our laboratory (129S1cCAST7/Lef). For all the genotypes of ES cells and mice described here, the maternally inherited allele is always presented first. The nomenclature for the new TAMERE alleles are Del(7Ascl2-Ins2)1Lef (synonym Del7AI) for the deletion and Dp(7Ascl2-Ins2)1Lef (synonym Dp7AI) for the duplication allele.

Genotyping

The nucleotide sequences for all primers used in this study are given in the Supplementary Material, Table S1. The segregation of the Sycp1-Cre transgene was monitored using a PCR-based Cre-specific genotyping with the PCR oligos MCKCre5′ and MCKCre3′ and annealing at 50°C. M2 to M2loxP excision from Sycp1-Cre transgenic males was detected by PCR amplification from the PGK promoter (forward primer PGK2) to downstream Ascl2 sequences (reverse primer MR1). PCR products of 1.3 kb (M2loxP) and 2.5 kb (M2) are obtained using the long-range Expand Taq (Roche) on purified genomic DNA samples from tails or ES cells. Analysis of the TAMERE products used the PCR primers F1, R1, F2 and R2 (Fig. 2B). In the duplication allele Dp7AI, a new junction generated between the 5′ of Ins2 and the 3′ end of Ascl2 is detected by a highly specific PCR reaction using the primers F2 and R1 (Taq, annealing at 60°C, 225-bp product). The PCR genotyping of Del7AI on tail genomic DNA from TAMERE (Fig. 2C) is based on the detection of a new Ascl2-neo junction by Expand Taq amplification with primers F1 and R2 (840 bp). For genotyping of Del7AI mice, we simply amplified the Ascl2-PGK junction (the Δ5′ PCR reaction), as described (30). Homozygous mutant embryos were detected by a positive Δ5′ PCR reaction and absence of a wild-type band for a PCR reaction spanning the I2 allele (I2wt reaction).

Two SNP detection assays based on genomic PCR and RFLP were used in the experiments involving the CAST mice. The assay for the SNP within exon 3 of Ascl2 has been described (30). A 511-bp PCR product from intron 2 to exon 3 (primers in2F1 and 726R) was digested with HpaII to yield fragments of 217, 198 and 96 bp for M. m. musculus alleles instead of 313 and 198 bp for CAST. The Ascl2 3′ flanking SNP, internal to the deletion was assayed by Expand Taq amplification with primers M2F and MR1 and digestion of the 618-bp product with XbaI, which cut only the CAST allele (468 and 150 bp bands).

DNA FISH

Three PAC clones from the RPCI-21 (129S6) mouse genomic DNA library and spanning the distal Chr 7 domain were used for DNA FISH experiments: the clones span the loci H19 (29G9), Th (231E13), and Phlda2 (210N7) (57). PAC DNA was purified using the Large-Construct kit (Qiagen) and directly labelled with fluorescein isothiocyanate (H19, Phlda2) or Texas Red (Th) by nick translation (Vysis, Downers Grove, IL, USA). Metaphase spreads were obtained from primary embryonic fibroblasts derived from E14.5 embryos heterozygous for the Del7AI allele (+/Del7AI genotype) or their wild-type littermate. Images were acquired and analyzed with an SD 200 Spectral Bio-imaging System (ASI Ltd, MigdalHaemek, Israel) attached to a Zeiss Axioplan 2 microscope (Carl Zeiss, Canada).

RNA purification and analysis

For the analysis of imprinted expression of the H19 and Kcnq1ot1 ncRNA, conceptuses from reciprocal crosses between CAST/Ei (or 129S1cCAST7/Lef) mice and +/Del7AI heterozygotes were recovered at E13.5 as described (54). Embryonic RNA was purified using Trizol (Invitrogen) and processed for RT–PCR by random priming using the reverse transcriptase SuperScript II (Invitrogen) on DNase-treated RNA (RQ1, Promega). The allele-specific H19 analysis was based on a polymorphic Cac8I site present in the castaneus RT–PCR product (58). The 641-bp PCR products were digested with BtsXI and Cac8I to yield bands of 418 and 223 bp for the domesticus allele but 277, 223 and 131 for castaneus. The expressed Kcnq1ot1 StuI polymorphism was kindly provided by Melissa Mann. The 465-bp PCR products were digested with SspI and StuI. The bands are of 402 and 63 bp for domesticus, 252, 150 and 63 bp for castaneus. All primer sequences are in Supplementary Material, Table S3.

DNA bisulfite modification and sequencing

Genomic DNA samples isolated from reciprocal CAST/DelTel7 and DelTel7/CAST E9.5 embryos were subjected to bisulfite modification, PCR amplification, subcloning and sequencing as described (9). For IC1 (H19 DMR, Fig. 5A), nested PCR reactions encompassing known domesticus/castaneus SNPs were carried out as described previously (but see Table S3 for primer modifications), with primers BMsp2t1+BHha1t3, followed by BMsp2t2.2+BHha2t4.2; note that these two primers were modified from the published sequence to reflect correction with the current genomic sequence of IC1 in our strains (59). For IC2 (KvDMR1, Fig. 6D), the assays were performed as described (4), except that we did not use the outer forward primer Kcnq1ot1 OF since it overlaps with a CpG. We used a semi-nested approach, with primers Kcnq1ot1 IF+Kcnq1ot1 OR followed by Kcnq1ot1 IF+Kcnq1ot1 IR (4). The parental alleles are distinguished by an A50C transversion, identified by sequencing CAST genomic DNA (dom A50/cast C50). PCR products were cloned into pGEM-T vector (Promega) and transformed into TOP10 cells (Invitrogen). Primer sequences are available in Supplementary Material, Table S3. For each DNA sample, we performed two independent bisulfite treatments; each bisulfite-treated DNA was amplified in two independent PCR reactions. We sequenced three to five cloned stands per PCR reaction and analyzed the data using BiQ Analyzer (60).

SUPPLEMENTARY MATERIAL

Supplementary Material is available at HMG online.

FUNDING

This work was supported in part by the Canadian Institutes of Health Research (MOP-82863 to L.L. and FRN 13687 to A.N.). L.L. holds a Canada Research Chair and was also supported by a Michael Smith Foundation for Health Research (MSFHR) Scholar Award.

ACKNOWLEDGEMENTS

We thank Jeff Mann and Mellissa Mann for information on the Ascl2 and Kcnq1ot1 M. m. castaneus polymorphism, respectively, Michael Higgins for providing the Chr 7 PAC clones, and François Cuzin for kindly providing the Sycp1-Cre transgenic mice.

Conflict of Interest statement. None declared.

REFERENCES

1
Verona
R.
Mann
M.R.
Bartolomei
M.
Genomic imprinting: intricacies of epigenetic regulation in clusters
Annu. Rev. Cell. Dev. Biol.
 , 
2003
, vol. 
19
 (pg. 
237
-
259
)
2
Maher
E.R.
Reik
W.
Beckwith–Wiedemann syndrome: imprinting in clusters revisited
J. Clin. Invest.
 , 
2000
, vol. 
105
 (pg. 
247
-
252
)
3
Weksberg
R.
Smith
A.C.
Squire
J.
Sadowski
P.
Beckwith–Wiedemann syndrome demonstrates a role for epigenetic control of normal development
Hum. Mol. Genet.
 , 
2003
, vol. 
12
 
Spec no. 1
(pg. 
R61
-
R68
)
4
Engemann
S.
Strödicke
M.
Paulsen
M.
Franck
O.
Reinhardt
R.
Lane
N.
Reik
W.
Walter
J.
Sequence and functional comparison in the Beckwith–Wiedemann region: implications for a novel imprinting centre and extended imprinting
Hum. Mol. Genet.
 , 
2000
, vol. 
9
 (pg. 
2691
-
2706
)
5
Onyango
P.
Miller
W.
Lehoczky
J.
Leung
C.T.
Birren
B.
Wheelan
S.
Dewar
K.
Feinberg
A.
Sequence and comparative analysis of the mouse 1-megabase region orthologous to the human 11p15 imprinted domain
Genome Res.
 , 
2000
, vol. 
10
 (pg. 
1697
-
1710
)
6
Paulsen
M.
El-Maarri
O.
Engemann
S.
Strödicke
M.
Franck
O.
Davies
K.
Reinhardt
R.
Reik
W.
Walter
J.
Sequence conservation and variability of imprinting in the Beckwith–Wiedemann syndrome gene cluster in human and mouse
Hum. Mol. Genet.
 , 
2000
, vol. 
9
 (pg. 
1829
-
1841
)
7
Bartolomei
M.
Webber
A.
Brunkow
M.
Tilghman
S.
Epigenetic mechanisms underlying the imprinting of the mouse H19 gene
Genes Dev.
 , 
1993
, vol. 
7
 (pg. 
1663
-
1673
)
8
Ferguson-Smith
A.C.
Sasaki
H.
Cattanach
B.M.
Surani
M.A.
Parental-origin-specific epigenetic modification of the mouse H19 gene
Nature
 , 
1993
, vol. 
362
 (pg. 
751
-
755
)
9
Tremblay
K.D.
Saam
J.R.
Ingram
R.S.
Tilghman
S.
Bartolomei
M.
A paternal-specific methylation imprint marks the alleles of the mouse H19 gene
Nat. Genet.
 , 
1995
, vol. 
9
 (pg. 
407
-
413
)
10
Thorvaldsen
J.L.
Duran
K.L.
Bartolomei
M.
Deletion of the H19 differentially methylated domain results in loss of imprinted expression of H19 and Igf2
Genes Dev.
 , 
1998
, vol. 
12
 (pg. 
3693
-
3702
)
11
Bell
A.C.
Felsenfeld
G.
Methylation of a CTCF-dependent boundary controls imprinted expression of the Igf2 gene
Nature
 , 
2000
, vol. 
405
 (pg. 
482
-
485
)
12
Hark
A.T.
Schoenherr
C.J.
Katz
D.J.
Ingram
R.S.
Levorse
J.M.
Tilghman
S.
CTCF mediates methylation-sensitive enhancer-blocking activity at the H19/Igf2 locus
Nature
 , 
2000
, vol. 
405
 (pg. 
486
-
489
)
13
Srivastava
M.
Hsieh
S.
Grinberg
A.
Williams-Simons
L.
Huang
S.P.
Pfeifer
K.
H19 and Igf2 monoallelic expression is regulated in two distinct ways by a shared cis acting regulatory region upstream of H19
Genes Dev.
 , 
2000
, vol. 
14
 (pg. 
1186
-
1195
)
14
Szabó
P.
Tang
S.H.
Rentsendorj
A.
Pfeifer
G.P.
Mann
J.R.
Maternal-specific footprints at putative CTCF sites in the H19 imprinting control region give evidence for insulator function
Curr. Biol.
 , 
2000
, vol. 
10
 (pg. 
607
-
610
)
15
Lee
M.P.
DeBaun
M.R.
Mitsuya
K.
Galonek
H.L.
Brandenburg
S.
Oshimura
M.
Feinberg
A.
Loss of imprinting of a paternally expressed transcript, with antisense orientation to KVLQT1, occurs frequently in Beckwith–Wiedemann syndrome and is independent of insulin-like growth factor II imprinting
Proc. Natl. Acad. Sci. USA
 , 
1999
, vol. 
96
 (pg. 
5203
-
5208
)
16
Smilinich
N.J.
Day
C.D.
Fitzpatrick
G.V.
Caldwell
G.M.
Lossie
A.C.
Cooper
P.R.
Smallwood
A.C.
Joyce
J.A.
Schofield
P.N.
Reik
W.
, et al.  . 
A maternally methylated CpG island in KvLQT1 is associated with an antisense paternal transcript and loss of imprinting in Beckwith–Wiedemann syndrome
Proc. Natl. Acad. Sci. USA
 , 
1999
, vol. 
96
 (pg. 
8064
-
8069
)
17
Fitzpatrick
G.
Soloway
P.
Higgins
M.
Regional loss of imprinting and growth deficiency in mice with a targeted deletion of KvDMR1
Nat. Genet.
 , 
2002
, vol. 
32
 (pg. 
426
-
431
)
18
Ainscough
J.F.
Koide
T.
Tada
M.
Barton
S.
Surani
M.A.
Imprinting of Igf2 and H19 from a 130 kb YAC transgene
Development
 , 
1997
, vol. 
124
 (pg. 
3621
-
3632
)
19
Cerrato
F.
Sparago
A.
Di Matteo
I.
Zou
X.
Dean
W.
Sasaki
H.
Smith
P.
Genesio
R.
Bruggemann
M.
Reik
W.
Riccio
A.
The two-domain hypothesis in Beckwith–Wiedemann syndrome: autonomous imprinting of the telomeric domain of the distal chromosome 7 cluster
Hum. Mol. Genet.
 , 
2005
, vol. 
14
 (pg. 
503
-
511
)
20
Giddings
S.J.
King
C.D.
Harman
K.W.
Flood
J.F.
Carnaghi
L.R.
Allele specific inactivation of insulin 1 and 2, in the mouse yolk sac, indicates imprinting
Nat. Genet.
 , 
1994
, vol. 
6
 (pg. 
310
-
313
)
21
Guillemot
F.
Caspary
T.
Tilghman
S.
Copeland
N.G.
Gilbert
D.J.
Jenkins
N.A.
Anderson
D.J.
Joyner
A.L.
Rossant
J.
Nagy
A.
Genomic imprinting of Mash2, a mouse gene required for trophoblast development
Nat. Genet.
 , 
1995
, vol. 
9
 (pg. 
235
-
242
)
22
Patton
S.J.
Luke
G.N.
Holland
P.W.
Complex history of a chromosomal paralogy region: insights from amphioxus aromatic amino acid hydroxylase genes and insulin-related genes
Mol. Biol. Evol.
 , 
1998
, vol. 
15
 (pg. 
1373
-
1380
)
23
Paulsen
M.
Khare
T.
Burgard
C.
Tierling
S.
Walter
J.
Evolution of the Beckwith–Wiedemann syndrome region in vertebrates
Genome Res.
 , 
2005
, vol. 
15
 (pg. 
146
-
153
)
24
Giannoukakis
N.
Deal
C.
Paquette
J.
Goodyer
C.G.
Polychronakos
C.
Parental genomic imprinting of the human IGF2 gene
Nat. Genet.
 , 
1993
, vol. 
4
 (pg. 
98
-
101
)
25
Moore
G.E.
Abu-Amero
S.N.
Bell
G.
Wakeling
E.L.
Kingsnorth
A.
Stanier
P.
Jauniaux
E.
Bennett
S.T.
Evidence that insulin is imprinted in the human yolk sac
Diabetes
 , 
2001
, vol. 
50
 (pg. 
199
-
203
)
26
Ohlsson
R.
Nyström
A.
Pfeifer-Ohlsson
S.
Töhönen
V.
Hedborg
F.
Schofield
P.
Flam
F.
Ekström
T.J.
IGF2 is parentally imprinted during human embryogenesis and in the Beckwith–Wiedemann syndrome
Nat. Genet.
 , 
1993
, vol. 
4
 (pg. 
94
-
97
)
27
Schulz
R.
Menheniott
T.R.
Woodfine
K.
Wood
A.
Choi
J.
Oakey
R.
Chromosome-wide identification of novel imprinted genes using microarrays and uniparental disomies
Nucleic Acids Res.
 , 
2006
, vol. 
34
 pg. 
e88
 
28
Shirohzu
H.
Yokomine
T.
Sato
C.
Kato
R.
Toyoda
A.
Purbowasito
W.
Suda
C.
Mukai
T.
Hattori
M.
Okumura
K.
Sakaki
Y.
Sasaki
H.
A 210-kb segment of tandem repeats and retroelements located between imprinted subdomains of mouse distal chromosome 7
DNA Res.
 , 
2004
, vol. 
11
 (pg. 
325
-
334
)
29
Zhao
Z.
Tavoosidana
G.
Sjölinder
M.
Göndör
A.
Mariano
P.
Wang
S.
Kanduri
C.
Lezcano
M.
Sandhu
K.S.
Singh
U.
, et al.  . 
Circular chromosome conformation capture (4C) uncovers extensive networks of epigenetically regulated intra- and interchromosomal interactions
Nat. Genet.
 , 
2006
, vol. 
38
 (pg. 
1341
-
1347
)
30
Oh
R.
Ho
R.
Mar
L.
Gertsenstein
M.
Paderova
J.
Hsien
J.
Squire
J.A.
Higgins
M.
Nagy
A.
Lefebvre
L.
Epigenetic and phenotypic consequences of a truncation disrupting the imprinted domain on distal mouse chromosome 7
Mol. Cell. Biol.
 , 
2008
, vol. 
28
 (pg. 
1092
-
1103
)
31
Nagy
A.
Rossant
J.
Nagy
R.
Abramow-Newerly
W.
Roder
J.C.
Derivation of completely cell culture-derived mice from early-passage embryonic stem cells
Proc. Natl. Acad. Sci. USA
 , 
1993
, vol. 
90
 (pg. 
8424
-
8428
)
32
Hérault
Y.
Rassoulzadegan
M.
Cuzin
F.
Duboule
D.
Engineering chromosomes in mice through targeted meiotic recombination (TAMERE)
Nat. Genet.
 , 
1998
, vol. 
20
 (pg. 
381
-
384
)
33
Tarchini
B.
Huynh
T.H.
Cox
G.A.
Duboule
D.
HoxD cluster scanning deletions identify multiple defects leading to paralysis in the mouse mutant Ironside
Genes Dev.
 , 
2005
, vol. 
19
 (pg. 
2862
-
2876
)
34
Chung
S.S.
Cuzin
F.
Rassoulzadegan
M.
Wolgemuth
D.J.
Primary spermatocyte-specific Cre recombinase activity in transgenic mice
Transgenic Res.
 , 
2004
, vol. 
13
 (pg. 
289
-
294
)
35
Vidal
F.
Sage
J.
Cuzin
F.
Rassoulzadegan
M.
Cre expression in primary spermatocytes: a tool for genetic engineering of the germ line
Mol. Reprod. Dev.
 , 
1998
, vol. 
51
 (pg. 
274
-
280
)
36
Constância
M.
Hemberger
M.
Hughes
J.
Dean
W.
Ferguson-Smith
A.
Fundele
R.
Stewart
F.
Kelsey
G.
Fowden
A.
Sibley
C.
Reik
W.
Placental-specific IGF-II is a major modulator of placental and fetal growth
Nature
 , 
2002
, vol. 
417
 (pg. 
945
-
948
)
37
DeChiara
T.M.
Robertson
E.J.
Efstratiadis
A.
Parental imprinting of the mouse insulin-like growth factor II gene
Cell
 , 
1991
, vol. 
64
 (pg. 
849
-
859
)
38
Hatada
I.
Mukai
T.
Genomic imprinting of p57KIP2, a cyclin-dependent kinase inhibitor, in mouse
Nat. Genet.
 , 
1995
, vol. 
11
 (pg. 
204
-
206
)
39
Yan
Y.
Frisén
J.
Lee
M.
Massagué
J.
Barbacid
M.
Ablation of the CDK inhibitor p57Kip2 results in increased apoptosis and delayed differentiation during mouse development
Genes Dev.
 , 
1997
, vol. 
11
 (pg. 
973
-
983
)
40
Zhang
P.
Liégeois
N.J.
Wong
C.
Finegold
M.
Hou
H.
Thompson
J.C.
Silverman
A.
Harper
J.W.
DePinho
R.A.
Elledge
S.J.
Altered cell differentiation and proliferation in mice lacking p57KIP2 indicates a role in Beckwith–Wiedemann syndrome
Nature
 , 
1997
, vol. 
387
 (pg. 
151
-
158
)
41
Frank
D.
Fortino
W.
Clark
L.
Musalo
R.
Wang
W.
Saxena
A.
Li
C.M.
Reik
W.
Ludwig
T.
Tycko
B.
Placental overgrowth in mice lacking the imprinted gene Ipl
Proc. Natl. Acad. Sci. USA
 , 
2002
, vol. 
99
 (pg. 
7490
-
7495
)
42
Guillemot
F.
Nagy
A.
Auerbach
A.
Rossant
J.
Joyner
A.L.
Essential role of Mash-2 in extraembryonic development
Nature
 , 
1994
, vol. 
371
 (pg. 
333
-
336
)
43
Kobayashi
K.
Morita
S.
Sawada
H.
Mizuguchi
T.
Yamada
K.
Nagatsu
I.
Hata
T.
Watanabe
Y.
Fujita
K.
Nagatsu
T.
Targeted disruption of the tyrosine hydroxylase locus results in severe catecholamine depletion and perinatal lethality in mice
J. Biol. Chem.
 , 
1995
, vol. 
270
 (pg. 
27235
-
27243
)
44
Zhou
Q.Y.
Quaife
C.J.
Palmiter
R.D.
Targeted disruption of the tyrosine hydroxylase gene reveals that catecholamines are required for mouse fetal development
Nature
 , 
1995
, vol. 
374
 (pg. 
640
-
643
)
45
Ager
E.I.
Pask
A.
Gehring
H.M.
Shaw
G.
Renfree
M.
Evolution of the CDKN1C-KCNQ1 imprinted domain
BMC Evol. Biol.
 , 
2008
, vol. 
8
 pg. 
163
 
46
Paulsen
M.
Davies
K.R.
Bowden
L.M.
Villar
A.J.
Franck
O.
Fuermann
M.
Dean
W.L.
Moore
T.F.
Rodrigues
N.
Davies
K.E.
, et al.  . 
Syntenic organization of the mouse distal chromosome 7 imprinting cluster and the Beckwith–Wiedemann syndrome region in chromosome 11p15.5
Hum. Mol. Genet.
 , 
1998
, vol. 
7
 (pg. 
1149
-
1159
)
47
Yokomine
T.
Structural and functional analysis of a 0.5-Mb chicken region orthologous to the imprinted mammalian Ascl2/Mash2-Igf2-H19 region
Genome Res.
 , 
2005
, vol. 
15
 (pg. 
154
-
165
)
48
Duvillié
B.
Bucchini
D.
Tang
T.
Jami
J.
Pàldi
A.
Imprinting at the mouse Ins2 locus: evidence for cis- and trans-allelic interactions
Genomics
 , 
1998
, vol. 
47
 (pg. 
52
-
57
)
49
Duvillié
B.
Cordonnier
N.
Deltour
L.
Dandoy-Dron
F.
Itier
J.M.
Monthioux
E.
Jami
J.
Joshi
R.L.
Bucchini
D.
Phenotypic alterations in insulin-deficient mutant mice
Proc. Natl. Acad. Sci. USA
 , 
1997
, vol. 
94
 (pg. 
5137
-
5140
)
50
Leroux
L.
Desbois
P.
Lamotte
L.
Duvillié
B.
Cordonnier
N.
Jackerott
M.
Jami
J.
Bucchini
D.
Joshi
R.L.
Compensatory responses in mice carrying a null mutation for Ins1 or Ins2
Diabetes
 , 
2001
, vol. 
50
 
Suppl 1
(pg. 
S150
-
S153
)
51
Lüdecke
B.
Dworniczak
B.
Bartholomé
K.
A point mutation in the tyrosine hydroxylase gene associated with Segawa's syndrome
Hum. Genet.
 , 
1995
, vol. 
95
 (pg. 
123
-
125
)
52
Andrews
S.
Wood
M.
Tunster
S.
Barton
S.
Surani
M.
John
R.
Cdkn1c (p57Kip2) is the major regulator of embryonic growth within its imprinted domain on mouse distal chromosome 7
BMC Dev. Biol.
 , 
2007
, vol. 
7
 pg. 
53
 
53
John
R.M.
Ainscough
J.F.
Barton
S.C.
Surani
M.A.
Distant cis-elements regulate imprinted expression of the mouse p57(Kip2) (Cdkn1c) gene: implications for the human disorder, Beckwith–Wiedemann syndrome
Hum. Mol. Genet.
 , 
2001
, vol. 
10
 (pg. 
1601
-
1609
)
54
Nagy
A.
Gertsenstein
M.
Vintersten
K.
Behringer
R.
Manipulating the Mouse Embryo: A Laboratory Manual
 , 
2003
Cold Spring Harbor, N.Y
Cold Spring Harbor Laboratory Press
55
Sauer
B.
Manipulation of transgenes by site-specific recombination: use of Cre recombinase
Methods Enzymol.
 , 
1993
, vol. 
225
 (pg. 
890
-
900
)
56
Niwa
H.
Yamamura
K.
Miyazaki
J.
Efficient selection for high-expression transfectants with a novel eukaryotic vector
Gene
 , 
1991
, vol. 
108
 (pg. 
193
-
199
)
57
Day
C.D.
Smilinich
N.J.
Fitzpatrick
G.V.
deJong
P.J.
Shows
T.B.
Higgins
M.J.
The imprinted domain in mouse distal Chromosome 7: reagents for mutagenesis and sequencing
Mamm. Genome
 , 
1999
, vol. 
10
 (pg. 
182
-
185
)
58
Thorvaldsen
J.L.
Mann
M.R.
Nwoko
O.
Duran
K.L.
Bartolomei
M.
Analysis of sequence upstream of the endogenous H19 gene reveals elements both essential and dispensable for imprinting
Mol. Cell. Biol.
 , 
2002
, vol. 
22
 (pg. 
2450
-
2462
)
59
Davis
T.L.
Trasler
J.M.
Moss
S.B.
Yang
G.J.
Bartolomei
M.
Acquisition of the H19 methylation imprint occurs differentially on the parental alleles during spermatogenesis
Genomics
 , 
1999
, vol. 
58
 (pg. 
18
-
28
)
60
Bock
C.
BiQ Analyzer: visualization and quality control for DNA methylation data from bisulfite sequencing
Bioinformatics
 , 
2005
, vol. 
21
 (pg. 
4067
-
4068
)

Author notes

Present address: Department of Pathology and Molecular Medicine, Queen's University, Kingston, ON, Canada, K7L 3N6.