The Birth of Animal Development: Multicellularity and the Germline

The evolution of multicellular animals has been attributed to many kinds of selective advantage; here I suggest that the evolution of somatic cells to feed and protect the germline was central to the appearance of animals. This would have been driven by selection for extreme anisogamy--the evolution of sperm and egg. Evidence is adduced from the germline stem cells of simple animals (defining germline as any cell that normally produces the next generation via the sexual process) and from the control circuitry ubiquitous in animal germlines. With the soma and its elaboration came animal development, as we understand it.

Protein interactions in Xenopus germ plasm RNP particles

Hermes is an RNA-binding protein that we have previously reported to be found in the ribonucleoprotein (RNP) particles of Xenopus germ plasm, where it is associated with various RNAs, including that encoding the germ line determinant Nanos1. To further define the composition of these RNPs, we performed a screen for Hermes-binding partners using the yeast two-hybrid system. We have identified and validated four proteins that interact with Hermes in germ plasm: two isoforms of Xvelo1 (a homologue of zebrafish Bucky ball) and Rbm24b and Rbm42b, both RNA-binding proteins containing the RRM motif. GFP-Xvelo fusion proteins and their endogenous counterparts, identified with antisera, were found to localize with Hermes in the germ plasm particles of large oocytes and eggs. Only the larger Xvelo isoform was naturally found in the Balbiani body of previtellogenic oocytes. Bimolecular fluorescence complementation (BiFC) experiments confirmed that Hermes and the Xvelo variants interact in germ plasm, as do Rbm24b and 42b. Depletion of the shorter Xvelo variant with antisense oligonucleotides caused a decrease in the size of germ plasm aggregates and loosening of associated mitochondria from these structures. This suggests that the short Xvelo variant, or less likely its RNA, has a role in organizing and maintaining the integrity of germ plasm in Xenopus oocytes. While GFP fusion proteins for Rbm24b and 42b did not localize into germ plasm as specifically as Hermes or Xvelo, BiFC analysis indicated that both interact with Hermes in germ plasm RNPs. They are very stable in the face of RNA depletion, but additive effects of combinations of antisense oligos suggest they may have a role in germ plasm structure and may influence the ability of Hermes protein to effectively enter RNP particles.

Localisation of RNAs into the germ plasm of vitellogenic Xenopus oocytes

We have studied the localisation of mRNAs in full-grown Xenopus laevis oocytes by injecting fluorescent RNAs, followed by confocal microscopy of the oocyte cortex. Concentrating on RNA encoding the Xenopus Nanos homologue, nanos1 (formerly Xcat2), we find that it consistently localised into aggregated germ plasm ribonucleoprotein (RNP) particles, independently of cytoskeletal integrity. This implies that a diffusion/entrapment-mediated mechanism is active, as previously reported for previtellogenic oocytes. Sometimes this was accompanied by localisation into scattered particles of the “late”, Vg1/VegT pathway; occasionally only late pathway localisation was seen. The Xpat RNA behaved in an identical fashion and for neither RNA was the localisation changed by any culture conditions tested. The identity of the labelled RNP aggregates as definitive germ plasm was confirmed by their inclusion of abundant mitochondria and co-localisation with the germ plasm protein Hermes. Further, the nanos1/Hermes RNP particles are interspersed with those containing the germ plasm protein Xpat. These aggregates may be followed into the germ plasm of unfertilized eggs, but with a notable reduction in its quantity, both in terms of injected molecules and endogenous structures. Our results conflict with previous reports that there is no RNA localisation in large oocytes, and that during mid-oogenesis even germ plasm RNAs localise exclusively by the late pathway. We find that in mid oogenesis nanos1 RNA also localises to germ plasm but also by the late pathway. Late pathway RNAs, Vg1 and VegT, also may localise into germ plasm. Our results support the view that mechanistically the two modes of localisation are extremely similar, and that in an injection experiment RNAs might utilise either pathway, the distinction in fates being very subtle and subject to variation. We discuss these results in relation to their biological significance and the results of others.

Poc1A and Poc1B act together in human cells to ensure centriole integrity

Proteomic studies in unicellular eukaryotes identified a set of centriolar proteins that included proteome of centriole 1 (Poc1). Functional studies in these organisms implicated Poc1 in centriole duplication and length control, as well as ciliogenesis. Using isoform-specific antibodies and RNAi depletion, we have examined the function of the two related human proteins, Poc1A and Poc1B. We find that Poc1A and Poc1B each localize to centrioles and spindle poles, but do so independently and with different dynamics. However, although loss of one or other Poc1 protein does not obviously disrupt mitosis, depletion of both proteins leads to defects in spindle organization with the generation of unequal or monopolar spindles. Our data indicate that, once incorporated, a fraction of Poc1A and Poc1B remains stably associated with parental centrioles, but that depletion prevents incorporation into nascent centrioles. Nascent centrioles lacking both Poc1A and Poc1B exhibit loss of integrity and maturation, and fail to undergo duplication. Thus, when Poc1A and Poc1B are co-depleted, new centrosomes capable of maturation cannot assemble and unequal spindles result. Interestingly, Poc1B, but not Poc1A, is phosphorylated in mitosis, and depletion of Poc1B alone was sufficient to perturb cell proliferation. Hence, Poc1A and Poc1B play redundant, but essential, roles in generation of stable centrioles, but Poc1B may have additional independent functions during cell cycle progression.

Pix proteins and the evolution of centrioles

We have made a wide phylogenetic survey of Pix proteins, which are constituents of vertebrate centrioles in most eukaryotes. We have also surveyed the presence and structure of flagella or cilia and centrioles in these organisms, as far as is possible from published information. We find that Pix proteins are present in a vast range of eukaryotes, but not all. Where centrioles are absent so are Pix proteins. If one considers the maintenance of Pix proteins over evolutionary time scales, our analysis would suggest that their key function is to make cilia and flagella, and the same is true of centrioles. Moreover, this survey raises the possibility that Pix proteins are only maintained to make cilia and flagella that undulate, and even then only when they are constructed by transporting ciliary constituents up the cilium using the intraflagellar transport (IFT) system. We also find that Pix proteins have become generally divergent within Ecdysozoa and between this group and other taxa. This correlates with a simplification of centrioles within Ecdysozoa and a loss or divergence of cilia/flagella. Thus Pix proteins act as a weathervane to indicate changes in centriole function, whose core activity is to make cilia and flagella.

Pix1 and Pix2 are novel WD40 microtubule-associated proteins that colocalize with mitochondria in Xenopus germ plasm and centrosomes in human cells

In many animals, the germ line develops from a distinct mitochondria-rich region of embryonic cytoplasm called the germ plasm. However, the protein composition of germ plasm and its formation remain poorly understood, except in Drosophila. Here, we show that Xpat, a recently identified protein component of Xenopus germ plasm, interacts via its C-terminal domain with a novel protein, xPix1. Xpat and xPix1 are co-expressed in ovaries, eggs and early embryos and colocalize to the mitochondrial cloud and germ plasm in stage I and stage VI oocytes, respectively. Although Xpat appears unique to Xenopus, Pix proteins, which contain an N-terminal WD40 domain and C-terminal coiled-coil, are widely conserved. In humans, two proteins, Pix1 and Pix2, are expressed at varying levels in different cancer cell lines. Importantly, as well as localizing to mitochondria, human Pix proteins localize to centrosomes and associate with microtubules in vitro and in vivo. Although, Pix proteins are stably expressed through the cell cycle, Pix2 concentrates on microtubule structures in mitosis and microinjection of Pix antibodies interferes with cell division. Based on these data, we propose that Pix1 and Pix2 are microtubule-associated adaptor proteins that likely contribute to a range of developmental and cell division processes.

The efficiency of Xenopus primordial germ cell migration depends on the germplasm mRNA encoding the PDZ domain protein Grip2

A microarray analysis of vegetal pole sequences in the egg and early Xenopus laevis embryo identified Unigene Xl.14891 as a vegetally localized RNA. Analysis of the Xenopus tropicalis genome showed this Unigene to be localized near the 3' end of the Grip2 (glutamate receptor interacting protein 2) transcription unit. RACE showed that the Unigene represented the 3' UTR of Grip2 mRNA. Grip2 mRNA is present in the mitochondrial cloud of late pre-vitellogenic oocytes and then in the germplasm through oogenesis and early development until tailbud tadpole stages. Interference with Grip2 mRNA translation using two antisense morpholino oligos (MOs) impairs primordial germ cell (PGC) migration to the germinal ridges. Both MOs also inhibit swimming movements of the tailbud tadpole, known to involve glutamate receptors. We conclude that Grip2 has several functions in the embryo, including enabling efficient PGC migration.

Regulation of the Xenopus Xsox17alpha(1) promoter by co-operating VegT and Sox17 sites

The gene encoding the Sox F-group transcription factor Xsox17α₁ is specifically expressed throughout the entire region of the Xenopus blastula fated to become endoderm, and is important in controlling endodermal development. Xsox17α₁ is a direct target of the maternal endodermal determinant VegT and of Sox17 itself. We have analysed the promoter of the Xenopus laevis Xsox17α₁ gene by transgenesis, and have identified two important control elements which reside about 9 kb upstream at the start of transcription. These elements individually drive transgenic endodermal expression in the blastula and gastrula. One contains functional, cooperating VegT and Sox-binding consensus sites. The Sox sites in this region are occupied in vivo. The other responds to TGF-β signals like Activin or Nodals that act through Smad2/3. We propose that these two regions co-operate in regulating the early endodermal expression of the Xsox17α₁ gene.

The protein encoded by the germ plasm RNA Germes associates with dynein light chains and functions in Xenopus germline development

Germ plasm plays a prominent role in germline formation in a large number of animal taxons. We previously identified a novel maternal RNA named Germes associated with Xenopus germ plasm. In the present work, we addressed possible involvement of Germes protein in germ plasm function. Expression in oocytes followed by confocal microscopy revealed that the EGFP fused to Germes, in contrast to the free EGFP, co-localized with the germ plasm. Overexpression of intact Germes and Germes lacking both leucine zipper motifs (GermesDeltaLZs) resulted in a statistically significant reduction of the number of primordial germ cells (PGCs). Furthermore, the GermesDeltaLZs mutant inhibited PGC migration and produced abnormalities in germ plasm intra-cellular distribution at tailbud stages. To begin unraveling biochemical interactions of Germes during embryogenesis, we searched for Germes partners using yeast two-hybrid (YTH) system. Two closely related sequences were identified, encoding Xenopus dynein light chains dlc8a and dlc8b. Tagged versions of Germes and dlc8s co-localize in VERO cells upon transient expression and can be co-immunoprecipitated after injection of the corresponding RNAs in Xenopus embryos, indicating that their interactions occur in vivo. We conclude that Germes is involved in organization and functioning of germ plasm in Xenopus, probably through interaction with motor complexes.

Global analysis of the transcriptional network controlling Xenopus endoderm formation

A conserved molecular pathway has emerged controlling endoderm formation in Xenopus zebrafish and mice. Key genes in this pathway include Nodal ligands and transcription factors of the Mix-like paired homeodomain class, Gata4-6 zinc-finger factors and Sox17 HMG domain proteins. Although a linear epistatic pathway has been proposed, the precise hierarchical relationships between these factors and their downstream targets are largely unresolved. Here, we have used a combination of microarray analysis and loss-of-function experiments to examine the global regulatory network controlling Xenopus endoderm formation. We identified over 300 transcripts enriched in the gastrula endoderm, including most of the known endoderm regulators and over a hundred uncharacterized genes. Surprisingly only 10% of the endoderm transcriptome is regulated as predicted by the current linear model. We find that Nodal genes, Mixer and Sox17 have both shared and distinct sets of downstream targets, and that a number of unexpected autoregulatory loops exist between Sox17 and Gata4-6, between Sox17 and Bix1/Bix2/Bix4, and between Sox17 and Xnr4. Furthermore, we find that Mixer does not function primarily via Sox17 as previously proposed. These data provides new insight into the complexity of endoderm formation and will serve as valuable resource for establishing a complete endoderm gene regulatory network.

Xenopus Xpat protein is a major component of germ plasm and may function in its organisation and positioning

In many animals, including Drosophila, C. elegans, zebrafish and Xenopus, the germ line is specified by maternal determinants localised in a distinct cytoplasmic structure called the germ plasm. This is consists of dense granules, mitochondria, and specific localised RNAs. We have characterised the expression and properties of the protein encoded by Xpat, an RNA localised to the germ plasm of Xenopus. Immunofluorescence and immunoblotting showed that this novel protein is itself a major constituent of germ plasm throughout oogenesis and early development, although it is also present in other regions of oocytes and embryos, including their nuclei. We found that an Xpat-GFP fusion protein can localise correctly in cultured oocytes, in early oocytes to the 'mitochondrial cloud', from which germ plasm originates, and in later oocytes to the vegetal cortex. The localisation process was microtubule-dependent, while cortical anchoring required microfilaments. Xpat-GFP expressed in late stage oocytes assembled into circular fields of multi-particulate structures resembling endogenous fields of germ plasm islands. Furthermore these structures could be induced to form at ectopic sites by manipulation of culture conditions. Ectopic Xpat-GFP islands were able to recruit mitochondria, a major germ plasm component. These data suggest that Xpat protein has an important role in Xenopus germ plasm formation, positioning and maintenance.

Early endodermal expression of the Xenopus Endodermin gene is driven by regulatory sequences containing essential Sox protein-binding elements

The Endodermin gene is expressed in the early endoderm and the Spemann organizer of Xenopus embryos. It has previously been shown to be a direct target of the early endodermal transcription factor Xsox17 (Clements et al., 2003, Mech Dev 120:337-348). Here we identify two adjacent control elements in the Endodermin promoter; these drive transcription of the gene in late-gastrula endoderm and contain consensus Sox-binding sites. We have analyzed one element in detail and show that it responds directly to Xsox17 and that the Sox sites are essential for endodermal expression in transgenic embryos. However, flanking regions on both sides are also essential, indicating that Xsox17 acts in concert with several DNA-binding partners.

VegT induces endoderm by a self-limiting mechanism and by changing the competence of cells to respond to TGF-beta signals

The maternal determinant VegT is required for both endoderm and mesoderm formation by the Xenopus embryo. An important downstream mediator of VegT action is Xsox17, which has been proposed to be induced in cell-autonomous, then signal-dependent phases. We show that Xsox17 is a direct VegT target, but that direct induction of Xsox17 by VegT is rapidly inhibited. This inhibition is relieved by TGF- beta signalling, to which the future endoderm cell is sensitised by VegT, resulting in the observed dependence on cell contact for maintained Xsox17 expression. We propose that this change in regulation is a consequence of a VegT-induced repressor, inhibiting direct induction of early endoderm markers by VegT, and contributing to the formation of the boundary of the endodermal domain.

Redundant early and overlapping larval roles of Xsox17 subgroup genes in Xenopus endoderm development

We have used antisense morpholino oligos to establish the developmental roles of three Xsox17 proteins in Xenopus development (Xsox17alpha(1), alpha(2) and beta). We show that their synthesis can be inhibited with modest amounts of oligo. The inhibition of each individually produces defects in late midgut development. Loss of activity of the Xsox17alpha proteins additionally inhibits hindgut formation, and inhibiting Xsox17alpha(1) disrupts foregut development with variable penetrance. When all Xsox17 activity is inhibited cell movements are halted during late gastrulation and the transcription of several endodermally expressed genes is reduced. Thus the Xsox17 proteins have redundant roles in early development of the endoderm and partly distinct roles during later organogenesis.

Initiation and early patterning of the endoderm

We review the early stages of endoderm formation in the major animal models. In Amphibia maternal molecules are important in initiating endoderm formation. This is followed by successive signaling events that establish and then pattern the endoderm. In other organisms there are differences in endodermal development, particularly in the initial, prephylotypic stages. Later many of the same key families of transcription factors and signaling cassettes are used in all animals, but more work will be needed to establish exact evolutionary homologies.

Changes in embryonic cell fate produced by expression of an endodermal transcription factor, Xsox17

Many molecules induce the ectopic expression of tissue-specific genes in Xenopus embryos. Conversely, interfering with their activity disrupts patterns of gene expression, implicating them in normal development. Does this mean that they control cell fate (i.e. position, as well as differentiation)? Xsox17alpha and beta can induce ectopic expression of endodermal markers; inhibiting their function suppresses expression of endodermal marker genes in the developing gut (Cell 91 (1997) 397). Here we show the effect of these manipulations on cell lineage. Expressing Xsox17 in a cells normally fated to become ectoderm causes their descendants either to relocate into the embryonic gut or to die at a late developmental stage. Conversely, disrupting Xsox17 activity in cells normally fated to be endodermal causes them to enter mesodermal and ectodermal lineages.

Mode of action of VegT in mesoderm and endoderm formation

mRNA encoding the T-box transcription factor VegT is located throughout the vegetal pole of the Xenopus egg and is believed to play an important part in endoderm and mesoderm formation. We find that VegT generates endoderm both by cell-autonomous action and by generating TGF-beta family signals, the latter being entirely responsible for its mesoderm-inducing activity. Signalling molecules induced cell-autonomously by VegT include derriere, Xnr4 and activin B. Xnr1 and Xnr2 are also induced, but primarily in a non-autonomous manner. All of these signalling molecules are found in the blastula and gastrula vegetal pole and induce both endoderm and mesoderm in the animal cap assay, and hence are good candidates both for the endogenous zygotic mesoderm-inducing signal and for reinforcing the vegetal expression of endoderm markers.

Xpat, a gene expressed specifically in germ plasm and primordial germ cells of Xenopus laevis

cDNAs specific to vegetal poles of Xenopus gastrula embryos were used as a probe to screen a gastrula vegetal pole cDNA library. One of the novel clones isolated had an RNA expression pattern consistent with it being a component of germ plasm and it was thus named Xpat (Xenopus primordial germ cell associated transcript). The open reading frame encodes a 35 kDa protein with no clear homologies. The RNA is localised to the vegetal pole throughout oogenesis and early cleavage. During gastrulation cells containing this message move internally and at tailbud stages they migrate in an antero-dorsal direction. Xpat mRNA is not detectable once the dorsal mesentery forms. We show that the 3'-UTR is required and is sufficient for localisation of exogenous RNA to the vegetal pole of oocytes. We propose that Xpat UTR-containing transcripts can be localised by the Vg1 or late pathway of mRNA localisation during stage III of oogenesis, but endogenous Xpat appears to be localised earlier by a mitochondrial cloud mechanism similar to that proposed for Xcat-2.

Involvement of NF-kappaB associated proteins in FGF-mediated mesoderm induction

In this report, we have used mRNA injection to study the action of mutants of XrelA, a Xenopus homolog of the RelA (p65) component of NF-kappaB, on the induction of mesoderm in Xenopus embryos. A region of the rel homology domain of XrelA was deleted to create XrelA deltaSP, which retains the dimerization and activation domains, but no longer binds to DNA. We also made an analogous derivative of mammalian NF-kappaB1 (p50). We show that both constructs have dominant inhibitory activity. When message encoding either is injected into eggs or oocytes, DNA binding of rel family members is suppressed, as is transactivation of a kappaB-dependent promoter in embryos. Expression of XrelA deltaSP in animal caps blocks the induction of mesoderm by bFGF. In addition, this mutant prevents elongation movements generated by activin, but has little effect on posterior dorsal cytodifferentiation, which in marked contrast is blocked by inhibition of the FGF signal transduction pathway between the receptor and MAP kinase. The specificity of the XrelA deltaSP effect on FGF signaling is shown by rescue of mesodermal marker expression when XrelA deltaSP is co-expressed with a specific rel inhibitor. The target of these dominant negative constructs seems to be neither XrelA itself, nor p50, but rather some other molecule with which XrelA, rather than NF-kappaB1, heterodimerizes. We show that XrelA deltaSP blocks FGF induction of mesoderm downstream of MAP kinase and Xbra expression. Thus it prevents the maintenance of Xbra expression by inhibiting its autoregulation by embryonic FGF (eFGF). We suggest that XrelA deltaSP differs from other reported inhibitors of FGF signaling because it inhibits only gastrula stage FGF signaling and not the maternally programmed signaling at the blastula stage. Our results therefore suggest that zygotic FGF action is required for cell movements rather than dorsal differentiation.

Xsox17alpha and -beta mediate endoderm formation in Xenopus

We have isolated two Xenopus relatives of murine Sox17 expressed in gastrula presumptive endoderm. Xsox17alpha and -beta expression can be induced in animal caps by activin, but not by FGF. Ectopic expression of these genes in animal caps induces the expression of endoderm markers; this induction is blocked by overexpression of a fusion of the Xsox17beta HMG domain to the Drosophila Engrailed repressor domain, as is induction of endoderm markers by activin and the expression of endodermal markers in whole embryos and isolated vegetal poles. These experiments, as well as the effects of the mRNAs on embryo phenotypes, suggest that the Xsox17 genes mediate an activin-induced endoderm differentiation pathway in animal caps and are involved in normal endoderm differentiation in embryos.

N-acetyl-cysteine causes a late re-specification of the anteroposterior axis in the Xenopus embryo

N-acetyl cysteine is an agent which has been shown to interrupt signal transduction processes linking a wide range of stimuli to the activation of NF-kappa B in mammalian cells. We have investigated its effect on the early development of Xenopus embryos by injecting it into blastulae, using concentrations comparable to those effective on cultured cells. High concentrations at the late blastula or early gastrula stage suppress posterior and enhance anterior development, yielding embryos with enlarged cement glands and otherwise consisting of little except head in extreme cases. Reducing the amount of N-acetyl cysteine injected leads to progressively more posterior structures developing. Injection into one- or two-cell embryos gives similar phenotypes, but of reduced severity and the cement gland is not so enlarged. Explants of animal cap cells taken several hours after injection develop to give large amounts of cement gland material. We have examined the expression of a number of genes in the anteriorised embryos. Posterior markers and Xsna are reduced. Noggin and Goosecoid mRNA are up-regulated through the gastrula and persist at these levels until at least the late neurula stage, whereas in controls Noggin is much lower and Goosecoid is absent at these stages. The most anteriorised phenotype may be a consequence of this changed expression.

Developmental effects of over-expression of normal and mutated forms of a Xenopus NF-kappa B homologue

High level over-expression of XrelA1, a homologue of the p65 sub-unit of NF-kappa B and of Drosophila dorsal, arrests Xenopus development at the gastrula stage, producing a reduction in the levels of expression of various genes of developmental interest without general reduction in transcription or cessation of cell division. There is little Goosecoid expression, even though a dorsal lip forms. At lower levels XrelA1 mRNA primarily produces disruption of the mid-dorsal axis. A dominant interference gene product, delta 222, produces mainly posterior, but also anterior abnormalities. On the basis of these results we postulate that the role of XrelA1 in the vertebrate embryo is unlikely to be in dorsoventral development, but more likely in the formation of the termini.

XrelA, a Xenopus maternal and zygotic homologue of the p65 subunit of NF-kappa B. Characterisation of transcriptional properties in the developing embryo and identification of a negative interference mutant

We have isolated two clones (XrelA.1 and XrelA.2) from Xenopus ovary representing differentially processed mRNAs homologous throughout their translated regions to the mammalian p65 subunit of NF-kappa B. The transcripts are ubiquitously present throughout development, but are most abundant in late blastulae and gastrulae. Overproduced protein shows nuclear localisation in both oocytes and early embryos. The XrelA.2 product bound to DNA as an oligomer which was not detected in the normal embryo. Two endogenous kappa B-binding complexes were present, showing no stage-specific variation, although one was relatively deficient in posterior regions of the early neurula. They were not disrupted by dimerization with over-expressed XrelA, suggesting that they were not produced by NF-kappa B/Rel/dorsal family members. The transcriptional properties of the cloned XrelA were assayed in intact embryos by co-injecting XrelA mRNA and a linear HIV LTR-driven CAT reporter gene. CAT levels were stimulated 20-30-fold by XrelA mRNA levels in the 100 pg range, and this was wholly dependent on NF-kappa B binding sites, and largely dependent on those for SP-1. These results were remarkably reproducible and show that quantitative analysis of transcription factor function is possible in intact developing Xenopus embryos A mutant lacking the transcriptional activation domain antagonised co-injected wild-type XrelA, providing a potential dominant negative p65 mutant for interfering with NF-kappa B function in analysing NF-kappa B function in normal development.

The MyoD binding site is dispensable for cardiac actin gene expression in the somites of later stage Xenopus embryos

We have studied the transcription directed by the promoter of a Xenopus borealis cardiac actin gene fused to a globin reporter sequence after injection into X. laevis embryos. This promoter is known to be effective specifically in somites of neurula stage embryos, and work on a very similar X. laevis gene has indicated the importance of an upstream MyoD binding site (E box) in this process. We show that, although in the absence of the MyoD site there is a small amount of tissue-specific expression in neurulae, transcription is abundant in the tailbud embryo a few hours later. Thus, although the E box is initially essential, other sequences must adopt the same role soon afterwards.

Determination of the sequence requirements for the expression of a Xenopus borealis embryonic/larval skeletal actin gene

In this study, we demonstrate that all sequences necessary and sufficient for the expression of a Xenopus borealis alpha 3B embryonic/larval skeletal actin gene, reside in a 156-nucleotide fragment of the promoter that spans nucleotides -197 to -42. This region of the promoter contains three imperfect repeats of the CC(A/T)6GG (CArG) box motif that have been demonstrated to be important in the expression of other sarcomeric actin genes. Deletion of the actin promoter, using Xenopus microinjection techniques as a transient assay system for promoter activity, shows that the most distal CArG box (CArG box 3) is essential for the full expression of the gene. Under our assay conditions, the most proximal CArG box (CArG box 1) exhibits two binding activities using bandshift analysis. One of these binding activities contains components antigenically related to a serum-response factor (transcription factor), whilst the second does not. In contrast, CArG box3 produces only a single retarded band using electrophoretic mobility-shift analysis. Although the shifted complex coelectrophoreses with the CArG box 1/serum-response factor complex, the band produced by CArG box3 appears to be distinct from SRF. In addition to the CArG motifs, a further upstream regulatory element has been identified in the actin promoter between nucleotides -197 and -167. In the actin promoter, a downstream region can apparently fulfil this function.

Regulation of expression of a Xenopus borealis embryonic/larval alpha 3 skeletal-actin gene

We have isolated a genomic clone, related in sequence to the skeletal-actin gene sub-family. It is expressed in the skeletal muscle of embryos from the neurula stage onwards and in tadpoles, but not in adults. The equivalent Xenopus laevis gene is expressed as a major transcript in adult muscle, as well as at earlier stages. The intron/exon structure is typical of vertebrate skeletal-actin genes, as is the possession of multiple copies of three serum-response elements in the promoter of this gene. The Xenopus actin and beta-globin genes were fused in their second introns. This construct, which contained 2.4 kb of upstream sequence, was injected into fertilized eggs at the two-cell stage. It showed the normal pattern of tissue-specific transcription. Thus all of the information necessary for appropriate expression of this actin gene in the embryo is contained in the region that extends from a point 2.4 kb upstream of transcription initiation to the centre of the second exon. A series of enhancer constructs were made in which upstream regions of the actin gene were placed upstream of a X. laevis beta-globin gene. The region immediately adjacent to the promoter, containing the three serum-response elements, was able to drive muscle-specific expression, and there was also a general enhancement of transcription by regions further upstream.

Mutational analysis of the galactose binding ability of recombinant ricin B chain

Ricin B chain (RTB) is an N-glycosylated galactose-specific lectin which folds into two globular domains. Each domain binds one galactoside. The x-ray crystallographic structure has shown that the two binding sites are structurally similar and contain key binding residues which hydrogen bond to the sugar, and a conserved tripeptide, Asp-Val-Arg. We have used oligonucleotide site-directed mutagenesis to change either the binding residues or the homologous tripeptide in one or other or in both of the sites. The 5' signal sequence and RTB coding region were excised from preproricin cDNA and fused in frame to generate preRTB cDNA. Transcripts synthesized in vitro from wild-type or mutant preRTB cloned into the Xenopus transcription vector pSP64T using SP6 RNA polymerase, were microinjected into Xenopus oocytes. The recombinant products were segregated into the oocyte rough endoplasmic reticulum and core-glycosylated, and the N-terminal signal peptide was removed. Mutating sugar binding sites individually did not abrogate the lectin activity of RTB. When both sites were changed simultaneously, RTB was produced which was soluble and stable but no longer able to bind galactose. Changing the Asn residues of the two RTB N-glycosylation sites to Gln showed that oligosaccharide side chains were essential for both the stability and biological activity of recombinant RTB.

The effects of N-glycosylation on the lectin activity of recombinant ricin B chain

Soluble, biologically-active recombinant ricin B chain has been produced by expressing B chain-encoding DNA in heterologous eukaryotic or prokaryotic hosts. N-Glycosylated recombinant ricin B chain expressed in Xenopus oocytes bound to both immobilized asialofetuin and immobilized lactose. Non-glycosylated ricin B chain expressed in either E. coli or in tunicamycin-treated oocytes did not bind to immobilized lactose. However, it did bind to asialofetuin, and increasing concentrations of free lactose did not reduce this asialofetuin binding dramatically, in contrast to the effect of free lactose on the binding of either glycosylated recombinant B chain or native ricin B chain.

Spatial aspects of neural induction in Xenopus laevis

A monoclonal antibody, 2G9, has been identified and characterised as a marker of neural differentiation in Xenopus. The epitope is present throughout the adult central nervous system and in peripheral nerves. Staining is first detected in embryos at stage 21 in the thoracic region. By stage 29 it stains the whole central nervous system, except the tail tip. The epitope is present in a 65K Mr protein, and includes sialic acid. The antibody also reacts with neural tissue in mice and axolotls and newts. 2G9 was used to show that both notochord and somites are capable of neural induction, and the stimulus is present as late as stage 22. Attempts to demonstrate the induction of nervous system by developing nervous system (homoiogenetic induction) were unsuccessful. The view that the lateral extent of the nervous system might be determined by that of the inductive stimulus is discussed. Neural induction was detected as early as stage 10 and occurs in embryos without gastrulation and without cell division from stage 7 1/2.

The organisation and expression of histone genes from Xenopus borealis

We have isolated genomic clones from Xenopus borealis representing 3 different types of histone gene cluster. We show that the major type (H1, H2B, H2A, H4, H3), present at about 60-70 copies per haploid genome (1), is tandemly reiterated with a repeat length of 15 kb. In situ hybridization to mitotic chromosomes shows that the majority of histone genes in Xenopus borealis are at one locus. This locus is on the long arm of one of the small sub-metacentric chromosomes. A minor cluster type with the gene order H1, H3, H4, H2A is present at about 10-15 copies. The genome also contains rare or unique cluster types present at less than 5 copies having other types of organisation. An isolate of this type had the gene order H1, H4, H2B, H2A, H1 (no H3 cloned). Microinjection of all of the clones into Xenopus laevis oocyte nuclei shows that most of the genes present are functional or potentially functional and a number of variant histone proteins have been observed. S1 mapping experiments confirm that the genes of the major cluster are expressed in all tissues and at all developmental stages examined.

Cytoskeletal actin gene families of Xenopus borealis and Xenopus laevis

We have sequenced the coding and leader regions, as well as part of the 3' untranslated region, of a Xenopus borealis type 1 cytoskeletal actin gene [defined according to the arrangement of acidic residues at the N-terminus; Vandekerckhove et al. (1981) J Mol Biol 152:413-426]. The encoded amino acid sequence is the same as the avian and mammalian beta (type 1) cytoskeletal actins, except for an isoleucine at position 10 (as found in the mammalian gamma cytoskeletal actins), and an extra amino acid, alanine, after the N-terminal methionine. Five introns were found, in the same positions as those of the rat and chicken beta-actin genes. The 5' and 3' untranslated regions resemble those of the human gamma (type 8) cytoskeletal actin gene more closely than the mammalian beta genes. Primer extension showed that this type 1 gene is transcribed in ovary and tadpole. Sequencing of primer extension products demonstrated two additional mRNA species in X. borealis, encoding type 7 and 8 isoforms. This contrasts with the closely related species Xenopus laevis, where type 4, 5, and 8 isoforms have been found. The type 7 isoform has not previously been found in any other species. The mRNAs of the X. borealis type 1 and 8 and X. laevis type 5 and 8 isoforms contain highly homologous leaders. The X. borealis type 7 mRNA has no leader homology with the other mRNA species and, unlike them, has no extra N-terminal alanine codon. The evolutionary implications of these data are discussed.

The development of an assay to detect mRNAs that affect early development

We have constructed an assay to identify developmental effects of injected RNA molecules. The RNA is injected into the animal pole region of a 2- to 8-cell embryo. At the blastula stage, the animal cap is removed and its development in isolation tested. In controls, only epidermis is produced, but several of the injected RNA preparations, though not all, also form dorsal and ventral mesoderm and nervous tissue. This assay should be suitable for selecting cDNA clones complementary to mRNAs that direct development.

Tissue-specific expression of actin genes injected into Xenopus embryos

We have isolated a complete Xenopus borealis cardiac actin gene, which is normally expressed in the myotomes and heart of the embryo and tadpole. After injection into the zygote, this cloned gene becomes distributed throughout the embryo, but it is expressed almost wholly in the myotomes. The same wide distribution of injected DNA but spatially restricted pattern of expression is found with a fusion between the first two actin gene exons and the last exon of a mouse beta-globin gene. By contrast, a histone-globin fusion gene is expressed fairly uniformly in all regions. We discuss the special advantages of using Xenopus in studies of tissue-specific gene expression from injected, cloned genes in early development.

Development of the ectoderm in Xenopus: tissue specification and the role of cell association and division

When do ectodermal cells become specified to form epidermis, that is, become committed to form epidermis when isolated? Animal pole explants of Xenopus cultured in saline from the 8-cell stage onwards express a specific epidermal antigen, whereas vegetal explants do not. The isolated outer layer of ectoderm formed by stage 7 is almost completely specified, the inner layer is only partially so. When are cell associations and cell divisions necessary for epidermal differentiation? Embryos that were either disaggregated or incubated in cytochalasin B after the midblastula stage do not require cell interactions, Ca2+, or cell divisions for epidermal differentiation to occur. Inhibition of mRNA transcription with actinomycin D shows that the epidermal antigen is certainly transcribed by the late gastrula stage (stage 12).

Polyadenylation of histone mRNA in Xenopus oocytes and embryos

Oogenesis of amphibians is an atypical situation in which histone mRNA is polyadenylated. The poly(A) tract on H4 mRNA has been examined by S1 nuclease analysis. Throughout oogenesis the poly(A) tract is very short, and nonexistent on some mRNA molecules. The poly(A) tract is completely removed during maturation of the oocyte, and is absent in embryos and cultured cells.

Are there major developmentally regulated H4 gene classes in Xenopus?

Primer extension analysis has been used to study the principal H4 mRNAs present at different developmental stages and in several adult tissues of Xenopus borealis and X. laevis. In X. borealis a single sequence class predominates in oocytes, tadpoles and cultured fibroblasts. There is also a polymorphic minor type which shows no developmental regulation. The primer extension bands obtained from adult liver and kidney RNA appear to be the same as ovary and therefore these tissues almost certainly contain the same major H4 mRNA species. This is confirmed by S1 mapping of the 3' end of the mRNA. Thus for H4 genes in X. borealis there is no evidence of the kind of switches in histone gene expression seen in sea urchins or certain protostomes. The situation in X. laevis is complicated by considerably higher gene variability both within and between individuals. Nevertheless, in this species, as in X. borealis, there seems to be no major developmental switch in the regulation of H4 gene expression, a conclusion that also holds for an H1C and an H3 gene.

Nucleotide sequences of H1 histone genes from Xenopus laevis. A recently diverged pair of H1 genes and an unusual H1 pseudogene

Four clones containing H1 histone gene sequences were previously isolated from a Xenopus laevis genomic library (1) and we now present the complete nucleotide sequences of these H1 genes and their flanking regions. Two of these genes code for minor H1 proteins, probably H1C, when expressed in the oocyte transcription/translation system and are present on clones with almost identical overall organization. However, at the nucleotide level these genes differ in showing base insertions and deletions, as well as substitutions. A third gene sequence which is more related to the major X. laevis H1A, corresponds to the 3' two thirds of an H1 gene. This gene has in place of a 5' coding region at least 1800 bp of apparently noncoding sequence, some of which is A-T rich. The junction does not correspond to the consensus sequence of an intron/exon boundary and therefore this H1 sequence is more likely to represent a pseudogene. Comparisons of the coding and flanking regions of these X. laevis H1 genes indicate the kind of differences which can occur among H1 subtypes within a species. A region of homology noted in the 3' noncoding portion of vertebrate histone genes is discussed in relation to the mechanism of termination of transcription.

Histone gene number and organisation in Xenopus: Xenopus borealis has a homogeneous major cluster

Using a Xenopus laevis H4 cDNA clone as a probe we have determined that the numbers of H4 histone genes in Xenopus laevis and Xenopus borealis are approximately the same. These numbers are dependent on the hybridization stringency and we measure about 90 H4 genes per haploid genome after a 60 degrees C wash in 3 X SSC. Using histone probes from both Xenopus and sea urchin we have studied the genomic organization of histone genes in these two species. In all of the X.borealis individuals analyzed about 70% of the histone genes were present in a very homogeneous major cluster. These genes are present in the order H1, H2B, H2A, H4 and H3, and the minimum length of the repeated unit is 16kb. In contrast, the histone gene clusters in X.laevis showed considerable sequence variation. However two major cluster types with different gene orders seem to be present in most individuals. The differences in histone gene organization seen in species of Xenopus suggest that even in closely related vertebrates the major histone gene clusters are quite fluid structures in evolutionary terms.

Organization and expression of cloned histone gene clusters from Xenopus laevis and X. borealis

We have isolated several clones containing Xenopus histone genes from genomic libraries of X. laevis and X. borealis DNA. Each genomic clone has been mapped and the positions of 26 histone genes in seven laevis clones and 5 histone genes in one borealis clone have been determined. In laevis, the histone gene clusters show considerable variation in gene order within a single individual. When the cloned DNAs were microinjected into the nucleus of Xenopus oocytes, expression of cloned genes at the transcriptional and translational level was readily detectable. Previously unknown histone variants were revealed by the microinjection experiments.

The DNase I sensitivity of Xenopus laevis genes transcribed by RNA polymerase III

Since the initial discovery that the DNase I sensitivity of the globin genes in different cell types correlates with globin gene expression, this relationship has been shown to hold true for a variety of genes, including the genes for ovalbumin, conalbumun, alpha- and beta-globin in chicken, several heat-shock proteins in Drosophila, the r-chromatin of Tetrahymena and the viral polyoma minichromosome. Although genes transcribed by RNA polymerases I and II have been studied extensively, the genes transcribed by RNA polymerase III have not. We have therefore investigated the DNase I sensitivity of transfer RNA (tRNA) and oogenetic 5S RNA genes in the liver and erythocyte nuclei of Xenopus laevis. The oogenetic 5S genes are not transcribed in any known somatic cell, and tRNA genes are transcribed in the hepatocyte but are inactive in the erythrocyte. We show here that, although in these two cell types the correspondence between DNase I sensitivity and gene transcription holds good for globin and the ribosomal genes, the tRNA and oogenetic 5S genes are DNase I sensitive in both liver and erythrocyte nuclei. Thus for the genes transcribed by polymerase III the correspondence of sensitivity and expression breaks down.

H3 and H4 histone cDNA sequences from Xenopus: a sequence comparison of H4 genes

Ovarian poly (A) + RNA from Xenopus laevis and Xenopus borealis was used to construct two cDNA libraries which were screened for histone sequences. cDNA clones to H4 mRNA were obtained from both species and an H3 cDNA clone from Xenopus laevis. The complete DNA sequences of these clones have been determined and are presented. These new sequences are compared with other H3 and H4 DNA sequences both in the coding and 3' noncoding regions. We find that there is considerable non-random codon usage in ten H4 genes. In addition there are some sequence similarities in the 3' noncoding regions of H3 and H4 genes.

Stability of non-polyadenylated viral mRNAs injected into frog oocytes

The naturally non-polyadenylated mRNAs of reovirus were shown to have a half-life in excess of 3 days when injected into Xenopus laevis oocytes. Hybridization analysis and gel electrophoresis showed that all 10 mRNA species had a similar high stability, despite being translated at widely differing rates. We have confirmed previous findings indicating a role for the 5'-terminal cap structure in determining reovirus mRNA stability [Furuichi et al. (1977) Nature (Lond.) 266, 235-239]. The significance of these results in relation to a general role for poly(A) in messenger RNA function is discussed.

Paternal gene expression in developing hybrid embryos of Xenopus laevis and Xenopus borealis

We have studied protein synthesis in the viable hybrid Xenopus laevis (female) X Xenopus borealis (male) using 2D gel electrophoresis. Fourteen borealis-specific proteins were studied. Two of these proteins appeared by the gastrula stage, five in the gastrula and the rest later. Where homologous laevis proteins were tentatively identified, androgenetic haploid hybrids were used to study whether the protein was encoded by stored maternal mRNA, and how long this mRNA persisted. The two proteins appearing in blastulae were probably initially coded by stored maternal mRNA. This was not detectable by the tailbud-tadpole stage, and presumably had been destroyed.

Actin synthesis during the early development of Xenopus laevis

Cytoskeletal beta and gamma-actin are amongst the most rapidly made proteins of oocytes, blastulae and late embryonic stages of Xenopus laevis but, relative to other proteins, the rate of synthesis is low in the egg or cleaving embryo, although the messenger RNA is present in comparable amounts at the different stages. Actin synthesis therefore involves post-transcriptional regulation. alpha-Actin, the actin class characteristics of straited muscle cells, is first detectable in late gastrulae and it is an abundant newly synthetized protein from the neurula stage onwards. mRNA template activity for this protein is not detectable before the gastrula stage. Thus alpha-actin synthesis probably reflects new gene action, confined to part of the embryo, for alpha-actin only appears in the section which includes presumptive skeletal muscle cells. It therefore constitutes the earliest cyto-specific protein so far demonstrated in Amphibia. When tadpole tail poly(A)-containing mRNA is injected into oocytes and eggs alpha-actin synthesis is seen in both cases. Extensive evidence for the identification of the actins is presented. This is based on location of synthesis, DNase-I binding and partial peptide mapping.