Regulation of the fibroblast growth factor receptor in early Xenopus ...

Proc. Nadl. Acad. Sci. USA Vol. 87, pp. 8365-8369, November 1990 Developmental Biology

Regulation of the fibroblast growth factor receptor in early Xenopus embryos (mesoderm induction/tyrosine kinase/embryonic development/Ca2+ efflux/activin)

THOMAS J. MUSCI*t, ENRIQUE AMAYA*, AND MARC W. KIRSCHNER* *Department of Biochemistry and Biophysics and the San Francisco, CA 94143

tDepartment of Obstetrics, Gynecology and Reproductive Sciences, University of California,

Contributed by Marc W. Kirschner, July 26, 1990

Recent evidence suggests that fibroblast ABSTRACT growth factor (FGF) is a primary mesoderm inducer in Xenopus development. We have isolated a full-length cDNA clone for the Xenopus FGF receptor. Like other FGF receptors, the Xenopus homolog is a membrane-spanning protein with a split intracellular tyrosine kinase domain. The Xenopus FGF receptor mRNA is present as a maternal message whose levels are constant through early development. There is no specific regional localization of the transcript by analysis of FGF receptor mRNA levels in microdissected embryonic tissue. In isolated animal-pole blastomeres, FGF receptor mRNA declines over 16 hr in culture and this loss can be prevented by incubation with FGF or activin. Despite the presence of the FGF receptor mRNA in the oocyte, oocytes in culture do not respond to added FGF. However, injection of exogenous Xenopus FGF receptor transcripts into oocytes does generate a functional response to FGF. Our data suggest that posttranscriptional mechanisms regulate the FGF receptor in the oocyte and early embryo and further suggest that mesoderm-inducing factors influence receptor mRNA levels during the time of early tissue formation.

Embryogenesis in vertebrate organisms depends on inductive interactions between the endoderm and ectoderm to form mesoderm. In frogs, this primary patterning event begins prior to gastrulation, when signals emanating from cells in the vegetal region cause the formation of prospective mesoderm in cells of the equatorial region. In Xenopus laevis, there is evidence that polypeptide growth factors, in particular members of the fibroblast growth factor (FGF) and transforming growth factor p (TGF-f3) families, are the inducing agents. Purified basic FGF (bFGF), as well as other members of the FGF family, will induce several types of mesodermal tissues (e.g., muscle) in vitro (1, 2). Members of the TGF-,8 family also induce mesoderm, though the response is quantitatively and qualitatively different from that of FGF (3, 4). Maternal mRNA and abundant levels of bFGF protein are present in the Xenopus oocyte and early embryo (5, 6). Although these observations, particularly for FGF, establish that several parts of the signaling system exist in the earliest stages of Xenopus development, a satisfactory understanding of the role of growth factors in mesoderm induction is still lacking. To use an extracellular factor as a signal for induction, the embryo must express a receptor for the factor. Gillespie et al. (7) have reported the high-affinity binding of labeled acidic FGF to Xenopus embryos and increased binding in regions that are destined for a mesodermal fate (7). It is not clear whether such regional binding represents binding to transmembrane receptors or binding to other cellular components, especially matrix components, for which FGF is known to have high affinity (8). Bona fide transmembrane receptors for The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. 8365

FGF recently have been cloned and characterized in the chicken (9), human (10), and mouse (11, 12) and shown to encode an extracellular domain for FGF binding and an intracellular tyrosine kinase. We now have isolated a cDNA for the Xenopus homolog of the FGF receptor: and studied its regulation.

MATERIALS AND METHODS Isolation of cDNA Clones. A Agt1O library of Xenopus oocyte cDNA (13) was screened at low hybridization stringency using a 32P-labeled 400-base-pair (bp) fragment of the region encoding the extracellular domain of the chicken FGF receptor (ref. 9; a gift of P. Lee and L. T. Williams, University of California, San Francisco). A 1.6-kilobase (kb) clone was isolated from a purified phage and used to rescreen the same library at higher stringency to obtain a more complete clone of length 2.9 kb. To obtain 5' sequence further upstream than the 2.9-kb clone, single-stranded cDNA was produced from embryonic RNA, as described by Frohman et al. (14). Five micrograms of total RNA from stage 11 embryos was reverse-transcribed using 20 pmol of a 22-base antisense oligonucleotide from the extracellular domain 3' to the first immunoglobulin-like region (5'-CATCATCTTCATCATCGTCCTC-3') with 100 units ofMoloney murine leukemia virus reverse transcriptase (Bethesda Research Laboratories) for 2 hr at 41'C. Excess primer was removed with the GeneClean kit (BiolOl, San Diego, CA) and products were tailed with dATP by using 15 units of terminal deoxynucleotidyltransferase (Bethesda Research Laboratories) as described (14). Tailed singlestranded cDNAs were then amplified by the polymerase chain reaction (PCR) with the following primers: Xenopus FGF receptor 5' (5'-GGATCTCCTCCCCTGTTATGCG-3', 30 pmol); (dT)17 adaptor (10 pmol), and adaptor (25 pmol) as reported by Frohman et al. (14). The DNA thermal cycler (Perkin-Elmer/Cetus) step program was 40 cycles as follows: 94°C, 40 sec; 55°C, 1 min; 72°C, 1 min. One-tenth of the PCR products was electrophoresed in a 2% NuSieve (FMC Products)/1% agarose gel, transferred to Duralon-UV nylon membrane (Stratagene), and hybridized at high stringency with a 32P-labeled, 300-bp EcoRI-HindlIl fragment that represents the most 5' end of the 1.6-kb clone. A single hybridizing band at -450 bp was purified and blunt-end-ligated into the EcoRV site of pBluescript II (Stratagene) for clone amplification and sequencing by published protocols (15). DNA Sequencing. All DNA sequencing was performed on double-stranded, cesium chloride-purified plasmid DNA (15) by using chain-termination methods (16) with Sequenase Abbreviations: FGF, fibroblast growth factor; bFGF, basic FGF; TGF-P, transforming growth factor ,B; PCR, polymerase chain reaction; RACE, rapid amplification of cDNA ends. tThe sequence reported in this paper has been deposited in the EMBL/GenBank data base (accession no. M37201).

8366

Developmental Biology: Musci et al.

(United States Biochemical) according to the manufacturer's protocol. Nucleic acid sequences were entered into EUGENE (Baylor College of Medicine, Houston, TX) for sequence analysis. Constructs for Oocyte Translation. The 2.9-kb Xenopus FGF receptor clone was inserted into a modified pSP64T vector (ref. 17; containing 5' and 3' untranslated sequence of the Xenopus A-globin gene) for use in RNA transcription. The start codon and signal sequence taken from chicken FGF receptor cDNA (gift of P. Lee) was spliced to the 5' end of the 2.9-kb clone. To accomplish splicing, novel restriction sites in the most 5' Xenopus sequence were created by PCR-generated mutagenesis. A 27-nucleotide sense primer beginning at nucleotide + 133 was used in PCR amplifications to convert the proline at position 46 to a leucine residue, creating a Pvu II site matching the homologous position in the chicken sequence. The resulting translated protein included the first 51 amino acids of the chicken FGF receptor fused to amino acids 46-812 of the Xenopus receptor. Constructs were translated in a wheat germ lysate (Promega) with [355]methionine and analyzed by SDS/PAGE (18). 45Ca2' Efflux Assay and Oocyte INjections. Oocytes were obtained and microinjected as described (19). They were incubated for 48 hr at 16'C in modified Barth's saline (MBSH) with bovine serum albumin (1 mg/ml), penicillin, and streptomycin (19). 45Ca2+ efflux was assayed essentially as described by Williams et al. (20). Healthy oocytes were washed in Ca2+-free lx MBSH with bovine serum albumin, penicillin, and streptomycin (solution B) and were transferred to 24-well dishes at 10-15 oocytes per well in 0.5 ml of Ca2+-free MBSH. 45Ca2+ was added to each well at a final concentration of 50 ,Ci/ml (1 p.Ci = 37 kBq) and incubated at 16'C for 3 hr. After 45Ca2+ loading, the oocytes were extensively washed, damaged oocytes were removed, and the remaining oocytes were transferred to clean wells at 8 oocytes per well. Aliquots of medium were removed for scintillation counting of 45Ca2+ at 10-min intervals and replaced with equal amounts of saline. Recombinant Xenopus bFGF (5) was used for all assays at a final concentration of 30 ng/ml. Medium collected at each time point was placed in 5 ml of Ecolume (ICN) scintillation fluid. The average of triplicate samples was expressed as 45Ca2+ cpm released. After 60 min of medium collection, oocyte samples were taken for scintillation counting to normalize for 45Ca2+ loading. Embryo Dissections. Xenopus eggs were obtained and fertilized as described (21). Manual dissections were carried out under a microscope to isolate either animal caps from stage 8 embryos (22) or various regions (animal, marginal, vegetal, dorsal, lateral, and ventral) from a range of embryonic stages. The animal caps were dissected free from underlying marginal and vegetal blastomeres in stage 8 embryos and incubated in modified Ringer's solution (21) with penicillin and streptomycin at 20°C with the addition of the following factors for up to 16 hr: purified porcine activin (gift of W. Vale, Salk Institute), 2 ng/ml; recombinant Xenopus bFGF (5), 30 ng/ml; and TGF-,61 and TGF-f32 (R+D Systems), 20 ng/ml. At selected times of incubation, animal caps were removed and frozen in liquid nitrogen for later isolation of total RNA. Isolation and Analysis of RNA. Animal-cap explants, embryo dissections, or whole embryos were homogenized and RNA was isolated as described (23) for RNase protection analyses (17). RNase protection analysis was performed on each sample with a specific probe for the Xenopus FGF receptor and either Xenopus elongation factor la (24) or cardiac-specific actin (which can also be used to detect cytoskeletal actin; ref. 25) as controls. A 300-base antisense transcript (EcoRI-HindIII fragment) of the Xenopus FGF receptor cDNA in pBluescript II vector was synthesized in the presence of [32P]UTP with T3 RNA polymerase (Strata-

Proc. Natl. Acad. Sci. USA 87 (1990)

gene). Hybridizations were done in the presence of 80% formamide at 450C for 16 hr. Samples were then subjected to digestion with RNase A (Sigma) at 10 pug/ml for 40 min at 370C. Following proteinase K digestion of RNase and precipitation of RNA, samples were electrophoresed in an 8% acrylamide/urea denaturing gel as described (17).

RESULTS Xenopus FGF Receptor Sequence. Initial screening of the oocyte cDNA AgtlO library with a heterologous probe resulted in the isolation of a 1.6-kb cDNA that, by homology to the chicken FGF receptor (9), represented approximately three-fourths of the coding region of the mature protein and lacked coding sequences on both the 3' and the 5' end. This clone was then used to rescreen the oocyte library under stringent conditions, yielding a 2.9-kb cDNA. This cDNA contained a 2.2-kb open reading frame including, by comparison with the chicken sequence, the entire tyrosine kinase and cytoplasmic tail but did not contain a 5' ATG start codon and signal sequence. At the 3' end it had a termination codon following the kinase domains in a position homologous to the chicken sequence. To obtain the complete 5' end of the gene we used a PCR protocol termed RACE (rapid amplification of cDNA ends) as reported by Frohman et al. (14). Singlestranded cDNA was generated from (stage 11) embryonic RNA primed with an antisense oligonucleotide from the 5' region of the 2.9-kb clone. Following PCR amplification using the cDNA as template, a 450-bp fragment was purified from RACE reaction products. This fragment included an ATG start codon followed by a deduced hydrophobic signal sequence and the previously cloned 5' end of the Xenopus FGF receptor cDNA. The entire deduced amino acid sequence is shown in Fig. 1. The 5' overlap region obtained from the RACE clone was found to encode the same amino acid residues as the clones isolated from the oocyte library except for four differences. These differences probably reflect allelic variation. The open reading frame of 2.4 kb encodes 812 amino acids to yield a protein of =90 kDa. The Xenopus FGF receptor has features identical to those found in the chicken bFGF receptor (9) and is similar in overall structure to other transmembrane growth factor receptors (26). Like the chicken FGF receptor it has three immunoglobulin-like domains in the putative extracellular portion, an unusual region of eight consecutive acidic amino acid residues between the first and second immunoglobulin-like domains, a relatively long juxtamembrane region of 87 residues that separates the transmembrane region from the start of the tyrosine kinase, and a tyrosine kinase sequence that is split by an insertion of 14 amino acids. Comparison with the chicken sequence (Fig. 1) shows overall 81% identity, with the highest levels of sequence similarity in the kinase domain and lower levels in the immunoglobulin-like domains and in the kinase insertion region. One minor difference in the Xenopus cDNA compared with the chicken is the deletion of 3 amino acids in the region between the conserved acidic residues and the second immunoglobulin-like domain. This is also an apparent area of variability between the chicken and human forms that have been analyzed (27). Functional Analysis of Xenopus FGF Receptor. To test whether the isolated cDNA represented a functional form of the FGF receptor, we translated the receptor mRNA in oocytes (20). An assay for the FGF receptor is based on the fact that several transmembrane receptors mobilize intracellular Cal' stores upon addition of the ligand, resulting in an efflux of Ca". This effiux can be measured after preloading oocytes expressing the FGF receptor with 45Ca2+. In our initial studies the 5' signal sequence of the Xenopus clone had not been obtained, and therefore we used an mRNA that was a fusion between the chicken initiation and

Developmental Biology: Musci et al. mouse

MFSGMSLLLWGVLLGAALSVARPPSTLPDEVAPKT-KTEVEPYSAQPGDRITLQCRLRED D LLQ R QAL A-NI SH H TWRC I A VT T A AP D EQAQ WGVPV SLLVH LLQ RV AP WGWKC LF A VT T CT

Xenopus

VQSINWVKNGVQLSETNRTRITGEEIQISNAGPEDNGVYACVTNGPSRTYTVLCSVNVSD

Xenopus

chicken

61

RD LRD

mouse

8000

GSE TYF GSD TYF

M S V SS

VEVRDRV SGL VEVRDSIPA SGL

P N V S

XFR

mouse

121 ALPSAEDDDEDDDNSSSEEKAAENSKPNR----PLWSHPEKMEKKLAVPAAKTVKFRCP K Q-AVA Y TY DE E D T K S RRPVA Y TS ETD T DD -

Xenopus

ANGTPTPTLRWLKNNRAFQQDQRIGGYKVRSQTWSLIMDSVVPSDKGNYTCIVENKYGAI

chicken

SG SS

Xenopus chicken

10000O

*

chicken

.

mouse

GKE KP H GKE KP H

YA YA

E

I

') 4000 + cu

S S

°

241 Xenopus chicken mouse

NHTYQLDVVERSPHRPILQAGLPANTSVTVGTTAEFSCKVYSDPQPHIQWLRHIEINGSR

Xenopus

VASDGFPYVEILKTAGVNTSDKDMEVLHLRNVTFEDAGQYTCLAANSIGISHHSAWLTVL

KT AL SNV V KT AL SNV

HPS

K K

M

6000

x

181 N N

2000' vI

chicken mouse

IGP NL IGP NL

Xenopus chicken mouse

EATEQS

T T

Q

E E

S S

E E

G GN

it~~~

K K

V V

301

Q

10

Y Y

MMT

EALEER VMT

Y Y

V FLI C FLI C LGSV

STT SGT

T S

H

L M

SIPVRRQVTVSGDSSSSMNSGVILVR--RLSSSGTPMLSGLSEYELPEDPRNEVARDRLI

chicken

L L

mouse

A A

M L

A

LP LP

A V A VS

PS PS

V

481

LGKPLGEGCFGQVVMAEAIGLDKEKPNKVTKVAVKMLKSDASEKDLSDLISEMEMKMIG

chicken

L L

mouse


D D

T T

R R

541 KHKNI INLLGACTQDGPLYVIVEYTSKGNLREYLRARRPPAMEYCYNPTCVPDQLLSFKD Q RI EEQ A G Q A SHN EEQ S G 601


LVSCAYQVARGMDYLASKKCIHRDLAARNVLVTEDNIMKIADFGLARDIHHIDYYKKTTN


GRLPVKWMAPEALFDRIYTHQSDVWSFGVLLWEIFTLGGSPYPGVPMEELFKLLKEGHRM

Xenopus chicken

DKPTNCTNELYMMMKDCWHAMPSQRPTFNQLVEDLDRILALSSNQEYLDLSMPVNQYSPC

V V

E E

661 V V 721 mouse


40

50

t Time bFGF added

(min)

H20 H2 XFD

60

KVEDNKPALLASPLQLEII IYCTGAAFVSAMVVTIIIFKMKHPSKKSDFNSQLAVHKLAK

421

Xenopus

--*---

30

20

L

361

Xenopus

8367


S S

R R

V V

K K

V MT H V LT

V LD I LD

G S

FIG. 2. Effect of bFGF on Ca+2 efflux from oocytes injected with Xenopus FGF receptor RNA (XFR), Xenopus FGF receptor cytoplasmic deletion mutant RNA (XFD), or water. All points indicate 45Ca+2 released into the medium and are the means of triplicate wells (8 oocytes per well) from a single representative experiment. Recombinant Xenopus bFGF was added at 30 ng/ml.

hyde/agarose gels (1), we estimate the full-length transcript to be -3.5 kb, which is in agreement with published estimates of the size of the chicken FGF receptor (9). There was no observable change in the size or in the levels of the Xenopus FGF receptor mRNA, indicating either persistence of maternal mRNA throughout gastrulation or new transcription at levels just balancing degradation. To determine the levels of FGF receptor mRNA more accurately through these early stages, RNase protection analysis was performed, normalizing for total RNA per embryo from stage 1 through stage 28. A

781 FPDTRSSTCSSGEDSMFSHDPLPDEPCLPKY----SNGGLKKR V A RCPPH- H A RH D E E V RHPTQLANSG RR

FIG. 1. Comparison of the amino acid sequences of the Xenopus, chicken, and mouse forms of the FGF receptor. The sequences were aligned using the EUGENE program except for the most C-terminal region, where gaps were placed by the authors to preserve homology. Gaps are noted with dashes. Stars above amino acid correspond to differences obtained in the 5'-end clone generated from PCR (see text) (Q -* R, R -e T, I -- V, N -e S). The sequences of the chicken and mouse cDNAs shown are from Lee et al. (9) and Reid et al. (11), respectively.

signal sequence and the Xenopus receptor sequence. As a control to show the requirement for the tyrosine kinase domain, an additional construct was made that had a deletion of the cytoplasmic tyrosine kinase domain. Oocytes injected with Xenopus FGF receptor mRNA (XFR) showed a liganddependent 45Ca2' efflux in response to FGF compared with the baseline steady-state efflux (Fig. 2). The deletion mutant (XFD) and water-injected controls showed no stimulated Ca2' release in response to FGF. All mRNA constructs synthesized peptides of the expected size (data not shown). Developmental Expression and Localization of the FGF Receptor mRNA. We examined the expression and localization of FGF receptor mRNA through early stages of embryonic development. When equal amounts of polyadenylylated RNA were probed with a partial cDNA clone (1.6 kb) of the Xenopus FGF receptor, a single transcript was detected at constant levels from the mature oocyte to late neurula stages (Fig. 3A). Based on DNA markers run in parallel, and the correction for anomalous migration of DNA in formalde-

4.3 -

B

K\

L

Cb

R ('R

t

^

-

*

2.0-

N'l

-N

('

C- ~

C~r

4, R R,

Z~

XFGFR ....iW ....im ~

1.0 0.8 0.8 0.7 0.7 0.8 0.7 0.6 FIG. 3. FGF receptor mRNA levels in the oocyte and early stage embryos. (A) Northern blot analysis of FGF receptor transcript size. Poly(A)+ RNA from 50 oocytes or embryos at each stage was isolated, separated in a formaldehyde/agarose gel, and transferred to nylon membrane. A 1.6-kb cDNA (XFR/B4) probe representing approximately two-thirds of the coding region was hybridized under stringent conditions. Size markers at left are in kilobases. (B) RNase protection analysis of FGF receptor mRNA levels in total RNA from embryos (one-cell stage through stage 28). The Xenopus FGF receptor (XFGFR)-specific band intensity was read by laser densitometry scanning and calculated areas are expressed as the ratio of

stage n to stage 1.

8368

Developmental Biology: Musci et al.


Table 1. Regional FGF receptor mRNA levels relative to whole embryo Relative mRNA level Stage Animal Marginal Vegetal Dorsal Lateral Ventral 4 1.5 1.2 1.3 8 1.0 0.9 1.2 1.7 1.5 1.5 11 0.3 1.3 Laser densitometry scan of each autoradiogram FGF receptor band (calculated area) is normalized to intensity of elongation factor la mRNA and expressed as a ratio relative to FGF receptor mRNA level in whole embryo; -, not assayed.

The total RNA per embryo, which is mostly rRNA, does not change appreciably during this time (28). Our analysis showed that levels of protected FGF receptor message, compared with the level in the one-cell embryo, varied little (Fig. 3B). Though the FGF receptor mRNA levels were relatively constant through early stages, we wished to know whether specific regions of the embryo differed in concentration of message. Total RNA was isolated from portions of embryos dissected at stage 4 (8-cell stage), stage 8 (midblastula), and

A

U) cc%

U1)

>._ C

U1)

C-

0

2

4

8

stage 11 (midgastrula) and FGF receptor mRNA was quantitated by RNase protection. The total level of mRNA in each region was normalized to the amount of elongation factor la mRNA, a transcript that has been shown to be distributed uniformly (23). The normalized levels of FGF receptor RNA were found to be rather uniform in different parts of the embryo at any given stage (Table 1). In animal and vegetal poles in stage 4 embryos and between animal pole, marginal zone and vegetal pole in stage 8 and 11 embryos, the level of FGF receptor mRNA was found to be identical within the -30o error of the RNase protection assay. At the midgastrula stage, when isolation of dorsal, lateral, and ventral tissues is possible, there was again little regional difference in transcript level; at this stage there was a small deficit in the animal region. FGF Receptor mRNA Levels Are Regulated by MesodermInducing Growth Factors. Although there was no appreciable change in FGF receptor mRNA in early stages (Fig. 3), there was a small decrease in the receptor mRNA in the animal cap at midgastrula stage (stage 11). When animal caps were isolated at stage 8, there was a gradual decrease in FGF receptor mRNA after 4 hr in culture (Fig. 4A). By 16 hr, a time at which complete embryos have several somites, there was no detectable FGF receptor mRNA in the isolated animal-cap masses. Animal caps at this stage have formed an epidermis and have increased their complement of cytoskeletal actin. Consistent with their epidermal phenotype, there was no detectable cardiac actin transcription (data not shown). In contrast, treatment with the potent mesoderm inducer activin (4) maintained FGF receptor mRNA at the initial level for up to 16 hr in culture (Fig. 4B and Fig. 5). During this period there was extensive synthesis of cardiac actin mRNA. bFGF is a less potent inducer of muscle than activin and it is generally thought that bFGF induces more ventral forms of mesoderm. Consistent with this we observed a partial loss of FGF receptor mRNA and a stabilization at a lower level than found with activin (Fig. 5). TGF-/31 alone has been shown not to induce muscle (1) and it seemed not to maintain levels of Xenopus FGF receptor mRNA. TGF-,82 had an effect similar to that of activin.

16

DISCUSSION

Time (hrs)

B 151

The ability of specific cells within the embryo to respond to inductive signals is termed competence. In Xenopus, the

* FGFR 1 Cytoskeletal Actin Cardiac Actin

U)

10' activin lY \t____

< 1.0'

6

\ U)

0

2

4

8

16

0

C2? ~ ~ ~ ~

bFGF

-

~ ~~0TGFB1

Time (hrs) FIG. 4. FGF receptor mRNA expression in isolated animal-cap blastomeres in explant culture. Animal caps were dissected at stage 8 and cultured at 16°C for 16 hr in the absence (A) or presence (B) of growth factor (purified porcine activin, 2 ng/ml). Ten animal caps were recovered at each time point and total RNA was isolated. RNA was analyzed using RNase protection for the presence of FGF receptor (FGFR) transcripts (two-thirds of total RNA) and cytoskeletal and muscle-specific (cardiac) actin transcripts (one-third of total RNA). Levels of protected fragments of each message are depicted as intensity of individual bands as read by laser scanning. Stars indicate no detectable mRNA.

Time (hrs) FIG. 5. FGF receptor mRNA levels in animal caps in culture with growth factors. RNase protection analysis was performed with total RNA from seven animal caps in isolated culture for up to 16 hr. RNA was hybridized with Xenopus FGF receptor-specific antisense transcript. Autoradiograms were scanned by laser and band densities are expressed as the ratio of each time point relative to isolated animal caps at zero time. Growth factors were added at the following final concentrations: Xenopus bFGF, 30 ng/ml; porcine activin, 2 ng/ml; TGF-,31 and TGF-f32, 20 ng/ml.

Developmental Biology: Musci et A animal-pole blastomeres in explant culture have a finite window of competence for mesoderm induction, by bFGF, beginning at the 64-cell stage and ending at the 1024-cell stage (29, 30). As reported by Gillespie et al. (7), this period is paralleled by increasing levels of membrane-bound FGF receptor over the first hours of cleavage, reaching a maximum concentration at the 1024-cell stage (7). The appearance of this receptor at the 64-cell stage cannot be caused by new transcription, which only begins in the embryo at the midblastula transition 3 hr later (22). In this paper, we report that the mRNA for the FGF receptor exists as a maternally derived mRNA in the oocyte and fluctuates very little during early development. Despite the presence of FGF receptor mRNA, the oocyte does not seem to be producing functional receptor molecules on its surface, since it shows no response to added bFGF. However, the oocyte seems to have all of the other necessary components of the signaling system, since injection of an in vitro transcribed Xenopus FGF receptor mRNA allows the oocyte to respond to bFGF by rapidly releasing Ca2'. The presence of the receptor mRNA and the absence of functional receptor in the oocyte suggest control of receptor translation or compartmentalization of the translated product. The injected mRNA may not be subject to translational control, since it is lacking untranslated sequences on the 5' end. The level of FGF receptor mRNA is stable through early development. There is no evidence for a loss of receptor mRNA corresponding to the observed loss of competence at the midblastula stage. No change occurs in the level or size of the mRNA after zygotic transcription begins at that stage. The continued presence of the FGF receptor mRNA past the period of competence to form mesoderm is not surprising, since the receptor protein continues to be present, as shown by binding studies (7). The FGF receptor may serve other functions after the loss of competence for mesoderm induction. For example mesodermal cells may require secondary inductions to generate a full range of tissues or may utilize FGF for angiogenesis or mitogenesis. Isolated ventral mesoderm cells from Xenopus embryos continue to respond to FGF with enhanced cell motility long after the period of competence for mesoderm induction (J. Gerhart and M.W.K., unpublished observations). This suggests that competence itself is regulated by processes downstream of the receptor. For example, changes in second-messenger pathways may modify the interpretation of the FGF signal as development proceeds. In explanted animal caps the mRNA for the FGF receptor is maintained by activin and TGF-32; lower levels of Xenopus FGF receptor mRNA are maintained by bFGF. Interestingly, TGF-f31, which causes cells to assume a mesenchymal phenotype and produce extracellular matrix but does not induce muscle, does not prevent the loss of the XGF receptor mRNA. The level of receptor mRNA present in animal caps after 16 hr corresponds best to dorsal mesoderm differentiation rather than general mesoderm formation and may reflect the need for muscle or other dorsal tissues to maintain their ability to respond to bFGF. The loss of the Xenopus FGF receptor mRNA in untreated animal caps may merely reflect the differentiation of cells into epidermis. We cannot say at this time whether the maintenance of the initial levels of Xenopus FGF receptor mRNA is due to stabilization of the mRNA or to new transcription balancing degradation. In summary, the regulation of FGF and its receptor in the early Xenopus embryo is likely to occur on the translational or posttranslational level. At later developmental stages, other signaling pathways may exert their activities directly on the transcription of FGF receptor and FGF mRNA or on the stability of these transcripts. Understanding these biosynthetic and signaling regulatory loops is likely to be important


8369

in understanding the process by which the embryo generates further complexity. We thank Teresita Bernal, Pascal Stein, and Tina Lee for essential expert technical assistance. We are grateful to Drs. P. Lee, D. Johnson, and L. T. Williams for advice and for providing the chicken FGF receptor cDNA clone and to Drs. D. Kimelman and J. Gerhart for helpful discussions during the completion of these studies. This work was supported by the National Institute of General Medical Studies and National Heart, Lung, and Blood Institute (Program of Excellence in Molecular Biology). E.A. was supported by a National Science Foundation graduate fellowship. T.J.M. is the recipient of a Reproductive Scientist Training Award supported by the National Institute of Child Health and Human Development and GynoPharma, Inc.

1. Kimelman, D. & Kirschner, M. (1987) Cell 51, 869-877. 2. Slack, J. M. W., Darlington, B. G., Heath, J. K. & Godsave, S. F. (1987) Nature (London) 326, 197-200. 3. Rosa, F., Roberts, A. B., Danielpour, D., Dart, L. L., Sporn, M. B. & Dawid, I. B. (1988) Science 239, 783-785. 4. Smith, J. C., Price, B. M. J., Van Nimmen, K. & Huylebroeck, D. (1990) Nature (London) 345, 729-731. 5. Kimelman, D., Abraham, J. A., Haaparanta, T., Palisi, T. M. & Kirschner, M. (1989) Science 242, 1053-1056. 6. Slack, J. M. W. & Isaacs, H. V. (1989) Development 105, 146-153. 7. Gillespie, L. L., Paterno, G. D. & Slack, J. M. W. (1989) Development 106, 203-208. 8. Gospodarowicz, D., Ferrara, N., Schweigerer, L. & Neufeld, G. (1987) Endocrine Rev. 8, 95-114. 9. Lee, P. L., Johnson, D. E., Cousens, L. S., Fried, V. A. & Williams, L. T. (1989) Science 245, 57-60. 10. Isacchi, A., Bergonzoni, L. & Sarmientos, P. (1990) Nucleic Acids Res. 18, 1906. 11. Reid, H. H., Wilks, A. F. & Bernard, 0. (1990) Proc. Natl. Acad. Sci. USA 87, 1596-1600. 12. Mansukhani, A., Moscatelli, D., Talarico, D., Levytska, V. & Basilico, C. (1990) Proc. NatI. Acad. Sci. USA 87, 4378-4382. 13. Rebagliati, M. R., Weeks, D. L., Harvey, R. P. & Melton, D. A. (1985) Cell 42, 769-777. 14. Frohman, M. A., Dush, M. K. & Martin, G. R. (1988) Proc. Natl. Acad. Sci. USA 85, 8998-9002. 15. Sambrook, J., Fritsch, E. F. & Maniatis, T. (1989) Molecular Cloning:A Laboratory Manual (Cold Spring Harbor Lab., Cold Spring Harbor, NY), 2nd Ed. 16. Sanger F., Nicklen, S. & Coulson, A. R. (1977) Proc. Natl. Acad. Sci. USA 74, 5436-5467. 17. Melton, D. A., Krieg, P. A., Rebagliati, M. R., Maniatis, T., Zinn, K. & Green, M. R. (1984) Nucleic Acids Res. 12, 70357056. 18. Laemmli, U. (1970) Nature (London) 227, 680-685. 19. Cyert, M. S. & Kirschner, M. W. (1988) Cell 53, 185-195. 20. Williams, J. A., McChesney, D. J., Calayag, M. C., Lingappa, V. R. & Logsdon, C. D. (1988) Proc. Natl. Acad. Sci. USA 85, 4939-4943. 21. Newport, J. & Kirschner, M. W. (1982) Cell 30, 675-686. 22. Nieuwkoop, P. D. & Faber, J. (1967) Normal Table ofXenopus Laevis (Daudin) (North-Holland, Amsterdam). 23. Krieg, P. & Melton, D. (1984) Nucleic Acids Res. 12, 70577070. 24. Krieg, P. A., Varnum, S. M., Wormington, W. M. & Melton, D. A. (1989) Dev. Biol. 133, 93-100. 25. Sargent, T. D., Jamrich, M. & David, I. B. (1986) Dev. Biol. 114, 238-246. 26. Yarden, Y. & Ullrich, A. (1988) Annu. Rev. Biochem. 57, 443-478. 27. Johnson, D., Lee, P., Lu, J. & Williams, L. T. (1990) Mol. Cell. Biol. 10, 4728-4736. 28. Davidson, E. H. (1986) Gene Activity in Early Development (Academic, Orlando, FL). 29. Slack, J. M., Isaacs, H. V. & Darlington, B. G. (1988) Development 103, 581-590. 30. Grainger, R. M. & Gurdon, J. B. (1989) Proc. Natl. Acad. Sci. USA 86, 1900-1904.