flg during gastrulation and segmentation in the mouse embryo

DEVELOPMENTAL

BIOLOGY

152,75-88 (1%8)

Expression of the Fibroblast Growth Factor Receptor FGFR-l/fig during Gastrulation and Segmentation in the Mouse Embryo TERRY P. YAMAGUCHI,*.'RONALD A. CONLON,~ANDJANETROSSANT*

Recent evidence from studies in both amphibians and mammals suggest that the fibroblast growth factor (FGF) family of signaling molecules and their receptors may play regulatory roles during early embryogenesis. We have used both standard and whole-mount i?/ situ hybridization techniques to analyze the temporal and spatial expression patterns of the murine fibrohlast growth factor receptor-l (FGFR-1) in order to help define the role of FGFs in the processes of gastrulation and segmentation. FGFR-I transcripts were detected in the primitive cctoderm of the egg cylinder emhrgo but not in the primitive rndoderm or ectoplacental cone. During gastrulation, FGFR-I mRNA were expressed at high levels in the migrating embryonic mesoderm of the mid-streak-stage embryo. Late-streak-stage embryos displayed strong expression in both the embryonic cctoderm and mesoderm. Within the ectodermal lintage, FGFR-1 mRNA later became localized to the neural ectoderm during its formation and continued to be expressed at high levels throughout neural development. In the mesodermal lineage, FGFR-1 transcripts became concentrated in the posterior medial mesodcrm of the embryo as it condensed to form paraxial mesoderm. The most striking expression patterns were observed before and during segmentation when FGFR-1 was strongly expressed in the presomitic mesodcrm and the rostra1 half of the newly formed somites. The patterns of expression are consistent with a role for FGFR-I in posterior mesoderm formation. FGFR-1 may also play significant roles in the formation of neural ectoderm and the c 1992 Academic Press, Inc. early events that establish compartments within the developing somites.

Genetic control of embryonic development has been shown to involve complex interactions between nuclear transcription factors expressed in particular cell types. However, in all species studied, cell-cell interactions also play important roles in establishing the basic body plan. It is clear therefore that extracellular signaling molecules will also play fundamental roles in regulating development. Recent genetic evidence has suggested that peptide growth factors and their receptor tyrosine kinases (RTKs) are important components of developmental signaling pathways. For instance, mutations within the c-kit proto-oncogene, an RTK encoded by the Wlocus (Chabot et (II., 1988; Geissler et ul., 1988) perturb the development of stem cells in the hematopoietic, neural crest, and germ cell lineages of the mouse (for review see Reith and Bernstein, 1991). Moreover, analysis of several Dmsophila and C. e1e.q~~~developmental mutants have revealed defects in genes encoding peptide growth factors or their RTKs (for review see Pawson and Bernstein, 1990).

’ To xvhom correspondence

should be addressed.

Functional evidence for the importance of peptide growth factors in development also comes from experimental studies in Xer~~1rt.s. Members of both the FGF and transforming growth factor-p (TFG-/j) families arc expressed in Xer~ol)~rs embryos at the time of mesoderm induction and have mesoderm-inducing activity when applied to cultured animal cap explants (for review, Smith, 1989). Interestingly, these peptide growth factors can lead to altered expression of the homcobox genes ,%fkl (Rosa, 1989), XhoxS (Ruiz i Altaba and Melton, 19891, XHboxl, and XlHl,o~6 (Cho and De Robertis, 1990), suggesting a mechanism for how peptide growth factors can modulate positional information. There are also qualitative differences in the type of mesoderm induced by these families of growth factors; FGF is thought to induce posterior/ventral mesoderm (Green et (xl., 1990; Slack Tut
76

DEVELOPMENTALBIOLOGY

(aFGF), and FGF-2 (bFGF), as well as FGF-3 (i&-2), FGF-4 (h&K-FGF), FGF-5 (see Burgess and Maciag, 1989), FGF-6 (Marics et al., 1989), and FGF-7 (KGF) (Finch et al, 1989). To date, four human FGF receptors (FGFR-1 to FGFR-4) belonging to the RTK gene superfamily have been cloned, and homologues from mouse, chick, and frog are being rapidly identified (Lee et al., 1989; Ruta et ab, 1989; Dionne et al., 1990; Mansukhani et al., 1990; Musci et al., 1990; Pasquale, 1990; Reid et al., 1990; Safran et al., 1990; Friesel and Dawid, 1991; Keegan et al., 1991; Miki et ab, 1991; Partanen et al., 1991; Stark et ucl.,1991). The biological activities of most of the FGF family members are not well-defined, nor is it clear whether they functionally bind to the same or distinct receptors. Indeed, FGF-1, FGF-2, and FGF-4 each bind to both FGFR-1 and FGFR-2 with high affinity (Dionne et al., 1990). While spontaneous mutations of the FGF family of signaling molecules and their receptors have not been identified in mammals, they seem likely to play important roles in regulating mammalian embryogenesis. Both FGF-1 and FGF-2 regulate the proliferation, migration, differentiation, and survival of cells of mesodermal and neurectodermal origin (for review, see Burgess and Maciag, 1989). However, there is little information available on the expression of the growth factors or their receptors during embryogenesis on which to base hypotheses as to their developmental role. FGF-2 peptide distribution has been studied immunohistochemically in the late gestation rat fetus (Gonzalez ef al., 1990) and the chick embryo (Joseph-Silverstein et al., 1989). The chick studies indicated that FGF-2 was expressed in the somitic myotome, limb bud muscle, and the developing heart. Detailed analysis of expression during early mouse postimplantation development has not been performed. In situ hybridization studies of the developmental expression of mouse id-2 (Wilkinson et al., 1988) and FGF-5 (Haub and Goldfarb, 1991; Hebert et crl., 1991) revealed that these factors are expressed in contrasting patterns during gastrulation. id-2 mRNA is expressed in parietal extraembryonic endoderm and in the primitive streak (Wilkinson et al., 1988) while transcripts encoding FGF-5 are expressed in the primitive ectoderm adjacent to int-2 expression in the streak (Haub and Goldfarb, 1991). We have defined the tissues that express FGFR-1 in the early postimplantation mouse embryo by both standard and whole-mount in situ hybridization in an attempt to further define the role that the FGF signaling pathways play in early mammalian development. The spatially restricted patterns observed are consistent with a role for the receptor and its associated ligands in mesoderm formation and subsequent neural development. Striking patterns of expression during early so-

VOLUME152.1992

mitogenesis suggest an important role for FGFR-1 segmentation of the paraxial mesoderm. MATERIALS

AND

in

METHODS

Embryos Mouse embryos at various stages of gestation were obtained by mating random-bred CD1 animals (Charles River Canada, Montreal). The day the vaginal plug was detected was considered to be embryonic day 0.5 (E0.5).

FGFRl cDNA Probes Antiphosphotyrosine antibodies were used to screen an erythroleukemic cell Xgtll cDNA expression library in a screen for functional tyrosine kinases (Ben-David et al., 1991). This screen produced, among others, a 2-kb partial cDNA representing nt 862 (-43 nt 5’ of the transmembrane domain) to nt 2917 (-66 nt upstream of the end of the 3’ untranslated sequence) of the previously described mouse basic FGFR (Mansukhani et al., 1990). This clone included the entire catalytic kinase domain. A smaller l-kb probe consisting predominantly of the 3’ untranslated region was generated by cutting the partial cDNA with BylII at nt 1851. Both probes revealed identical patterns of expression.

In Situ hybridixatiow Ir, situ hybridizations were performed essentially as described by Frohman et ul. (1990) with the following modifications. Hybridizations were performed overnight at 50°C. All washes were carried out using 0.1%’ mercaptoethanol as a reducing agent. Final washes in 2~ SSC and 0.1~ SSC were performed for 30 min at 63°C. Emulsion-coated slides were exposed for 2 to 10 days at 4°C. After developing slides, sections were stained with either toluidine blue or hematoxylin and eosin. We did not observe any specific hybridization using the FGFR-1 sense strand as a negative control (data not shown). Whole-mount in situ hybridizations were performed using a modification of existing procedures (Tautz and Pfeifle, 1989; Hemmati-Brivanlou et al., 1990; Conlon and Rossant, in preparation). The hybridization of single-stranded RNA probes labeled with digoxigenin was detected with antidigoxigenin antibodies coupled to alkaline phosphatase. Details of the procedure are available upon request. RESIJLTS

IU Situ Hybridixutim Anulysis of Embryonic FGFR-1 E,rpressio?~during Gastrulation und Neurulatim Northern analysis using a 2-kb partial cDNA as a probe revealed that a single FGFR-1 transcript of 4.3 kb

was first detected in the E9.5 embryo (data not shown). To investigate the temporal and spatial expression patterns of FGFR-1 during early mouse embryogenesis, irk situ hybridization experiments were performed. Paraffin sections of developmentally staged embryos were hybridized with a radiolabeled FGFR-1 probe. Identical patterns of expression were observed when using either a probe derived from the 3’ untranslated region of the cDNA or one that also included both the transmembrane domain and the catalytic kinase domain (for details see Materials and Methods). These probes would not distinguish between putative receptor isoforms, generated by alternative splicing, that vary in their extracellular immunoglobulin-like (Ig-like) ligand-binding domains (Mansukhani et al., 1990; Reid et al., 1990; Bernard ef (xl., 1991) or contain type 1 and type 2 COOHterminal motifs (Hou ct ul., 1991). FGFR-1 was highly expressed in the maternal decidua at 6.5 days (Fig. la and lb) but was expressed at much lower levels in the egg cylinder itself, where transcripts were confined to the embryonic ectoderm. No evidence for regionalization of expression within the ectodcrm was observed at this stage (Fig. la and lb). We failed to detect any signal above background in visceral endoderm and ectoplacental cone. Once gastrulation occurred, the patterns of FGFR-1 expression became more complex. The establishment of the basic body plan begins with the formation of the anterior-posterior axis and is manifested within the primitive ectoderm by the ingression of cells at the posterior end of the embryo through the primitive streak. Embryonic mesodermal cells from the streak migrate laterally and anteriorly between ectoderm and endoderm giving rise to the third germ layer. Migration of mesodermal cells into the extraembryonic region results in a mesodermal contribution to the allantois, visceral yolk sac, amnion, and chorion. Whole-mount in sif~r hybridization of early streak-stage (kE6.5) embryos failed to reveal any detectable signal in the initially forming mesoderm (Fig. 2a) or in embryonic ectoderm. These data suggest that, in our hands, standard in s&r hybridization techniques detected low-level expression more readily than the whole-mount procedure. FGFR-1 transcripts were first shown by whole-mount irl situ hybridization to be concentrated in the posterior mesoderm of the mid-streak-stage embryo (-E7) lateral to the primitive streak (Fig. 2b). As gastrulation proceeded, the domain of FGFR-1 expression in the mesoderm increased in intensity and expanded anteriorly, concomitant with the anterior migration of mesodermal cells from the primitive streak (Fig. 2~). By -E7.25, late-streak-stage embryos displayed high levels of expression throughout the embryonic ectoderm and the mesoderm (Fig. 2d). We could not detect FGFR-1 ex-

pression in extraembryonic tissues at any of these stages by whole-mount in situ hybridization. Once mesoderm has extended to the anterior end of the embryo, the overlying anterior ectoderm forms the neural plate, presumably due to mesodermal influences. Whole-mount ivya&u hybridization of E7.5-7.75 embryos showed that FGFR-1 transcripts continue to be strongly expressed in the embryonic ectoderm (Fig. 3a). However, by this stage, a striking regionalization of FGFR-1 transcripts within the ectoderm was evident. FGFR-1 transcripts were not detected in the posterior proximolateral ectoderm of preheadfold- and headfold-stage embryos (Fig. 3a and 3b). The regions of the ectoderm expressing FGFR-1 in preheadfold embryos probably correspond to ectoderm fated to become neurectoderm, whereas nonexpressing regions become surface ectoderm (Tam, 1989). FGFR-1 mRNA continued to be expressed at high levels in the forming neural plate of headfold stage embryos (Fig. 3b-3d), but no obvious pattern was observed within the neurectoderm at this stage. Dissection of labeled whole-mount late-streak-stage embryos confirmed that FGFR-1 was initially expressed in anterior mesoderm of preheadfold embryos ventral to the future headfolds (data not shown) but hybridization to both sectioned and whole-mount cmbryos showed that transcripts were low to undetectable in the anterior mesoderm by the time the headfolds appeared (Figs. lc, Id, and 3b). In headfold-stage embryos, FGFR-1 transcripts within the mesoderm were localized to the posterior half of the embryo, just, lateral to the primitive streak (Fig. 3b-3d). The FGFR-1 transcripts were observed to be most abundant close to the midline although transcripts were absent from the midline cells of the prirnitive streak and the cells overlying the archenteron, i.e., the node (Fig. 3c and 3d). The distribution of FGFR-1 transcripts within the mesoderm showed an anterior boundary rostra1 to the archenteron. This concentration of FGFR-1 expression in the midline is coincident with the beginning of the development of the paraxial mesoderm, which will later form the somites. Interestingly, FGFR-1 expression was confined to the embryonic mesoderm; transcripts were not found in the extraembryonic mesoderm of the yolk sac or allantois (Fig. 3a and 3b). Again, transcripts were not detected in visceral or parietal endoderm. FGFR-1 continued to be expressed at high levels in the maternal decidua throughout these stages (data not shown).

The primitive streak regresses and mesodermal segmentation begins (-E8) with the formation of the re-

78

DEVELOPMENTAL BIOLOGY

VOLUME 152.1992

FIG. 1. Standard in situ hybridization analysis of FGFR-1 expression in the mouse embryo. A, C, E, G are bright-field views; B, D, F, H are dark-field views showing autoradiography signals. (A, B) ITLsitar hybridization of antisense FGFR-1 RNA to a section through E6.5 egg cylinder demonstrating expression restricted to the embryonic ectoderm. Decidual cells also strongly express FGFR-1 mRNA. (C, D) A sagittal section of an early head-fold-stage embryo (-E7.75) reveals particularly strong expression in the distal mesoderm lateral to the primitive streak. (E, F) Mid-sagittal section of El1 embryo demonstrating predominant FGFR-1 expression in the brain, spinal cord, and somitic sclerotome. (G, H) Mid-sagittal section of E12.5 embryo. Note expression throughout the length of the vertebral column and somites, kidney, and craniofacial mesenchyme. Abbreviations: embryonic ectoderm (ee), visceral endoderm (ve), maternal decidua (dec), amnion (am), embryonic mesoderm (em), visceral yolk sac (vys), headfold (hf), lateral ventricle (v), sclerotome (XI), dorsal root ganglia (drg), metanephros (met). A, B: Bar = 50pm. C, D: Bar = 100 Wm. E-H: Bar : 1 mm.

YAMAGUCHI,CONLON,AND ROSSANT

E,cpres.s;o,/ of Fibroblast

Growth F~rctor Rcwplor

FGFR-l/j,g

FIG. 2. Whole-mount i?csitlc hybridization analysis of FGFR-1 expression duringgastrulation. In all cases, embryos are viewed laterally, with anterior to the left, primitive streak to the right, and ventral down. (A) Early streak-stage embryo (-E6.5). FGFR-1 transcripts were not detectable at this stage. (B) Mid-streak-stage embryo. Note small domain of expression in the posterior mesoderm lateral to the primitive streak. (C) Later mid-streak embryo. The domain of expression has moved anteriorly and increased in intensity. (D) Late-streak-stage embryo. Strong expression throuphout both embryonic ectoderm and mesoderm. Bar = 100 pm.

peated mesodermal structures of the trunk known as somites. Somites are formed by the sequential epithelialization of the rostra1 end of the mesenchymal presomitic mesoderm. The somites subsequently lose their epithelial character and subdivide into dermamyotome and sclerotome; the former gives rise to dermis and skeletal muscle of the trunk and limbs and the latter to the vertebral column (for review, see Stern, 1990). Expression of FGFR-1 continued in both mesodermal and ectodermal lineages during segmentation. Striking expression patterns were observed in the presomitic mesoderm and the forming somites at E8.5. Standard in situ hybridization techniques revealed that FGFR-1 transcripts were abundant in the most rostra1 part of the presomitic mesoderm immediately caudal to the lastformed somite (Fig. 4a and 4b). Transcripts here were confined to a block of presomitic mesodermal cells approximately the size of one somite. Thus the posterior mesodermal expression on late Day 7 has become even more restricted to the rostra1 end of the presomitic mesoderm. FGFR-1 was also strongly expressed in the newly forming epithelial somites, but only in the rostra1 half (Fig. 4a and 4b). Restriction of expression to the rostra1 half somites was clearest in the two most recently formed somites and was less easily detectable in more rostra1 somites. Whole-mount 6~ sits analysis of Day 8 embryos supported the observations made by conventional in situ hybridization techniques. Examination of early l- to 2-

somite embryos labeled with FGFR-1 probes revealed a patch of FGFR-1 expression in the presomitic mesoderm marking the site of formation of the next somite prior to its overt segmentation (Figs. 5a and 3e). Figure 5b demonstrates expression of FGFR-1 mRNA in the rostra1 half of the most caudal newly formed epithelial somite (indicated by arrow) as well as continued expression in the presomitic mesoderm of a lo-somite embryo. This is documented in a high-power magnification of the dorsal aspect of the developing somites of a similar lo-somite embryo (Fig. 3f). Within this rostra1 expression domain, staining appeared to be more intense on the medial side. These expression patterns are even clearer in the 15-somite embryo (Fig. 5~). It was also obvious at this stage that the more mature (rostral) somites were diffusely labeled. The intensity of expression in the rostra1 half somites decreased prior to the loss of their epithelial arrangement and the subsequent formation of the sclerotome. The domains of FGFR-1 expression became more extensive between 8.5 and 9.5 days of gestation. By E9.5, the dorsal and anterior aspects of the two newest somites of the embryo were labeled. In addition, the more rostra1 somites displayed dorsal expression in the forming dermatome (Fig. 5d, and see below). There was also widespread, lateral FGFR-1 expression throughout the 8.5- to 9.5-day embryos. This appeared to be specific signal and not background since the heart was clearly not labeled in any of the embryos examined. Furthermore,

80

DEVELOPMENTALBIOLOGY

V0~~~~152,1992

-.--..= /.,

D

E FIG. 3. Whole-mount in sits hybridization analysis of FGFR-1 expression during neurulation. (A) Lateral view of E7.5 embryo, anterior to the left, ventral dorm. Expression is in both embryonic ectoderm and mesodcrm hut not in proximolateral ecto- or tnesodcrm or extracmbryonic tissues. (B) Lateral view of early headfold-stage embryo (-E7.75). (C) Posterolateral view of headfold stage emhryo. Note strong expression in neurectoderm of head folds and in the primitive streak. (D) Posterior view of similar stage embryo as shown in C. Note expression throughout neural plate and developing paraxial mesoderm. (E) Dorsal view of l-somite emhryo ( -EX) displaying expression in presomitic mesoderm. (F) Similar dorsal view of lo-somite embryo. Note stronger expression in rostra) half of the newly formed somite. Abbreviations: allantois (al), amnion (am), proximolateral mesoderm (plm), headfolds (hf), embryonic ectoderm (ee), embryonic mrsoderm (m), visceral endoderm (ve), node (n), cpithelial somites (es), presomitic mesoderm (psm), developing paraxial mesoderm (pm), neural tube (nt), neural groove (nx), dermatomc (do, rostra1 (r) and caudal (CI half somites. Bar = 100 pm. FIG. 5. Whole-mount iw sitrr hybridization analysis of FGFR-1 expression in segmenting IWE9.5 embryos. For all embryos, rostra1 is to the left and dorsal is up. (A) Lateral view of 2-somite embryo ( -E8). (B) lo-somite embryo. (C) -5somite embryo. Note high levels of FGFR-1 expression in developing forebrain, presomitic mesoderm, and forming somites. Arrow denotes FGFR-I expression in the rostra1 half of the newest somite. (D) E9.5 emhryo. Strong FGFR-1 expression is obvious in the developing forelimb bud and the branchial arches. Patchy staining in the anterior head region is artifactual. (E) Control ,Trc-” staining in an embryo of similar stage to that shown in (C). Abbreviations: hcadfold (hf), somitrs (SJ, presomitic mesoderm (psm). A-C, E: Bar = 200 Qrn. D: Bar = 250 pm.

82

DEVELOPMENTALBIOLOGYV0~~~~152.1992

C FIG. 4. FGFR-1 expression in the developing somites. A and C are bright-field views, B, D, and E are dark field. (A, B) Longitudinal section through presomitic mesoderm and forming somites of E8.5 embryo. Note strong expression in rostra1 half of newly formed somite. (C, D) Frontal sections through tail somites of El1 embryo hybridized with a FGFR-1 antisense probe. (E) A serial section hybridized with a myogenin probe to delineate the myotome. Abbreviation: somites (s), presomitic mesoderm (psm), dermatomc (d), myotome (m), sclerotome (~1). Bar = 100 pm. control embryos of similar stages hybridized with the highly spatially regulated gene En-2 did not show widespread diffuse staining (Fig. 5e). Expression of FGFR-1 remained high in the somites as they began to differentiate into dermamyotome and sclerotome. Using the myogenic determination gene, myogenin, as a marker for the myotome (Sassoon ef al., 1989) we found via standard in situ hybridization analysis that at El1 FGFR-1 displays a pattern of expression (Fig. 4c and 4d) that is complementary to that of myogenin (Fig. 4e) and therefore conclude that FGFR-1 is expressed in the dermatome and sclerotome.

FGFR-1 expression was confined to perichondrium and osteoblasts in the anterior part of the sclerotomederived vertebral column by E11.5-12.5, while the centers of ossification were negative (Fig. le-lh). FGFR-1 expression remained strong in the less well differentiated sclerotome tissues of tail somites. Interestingly, at these later stages (E11.5-E12.5), FGFR-1 expression marked the caudal half of the forming vertebrae. In summary, the expression pattern of FGFR-1 during somitogenesis is rather complex, with transient early expression demarcating the site of condensation of new somites in the presomitic mesoderm. Expression

83

YAMAGUCHI, CONLON, AND ROSSANT

patterns initially marked rostra1 domains of the forming somites but subsequent expression was restricted to the caudal half of the forming vertebrae. Both the dermatome and sclerotome, but not the myotome, expressed FGFR-1. FGFR-1 appears to be an early molecular marker for rostral-caudal regionalization during somitogenesis. Expression

qf’ FGFR-1 Transcripts

after Neurulation

FGFR-1 continued to be expressed in the derivatives of the embryonic ectoderm after neurulation. As mentioned previously, transcripts were found at high levels in the forming neural plate (Fig. 3b-3d). By E8, strong expression was obvious throughout the neural folds (Fig. 3e). The highest levels of FGFR-1 neural expression were in the developing forebrain around E8.5 (Fig. 5a-5c). FGFR-1 levels appeared to gradually decline posteriorly within the neural tube (Fig. 3f). No expression of FGFR-1 was detected in the heart at this stage (Fig. 5b and 5~). Similar results were seen in the E9.5 embryo by both standard and whole-mount hybridization analysis. Transcripts continued to be expressed at high levels in the developing brain, particularly in the developing diencephalon (Fig. 5d). The mesenchyme of the first two branchial arches were intensely labeled (Fig. 5d). FGFR-1 was also strongly expressed in the mesenchyme of the budding forelimbs (Fig. 5d). Again, little or no expression was detected in the heart. FGFR-1 continued to be expressed at high levels in the brain and spinal cord of the El1 embryo (Fig. le and lf). However, expression became restricted to the proliferative ependymal layer of the lateral and fourth ventricles of the brain by E12.5 (Fig. lg and lh). No labeling of dorsal root ganglia was observed (Fig. le-lh). Strong labeling was observed in the craniofacial regions of 12.5-day embryos, particularly the mesenchyme of the nasal and maxillary area as well as the tongue and mandible (Fig. lg and lh). Intense labeling was also seen in the kidney while lungs and intestines stained at lower levels and more diffusely. Again, no expression was observed in the heart or the liver at this stage. Our observations of FGFR-1 expression at 12.5 days of gestation are generally consistent with previously published reports in which FGFR-1 expression was analyzed from 12.5 days onward (Safran et (II., 1990; Wanaka et ul., 1991).

The observed tissue-specific and spatially-restricted patterns of FGFR-1 expression during early postimplantation development suggest that the FGFs are not simply mitogens but may also play roles in a number of

important embryonic events. These events include mesoderm induction, neural induction, and somite formation and patterning. FGFR-1 and Mesoderm Induction Previous studies in amphibians have suggested a role for the FGF signaling pathway in the induction of mesoderm. Classic explant-recombination experiments have indicated that a signal from vegetal pole cells induces overlying animal cap cells to become mesoderm (Nieuwkoop, 1969). The nature of the signaling molecules involved has been a subject of much recent research. Several mesoderm-inducing factors (MIFs) have been identified, all of which are related to either FGF or TGF-/I (for review see Smith, 1989). bFGF is a potent MIF and both bFGF and its receptor are expressed in the frog blastula at a time when cells are competent to form mesoderm (Gillespie et al., 1989; Kimelman ef crl., 1988; Musci et al., 1990). The process of primary mesoderm induction is not clearly understood in the mouse. However, the closest equivalents of animal cap and vegetal pole cells are the primitive ectoderm and primitive endoderm of the early egg cylinder. The localization of FGFR-1 to the primitive ectoderm but not the primitive endoderm would be consistent with its involvement in receiving a mesoderm-inducing signal from the underlying endoderm. Recent data from the chick, however, suggests that there may already be a disperse subpopulation of cells in the epiblast committed to mesoderm prior to primitive streak formation (Stern and Canning, 1990), although lineage analysis in the mouse does not. support this concept (Lawson et ul., 1991). If FGFs were involved in establishing this subpopulation one might expect to see mosaicism of a component of the FGF signaling pathway. The resolution of the iv sifzc hybridization approach is not sufficient to determine whether mosaicism of FGFR-1 exists. Recent experiments in Xenopus involving the overexpression of a dominant negative FGFR-1 construct, in fact,, support a role for FGFs in development of the later posterior mesoderm that gives rise to the somites, rather than in the initial mesoderm that involutes through the dorsal lip of the blastopore (Amaya et crl., 1991). In the mouse, this would be equivalent to suggesting that FGFs are not involved in the formation of the first mesoderm to migrate through the primitive streak but rather in the later forming paraxial mesoderm that comes to lie posterior to the head region. The strong expression we observe in the condensing paraxial mesoderm suggests such a role for FGFR-1 in the development of the posterior mesoderm in the mouse. FGFR-1 m/d Sornitogew.sis The localization of FGFR-1 expression to the condensing paraxial mesoderm of the headfold-stage em-

84

DEVELOPMENTALBIOLOGY

bryo is not the only stage at which an association between FGFR-1 and axial mesoderm development occurs. We observed striking patterns of FGFR-1 expression during the establishment of the somites. FGFR-1 was highly expressed in the most rostra1 domain of the presomitic mesoderm throughout somitogenesis and in the rostra1 half of the first two epithelializing somites. The pattern of FGFR-1 expression in the presomitic mesoderm suggests that FGFR-1 expression could demarcate the future segmental unit. Studies designed to investigate the mechanisms responsible for segmentation suggest that the cell division cycle may play a role (Primmett et al., 1988, 1989). Brief heat-shock treatments result in multiple cyclic somitic anomalies (Primmett et al., 1988,1989). These treatments alter segments qualitatively and quantitatively as well as the rostrocaudal proportions of the sclerotome (Primmett et al., 1988). In addition, a high mitotic index is found at the rostra1 end of the segmental plate of the chick (Stern and Bellairs, 1984) in a domain that suggests an overlap with the pattern of FGFR-1 expression. Given that FGFR-1 can mediate proliferative signals it is possible that the FGF signaling pathway controls the proliferation of cells of the presomitic mesoderm. It may be possible to obtain experimental evidence for this by manipulating the FGF signaling pathway in in vitro explants of presomitic mesoderm and by determining whether treatments such as heat shock can alter FGFR-1 function or expression. The early distinction between rostra1 and caudal halves of the newly formed somites shown by FGFR-1 expression is interesting in light of the extensive evidence in the chick that this is an important developmental boundary. Functional differences between rostra1 and caudal half somites are best demonstrated by the observation that both neural crest cells and spinal cord motor axons migrate out across the rostra1 half only (Keynes and Stern, 1984). It has been argued that both rostral-caudal distinctions as well as the determination of the fate of cells to become dermomyotome and sclerotome are established during the process of somite epithelialization (for discussion see Keynes and Stern, 1988; Stern et al., 1988). Grafting experiments demonstrated that when newly formed rostra1 or caudal half somites were transplanted to new sites, the half sclerotome that subsequently formed maintained the properties of the half of its origin (Stern and Keynes, 1987). There is little evidence available on the genetic control of this early rostral-caudal distinction. In chick somites, a number of proteins that are differentially expressed in the rostral or caudal half have been identified (for review, see Keynes and Stern, 1988; Norris et al., 1989). However, most of these proteins do not show rostral-caudal differences in expression until after neural crest migration

V0~~~~152,1992

has already begun. They cannot therefore account for the establishment of functional differences between half somites. The restricted pattern of FGFR-1 expression in the rostra1 half of the newly formed epithelial somite during somitogenesis shows that FGFR-1 is an early molecular marker for rostral-caudal somitic differences and indicates that it may play a functional role in this process. Later in development, FGFR-1 is expressed in the caudal half of the forming vertebrae, consistent with the resegmentation hypothesis that proposes that a vertebra arises from the fusion of the sclerotome derived from the rostra1 half of one somite with the caudal half of the preceding somite (although see Stern, 1990).

FGFR-1 and Neural Induction

FGFR-1 expression is not only associated with mesodermal differentiation but also shows an interesting association with the formation of the neural plate. As noted above, FGFR-1 is expressed at detectable levels throughout the ectoderm prior to gastrulation and persists within the ectoderm layer as mesoderm passes through the primitive streak. However, by the late primitive streak stage, when mesoderm has migrated to the anterior of the embryo, intense ectodermal expression of the FGFR-1 gene shows a clear boundary within the embryonic region (Fig. 3a). The ectoderm closest to the amnion expresses only low levels of FGFR-1 while more distal ectoderm shows higher expression. The domain of low expression is wedge shaped, with the larger dimension at the posterior end of the embryo. As the head folds form, FGFR-1 expression is clearly associated with the developing neurectoderm at the anterior end and not with the nonneural posterolateral ectoderm. Grafting studies in mouse embryos in culture (Tam, 1989) indicate that the FGFR-l-expressing area in the preheadfold-stage embryo is the precursor region for the neural plate. The rest of the ectoderm gives rise to surface ectoderm. FGFR-1 appears to be delineating the future neural plate before overt signs of its differentiation. Formation of neurectoderm is thought to involve an inductive signal from underlying mesoderm (Spemann, 1938) and the localization of FGFR-1 would be consistent with it playing a role in receiving this inductive signal. The phenotype of Xenopus embryos expressing a dominant-negative FGFR-1 construct does not support this hypothesis since head structures, including neural structures, are apparently normal (Amaya et aZ., 1991). However, since nothing is known of the precise localization of FGFR-1 in Xenopus, one cannot exclude the possibility that FGFR-1 is required for neural induction in mammals but not in Xenopus.

YAMAGUCHI,CONLON,ANDROSSANT

FGFR-1 Expression

Expression

iz Later Tissues

As development proceeded, FGFR-1 expression was observed in a wide variety of tissues, indicating possible multiple roles for this receptor. FGFR-1 expression patterns in chondrogenic tissues such as the sclerotome, craniofacial mesenchyme, and developing limb bud mesenchyme imply that it could regulate the chondrogenesis of mesenchymal cells. Our findings essentially confirm and extend that of Wanaka and others (1991) who recently reported an in situ analysis of FGFR-1 expression during late stages of rat development. Within the central nervous system, FGFR-1 expression subsequently became localized to neuronal precursors within the proliferative ependymal layer in agreement with others (Heuer ef al., 1990; Wanaka et ab, 1990). However, two additional observations from our results should be noted. First, we could not detect FGFR-1 transcripts in the developing heart at the stages we examined. It has been previously demonstrated that FGFR-1 was expressed in the 14.5-day heart (Wanaka et al., 1991); thus, the onset of FGFR-1 expression appears to be relatively late in the development of cardiac muscle. Second, we could not demonstrate significant levels of FGFR-1 transcripts in early skeletal muscle precursors, i.e., the somitic myotome (Fig. 4d). However, high levels of the myogenic determination gene, myogenin, were detected (Fig. 4e; Sassoon ef trl, 1989). Transcripts were only found in muscle at later stages of development, particularly in the muscle surrounding the vertebral column (Fig. lh, see also Wanaka et a.h, 1991) and in postnatal (data not shown) and adult skeletal muscle (Moore et al., 1991). These findings suggest that FGFR-1 is involved in relatively late stages of muscle development. However, it is also possible that two different populations of myogenie precursor, differentially expressing FGFR-1, could exist during muscle development. It has been demonstrated in the developing chick limb bud that at least two distinct classes of myogenic precursor can be distinguished based on both their time of migration into the limb bud (Seed and Hauschka, 1984) as well as on their requirement for FGF in order to differentiate in vitro (Seed and Hauschka, 1988). Moreover, we know that FGF suppresses the expression of two myogenic determination genes, myogenin and MyoDl (Brunetti and Goldfine, 1990; Vaidya et al., 1989) and that FGFR levels are downregulated during myogenic differentiation in ~tifro (Moore et al., 1991; Olwin and Hauschka, 1988). Interestingly, Stark et a,Z.(1991) have recently observed that FGFR-4, in contrast to FGFR-1, is expressed in the early myotome. Taken together, these observations suggest that the FGF signaling pathway participates in the regulation of skeletal myogenesis of specific myogenic precursors. It should also be noted that it is unlikely, in

of Fibroblast

Grouth

85

Factor Receptor FGFR-l/fig

the embryo, that FGFR-1 mediates any of the well-documented effects of FGF family members on vasculo- or angiogenesis (Burgess and Maciag, 1989) as FGFR-1 was not expressed in yolk sac mesoderm or blood islands nor in any endothelial cells lining the heart or blood vessels. The FGF Ligand-Receptor Development

Families

in Early

Mouse

In order to claim that localized FGFR-1 transcription is causally related to a number of important developmental events, one needs also to know the distribution of its cognate ligand. Unfortunately, it is not clear what the physiologically relevant ligand is. There are multiple ligands and receptors in the FGF family of signaling molecules and the pharmacology of ligand-receptor interactions is complex. This issue is further complicated by the finding that FGFR-1 mRNA is alternately spliced, with at least two different forms of FGFR-1 existing in the mouse (Bernard et aZ., 1991) and human (Johnson et a,l., 1990). Similar results have been found for FGFR-2 (Dionne et al., 1990; Miki et al., 1991). The longer form of mouse FGFR-1 (FGFR-IL) contains three Ig-like extracellular domains while the shorter form (FGFR-1s) lacks the first Ig-like domain (Reid et al., 1990). Interestingly, only FGFR-1L was expressed in the developing mouse brain, while cardiac and skeletal muscle, as well as several cell lines, expressed both forms of FGFR-1 (Bernard et al., 1991). The probes used in our study detect all FGFR-1 molecules possessing catalytic tyrosine kinase domains and which are therefore capable of activating downstream signaling pathways. Future plans include further defining the expression patterns of FGFR-1 by using probes specific for alternately spliced mRNAs. Ligand binding studies revealed that mouse FGFR-1L bound both aFGF and bFGF (Safran et al., 1990) while FGFR-1S was activated by bFGF and K-FGF (Mansukhani et al., 1990). Both human FGFR-1L and FGFR-1S (Johnson et al., 1990) as well as FGFR-2L bound aFGF, bFGF, and K-FGF with high affinity (Dionne et al., 1990). However, bFGF did not bind to FGFR-2S (Miki et al., 1991). FGFR-3 was activated by both aFGF and bFGF (Keegan et al., 1991) while the recently cloned FGFR-4 bound aFGF with high affinity, but not bFGF (Partanen et al., 1991). Clearly, more detailed in vitro studies on the relative binding affinities of the various ligands for each individual receptor are needed. Such studies, however, still do not prove that ligand and receptor actually interact in viva. To further complicate matters, studies indicate that heparan sulfate proteoglycans (HSPGs) are required in order for bFGF to bind to its high-affinity receptor (Yayon et al., 1991). A detailed

86

DEVELOPMENTALBIOLOGYVOLUME152,1992

analysis of the distribution of ligands, HSPGs and receptors during embryogenesis should help reveal the intrinsic ligand-receptor interactions. The expression patterns of FGFR-4 have been recently determined and they reveal that FGFR-4 was expressed mainly in tissues where FGFR-1 was not (Stark et al., 1991; Yamaguchi et al, unpublished data). At E&5-E9.5, FGFR-4 was expressed in the definitive endoderm of the developing gut, the endodermal component of the yolk sac, and the myotomal compartment of the somites. Expression continued in endodermal derivatives of the intestine, liver, and lung, as well as in skeletal muscle by E14.5 (Stark et al., 1991). FGFR-4 and FGFR-1 were apparently coexpressed in the kidney and cartilage of the 14.5-day embryo. In contrast to FGFR-1 transcripts, FGFR-4 mRNA was never detected in the central nervous system. There is little data available on the distribution of FGF family members in the early stages of postimplantation development to aid in determining the relevant ligand(s) for FGFR-1. RNA in situ hybridization data is available for int-2 and FGF-5. However, the expression profile of neither growth factor is consistent with it acting as an autocrine or paracrine ligand for FGFR-1. During gastrulation (E7), FGF-5, like FGFR-1, was expressed throughout the embryonic ectoderm (Haub and Goldfarb, 1991;Hebert et al., 1991). FGF-5, however, was also expressed in visceral endoderm and the primitive streak. FGFR-1, like FGF-5, was subsequently expressed (E7.5) in the distal midline mesoderm, but FGF-5 was also expressed in the head process (Hebert ef al., 1991). int-2, by contrast, was expressed in extraembryonic endoderm and throughout the primitive streak including newly formed embryonic and extraembryonic mesodermal cells migrating out of the streak (Wilkinson et ul., 1988). Expression in the mesodermal cells was confined to the migratory period only. Direct comparison of FGF-5 and int-2 expression patterns at E7.5 indicate that FGF-5 expression in the ectoderm is adjacent to id-8 expression in the mesoderm (Haub and Goldfarb, 1991). At later stages of development FGF-5 expression appears to be complementary to FGFR-1 with predominant FGF-5 expression in lateral mesoderm, somitic myotome, facial muscle, and specific sites within the limb (Haub and Goldfarb, 1991). km-2 is also expressed in the cerebellum, retina, teeth mesenchyme, and ear during later fetal development (Wilkinson et al., 1989). It is not known whether either or both FGF-5 and int-2 can functionally bind to FGFR-1. Furthermore, the distribution of the factors that have been shown to bind FGFR-1 in vitro, namely aFGF, bFGF, and K-FGF, have not been determined in the early embryo. In order to further elucidate the role of FGFs and their receptors in the regulation of early developmental

processes, several lines of investigation must be pursued. As mentioned above, detailed characterization of the expression patterns of all the FGF ligands and their receptors should be performed. The relative binding affinities of the FGF ligands for each receptor must be assessed. Informative phenotypes may be generated by disrupting the genes encoding the FGFs and their receptors by gene targeting or by expressing dominant negative constructs in mammalian embryos. Finally, the generation of FGFR-l- ES cell lines should allow for the assessment of their in vitro differentiative abilities, as well as provide a vehicle for characterization of putative downstream genes. The authors thank Ken Letwin, Yaacov Ben-David, and Tony Pawson for the FGFR-1 partial cDNA. T.P.Y. would like to dedicate this paper to the memory of Joe A. Connolly. This work was supported by the National Cancer Institute of Canada. J.R. is a Terry Fox Cancer Research Scientist of the NCIC and a Howard Hughes International Scholar. R.A.C. holds a Medical Research Council of Canada Fellowship. Note odded in ,roof: Since this manuscript was submitted, two papers have been published describing the expression patterns of both FGFR-1 and FGFR-2 in the mouse by standard in situ hybridization (Orr-Urteger ef rcL, Development 113, 1991; Peters et cd.,Development, 114,1992). The expression pattern described for FGFR-1 is consistent with our findings at later stages of development; we provide a detailed analysis of the expression pattern in the early postimplantation embryo.

REFERENCES Amaya, E., Musci, T. J., and Kirschner, M. W. (1991). Expression of a dominant negative mutant of the FGF receptor disrupts mesoderm formation in Xe?loprls embryos. Cell 66, 257270. Ben-David, Y., Letwin, K., Tannock, L., Bernstein, A., and Pawson, T. (1991). A mammalian protein kinase with potential for serine/ threonine and tyrosine phosphorylation is related to cell cycle regulators EMBO J. 10,317-325. Bernard, O., Li, M., and Reid, H. H. (1991). Expression of two different forms of fibroblast growth factor receptor 1 in different mouse tissues and cell lines. Pmt. Nutl. Acad. Sci. USA 88, 7625-7629. Brunetti, A., and Goldfine, I. D. (1990). Role of myogenin in myoblast differentiation and its regulation by fibroblast growth factor. J. Bid Ch.em. 265,5960-5963. Burgess, W. H., and Maciag, T. (1989). The heparin-binding (fbroblast) growth factor family of proteins. Annu. Rev. Biochem. 58, 575-606. Chabot, B., Stephenson, D. A., Chapman, W. M., Besmer, P., and Bernstein, A. (1988). The proto-oncogene c-kit encoding a transmembrane tyrosine kinase receptor maps to the mouse W locus. Nature 335, 88-89. Cho, K. W. Y., and De Robertis, E. M. (1990). Differential activation of Xenopus homeo box genes by mesoderm-inducing growth factors and retinoic acid. Genes Deu. 4,1910-1916. Dionne, C. A., Crumley, G., Bellot, F., Kaplow, J. M., Searfoss, G. R. M., Burgess, W. H., Jaye, M., and Schlessinger, J. (1990). Cloning and expression of two distinct high-affinity receptors cross-reacting with acidic and basic fihroblast growth factors. EMBO J. 9, 26852692.

Finch, P. W., Ruhin, J. S., Miki, T., Ron, D., and Aaronson, S. A. (1989). Human KGF is FGF-related with properties of a paracrine effector of epithelial cell groivvth. Scicww 245, 752-755. Fricsel, R., and David, I. 8. (1991). cDNA cloning and developmental expression of fihrohlast growth factor receptors from X
IO, 172% 4’i:Hi.

Josc,l)h-Silverstein. J., Consigli, S. A., Lyser, K. M., and Ver F’ault, C. (1989). Basic fihroblast growth factor in the chick embryo: Immunolocalization to striated muscle cells and their precursors. ,/ Cc,// Bid. 1OX, 2d%-Zl6li. Kwgan, K., Johnson, J). E., Williams, L. T., and Hagman, M. J. (1991). Isolation of an additional memhcr of the fibroblast growth factor rrwptor farnil\-, F(:PR-3. I’u~.. ;l’,t//. ;lccctl. Sci. I’,%~188, 1095-1099. Ktyrws, R. .J.. and Stern, V. D. (198-1). Se)rmcntation in the vertebrate nervous system. .\Tf/t~i~ 310, 786-7811. Kcyncts, R. .J., and Stern, (‘. D. (1988). Mwhanisms of vertebrate seg:nientation. 1)(~,,f,l,,/J,i/f,/// 103, 11:3-4214. Kimelman, D., Abraham, J. A., Haaparanta. T., Palisi. T. M., and Kirschncr, M. W. (1988). The presence of fib&last growth factor in the fro(l tyg: Its role as a natural mesoderm inducer. Scicww 242, 1OS- 1056. Lawson, Ii. ‘1.. Mentws, ,J. J., and I’rdersen, R. A. (1991). Clonal analysis of cpihlast fntc during germ layr formation in the mouse cmhryo. /)c,r,c,lo/,rr,c,r/t 113, X91-911. Leca. I’. I,., Johnson, D. E.. (‘ouwns, L. S., Fried, V. A., and Williams, I,. T. (19XI~). Purification anti complementary L)NA cloning of a wwpror for basic tibrohlast growth factor. Sciwrw 245, 57-60. Lyon. RI. F.. and Scarlv. A. G., Etls. (1990). “Genetic Variants and

Strains of the Laboratory Mouse,” 2nd ed. Oxford ITniv. Press, IJnited Kingdom. Mansukhani, A., Moscatelli, D., Talarico, D., Levgtska. V., and Basilice, C. (1990). A murine fibroblast growth factor (FGF) receptor expressed in CHO cells is activated by basic FGF and Kaposi FGF. Pror. l’ccf/. AU//l. sci. IX.1 87, 4378-4382. Marics, I., Adelaide, J., Raybaud, F., Mattei, M.-G., Coulier, F., Planche, J., De Lapeyricre, O., and Birnhaum, L). (19X9). Charactcrization of the HST-wlated FGF-6 gene, a new member of the fibroblast growth factor gene family. Ot~c.oqc~~~~~ 4, 335-310. Miki, T., Fleming. T. P., Bottaro, D. P., Ruhin, J. S., Ron, D.. and Aaronson, S. A. (l!Wl). Erpression cDNA cloning of the KGF receptor by creation of a transforming autocrine loop. ,Scj~ilw 251, 7%75. Moore, J. W., Dionne, C., Jayc, ICI., and Swain, J. L. (1991). The mRNlZs encoding acidic FGF, basic FGF, and FGF receptor are coordinateI! downre~ulated during myogvnic differentiation. I)(,/‘c,/,,l,tr/t’,ct 11 1, 741-748. Musci, T. J., Amaya, E., and Kirschner, M. W. (1990). Kegulation of the lihroblast growth factor receptor in early X<~twf~~s embryos. f’vrw. h’utl. Aurtl. Sri. 1S.4 87, 8Xj.i~X369. Nieu\vkoop, I’. D. (1969). The formation of mesodrrm in urodelean amphibians. 1. Induction by the endoderm. Il’i/l/c~/w Roui Aw//. E,/f /P,~I(W//. o/;c/. 162, :3.11~:~7:~. Norris, LV. E., Strrn, (‘. I).. and Keynes, R. J. (1989). Molecular ditl’ev ences betncvn the rostra1 and caudal halves of the sclt~rotomc in thv chick embryo. I)(~,~(,/,JI,,,/(,,/I 10.5, 511-518. Olwin. B. B.. and Hauschka, S. D. (198ti). Cell surfaw fihrohlast growth factor and cspitlermal yrowth factor receptors are pt’rmwnently lost during skeletal muscle terminal differentiation in c.uturc. J. C’f,l/ Rioi. 107, 761 m7611. Fartanen, J., Makrla. 1’. P., Ewola. E., Korhoncq J., llirwnc~n, H.. Claesson-~~~l:c,lcli, L., and Alitalo. K. (1991 1. FGFR-.I, a novel acidic fibroblast gro\vth factor reccaptor with a distinct esprcssion ~)atern. ELZI&3 .I 10, 1:317-135,1. Pasqualc, E. Ii. (1990). A distinctive family of embryonic protcin-tyosine kinase rcwptors. !“~YJv.,\irtl. .,Icvt/. SC;. TN.4 87, 5Xl%Z3lii. Pawson, T., and Bernstein, A. (1990). Receptor tyrosine kinascs: Gchnrtic evidence for their role in Llw~sopltilu and mouse dwelopmenl. Twutls C;c~~ifd6, X0-i356. Primmvtt. D. R. S., Norris, IV. E.. Carlson, (;. J.. Keynes, R. d.. and Stern, (‘. I). (19%)). Periodic segmental anomalies induced by brat shock in thr chick embryo art’ associated with the wll cycle. Ikzr~>/r)(,U/cat/l 105, 119-1x0. Primmvtt, D. K. S., Stern, C. D., and KeyncBs. R. J. (1988). Howt shock causes rqwated scymt~ntal anomalies in the chick ~~mhryo. I)cz,~/r~,w /r/c,ttt 104, :x31-:3x1. Reid, II. H., Wilks. A. F., and Bernard, 0. (1!)!10). ‘I’lvo forms ol’ thv basic lihrohlast Ero\vth factor receptor-like mRNA arr expressed in the developing mouse brain. PYW. ,\‘crt/. ..Icotl. SC,/. I3.11 87, 159& 1600. Rrith, A. D., and Bvrnstvin, X. I 1991 ). Molecular biology of the Ifvantl Sfwl loci. I>/ “Genome Analysis Volume 3: Genes and Phenotypes.” Cold Spring Harbor I,aboratory Press, (‘old Spring Harbor, NY. Rosa, F. M. (1989). ;Vi.r. I. a homeohox mRNA inducible hy mcsoderm inducers. is espreswd mostly in the presumptive cndoderm;~l cells of Xenopus embryos. (‘(,/I 57, 96j~97-1. Ruiz i Altaba. A. (1991 ). \‘crtehrate dc~velopmcnt: An cmer~innsynt htssis. ?‘wtit/s C;c~ilc+.7, %7&2X0. Ruiz i Altaba, A.. and Melton, D. A. I 1989). Interaction hct\vwn pep tide growth factors and homeobox genes in the wtablishmc~nt of antrro-l)osteriot, polarity in frog embryos. ‘VUIU~(J 341, :3:3-:3X. Ruta. M., Ihlrgcw, \V., (;ivol, L)., Epstein, J., Neiger, N., Kal)lo\r, J., Crumlcy, G.. Dionne, (‘.. Jays. M., and Schlcssingcr, J. (11lX9~.Rwep tor for acidic lihrohlast yrowth factor. iu rc~lawd to thv tyrosinc,

88

DEVELOPMENTALBIOLOGY

kinase encoded by the,tin.s-like gene (FLG). Pwc. Ntrtl. dccld. bi. 1JS.l 86, 8722-8726. Safran, A., Avivi, A., Orr-Urtereger, A., Neufeld, G., Lonai, P., Givol, D., and Yarden, Y. (1990). The murinejggene encodes a receptor for libroblast growth factor. Owogew 5, 635-643. Sassoon, D., Lyons, G., Wright, W. E., Lin, V., Lassar, A., Weintraub, H., and Buckingham, M. (1989). Expression of two myogcnic regulatory factors myogenin and MyoDl during mouse embryoEenesis. Nut?rrc 341, 303-307. Seed, J., and Hauschka, S. D. (1984). Temporal separation of the migration of distinct rnk-ogenic precursor populations into the develop ing chick wing bud. 1)~~~.Biol. 106, 389-393. Seed. J., and Hauschka, S. D. (1988). Clonal analysis of vertebrate myopcnesis. VIII. Fibroblast growth factor (FGF)-dependent and FGF-independent muscle colony types during chick wing devclopment. DCP. Bid. 128, 40-49. Slack, J. M. W., Darlington, B. G., Heath, J. K., and Godsave, S. F. (198’7). Mesoderm induction in early Xf,t/o~~u.s embryos by heparinbinding growth factors. ,hJrrflrw 326,197. Smith, J. C. (1989). Mesoderm induction and mesoderm-inducing factors in early amphibian development. Dewlopnrer/f 105, 665-667. Smith, J. C., Price, 8. M. J., Green, J. B. A., Weigel, D., and Herrmann, B. G. (1991). Expression of a Xwcol,rrs homolog of Brtrch!ycv!/ (T) is an immediate-early response to mcsoderm induction. Cc>II647, 7% XT. Sprmann, H. (1938). “Embryonic Development and Induction.” Yale Univ. Press, New Haven. Stark, K. L., McMahon, J. -4.. and McMahon, A. P. (1991). FGFR-4, a new member of the fibroblast growth factor receptor family, expressed in the definitive endodcrm and skeletal muscle lineages of the mouse. Dew/opvrc,t,f 113, 64-651. Stern, C. D. (1990). Two distinct mechanisms for segmentation? SPY. Dets. Bid. 1, 109-116. Stern, r. D., and Bellairs, R. (1984). Mitotic activity during somite segmentation in the early chick embryo. ;lr/cct. Embr,gol 169, 97-102. Stern, C. D., and Canning, D. R. (1990). Origin of cells giving rise to mesoderm and endoderm in chick embryo. ,Vctfrcw 343, 273-275.

VOLUME~~~71992 ..

Stern, C. D., Fraser, S. E., Keynes, R. J., and Primmett, D. R. N. (1988). A cell lineage analysis of segmentation in the chick embryo. Dcwc~lqrrcmd 104, 231-244. Stern, C. D., and Keynes, R. J. (1987). Interactions between somite cells: The formation and maintenance of segment boundaries in the chick embryo. Dtwlopncwf 99, 261-272. Tam, P. P. L. (1989). Regionalisation of the mouse embryonic ectodcrm: Allocation of prospective ectodermal tissues during gastrulation. Dc~~f4opnwtit 107, 55-67. Tautz, D., and Pfcifle, C. (1989). A non-radioactive irr situ hgbridization method for the localization of specific RNAs in I)ro.so~~hi/tr embryos reveals translational control of the segmentation gene hu,/chhdi. (%ron/oso,ntr 98, 81-85. Thomsen, G., Woolf, T., Whitman, M., Sokol, S., Vaughan, J., Vale, W., and Melton, D. A. (1990). Activins are expressed early in Xf,rro/>?l,s embryogenesis and can induce axial mesoderm and anterior structures. (%I/ 63, 485-493. Vaidya, T. B., Rhodes, S. J., J., T. E., and Konieczny, S. F. (1989). Fibroblast growth factor and transforming growth factor p repress transcription of the myogenic regulatory gene MyoDl. Mol. &ll. B&l. 9, 3576-3579. Wanaka, A., Johnson, E. M., and Milbrandt, J. (1990). Localization of FGF receptor mRNA in the adult rat central nervous system by iv sit!{ hybridization. ,l’(~,~rr,r/ .5, 267-281. Wanaka, A., Milbrandt, J., and Johnson, E. M., Jr. (1991). Expression of FGF receptor gene in rat development. D~wlopnrcrcf 111,455-468. Wilkinson, D. G., Peters, G., Dickson, C., and McMahon, A. P. (1988). Expression of the FGF-related proto-oncogenc it/f-L during gastrulation and neurulation in the mouse. E&‘BO J. 7, 69-695. Wilkinson, D. G., Bhatt, S., McMahon, A. P. (1989). Expression patterns of the FGF-related proto-oncogene iuf-2 suggests multiple roles in fetal development. Dc~w~oprr~c~rcf105, 131L136. Yayon, A., Klagsbrun, M.. Esko, J. D., Leder, P., and Ornitz, D. M. (1991). Cell surface, heparin-like molecules are required for binding of basic fibroblast growth factor to its high affinity receptor. (‘(~11 64, 841L848.