Abnormalities of developmental cell death in Dad1-de®cient mice Kiyomasa Nishii1,*, Teruhisa Tsuzuki2,a, Madoka Kumai1, Naoki Takeda3,b, Hideya Koga1, Shinichi Aizawa3, Takeharu Nishimoto4 and Yosaburo Shibata1 1
Department of Developmental Molecular Anatomy, Faculty of Medicine, Kyushu University, 3-1-1 Maidashi, Fukuoka 812-8582, Japan 2 Medical Institute of Bioregulation, Kyushu University, 3-1-1 Maidashi, Fukuoka 812-8582, Japan 3 Department of Morphogenesis, Institute of Molecular Embryology and Genetics, Kumamoto University School of Medicine, 2-2-1 Honjo, Kumamoto 860-0811, Japan 4 Department of Molecular Biology, Kyushu University, 3-1-1 Maidashi, Fukuoka 812-8582, Japan
Abstract Background: Dad1, the defender against apoptotic cell death, comprises the oligosaccharyltransferase complex and is well conserved among eukaryotes. In hamster BHK21-derived tsBN7 cells, loss of Dad1 causes apoptosis which cannot be prevented by Bcl-2. Results: To determine the role of Dad1 function in vivo, we prepared by gene targeting, mice harbouring a disrupted Dad1 gene. Homozygous mutants died shortly after they were implanted with the characteristic features of apoptosis. In an in vitro blastocyst culture system, Dad1-null cells displayed
Introduction Dad1 was originally identi®ed as a highly hydrophobic protein, whose defect causes apoptosis in tsBN7 cells. It has been proposed that Dad1 regulates the apoptotic pathway, either downstream of or independent of, Bcl2, because this process cannot be prevented by Bcl-2 (Nakashima et al. 1993). In Caenorhabditis elegans, the expression of human DAD1 inhibits developmentally programmed cell death, indicating the existence of a conserved molecular mechanism in which Dad1 participates (Sugimoto et al. 1995). Indeed, homologues Communicated by: Tim Hunt * Correspondence: E-mail:
[email protected]. ac.jp Present addresses: aDepartment of Medical Biophysics and Radiation Biology, Faculty of Medicine, Kyushu University, 3-1-1 Maidashi, Fukuoka 812-8582, Japan; bCenter for Animal Resources and Development, Kumamoto University, 4-24-1 Kuhonji, Kumamoto 862-0976, Japan. q Blackwell Science Limited
abnormalities which were comparable to those obtained in vivo. However, oligosaccharyltransferase activity was apparently retained even after the Dad1-null cells were destined to die. Some liveborn heterozygous mutants displayed soft-tissue syndactyly. Mild thymic hypoplasia was also indicated in heterozygotes. Conclusion: These results suggest the involvement of the Dad1 gene in the acquisition of a common syndactyly phenotype, as well as in the control of programmed cell death during development.
of Dad1 were isolated from many eukaryotic organisms and proved to be highly homologous to each other. Even with the yeast homologue Ost2p, amino acid sequences are 40% identical and 65% similar to those of human DAD1 (Silberstein et al. 1995). Although the temperature-sensitive apoptosis of tsBN7 cannot be prevented by yeast OST2 cDNA, it can be prevented by Xenopus, C. elegans, Arabidopsis thaliana and rice, in addition to human DAD1 (Sugimoto et al. 1995; Gallois et al. 1997; Makishima et al. 1997; Tanaka et al. 1997a; Nakashima et al. 1993). Surprisingly, Ost2p comprises the essential «-subunit of the yeast oligosaccharyltransferase complex (Silberstein et al. 1995). Mammalian Dad1 also seems to constitute the oligosaccharyltransferase complex, as revealed either by protein-crosslinking or by the yeast two-hybrid assay (Fu et al. 1997; Kelleher & Gilmore 1997). Moreover, tsBN7 cells display a defect in N-linked glycosylation at the nonpermissive temperature, resulting in apoptosis (Makishima et al. 1997). Although these ®ndings have provided us with extensive Genes to Cells (1999) 4, 243±252
243
K Nishii et al.
knowledge concerning the biochemical and genetic studies of Dad1 in vitro, its roles in animal development have not yet been clari®ed. In order to determine the role of Dad1 function in vivo, we prepared by gene targeting, mice harbouring a disrupted Dad1 gene. In this study, we have shown that Dad1 has signi®cant roles to play during mouse development. Soon after implantation, mice homozygous for the Dad1 mutation (Dad1 / ) display severe growth impairment and die, showing the characteristics of apoptosis. Although heterozygous mutants (Dad1/ ) are almost indistinguishable from their wild-type littermates, a signi®cant portion of them display softtissue syndactyly. Mild thymic hypoplasia is also described as a heterozygous phenotype.
Results Targeting of the Dad1 gene We cloned the mouse Dad1 genomic DNA and constructed a targeting vector which disrupts the Dad1 open-reading frame with the neor gene at a site
upstream from the ®rst membrane-spanning domain (Fig. 1A) (Deng et al. 1993). A pair of herpes simplex virus thymidine kinase cassettes was engineered to ¯ank the Dad1 genomic region in order to increase targeting ef®ciency (Mansour et al. 1988; Rancourt et al. 1995). The linearized targeting vector was electroporated into CCE embryonic stem cells. Correctly targeted clones were identi®ed at the remarkably high incidence of 36/72 drug-resistant clones (Fig. 1B) (Wurst & Joyner 1993). Three lines of germ-line chimeras were obtained and used for the experiments (Papaioannou & Johnson 1993). The phenotypes were similar among these lines.
Dad1 / mice die at the early postimplantation period Dad1/ mice were viable and fertile. When these Dad1/ animals were intercrossed, however, no Dad1 / null mutants were obtained from the viable pups, indicating embryonic lethality. Embryos from the timed matings of heterozygous intercrosses were genotyped by the polymerase chain reaction (PCR)
Figure 1 Targeting of the Dad1 gene. (A) Schematic representation of the Dad1targeted disruption strategy. The Dad1coding exons are shown in black boxes. Ev, EcoRV; H, HindIII; Sc, SacI. (B) Southern blot analysis of the DNA isolated from the germ-line-transmitted embryonic stem cell lines. An EcoRV digest hybridized with probe B (30 ¯anking) and a SacI digest hybridized with probe A (50 internal) yielded a wild-type band and a mutant band indicated in (A). wt, wild-type. (C) PCR-mediated genotyping of the E8.5 embryos and the blastocyst embryos (E3.5). The primers are shown by arrowheads in (A). Wild-type 0.36 kb and mutant 0.69 kb bands are clearly distinguishable. / , wild-type; / , heterozygote; / , homozygote; N, negative control.
244
Genes to Cells (1999) 4, 243±252
q Blackwell Science Limited
Dad1-de®cient mice Table 1 Classi®cation of offspring from heterozygous intercrosses
Histology
E3.5 E6.5 E7.5 E8.5 E10.5 5 weeks
Genotype
Normal
Abnormal
18 18
4 3² 2 7²
/ 7 20 2³ 11 109
/
/
P-valuea
14
11
0.472
26 29³ 17 14³ 180
0 2³ 0 0
< 0.001 0.007 < 0.001
* Determined from the x2-values to test the null hypothesis of a 1:2:1 ratio. ² Numbers indicate abnormal implantation sites ®lled with apoptotic cells. ³ Numbers indicate resorption.
method using the foetal membrane of embryonic day 10.5 embryos (E10.5) and the embryo proper of E8.5 and E3.5. Empty deciduae lacking any discernible embryos indicated resorption after implantation. Small pink-coloured debris was sometimes found in place of the embryos. In these cases, a piece of the extraembryonic tissue or debris was analysed by the PCR method. Among the 33 resorptions at E8.5, only two null mutants were detected, whereas the Dad1 / genotype was commonly found in the E3.5 blastocyst at the expected Mendelian ratio, with no obvious abnormality (Fig. 1C and Table 1). To understand the phenotypes of Dad1 / , the E6.5
and E7.5 decidual swellings from the heterozygous intercrosses were serially sectioned and analysed histologically (Kaufman 1992; Hogan et al. 1994). Abnormal embryos comprised 28% (E6.5) and 33% (E7.5) of the examined samples, and were presumed to be Dad1 / (Table 1). At E6.5, the Dad1 / embryos were distinguishable by their small size (Fig. 2A,D). Compared with the normal control, the proamniotic cavity was narrow and the morphological distinction between the embryonic and extra-embryonic tissue was not clear, indicating growth retardation in Dad1 / . Cells in mitosis, however, could be observed easily. In contrast, at E7.5, almost all of the abnormal implantation sites
Figure 2 Dad1 / phenotype. (A,D) Haematoxylin-eosin-stained sections through E6.5 normal (A) and presumptive Dad1 / (D) embryos. The ectoderm and the endoderm in (D) are indicated by an asterisk and an arrow, respectively. This Dad1 / embryo was retarded in growth, reminiscent of embryos in earlier stages. (B,E) Haematoxylin±eosin-stained sections through E7.5 normal (B) and presumptive Dad1 / (E) embryos. The boxed area in (E) is shown at a higher magni®cation in the inset. Note the pyknotic cells and the complete loss of embryonic structure in Dad1 / . (C,F) A TUNEL assay near the sections corresponding to (B,E). a, amniotic cavity; ec, ectoplacental cavity; EmEc, embryonic ectoderm; EmEn, embryonic endoderm; ex, exocoelomic cavity; ExEc, extra-embryonic ectoderm; ExEn, extraembryonic endoderm; pa, proamniotic cavity. Bars 100 mm. q Blackwell Science Limited
Genes to Cells (1999) 4, 243±252
245
K Nishii et al.
showed bleeding and were occupied by the disordered arrangement of abnormal cells with pyknotic chromatin (Fig. 2B,E). These cells were labelled positive by a TdTmediated dUTP-biotin nick end-labelling (TUNEL) assayÐone of the methods used for detecting apoptotic cells in situ (Fig. 2C,F) (Gavrieli et al. 1992). The transitional state where the abnormal embryos had begun to break down was not observed, indicating that this process had progressed rapidly.
Dad1 / embryos from blastocysts undergo apoptotic death in the presence of N-glycans in vitro We examined the blastocyst outgrowth in an in vitro culture systemÐan alternative method for studying early postimplantational development. It enabled us to follow the developmental consequences of the same embryo, free from external maternal factors which contained materials from the wild-type Dad1 allele. Genotypes were subsequently determined after culture. All blastocysts examined were indistinguishable, indicating that the Dad1 / embryos were morphologically normal before implantation (Fig. 3A±C). About half of the Dad1 / blastocysts underwent apparently normal hatching, implantation, and initial outgrowth of trophoblasts (Fig. 3B,E). However, at culture days 2±5, they were suddenly detached from the culture dish (Fig. 3H,K), reminiscent of E7.5 Dad1 / embryos in vivo (Fig. 2E). In contrast, the other Dad1 / embryos displayed notable growth impairment (Fig. 3F,I,L). They remained almost unchanged in shape throughout the culture. It is probable that embryos showing severe growth impairment, and which resemble E6.5 Dad1 / embryos (Fig. 2D), were probably resorbed through apoptotic death soon after they were implanted, as has been observed in vivo. Sometimes Dad1/ embryos showed Dad1 / like growth impairment (Fig. 3M), but all of the wild-type blastocysts underwent normal development. Taken together, the culture experiment gave results that were comparable to those obtained in vivo, indicating that factors outside the embryo are unlikely to be related to lethality. To con®rm that apoptosis would really take place, we carried out an annexin V-binding and propidium iodide-uptake assay at culture day 3. This assay detects the loss of phospholipid asymmetry of the plasma membrane, which is an early event in apoptosis, by binding annexin V to the phosphatidylserine residue exposed at the outer plasma membrane lea¯et (for 246
Genes to Cells (1999) 4, 243±252
review, see van Engeland et al. 1998). This assay was performed for three litters. Embryos of six Dad1 / mice existed and two of them displayed a typical annexin V-positive appearance. As shown in Fig. 3N, early apoptotic cells were detected extensively at the inner cell mass region from the apparently normal Dad1 / embryo (compare this ®gure with Fig. 4B which shows the same specimen at lower magni®cation). We next investigated whether oligosaccharyltransferase activity was retained in Dad1 / embryos. If we consider that Dad1 constitutes the oligosaccharyltransferase complex, then it is very likely that disruption of the Dad1 gene would lead to a complete loss of Nglycans. In order to determine whether Dad1 / embryos lacked N-glycans at their death, cells in the blastocyst culture were cytochemically labelled with the lectin Phaseolus vulgaris erythroagglutinin (E-PHA). The reason why we chose this lectin for detecting N-glycans is that it is one of the most reliable, speci®c lectins discovered to date. The E-PHA binding capability of Mgat1-null embryos is lost, indicating that this lectin does not bind to any substances other than hybrid and complex N-glycans (Campbell et al. 1995). Loss of oligosaccharyltransferase activity implies the absence of both hybrid and complex N-glycans, and accordingly, this should lead to a lack of capability with regard to E-PHA binding. Surprisingly, however, the E-PHA-binding capability of Dad1 / was apparently retained, even after they were directed to die (Figs 3N and 4). The staining pattern was unaltered by the preceding annexin V-binding assay, and virtually the same result was obtained by the concanavalin A lectin (data not shown). It should be noted, however, that the intensity of the signal seemed to be weaker than in the case of the heterozygous and wild-type controls.
The number of mutant animals deviates signi®cantly Table 1 shows the distribution of each genotype. Despite apparently normal viability, the number of Dad1/ mice was signi®cantly less than the expected number as deduced from the Mendelian segregation. In contrast to the expected 1:2 ratio, the observed ratio was 1:1.6, which continued after E8.5 and which was signi®cantly different (x2 4.48, d.f. 1, P 0.034). In the case of backcrossing, a ratio of 1:1 was expected, whereas the ratio we obtained was 1 (wild-type):0.83 (Dad1/ ) (x2 4.68, d.f. 1, P 0.031). It would therefore seem that about 20% of Dad1/ embryos die during development. Resorption of Dad1/ was often observed, and seems to represent a heterozygous q Blackwell Science Limited
Dad1-de®cient mice
phenocopy (Table 1). As expected, Dad1 / like growth impairment was observed in four of the 21 cultured Dad1/ blastocysts (Fig. 3M).
Some Dad1/ mice show soft-tissue syndactyly We found that Dad1/ mice frequently displayed softtissue syndactyly, which almost always comprised of a web remaining between the second and third digits
(Fig. 5A). The ®nger bones visualized by X-ray photographs or cleared alizarin red staining were morphologically normal (data not shown). The hindlimbs of 14% of the Dad1/ mice were affected (left only: 30%; right only: 5%; both: 65%), as well as the forelimbs of 5% (left only: 25%; right only: 13%; both: 63%). All the pups with forelimb deformity possessed an additional hindlimb abnormality. The frequency of syndactyly in Dad1/ mice was signi®cantly high among F1 ´ F1 heterozygous intercrosses (x2 5.36,
Figure 3 Representative outgrowths of cultured blastocysts. The day 0 (A±C), 2 (D±F), 3 (G±I), and 4 ( J±L) embryos of wild-type (A, D, G, J) and Dad1 / (B, E, H, K and C, F, I, L) cultured in vitro. Dad1 / blastocysts appeared to be morphologically normal with characteristic inner cell mass and trophectoderm (B, C). About half of the Dad1 / embryos were indistinguishable from the wild-type embryos until they suddenly became detached from the culture dish (B, E, H, K). The other Dad1 / embryos showed a severe outgrowth defect (C, F, I, L). Occasionally, Dad1/ embryos showed a comparable outgrowth defect to Dad1 / (M: culture day 3 embryo). An annexin V-binding and propidium iodide-uptake assay (N: culture day 3 Dad1 / embryo) indicated that apoptosis had occurred in the inner cell mass region, while the surrounding cells were still intact. Green, which represents the annexin V binding, marks a membrane lipid change that occurs as an early apoptotic event, while red, which represents propidium iodide uptake, labels DNA after the membrane has lost its integrity, a later apoptotic event. Images from phase contrast and ¯uorescence microscopy are merged. icm, inner cell mass; tb, trophoblast; te, trophectoderm. Bars 100 mm. q Blackwell Science Limited
Genes to Cells (1999) 4, 243±252
247
K Nishii et al.
Figure 4 Lectin staining of culture day 3 embryos. (A, B) Phase contrast images of Dad1/ (A) and Dad1 / (B) embryos. (C, D) E-PHA binding was detected under ¯uorescence microscopy (C: Dad1/ , D: Dad1 / ). Note that the Dad1 / sample was the same as in Fig. 3N. At the lectin-staining stage, the annexin V signal had disappeared and the propidium iodide signal could be virtually ignored (data not shown). Abbreviations are the same as those used in Fig. 3. Bar 100 mm.
d.f. 1, P 0.021). Interestingly, as shown in Fig. 5B, the penetration of syndactyly varied according to the background. This suggests that other factors might be required to cause syndactyly in Dad1/ mice.
Dad1/ mice show mild thymic hypoplasia We examined other abnormalities among Dad1/ mice. The total body weight of 5-week-old pups was similar between wild-type and Dad1/ (data not shown). The weight of the organ was standardized as the ratio of its weight relative to total body weight (Table 2). In contrast to the spleen and the kidney, the thymus of the Dad1/ mice weighed about 10% less than in the wild-type in both sexes. A histological section through the thymus revealed no difference, except for the relatively small size of the Dad1/ (data not shown). This mild thymic hypoplasia may have been due either to a disruption to the cis element of the T-cell receptor a/d constant gene close to the Dad1 locus (Hong et al. 1997; Wang et al. 1997), or to the positive effect on the apoptotic selection process in the thymus by the partial loss of Dad1, or both. 248
Genes to Cells (1999) 4, 243±252
Figure 5 Limb deformity in Dad1/ .(A) Two Dad1/ males of the same litter from an F1 ´ F1 intercross. Soft-tissue syndactyly of the hand/foot between the second and third digits is visible (black mouse). Inset is another view of this phenotype. (B) The penetrance of syndactyly is demonstrated by a bar chart. The occurrence ratio was signi®cantly higher for Dad1/ progeny from F1 ´ F1 intercrosses. However, the ratio differed among the strains. Note that the phenotype of wild-type animals was very weak in each case. The number of mice used to generate the data are indicated under the genotypes. F1, offspring from crosses of chimeras with C57Bl/6 females. B6 ´ F1, offspring from backcrosses of Dad1/ males with C57Bl/6 females; F1 ´ B6, offspring from backcrosses of C57Bl/6 males with Dad1/ females.
Discussion In this report, we have described how the functional Dad1 gene is required for ontogenetic processes in the mouse embryo. Our observation of early lethality is not surprising, when one considers that Dad1 is essential for cell survival even in yeasts and the cultured tsBN7 cell line. In tsBN7 cells, a loss of Dad1 leads to lethality within 2 days (Nakashima et al. 1993). It should be noted, however, that our results demonstrate that Dad1 / null mutants can survive much longer than tsBN7, although ®nal death by apoptosis is a characteristic of both cases. This phenotype is quite different q Blackwell Science Limited
Dad1-de®cient mice Table 2 Weight analysis of Dad1/ mice relative to wild-type mice Male
Female
Organ
/ (N 19) (%)
/ (N 44) (%)
P-value*
/ (N 20) (%)
/ (N 34) (%)
P-value*
Thymus Spleen Kidney
0.400 6 0.099 0.423 6 0.092 0.765 6 0.057
0.354 6 0.071 0.395 6 0.100 0.742 6 0.066
0.039 0.299 0.182
0.539 6 0.111 0.499 6 0.278 0.657 6 0.042
0.484 6 0.079 0.427 6 0.083 0.667 6 0.068
0.038 0.166 0.590
The weight of the isolated tissue (g) was divided by the total body weight (g) of each animal and is indicated as a percentage. Data are means 6 SD. *Determined from the two-tailed Student's t-test.
from the previous reports which dealt with early postimplantation lethality (Spyropoulos & Capecchi 1994; FaÈssler & Meyer 1995; Feldman et al. 1995; Jones et al. 1995; Montes de Oca Luna et al. 1995; Stephens et al. 1995; Lim & Hasty 1996; Liu et al. 1996; Tsuzuki et al. 1996; Xanthoudakis et al. 1996; Murphy et al. 1997; Takeda et al. 1997). Dad1 / embryos displayed many cells containing fragmented, TUNEL-positive nuclei, ®ndings which differed markedly from those of the previous reports, indicating that apoptosis had occurred simultaneously in the embryo itself (Fig. 2). This was further con®rmed by the in vitro system, where early apoptotic cells were labelled annexin V-positive (Fig. 3). The manner in which Dad1 de®ciency leads to apoptotic death remains to be determined. Dad1 comprises the oligosaccharyltransferase complex. This enzyme catalyses the transfer of an oligosaccharide from the dolichol carrier to a susceptible asparagine residue en bloc on a nascent protein, providing the common precursor of N-linked oligosaccharides (for review, see Silberstein & Gilmore 1996). Thus, we can infer that disruption of the Dad1 gene should lead to an absence of N-glycans. However, oligosaccharyltransferase activity was apparently retained, even when the Dad1 / cells began to die. This result suggests two possibilities. One is that Dad1 / may be able to survive the implantation period through a persistence of maternally derived molecules, in a similar fashion to other genes, as discussed in various papers (Larue et al. 1994; Campbell et al. 1995; Lim & Hasty 1996; Tsuzuki et al. 1996; Ioffe et al. 1997; Murphy et al. 1997). The other possibility is that the Dad1 protein comprises the subunit of oligosaccharyltransferase, but that it is not an essential component for oligosaccharide transfer. To resolve this issue, the existence of the Dad1 protein was examined immunocytochemically q Blackwell Science Limited
by a speci®c antibody; however, this attempt proved to be unsuccessful, probably because of the highly hydrophobic nature of the protein in vivo (Nakashima et al. 1993). Conditional inactivation of the Dad1 gene will no doubt eventually lead us to the solution. Impairment of the hybrid and complex classes of N-glycans by gene targeting of the Mgat1 locus leads to embryonic lethality at about E10.5 (Ioffe & Stanley 1994; Metzler et al. 1994). Mgat1 de®ciency leads to growth retardation, neural tube abnormalities, impairment of the vascular system, in addition to loss of N-acetylglucosaminyltransferase I activity. In Dad1 / , however, death occurs at an earlier stage. Dying embryos are distinguished by their severe growth retardation or by a complete resolution of structure through apoptosis. These phenotypic differences may be explained by the oligosaccharyltransferase model, in which a glycosylated signalling molecule might be aberrantly stimulated and cause apoptosis in Dad1-, but not in Mgat1-null cells. Modi®cations to the N-linked oligosaccharides take place within the Golgi apparatus, where three different versions of a processing pathway lead, respectively, to high-mannose, hybrid, and complex oligosaccharides. Accordingly, this signal should be transmitted via the high-mannose oligosaccharide proteins which do not require modi®cations of sugar residues by Mgat1-encoded N-acetylglucosaminyltransferase I. To our knowledge, granzyme B- and Fas ligand-mediated pathways seem to be possible candidates which satisfy these criteria (Grif®ths & Isaaz 1993; Trapani et al. 1996; Orlinick et al. 1997; Tanaka et al. 1997b). We have also demonstrated that two copies of the intact Dad1 allele are needed for normal limb formation. Signi®cant numbers of live-born Dad1/ displayed soft-tissue syndactyly. We could not, however, Genes to Cells (1999) 4, 243±252
249
K Nishii et al.
detect abnormal web-removing processes through the observation of embryos at mid- and late-gestational stages (data not shown). One possibility for this is the background effect of the mouse strain we used. With the mixed or pure 129/SvJ background, we found that only one of the 58 Dad1/ mice possessed the phenotype. Even with the progressed C57Bl/6 background, where extensive backcrossings were made, the penetration of syndactyly in Dad1/ mice was about 10% (data not shown), and this relatively low frequency seems to make it dif®cult to identify the abnormal embryos. Another less favoured possibility is that the web formation may occur at later stages. Our animal model, however, should be able to provide us with a molecular basis for studying the aetiology and pathogenesis of the syndactyly phenotype. It could be supposed that the haplo-insuf®ciency condition of the `defender against apoptotic cell death' gene would promote interdigital apoptosis; however, our results were the opposite, suggesting that Dad1 may be involved in uncovered biological processes (Jacobson et al. 1997). Recently, Hong et al. reported that the mouse Dad1 locus was found to be located in close proximity to the T-cell receptor a/d constant gene, and the large deletion of the locus control region was unexpectedly found to include Dad1 exon 3, resulting in embryonic lethality (Hong et al. 1997). No homozygous progeny were reported among the viable pups, although the expected number of heterozygous progeny was presented. In contrast to this ®nding, we determined the fate of homozygous progeny. Additionally, we found that heterozygous progeny did not appear in the expected Mendelian ratio, and often displayed softtissue syndactyly. These differences can probably be attributed to a different targeting strategy: in this experiment, the disruption site is aimed at exon 1, while the reported construct is aimed at the locus control region of the T-cell receptor a/d gene including untranslated exon 3. In the latter case, it is possible that a small number of functional transcripts still remain, making heterozygous phenotypes that are somewhat milder than those we obtained. If we consider that the granzyme B cluster, which consists of granzyme B-G, cathepsin G and mouse mast cell protease 2, exists at nearly the same chromosomal location as the T-cell receptor a/d gene (and the Dad1 gene), then it is likely that genes which have some relation to apoptosis may comprise a functional complex around the Dad1 locus (Pham et al. 1996). Further investigations into the Dad1 genomic structure and function would help us to fully understand the phenotypes of the Dad1 mutant. 250
Genes to Cells (1999) 4, 243±252
Experimental procedures Targeting vector Mouse Dad1 genomic DNA (129/SvJ origin) was obtained by the standard plaque hybridization technique from the lFIX II genomic library (Stratagene), using human DAD1 as the probe (Sambrook et al. 1989). An 8.5-kb HindIII fragment contained exon 1 and 50 4.8 kb and 30 3.4 kb ¯anking sequences. The unique SacI site was blunt-ended, and the XhoI linker was inserted. A SalI-XhoI fragment of polIIneopA, the neo resistance gene driven by the RNA polymerase II promoter and followed by the polyA signal, was inserted in reverse orientation (Deng et al. 1993). This construct disrupts the Dad1 open-reading frame at the 18th amino acid position. The HindIII sites of both ends were disrupted by polymerase ®lling-in, and two thymidine kinase genes were added at the ¯anking position (Mansour et al. 1988; Rancourt et al. 1995). The targeting vector was linearized by digestion with SalI prior to electroporation.
Southern blotting To identify mutant embryonic stem cells, genomic DNA was digested with EcoRV and analysed by Southern blotting. A PstIEcoRV 0.9 kb fragment (probe B in Fig. 1A) ¯anking the 30 region detected a 12.4-kb band in the wild-type, but an 8.5-kb band in the mutant allele. To further con®rm the 50 homologous recombination, the same blot was stripped and rehybridized with a HindIII-PstI 0.6 kb fragment (probe A in Fig. 1A) that detected a 12.4-kb band in the wild-type, and a 7.0-kb band in the mutant allele. Further Southern blotting analysis con®rmed this favourable targeting event (Fig. 1B; data not shown). A successful transmission of the mutant allele for progeny was con®rmed by analysing the tail-tip DNA using the same strategy. Loaded DNA consisted of 3 mg/lane.
PCR Tail-tip DNA was simply ampli®ed by four primer mixes, dadF2 (50 -GTCATGTCGGCGTCTGTGGTGTCC), dadR2 (50 -CATTTCTGAGCCTGCTTCCTGGAT), neoF2 (50 AATCGGCTGCTCTGATGCCG), and neoR2 (50 -ATAGAAG GCGATGCGCTGCG). Embryo DNA was ampli®ed by dadF1 (50 -CCGGTATCCGAAGTCCCCGTGTTC), dadR1 (50 -CAG TTTCAACTCCTGTTAGGCATTAGAA), neoF1 (50 -GATTG CACGCAGGTTCTCCG), and neoR1 (50 -CAAGAAGGCG ATAGAAGGCG). When the ®rst ampli®cation of embryonic DNA was not suf®cient, nested PCR was accomplished by a mixture of dadF2, dadR2, neoF2 and neoR2, using the ®rst-round PCR product as a template. Both PCR conditions were the same: initial denaturation at 94 8C for 2 min, followed by variable cycles at 94 8C for 1 min, 64 8C for 1 min and 72 8C for 1 min. The number of cycles ranged from 30 (simple PCR) to a total of 60 (nested PCR), depending on DNA content. q Blackwell Science Limited
Dad1-de®cient mice
Histological examinations Pregnant mice were perfusion-®xed with 4% paraformaldehyde/ phosphate-buffered saline (PBS). After ®xation, the decidual swellings were isolated and immersed in 4% paraformaldehyde/ PBS at 4 8C overnight, dehydrated through graded ethanol series, cleared in xylene, and then in®ltrated and embedded in paraf®n. Sections were obtained serially at 5 mm thickness in sagittal orientation (Kaufman 1992; Hogan et al. 1994), and stained with haematoxylin±eosin. Some of the paramedian sections were subjected to a TUNEL assay, prepared exactly as described (Gavrieli et al. 1992).
In vitro culture of blastocysts Intercross blastocysts were generated by pairing male and female heterozygotes. Blastocysts were isolated from plugged females at E3.5 as described (Hogan et al. 1994) with M2 medium. Blastocysts were independently cultured on tissue culture plates in embryonic stem cell culture medium containing leukaemia inhibitory factor and antibiotics at 37 8C in a humidi®ed atmosphere of 5% CO2. Blastocyst outgrowths were photographed daily to monitor their development. After 3±5 days of culture, they were subjected to an annexin V-binding and propidium iodide-uptake assay, followed by a lectin-staining procedure. Finally, their genotypes were determined by PCR.
Annexin V-binding and propidium iodideuptake assay Labelling of cells with annexin V-¯uorescein isothiocyanate (FITC) and propidium iodide was performed essentially as described (van Engeland et al. 1996). The cells were washed brie¯y with cold PBS; immersed in binding buffer (10 mM HEPES/NaOH pH 7.4, 150 mM NaCl, 5 mM KCl, 1 mM MgCl2, 1.8 mM CaCl2) with annexin V-FITC (1:100 dilution; Trevigen) and 5 mg/mL propidium iodide for 15 min in the dark at room temperature. The cells were washed twice for 2 min in binding buffer at room temperature before observation using ¯uorescence microscopy.
Lectin staining All steps were carried out at room temperature. Cells were ®xed with 4% paraformaldehyde/PBS for 15 min. After washing with PBS, nonspeci®c protein binding sites were blocked by incubation in 1% bovine serum albumin (BSA)/PBS for 10 min. Biotinylated E-PHA (10 mg/mL; Sigma) in 1% BSA/PBS was applied to the cells for 1 h. After washing three times for 5 min in PBS, Texas Red-streptavidin reagent (1:150 dilution; Vector)/PBS was applied for 1 h. The cells were washed three times for 5 min in PBS before observation by ¯uorescence microscopy.
Weight analysis Progeny from heterozygous intercrosses was used 35 days after birth. After their total body weight was determined, they were q Blackwell Science Limited
sacri®ced by cervical dislocation. The tail tips were cut off for genotyping, and the thymus, spleen and kidney were weighed.
Acknowledgements We thank N. Kinoshita and Y. Tominaga for assistance with the cell culture; M. Kamizono, T. Nakamine and T. Ohmura for animal care; S. Mouri for providing 129/SvJ mice; Y. Fukumaki, H. Sasaki, A. Iwaki and Y. Takihara for technical advice; S. Kono for statistical advice; P. Gallois, T. Makishima, K. Nakamura, D.E. Rancourt and K.T. Riabowol for discussion; K. Miller for revising the English used in the manuscript; and the Human Genome Center for supplying useful software packages. This work was supported by Grants-in-Aid for Scienti®c Research from the Ministry of Education, Science and Culture, Japan.
References Campbell, R.M., Metzler, M., Granovsky, M., Dennis, J.W. & Marth, J.D. (1995) Complex asparagine-linked oligosaccharides in Mgat1-null embryos. Glycobiology 5, 535±543. Deng, C., Thomas, K.R. & Capecchi, M.R. (1993) Location of crossovers during gene targeting with insertion and replacement vectors. Mol. Cell. Biol. 13, 2134±2140. van Engeland, M., Nieland, L.J.W., Ramaekers, F.C.S., Schutte, B. & Reutelingsperger, C.P.M. (1998) Annexin V-af®nity assay: a review on an apoptosis detection system based on phosphatidylserine exposure. Cytometry 31, 1±9. van Engeland, M., Ramaekers, F.C.S., Schutte, B. & Reutelingsperger, C.P.M. (1996) A novel assay to measure loss of plasma membrane asymmetry during apoptosis of adherent cells in culture. Cytometry 24, 131±139. FaÈssler, R. & Meyer, M. (1995) Consequences of lack of b1 integrin gene expression in mice. Genes Dev. 9, 1896±1908. Feldman, B., Poueymirou, W., Papaioannou, V.E., DeChiara, T.M. & Goldfarb, M. (1995) Requirement of FGF-4 for postimplantation mouse development. Science 267, 246±249. Fu, J., Ren, M. & Kreibich, G. (1997) Interactions among subunits of the oligosaccharyltransferase complex. J. Biol. Chem. 272, 29687±29692. Gallois, P., Makishima, T., Hecht, V., et al. (1997) An Arabidopsis thaliana cDNA complementing a hamster apoptosis suppressor mutant. Plant J. 11, 1325±1331. Gavrieli, Y., Sherman, Y. & Ben-Sasson, S.A. (1992) Identi®cation of programmed cell death in situ via speci®c labeling of nuclear DNA fragmentation. J. Cell Biol. 119, 493±501. Grif®ths, G.M. & Isaaz, S. (1993) Granzymes A and B are targeted to the lytic granules of lymphocytes by the mannose6-phosphate receptor. J. Cell Biol. 120, 885±896. Hogan, B., Beddington, R., Costantini, F. & Lacy, E. (1994) Manipulating the Mouse Embryo: a Laboratory Manual, 2nd edn. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press. Hong, N.A., Cado, D., Mitchell, J., Ortiz, B.D., Hsieh, S.N. & Winoto, A. (1997) A targeted mutation at the T-cell receptor a/d locus impairs T-cell development and reveals the presence of the nearby antiapoptosis gene. Dad1. Mol. Cell. Biol. 17, 2151±2157. Ioffe, E., Liu, Y. & Stanley, P. (1997) Complex N-glycans in Mgat1 null preimplantation embryos arise from maternal Mgat1 RNA. Glycobiology 7, 913±919. Genes to Cells (1999) 4, 243±252
251
K Nishii et al. Ioffe, E. & Stanley, P. (1994) Mice lacking N-acetylglucosaminyltransferase I activity die at mid-gestation, revealing an essential role for complex or hybrid N-linked carbohydrates. Proc. Natl. Acad. Sci. USA 91, 728±732. Jacobson, M.D., Weil, M. & Raff, M.C. (1997) Programmed cell death in animal development. Cell 88, 347±354. Jones, S.N., Roe, A.E., Donehower, L.A. & Bradley, A. (1995) Rescue of embryonic lethality in Mdm2-de®cient mice by absence of p53. Nature 378, 206±208. Kaufman, M.H. (1992) The Atlas of Mouse Development. London: Academic Press. Kelleher, D.J. & Gilmore, R. (1997) DAD1, the defender against apoptotic cell death, is a subunit of the mammalian oligosaccharyltransferase. Proc. Natl. Acad. Sci. USA 94, 4994±4999. Larue, L., Ohsugi, M., Hirchenhain, J. & Kemler, R. (1994) Ecadherin null mutant embryos fail to form a trophectoderm epithelium. Proc. Natl. Acad. Sci. USA 91, 8263±8267. Lim, D.S. & Hasty, P. (1996) A mutation in mouse rad51 results in an early embryonic lethal that is suppressed by a mutation in p53. Mol. Cell. Biol. 16, 7133±7143. Liu, C.Y., Flesken-Nikitin, A., Li, S., Zeng, Y. & Lee, W.H. (1996) Inactivation of the mouse Brca1 gene leads to failure in the morphogenesis of the egg cylinder in early postimplantation development. Genes Dev. 10, 1835±1843. Makishima, T., Nakashima, T., Nagata-Kuno, K., et al. (1997) The highly conserved DAD1 protein involved in apoptosis is required for N-linked glycosylation. Genes Cells 2, 129±141. Mansour, S.L., Thomas, K.R. & Capecchi, M.R. (1988) Disruption of the proto-oncogene int-2 in mouse embryoderived stem cells: a general strategy for targeting mutations to non-selectable genes. Nature 336, 348±352. Metzler, M., Gertz, A., Sarkar, M., Schachter, H., Schrader, J.W. & Marth, J.D. (1994) Complex asparagine-linked oligosaccharides are required for morphogenic events during postimplantation development. EMBO J. 13, 2056±2065. Montes de Oca Luna, R., Wagner, D.S. & Lozano, G. (1995) Rescue of early embryonic lethality in mdm2-de®cient mice by deletion of p53. Nature 378, 203±206. Murphy, M., Stinnakre, M.G., Senamaud-Beaufort, C., et al. (1997) Delayed early embryonic lethality following disruption of the murine cyclin A2 gene. Nature Genet. 15, 83±86. Nakashima, T., Sekiguchi, T., Kuraoka, A., et al. (1993) Molecular cloning of a human cDNA encoding a novel protein, DAD1, whose defect causes apoptotic cell death in hamster BHK21 cells. Mol. Cell. Biol. 13, 6367±6374. Orlinick, J.R., Elkon, K.B. & Chao, M.V. (1997) Separate domains of the human Fas ligand dictate self-association and receptor binding. J. Biol. Chem. 272, 32221±32229. Papaioannou, V. & Johnson, R. (1993) Production of chimeras and genetically de®ned offspring from targeted ES cells. In: Gene Targeting: a Practical Approach (ed. A.L. Joyner), pp. 107±146. New York: IRL Press. Pham, C.T., MacIvor, D.M., Hug, B.A., Heusel, J.W. & Ley, T.J. (1996) Long-range disruption of gene expression by a selectable marker cassette. Proc. Natl. Acad. Sci. USA 93, 13090±13095.
252
Genes to Cells (1999) 4, 243±252
Rancourt, D.E., Tsuzuki, T. & Capecchi, M.R. (1995) Genetic interaction between hoxb-5 and hoxb-6 is revealed by nonallelic noncomplementation. Genes Dev. 9, 108±122. Sambrook, J., Fritsch, E.F. & Maniatis, T. (1989) Molecular Cloning: a Laboratory Manual, 2nd edn. Cold Spring Harbor, New York: Cold Spring Harbor Laboratory Press. Silberstein, S., Collins, P.G., Kelleher, D.J. & Gilmore, R. (1995) The essential OST2 gene encodes the 16-kD subunit of the yeast oligosaccharyltransferase, a highly conserved protein expressed in diverse eukaryotic organisms. J. Cell Biol. 131, 371±383. Silberstein, S. & Gilmore, R. (1996) Biochemistry, molecular biology, and genetics of the oligosaccharyltransferase. FASEB J. 10, 849±858. Spyropoulos, D.D. & Capecchi, M.R. (1994) Targeted disruption of the even-skipped gene, evx1, causes early postimplantation lethality of the mouse conceptus. Genes Dev. 8, 1949±1961. Stephens, L.E., Sutherland, A.E., Klimanskaya, I.V., et al. (1995) Deletion of b1 integrins in mice results in inner cell mass failure and peri-implantation lethality. Genes Dev. 9, 1883±1895. Sugimoto, A., Hozak, R.R., Nakashima, T., Nishimoto, T. & Rothman, J.H. (1995) dad-1, an endogenous programmed cell death suppressor in Caenorhabditis elegans and vertebrates. EMBO J. 14, 4434±4441. Takeda, K., Noguchi, K., Shi, W., et al. (1997) Targeted disruption of the mouse Stat3 gene leads to early embryonic lethality. Proc. Natl. Acad. Sci. USA 94, 3801±3804. Tanaka, Y., Makishima, T., Sasabe, M., et al. (1997a) dad-1, a putative programmed cell death suppressor gene in rice. Plant Cell Physiol. 38, 379±383. Tanaka, M., Suda, T., Yatomi, T., Nakamura, N. & Nagata, S. (1997b) Lethal effect of recombinant human Fas ligand in mice pretreated with Propionibacterium acnes. J. Immunol. 158, 2303±2309. Trapani, J.A., Browne, K.A., Smyth, M.J. & Jans, D.A. (1996) Localization of granzyme B in the nucleus. A putative role in the mechanism of cytotoxic lymphocyte-mediated apoptosis. J. Biol. Chem. 271, 4127±4133. Tsuzuki, T., Fujii, Y., Sakumi, K., et al. (1996) Targeted disruption of the Rad51 gene leads to lethality in embryonic mice. Proc. Natl. Acad. Sci. USA 93, 6236±6240. Wang, K., Gan, L., Kuo, C.L. & Hood, L. (1997) A highly conserved apoptotic suppressor gene is located near the chicken T-cell receptor alpha chain constant region. Immunogenetics 46, 376±382. Wurst, W. & Joyner, A.L. (1993) Production of targeted embryonic stem cell clones. In: Gene Targeting: a Practical Approach (ed. A.L. Joyner), pp. 33±61. New York: IRL Press. Xanthoudakis, S., Smeyne, R.J., Wallace, J.D. & Curran, T. (1996) The redox/DNA repair protein, Ref-1, is essential for early embryonic development in mice. Proc. Natl. Acad. Sci. USA 93, 8919±8923.
Received: 7 January 1999 Accepted: 2 March 1999
q Blackwell Science Limited
