Download - Inductive effects of fibroblast growth factor and - Development

Development 103, 581-590 (1988)Printed in Great Britain © The Company of Biologists Limited 1988

581

Inductive effects of fibroblast growth factor and lithium ion on Xenopus

blastula ectoderm

J. M. W. SLACK, H. V. ISAACS and B. G. DARLINGTON

Imperial Cancer Research Fund, Developmental Biology Unit, Department of Zoologv, University of Oxford, South Parks Road, OxfordOX1 3PS, UK

Summary

We have studied the response of Xenopus blastulaectoderm to fibroblast growth factor and to lithiumion. The properties of acidic and basic FGF are verysimilar showing a 50 % induction level at l - lngml" 1

and a progressive increase of muscle formation up toconcentrations of 100-200 ng ml"1. The elongation ofexplants also shows a dose-response relationship. Theminimum contact requirement for induction of ecto-derm explants is about 90 min and the stage range ofectodermal competence extends from midblastula toearly gastrula, both these figures resembling thoseobtained in embryological experiments with vegetaltissue as the inducer. Lithium chloride concentrationswhich produce anteriorization of whole embryos haveno effect on isolated ectoderms unless accompanied byFGF. Simultaneous treatment with FGF and Li lead to

a marked enhancement of both elongation and muscleformation over that produced by FGF alone. Bycontrast, ventral marginal explants show increasedelongation and muscle formation if treated with lith-ium alone suggesting that they have already received alow-dose FGF treatment within the embryo. It isconcluded that endogenous FGF may be solely respon-sible for inducing the ventral mesoderm and thatdorsalization of ventral mesoderm to the level ofsomitic muscle might be achieved either by a very highlocal concentration of FGF in the dorsal region, or bythe action of a second, synergistic, agent in the dorsalregion.

Key words: heparin-binding growth factors, fibroblastgrowth factors, lithium ion, mesoderm induction, muscleinduction, organizer, dorsalization.

Introduction

The nature of the morphogens that control the spatialpattern of specification in the early embryo has longbeen a mystery. Interest in this problem was recentlystimulated by the discovery of a Xenopus cell linewhich secretes a mesoderm-inducing factor (Smith,1987), and shortly afterwards we reported that a smallgroup of pure heparin-binding growth factors(HBGFs) were active as mesoderm-inducing agentswhen tested on isolated ectoderm from Xenopusblastulae (Slack et at. 1987). They were basic fibro-blast growth factor (bFGF), acidic fibroblast growthfactor (aFGF) and embryonal carcinoma derivedgrowth factor (ECDGF). More recently, Kimelman& Kirschner (1987) have obtained evidence for thepresence of an mRNA of the bFGF type in Xenopusembryos. Together these results suggest that FGFmay be a morphogen responsible for inducing meso-derm during early Xenopus development, but in

order to assess the possible role of the endogenousFGF we need an accurate knowledge of the responseof competent cells. In the present paper, we havemade further studies with both a and bFGF, whichgive essentially identical results, and examined thefollowing features: (i) the amount of muscle formedat different concentrations, measured both histologi-cally and by Western blotting, (ii) the elongation ofthe explants that occurs while control embryos aregastrulating, (iii) the minimum contact time requiredfor induction and (iv) the range of stages over whichthe ectoderm remains competent. The results are allconsistent with a role for endogenous FGF as aninducer of ventral-type mesoderm in vivo.

However, one morphogen is not enough to accountfor the formation of pattern over a 2- or 3-dimensional embryo. A substance emitted from asource of any shape and spreading by diffusion wouldbe expected to induce concentric rings of tissue to

582 J. M. W. Slack and others

similar states of specification, the cell state mappingone-to-one onto the concentration gradient. Thespecification map of the early Xenopus mesoderm isnot like this and shows a localized region of extremedorsal character called the organizer and an extendedring of extreme ventral character (later formingmesenchyme and blood cells). The main intermediatemesodermal-derived tissue, the skeletal muscle of themyotomes, is, according to our results, induced laterby a secondary interaction between these two initialmesodermal primordia (Slack & Forman, 1980; Dale& Slack, 19876).

In the present paper, we show that LiCl, an agentwhich perturbs several second message pathways andhas a striking anteriorizing effect on whole embryos(Kao et al. 1976; Cooke & Smith, 1988), has anactivity complementary to the HBGFs in that it willdorsalize ventral mesoderm but not mesodermalizeectoderm. Application of the two agents together toectoderm explants will produce mesoderm inductionswith a high muscle content. Although we do not wishto suggest that lithium is a natural morphogen we feelthat these results strengthen the case for the existenceof morphogens additional to FGF and that thepattern of territories in the blastula becomes specifiedin response to a combination of morphogen concen-trations.

Materials and methods

FGFa and bFGF used for biological experiments were preparedfrom bovine brain as described by Gospodarowicz et al.(1984), Esch et al. (1985) and Lobb et al. (1986). The stepsare: ammonium sulphate precipitation, CM-Sephadex,Heparin Sepharose, and Mono S FPLC. After this theaFGF appeared completely pure by reverse-phase HPLCand by amino acid composition analysis; the bFGF stillcontained some minor impurities which could be resolvedby reverse-phase HPLC. The concentrations were deducedfrom the amino acid analyses of reverse-phase-purifiedmaterial and enabled us to determine the specific activitiesof both forms of FGF as 0-5-lxlO6 inducing unitsmg~'.For mesoderm-inducing factors generally, we have defined1 unit ml"1 as the concentration which is just sufficient tocause induction of an explant, or, if a population of explantsis treated, is sufficient to cause induction in 50 % of explants(Cooke et al. 1987; Godsave et al. 1988). Although somedifferences between a and bFGF have been reported intheir capacity as endothelial cell mitogens (Gospodarowicz& Cheng, 1986), their properties as inducing factors appearidentical.

Other substancesDibutyryl cAMP, dibutyryl cGMP, TPA and A23187 werepurchased from Sigma.

Western blotsThese were used for measuring the myosin content ofexplants induced by FGF. The samples, representing oneexplant, were separated on a 5 % gel (Laemmli, 1970),equilibrated for 30min in transfer buffer (Towbin et al.1979) containing 0-1 % SDS, and transferred to nitrocellu-lose for lih in a Hoefer Transphor apparatus. Thenitrocellulose sheets were blocked for 1 h in 0-15M-NaCl,OlM-Tris-HCl pH7-4 containing 1% w/v Milk powderand then stained with a 1/200 dilution of an antibody toXenopus muscle myosin (see Dale et al. 1985) in the samebuffer for lh. The form(s) of myosin recognized by thisantibody start to accumulate in somitic muscle from aboutstage 35 and are abundant by stage 40-41 when the culturesare terminated. The blots were washed 3xl0min in thesame buffer without milk but with 01 % Tween 20, thenreacted with 1/100125I-donkey anti-rabbit (AmershamInternational) in the same buffer followed by severalwashes, drying and autoradiogTaphy. The film was pre-flashed to an optical density of 01-0-2 and the bandsquantified by scanning with a Gelman DCD-16 gel scanner,or cutting out from the blot and counting in a gamma-counter. The procedure was calibrated by including in eachrun a set of tracks containing known amounts of Xenopusmyosin. This corrects for losses during transfer and for non-linearities in antibody binding or autoradiography andallows for quantitative comparison between inducedexplants and explants of somitic muscle from controlembryos. Where protein content of samples was measuredthis was by the Folin method on dialysed gel samples.

Embryological methodsXenopus embryos were obtained by artificial fertilization.Methods for fertilization, composition of salines, histologi-cal techniques and inducing factor assays are all given inGodsave et al. (1988). The stage series is that of Nieuwkoop& Faber (1967). Animal pole explants were dissected asdescribed in Dale et al. (1985) and ventral marginal zoneexplants, following vital staining of the dorsal side, as inDale & Slack (1987b). Embryos were dissected at latestage 7 (256cells) or early stage 8 (512cells). 2h treatmentswith LiCl or FGF were at room temperature (22°C) andculture at 25 °C.

For the measurement of lithium uptake, groups of 20whole animal hemispheres were exposed for 2 h. They werethen rinsed twice in NAM, homogenized in 0-1 ml 10%TCA and microfuged. The Li content of the supernatantwas measured by atomic emission spectroscopy using apreparation from unexposed explants as a blank. Thevolume of tissue was estimated by measurement of totalprotein in the TCA pellet and comparison with the proteincontent of a whole embryo (volume 1-44^1) processed inthe same way. The derived intracellular lithium concen-tration was 6-4 mM. Protein was measured by the Folinmethod.

Results

Dose-response curve for FGF

In our previous publication, we presented a dose-

Dose-response

0-17 0-5 1-7 5 17 50aFGF(ngmr')

Elongation

0-9-

0-8-EEs:

5 0-7'

0-6-

B

A

i20 40 60

aFGF(ngml-')100

Fig. 1. Dose-response curves for Xenopus ectodermsexposed to bovine aFGF. (A) The percentage of explantsinduced (filled circles) and the average myosincontent/explant (filled triangles). Data compiled from 111explants from 4 egg batches. (B) Lengths of explants afterovernight culture (control stage 22). Each pointrepresents a mean for 6 explants with standard errorshown.

response curve for bFGF (Slack et al. 1987). Here weshow results for aFGF (Fig. 1A) but with a muchmore accurate quantification of the amount of musclefound in the inductions. It is striking that the dose-response relationships of the acidic and basic forms ofFGF seem to be essentially identical. The upper curveshows the proportion of explants forming swollenvesicles after 3 days of culture. Histological studieshave shown that unswollen explants contain onlyatypical epidermis while swollen ones always containsome mesoderm, even if it is only a small amount ofmesenchyme and mesothelium. There is a very steeprise from 0 % to 100 % with 50 % of explants becom-

Inductive effects of fibroblast growth factor and lithium ion 583

ing induced at a concentration of l-2ngml ', orabout 100 pM. The lower curve shows the amount ofmuscle formed, measured by Western blotting (alsosee Fig. 3A,D). It may be seen that there is a plateauregion from about 2 to ISngml"1 over which most ofthe explants are induced but the muscle content is lowand above this concentration the muscle content risessteeply. Using the same method we have found thatthe myosin content of pure somitic muscle from 3-dayembryos is about 33 % of total protein and, taking theaverage protein content of an explant to be 7/xg, thisenables us to deduce that the average muscle contentof the explants treated at rJOngml"1 is 24%. Athigher concentrations than this some inhibition ofdifferentiation is found so the maximum musclecontent inducible with FGF is probably around thisvalue. We have also measured the muscle content ofmany specimens histologically by tracing onto graphpaper with a camera lucida. The outline of the muscleblock and of the whole specimen was traced fromevery fifth section and the volume of the muscle wasexpressed as a proportion of the whole specimen.This method gives a maximum muscle content ofabout 10 %, and since differentiated explants containmuch empty space we feel that this is consistent withthe 24 % figure from the biochemical measurements.It is known that the pigmented surface of the explantsis impermeable to inducing factors (Cooke etal. 1987)and so probably quite a high proportion of theexposed cells become committed to form muscle as aresult of high-dose exposure to FGF. Whether suchdoses can be regarded as physiological is anothermatter which is discussed below. In contrast to theubiquitous appearance of mesenchyme and meso-thelium, and the frequent appearance of muscle, onlyone case out of several hundred examined histologi-cally has contained any notochord even though doseshave ranged as high as 1 jugml"1. In this regard, FGFseems to differ in its biological effects from themesoderm-inducing factors of the XTC/WEHI classwhich induces notochord reproducibly (Smith, 1987;Godsaveer al. 1988).

Fig. IB shows the mean lengths of a group ofexplants exposed to aFGF overnight. Previous workhas shown that the elongation somewhat resemblesthe gastrulation movements of an intact embryo andrepresent the earliest known marker of mesoderminduction (Symes & Smith, 1987). In the case of FGF,we have found that the elongations commence duringgastrulation (from stage 10£) and may continue dur-ing the neurulation and subsequent elongation ofcontrol embryos. The length at a given time alwaysincreases with dose, probably reflecting a greaterproportion of induced cells within the explant. How-ever, the absolute values of the elongations showconsiderable variability both between individual


Contact time Start of competence

Stage

cases and between egg batches and are therefore lessuseful as quantitative measures of the response thanmuscle formation.

Ventral marginal zone (VMZ) explants from stage-7 to -8 embryos, which normally self-differentiate intoventral-type mesoderm, were also treated with aFGF(17ngml~') but showed no change in shape and noextra muscle formation.

Competence of the ectodermOur early experiments involving short-term treat-ments with FGF gave somewhat variable resultswhich were eventually ascribed to absorption of FGFonto the agar bases of the dishes used. When dishesare coated with electrophoretic grade agarose thisproblem does not arise and the results presented herewere obtained in this way. The minimum contactrequirement was determined by exposing stage-8ectoderm explants for different times then rinsing inNAM, culturing and scoring for vesicles. The resultsare shown in Fig. 2A and it may be seen that theproportion of inductions rises to a plateau after about90min. This is very similar to the estimate of1-5-2-5 h for the minimum contact time using vegetalpole tissue as the inducer (Gurdon et al. 1985), andsomewhat longer than the estimate of lOmin fromCooke et al. (1987) using conditioned medium fromXTC cells.

To determine the stage of onset of competence,ectoderm explants of different stages were exposed tobovine FGF for 90min (Fig. 2B). This showed thatthe response was feeble for the early stages but rose

Fig. 2. Competence of ectoderm to respondto 25ngml"' bFGF. (A) % inductionfollowing exposure of stage-8 ectodermexplants for different times (123 explants,3 egg batches). (B) % induction followingexposure of different stage ectoderms for90min (158 explants, 4 batches). (C) %induction following long-term exposure ofdifferent stage ectoderms (329 explants,15 egg batches).

to a high plateau by stage 7 (128 cells). To determinethe time of cessation of competence, explants fromdifferent stages were exposed to FGF indefinitely,and this showed that the major drop came betweenstages 9 and 10, around the beginning of gastrulation(Fig. 2C). Once again these results show good agree-ment with embryological experiments as Jones &Woodland (1987) have recently used heterochroniccombinations to estimate the commencement of com-petence between stages 6 and 7 and its disappearancebetween stages 10 and 10-5. We have also measuredthe muscle content of many of these specimens byhistological methods and found that the highestvalues for a given concentration of FGF were foundfor specimens treated between stage 7 and 9.

Effects of lithium ionLithium ion has long been known to have interestingeffects on the morphology of many embryo typesincluding amphibians (Lehmann, 1937; Masui, 1956)and it was thought desirable to re-examine its effectnow that the biochemical basis of early regionalspecification is being uncovered.

The standard solution of Li was one in which theNaCl of the NAM salts is entirely replaced by LiCl.The Li concentration is thus 0-1 M although sodiumion is still present at 9-2 mM from the sodium phos-phate buffer. This solution is toxic in the long termand so treatments were given for 2h at 22 °C whichhad no ill effects on cell viability or differentiation.

Several batches of intact embryos were treated inthis way starting at stages from 32 to 512 cells to see


1 2 3 4 5 1 2 3 4 5 6

P.1 2 2 3 4 5 6 7

Fig. 3. Synthesis of myosin byXenopiis blastula explants cultured for3 days. Bands are Xenopus musclemyosin heavy chain visualized byWestern blot using anti-myosin and125I anti-rabbit antibodies. Althoughgroups of explants were treated, eachlane contains an amount of sampleequivalent to one explant. Treatmentswere commenced at stage 7.(A) Lanes 1-5: Ectoderms treatedwith aFGF at 170, 50, 17, 5 andl-7ngml~'. (B) Lane 1: Ectodermcontrol; 2, 3: aFGF17ngmr' , 2h;4,5: LiC10-lM, 2h; 6: aFGF17ngmr' + LiC10lM, 2h.(C) Lane 1: ventral marginal zoneexplants, untreated. Lane 2: VMZexplants, 2h 0-lM-LiCl.(D) Lanes 1-7: Myosin standards 1000,750,500,250, 100, 50, 10 ng.

Fig. 4. A batch ofembryos treated for 2 hwith 0-lM-LiCl starting atstage 7. The embryos havebeen cultured for 3 daysand the majority showsevere anteriorization.

whether the 'lithium syndrome' described by Kao etal. (1986) could be reproduced. The results of thiswere extremely variable from batch to batch. Someshowed no effect at all, some showed only a few mildcases while others showed 100 % extreme anterioriza-tion (see Fig. 4). This variability was not due todifferences in stage at the beginning of treatment andwas not shown in explant experiments, suggestingthat its cause lies in a variation of ease of penetrationof the embryo's outer surface.

Ectoderm explants receiving the standard Litreatment were entirely unaffected and, like controls,formed only epidermis. However Li + FGF together

gave rise to extreme elongations and high levels ofinduced muscle (Figs 3 and 5). In terms of muscleformation, the lithium had an effect equivalent toraising the FGF concentration by a factor of about 10.The lithium enhancement is shown with both a andbFGF. Although simultaneous treatment was mosteffective, an effect was also found when the Li wasgiven for 2h before or after a 2h treatment with theFGF. The mechanism of Li action is still not reallyunderstood and it may affect more than one pathway,however the lack of requirement for simultaneoustreatment suggests that the responses to both FGFand Li probably have a time course of an hour or two

586 /. M. W. Slack and others

Fig. 5. Elongation of ectoderm explants treated at stage 7 and cultured overnight (control stage 22). (A) Untreated.(B)0-lM-LiC12h. (C)aFGF nngml"1 2h, showing slight elongation. (D) aFGF + LiCl 2h, showing substantialelongation. Bar, 0-4mm.

rather than a minute or two.In order to obtain an estimate of the intracellular

lithium concentration, 20 isolated animal hemi-spheres were exposed for 2h and then their lithiumcontent was measured by atomic emission spec-troscopy. This gave a value of 6-4 ITIM, a little higherthan that measured by Beckenridge et al. (1987) forintact embryos. The intracellular concentration isevidently much lower than the extracellular level of0-1 M which suggests that even the blastocoelic surfaceof blastula cells presents a considerable barrier to Lientry, but that the intracellular concentration isprobably high enough to affect both inositol lipid andcAMP metabolism (Drummond, 1987).

Although Xenopus ectoderm fails to respond toLiCl alone, explants of ventral marginal tissue re-sponded in a spectacular fashion giving extremeelongated structures and later forming almost solidmasses of muscle (Figs 3 and 6). It would appear,therefore, that ventral marginal zone has alreadyexperienced the equivalent of a low-dose FGF treat-ment, which is entirely consistent with the typicaldifferentiation patterns of such explants, differingfrom FGF inductions only in the additional presenceof blood cells.

Other factorsIn attempts to approach the second message level ofmechanism, the following compounds have beentested alone or with aFGF: diBu cAMP and diBucGMP (3-100 fig ml"'), A23187 (0-5 ̂ g ml"1), TPA(lOOngmP1), A23187 + TPA, but none have yielded

any inductions alone or enhancements of elongationor muscle formation with FGF.

Discussion

The results presented above on the dose-responsecurves for proportion of induced explants, elongationand muscle formation suggest that the inductiveproperties of the bovine FGFs resemble those of theventral and lateral parts of the vegetal hemisphere ofthe blastula (Dale & Slack, 19876). Furthermore, theresults obtained for minimum contact time, onset ofcompetence and decay of competence in the ecto-derm are similar to those obtained in the analogousembryological experiments (Gurdon et al. 1985; Jones& Woodland, 1987). Kimelman & Kirschner (1987)have found a bFGF-like mRNA in eggs and embryosand we ourselves have recently obtained evidence forthe presence of an HBGF protein which closelyresembles bovine bFGF in its specific activity, bio-logical effects and immunological reactivity(J.M.W.S., unpublished data). These considerationsmake us feel that it is legitimate to attempt to explainevents within the embryo by reference to in vitroexperiments with the heterologous factors. So, towhat extent can we explain the mesoderm inductionprocess by the known properties of the FGFs?

Mechanism of mesoderm inductionOn the basis of embryological experiments we havedistinguished three component events: induction ofthe ventral mesoderm, induction of the organizer and


Fig. 6. Dorsalization of ventral marginal zone explants by treatment with LiCl at stage 7. A,B after overnight culture;C,D after 3-day culture. (A) Control, bar, 0-5 mm. (B) Treated 2h LiCl showing substantial elongation, bar, 0-5 mm.(C) Section of control showing ventral mesoderm, bar, 50 fim. (D) Section of Li-treated VMZ showing massive muscleblock, bar, 50 JOTI.

dorsalization of part of the ventral mesoderm by theorganizer (reviewed Smith et al. 1985). A probablerole for the endogenous FGF is as the inducer of thering of ventral-type mesoderm extending over theventrolateral 270° or more of marginal zone circum-ference in the blastula. Experiments involving iso-lated ectoderm have shown that exposure to lowdoses of purified FGF alone is sufficient to inducemesenchyme and mesothelium although formation ofthe blood islands may involve a further influence fromthe endoderm (Maufroid & Capuron, 1985). The factthat ventral marginal zone explants can be dorsalizedby lithium alone suggests that they have alreadyexperienced the equivalent of a low-dose FGF treat-ment in the embryo.

On the other hand, it is probable that the endogen-ous FGF is not responsible for the induction of theorganizer. The organizer is defined as that part of themesoderm that will differentiate partly into noto-chord and that will dorsalize other mesoderm (Slack,1983), and it has been shown that it is induced only byendoderm from the most dorsal part of the blastula(Gimlich & Gerhart, 1984; Dale & Slack, 19876).Since FGF will not reliably induce notochord at anyconcentration, it is likely that the induction of theorganizer is due to something else. In our hands, bothLiCl and TGF/J together with FGF have yielded a few

notochord-containing inductions but this effect is notreproducible, so if the endogenous FGF is respon-sible for induction of the organizer it probably does soin conjunction with some other as yet unknownagent. Alternatively the organizer may perhaps beinduced by a mesoderm-inducing factor of theXTC/WEHI class (Smith, 1987; Godsave etal. 1988),which now seems likely to be another molecule of theTGF£ family called TGF0-2 (Rosa et al. 1988). It hasbeen shown that an organizer can be induced byXTC-MIF (Cooke etal. 1987) although there is as yetno evidence for the presence of such factors in theembryo.

The third process that we have defined by embryo-logical experiments is the dorsalization of the ventralmesoderm by the organizer to form striated muscleand pronephric tubules. Could endogenous FGF alsobe responsible for this? Because of the wealth ofmolecular and immunological markers that have beenavailable (Dale et al. 1985; Gurdon et al. 1985; Jones& Woodland, 1987) the discussion is most easilyconducted in terms of striated muscle. All of themuscle arising up to the tailbud stage in an amphibianembryo comes from the somites (Mohun et al. 1980,1984). Fate-mapping experiments have shown thatmost of the somites derive, in normal development,from the region occupied by blastomeres B2, 3, 4 and


C2, 3, 4 at the 32-cell stage, which becomes theventral and lateral marginal zone of the blastula(Dale & Slack, 1987a; Moody, 1987). But our explan-tation experiments have shown that this tissue is notspecified to form muscle until after the beginning ofgastrulation. Before this stage explants from theventrolateral 270° of marginal zone circumferenceform ventral-type mesoderm with few, if any, musclecells (Dale & Slack, 1987ft). Ventral marginalexplants will, however, form normal, or more thannormal, amounts of muscle when cultured togetherwith the organizer. We have previously called thisphenomenon 'dorsalization' (Slack & Forman, 1980;Smith et al. 1985; Dale & Slack, 1987b), and theresults presented above suggest two possible expla-nations for this process. First, it might be due to asecond exposure of the prospective somite cells to ahigh local concentration of FGF in the dorsal regionor, alternatively, the organizer may emit a distinctchemical signal with dorsalizing consequences, whichcan be mimicked by LiCl or TGF/3.

For the first possibility, our present results showthat high doses of FGF alone can lead to theformation of substantial amounts of muscle, this isabout 24% of the cells, and presumably a higherproportion of those exposed on the blastocoelicsurface of the explant. We know from fate mappingthat 20-50 % of cells derived from blastomeres B2, 3,4 and C2, 3, 4 become muscle and if this were solelydue to FGF these cells would have to be exposed to aconcentration of 100-200ngml"1 during dorsal-ization. The most-likely way this could be achievedwould be for most of the endogenous FGF to besequestered to extracellular material in the dorsalmarginal zone. Then, as prospective somite cellsmigrate dorsally during gastrulation, they would en-counter this high concentration and a sufficient pro-portion of them become specified as muscle. A dorsalconcentration of toluidine blue metachromatic extra-cellular material has been detected in Rana blastulaeby Johnson (1977), and it has been shown by Smith etal. (1982) that FGF can remain active when bound toextracellular matrix components.

As far as the second possibility is concerned,Kimelman & Kirschner (1987) suggest that adequatemuscle induction may be obtained in the presence ofTGF/5 and that the corresponding endogenous mol-ecule is the product of the veg 1 gene which shows amoderate sequence homology to human TGF/J(Weeks & Melton, 1987). It may be that bothexplanations are correct and that the organizer doesindeed emit a TGF/J-like factor which provokes ECMsecretion and hence local sequestration of FGF,although the published evidence suggests that there isno preferential dorsal location for the vegl message.A stimulation of ECM secretion by TGF/3 is well

documented in other systems (Massague, 1987).The dorsalizing role of LiCl on VMZ explants is

consistent with its effects on whole embryos (Kao etal. 1986; Cooke & Smith, 1988; our Fig. 4) whichresult in an overrepresentation of the anterior parts ofthe later embryo. Since the anterior parts normallyarise from the dorsal region of the blastula, theimplication is that the dorsal territories become largerthan usual in whole embryos treated with LiCl. Theactual mechanism of Li action at the molecular levelis still not clear. Most current interest has centred onthe inhibition of inositol monophosphatase (Hallcher& Sherman, 1980) which blocks the recycling of InsP3

to inositol. In brain this depresses the effects offactors acting via InsP3 by inositol depletion, but inother cell types it might perhaps enhance them byslowing the breakdown of InsP3 or by potentiating theformation of other inositol polyphosphates withsecond message activity (Drummond, 1987). How-ever, neither FGF nor TGF/J are known to act via thispathway (Magdalno et al. 1986) and a number ofother biochemical effects of lithium have been de-scribed including an inhibition of G proteins (Avissaret al. 1988), competition with Na for Na channels,competition with Mg for Mg-requiring enzymes, inhi-bition of adenyl cyclase and increased turnover ofcatecholamines (Hendler, 1978), so it is not reallypossible at present to make deductions about thebiochemical responses to induction based on thelithium effect alone.

In summary, the endogenous FGF of Xenopusembryos seems a good candidate as the sole agent forthe induction of ventral mesoderm. It is probably alsorequired for the induction of muscle from ventralmesoderm, perhaps together with a second factor.We have also shown that the FGFs and LiCl haveinductive effects on a quite different type of amphib-ian embryo, the axolotl (J.M.W.S. unpublisheddata), suggesting that these results may be generaliz-able to other amphibian early embryos. At presentthere is little evidence allowing us to speculate onwhether HBGFs are also morphogens in highervertebrate embryos, although one of the three factorsthat we have shown to possess inducing activity isECDGF derived from murine embryonal carcinomacells (Heath & Isacke, 1984), and the oncogene int-2,which is expressed in the mesoderm of the earlymouse embryo, has a strong sequence homology toboth forms of FGF (Dickson & Peters, 1987).

We should like to thank Dr A. Willis (MRC Immuno-chemistry Lab., Oxford) for the amino acid compositionanalyses and Mr M. Wormald (Inorganic Chemistry,Oxford) for arranging the lithium measurement.


References

AVISSAR, S., SCHREIBER, G., DANON, A. & BELMAKER,R. H. (1988). Lithium inhibits adrenergic andcholinergic increases in GTP binding in rat cortex.Nature, Lond. 331, 440-442.

BECKENRIDGE, L. J., WARREN, R. L. & WARNER, A. E.(1987). Lithium inhibits morphogenesis of the nervoussystem but not neuronal differentiation in Xenopuslaevis. Development 99, 353-370.

COOKE, J. & SMITH, E. J. (1988). The restrictive effect ofearly exposure to lithium upon body pattern inXenopus development, studied by quantitative anatomyand immunofluorescence. Development 102, 85-99.

COOKE, J., SMITH, J. C , SMITH, E. J. & YAQOOB, M.(1987). The organization of mesodermal pattern inXenopus laevis: experiments using a Xenopusmesoderm inducing factor. Development 101, 893-908.

DALE, L. & SLACK, J. M. W. (1987a). Fate map for the32-cell stage of Xenopus laevis. Development 99,527-551.

DALE, L. & SLACK, J. M. W. (19876). Regionalspecification within the mesoderm of early embryos ofXenopus laevis. Development 100, 279-295.

DALE, L., SMITH, J. C. & SLACK, J. M. W. (1985).Mesoderm induction in Xenopus laevis. A quantitativestudy using a cell lineage label and tissue specificantibodies. J. Embryol. exp. Morph. 89, 289-313.

DRUMMOND, A. H. (1987). Lithium and inositol lipid-linked signalling mechanisms. Trends inPharmacological Sciences 8, 129-133.

DICKSON, C. & PETERS, G. (1987). Potential oncogeneproduct related to growth factors. Nature, Lond. 326,833.

ESCH, F., BAIRD, A., LING, N., UENO, N., HILL, F.,GEUEROY, L., KLEEPER, R., GOSPODAROWICZ, D.,BOHLEN, P. & GUILLEMIN, R. (1985). Primary structureof bovine pituitary basic fibroblast growth factor (FGF)and comparison with the amino terminal sequence ofbovine brain acidic FGF. Proc. natn. Acad. Sci. U.S.A.82, 6507-6511.

GIMLICH, R. L. & GERHART, J. C. (1984). Early cellularinteractions promote embryonic axis formation inXenopus laevis. Devi Biol. 104, 117-130.

GODSAVE, S. F., ISAACS, H. & SLACK, J. M. W. (1988).Mesoderm inducing factors: a small group ofmolecules. Development 102, 555-566.

GOSPODAROWICZ, D. & CHENG, J. (1986). Heparinprotects basic and acidic FGF from inactivation. J. CellPhysiol. 128, 475-484.

GOSPODAROWICZ, D., CHENG, J., Lui, G. M., BAIRD, A.& BOHLEN, P. (1984). Isolation of brain fibroblastgrowth factor by heparin Sepharose affinitychromatography: identity with pituitary fibroblastgrowth factor. Proc. natn. Acad. Sci. U.S.A. 81,6963-6967.

GURDON, J. B., FAIRMAN, S., MOHUN, T. J. & BRENNAN,S. (1985). The activation of muscle specific actin genesin Xenopus development by an induction betweenanimal and vegetal cells of a blastula. Cell 41, 913-922.

HALLCHER, L. M. & SHERMAN, W. R. (1980). The effects

of lithium ion and other agents on the activity of myo-Inositol-1-phosphatase from bovine brain. J. biol. Chem.

255, 10896-10901.HEATH, J. K. & ISACKE, C. M. (1984). PC13 Embryonal

carcinoma derived growth factor. EMBO J. 3,2957-2962.

HENDLER, N. H. (1978). Lithium pharmacology andphysiology. In Handbook of Psychopharmacologyvol. 14 (ed. L. L. Iverson, S. D. Iverson & S. H.Snyder), pp. 233-273. New York: Plenum Press.

IGNOTZ, R. A. & MASSAGU£, J. (1986). Transforminggrowth factor-/3 stimulates the expression of fibronectinand collagen and their incorporation into theextracellular matrix. J. biol. Chem. 261, 4337-4345.

JOHNSON, K. E. (1977). Extracellular matrix synthesis inblastula and gastrula stages of normal and hybrid frogembryos. I. Toluidine blue and lanthanum staining.J. Cell Sci. 25, 313-322.

JONES, E. A. & WOODLAND, H. R. (1987). Thedevelopment of animal cap cells in Xenopus: a measureof the start of animal cap competence to formmesoderm. Development 101, 557-563.

KAO, K. R., MASUI, U. & EUNSON, R. P. (1986).Lithium induced respecification of pattern in Xenopuslaevis embryos. Nature, Lond. 322, 371-373.

KIMELMAN, D. & KIRSCHNER, M. (1987). Synergisticinduction of mesoderm by FGF and TGF/J and theidentification of an mRNA coding for FGF in the earlyXenopus embryo. Cell 51, 869-877.

LAEMMLI, U. K. (1970). Cleavage of structural proteinsduring assembly of the head of bacteriophage T4.Nature, Lond. 227, 680-685.

LEHMANN, F. E. (1937). Mesodermisierung despraesumptiven Chordamaterials durch Einwirkung vonLithiumchlorid auf die Gastrula von Triturus alpestris.Wilhelm Roux' Arch. EntwMech. Org. 136, 112-146.

LOBB, R. R., HARPER, J. W. & FETT, J. W. (1986).Purification of heparin binding growth factors. Anal.Biochem. 154, 1-14.

MAGDALNO, I., L'ALLEMAIN, G., CHAMBARD, J. C ,MOENNER, M. , BARRITAULT, D . & POUYSSEGUR, J.(1986). The mitogenic signalling pathway of fibroblastgrowth factor is not mediated through phosphoinositidehydrolysis and protein kinase C activation in hamsterfibroblasts. J. biol. Chem. 261, 16916-16922.

MASUI, Y. (1956). Effect of LiCl upon the organizer andthe presumptive ectoderm. Annotationes ZoologicaeJaponiensis 29, 75-78.

MAUFROID, J. P. & CAPURON, A. P. (1985). Ademonstration of cellular interactions during theformation of mesoderm and promordial germ cells inPleurodeles waltl. Differentiation 29, 20-24.

MOHUN, T. J., BRENNAN, S., DATHAN, N., FAIRMAN, S. &GURDON, J. B. (1984). Cell type specific activation ofactin genes in the early amphibian embryo. Nature,Lond. 311,716-721.

MOHUN, T. J., TILLY, R., MOHUN, R. & SLACK, J. M. W.(1980). Cell commitment and gene expression in theaxolotl embryo. Cell 22, 9-15.

MOODY, S. A. (1987). Fates of the blastomeres of the 32cell stage Xenopus embryo. Devi Biol. Ml, 300-319.

590 /. M. W. Slack and others

NIEUWKOOP, P. D. & FABER, J. (1%7). Normal Table ofXenopus laevis. Amsterdam: North Holland.

ROSA, F., ROBERTS, A. B., DANIELPOUR, D., DART, L. L.,SPORN, M. B. & DAWID, I. B. (1988). Mesoderm

induction in amphibians: the role of TGF/S-2 likefactors. Science 239, 783-785.

SLACK, J. M. W. (1983). From Egg to Embryo.Determinative Events in Early Development.Cambridge: Cambridge University Press.

SLACK, J. M. W., DARLINGTON, B. G., HEATH, J. K. &GODSAVE, S. F. (1987). Mesoderm induction in earlyXenopus embryos by heparin-binding growth factors.Nature, Lond. 326, 197-200.

SLACK, J. M. W. & FORMAN, D. (1980). An interactionbetween dorsal and ventral regions of the marginalzone in early amphibian embryos. /. Embryol. exp.Morph. 56, 283-299.

SMITH, J. C. (1987). A mesoderm inducing factor isproduced by a Xenopus cell line. Development 99,3-14.

SMITH, J. C , DALE, L. & SLACK, J. M. W. (1985). Cell

lineage labels and region specific markers in theanalysis of inductive interactions. /. Embryol. exp.Morph. 89 Supplement, 317-331.

SMITH, J. C , SINGH, J. P., LILLQUIST, J. S., GOON, D. S.& STILES, C. D. (1982). Growth factors adherent to cellsubstrate are mitogenically active in situ. Nature, Lond.296, 154-156.

SYMES, K. & SMITH, J. C. (1987). Gastrulationmovements provide an early marker of mesoderminduction in Xenopus laevis. Development 101,339-349.

TOWBIN, H., STAEHELIN, T. & GORDON, J. (1979).Electrophoretic transfer of proteins frompolyacrylamide gels to nitrocellulose sheets: procedureand some applications. Proc. natn. Acad. Sci. U.S.A.76, 4350-4354.

WEEKS, D. L. & MELTON, D. A. (1987). A maternalmessenger RNA localized to the vegetal hemisphere inXenopus eggs codes for a growth factor related toTGF(3. Cell 51, 861-867.

{Accepted 8 April 1988)