599

The evolution of multicellularity and early animal genomes Nina M Brooke and Peter WH HollandÃ

Several independent molecular datasets, including complete although some expressed sequence tag screens have mtDNA sequence, indicate that are most closely commenced (Hydra EST Database: http://hampson- related to multicellular animals. There is still confusion pc2.ics.uci.edu/blast/jf/; [1,2]) and gene-expression data concerning basal animal phylogeny, although recent data are accumulating (for cnidarians, these are being indicate that Placozoa are not degenerate cnidarians and hence assembled into a web-accessible database [3]). (along with sponges) occupy a pivotal position. The transition in evolution from diploblast to bilaterian animals is becoming better Models and mechanisms understood, with gene expression data arguing that cnidarians Some organisational principles are shared by different have forerunners of the anteroposterior and dorsoventral body groups of multicellular organisms, such as division of axes, and even a putative homologue of mesoderm. The labour, differentiation and cell communication. Wolpert homeobox and kinase gene families have been further analysed and Szathmary [4] propose that development from a single in basal animals, although more data are required to enable cell is an additional feature shared by all those multicellular detailed comparison with Bilateria. organisms that evolved complexity and diversity. Their reasoning is that an egg creates a genetic bottleneck in Addresses development, removing the potential for conflict between Department of Zoology, University of Oxford, South Parks Road, somatic cells, and allowing coordination to evolve. Oxford OX1 3PS, UK Ãe-mail: [email protected] When we consider molecular mechanisms responsible for that coordination, however, there is no reason to Current Opinion in Genetics & Development 2003, 13:599–603 suspect there will be much similarity between different multicellular lineages. After all, many different mole- This review comes from a themed issue on Genomes and evolution cules are suitable for use as signals, many types of Edited by Evan Eichler and Nipam Patel proteins could mediate adhesion or signal reception, and there are many ways to control differential gene 0959-437X/$ – see front matter expression. It is particularly strange, therefore, that ß 2003 Elsevier Ltd. All rights reserved. Lalucque and Silar [5] report a gene apparently confined DOI 10.1016/j.gde.2003.09.002 to multicellular , and absent from unicellular species, regardless of phylogenetic position. The function of the gene, NADPH oxidase, is unclear, but it may play a Abbreviations dpp Decapentaplegic role in signalling metabolic status between cells. Regard- GFP green fluorescent protein less of its function, however, the phylogenetic distribution mtDNA mitochondrial DNA is puzzling. Independent origins of the gene seems impos- PTK protein tyrosine kinase sible, horizontal transfer unlikely, and loss from unicel- rDNA ribosomal DNA lular taxa illogical because this represents the primitive condition. Further taxonomic sampling is needed to Introduction resolve the picture. Multicellularity has arisen many times in eukaryotic evolution, with independent origins in the lineages lead- Search for the metazoan ancestor ing to animals (Metazoa), green plants, fungi, cellular Over one hundred years ago it was proposed that slime moulds and several other taxa. To find molecular sponges, and by implication all Metazoa, evolved from changes that accompanied the evolution of multicellular- a group of unicellular protists, the choanoflagellates ity, it is necessary to compare genes or genomes between (Figure 2). This hypothesis, made on the basis of a multicellular organisms and their closest unicellular rela- striking structural resemblance of choanoflagellates to tives. Identifying these relatives is therefore an important the choanocyte cells of sponges, has been revisited with task; recent advances in this direction are discussed molecular phylogenetics. Primary sequence analysis of below. In addition, if we want to determine characters 18S rDNA failed to resolve the exact phylogenetic pla- shared across a multicellular group, we must compare cement of choanoflagellates, but protein-coding genes widely divergent species within that group, including are proving more informative. Snell et al. [6] cloned the descendants of the basal branches. For the Metazoa, this cytoplasmic hsp70 gene from a freshwater choanoflagel- means including data from sponges, placozoans (Figure 1), late, Monosiga ovata, and concluded that choanoflagel- cnidarians and ctenophores. Unfortunately, there are no lates are indeed closer to animals than fungi. King and complete genome sequences for any of these taxa, Carroll [7] isolated four additional protein-coding genes www.current-opinion.com Current Opinion in Genetics & Development 2003, 13:599–603 600 Genomes and evolution

Figure 1 Figure 2

Trichoplax adhaerens (Phylum: Placozoa). Note the irregular shape, without defined polarity or axes of symmetry. Photograph kindly provided by Ana Sara Monteiro. from a marine choanoflagellate, M. brevicollis;theirana- lyses endorse the above conclusion, as does phylogenetic analysis of concatenated mitochondrial protein-coding genes by Lang et al. who sequenced the complete mito- chondrial genome of M. brevicollis [8].Burgeret al. [9] analysed gene content in this mtDNA genome, noting A (Choanoeca perplexa). The species shown here is the presence of long intergenic regions and many more different from those discussed in the main text, but all choanoflagellate protein-coding genes than animal mtDNA. These species share structural similarities to the feeding cells of sponges. In authors suggest that mtDNA compaction, characteristic particular, notice the long central flagellum, surround by a collar of tentacles used to entrap bacteria. Individual bacterial cells can be of Metazoa, occurred concurrently with the origin of seen adhering to the collar. Photograph kindly provided by Barry multicellularity. Comparison of choanoflagellate, sponge Leadbeater. and Trichoplax mtDNA is needed to test this hypothesis.

The choanoflagellates comprise part of the phylum Choanozoa, a diverse group also including some non- The position of the phylum Placozoa, represented by flagellateandamoeboidprotists[10]. is Trichoplax adhaerens, is particularly relevant to discussions a stellate, apparently nonflagellate marine protozoan, about metazoan origins, because of its morphological which has been placed on the basis of morphology simplicity (Figure 1). A recent paper by Cavalier-Smith alongside the filose amoebae Nuclearia sp. in the Class and Chao [11], based upon primary sequence of 18S Cristidiscoidea. 18S rDNA phylogenies from Cavalier- rDNA, places Trichoplax either as a sister group to Cni- Smith and Chao [11] strongly challenge this view and daria, or even within the Cnidaria close to Medusozoa place Ministeria as sister either to the choanoflagellates or (comprising Hydrozoa, Scyphozoa, and Cubozoa). These to animals. authors suggest that Trichoplax may have evolved from a scyphozoan (jellyfish) ancestor. This seems unlikely, how- Phylogeny of basal animals ever, in the light of two novel and informative molecular There is still much confusion concerning the early evolu- characters studied by Ender and Schierwater [14]. These tionary history of animals, and morphological characters researchers report that Trichoplax possesses a circular often contradict molecular data [12]. Aleshin and Petrov mtDNA molecule, plesiomorphic with all metazoans, [13] analysed the variable V7 and V2 regions of 18S rRNA including basal Cnidaria (Anthozoa), whereas medusozo- and coded secondary structure characters for cladistic ans possess linear mtDNA molecules. Their second mar- analysis. They suggest that Cnidaria are a sister group ker, the predicted secondary structure of 16S rRNA, shows to the Bilateria, that Ctenophora and Porifera (sponges) great disparity between Placozoa and Cnidaria with may be earlier derivatives of Metazoa, and that Placozoa respect to the number and length of stem and loop regions. have a close affiliation with Cnidaria. It seems unlikely that Trichoplax is a derived cnidarian.

Current Opinion in Genetics & Development 2003, 13:599–603 www.current-opinion.com The evolution of multicellularity and early animal genomes Brooke and Holland 601

Whereas Trichoplax remains a key animal for understand- conclusion is tentative until further comparative and func- ing the evolution of multicellularity, the same is not true tional analyses are conducted; indeed, Scholz and Tech- for Myxozoa. Once thought to be protists, and then nau [21] report a somewhat different expression for elevated to within the basal animals, a recent finding Brachyury in another cnidarian (an anthozoan or sea ane- supports the view that they are actually highly degenerate mone, Nematostella), including exclusion from the differ- bilaterians. Monteiro et al. [15] found that the 18S rDNA entiated mesenteric muscle tissue. The two studies are not sequence of one myxozoan is almost identical to that of a necessarily contradictory, however, because the Podocoryne rare and enigmatic worm, known as Buddenbrockia. Find- work included the medusa stage, while the Nematostella ing a myxozoan with muscles and a worm-like body (i.e. work studied the polyp stage only (medusae being absent Buddenbrockia) strongly indicates bilaterian affinity for in Anthozoa). The real possibility exists that mesoderm (or this phylum. a cell layer rather like it) is older than once thought.

It is clear that 18S rDNA primary sequence data con- A second topic that has come under repeated analysis is tinues to resolve some important nodes, but it has been body axis homology. The cnidarian and ctenophore body found wanting for some ‘deep’ phylogenetic issues [16]. plans are not precisely radial in many species or devel- More resolving power is coming from multiple protein- opmental stages, but do they have any axes equivalent to coding genes and structural genomic characters, some of bilaterian anteroposterior and dorsoventral axes? In Dro- which have been mentioned here [9,13,14].Other sophila and vertebrates, the decapentaplegic (dpp) system possible markers include derived gene fusions, such is involved in dorsoventral fate specification, with DPP as that used to root the eukaryotic tree [17],andintron (or BMP2/4) specifying dorsal in Drosophila and ventral in positions [18]. vertebrates. Hayward et al. report cloning of a member of this gene family from the coral Acropora [22]. This is a Germ layers and axes well-conserved orthologue, with high sequence similarity Most classic zoology textbooks divide the Metazoa (some- to vertebrate BMP2 and BMP4, and the ability to mimic times excluding sponges) into diploblasts and triploblasts, DPP activity when overexpressed in Drosophila. In coral following the century-old proposals of von Baer and development, the gene is expressed in the ectoderm close Lankester. The diploblasts, containing cnidarians and cte- to the blastopore. Most intriguingly, the mRNA is not nophores (comb jellies), have two germ layers (ectoderm uniform around the main axis but is localised in an arc and endoderm), whereas the supposedly more ‘advanced’ next to the blastopore, revealing cryptic bilateral sym- triploblasts have an additional third layer (mesoderm) metry in the embryo. This hints that bilateral symmetry sandwiched in between. Radiata and Bilateria are alter- may be older than previously thought, predating the native terms for the same groupings, referring to the view divergence of Cnidaria and Bilateria. More spatially that cnidarians and ctenophores are radially symmetrical, restricted genes must be examined to test this hypothesis, whereas triploblasts have a bilateral plane of symmetry, and to determine if the forerunner of the dorsoventral axis with distinct anteroposterior and dorsoventral axes. These runs along the oral–aboral axis or perpendicular to it. views, although well entrenched, have come under scru- tiny in recent years. Key questions are whether cnidarians Careful analysis of gene expression patterns in Cnidaria or ctenophores have cells homologous to mesoderm, (reviewed in [23]) and ctenophores [24] also suggests whether they are radial (some species and life cycle stages homology between the diploblast oral–aboral axis and the are clearly not), and which form is ancestral. bilaterian anteroposterior axis, although which end cor- responds to which is controversial. In a thought-provoking paper, Spring et al. [19] ask whether cnidarians have a tissue layer homologous to Homeobox genes mesoderm. Many cnidarians do have striated and smooth Although homeobox genes are found in all eukaryotes, muscle cells at the medusa stage of development (a bell- the PRD and ANTP classes have only been found in shaped body form). These cells develop from the animal genomes. Of these, the Hox and ParaHox genes ‘entocodon’ that separates from the ectodermal layer of are of particular interest because of their roles in embryo- the medusa, finally becoming clearly demarcated by an nic patterning of Bilateria. Similar genes have been extracellular matrix layer. To assess possible homology cloned from Cnidaria but not sponges (although sponges with the mesoderm of triploblasts, Spring et al. cloned have NK-related genes [25]). Sequence analysis and orthologues of Brachyury, Mef2 and the snail/slug gene linkage data from Southern blotting suggest the some family from the hydrozoan Podocoryne, while Mu¨ller et al. of the cnidarian genes are probably true Hox and ParaHox [20] cloned a putative MyoD homologue. In situ hybridisa- genes [26,27]. The best-studied is the Cnox-2 gene [1,28], tion revealed expression in the myogenic lineage of the the putative orthologue of a bilaterian ParaHox gene, Gsx. medusa (and a few other sites), supporting the hypothesis Cnox-2 is spatially expressed in the ectoderm of Hydra, that cnidarians and triploblasts have homologous devel- although it is intriguing that a Cnox-2–GFP construct, opmental pathways for building muscle systems. This introduced by electroporation into adult Hydra, also www.current-opinion.com Current Opinion in Genetics & Development 2003, 13:599–603 602 Genomes and evolution

drives reporter expression in endoderm [29]. This may supposedly simpler body layouts. Even here, some major imply that ancestral endodermal expression is secondarily questions remain unanswered, such as whether the suppressed in Hydra through a repressor element missing homeobox gene family diversified further in Metazoa from the construct. Discerning the original, or fundamen- to carry out new patterning roles. To answer these ques- tal, role of Cnox-2 is difficult, not least because the same tions, further comparative genomic and developmental gene is expressed in distinct patterns in different cnidar- biology work is vital. ians [28]. Schierwater et al. speculate that this may reflect complex evolution of the gene, whereby the same homeo- Update box sequence was combined with different protein mod- Very recently, King and colleagues [33] report three ules in different lineages [28]. further tyrosine kinase genes from the choanoflagellate Monosiga brevicollis, in addition to the one previously Signalling reported [7], plus C-type lectins and cadherins. One of Cell–cell interactions are crucial for the development of the newly reported tyrosine kinase genes encodes a multicellular animals. It might be expected, therefore, predicted transmembrane protein; a second resembles that genes encoding signals, receptors and downstream the Src family of non-receptor kinases from animals. effectors either arose or were elaborated during the evo- Pharmacological inhibition of Src or general kinase activ- lution of multicellularity. Of particular interest are the ity interferes with choanoflagellate proliferation. These receptor tyrosine kinases and other kinases involved in findings add further support to the view that signalling signal transduction. Suga et al. [30] previously cloned and adhesion between cells predates metazoan origins. multiple members of the protein tyrosine kinase (PTK) Moving into the animals, we have discussed the confusion ‘animal-specific’ gene family from the sponge Ephydatia concerning the original role of Cnox-2, the putative cni- fluviatilis. From these partial cDNAs, they proposed that darian homologue of Gsx. Recent analyses by Finnerty the PTK subfamily duplications arose before the diver- et al. [34] strongly suggest that Cnox-2 and Gsx are indeed gence of sponges from other animals. More recently, upon orthologous; these authors also describe developmental complete characterisation of these genes [31], they expression of the Cnox-2 orthologue in Nematostella (a sea showed that domain shufflings, which generated the anemone). Expression is localised to the posterior pole of distinct domain organisations characterising each PTK the moving planula larva, which develops into the oral subfamily, were also complete in early animal evolution. pole of the sea anemone. Even with these data, however, A diversity of serine/threonine kinases was also generated the variability of Cnox-2 expression between species early in animal evolution, as revealed by cloning of PKB means that comparison between cnidarian and bilaterian and PRK family genes in Hydra [32]. body axes remains tentative.

A paper by King and Carroll [7] provides the first evidence Acknowledgements of a receptor tyrosine kinase gene outside of the Metazoa. We thank Ana Sara Monteiro and Barry Leadbeater for the figures. The authors’ work on the origins of multicellularity is funded by the Their cloning of MBRTK1 from a choanoflagellate has Human Frontiers Science Program. significant implications, and demonstrates that at least one family of signal receptor molecule evolved before the References and recommended reading origin of animals. It will be fascinating to learn the Papers of particular interest, published within the annual period of diversity of this gene family in Choanozoa, and whether review, have been highlighted as: RTK diversification accompanied the evolution of multi-  of special interest  of outstanding interest cellularity. 1. Ball EE, Hayward DC, Reece-Hoyes JS, Hislop NR, Samuel G, Saint R, Harrison PL, Miller DJ: Coral development: from classical Conclusions embryology to molecular control. Int J Dev Biol 2002, Despite some attempts to find general principles under- 46:671-678. lying the evolution of multicellularity, more insight is 2. Yang Y, Cun S, Xie X, Lin J, Wei J, Yang W, Mou C, Yu C, Ye L, Lu Y et al.: EST analysis of gene expression in the likely to come if the different multicellular lineages are tentacle of Cyanea capillata. FEBS Lett 2003, 538:183-191. considered separately. For example, now that Choanozoa 3. Ryan JF, Finnerty JR: CnidBase: the cnidarian evolutionary are confirmed as the sister group to animals, these are database. Nucleic Acids Res 2003, 31:159-163. clearly key organisms for comparison to sponges, Tricho- 4. Wolpert L, Szathmary E: Evolution and the egg. Nature 2002, plax and Cnidaria. Such comparison should reveal the 420:745. molecular correlates of multicellularity in the animal 5. Lalucque H, Silar P: NADPH oxidase: an enzyme for lineage. A next step, and one for which progress has been multicellularity. Trends Microbiol 2003, 11:9-12. virtually non-existent, is to tackle the origin of symmetry, 6. Snell EA, Furlong RF, Holland PWH: Hsp70 sequences the diploblast germ layers and gastrulation — characters indicate that choanoflagellates are closely related to animals. Curr Biol 2001, 11:967-970. that are not automatic consequences of multicellularity. 7. King N, Carroll SB: A receptor tyrosine kinase from By contrast, progress has been made in understanding the choanoflagellates: molecular insights into early animal evolution of bilateral symmetry and the mesoderm, from evolution. Proc Natl Acad Sci USA 2001, 98:15032-15037.

Current Opinion in Genetics & Development 2003, 13:599–603 www.current-opinion.com The evolution of multicellularity and early animal genomes Brooke and Holland 603

8. Lang BF, O’Kelly C, Nerad T, Gray MW, Burger G: The closest 22. Hayward DC, Samuel G, Pontynen PC, Catmull J, Saint R, Miller DJ,  unicellular relatives of animals. Curr Biol 2002, 12:1773-1778.  Ball EE: Localized expression of a dpp/BMP2/4 ortholog in a This paper reports the first full mitochondrial DNA sequence from a coral embryo. Proc Natl Acad Sci USA 2002, 99:8106-8111. choanoflagellate. Phylogenetic analysis of concatenated protein seq- A gene from the dpp/Bmp2/4 family is expressed at one pole of coral uences confirms that choanoflagellates are close relatives of multicellular embryos, in an arc or crescent shape. Bmp2 and Bmp4 mark ventral in animals, consistent with the proposal made by Henry James-Clark in vertebrates, while their pro-orthologue dpp marks dorsal in Drosophila; 1866, and also with trees based on some nuclear genes. The paper also hence, this suggests that the cnidarian oral–aboral axis has some sim- suggests that the parasitic Ichthyosporea are a little more evolutionarily ilarity in patterning mechanism to the dorsoventral axis of bilaterians. The distant, contrary to earlier views. fact that the expression is arc shaped, rather than doughnut shaped, also hints at the presence of a cryptic bilateral axis. 9. Burger G, Forget L, Zhu Y, Gray MW, Lang BF: Unique mitochondrial genome architecture in unicellular relatives of 23. Meinhardt H: The radial-symmetric Hydra and the evolution of animals. Proc Natl Acad Sci USA 2003, 100:892-897. the bilateral body plan: an old body became a young brain. Bioessays 2002, 24:185-191. 10. Mendoza L, Taylor JW, Ajello L: The class Mesomycetozoea: a heterogeneous group of microorganisms at the animal-fungal 24. Yamada A, Martindale MQ: Expression of the ctenophore brain boundry. Annu Rev Microbiol 2002, 56:315-344.  factor 1 forkhead gene ortholog (ctenoBF-1) mRNA is restricted to the presumptive mouth and feeding apparatus: 11. Cavalier-Smith T, Chao EE: Phylogeny of Choanozoa, Apusozoa, implications for axial organisation in the Metazoa. Dev Genes and other protozoa and early megaevolution. Evol 2002, 212:338-348. J Mol Evol 2003, 56:540-563. The comb jellies, or ctenophores, have been sorely neglected by mole- cular developmental studies, even though they are relatively common 12. Martindale MQ, Finnerty JR, Henry JQ: The Radiata and the animals, long thought to be key to understanding the evolutionary transi- evolutionary origins of the bilaterian body plan. Mol Phylogenet tion from diploblast to bilaterian body plans. This paper reports the first Evol 2002, 24:358-365. full-length nuclear protein-coding gene cloned from a ctenophore (a 13. Aleshin VV, Petrov NB: Molecular evidence of regress in homologue of the forkhead family gene BF-1) and its expression in the evolution of Metazoa. Zh Obshch Biol 2002, 63:195-208. developing mouth and tentacular feeding apparatus. This paper marks the start of ctenophore molecular embryology. 14. Ender A, Schierwater B: Placozoa are not derived cnidarians:  evidence from molecular morphology. Mol Biol Evol 2003, 25. Coutinho CC, Fonseca RN, Mansure JJ, Borojevic R: Early steps in 20:130-134. the evolution of multicellularity: deep structural and functional Predicted secondary structure of rRNA, and presence of linear mtDNA, homologies among homeobox genes in sponges and higher implies that Trichoplax is neither a cnidarian larva nor a degenerate metazoans. Mech Dev 2003, 120:429-440. cnidarian. This overturns a long-standing misconception about the nature 26. Ferrier DEK, Holland PWH: Ancient origins of the Hox gene of Trichoplax, and resurrects this animal to a pivotal place in the history of cluster. Nat Rev Genet 2001, 2:33-38. our own multicellular lineage. 27. Finnerty JR: Cnidarians reveal intermediate stages in the 15. Monteiro A, Okamura B, Holland PWH: Orphan worm finds a evolution of Hox clusters and axial complexity. Am Zool 2001, home: Buddenbrockia is a myxozoan. Mol Biol Evol 2002, 41:608-620. 19:968-971. 28. Schierwater B, Dellaporta S, DeSalle R: Is the evolution of Cnox-2 16. Petrov NB, Aleshin VV: Conditionally neutral phylogenetic Hox/ParaHox genes ‘multicoloured’ and ‘polygenealogical’? markers of major taxa: a new aspect of the evolution of Mol Phylogenet Evol 2002, 24:374-378. macromolecules. Russ J Genet 2002, 38:877-894. 29. Mijkovic M, Mazet F, Galliot B: Cnidarian and bilaterians 17. Stechmann A, Cavalier-Smith T: Rooting the eukaryote tree by promoters can direct GFP expression in transfected Hydra. using a derived gene fusion. Science 2002, 297:89-91. Dev Biol 2002, 246:377-390. 18. Mu¨ ller WE, Bohm M, Grebenjuk VA, Skorokhod A, Mu¨ ller IM, 30. Suga H, Koyanagi M, Hoshiyama D, Ono K, Iwabe N, Kuma K, Gamulin V: Conservation of the positions of metazoan introns Miyata T: Extensive gene duplication in the early evolution of from sponges to humans. Gene 2002, 295:299-309. animals before the parazoan-eumetazoan split demonstrated 19. Spring J, Yanze N, Jo¨ sch C, Middel AM, Winninger B, Schmid V: by G proteins and protein tyrosine kinases from sponge and hydra. J Mol Evol 1999, 48:646-653.  Conservation of Brachyury, Mef2 and Snail in the myogenic lineage of jellyfish: a connection to the mesoderm of Bilateria. 31. Suga H, Kotoh K, Miyata T: Sponge homologs of vertebrate Dev Biol 2002, 244:372-384. protein tyrosine kinases and frequent domain shufflings in the A careful analysis of gene-expression patterns in the often-ignored early evolution of animals before the parazoan-eumetazoan medusa stage of cnidarian development reveals fascinating similarities split. Gene 2001, 280:195-201. to mesoderm. Brachyury, Mef2 and Snail/slug related genes are all expressed in the myogenic lineage of the medusa of a jellyfish-like 32. Herold M, Cikala M, MacWilliams H, David CN, Bo¨ ttger A: Cloning hydrozoan. and characterisation of PKB and PRK homologs from Hydra and the evolution of the protein kinase family. Dev Genes Evol 20. Mu¨ ller P, Seipel K, Yanze N, Reber-Mu¨ ller S, Streitwolf-Engel R, 2002, 212:513-519. Stierwald M, Spring J, Schmid V: Evolutionary aspects of developmentally regulated helix-loop-helix transcription 33. King K, Hittinger CT, Carroll SB: Evolution of key cell signalling factors in striated muscle of jellyfish. Dev Biol 2003, and adhesion protein families predates animal origins. 255:216-229. Science 2003, 301:361-363. 21. Scholz CB, Technau U: The ancestral role of Brachyury: 34. Finnerty JR, Paulson D, Burton P, Pang K, Martindale MQ: expression of NemBra1 in the basal cnidarian Nematostella Early evolution of a homeobox gene: the parahox Gsx in the vectensis (Anthozoa). Dev Genes Evol 2003, 212:563-570. Cnidaria and the Bilateria. Evol Dev 2003, 5:331-345.

www.current-opinion.com Current Opinion in Genetics & Development 2003, 13:599–603