Keywords

These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.

7.1 Introduction

For decades, Xenopus oocytes and embryos have been used as an experimental system for the analysis of a wide range of molecular, cellular, and developmental processes. Since hundreds of oocytes can be obtained from a single female, oocyte extracts have been used extensively for biochemical analyses of cell cycle regulation, chromatin assembly, and the dynamics of microtubule/spindle assembly. Intact oocytes may be cultured for several days and oocyte maturation may be induced by treatment with progesterone. Microinjection of specific mRNAs in immature oocytes has been commonly used to evaluate protein function and characterize structure–function relationships for ion channels and other membrane proteins.

Xenopus embryos are suitable for both mechanistic and functional studies of embryonic development; they are perhaps the primary vertebrate model system for investigations of molecular and biochemical mechanisms underlying early developmental processes. Over- or mis-expression of specific genes is easily accomplished by microinjection of exogenous RNAs, and these can be targeted to specific cells, or blastomeres, within the early embryo. Loss-of-function analyses are generally carried out via introduction of modified oligonucleotides, either morpholino oligonucleotides or 2-o-methyl oligonucleotides, directed against either the translation start site or a splice acceptor site, as will be discussed in more detail below. These modified oligonucleotides may be introduced into the immature oocyte by host-transfer methods to disrupt maternal expression of a given gene, or into the early embryo, to inhibit zygotic gene function.

Embryos develop normally in a defined saline solution, and tissue isolates can be easily cultured to monitor cell behavior or to study terminal differentiation. These embryos are amenable to microsurgical manipulation, and such studies have yielded fundamental insights into mechanisms of developmental commitment, and the establishment of spatial organization in vertebrate embryos. The ability to introduce isolated nuclei via microinjection into enucleated oocytes led to the first clear demonstration of genomic equivalence, establishing the conceptual foundation for current studies of stem cells and induced pluripotent (iPS) cells.

Experimental studies in Xenopus have provided major advances in our understanding of mechanisms of cell signaling, translational control, transcriptional regulation, early embryonic development, organogenesis, and regeneration. Until recently, these studies have used Xenopus laevis almost exclusively (Fig. 7.1). Over the last decade, however, Xenopus tropicalis has emerged as a model system for genetic analyses of early development, providing the potential to integrate genetic approaches with biochemical and molecular strategies in a relatively inexpensive and fast-developing vertebrate model system.

Fig. 7.1
figure 0071

Key stages in the life cycle of Xenopus laevis. Stages shown include a 16-cell embryo, midgastrula, late neurula, tadpole, and adult

Full size image

7.2 Evolutionary Origins and Genome Organization

X. tropicalis is a true diploid, with a genome size estimated at ~1.7 GB and 20 chromosomes (2n). In contrast, X. laevis is considered an allotetraploid, with a genome size of 3.1 GB and 36 chromosomes (2n). Both the oocytes and the adults of X. tropicalis are considerably smaller than X. laevis, and the generation time is shorter as well: X. tropicalis reaches sexual maturity in 4–6 months, compared with up to 2 years in X. laevis (see Table 7.1).

Table 7.1 Key characteristics of Xenopus laevis vs. Xenopus tropicalis
Full size table

The Xenopus genus consists of two clades: the larger clade that includes X. laevis and at least 15 other species, and a second clade that includes the diploid X. tropicalis and the tetraploid X. epitropicalis (Evans 2008). While this second clade is sometimes given the genus name Silurana, both clades are thought to originate from a common ancestor that existed ~50 million years (de Sá and Hillis 1990; Hellsten et al. 2007; Evans 2008). In this scenario, Xenopus laevis originated from the hybridization of two closely related species, yielding a stable allotetraploid genome retaining a large part of the chromosomal complement of both species. This allotetraploidization event is thought to have occurred approximately 40 million years ago (Bisbee et al. 1977; Graf and Kobel 1991; Evans et al. 2004). Hybridizations of these hybrids have occurred during the subsequent interval, resulting in allooctaploid (e.g., X. wittei) and allododecaploid (e.g., X. ruwenzoriensis) members of the Xenopus genus, along with the occurrence of other allotetraploid species (e.g., X. gilli, X. borealis). In X. laevis, early estimates (Bisbee et al. 1977) suggest that approximately half of the duplicated genes are retained, while later analysis has identified A and B copies, or homeolog pairs, of roughly 75% of protein-coding genes (Harland et al. personal communication). Surviving duplicated genes may undergo functional divergence: either sub-functionalization, as the original gene function is partitioned between the copies, or neo-functionalization, as one copy acquires a novel function. It is not clear what level of functional divergence would be expected in an allotetraploid like X. laevis, and while one study suggests that it was limited to <20% (Morin et al. 2006), a second concludes that over 50% of homeolog pairs are affected in this manner (Hellsten et al. 2007). Sequence comparisons indicate that genes with functional homeologs evolve at a higher rate relative to genes that are present as only one functional copy or to their X. tropicalis orthologs (Hellsten et al. 2007).

Most laboratory colonies of X. tropicalis have been generated with animals of either the Nigerian or Ivory Coast strains; both strains have been inbred for multiple generations. Ivory Coast strains include both the IC and TGA strains: the IC strain, like the Nigerian strain, has been inbred from animals originally obtained by Marc Kirschner, while the TGA strain is traced back to a colony that originated with Louis Du Pasquier at the Basel Institute for Immunology.

7.3 Genomics Resources for Xenopus

7.3.1 The X. tropicalis Genome Project

Sequencing of the X. tropicalis genome began in 2002 at the Joint Genome Institute (JGI), using DNA from a Nigerian female, inbred for seven generations. The genome was assembled using JGI’s JAZZ assembler (Aparicio et al. 2002), culminating in the release of assembly v4.1 (Hellsten et al. 2010). This assembly, representing 7.6X coverage, includes approximately 1.36 Gb over 19,759 scaffolds; half of the X. tropicalis genome is represented in the first 272 scaffolds, which are 1.56–7.82 Mb in size. Long-range sequential ordering of scaffolds was based on bacterial artificial chromosome (BAC) end-sequencing data, the X. tropicalis meiotic map (Wells et al. 2011), and synteny with other vertebrate genomes, although this did not form part of the publicly released version of the genome. Over 97% of known Xenopus tropicalis cDNAs were mapped at >95% identity over >50% of their length in this assembly. Regions showing conserved synteny with human and/or chicken genomes are distributed throughout the genome. The genome assembly v4.1 has been extensively annotated, and over 20,000 gene models have been identified; these are available on Xenbase. As of this writing, v4.1 is available at the University of California Santa Cruz (UCSC) Genome Browser, National Center for Biotechnology Information (NCBI), and Ensembl and Xenbase Web sites.

Limited additional sequence data have become available since the release of v4.1, and subsequent assemblies have been generated using the ARACHNE assembler at the Hudson-Alpha Institute. The most recent assembly, v7.1, has been informally released into the public domain, has been annotated with gene models, and is now available at Xenbase (for review see Gilchrist 2012). Unlike the previous assembly, this has been released as a substantially chromosomal-scale assembly, using the meiotic map and syntenic alignments with chicken to order the scaffolds into super-scaffolds. The assembly contains just under 1.4 Gb of sequence data, and the super-scaffolds have been numbered to correspond to the ten X. tropicalis chromosomes, although chromosomes 3, 5, and 8 are fragmented into two or three sections. Figure 7.2 shows a karyotype for X. tropicalis and the sizes of the corresponding super-scaffolds. Over 95% of the assembled sequence is included in scaffolds >50 kb in length. Most of the short-range discrepancies between the meiotic map and the assembled sequence that characterized earlier assemblies have been eliminated. There are two exceptions: the short arm of Chromosome 7 (Chr7p) in which discrepancies between the map and the assembled sequence persist, and the short arm of Chromosome 2 (Chr2p), which is not represented on the meiotic map (see below). The assembly is now available on Xenbase, both for download and through the genome browser, but at the time of writing it is not yet available on other public genome browsers.

Fig. 7.2
figure 0072

Xenopus tropicalis chromosomes and the sizes of the corresponding v7.1 super-scaffolds. Chromosome spreads were prepared as described (Wells et al. 2011) and stained with propidium iodide.

Full size image

7.3.2 BAC Resources

BAC libraries have proven invaluable for genome assembly, functional analysis of promoters and enhancers, and investigations of gene structure. BAC libraries have been generated for X. laevis and X. tropicalis by the Children’s Hospital of Oakland Research Institute (CHORI), the Institute of Systems Biology (ISB), and Amplicon Express (Pullman WA), among others (Table 7.2). The Adiopoudoume library, generated by Nicolas Pollet from Ivory Coast strain frogs in collaboration with Amplicon Express (Pullman WA), provides an essential complement to the other BAC libraries, which were generated from Nigerian strain individuals. Clones from the CHORI-216 library were end-sequenced and fingerprinted to aid in genome assembly; end-sequencing of clones from the Adiopoudoume library is in progress (Pollet, personal communication). Each of these libraries is publicly/commercially available, either through CHORI (CHORI libraries, ISB library) or through Amplicon Express/Genomex (Adiopoudoume library).

Table 7.2 Xenopus BAC libraries
Full size table

Additional long-range physical mapping of the X. tropicalis genome is in progress via a pilot HAPPY mapping project (Jiang et al. 2011). HAPPY mapping, a PCR-based method for assessing representation of specific sequences within a panel of fragmented haploid DNA, has not yet been used extensively for mapping of metazoan genomes. As Jiang and colleagues point out, however, the adaptation of HAPPY mapping strategies to whole-genome amplification, and massively parallel sequencing should provide a highly effective approach for the generation of long-range physical maps in Xenopus.

7.3.3 A Meiotic Map for X. tropicalis

A 2,886 simple sequence-length polymorphism (SSLP) meiotic map of the X. tropicalis genome has been generated (Wells et al. 2011). The map is 1,658 cM in size, with a theoretical limit of resolution of 0.26 cM. The ten major linkage groups identified in this map correspond to the ten chromosomes; the four additional minor linkage groups have also been assigned to specific chromosomes. The map has a minimal genomic representation of 66% of the X. tropicalis genome, based on comparisons with X. tropicalis Assembly 4.1, i.e., the scaffolds containing the SSLP marker sequences comprise 66% of the total genomic sequence.

Cytogenetic analyses indicate that this map spans nearly the entire length of most of the chromosomes, with the exception of the p arm of Chr2. This region is not represented in the meiotic map, and SSLPs from genome scaffolds assigned to this region by cytogenetic mapping show markedly reduced levels of polymorphism. Ongoing genetic analyses indicate that levels of recombination may be reduced in this region (Zimmerman, personal communication), which distorts genetic distances relative to the map as a whole. Moreover, these genome scaffolds are underrepresented among the SSLPs, given their size. These observations suggest that Chr2p may include one or more large regions of significantly reduced polymorphism, as might arise from a recent selective sweep. The gene content of Chr2p may be clarified by ongoing studies that use laser dissection to isolate this chromosomal region, which will then be sequenced (Krylov and Pollet, personal communication). Further investigation into these and other chromosomal aspects of genome organization will benefit from higher-resolution cytogenetic mapping using the highly extended “lampbrush” chromosomes obtained from oocytes in diplotene (st. IV–VI) (Penrad-Mobayed et al. 2009).

7.4 Genetic Strategies for X. Tropicalis

7.4.1 Forward Genetic Screens

Forward genetic screens in X. tropicalis have identified a number of developmentally significant mutations (Goda et al. 2006); some of these have been shown to correspond to genes associated with human diseases (Zimmerman, personal communication). Additional gynogenetic screens have sought to uncover recessive mutations existing in the major laboratory strains (e.g., Noramly et al. 2005); mutations revealed in these screens produce a range of organ-specific defects. Gynogenetic screens are accomplished by suppression of polar body formation, generally via an early cold shock of the oocytes; since the regions flanking the centromeres are homozygous, this strategy also provides the basis for gynogenetic mapping of mutations arising in mutagenized females, described below.

A complementary, reverse genetic approach has used targeting-induced local lesions in genomes, or TILLING, to retrieve mutations in genes on interest from progeny of mutagenized animals. TILLING has been used extensively in many plant systems, as well as in Drosophila and Zebrafish (Winkler et al. 2011). Although many of the X. tropicalis mutations uncovered to date via pilot TILLING efforts have been silent with respect to protein sequence (Showell et al. 2011), larger-scale TILLING efforts now in progress are expected to identify mutations that yield specific phenotypes.

Mapping of mutations identified by forward genetic screens has been accomplished using the SSLP map (Wells et al. 2011), often in combination with gynogenetic mapping (Khokha et al. 2009) and additional amplified fragment length polymorphisms (AFLPs). More recent efforts take advantage of exon capture strategies, in which fragments of genomic DNA containing protein-coding exons and their flanking regions are captured using anchored oligonucleotide and sequenced to identify mutations.

7.4.2 Transgenesis Systems

Since the initial demonstration of transgenesis via nuclear injection (Kroll and Gerhart 1994), methods for transgenesis in Xenopus have undergone substantial development. It is relatively simple to generate hundreds of transgenic embryos from a single fertilization to carry out “F0” experiments, using either X. laevis or X. tropicalis. The restriction enzyme-mediated insertion (REMI) method (Kroll and Amaya 1996) involves incorporation of exogenous DNAs in isolated sperm nuclei treated with restriction enzymes. Although this method is still frequently used, several modifications have simplified this technique, in particular, the elimination of the restriction enzyme treatment for intracytoplasmic sperm injection (ISCI) (Sparrow et al. 2000). An alternative system uses IsceI “Meganuclease” (Ogino et al. 2006; Pan et al. 2006); this method is highly sensitive to the activity of the enzyme. Transgenesis based on the phiC31 integrase has also been reported (Allen and Weeks 2009). Chesneau et al. (2008) offer a superb review comparing the merits of these transgenesis strategies.

Transposon-mediated transgenesis has been developed using the Tol2 and Sleeping beauty (SB) transposable elements. Since the overall frequency of insertion is lower, the likelihood of multiple insertions is reduced; thus, transposon-mediated systems are well suited for the generation of transgenic lines. They are also ideal for insertional mutagenesis screens, such as “gene trap” or “enhancer trap” approaches, especially since the capacity for remobilization of transposons, coupled with the long lifespan of Xenopus, extends the potential duration of such screens. Remobilization of stably integrated SB transposons and subsequent re-integration at distinct sites can occur at both long and short range (Yergeau et al. 2011). Since insertional frequencies are both higher and more consistent with the sperm nuclei-based methods, however, these methods are more suitable than transposon-mediated transgenesis for “F0” experiments.

7.5 The Impact of Genomics on the Investigation of Gene Regulation and Function

Sequencing and assembly of the X. tropicalis genome, and the development of additional genomics resources for both X. tropicalis and X. laevis, have opened the door to a wide range of experimental analyses. This has enabled a number of different experimental approaches: large-scale, systematic investigation of gene function through over- and under-expression screens; exploration of genetic relationships through ChIP analysis; expanding our understanding of epigenetic regulation, genome structure, and organization; and post-transcriptional mechanisms of gene regulation.

7.5.1 Functional Screens

Xenopus embryos have been used extensively for functional molecular screens, and the development of genomic tools and resources has greatly expanded the opportunities for large-scale screens of gene function based on either gain- or loss-of-function. Gain-of-function screens were generally carried out by microinjection of pooled cDNAs; these screens were critical in the identification of many critical regulators of early development (e.g., noggin, Smith and Harland 1992).

The use of morpholino oligonucleotide-based “knockdown” strategies has led to loss-of-function screens for genes that regulate early development in X. tropicalis. The morpholino is a short (~20 bp) synthetic oligonucleotide (“oligo”) designed to reverse complement the RNA sequence of protein coding genes. The oligo is injected into the embryo at an early stage and binds tightly to the RNA, substantially disabling its function. This can be done in two ways: a “translation blocking” morpholino can be designed to complement the 5′ untranslated region (UTR) around the initiating methionine, which prevents progress of ribosomal complex along the mRNA, and thus inhibits protein production. Alternatively, it can be designed to complement the region across (typically) the first exon/intron junction and interfere with normal splicing. The two methods have different advantages. Splice blocking allows rescue experiments to be performed with cloned mRNA reagents, without having to make special constructs (e.g., “naked” open reading frames, or ORFs), while translation blocking would be more useful to disrupt the function of maternal mRNAs (which no longer require splicing). The morpholino can be simply injected into the early embryo, which is subsequently monitored for defective development. The effects can persist for many hours, although they will be eventually be diluted out as the number of cells continues to increase. In addition, an embryo can provide its own control if (for example) only one cell is injected at the two-cell stage. Although this is not a true “knockout” of the targeted gene as some protein may still be produced, results can nevertheless be impressive (for example, Kenwrick et al. 2004; Rana et al. 2006). In excess, morpholinos can be toxic to the embryo, so a rescue experiment is useful. Nevertheless, this is a quick and relatively cheap way to compromise gene function in early vertebrate development, and it can easily be extended to perform “knockdowns” of multiple genes in concert (e.g., Khokha et al. 2005).

7.5.2 Analysis of Epigenetic Regulation

Xenopus embryos are well suited to studies of transcriptional regulatory mechanisms, and the availability of the X. tropicalis genome provides unparalleled opportunities for genome-wide analyses of epigenetic regulation during early embryonic development in vivo. Large-scale epigenetic regulatory mechanisms can be addressed using chromatin immunoprecipitation (ChIP) followed by microarray analysis (ChIP-chip) or sequencing (ChIP-Seq). Nimblegen tiling arrays representing either promoter regions (based on proximity to 5′ ends of annotated genes) or repeat-masked whole-genomic sequence have been used to assess the distribution of individual histone modifications associated with transcriptional activation (e.g., trimethylation of Histone H3 Lys4, H3K4me3) (Akkers et al. 2010). The H3K4me3 mark becomes widespread within the Xenopus genome shortly after the midblastula transition (MBT) (Akkers et al. 2009). Distribution of the H3K4me3 mark strongly overlaps with that of TATA Binding Protein (TBP) binding; since TBP is known to bind to promoter sequences, this correlation confirms the H3K4me3 mark as a specific identifier of active promoters in early Xenopus embryos. Moreover, there is significant agreement between the results of both the promoter-specific and genome-wide arrays, validating the use of the more limited promoter arrays.

7.5.3 Translational Control and Noncoding RNAs

Xenopus oocytes and early embryos have provided an outstanding experimental system for the analysis of translational control in vivo, revealing several distinct phases characterized by distinct mechanisms of post-transcriptional gene regulation. First, over the course of oogenesis, transcripts are sorted and localized through the early message transport organizer (METRO) and the late pathways (Kloc and Etkin 1995; reviewed in King et al. 2005). Large-scale changes in polyadenylation during oocyte maturation alter patterns of transcript stability and availability for translation (reviewed in Richter and Lasko 2011). PolII-dependent transcription is almost entirely silent for the first 12 rounds of cleavage until the MBT; at this time, the cells shift to a conventional cell cycle, acquiring G1 and G2 phases, cell division becomes asynchronous, and large-scale zygotic transcription is initiated (reviewed in Maller et al. 2001). Recent work suggests that this transition is mediated by phosphorylation of the DNA replication checkpoint protein Claspin, which occurs when the nuclear/cytoplasmic ratio increases above a threshold (Gotoh et al. 2011). The maternal transcripts responsible for early development are subsequently deadenylated following the initiation of gastrulation, and this deadenylation is governed by the miRNA family miR-430/427/302 (Giraldez et al. 2006; Lund et al. 2009).

7.5.4 Analysis of Noncoding RNAs in Xenopus

MicroRNAs, ~22-nucleotide RNAs that contribute to the regulation of translation and RNA stability, have been investigated in Xenopus via multiple approaches, including massively parallel sequencing studies, bioinformatic analyses, and in situ hybridization screens. Although an early cloning-based study identified approximately 30 microRNAs (miRNAs) expressed in early embryos (Watanabe et al. 2005), a bioinformatic characterization of miRNAs within the X. tropicalis genome has identified over 140 miRNAs (Tang and Maxwell 2008); most of these are intronic miRNAs, while <20% are intergenic. Over 190 Xenopus miRNAs are listed in miRBase (http://www.miRBASE.org); many of these are expressed in spatially dynamic patterns during embryogenesis (Walker and Harland 2008). The high degree of sequence conservation between X. tropicalis and X. laevis will facilitate the use of microRNA sequences initially identified in X. tropicalis for functional studies that are more extensively developed in X. laevis. A sequencing-based characterization of small RNAs in the X. tropicalis oocytes has identified several classes of small RNAs, including miRNAs, piwi protein-associated RNAs (piRNAs), and endogenous short interfering RNAs (siRNAs) (Armisen et al. 2009).

Other noncoding RNAs have been investigated most extensively in oocytes. The noncoding RNA Xlsirts, along with the VegT transcript, has been shown to be required for organization of the vegetal cytokeratin network in Xenopus oocytes; both transcripts are necessary for germ cell development (Kloc et al. 2005). More recent microarray analyses have identified an additional noncoding RNA associated with the germ plasm (Cuykendall and Houston 2010).

Piwi protein-associated RNAs (piRNAs) are 24–30 nucleotide RNAs that are required for silencing of retrotransposons and genome stability within the germ line; they are also known to function in other contexts (reviewed in Klattenhoff and Theurkauf 2008). Massively parallel sequencing studies have revealed expression of piRNAs in oocytes (Armisen et al. 2009) and gastrulae (Robine et al. 2009). They have also been detected, along with expressed Tc1/mariner transposable element sequences, in Xenopus embryonic neural tissue, although these piRNAs appear to be derived from the expressed transposons themselves (Faunes et al. 2011). A surprisingly large proportion of the piRNAs identified in oocytes are 3′ UTR fragments apparently derived from cleavage of mRNAs encoding proteins mediating developmental or other regulatory processes (e.g., DNA-binding proteins, kinases), which are not themselves highly expressed (Robine et al. 2009).

7.5.5 Functional Analysis of miRNAs

MiRNAs have been shown to carry out essential functions in oogenesis and early Xenopus development. Recent work shows that miR-16 is required to maintain Xenopus oocytes in a quiescent immature state (Mortensen et al. 2011). Surprisingly, miR-16 acts to upregulate translation of the target gene, Myt1 kinase; translational activation by miRNAs is also observed in quiescent mammalian cells (Vasudevan et al. 2007). Targeting of the nodal/activin receptor Acvr2A by miR-15 and miR-16 in Xenopus embryos leads to spatial restriction of the Organizer; loss-of-function of these miRNAs leads to an expansion of the Organizer (Martello et al. 2007). MiRNA-dependent regulation of nodal signaling is a conserved feature of vertebrates, relying upon a family of miRNAs (miR-430/427/302) that share a common “seed” sequence to regulate translation of nodal ligands and antagonists in anamniote embryos and human embryonic stem cells (Rosa et al. 2009). Recent work indicates that miR-9 has distinct, regionalized functions during neurogenesis in Xenopus: it regulates proliferation via its effects on hairy transcripts, but it is also required to inhibit accumulation of p53 in the forebrain, thus preventing apoptosis (Bonev et al. 2011). Additional functional studies in Xenopus have shown that miRNAs are critical for cell survival and differentiation during retinal development (Decembrini et al. 2009; Walker and Harland 2009), the timing of pronephric differentiation (Agrawal et al. 2009), and the regulation of cell division and differentiation within the embryonic skeletal muscle (Chen et al. 2006).

Recent work by Lund et al. (2011) demonstrates that levels of Argonaute 2 (Ago2) protein are very low, and essentially limiting for microRNA activity, prior to the MBT; as a result, introduction of RNA oligonucleotides intended as siRNAs or miRNAs are not incorporated into Ago2 complexes and may be degraded. Introduction of exogenous Ago2 promotes persistence of RNA oligonucleotides and may aid in the development of siRNA-based approaches. In support of this possibility, two studies (Chen et al. 2009; Pan et al. 2010) have already shown that transgenic expression of Argonaute increases the efficiency of short hairpin RNAs (shRNAs) in knocking down expression of endogenous genes. Further development of these approaches should expand the range of loss-of-function analyses in Xenopus.

7.6 Prospects

The development of genomics resources has produced a wealth of opportunities to carry out systems-level investigations of gene regulation and function in vivo. These studies will allow us to understand the interplay between the well-described signaling pathways that govern early development, such as wnt or nodal signals, with the chromatin modification networks responsible for epigenetic regulation (e.g., Blythe et al. 2010), or other fundamental transcriptional regulators (Yoon et al. 2011). This avenue of study will generate new insight into fundamental mechanisms of developmental commitment. These opportunities extend to post-transcriptional mechanisms of gene regulation, as well as gene function in such diverse areas as cell cycle control (e.g., Mochida et al. 2010) and metabolic regulation (Jorgensen et al. 2009). In view of the range of molecular, biochemical, and developmental approaches currently in use with Xenopus, the continued expansion of genomics resources will further our understanding of basic mechanisms underlying vertebrate development and cell function.