Novel HOX, POU and FKH genes expressed during ... - BioMedSearch

3990-3996 Nucleic Acids Research, 1994, Vol. 22, No. 19

.:) 1994 Oxford University Press

Novel HOX, POU and FKH genes expressed during bFGF-induced mesodermal differentiation in Xenopus 'Michael W.King* and Matthew J.Moore Indiana University School of Medicine, Department of Biochemistry and Molecular Biology, Terre Haute Center for Medical Sciences, Terre Haute, IN 47809, USA Received April 19, 1994; Revised and Accepted August 3, 1994

ABSTRACT Cells from the cap of the animal hemisphere of the early Xenopus embryo are determined to form ectodermal lineages. When these cells are explanted and cultured in the presence of various growth factors a change in fate to cells of mesodermal lineage can be observed. Proteins of the fibroblast growth factor (FGF) family belong to this class of fate altering compounds. The ability of FGFs to change animal cap cell fate is in part due to an alteration in the program of genes expressed in these explanted cells. Several genes that are known to be pattern regulating in other systems have been shown to be induced by FGFs in the animal cap assay. We have utilized a PCR-based sib-selection and cloning protocol to identify a large number of cDNAs of the HOX, POU and FKH families that are present in animal caps very early during bFGF-induced mesodermal differentiation. A total of 11 different HOX, 7 POU and 4 FKH cDNAs were identified in an induced animal cap cDNA library. In several cases, pairs of highly related sequence variants were identified that presumably represent expression from the duplicated alleles of the ancestrally tetraploid Xenopus genome. In this report we characterize the temporal and spatial expression of three novel Xenopus genes during early development as well as during bFGF-induced mesodermal differentiation.

INTRODUCTION The ability of one type of cell (or cells) of an early embryo to alter the fate of another cell (or cells) is known as embryonic induction. The earliest detectable inductive event during embryogenesis is that of mesodermal cell types from ectoderm (1,2). This change in fate of ectodermal cells occurs through the release of factors from vegetal cells of the embryo that reside near the marginal zone. This inductive event was first identified and characterized by transplantation experiments with developing embryos (1,3). A small subset of cells that lie near the dorsal lip of the blastopore at the onset of gastrulation exhibit the ability to induce a complete secondary body axis when transplanted to *To whom correspondence should be addressed

+UO4867, U04870 and U04872

EMBL accession nos+

the ventral side of a donor embryo (1). In the developing Xenopus embryo these cells are known as the 'organizer' region. When cultured in isolation, cells from the animal cap region of early Xenopus embryos differentiate into ciliated epidermis indicative of the ectodermal lineage. However, these animal cap cells will undergo a change of fate when cultured in contact with vegetal hemisphere cells (4). Significant amounts of mesodermal cell types are observed including muscle and notochord (4). This mesodermal induction event can be studied in vitro using the explanted animal cap ectodermal cells of the early Xenopus embryo. Proteins of the fibroblast growth factor family including basic (bFGF), acidic (aFGF), KGF and the oncogene int-2 induce predominantly ventral type mesoderm characterized as mesenchyme and mesothelium, although some muscle is also observed (4-10). The results of in vitro mesoderm induction experiments with bFGF are suggestive of a natural role for this growth factor in normal differentiation of the mesoderm. Subsequent studies demonstrated the presence of the RNA for bFGF as well as bFGF protein in the early embryo (6,11). The presence of bFGF in the early embryo at the correct time and spatial location implicated this growth factor in the normal process of mesoderm differentiation. Several forms of FGF protein have been detected in the early Xenopus embryo (6,11-13). The effects of bFGF during early development (assayable in animal cap induction studies), are likely to be mediated through the interaction of several signaling pathways. The predominant pathway for FGF signaling appears to be the pathway coupled to activation of Raf-l (14). These pathways are coupled to the activation of selective members of the HOX family of genes (15-17) as well as other growth and pattern regulatory proteins such as members of the POU (18), PAX (19) and FKH genes (20). Expression of the Xenopus HOX genes Xhox3, XlHbox6 and Xnot is induced early during bFGF mediated mesoderm induction (15,21,22). Expression of Xhox3 and XlHbox6 is also affected by bFGF treatment of animal caps (15,21,23). However, maximal induction is not observed until control embryos have reached the late gastrula to early neurula stages (embryo stages 13-15, ref. 48). Treatment of animal caps with bFGF has no effect on the

Nucleic Acids Research, 1994, Vol. 22, No. 19 3991 expression of several other Xenopus HOX genes e.g., Mix. 1, gsc and Xliml (24-26). One likely explanation for the observations of differing inductive capacities of bFGF is that different programs of gene expression are initiated. The complex nature of the structures induced by bFGF suggests that no single gene function induced by this growth factor is responsible for all of the observed developmental changes. It is likely that the bFGF-induced signal transduction pathway effects changes in the expression of multiple genes. These genes are likely to be of multiple classes such as the HOX, POU, and FKH families of transcription factors as well as other growth and differentiation modulating genes whose products interact with cell surface receptors such as the TGF-3, Xwnt and BMP families of genes. We have extended the examination of the expression of transcription factor genes whose products are known to regulate early embryonic patterning. We have focused upon genes of the HOX, POU and FKH families. The characterization of the expression of these genes has focused on the early time frame of in vitro mesoderm induction in response to bFGF treatment of Xenopus animal cap explants. The results of these experiments have demonstrated that numerous HOX, POU and FKH genes are expressed in bFGF-induced animal caps very early in the differentiation process. Several novel Xenopus homologs of these

pattern regulating genes have been identified in this study. MATERIALS AND METHODS Mesoderm induction Induction of mesoderm differentiation in Xenopus animal cap ectodermal cells was carried out essentially as described (4,27). All inductions were done with recombinant human bFGF. The protein was purified, by heparin-Sepharose chromatography, from COS-7 cells expressing the l8kDa form of human bFGF (a kind gift from Dr Robert Florkiewicz). Animal caps were dissected from stage 8-8.5 embryos ensuring that no vegetal cells remained adherent. The explants were placed into 75 % NAM/0. 1mM EDTA/0. 1 % BSA (28) in microtiter dishes coated with 10% NAM/l % noble agar (29). Explants were then incubated at 20°C. Induction of mesodermal differentiation was carried out in the above buffer containing either Sng/ml bFGF or 12.5ng/ml activin A. Initial induction experiments were allowed to progress until control non-dissected embryos reached stage 20. At this time it is possible to identify gastrulation-like rearrangements in the induced explants but not controls (30). PCR-based isolation of pattern regulating cDNAs expressed in bFGF-induced animal caps Degenerate oligonucleotide primers were designed to amplify a conserved region of the homeodomain of HOX genes (QT/LL/FEL/FEKE/D and I/VKI/VWFQNR, ref. 12), the POUhomeodomain (KRK/RK/RRTS/TI and WFCNR/SR/SQK/N, ref. 31) and the fork head domain (LIT/MIMAIQ/N and NSIRHN/SLS, ref. 20). These primers were then used to isolate (by sib-selection) cDNAs from a library of cDNAs prepared from induced animal cap RNA or to isolate PCR products generated using the cDNA library as template. This library was prepared from RNA isolated from animal caps following mesoderm induction in response to 5ng/ml bFGF. The animal caps were incubated in the presence of bFGF until control non-dissected

embryos had reached stage 10. This allowed for the isolation of transcripts from genes induced during the early stages of mesoderm induction. Isolation of cDNAs from the induced library was carried out by PCR-based sib-selection (32). The cDNA library was plated at a density of 30,000 pfu on each of ten 150mm agar dishes and allowed to grow until plaques began to contact each other. The phage were eluted from each plate by overlaying the bacterial lawn with lOml of SM buffer. Debris was removed from each individual plate lysate by centrifugation at 6,500 xg for 10 min. A 1/l aliquot of each lysate was used as a template directly in a standard PCR. For all three primer pairs, the reactions were carried out as follows: 95°C for 1 min, 58°C for 1.5 min and 72°C for lmin for 30 cycles. Aliquots of the reaction products were analyzed on agarose gels to assess the success of amplification of the expected sized products. Lysates that positively amplified the correctly sized product were used for subsequent rounds of selection. Each round of PCR utilized 1,1d of the plate lysates. The second round of sib-selection was carried out by plating five to ten 100mm plates at a density of 3,000 to 5,000 pfu/plate. Phage lysates were prepared by overlaying the lawns with 3ml of SM buffer. The third round involved five plates at a density of 100 to 300 pfu/plate. The final round of selection involved plating at 25 to 50 pfu/plate. At this final stage of screening, single plaques were transferred to 100/1d of SM buffer and a 3y1 aliquot was used for the PCR. In addition to the sib-selection of individual cDNAs, several genes expressed in bFGF-induced animal caps were detected and characterized by simply cloning and sequencing of PCR products in pCR-fl (Stratagene, Inc.).

PCR analysis of gene expression Expression of the different pattern regulating genes isolated in this study was assayed by a PCR-based method (33 -35) utilizing total RNA extracted from embryos, microdissected sections of embryos or control animal caps and animal caps treated with Sng/ml bFGF or 12.5ng/ml activin A. Immediately following microdissection or at various times following bFGF and activin A induction, 6 embryo sections or 6 explants (for each assay) were lysed in guanidinium-isothiocyanate buffer (4M guanidinium-isothiocyanate, 1OOmM Tris-HCI pH 7.5, 1% 3mercaptoethanol). and the poly(A+) RNA fraction purified from 1/2 of each lysate (thereby utilizing the equivalent of 3 sections or explants in each RT reaction) by spotting the lysate onto mAP flter squares (Amersham, Inc.). The filters were washed and the RNA eluted into a 30O1t RT reaction containing random hexamers (S0pmols) as primers. For amplification of RNA from developmentally staged whole embryos, 1,ug of total RNA was used in a 30,u1 RT. Following the RT reaction, 1,ud aliquots were removed and added to a PCR containing specific oligos (S0pmols). The first strand cDNA was amplified in the presence of a trace amount of 32P-dCTP (0.5ACi). PCRs were performed with sufficient cycles to allow linear amplification of the cDNA to take place (20-25 cycles). This procedure allows for quantitative evaluation of the results of the RT-PCR. Aliquots (5,1l) of the PCRs were electrophoresed in PAGE gels. The gels were dried and exposed to film. The linearity of PCR amplification of first strand cDNA was verified utilizing synthetic transcripts. A dilution series ranging from 0.2 pg to 2000 pg of each transcript was used in the first strand cDNA synthesis. The 30al RT reaction contained 500sM

3992 Nucleic Acids Research, 1994, Vol. 22, No. 19 each dNTP, 100 pmole of random hexanucleotide primers and Moloney murine leukemia virus reverse transcriptase (33 -35). Control reactions omitted the random hexamers. A 1lad aliquot from the first strand reaction was used for PCR amplification (25f4 reaction) using 50 pmole each of 5'- and 3'- specific primers corresponding to a region of the synthetic RNAs utilized. To allow quantitative evaluation of the products, each PCR reaction included 0.5ytCi of 32P-dCTP. Following amplification, a 1/d aliquot of the PCR product was analyzed by gel electrophoresis and autoradiography. Some PCR assays exceeded the linear amplification range in order to detect low level expression in some tissues or embryo stages.

RESULTS Pattern regulating genes expressed in bFGF-induced animal caps Degenerate oligonucleotide primer pairs were used to isolate, by PCR, HOX, POU and FKH cDNAs (or PCR products) from a Xenopus cDNA library prepared from animal cap explants

induced to differentiate into mesodermal tissue in response to Sng/ml bFGF. A total of 168 different HOX cDNAs or PCR clones, 78 POU cDNAs or PCR clones and 96 FKH cDNAs or PCR clones were characterized in this study. Sequencing within the conserved domains of isolated cDNAs or cloned PCR products was accomplished with the use of the original PCR primers or universal sequencing primers, respectively. The results of this analysis are presented in Table 1 which indicates the particular type of clone isolated as well as the frequency each was encountered in the analysis. There were 11 different HOX cDNAs identified in this study. Several HOX sequences were isolated that likely represent RNAs expressed from both of the duplicated alleles that we would expect to be present in the ancestrally tetraploid genome of Xenopus laevis. These include XlHbox2, XlHbox4, XlHbox6 and XlHbox8 which are previously reported Xenopus HOX genes as well as XHox7, which represents a novel Xenopus HOX gene related to the chicken CHox7 and mouse MMoxA and MMoxB genes (Figure 1, refs. 37,47). The amino acid sequences, within the homeodomain, deduced for the two different forms of

Table 1. Pattern regulating genes in bFGF-induced animal caps HOX Name

XlHboxl XlHbox2(A) XlHbox2(B, MM3) XlHbox4(A) XlHbox4(B) XlHbox5 XlHbox6(A) XlHbox6(B)

XlHbox8(A+B)a Xhox IA Xcadl Xnot

XHox7(A+B)b XHox-b2

% of Clones 2 16 7 9 5 16 11 9 5 3 2 3 11 2

POU Name

% of Clones

XlOct-l XlOct-25 XlOct-60 Xloctl-13 XlOct-91 XlOct-92 XBrn-3

31 4 15 27 4 15 4

FKH Name

XFD2'

XFD4(A) XFD4(B) XFLIP

XHNF-3ac

% of Clones 3 47 22 6 22

The percent of clones isolated for XlHbox8(A) and (B) as well as XHox7(A) and (B) are included together since the deduced coding regions within the conserved portion of the homeodomain are unchanged by the nucleotide differences. cThe Xenopus HNF-3a cDNA cloned in this study is identical to the cDNA identified as XFKH2 (36) except that an additional 54 bp are present at the 5' end of our cDNA (not shown). a,b

HOX-Homeodomain XHox7 CHox7

NRRRRTAFTSEQLLELEKEFHCKKYLSLTERSQIAHVLKLSEVQVKIWFQNRRAKWKRVK S-----------------------------------A---------------------I-

POU-Homeodomain XBrn-3 RBrn-3

KRKKRTSIAAPEKRSLEAYFAIQPRPSSEKIAAIAEKLDLKKNVVRVWFCNRR Q -----------V--------------

FKH Domain XFLIP DmFD3

LITMAILQSPHKKLTLSGICDFIMSKFPYYKDKFPAWQNSIRHNL ----------------------------------R-

Figure 1. Amino acid sequences of conserved domains of novel Xenopus HOX, POU and FKH cDNAs. The amino acid sequences of the conserved regions of the three clones characterized in this study are shown in comparison to their closest mammalian or Drosophila relatives. XHox 7 is a relative of the chicken HOX gene, CHox7 (37). XBrn3 is a relative of the mammalian POU gene, brn-3 (31). XFLIP is related to the Drosophila FKH sequence, FD3 (42)

_~XFLIP

Nucleic Acids Research, 1994, Vol. 22, No. 19 3993 W

0

tOO

LO

,.-

_-

u)

Un)

_

a)

it

CN

en

U)

Un

Co)

'0 G)

N

~- LL )

U

)

c LL

XHox7

I-

Z 0

XBrn-3

0L

Zn

c ._

v

:

XHox7

XFLIP .-- XMax 2 --

I

D Vn An Vg

st.16

st.12

st.10

..

A

P

A

P

Figure 4. Expression of XHox7, XBrn-3 and XFLIP in animal cap explants in response to bFGF. Animal caps were dissected from stage 8 embryos and incubated in the presence of buffer (non-induced), Sng/ml bFGF or 12.5ng/ml activin A. Expresssion of XHox7, XBrn-3 and XFLIP was then assyed by RT-PCR. NT (no template) is a control RT reaction from uninduced explant RNA to which no random hexamers were added. Non-induced, bFGF-induced explants and activin-induced explants are indicated above their respective lanes. ODC represents omithine decarboxylase.

family, XlOct-1 and Xloctl-13 (Table 1). One of the Xenopus POU cDNAs characterized is highly related to XlOct9l (39,40) and has been called, XlOct92 (see Table 1). XlOct92 exhibits

II

I .. I

Vg

_ *wl__~-*- 0 DC

XMaxl

Figure 2. Developmental expression of XHox7, XBrn-3 and XFLIP. The relative level of expression of XHox7, XBm-3 and XFLIP in early embryos was determined by RT-PCR. The PCR products were trace-labeled by the inclusion of 32P-dCTP in the reaction. The products were separated by acrylamide gel electrophoresis, the gel dried and exposed to X-ray film. The embryonic stages analyzed are indicated above their respective lanes. UFE = unfertilized egg. NT (no template) represents an RT reaction from embryo stage 10 RNA carried out without the addition of random hexamer primers. Expression of XMaxl and XMax2 (34) was assayed as internal controls for RNA loading in the RT reaction.

I

-XBrn-3

Vg

-

XHox7

90% amino acid sequence identity to XlOct9l over a region

XBrn-3

spanning 236 amino acid (including sequences outside the POUdomain) which suggests that XlOct92 represents expression from the duplicated XlOct9l allele. A novel Xenopus POU cDNA identified in this study is the homolog of the murine class IV

XFLIP

gene,

-Xc-Myb

Brn-3 (Figure 1; ref. 3 1) and has been given the

the POU-homeodomain (Figure 1, ref. 31).

Four different FKH cDNAs

-

XM yoDb

Figure 3. Expression of XHox7, XBm-3 and XFLIP in regic sectlons tn-sPeclaic relative level of of early embryos. Early embryos were micro-dissected and tthe expression of XHox7, XBrn-3 and XFLIP was assayed by RT--PCR. The sections dissected at each stage represent surface regions of the embryoss. Sections analyzed at stage 10 include the dorsal lip (D), ventral lip (Vn), animal c ap (An) and vegetal cap (Vg). Stage 12 and 16 sections were anterior dorsal (A ), posterior dorsal (P) and ventral (Vg). Expression of zygotic XMyoDb (43) and IXc-Myb (35) included as region specific controls. were

XlHbox8 and XHox7 are identical, whereas, the amino acid sequences for the two forms of XlHbox2 differ by one amino acid and XlHbox4 and XlHbox6 differ by 2 anriino acids (not shown). The sequence identified in this study as XlHbox2(B) is identical to the previously published Xenopus HOXc protein, MM3 (38). An additional novel Xenopus HOX cDNA Mvas isolated that is the Xenopus equivalent of the mammalian Hox -b2 locus (12). The XHox-b2 cDNA has not been characterizedl further in this study. A total of 7 different POU cDNAs were isolate d in this study. The most frequently isolated POU cDNAs belonjged to the class

name

XBrn-3. Within the conserved POU-homeodomain, XBrn-3 differs from the rat Brn-3 sequence at only two positions within were

identified in this study.

XFD4, a class II FKH gene (41), was the most frequently isolated

(Table 1). Two closely related XFD4 cDNAs, XFD4(A) and XFD4(B) which differ at two amino acid positions within the fork head domain were isolated in this study (not shown). The Xenopus homolog of the mammalian class I FKH gene, HNF-3a also isolated in this study and is identified XHNF-3a (20) (Table 1). The Xenopus HNF-3a homolog has also been identified XFKH2 (36). One novel Xenopus FKH cDNA was isolated and identified as XFLIP (Xenopus forkhead expressed in the dorsal lip). XFLIP is most related to the forkhead domain of the Drosophila FD3 sequence (Figure 1, ref. 42). as

Expression of XHox7, XBrn-3 and XFLIP during normal development Expression of XHox7, XBrn-3 and XFLIP was examined during early development utilizing an RT-PCR based assay (see Methods). The RT-PCR assay is a rapid and efficient technique for analyzing the expression of genes provided care is taken to ensure amplification is the result of the presence of the specific transcript of interest (33 -35,43). We utilized primer pairs that were specific to nucleotide sequences for each of the three genes analyzed in this study. The equivalence of RNA loading in the RT reactions was assessed by PCR amplification of XMax RNAs previously shown to be equivalently expressed at all stages of early development (34).

3994 Nucleic Acids Research, 1994, Vol. 22, No. 19 The expression of XHox7 and XBrn-3 are non-maternal, whereas, a very low level of maternal XFLIP RNA is detectable by this assay (Figure 2). In some reactions with the XBrn-3 primers, non-specific products that are slightly larger and smaller in size than the specific PCR product can be observed (see UFE lane for XBrn-3 in Figure 2). These fragments were shown to be unrelated to XBrn-3 since an internal primer was incapable of hybridizing to the incorrectly sized PCR products in Southern analyses (not shown). Expression of XHox7 is detectable at midblastula, with maximal expression seen in embryo stages 11-19 (Figure 2). The level of XBrn-3 expression is quite low and biphasic in early embryos (Figure 2). XBrn-3 expression is first detected at midblastula, is undetectable to barely detectable in stage 11 embryos and is re-expressed by the early neural fold stage (stage 15). Maximal expression of XBrn-3 is observed by embryo stage 42 (Figure 2). Although a low level of detectable XFLIP expression is seen in the unfertilized egg, this maternal RNA is degraded prior to midblastula (Figure 2). Zygotic XFLIP expression begins at midblastula and reaches peak levels by the initial gastrulation stage (stage 10) through the initial neural tube stage (stage 19) before declining again by embryo stage 34 (Figure 2).

Spatial distribution of XHox7, XBrn-3 and XFLIP RNAs in

early embryos An initial estimation of the early spatial distribution of XHox7, XBrn-3 and XFLIP RNAs was carried out using the RT-PCR method on RNA extracted from microdissected sections of stage

10 (initial gastrula), 12 (medium yolk plug) and 16 (mid-neural fold) embryos. Each section taken represented primarily external embryonic regions of identical size. The sections from the stage 10 embryos were the animal cap, vegetal cap, dorsal blastoporal lip and ventral blastoporal lip. Dissection of stage 12 and 16 embryos isolated anterior and posterior dorsal sections and a ventral vegetal section. The dorsal sections of stage 12 and 16 embryos contained a portion of the somitic mesoderm and lateral plate mesoderm. For this reason we chose to use expression of XMyoDb (43) and Xc-Myb (35) as internal controls in RNA from these sections since both are expressed in the presomitic and lateral plate mesoderm. Additionally, XMyoDb is not expressed in the ventral vegetal region (43). The level of RNA assayed in the vegetal sections at each stage is less than in the animal hemisphere sections due to the larger size of the vegetal cells and the presence of large yolk platelets. XHox7 RNA is found predominantly in the dorsal lip and animal cap region of the stage 10 embryo relative to the ventral lip and vegetal cap regions (Figure 3). By stage 12 expression is near equally distributed dorsal and ventrally, within the sections analyzed. In stage 16 embryos expression of XHox7 predominates in the anterior dorsal section relative to the posterior dorsal and ventral sections analyzed (Figure 3). XBrn-3 RNA is not detectable in the dorsal lip in stage 10 embryos but is detected in the ventral lip and animal and vegetal cap sections (Figure 3). Like XHox7, XBrn-3 RNA is detectable at higher levels in the anterior dorsal sections of stage 16 embryos relative to the posterior dorsal and ventral sections (Figure 3). Expression of XFLIP is essentially localized to the dorsal lip section (Figure 3). The low level of detectable XFLIP RNA seen in the ventral lip and animal cap sections likely reflects the fact that these sections contained cells representing the extreme outer limits of XFLIP RNA distribution in embryos at stage 10. Unlike

XHox7 and XBrn-3, XFLIP expression is restricted to the dorsal sections of the embryo that were analyzed (Figure 3). Expression of XFLIP is observed both anteriorly and posteriorly in the stage 12 and 16 sections with predominant expression seen in the stage 16 posterior dorsal section (Figure 3).

Expression of XHox7, XBrn-3 and XFLIP during in vitro mesoderm induction Since each of the three sequences characterized in this report were originally identified in a cDNA library prepared from bFGFinduced animal cap RNA we assayed for these RNAs in noninduced and induced animal caps. Animal caps were dissected from stage 8-8.5 embryos and treated with either buffer alone, Sng/ml bFGF or 12.5ng/ml activin A and incubated until control non-dissected embryos reached stage 12 (identified as stage 12 caps). The concentration of activin A used in this study was chosen since we could consistently detect induction of XMyoDb in induced stage 12 animal caps (not shown). Caps were pooled (6 each) and expression of XHox7, XBrn-3 and XFLIP was assayed using the RT-PCR protocol (see Methods). Expression of ornithine decarboxylase (ODC) was used as a control since its expression does not change in induced animal caps (Dr Anton Neff, personal communication). Expression of XHox7 is inducible with bFGF and activin A (Figure 4). Expression of XBrn-3 and XFLIP is only marginally inducible in treated animal caps (Figure 4). XBrn-3 RNA is not detectable in the uninduced animal caps and only an extremely low level of XFLIP RNA is detectable (Figure 4). The level of XBrn-3 RNA is higher in bFGF versus activin A induced animal caps (Figure 4). Unlike XHox7 and XBrn-3 RNAs, which are still detectable at elevated levels in induced explants, relative to uninduced, 24 hrs. post-induction (embryo stage 20), XFLIP RNA is undetectable by stage 20 whether or not the explants were induced (not shown).

DISCUSSION The expression of numerous pattern regulating genes has been shown to be affected by growth factor addition to cells explanted from the animal cap of early Xenopus embryos. The classes of regulated genes include the HOX genes (15,21,23), the POU genes (40) and the FKH genes (41,44,45). The majority of genes, assayed for growth factor inducibility in animal caps, have been characterized for responses to mesoderm induction by the TGF,B family of proteins. We are interested in the molecular control of mesodermal differentiation exerted by bFGF, in particular, control exerted upon classes of genes known to affect patterning in the early embryo. Therefore, we sought to identify and characterize the expression of HOX, POU and FKH genes in animal cap explants in response to mesoderm induction by bFGF. In particular we are interested in the very early molecular changes that occur in the inductive process. To this end we isolated numerous cDNAs of these gene families by PCR screening of a library produced from bFGF-induced animal caps. We were able to identify 11 different HOX, 7 different POU and 4 different FKH cDNAs as being expressed in stage 10 animal caps (i.e. animal caps that were cultured until undissected embryos reached stage 10) treated with Sng/ml bFGF. Not all members of any class could be expected to be isolatable by this technique given the limits of oligonucleotide degeneracy in the

Nucleic Acids Research, 1994, Vol. 22, No. 19 3995 PCR. For example, Xhox3 has been shown to be inducible by both bFGF and TGF-,B family molecules (21), yet we did not find this gene in our screening of 168 different HOX PCR clones (Table 1). Several of the RNAs isolated in this study may be present in uninduced and induced animal caps at equivalent levels but we have not undertaken a characterization of inducibility of all the cDNAs identified. Both XHox7, and XFLIP RNAs are present in uninduced animal caps (Figure 4). The animal cap induction assay can also lead to the isolation of abnormally expressed RNAs. For instance, expression of the HOX gene, Mix.1 is inducible by TGF-3 in animal caps but is normally expressed only in the endodermal cells of the early embryo (24). We have identified another HOX gene that is perhaps abnormally expressed in early animal caps induced with bFGF, the XlHbox8 gene (Table 1). Normal expression of XlHbox8 is not detected until around embryo stage 33 and is restricted to a narrow band of endodermal tissue from which the pancreatic primordium and duodenum arise (46). Whether this gene is responsive to bFGF signals in the intact embryo is unknown but would be an important question to answer. The XHox7 gene is related in sequence to the chicken CHox7 gene (37), however, only within the homeodomain. Conservation, within the homeodomain, of XHox7 is also observed with the PCR products MMoxA and MMoxB isolated from mouse (47). CHox7 expression is biphasic with peak expression observed during somitogenesis and organogenesis (37). XHox7 is expressed only from the zygotic genome beginning at midblastula (stage 9; Figure 2) reaching a peak level of expression during the neurulation stages of development and remaining high until at least the initial neural tube stage (Figure 2). Expression of XHox7 was not assayed between stages 19 and 34 so it is not known whether the decrease observed between these two stages occurs gradually. Expression of MMoxA and MMoxB is highest in the telencephalon of the 13.5 day mouse embryo (47). In microdissected sections of early Xenopus embryos the highest level of XHox7 RNA was detectable in the anterior dorsal section of the mid-neural fold stage embryo (stage 16, Figure 3). This region of the mid-neural fold stage embryos contributes to the development of the telencephalon and could, therefore, account for the elevated level of XHox7 RNA relative to the other regions of the embryo assayed in this report. XBrn-3 is a non-maternal RNA first detectable at midblastula (Figure 2). This is consistent with the non-maternal nature of rat Brn-3 (31). Rat Brn-3 exhibits a restricted spatial pattern of expression being detectable only in the brain and spleen (31). Rat Brn-3 was the only brain derived POU RNA detectable in sensory ganglion cells which are derived from the neural crest (31). This is a pattern of expression highly conserved with that of unc86 expression in C. elegans (47). In Xenopus embryos at the mid-neural fold stage (stage 16) the neural crest is localized to the anterior and middle regions of the neural primordium (48). The pattern of XBrn-3 RNA distribution in sections from stage 16 embryos is consistent with the potential for expression of XBrn-3 within the neural crest cell population (Figure 3). Clearly, at this stage other cell types are also expressing XBrn-3. The highest levels of XBrn-3 RNA detected in this study are in stage 42 embryos (Figure 2). Near stage 42-43 a particular group of mesechymal cells begins accumulating in the dorsal mesentary near the anterior end of the stomach (48). It is from

these cells that the adult spleen derives. Given that rat Brn-3 is expressed only in the adult brain and spleen (31) the possibility exists that a portion of the increase in XBrn-3 expression seen in stage 42 embryos is related to events signaling differentiation of the primordial spleen. Expression of XHox7 is seen in the organizer region as well as in the animal cap of initial gastrula stage embryos (stage 10, Figure 3). XFLIP expression is almost exclusively localized to the organizer region at the initial gastrula stage. This pattern of expression of these two genes suggests they are expressed in the intact embryo in a temporal and spatial pattern that suggests they have important roles in the early patterning of the Xenopus embryo. Although XHox7 responds to growth factor addition to animal caps to a greater extent than either XBrn-3 or XFLIP, all three sequences are expressed in animal caps in response to mesoderm induction mediated by bFGF. The results presented in this report indicate that these genes are likely candidates for early pattern regulating genes in the Xenopus embryo and that they are responsive to signals elicited by bFGF and/or TGF-,3 family proteins in vivo.

ACKNOWLEDGEMENTS The authors wish to hank Drs Anton Neff and Hea-Moon Chung for assistance with embryo microdissection, Drs Anton Neff and Makoto Asashima for the kind gift of activin A, and Dr Gary Stuart for critical review of this manuscript. MJM was a visiting undergraduate research fellow from UCSD.

REFERENCES 1. Spemann, H. (1938). Embryonic development and induction. New Haven Connecticut: Yale University Press. 2. Jacobson, A.G. (1966). Science 152, 25-34. 3. Nieuwkoop, P.D. (1969). Roux's Arch. Dev. Biol. 162, 341-373. 4. Slack, J.M.W., Darlington, B.G., Heath, J.K. and Godsave, S.F. (1987). Nature 326, 197-200. 5. Slack, J.M.W., Isaacs, H.V. and Darlington, B.G. (1988). Development 103, 581-590. 6. Slack, J.M.W., Darlington, B.G., Gillespie, L.L., Godsave, S.F., Isaacs, H.V. and Paterno, G.D. (1989). Development 107, 141-148 7. Paterno, G.D., Gillespie, L.L., Dixon, M.S., Slack, J.M.W. and Heath, J.K. (1989). Development 106, 79-83. 8. Zimmerman, K.A., Yancopoulus, G.D., Collum, R.G., Smith, R.K., Paterno, G.D., Gillespie, L.L., Dixon, M.S., Slack, J.M.W. and Heath, J.K. (1989). Development 106, 79-83. 9. Jessel, T.M. and Melton, D.A. (1992). Cell 68, 257-270. 10. Kimelman, D. and Maas, A. (1992). Development 114, 261-269. 11. Kimelman, D., Abraham, J.A., Haaparanta, T., Palisi, T.M. and Kirschner, M.W. (1988). Science 242, 1053-1056. 12. Kappen, C., Schughart, K. and Ruddle, F.H. (1993). Genomics 18, 54-70. 13. Tannahill, D., Isaacs, H.V., Close, M.J., Peters, G. and Slack, J.M.W. (1992). Development 115, 695-702. 14. MacNicol, A.M., Muslin, A.J. and Wiliams, L.T. (1993). Cel 73,571-583. 15. Cho, K.W.Y. and De Robertis, E.M. (1990). Genes and Development 4, 1910-1916. 16. Christian, J.L., Olson, D.J. and Moon, R.T. (1992). EMBO J. 11, 33-41. 17. Whitman, M. and Melton, D.A. (1992). Nature 357, 252-254. 18. Rosenfeld, M.G. (1991). Genes and Dev. 5, 897-907. 19. Chalepakis, G. Fritsch, R., Fickenscher, H., Deutsch, U., Goulding, M. and Gruss, P. (1991) Cell 66, 873-884. 20. Sasaki, H. and Hogan, B.L.M. (1993). Development 118, 47-59. 21. Ruiz i Altaba, A. and Melton, D.A. (1989). Nature 341, 33-38. 22. von Dassow, G., Schmidt, J.E. and Kimelman, D. (1993). Genes and Dev. 7, 355-366. 23. Green, J.B.A., New, H.V. and Smith, J.C. (1992). Cell 71, 731-739. 24. Rosa, F.M. (1989). Cell 57, 965-974.

3996 Nucleic Acids Research, 1994, Vol. 22, No. 19 25. Cho, K.W.Y., Blumberg, B., Steinbeisser, H. and De Robertis, E.M. (1991). Cell 67, 1111-1120. 26. Tadano, T., Otani, H., Taira, M. and Dawid, I.B. (1993). Developmental Genetics 14, 204-211. 27. Rosa, F., Roberts, A.B., Danielpour, D., Dart, L.L., Sporn, M.B. and Dawid, I.B. (1988). Science 239, 783-785. 28. Slack, J.M.W. (1984). J. Embryol. exp. Morph. 80, 289-319. 29. Green, J.B.A., Howes, G., Symes, K., Cooke, J. and Smith, J.C. (1990). Development 108, 173-183. 30. Symes, K. and Smith, J.C. (1987). Development 101, 339-349. 31. He, X., Treacy, M.N., Simmons, D.M., Ingraham, H.A., Swanson, L.W. and Rosenfeld, M.G. (1989). Nature 340, 35-42. 32. Amaravadi, L. and King, M.W. (1994). BioTechniques 16, 4-7. 33. Banker, D.E., Bigler, J. and Eisenman, R.N. (1991). Mol. Cell. Biol. 11, 5079-5089. 34. King, M.W., Blackwood, E.M. and Eisenman, R.N. (1993). Cell Growth and Differentiation 4, 85-92. 35. Amaravadi, L. and King, M.W. (1994). Oncogene 9, 971-974. 36. Bolce, M.E., Hemmati-Brivanlou, A. and Harland, R.M. (1993) Dev. Biol. 160, 413-423. 37. Fainsod, A. and Gruenbaum, Y. (1989). FEBS Lett. 250, 381-385. 38. Miller, M.M., Carrasco, A.E. and DeRobertis, E.M. (1984). Cell 39, 157- 162. 39. Hinkley, C.S., Martin, J.F., Leibham, D. and Perry, M. (1992). Mol. Cell Biol. 12, 638-649. 40. Frank, D. and Harland, R.M. (1992). Development 115, 439-448. 41. Knochel, S., Lef, J., Clement, J., Klocke, B., Hille, S., Koster, M. and Knochel, W. (1998). Mech. Dev. 38, 157-165. 42. Hacker, U., Grossinklaus, U., Gehring, W.J. and Jickle, H. (1992). Proc. Natl. Acad. Sci. U.S.A. 89, 8754-8758. 43. Rupp, R.A. and Weintraub, H. (1991). Cell 65: 927-937. 44. Dirksen, M.L. and Jamrich, M. (1992). Genes and Dev. 6, 599-608. 45. Ruiz i Altaba, A. and Jessell, T.M. (1992). Development 116, 81-93. 46. Wright, C.V.E., Schnegelsberg, P. and DeRobertis, E.M. (1988). Development 104, 787-794. 47. Finney, M., Ruvkun, G. and Horvitz, H.R. (1988). Cell 55, 757-769. 48. Nieuwkoop, P.D. and Faber, J. (1967). Normal Table of Xenopus laevis (Daudin) North Holland Publishing Co. Amsterdam.