THE JOURNAL OF EXPERIMENTAL ZOOLOGY 254286-295 (1990)
ENU-Induced Allele of Brachyury ( Tktl ) Exhibits a Developmental Lethal Phenotype Similar to the Original Brachyury ( T ) Mutation MONICA J. JUSTICE AND VERNON C. BODE Division of Biology, Kansas State University, Manhattan, Kansas 66506 ABSTRACT New alleles of brachyury (Tkt', Tk") were induced in the mouse complete tW5 haplotype by ethylnitrosourea (ENU). Like the original brachyury (2') mutation, the new alleles cause a short-tailed phenotype in heterozygotes, and interact with the t complex tail interaction factor (tct)in trans to cause phenotypically tailless mice. Because ENU is mainly a point mutagen, it is important to determine that the new alleles are homozygous embryonic lethal mutations like the original T allele, and to characterize their embryonic lethal phenotype. Moreover, the Tktl mutation maps to a n inverted position relative to quaking (qh) in t haplotypes as compared with its position on normal chromosome 17. The T"' allele was separated from the resident tW"lethal gene, tclw5,by recombination, allowing embryology studies to be performed. Embryological analyses show that the T"' allele is nearly identical to the classic T allele. At 9 and 10 days of development, homozygous TktllTktlembryos are grossly abnormal with properties including 1) irregular, disorganized somite pairs, 2) a shortened posterior end of the embryo, 3) a n irregular neural tube, and 4) an abnormal notochord. In addition, 10 day-old abnormal embryos have anterior limb buds that point dorsally rather than ventrally, and are smaller than normal littermates. We conclude that the Tkt' mutation is a valuable allele for both mapping and molecular characterization of the brachyury locus.
The brachyury ( T )mutation on mouse chromosome 17 has been studied extensively because of its effect on embryonic spinal development (Chesley, '35; Gluecksohn-Schoenheimer, '44;Gruneberg, '58; Spiegelman, '76; Yanagisawa et al., '81). Heterozygous T / + mice are short-tailed, and homozygous TIT mice die in midgestation. The original T mutation was found during an Xirradiation experiment (Dobrovolskaia-Zavadskaia, '27); however, it is uncertain whether the mutation was induced or arose spontaneously. Tissue analyses have suggested that the original T allele has a primary effect on the notochord, with secondary effects on the neural tube and somites (Chesley, '35; Gruneberg, '58). Electron microscopic analysis revealed that the basal lamina between the neural tube and notochord is absent in TIT animals, allowing contact between the two tissues (Spiegelman, '76). By culturing embryos from matings of TI+ heterozygotes in the extraembryonic coelem of chick embryos, Gluecksohn-Schoenheimer ('44)demonstrated that the allantoides of TIT embryos were shortened. She concluded that TIT embryos do not establish proper fetal-maternal circulation, resulting in early embryonic death. Numerous alleles of brachyury have been re0 1990 WILEY-LISS, INC.
ported; many were induced by X-irradiation and are probably deletions (Selby, '73; MacMurray and Shin, '88).The ThPand TtorZspontaneous mutations are large deletions (Moutier, '73; Johnson, '75; Silver et al., '83). Thus, the short-tail phenotype in heterozygotes can be a result of haploinsufficiency. Descriptive analyses of three of the deletion mutations indicate that they have a broad range of developmental anomalies, suggesting that these deletions uncover several genes affecting embryogenesis (Babiarz, '83). Two new alleles of brachyury (T"' and Tkt4) were selected in this laboratory after ethylnitrosourea (ENU) mutagenesis and screening (Justice and Bode, '86). Each was induced in the complete tW5haplotype which exhibits characteristics typical of mutant t haplotypes (reviewed by Gluecksohn-Waelsch and Erickson, '70; Bennett, '75; Klein and Hammerberg, '77; Sherman and Wudl, '77; Lyon, '81; Frischauf, '85; Silver, '81, '85). These characteristics include effects on embryonic development, male fertility, male transReceived March 23, 1989; revision accepted October 9, 1989. The current address of Dr. Monica J. Justice is Mammalian Genetics Laboratory, NCI-Frederick Cancer Research Facility, Frederick, MD 21701. Address reprint requests there.
DEVELOPMENTAL DEFECTS OF THE Tkt' ALLELE
mission ratio, and meiotic recombination. The complete t haplotypes exhibit a non-Mendelian transmission ratio from heterozygous males; thus 90-100% of the offspring from a male heterozygous for a complete t haplotype will have the t chromosome rather than its normal homolog. In addition, a complete t haplotype includes a mutation called t complex tail interaction factor (tct) which interacts with T to cause a tailless phenotype. The tW5haplotype includes this mutation as a part of the complete t haplotype. The new Tktl and TKt4mutations were induced in cis with the resident tct mutation, and cause heterozygous Tktl/+and Tkt4/+mice t o have a short tail (Justice and Bode, '86). As expected of alleles of T , both Tktlltctand Tkt4/tctmice are tailless. Each complete t haplotype contains a recessive embryonic lethal gene that can belong to one of several complementation groups. Lethal genes within complementation groups map t o distinct loci within the t complex (Artzt, '84). The Tktl allele was genetically separated from the resident lethal gene, tcZW5,of the tW5haplotype by recombination (Justice and Bode, '88a). The genetic crosses were designed t o determine the order of T and quaking ( q k ) in t haplotypes, and utilized various markers independently induced within the complete tW5haplotype by ENU (Justice and Bode, '86, '88a). For mapping, the complete $2 haplotype and the complete tW5haplotype were used in trans. Thus, a recombinant chromosome, Tktl + tell2 ~ - 2 t 1 2+ tclw5 t f k t l , was derived from a normal crossover between two complete t haplotypes (Fig. 1A; Justice and Bode, '88a), and contains the Tktl mutation within a complete t haplotype without a resident t lethal gene. The absence of a resident t lethal gene allows the initiation of embryological studies on the new Tktl allele. Although mating crosses indicated that the Tktl allele failed to complement T , we could not determine that this mutation had a homozygous embryonic lethal phenotype without first separating it from the resident lethal gene, tcZW5,of the tW5haplotype. Because ENU is mainly a point mutagen, the ENU-induced Tkt' and Tkt4 alleles may have properties that are different from the original T allele. Moreover, it is important to determine that the Tktlshort-tailed mutation is a true allele of T ; it maps t o an inverted position relative to qk in t haplotypes as compared with its position in a normal chromosome 17 (Justice and Bode, '88a). Interestingly, the fused (Fu) mutation which affects tail phenotype in heterozygotes and exhibits homozygous embryonic lethality maps t o this position in a normal chromo-
287
some 17 (Gluecksohn-Shoenheimer,'49; Dunn and Gluecksohn-Waelsch, '53). These embryological studies establish that the new Tktlallele is a lethal allele of T; the Tktlmutation fails to complement a T mutation. Furthermore, the Tktlmutation has developmental anomalies similar t o the classic allele of T . The new Tktl and Tkt4alleles are important additions t o the collection of T alleles, for they are likely t o be simple mutations t o the short-tailed phenotype. Once the T locus is cloned and analyzed molecularly, these mutations will be important to demonstrate how a simple change can detrimentally alter the locus.
MATERIALS AND METHODS Mouse strains and mutants The CBAiCaJ and C57BL16J strains, obtained from The Jackson Laboratory (Bar Harbor, Maine), were at the F23 and F24 generations of inbreeding, respectively, at Kansas State University at the time of these experiments. The BTBR T t f / + tf strain, obtained from Dr. William F. Dove (McArdle Cancer Institute, Madison, Wisconsin), was at the F8 generation of inbreeding. The T + t f / + tW5 + strain, obtained from Dr. Michael Sherman (The Roche Institute for Molecular Biology, Nutley, New Jersey), was at the F22 generation of inbreeding at the time of mutagenesis. The Tktlmutation was induced in the complete tW5haplotype (tct +tcz12 H-2tW5tcZW5 + t f ; Justice and Bode, '86). The strain used for embryology studies, + qk Tkt1 + tc112 ~ - 2 t 1 2+ tclw5 t f k t l / + T q k + stf + H-2' (see Fig. l), was maintained on a C57BL/6J background, and was at the F7 generation of inbreeding at the time of these experiments. The tufted ( t f ) and i f k t 1 mutations are homozygous recessive viable mutations that cause cycles of hair loss in homozygotes. The strain carrying the recombinant chromosome used as a control, tct + tcz12 H-2t12+ tczw5 tfkt'/ T + + t f + H-2b, was maintained by brothersister matings on a C57BL/6J background, and was at the F6 generation of inbreeding at the time of these experiments. Derivation of a recombinant chromosome lacking the tdW5and tc1" lethal genes The recombinant chromosomes already described were obtained from mapping crosses designed to determine the order of T and quaking ( q k ) in t haplotypes. A complete t12 haplotype (tct
M.J. JUSTICE AND V.C. BODE
288
A.
;.
. ,o----L_ -,
X
t
+
+
tf
+
H-2&
t
+
t
tf
+
H-Zk
0 0
+
T
X
X
ENU NT p.c.
(.
tf
t
ethylnitrosourea notochord post coitum
TktZ t f k t 1
H-Zb
. . .
. . 0 X . . . .~ +l - /
Abbrev. S tct ,fktZ
. . . . . .
..
somite
tct
+ tcZ12
Tktl
H - Z t f 2 + tclw5
+ tdZ2
tfktl
H-2t12 + tclw5 q k t f
Fig. 1. A: The mating used to derive a recombinant chromosome containing Tktf,but lacking the tc1" and tclw5 lethal genes. The X indicates where the crossover occurred. B: The mating used to derive a recombinant chromosome lacking the tc1" and tclw5 lethal genes, as well as Tkt'. The X's
indicate where the crossovers occurred; two recombination events were required to explain this recombinant chromosome (Justice and Bode, '88a). The open boxes indicate the region included in the complete t haplotype. The female is shown first in each cross.
tC112H-2t1.2 + t C l W 5 + t f ) was used heterozygous with tw5for mapping. As an initial step in generating recombinants suitable for mapping T , qlz, tufted ( t f ) , and the H-2 complex, tailless female mice, Tktl + H-Ztw5tclw5 + Itct tc112 H-Zt12 + tfktl,were mated t o tfltf males (Fig. 1A). Short-tailed, tufted recombinant animals were selected, microcytotoxicity tested for H-2 haplotype, and progeny tested for the lethal genes t d 2 and tclw5,belonging to the t12 and tw5haplotypes, respectively. The lethal genes belonging to these two haplotypes
map approximately 1.5 cM apart (Artzt, '84; Justice and Bode, '88a). Thus, it is possible to obtain recombinants that lack both tcZ12 and tcZw5,but retain the Tktl mutation within the context of a complete t haplotype. A recombinant chromosome was derived that carried the dominant shorttailed mutation, Tktl,but lacked the tclw5and tc1I2 lethal genes. In addition, this recombinant Tktl + tc112 H-2t12 + tfktl chromosome exhibited high transmission from heterozygous males (93%; N = 86). To obtain animals for the embryological
DEVELOPMENTAL DEFECTS OF THE Tkt' ALLELE
+
289
studies, short-tailed mice, + q k Tktl tcz12 H-2t12 tft'/+T qk + stf + H-2b were mated t o (C57BL/6J x CBA/CaJ)F1 normal-tailed mice. Subsequently, the short-tailed mice from this mating were intercrossed, and the embryos were examined. For clarity in genetic derivations, the Haw3' haplotype belonging to the tW5haplotype is designated H-ztw5.Likewise, the H-2w28haplotype belonging t o the t12 haplotype is designated H-2t12. The recombinant chromosome Tkt' + tcz12 H-2t'2 + tclw5 tf kt' will be referred to as Tktltf ktl throughout the paper.
scribed by Theiler ('72). If the embryo was being resorbed and somite number was not obtained, other characteristics were observed to give the approximate stage. Extreme degeneration of embryonic tissue or empty implantation sites were classified as resorbed.
Control matings Complementation between different t haplotypes is not complete (Shin et al., '83; Mains, '86). It was not clear a priori what phenotype embryos homozygous for a complete t haplotype, but lacking a t lethal gene, might exhibit. Thus, it was important t o include a control for the t haplotype genetic background in this study. Another recombinant chromosome was derived in the mapping studies that suited this purpose (Fig. 1B). The tct + td12 H-2t12 + t ~ l w 5tfktl chromosome lacks the Tkt' mutation and a resident t complex lethal gene, but has the properties of a complete t haplotype. This chromosome (tct + tc112 H-2t12 + tczw5 tfktl)was derived by recombination in tct H-.PW5 tcZw5 tf"'ltct tcZ" H-Zt" + + females (Fig. lB), and will be referred to as tct tfkt' throughout the manuscript. To obtain animals for the embryological studies, tailless mice, T +/tct tfktl, were mated to (C57BL/6J x CBA/CaJ)Fl normaltailed animals. Normal-tailed mice from this cross were intercrossed, and the embryos were examined. Again, the tct tf"' chromosome exhibits high transmission from heterozygous males (96%; N = 45).
RESULTS Abnormal Tktl/Tktlembryos have
+ tclw5
+
Embryology Females were placed with the appropriate male in the late afternoon and checked for a vaginal plug every morning over successive days. The day of the plug was designated day 0 of embryonic development. Pregnant mice were sacrificed; the uteri were removed and placed in PhosphateBuffered Saline (PBS), pH 7.4. Some embryos were removed by dissection and placed in fresh PBS for observation and photography. Other embryos were removed from the uterus and fixed and embedded with decidual tissue intact. Embryos were staged by somite pair number and external characteristics when appropriate. The developmental stages given are those previously de-
His tology Embryos were fixed in Bouin's solution for 24 hours, embedded in paraffin, and serially sectioned at a thickness of 7 or 8 microns. Sections were stained with hematoxylin and eosin.
developmental defects similar to TIT embryos Matings of Tkt' tfkt'/+ + x Tktl tfktl/+ + animals revealed no tufted animals in the offspring, indicating that a lethal gene was retained in the t haplotype in the absence of the t d 2 and tcZw5mutations. The embryos from these matings were examined t o determine the lethal phenotype. Abnormalities in embryos from matings of heterozygotes were first seen at 8-8.5 days post coitum (P.C.)(Table 1A). The abnormal embryos at this stage were detected as shortened embryos with abnormally small allantoides (Fig. 2A). By 9-9.5 days P.c., the abnormal embryos were easily distinguished from normal littermates by irregular, indistinct somite pairs, a shortened posterior end of the embryo, and an irregular neural tube (Fig. 2B). At this stage the anterior limb buds are just beginning t o form and appeared normal. Embryos observed at 10-10.5 days, however, had anterior limb buds that pointed dorsally rather than ventrally, exhibited an absence of the posterior end of the embryo, and had no distinct, organized somitic tissue (Fig. 2C,D). Posterior limb buds did not develop. All abnormal embryos at 9-10.5 days were smaller than normal littermates. Histologically, abnormal embryos at 9-10.5 days were characterized by an irregular neural tube (Fig. 3A,B) and irregular, unorganized somites. In some sections, the neuroepithelium was hypoplastic (Fig. 3C-E). In other areas, the neuroepithelium was thickened and appeared to merge with the surrounding mesenchyme. In most sections of embryos that were classified as abnormal, the notochord or chordamesoderm was not detected or appeared to have merged with the neural tube (Fig. 3D). In addition, an unusually thick layer of mesenchyme often overlaid the neural tube in embryos classified as abnormal
M.J. JUSTICE AND V.C. BODE
290
TABLE 1 . Numbers and types of embryos observed on sequential days during embryonic development' Female x Male
Day
Normal
Abnormal
Resorbed'
A. Embryological observations of litters from matings of Tkt' heterozygotes Tkt' tfkt'l+ + x Tktltfkt'/+ + 6-6.5 17 0 2 7-7.5 20 2 1 8-8.5 36 17 2 9-9.5 30 31 3 10-10.5 16 17 0 11-11.5 27 19 5 B. Embryological observations of litters from control matings + + I + + x Tktlt f k t l / + + 11-13 41 0 1 Tktltfk"l+ + x + + I + + 11-11.5 11 0 0 tct t f k t ' / _ t + x tct tfktl/+ + 8-8.5 0 0 7 9-9.5 35 0 2 10-10.5 25 0 3 11-11.5 22 0 4 13-13.5 0 4 35 C. Embryological observations of litters from matings of Tktl tfkt'l+ + with T tf/+ tf T t f / + tf x Tkt' tfktll+ + 8-8.5 9 0 1 9-9.5 4 7 0 10-10.5 12 12 0 7 3 1 11-11.5
Total
70abnormal or resorbed
19 23 55 64 33 51
11 15 35 53 52 47
42 11 7 37 28 26 39
2 0 0 5 11 15 10
10 11 24 11
10 64 50 36
+
Tk'tfk"l+ + and tct tfk"/ + males transmit the chromosome bearing TktZand tfk" at a frequency of 93-96% (see Materials and Methods), so approximately 47% abnormal homozygous embryos would be expected from any linked lethal allele in matings of heterozygotes. 'Extreme degeneration of embryonic tissue and empty implantation sites were classified as resorbed.
(Fig. 3D). All abnormal embryos observed at 11.0-11.5 days were classified as dead, assessed by lack of beating hearts. Litters observed on later days had a high number of resorbed embryos (Table 1A). Seven of 48 abnormal embryos observed on days 9-10.5 had a cleavage through the brain, as though the head folds had not closed, and one observed on day 8 had two pairs of head folds. Three abnormal embryos had enlarged pericardiums. Controls show that the defects observed are due to mutation of Tktl Two types of controls were dissected. First, Tktl tfktl/+ + heterozygotes were mated with (C57BL16J x CBA/CaJ)Fl animals (Table lB), and embryos were observed at various stages of development. Only one resorbed early implantation site was noted in 53 embryos examined. As another control, embryos from matings of tct tfktll + heterozygotes were observed to determine what abnormalities might be attributed t o homozygosity of the t chromosome (Table 1B). Thirteen abnormal embryos were observed in 138 total embryos examined. Again, the abnormal embryos observed in these matings were early implantation lethals. Implantation sites observed at 8-9.5 days contained abnormal trophectoderm;
+
on days 10-13.5, the implantation sites appeared empty or were resorbed. This semi-lethal phenotype was different from the lethal phenotype exhibited by Tktl homozygotes; the tct tfktlltct tfktl semi-lethal period occurred much earlier, probably very soon after implantation, since very little embryonic tissue could be found in these abnormal implantation sites. Obviously, when t haplotypes are homozygous, any lethal factors present cause developmental defects that are distinct from those caused by the Tktl mutation. Some of these abnormal early implantation sites were also Seen in Tktl tfktll+ + Tkt1 tfktl/+ + matings, as expected. The Tkt'mutation fails to complement a T mutation To demonstrate absence of complementation in the embryo with the classic allele of T , BTBR T tfl+ tf animals were mated with Tkl tfktl/+ + heterozygotes (Table lC), and litters were examined at various stages of development. The abnormal embryos had no detectable differences from the Tktl tfktl homozygotes: T tflTktl tfktl abnormal embryos had irregular neural tubes, indistinct somites, abnormal or absent notochords, and anterior limb buds that pointed dorsally rather than ventrally. Although few animals were observed
DEVELOPMENTAL DEFECTS OF THE Tk"ALLELE
Fig. 2. A A brightfield photomicrograph of four embryos from a dissection of a litter from a mating of Tkt' tfkl*l+ + heterozygotes at 8 days p.c. The embryo on the left is normal: note the neural development and allantois. The second embryo from the left was classified as abnormal and is in the yolk sac. The two embryos on the right are abnormal: note the small or nearly absent allantoides. The embryo on the far right appears to have abnormal head folds. B: An abnormal embryo from a litter of Tkfrt f k t l l + + heterozygotes a t 9 days
at days 8-8.5, the dorsal blebs observed in TIT homozygous abnormal embryos were not observed in T +lTkt' tfkt2abnormal embryos.
291
p.c. The arrow points to the irregular neural tube. C: A darkfield photomicrograph of an abnormal homozygous Tkt'I TktJembryo on the right with a normal day 10 littermate on the left. Note the absence of somites in the abnormal embryo, the absence of the posterior end, and the smaller size. D: A different abnormal homozygous Tkt'/Tkt'embryo from the same litter as C. Note the abnormal hyperplasia of cells on the dorsal posterior surface of the closed neural tube.
of brachyury (TI, and exhibit similar features. Both the original T allele and the ENU-induced Tkt' allele have abnormal, diffuse somitic tissue, absence of the posterior end of the embryo, anteDISCUSSION rior limb buds that point dorsally rather than ventrally, irregular neural tubes, shortened alThe results show that the new Tktlallele of T is lantoides, and abnormalities of the notochord. a recessive embryonic lethal mutation. We observed a total of 84 abnormal embryos from days However, three differences between the alleles 8-11.5 that exhibited the abnormal phenotype. In were noted: 1) TIT homozygotes had transient a total of 203 embryos, this is not significantly dorsal blebs at 8-8.5 days (Chesley, '35; Spiegeldifferent from the expected value of 94 (x2 = man, ,761, whereas TktllTktlhomozygotes did not; 1.98), given a male transmission ratio of 93%. 2) abnormal TIT embryos were the same size as (Note that about 47% abnormal embryos are ex- normal littermates (Chesley, '35; Spiegelman, heterozy- ,761, except for the shortened posterior end, pected from matings of Tkt' tfkt'l+ gotes because of high male transmission ratios.) whereas abnormal T"' ITkt1embryos were shortHomozygous Tkt' embryos are abnormal at the ened or retarded at 8 days and smaller than norsame stage of development as the original allele mal littermates at later developmental stages;
+
292
M.J. JUSTICE AND V.C. BODE
Fig. 3. A, B: Frontal sections taken from the dorsal side of embryos from Tktl t f k t l l + + x Tktl t f h t ' / + + matings on day 9.5 p.c. A: A normal embryo with normal well-defined, organized somite pairs, and a normal neural tube. B: An abnormal embryo with unorganized, irregular somite pairs, and an irregular neural tube. C-E: Transverse sections of embryos from Tkt' tfkt'l+ + heterozygotes on day 9 p.c. C: A normal neural tube (NT) and notochord a t the level of the
pharynx. The arrow points to the normal notochord. D: A similar section of an abnormal embryo. Note the absence of the notochord, the abnormal placement of the somite (S), the abnormal neural tube, and abnormally thick layer of mesenchyme over the neural tube. E: A section from the level of the hind gut of a normal embryo showing normal somite and notochord.
and 3) some abnormal TIT embryos have exhibited a neural hyperplasia or neural tube duplication (Spiegelman, '76; Cogliatti '86), whereas this apparent duplication was not observed in TktllTktl homozygotes. The neural hyperplasia has been interpreted differently by investigators as appearing as though chordamesoderm had differentiated into neural tissue (Spiegelman, '761, or caudal duplication of the neural tube (Cogliatti, '86). The neural hyperplasia has been observed in both TI + and TIT embryos, but has exhibited incomplete penetrance (Spiegelman, '76; Cogliatti, '86). Two TktllTktlembryos observed at day 9-9.5 had what
could be interpreted as a second rudimentary neural tube near the caudal end of the embryo; however, it was not as distinct as what has been previously reported, and occurred at a very low frequency. Thus, the neural hyperplasia appears not to be as common in the Tkt11Tkt2 embryos as in TIT embryos. The lack of dorsal blebs and lack of neural hyperplasia in TktllTktlembryos may be because the Tktl mutation is less detrimental to the embryos than the T mutation. For example, the T mutation may be a small deletion, and an additional locus included in the deleted region could be re-
DEVELOPMENTAL DEFECTS OF THE T"" ALLELE
sponsible for the developmental differences. Alternatively, the observed dissimilarities may be due t o differences in genetic background of the studies. Once the T locus is molecularly cloned, molecular analyses combined with the embryological observations may determine how these mutations have a slightly different effect on embryos. The absence or partial absence of a notochord was noted in this study and in studies on the classic T allele (Chesley, '35; Gruneberg, '58; Spiegelman, '76). The development of the neural tube has been correlated with the presence of a notochord (Spemann and Mangold, '24), and Chesley ('35) speculated that the notochordal defect was primary and the neural tube defects were secondary. Spiegelman ('76) subsequently showed that the basal lamina was absent between the neural tube and notochord, allowing cell contacts that do not normally occur. Thus, the abnormal notochord in TIT homozygotes may be a result of the abnormal cell interactions, or lack of some other factor, and not a primary defect. The presence or absence of a basal lamina surrounding the neural tube and the notochord was not examined in Tkt' homozygotes. However, the neuroepithelium that was often thickened and merged with surrounding mesenchyme observed in TktlITkt1 homozygotes is consistent with what would be expected if the basal lamina surrounding the neural tube was defective. These embryological studies were carried out on a mutation induced within a t haplotype, demonstrating that new embryonic lethal mutations induced within t haplotypes can be examined embryologically if they are genetically separated from resident lethal genes and if appropriate controls are included in the study. Obviously, when the t I 2 and tW5 haplotypes are homozygous, any lethal factors other than the resident embryonic lethal genes cause defects that are distinct from those caused by the Tktl mutation. Other mutations induced by ENU mutagenesis in a t haplotype have been genetically separated from resident lethal alleles, allowing embryology studies to be performed (Justice and Bode, unpublished results). For example, new lethal alleles of quaking (qk"' and qk"') were induced in the tW5haplotype. Before the ENU mutagenesis study, one viable, myelin-deficient qk allele identified the quaking locus. Four independent ENU-induced alleles of quaking are recessive embryonic lethal mutations (Justice and Bode, '88b; Shedlovsky et al., '88). Thus, early embryonic death is an impor-
293
tant characteristic of the qk locus (Justice and Bode, '88b). The lethal qk"' allele has been genetically separated from the resident tcZ1' and tcZW5 lethal genes during the mapping studies; however, initial studies indicate that the qk"' allele does not exhibit as distinctive a lethal phenotype as the Tktl mutation (Justice and Bode, '88b). It will be essential to differentiate the effects of a t haplotype genetic background from the effects of the mutant gene in embryological studies of the qkktl allele. These studies on the Tkt' allele thus demonstrate the potential to study the effect of other mutations induced in a t haplotype on embryogenesis. ENU likely causes single base changes, whereas X-rays commonly cause deletions which may involve a large part of one gene or several genes (Vogel and Natarjan, '79; Peters et al., '85; Popp et al., '83;Russell, '82; Ehling and Favor, '84). Thus, a point mutation more probably exhibits a phenotype that reflects a simple change at a locus, whereas an X-ray-induced mutation may reflect more complex changes. It is important t o have numerous alleles of T t o analyze the cellular function of the gene product of the brachyury locus. Interactions of the various alleles at the T locus suggest an apparent gradient of T gene function associated with gene dosage along the length of the body axis (MacMurray and Shin, '88). Once the locus is cloned, it will be useful t o have presumed simple mutations to the shorttailed phenotype to analyze this gradient hypothesis. It is possible that tct is the most simple mutation at this locus, and T is a more severe change. In another experiment, V. Bode has examined 3,589 F1 gametes after ENU mutagenesis of spermatogonial stem cells for a mutation t o brachyury in a normal chromosome 17 without success (V. Bode, unpublished results). In addition, over 7,000 F1 animals carrying a normal chromosome 17 were observed for a mutation to the short-tailed phenotype in other ENU mutagenesis screens in this laboratory (J. McDonald and V. Bode, unpublished results). No heritable mutations to a dominant abnormal tail phenotype were found. One brachyury mutation (TRP)was found in an ENU mutagenesis screen in another laboratory; however, it was not clear whether TRP was induced by ENU or occurred spontaneously (Mann et al., '87). Two independent mutations to brachyury were obtained in a t haplotype in 4,869 gametes screened after ENU mutagenesis (Justice and Bode, '86). Furthermore, a tct mutation was induced in a normal
294
M.J. JUSTICE AND V.C. BODE
chromosome 17 by ENU in 2,215 gametes screened (Bode, '84).It is possible that brachyury is difficult to induce with a point mutagen, and that the presence of the tct mutation within t haplotypes makes mutation t o the short-tailed phenotype more probable in a t haplotype as opposed to a normal chromosome 17. Molecular studies are needed t o resolve this disparate observation and t o determine the developmental function of the wild-type T locus. The ENU-induced alleles of brachyury (Tkt' and Tk"4,,along with the spontaneous and irradiation-induced alleles of brachyury, will be valuable in future molecular studies of this intriguing locus.
ACKNOWLEDGMENTS We thank Nancy Hedrick for excellent technical assistance. We also thank Dr. Linda Siracusa, Dr. Peter Donovan, and an anonymous reviewer for the critical reading of this manuscript. This work was supported by grant HD 15354 from the National Institutes of Child Health and Human Development t o V.C.B. LITERATURE CITED Artzt, K. (1984) Gene mapping within the Tit complex of the mouse. 111: t-lethal genes are arranged in three clusters on chromosome 17. Cell, 39:565-572. Babiarz, B. (1983) Deletion mapping of the Tit complex: Evidence for a second region of critical embryonic genes. Dev. Biol., 95:342-351. Bennett, D. (1975) The t-locus of the mouse. Cell, 6:441-454. Bode, V. (1984) Ethylnitrosourea mutagenesis and the isolation of mutant alleles for specific genes located in the t region of mouse chromosome 17. Genetics, 108:457-470. Chesley, P. (1935) The development of the short-tailed mutant in the house mouse. J . Exp. Zool., 70:429-459. Cogliatti, S. (1986) Diplomyelia: Caudal duplication of the neural tube in mice. Teratology, 34:343-352. Dobrovolskaia-Zavadskaia, N. (1927) Sur la mortification spontanee de la queue chez la souris nouveau-nee et sur l'existence d'un caractere facteur hereditaire, non viable, C. R. SOC.Biol. (Paris), 97:114-119. Dunn, L.C., and S. Gluecksohn-Waelsch (1953) A genetical study of the mutation 'Fused' in the house mouse, with evidence concerning its allelism with a similar mutation 'Kink'. J. Genet., 52:383-391. Ehling, U.H., and J . Favor (1984) Recessive and dominant mutations in mice. In: Mutation, Cancer, and Malformation. E.H.Y. Chu and W.M. Generoso, eds. Plenum Press, New York, pp. 389-428. Frischauf, A.-M. (1985) The Tit complex of the mouse. Trends Genet., 4:lOO-103. Gluecksohn-Schoenheimer, S. (1944) The development of normal and homozygous Brachy (TIT) mouse embryos in the extraembryonic coelom of the chick. Proc. Natl. Acad. Sci. USA, 30:134-140. Gluecksohn-Schoenheimer, S. (1949) The effects of a lethal mutation responsible for duplications and twinning in mouse embryos. J . Exp. Zool., 110:47-76.
Gluecksohn-Waelsch, S., and R.P. Erickson (1970) The t-locus of the mouse: Implications for mechanisms of development. In: Current Topics in Development Biology. A.A. Moscona and A. Monroy, eds. Academic Press, New York, pp. 281315. Gruneberg, H. (1958) Genetical studies on the skeleton of the mouse. XXIII. The development of Brachyury and Anury. J. Embryol. Exp. Morphol., 6:424-443. Johnson, D.R. (1975) Further observations on the hairpintail (Thp)mutation in the mouse. Genet. Res., 24:207-213. Justice, M.J., and V.C. Bode (1986) Induction of new mutations in a mouse t-haplotype using ethylnitrosourea mutagenesis. Genet. Res., 47:187-192. Justice, M.J., and V.C. Bode (1988a) Genetic analysis of mouse t-haplotypes using mutations induced by ethylnitrosourea mutagenesis: The order of T and qk is inverted in t mutants. Genetics, 120:533-543. Justice, M.J., and V.C. Bode (198%) Three ENU-induced alleles of the murine quaking locus are recessive embryonic lethal mutations. Genet. Res., 51:95-102. Klein, J., and C. Hammerberg (1977) The control of differentiation by the T complex. Immunol. Rev., 33:70-104. Lyon, M.F. (1981) The t-complex and the genetic control of development. In: Biology of the House Mouse. R.J. Berry, ed. Academic Press, London, pp. 455-477. MacMurray, A,, and H.-S. Shin (1988) The antimorphic nature of the T' allele at the mouse T locus. Genetics, 120: 545-550. Mains, P.E. (1986) The cis-trans test shows no evidence for a functional relationship between two mouse t complex lethal mutations: Implications for the evolution of t haplotypes. Genetics, 114:1225-1237. Mann, E., R. Benz, D. Swiatek, and V.M. Chapman (1987) Recovery of a new allele at the T locus on chromosome 17. Mouse News Lett., 78:76. Moutier, R. (1973) Mouse News Lett., 48:38. Peters, J., S.J. Andrews, J.F. Loutit, and J.B. Clegg (1985) A mouse p-globin mutant that is an exact model of hemoglobin Rainier in man. Genetics, 110:709-721. Popp, R.A., E.G. Bailiff, L.C Skow, F.M. Johnson and S.E. Lewis (1983) Analysis of a mouse a-globin gene mutation induced by ethylnitrosourea. Genetics, 105:157-167. Russell, W.L. (1982) Factors affecting mutagenicity of ethylnitrousourea in the mouse specific-locustest and their bearing on risk estimation. In: Environmental Mutagens and Carcinogens: Proceedings of the Third International Conference on Environmental Mutagens. T. Sugimura, s. Kondo, and H. Takebe, eds. Alan R. Liss, Inc., New York, pp. 59-70. Selby, P. (1973) X-ray induced specific-locus mutation rates in newborn male mice. Mutat. Res., 18:63-75. Shedlovsky, A., T.R. King, and W.F. Dove, (1988) Saturation germ line mutagenesis of the murine t region including a lethal allele at the quaking locus. Proc. Natl. Acad. Sci. USA, 85:180-184. Sherman, M.I., and L.R. Wudl, (1977) T-complex mutations and their effects. In: Concepts in Mammalian Embryogenesis. M.I. Sherman, ed. MIT Press, Cambridge, pp. 136-234. Shin, H.-S., P. McCormick, K. Artzt, and D. Bennett (1983) Cis-trans test shows a functional relationship between nonallelic lethal mutations in the Tit complex. Cell, 33:925929. Silver, L.M. (1981) Genetic organization of the mouse t complex. Cell, 278:239-240.
DEVELOPMENTAL DEFECTS OF THE Tkt' ALLELE Silver, L.M. (1985) Mouse t haplotypes. Annu. Rev. Genet., 19:179-208. Silver, L.M., D. Lukralle, and J.I. Garrels (1983) To" is a novel, variant form of mouse chromosome 17 with a deletion in a partial t-haplotype. Nature, 301:422-424. Spemann, H., and H. Mangold (1924) Uber induktion von embryonalanlagen durch implantation artfremder organisatoren. Wilhelm Roux Arch. Entwickl. Mech. Org., 100: 599-638. Spiegelman, M. (1976) Electron microscopy of cell associations in T-locus mutants. In: Embryogenesis in Mammals. K. Elliott and M. O'Connor, eds. ElsevieriNorth Holland, Amsterdam, pp. 199-226.
295
Theiler, K. (1972) The House Mouse. Springer-Verlag, New York. Vogel, E., and A.T. Natarjan (1979) The relation between reaction kinetics and mutagenic action of mono-functional alkylating agents in higher eukaryotic systems. 1. Recessive lethal mutations and translocations in Drosophila. Mutat. Res., 62:51-100. Yanagisawa, K.O., H. Fujimoto, and H. Urushihara (1981) Effects of t h e Brachyury (2') mutation on morphogenetic movement i n the mouse embryo. Dev. Biol., 87:242-248.
ENU-Induced Allele of Brachyury ( Tktl ) Exhibits a Developmental Lethal Phenotype Similar to the Original Brachyury ( T ) Mutation MONICA J. JUSTICE AND VERNON C. BODE Division of Biology, Kansas State University, Manhattan, Kansas 66506 ABSTRACT New alleles of brachyury (Tkt', Tk") were induced in the mouse complete tW5 haplotype by ethylnitrosourea (ENU). Like the original brachyury (2') mutation, the new alleles cause a short-tailed phenotype in heterozygotes, and interact with the t complex tail interaction factor (tct)in trans to cause phenotypically tailless mice. Because ENU is mainly a point mutagen, it is important to determine that the new alleles are homozygous embryonic lethal mutations like the original T allele, and to characterize their embryonic lethal phenotype. Moreover, the Tktl mutation maps to a n inverted position relative to quaking (qh) in t haplotypes as compared with its position on normal chromosome 17. The T"' allele was separated from the resident tW"lethal gene, tclw5,by recombination, allowing embryology studies to be performed. Embryological analyses show that the T"' allele is nearly identical to the classic T allele. At 9 and 10 days of development, homozygous TktllTktlembryos are grossly abnormal with properties including 1) irregular, disorganized somite pairs, 2) a shortened posterior end of the embryo, 3) a n irregular neural tube, and 4) an abnormal notochord. In addition, 10 day-old abnormal embryos have anterior limb buds that point dorsally rather than ventrally, and are smaller than normal littermates. We conclude that the Tkt' mutation is a valuable allele for both mapping and molecular characterization of the brachyury locus.
The brachyury ( T )mutation on mouse chromosome 17 has been studied extensively because of its effect on embryonic spinal development (Chesley, '35; Gluecksohn-Schoenheimer, '44;Gruneberg, '58; Spiegelman, '76; Yanagisawa et al., '81). Heterozygous T / + mice are short-tailed, and homozygous TIT mice die in midgestation. The original T mutation was found during an Xirradiation experiment (Dobrovolskaia-Zavadskaia, '27); however, it is uncertain whether the mutation was induced or arose spontaneously. Tissue analyses have suggested that the original T allele has a primary effect on the notochord, with secondary effects on the neural tube and somites (Chesley, '35; Gruneberg, '58). Electron microscopic analysis revealed that the basal lamina between the neural tube and notochord is absent in TIT animals, allowing contact between the two tissues (Spiegelman, '76). By culturing embryos from matings of TI+ heterozygotes in the extraembryonic coelem of chick embryos, Gluecksohn-Schoenheimer ('44)demonstrated that the allantoides of TIT embryos were shortened. She concluded that TIT embryos do not establish proper fetal-maternal circulation, resulting in early embryonic death. Numerous alleles of brachyury have been re0 1990 WILEY-LISS, INC.
ported; many were induced by X-irradiation and are probably deletions (Selby, '73; MacMurray and Shin, '88).The ThPand TtorZspontaneous mutations are large deletions (Moutier, '73; Johnson, '75; Silver et al., '83). Thus, the short-tail phenotype in heterozygotes can be a result of haploinsufficiency. Descriptive analyses of three of the deletion mutations indicate that they have a broad range of developmental anomalies, suggesting that these deletions uncover several genes affecting embryogenesis (Babiarz, '83). Two new alleles of brachyury (T"' and Tkt4) were selected in this laboratory after ethylnitrosourea (ENU) mutagenesis and screening (Justice and Bode, '86). Each was induced in the complete tW5haplotype which exhibits characteristics typical of mutant t haplotypes (reviewed by Gluecksohn-Waelsch and Erickson, '70; Bennett, '75; Klein and Hammerberg, '77; Sherman and Wudl, '77; Lyon, '81; Frischauf, '85; Silver, '81, '85). These characteristics include effects on embryonic development, male fertility, male transReceived March 23, 1989; revision accepted October 9, 1989. The current address of Dr. Monica J. Justice is Mammalian Genetics Laboratory, NCI-Frederick Cancer Research Facility, Frederick, MD 21701. Address reprint requests there.
DEVELOPMENTAL DEFECTS OF THE Tkt' ALLELE
mission ratio, and meiotic recombination. The complete t haplotypes exhibit a non-Mendelian transmission ratio from heterozygous males; thus 90-100% of the offspring from a male heterozygous for a complete t haplotype will have the t chromosome rather than its normal homolog. In addition, a complete t haplotype includes a mutation called t complex tail interaction factor (tct) which interacts with T to cause a tailless phenotype. The tW5haplotype includes this mutation as a part of the complete t haplotype. The new Tktl and TKt4mutations were induced in cis with the resident tct mutation, and cause heterozygous Tktl/+and Tkt4/+mice t o have a short tail (Justice and Bode, '86). As expected of alleles of T , both Tktlltctand Tkt4/tctmice are tailless. Each complete t haplotype contains a recessive embryonic lethal gene that can belong to one of several complementation groups. Lethal genes within complementation groups map t o distinct loci within the t complex (Artzt, '84). The Tktl allele was genetically separated from the resident lethal gene, tcZW5,of the tW5haplotype by recombination (Justice and Bode, '88a). The genetic crosses were designed t o determine the order of T and quaking ( q k ) in t haplotypes, and utilized various markers independently induced within the complete tW5haplotype by ENU (Justice and Bode, '86, '88a). For mapping, the complete $2 haplotype and the complete tW5haplotype were used in trans. Thus, a recombinant chromosome, Tktl + tell2 ~ - 2 t 1 2+ tclw5 t f k t l , was derived from a normal crossover between two complete t haplotypes (Fig. 1A; Justice and Bode, '88a), and contains the Tktl mutation within a complete t haplotype without a resident t lethal gene. The absence of a resident t lethal gene allows the initiation of embryological studies on the new Tktl allele. Although mating crosses indicated that the Tktl allele failed to complement T , we could not determine that this mutation had a homozygous embryonic lethal phenotype without first separating it from the resident lethal gene, tcZW5,of the tW5haplotype. Because ENU is mainly a point mutagen, the ENU-induced Tkt' and Tkt4 alleles may have properties that are different from the original T allele. Moreover, it is important to determine that the Tktlshort-tailed mutation is a true allele of T ; it maps t o an inverted position relative to qk in t haplotypes as compared with its position in a normal chromosome 17 (Justice and Bode, '88a). Interestingly, the fused (Fu) mutation which affects tail phenotype in heterozygotes and exhibits homozygous embryonic lethality maps t o this position in a normal chromo-
287
some 17 (Gluecksohn-Shoenheimer,'49; Dunn and Gluecksohn-Waelsch, '53). These embryological studies establish that the new Tktlallele is a lethal allele of T; the Tktlmutation fails to complement a T mutation. Furthermore, the Tktlmutation has developmental anomalies similar t o the classic allele of T . The new Tktl and Tkt4alleles are important additions t o the collection of T alleles, for they are likely t o be simple mutations t o the short-tailed phenotype. Once the T locus is cloned and analyzed molecularly, these mutations will be important to demonstrate how a simple change can detrimentally alter the locus.
MATERIALS AND METHODS Mouse strains and mutants The CBAiCaJ and C57BL16J strains, obtained from The Jackson Laboratory (Bar Harbor, Maine), were at the F23 and F24 generations of inbreeding, respectively, at Kansas State University at the time of these experiments. The BTBR T t f / + tf strain, obtained from Dr. William F. Dove (McArdle Cancer Institute, Madison, Wisconsin), was at the F8 generation of inbreeding. The T + t f / + tW5 + strain, obtained from Dr. Michael Sherman (The Roche Institute for Molecular Biology, Nutley, New Jersey), was at the F22 generation of inbreeding at the time of mutagenesis. The Tktlmutation was induced in the complete tW5haplotype (tct +tcz12 H-2tW5tcZW5 + t f ; Justice and Bode, '86). The strain used for embryology studies, + qk Tkt1 + tc112 ~ - 2 t 1 2+ tclw5 t f k t l / + T q k + stf + H-2' (see Fig. l), was maintained on a C57BL/6J background, and was at the F7 generation of inbreeding at the time of these experiments. The tufted ( t f ) and i f k t 1 mutations are homozygous recessive viable mutations that cause cycles of hair loss in homozygotes. The strain carrying the recombinant chromosome used as a control, tct + tcz12 H-2t12+ tczw5 tfkt'/ T + + t f + H-2b, was maintained by brothersister matings on a C57BL/6J background, and was at the F6 generation of inbreeding at the time of these experiments. Derivation of a recombinant chromosome lacking the tdW5and tc1" lethal genes The recombinant chromosomes already described were obtained from mapping crosses designed to determine the order of T and quaking ( q k ) in t haplotypes. A complete t12 haplotype (tct
M.J. JUSTICE AND V.C. BODE
288
A.
;.
. ,o----L_ -,
X
t
+
+
tf
+
H-2&
t
+
t
tf
+
H-Zk
0 0
+
T
X
X
ENU NT p.c.
(.
tf
t
ethylnitrosourea notochord post coitum
TktZ t f k t 1
H-Zb
. . .
. . 0 X . . . .~ +l - /
Abbrev. S tct ,fktZ
. . . . . .
..
somite
tct
+ tcZ12
Tktl
H - Z t f 2 + tclw5
+ tdZ2
tfktl
H-2t12 + tclw5 q k t f
Fig. 1. A: The mating used to derive a recombinant chromosome containing Tktf,but lacking the tc1" and tclw5 lethal genes. The X indicates where the crossover occurred. B: The mating used to derive a recombinant chromosome lacking the tc1" and tclw5 lethal genes, as well as Tkt'. The X's
indicate where the crossovers occurred; two recombination events were required to explain this recombinant chromosome (Justice and Bode, '88a). The open boxes indicate the region included in the complete t haplotype. The female is shown first in each cross.
tC112H-2t1.2 + t C l W 5 + t f ) was used heterozygous with tw5for mapping. As an initial step in generating recombinants suitable for mapping T , qlz, tufted ( t f ) , and the H-2 complex, tailless female mice, Tktl + H-Ztw5tclw5 + Itct tc112 H-Zt12 + tfktl,were mated t o tfltf males (Fig. 1A). Short-tailed, tufted recombinant animals were selected, microcytotoxicity tested for H-2 haplotype, and progeny tested for the lethal genes t d 2 and tclw5,belonging to the t12 and tw5haplotypes, respectively. The lethal genes belonging to these two haplotypes
map approximately 1.5 cM apart (Artzt, '84; Justice and Bode, '88a). Thus, it is possible to obtain recombinants that lack both tcZ12 and tcZw5,but retain the Tktl mutation within the context of a complete t haplotype. A recombinant chromosome was derived that carried the dominant shorttailed mutation, Tktl,but lacked the tclw5and tc1I2 lethal genes. In addition, this recombinant Tktl + tc112 H-2t12 + tfktl chromosome exhibited high transmission from heterozygous males (93%; N = 86). To obtain animals for the embryological
DEVELOPMENTAL DEFECTS OF THE Tkt' ALLELE
+
289
studies, short-tailed mice, + q k Tktl tcz12 H-2t12 tft'/+T qk + stf + H-2b were mated t o (C57BL/6J x CBA/CaJ)F1 normal-tailed mice. Subsequently, the short-tailed mice from this mating were intercrossed, and the embryos were examined. For clarity in genetic derivations, the Haw3' haplotype belonging to the tW5haplotype is designated H-ztw5.Likewise, the H-2w28haplotype belonging t o the t12 haplotype is designated H-2t12. The recombinant chromosome Tkt' + tcz12 H-2t'2 + tclw5 tf kt' will be referred to as Tktltf ktl throughout the paper.
scribed by Theiler ('72). If the embryo was being resorbed and somite number was not obtained, other characteristics were observed to give the approximate stage. Extreme degeneration of embryonic tissue or empty implantation sites were classified as resorbed.
Control matings Complementation between different t haplotypes is not complete (Shin et al., '83; Mains, '86). It was not clear a priori what phenotype embryos homozygous for a complete t haplotype, but lacking a t lethal gene, might exhibit. Thus, it was important t o include a control for the t haplotype genetic background in this study. Another recombinant chromosome was derived in the mapping studies that suited this purpose (Fig. 1B). The tct + td12 H-2t12 + t ~ l w 5tfktl chromosome lacks the Tkt' mutation and a resident t complex lethal gene, but has the properties of a complete t haplotype. This chromosome (tct + tc112 H-2t12 + tczw5 tfktl)was derived by recombination in tct H-.PW5 tcZw5 tf"'ltct tcZ" H-Zt" + + females (Fig. lB), and will be referred to as tct tfkt' throughout the manuscript. To obtain animals for the embryological studies, tailless mice, T +/tct tfktl, were mated to (C57BL/6J x CBA/CaJ)Fl normaltailed animals. Normal-tailed mice from this cross were intercrossed, and the embryos were examined. Again, the tct tf"' chromosome exhibits high transmission from heterozygous males (96%; N = 45).
RESULTS Abnormal Tktl/Tktlembryos have
+ tclw5
+
Embryology Females were placed with the appropriate male in the late afternoon and checked for a vaginal plug every morning over successive days. The day of the plug was designated day 0 of embryonic development. Pregnant mice were sacrificed; the uteri were removed and placed in PhosphateBuffered Saline (PBS), pH 7.4. Some embryos were removed by dissection and placed in fresh PBS for observation and photography. Other embryos were removed from the uterus and fixed and embedded with decidual tissue intact. Embryos were staged by somite pair number and external characteristics when appropriate. The developmental stages given are those previously de-
His tology Embryos were fixed in Bouin's solution for 24 hours, embedded in paraffin, and serially sectioned at a thickness of 7 or 8 microns. Sections were stained with hematoxylin and eosin.
developmental defects similar to TIT embryos Matings of Tkt' tfkt'/+ + x Tktl tfktl/+ + animals revealed no tufted animals in the offspring, indicating that a lethal gene was retained in the t haplotype in the absence of the t d 2 and tcZw5mutations. The embryos from these matings were examined t o determine the lethal phenotype. Abnormalities in embryos from matings of heterozygotes were first seen at 8-8.5 days post coitum (P.C.)(Table 1A). The abnormal embryos at this stage were detected as shortened embryos with abnormally small allantoides (Fig. 2A). By 9-9.5 days P.c., the abnormal embryos were easily distinguished from normal littermates by irregular, indistinct somite pairs, a shortened posterior end of the embryo, and an irregular neural tube (Fig. 2B). At this stage the anterior limb buds are just beginning t o form and appeared normal. Embryos observed at 10-10.5 days, however, had anterior limb buds that pointed dorsally rather than ventrally, exhibited an absence of the posterior end of the embryo, and had no distinct, organized somitic tissue (Fig. 2C,D). Posterior limb buds did not develop. All abnormal embryos at 9-10.5 days were smaller than normal littermates. Histologically, abnormal embryos at 9-10.5 days were characterized by an irregular neural tube (Fig. 3A,B) and irregular, unorganized somites. In some sections, the neuroepithelium was hypoplastic (Fig. 3C-E). In other areas, the neuroepithelium was thickened and appeared to merge with the surrounding mesenchyme. In most sections of embryos that were classified as abnormal, the notochord or chordamesoderm was not detected or appeared to have merged with the neural tube (Fig. 3D). In addition, an unusually thick layer of mesenchyme often overlaid the neural tube in embryos classified as abnormal
M.J. JUSTICE AND V.C. BODE
290
TABLE 1 . Numbers and types of embryos observed on sequential days during embryonic development' Female x Male
Day
Normal
Abnormal
Resorbed'
A. Embryological observations of litters from matings of Tkt' heterozygotes Tkt' tfkt'l+ + x Tktltfkt'/+ + 6-6.5 17 0 2 7-7.5 20 2 1 8-8.5 36 17 2 9-9.5 30 31 3 10-10.5 16 17 0 11-11.5 27 19 5 B. Embryological observations of litters from control matings + + I + + x Tktlt f k t l / + + 11-13 41 0 1 Tktltfk"l+ + x + + I + + 11-11.5 11 0 0 tct t f k t ' / _ t + x tct tfktl/+ + 8-8.5 0 0 7 9-9.5 35 0 2 10-10.5 25 0 3 11-11.5 22 0 4 13-13.5 0 4 35 C. Embryological observations of litters from matings of Tktl tfkt'l+ + with T tf/+ tf T t f / + tf x Tkt' tfktll+ + 8-8.5 9 0 1 9-9.5 4 7 0 10-10.5 12 12 0 7 3 1 11-11.5
Total
70abnormal or resorbed
19 23 55 64 33 51
11 15 35 53 52 47
42 11 7 37 28 26 39
2 0 0 5 11 15 10
10 11 24 11
10 64 50 36
+
Tk'tfk"l+ + and tct tfk"/ + males transmit the chromosome bearing TktZand tfk" at a frequency of 93-96% (see Materials and Methods), so approximately 47% abnormal homozygous embryos would be expected from any linked lethal allele in matings of heterozygotes. 'Extreme degeneration of embryonic tissue and empty implantation sites were classified as resorbed.
(Fig. 3D). All abnormal embryos observed at 11.0-11.5 days were classified as dead, assessed by lack of beating hearts. Litters observed on later days had a high number of resorbed embryos (Table 1A). Seven of 48 abnormal embryos observed on days 9-10.5 had a cleavage through the brain, as though the head folds had not closed, and one observed on day 8 had two pairs of head folds. Three abnormal embryos had enlarged pericardiums. Controls show that the defects observed are due to mutation of Tktl Two types of controls were dissected. First, Tktl tfktl/+ + heterozygotes were mated with (C57BL16J x CBA/CaJ)Fl animals (Table lB), and embryos were observed at various stages of development. Only one resorbed early implantation site was noted in 53 embryos examined. As another control, embryos from matings of tct tfktll + heterozygotes were observed to determine what abnormalities might be attributed t o homozygosity of the t chromosome (Table 1B). Thirteen abnormal embryos were observed in 138 total embryos examined. Again, the abnormal embryos observed in these matings were early implantation lethals. Implantation sites observed at 8-9.5 days contained abnormal trophectoderm;
+
on days 10-13.5, the implantation sites appeared empty or were resorbed. This semi-lethal phenotype was different from the lethal phenotype exhibited by Tktl homozygotes; the tct tfktlltct tfktl semi-lethal period occurred much earlier, probably very soon after implantation, since very little embryonic tissue could be found in these abnormal implantation sites. Obviously, when t haplotypes are homozygous, any lethal factors present cause developmental defects that are distinct from those caused by the Tktl mutation. Some of these abnormal early implantation sites were also Seen in Tktl tfktll+ + Tkt1 tfktl/+ + matings, as expected. The Tkt'mutation fails to complement a T mutation To demonstrate absence of complementation in the embryo with the classic allele of T , BTBR T tfl+ tf animals were mated with Tkl tfktl/+ + heterozygotes (Table lC), and litters were examined at various stages of development. The abnormal embryos had no detectable differences from the Tktl tfktl homozygotes: T tflTktl tfktl abnormal embryos had irregular neural tubes, indistinct somites, abnormal or absent notochords, and anterior limb buds that pointed dorsally rather than ventrally. Although few animals were observed
DEVELOPMENTAL DEFECTS OF THE Tk"ALLELE
Fig. 2. A A brightfield photomicrograph of four embryos from a dissection of a litter from a mating of Tkt' tfkl*l+ + heterozygotes at 8 days p.c. The embryo on the left is normal: note the neural development and allantois. The second embryo from the left was classified as abnormal and is in the yolk sac. The two embryos on the right are abnormal: note the small or nearly absent allantoides. The embryo on the far right appears to have abnormal head folds. B: An abnormal embryo from a litter of Tkfrt f k t l l + + heterozygotes a t 9 days
at days 8-8.5, the dorsal blebs observed in TIT homozygous abnormal embryos were not observed in T +lTkt' tfkt2abnormal embryos.
291
p.c. The arrow points to the irregular neural tube. C: A darkfield photomicrograph of an abnormal homozygous Tkt'I TktJembryo on the right with a normal day 10 littermate on the left. Note the absence of somites in the abnormal embryo, the absence of the posterior end, and the smaller size. D: A different abnormal homozygous Tkt'/Tkt'embryo from the same litter as C. Note the abnormal hyperplasia of cells on the dorsal posterior surface of the closed neural tube.
of brachyury (TI, and exhibit similar features. Both the original T allele and the ENU-induced Tkt' allele have abnormal, diffuse somitic tissue, absence of the posterior end of the embryo, anteDISCUSSION rior limb buds that point dorsally rather than ventrally, irregular neural tubes, shortened alThe results show that the new Tktlallele of T is lantoides, and abnormalities of the notochord. a recessive embryonic lethal mutation. We observed a total of 84 abnormal embryos from days However, three differences between the alleles 8-11.5 that exhibited the abnormal phenotype. In were noted: 1) TIT homozygotes had transient a total of 203 embryos, this is not significantly dorsal blebs at 8-8.5 days (Chesley, '35; Spiegeldifferent from the expected value of 94 (x2 = man, ,761, whereas TktllTktlhomozygotes did not; 1.98), given a male transmission ratio of 93%. 2) abnormal TIT embryos were the same size as (Note that about 47% abnormal embryos are ex- normal littermates (Chesley, '35; Spiegelman, heterozy- ,761, except for the shortened posterior end, pected from matings of Tkt' tfkt'l+ gotes because of high male transmission ratios.) whereas abnormal T"' ITkt1embryos were shortHomozygous Tkt' embryos are abnormal at the ened or retarded at 8 days and smaller than norsame stage of development as the original allele mal littermates at later developmental stages;
+
292
M.J. JUSTICE AND V.C. BODE
Fig. 3. A, B: Frontal sections taken from the dorsal side of embryos from Tktl t f k t l l + + x Tktl t f h t ' / + + matings on day 9.5 p.c. A: A normal embryo with normal well-defined, organized somite pairs, and a normal neural tube. B: An abnormal embryo with unorganized, irregular somite pairs, and an irregular neural tube. C-E: Transverse sections of embryos from Tkt' tfkt'l+ + heterozygotes on day 9 p.c. C: A normal neural tube (NT) and notochord a t the level of the
pharynx. The arrow points to the normal notochord. D: A similar section of an abnormal embryo. Note the absence of the notochord, the abnormal placement of the somite (S), the abnormal neural tube, and abnormally thick layer of mesenchyme over the neural tube. E: A section from the level of the hind gut of a normal embryo showing normal somite and notochord.
and 3) some abnormal TIT embryos have exhibited a neural hyperplasia or neural tube duplication (Spiegelman, '76; Cogliatti '86), whereas this apparent duplication was not observed in TktllTktl homozygotes. The neural hyperplasia has been interpreted differently by investigators as appearing as though chordamesoderm had differentiated into neural tissue (Spiegelman, '761, or caudal duplication of the neural tube (Cogliatti, '86). The neural hyperplasia has been observed in both TI + and TIT embryos, but has exhibited incomplete penetrance (Spiegelman, '76; Cogliatti, '86). Two TktllTktlembryos observed at day 9-9.5 had what
could be interpreted as a second rudimentary neural tube near the caudal end of the embryo; however, it was not as distinct as what has been previously reported, and occurred at a very low frequency. Thus, the neural hyperplasia appears not to be as common in the Tkt11Tkt2 embryos as in TIT embryos. The lack of dorsal blebs and lack of neural hyperplasia in TktllTktlembryos may be because the Tktl mutation is less detrimental to the embryos than the T mutation. For example, the T mutation may be a small deletion, and an additional locus included in the deleted region could be re-
DEVELOPMENTAL DEFECTS OF THE T"" ALLELE
sponsible for the developmental differences. Alternatively, the observed dissimilarities may be due t o differences in genetic background of the studies. Once the T locus is molecularly cloned, molecular analyses combined with the embryological observations may determine how these mutations have a slightly different effect on embryos. The absence or partial absence of a notochord was noted in this study and in studies on the classic T allele (Chesley, '35; Gruneberg, '58; Spiegelman, '76). The development of the neural tube has been correlated with the presence of a notochord (Spemann and Mangold, '24), and Chesley ('35) speculated that the notochordal defect was primary and the neural tube defects were secondary. Spiegelman ('76) subsequently showed that the basal lamina was absent between the neural tube and notochord, allowing cell contacts that do not normally occur. Thus, the abnormal notochord in TIT homozygotes may be a result of the abnormal cell interactions, or lack of some other factor, and not a primary defect. The presence or absence of a basal lamina surrounding the neural tube and the notochord was not examined in Tkt' homozygotes. However, the neuroepithelium that was often thickened and merged with surrounding mesenchyme observed in TktlITkt1 homozygotes is consistent with what would be expected if the basal lamina surrounding the neural tube was defective. These embryological studies were carried out on a mutation induced within a t haplotype, demonstrating that new embryonic lethal mutations induced within t haplotypes can be examined embryologically if they are genetically separated from resident lethal genes and if appropriate controls are included in the study. Obviously, when the t I 2 and tW5 haplotypes are homozygous, any lethal factors other than the resident embryonic lethal genes cause defects that are distinct from those caused by the Tktl mutation. Other mutations induced by ENU mutagenesis in a t haplotype have been genetically separated from resident lethal alleles, allowing embryology studies to be performed (Justice and Bode, unpublished results). For example, new lethal alleles of quaking (qk"' and qk"') were induced in the tW5haplotype. Before the ENU mutagenesis study, one viable, myelin-deficient qk allele identified the quaking locus. Four independent ENU-induced alleles of quaking are recessive embryonic lethal mutations (Justice and Bode, '88b; Shedlovsky et al., '88). Thus, early embryonic death is an impor-
293
tant characteristic of the qk locus (Justice and Bode, '88b). The lethal qk"' allele has been genetically separated from the resident tcZ1' and tcZW5 lethal genes during the mapping studies; however, initial studies indicate that the qk"' allele does not exhibit as distinctive a lethal phenotype as the Tktl mutation (Justice and Bode, '88b). It will be essential to differentiate the effects of a t haplotype genetic background from the effects of the mutant gene in embryological studies of the qkktl allele. These studies on the Tkt' allele thus demonstrate the potential to study the effect of other mutations induced in a t haplotype on embryogenesis. ENU likely causes single base changes, whereas X-rays commonly cause deletions which may involve a large part of one gene or several genes (Vogel and Natarjan, '79; Peters et al., '85; Popp et al., '83;Russell, '82; Ehling and Favor, '84). Thus, a point mutation more probably exhibits a phenotype that reflects a simple change at a locus, whereas an X-ray-induced mutation may reflect more complex changes. It is important t o have numerous alleles of T t o analyze the cellular function of the gene product of the brachyury locus. Interactions of the various alleles at the T locus suggest an apparent gradient of T gene function associated with gene dosage along the length of the body axis (MacMurray and Shin, '88). Once the locus is cloned, it will be useful t o have presumed simple mutations to the shorttailed phenotype to analyze this gradient hypothesis. It is possible that tct is the most simple mutation at this locus, and T is a more severe change. In another experiment, V. Bode has examined 3,589 F1 gametes after ENU mutagenesis of spermatogonial stem cells for a mutation t o brachyury in a normal chromosome 17 without success (V. Bode, unpublished results). In addition, over 7,000 F1 animals carrying a normal chromosome 17 were observed for a mutation to the short-tailed phenotype in other ENU mutagenesis screens in this laboratory (J. McDonald and V. Bode, unpublished results). No heritable mutations to a dominant abnormal tail phenotype were found. One brachyury mutation (TRP)was found in an ENU mutagenesis screen in another laboratory; however, it was not clear whether TRP was induced by ENU or occurred spontaneously (Mann et al., '87). Two independent mutations to brachyury were obtained in a t haplotype in 4,869 gametes screened after ENU mutagenesis (Justice and Bode, '86). Furthermore, a tct mutation was induced in a normal
294
M.J. JUSTICE AND V.C. BODE
chromosome 17 by ENU in 2,215 gametes screened (Bode, '84).It is possible that brachyury is difficult to induce with a point mutagen, and that the presence of the tct mutation within t haplotypes makes mutation t o the short-tailed phenotype more probable in a t haplotype as opposed to a normal chromosome 17. Molecular studies are needed t o resolve this disparate observation and t o determine the developmental function of the wild-type T locus. The ENU-induced alleles of brachyury (Tkt' and Tk"4,,along with the spontaneous and irradiation-induced alleles of brachyury, will be valuable in future molecular studies of this intriguing locus.
ACKNOWLEDGMENTS We thank Nancy Hedrick for excellent technical assistance. We also thank Dr. Linda Siracusa, Dr. Peter Donovan, and an anonymous reviewer for the critical reading of this manuscript. This work was supported by grant HD 15354 from the National Institutes of Child Health and Human Development t o V.C.B. LITERATURE CITED Artzt, K. (1984) Gene mapping within the Tit complex of the mouse. 111: t-lethal genes are arranged in three clusters on chromosome 17. Cell, 39:565-572. Babiarz, B. (1983) Deletion mapping of the Tit complex: Evidence for a second region of critical embryonic genes. Dev. Biol., 95:342-351. Bennett, D. (1975) The t-locus of the mouse. Cell, 6:441-454. Bode, V. (1984) Ethylnitrosourea mutagenesis and the isolation of mutant alleles for specific genes located in the t region of mouse chromosome 17. Genetics, 108:457-470. Chesley, P. (1935) The development of the short-tailed mutant in the house mouse. J . Exp. Zool., 70:429-459. Cogliatti, S. (1986) Diplomyelia: Caudal duplication of the neural tube in mice. Teratology, 34:343-352. Dobrovolskaia-Zavadskaia, N. (1927) Sur la mortification spontanee de la queue chez la souris nouveau-nee et sur l'existence d'un caractere facteur hereditaire, non viable, C. R. SOC.Biol. (Paris), 97:114-119. Dunn, L.C., and S. Gluecksohn-Waelsch (1953) A genetical study of the mutation 'Fused' in the house mouse, with evidence concerning its allelism with a similar mutation 'Kink'. J. Genet., 52:383-391. Ehling, U.H., and J . Favor (1984) Recessive and dominant mutations in mice. In: Mutation, Cancer, and Malformation. E.H.Y. Chu and W.M. Generoso, eds. Plenum Press, New York, pp. 389-428. Frischauf, A.-M. (1985) The Tit complex of the mouse. Trends Genet., 4:lOO-103. Gluecksohn-Schoenheimer, S. (1944) The development of normal and homozygous Brachy (TIT) mouse embryos in the extraembryonic coelom of the chick. Proc. Natl. Acad. Sci. USA, 30:134-140. Gluecksohn-Schoenheimer, S. (1949) The effects of a lethal mutation responsible for duplications and twinning in mouse embryos. J . Exp. Zool., 110:47-76.
Gluecksohn-Waelsch, S., and R.P. Erickson (1970) The t-locus of the mouse: Implications for mechanisms of development. In: Current Topics in Development Biology. A.A. Moscona and A. Monroy, eds. Academic Press, New York, pp. 281315. Gruneberg, H. (1958) Genetical studies on the skeleton of the mouse. XXIII. The development of Brachyury and Anury. J. Embryol. Exp. Morphol., 6:424-443. Johnson, D.R. (1975) Further observations on the hairpintail (Thp)mutation in the mouse. Genet. Res., 24:207-213. Justice, M.J., and V.C. Bode (1986) Induction of new mutations in a mouse t-haplotype using ethylnitrosourea mutagenesis. Genet. Res., 47:187-192. Justice, M.J., and V.C. Bode (1988a) Genetic analysis of mouse t-haplotypes using mutations induced by ethylnitrosourea mutagenesis: The order of T and qk is inverted in t mutants. Genetics, 120:533-543. Justice, M.J., and V.C. Bode (198%) Three ENU-induced alleles of the murine quaking locus are recessive embryonic lethal mutations. Genet. Res., 51:95-102. Klein, J., and C. Hammerberg (1977) The control of differentiation by the T complex. Immunol. Rev., 33:70-104. Lyon, M.F. (1981) The t-complex and the genetic control of development. In: Biology of the House Mouse. R.J. Berry, ed. Academic Press, London, pp. 455-477. MacMurray, A,, and H.-S. Shin (1988) The antimorphic nature of the T' allele at the mouse T locus. Genetics, 120: 545-550. Mains, P.E. (1986) The cis-trans test shows no evidence for a functional relationship between two mouse t complex lethal mutations: Implications for the evolution of t haplotypes. Genetics, 114:1225-1237. Mann, E., R. Benz, D. Swiatek, and V.M. Chapman (1987) Recovery of a new allele at the T locus on chromosome 17. Mouse News Lett., 78:76. Moutier, R. (1973) Mouse News Lett., 48:38. Peters, J., S.J. Andrews, J.F. Loutit, and J.B. Clegg (1985) A mouse p-globin mutant that is an exact model of hemoglobin Rainier in man. Genetics, 110:709-721. Popp, R.A., E.G. Bailiff, L.C Skow, F.M. Johnson and S.E. Lewis (1983) Analysis of a mouse a-globin gene mutation induced by ethylnitrosourea. Genetics, 105:157-167. Russell, W.L. (1982) Factors affecting mutagenicity of ethylnitrousourea in the mouse specific-locustest and their bearing on risk estimation. In: Environmental Mutagens and Carcinogens: Proceedings of the Third International Conference on Environmental Mutagens. T. Sugimura, s. Kondo, and H. Takebe, eds. Alan R. Liss, Inc., New York, pp. 59-70. Selby, P. (1973) X-ray induced specific-locus mutation rates in newborn male mice. Mutat. Res., 18:63-75. Shedlovsky, A., T.R. King, and W.F. Dove, (1988) Saturation germ line mutagenesis of the murine t region including a lethal allele at the quaking locus. Proc. Natl. Acad. Sci. USA, 85:180-184. Sherman, M.I., and L.R. Wudl, (1977) T-complex mutations and their effects. In: Concepts in Mammalian Embryogenesis. M.I. Sherman, ed. MIT Press, Cambridge, pp. 136-234. Shin, H.-S., P. McCormick, K. Artzt, and D. Bennett (1983) Cis-trans test shows a functional relationship between nonallelic lethal mutations in the Tit complex. Cell, 33:925929. Silver, L.M. (1981) Genetic organization of the mouse t complex. Cell, 278:239-240.
DEVELOPMENTAL DEFECTS OF THE Tkt' ALLELE Silver, L.M. (1985) Mouse t haplotypes. Annu. Rev. Genet., 19:179-208. Silver, L.M., D. Lukralle, and J.I. Garrels (1983) To" is a novel, variant form of mouse chromosome 17 with a deletion in a partial t-haplotype. Nature, 301:422-424. Spemann, H., and H. Mangold (1924) Uber induktion von embryonalanlagen durch implantation artfremder organisatoren. Wilhelm Roux Arch. Entwickl. Mech. Org., 100: 599-638. Spiegelman, M. (1976) Electron microscopy of cell associations in T-locus mutants. In: Embryogenesis in Mammals. K. Elliott and M. O'Connor, eds. ElsevieriNorth Holland, Amsterdam, pp. 199-226.
295
Theiler, K. (1972) The House Mouse. Springer-Verlag, New York. Vogel, E., and A.T. Natarjan (1979) The relation between reaction kinetics and mutagenic action of mono-functional alkylating agents in higher eukaryotic systems. 1. Recessive lethal mutations and translocations in Drosophila. Mutat. Res., 62:51-100. Yanagisawa, K.O., H. Fujimoto, and H. Urushihara (1981) Effects of t h e Brachyury (2') mutation on morphogenetic movement i n the mouse embryo. Dev. Biol., 87:242-248.
