H I G H L I G H T S

A twisted tale of BMPantagonists…

D E V E LO P M E N TA L B I O LO G Y

A paper featured in last month’sHighlights showed the way in whichgenetics could be used to resolve alongstanding puzzle in signal trans-duction. This month, genetics pro-vides the puzzle — and (along withsome biochemistry) the solution.This is because three papers inNature explain the counterintuitiveobservation that an inhibitor ofBMP (bone morphogenetic protein)signalling is required for a peak ofBMP signalling activity that occursduring development in both inverte-brates and vertebrates.

In Drosophila, the dorsoventralgradient of Decapentaplegic (Dpp)— a fly orthologue of vertebrateBMPs — is regulated by severalsecreted factors, among them Shortgastrulation (Sog) and Twisted gas-trulation (Tsg). Previous evidencehad indicated that both Tsg and Sogact as antagonists of Dpp. However,their mutant phenotypes point to a

paradox. A peak in Dpp activity isrequired to specify a dorsal embryon-ic structure called the amnioserosa.As Dpp loss-of-function mutants lackthis structure, one would expect it tobe expanded in the absence of a Dppantagonist. But surprisingly, Sog andTsg loss-of-function fly embryos lackthe amnioserosa completely.

A solution to this puzzle nowemerges in the form of an unusualmodel for a morphogen gradient, inwhich Tsg has a dual role. First, Tsgstabilizes the interaction between Sogand Dpp and, thanks to its own dif-fusability, helps to distribute themthroughout the embryo, shaping thegradients of both molecules in doingso. Second, it facilitates Sog’s cleavageby a protease called Tolloid (Tld).Near the source of Sog on theembryo’s ventral side, Tsg and Tldcannot keep up with the fresh supplyof Sog, so here most Dpp is bound inSog complexes. However, more dor-

sally, away from the Sog source, Tsg isno longer swamped by high Sog lev-els and, with the help of Tld, it con-tributes to Sog cleavage and to therelease of active Dpp. These twinfunctions of Tsg — transport andrelease of Dpp — are required for thedorsal peak of Dpp activity and forthe formation of the amnioserosa.

So what happens in vertebrates?It turns out that vertebrate TSGbinds directly to chordin (the verte-brate homologue of Sog) and to BMP4 (Dpp), and increaseschordin’s affinity for BMP4.Overexpression of tsg mRNA inzebrafish results in a block in BMP

Behavioural ethic-ette On 19 March 2001,TheNuffield Council on Bioethicsposted a public consultationdocument on its Web site toseek public opinion on theethical, legal and socialimplications of research intothe genetics of variation inhuman behaviour. Over1,000 copies of thisdocument have also beensent to interested individualsand UK organizations,including academics,voluntary organizations,religious groups andmembers of parliament.

The responses to thisdocument will inform, andprovide discussion materialfor, a panel of expertsassembled by the Council toconsider the implications andapplications of this area ofgenetics research. Some ofthe issues they will discussinclude: the ethics ofundertaking research into thegenetics of human behaviouron human participants; theimplications of using thisresearch to develop genetictests for certain behaviouralcharacteristics and their useby, for example, employersand insurance companies;and the way in which geneticinformation might influenceour perception of those withparticular behavioural traits.

The consultationdocument begins with a briefintroduction to behaviouralgenetics research — what itis and why it is studied —and then goes on to discusssome of the ethical andsocial issues that are raisedby this research. Throughoutthe document, key questionsare posed to provide aframework around whichrespondents can formulatetheir responses andopinions, some of which willbe collated by the panel intheir final policy document,which is due to be publishedin early 2002.

So what do you do tobecome involved? Visit theNuffield Council’s Web site,download the policydocument and return yourcomments to the Councilbefore 31 July 2001.

Jane Alfred

WEB WATCH

Unravelling the role of BMPantagonists in controlling and fine tuning embryogenesis is a hottopic in developmental biology at present (see accompanyingHighlight). And who’d havepredicted that getting to thebottom of a mouse with no trunkmight add further to the twists inthis field? For that’s what ElizabethLacy and colleagues might havedone by identifying a novel, BMP-antagonist-like, gene that whenmutated abolishes trunkdevelopment in mouse embryos.

The recessively lethal amnionless(amn) mutation, generated by atransgene insertion, disrupts themiddle region of the primitivestreak that gives rise to trunk

mesoderm. After many years ofwork on amn, Lacy and colleaguesmade their breakthrough whenthey found a BAC that rescued the amn phenotype. Their analysisshowed that this BAC spanned the insertion site and containedthree genes, but only one of themin its entirety. Lacy’s team founddeletions in this gene in amnmutants at the transgeneintegration site and confirmed its role by knocking it out toreproduce the amn mutantphenotype.

The amn gene encodes a noveltype I transmembrane protein, theextracellular domain of whichshows sequence similarity to thecysteine-rich (CR) domains of

BMP inhibitors such as Sog, Tsgand chordin (see accompanyingHighlight). These CR domainsmediate the activity of BMPantagonists by binding BMPs andsequestering them away from theirreceptors. In a comparison ofmouse amn to its fly and humanhomologues, the authors found itwas this amino-terminal, CR-containing region that was mosthighly conserved.

Expression studies showed thatamn is exclusively expressed in the extra-embryonic visceralendoderm (VE) tissue, butcuriously throughout the VE andnot just where it overlies the middleprimitive streak. Furthermore, amnis expressed on the apical surface ofthe VE, which faces away from theembryonic epiblast (see picture).This raises the question of howAmn mediates its effects withoutdirectly interacting with this

… and a new family member?

D E V E LO P M E N TA L B I O LO G Y

328 | MAY 2001 | VOLUME 2 www.nature.com/reviews/genetics

© 2001 Macmillan Magazines Ltd

H I G H L I G H T S

signalling, and, as expected, block-ing Tsg function results in an expan-sion of BMP signalling, also seen in chordino (chordin) mutants.Furthermore, ventral injections ofBMP antagonists in Xenopus resultin an induction of a secondary dor-sal axis, although Tsg achieves thisonly when co-injected with chordin.Tsg’s diffusability might be crucialto its function, because its mem-brane-tethered form alone caninduce an extra axis.

The connection between Tsg andTld that is seen in flies has not been so clearly established in vertebrates,but the cooperative action between

Tsg and Sog (chordin) seems highlyconserved. This twisting tale providesan interesting example of how a phe-notype can lead the unwary astray.

Magdalena Skipper

References and linksORIGINAL RESEARCH PAPERS Scott, I. C. et al. Homologues of Twisted gastrulation areextracellular cofactors in antagonism of BMPsignalling. Nature 410, 475–478 (2001) | Ross, J. J. et al. Twisted gastrulation is aconserved extracellular BMP antagonist. Nature 410, 479–483 (2001) | Chang, C. et al.Twisted gastrulation can function as a BMPantagonist. Nature 410, 483–487 (2001)FURTHER READING Harland, R. M. et al. A twist on embryonic signalling. Nature 410,423–424 (2001)WEB SITE Daniel Greenspan’s lab | LawrenceMarsh’s lab | Ali Brivanlou’s lab

IN BRIEF

Male-to-female sex reversal in mice lacking fibroblastgrowth factor 9. Colvin, J. S. et al. Cell 104, 875–889 (2001)

Fibroblast-growth-factor (FGF) signalling has been implicated in many aspects of development; and now Colvin et al. describe a new function for Fgf9 in testicular embryogenesis. Their analysis of Fgf9 knockout mice identified a variable gonadalphenotype, ranging from testicular hypoplasia to complete sexreversal. They found that Fgf9 acts early in the commitment tomale development and regulates many Sry-dependent processes,including Sertoli-cell differentiation and mesonephric-cellmigration. The authors suggest that FGF signalling could inducemale development in the absence of Sry — an important findingas many mammals lack Sry — and that FGF9 mutations mightcontribute to human sex reversal.

An alternative mode of translation permits production of a variant NBS1 protein from the common Nijmegenbreakage syndrome allele.Maser, R. S. et al. Nature Genet. 27, 417–421 (2001)

Around 90% of patients with Nijmegen breakage syndrome(NBS) are homozygous for a small deletion that prematurelytruncates the NBS1 transcript. As loss of Nbs1 is lethal in mice,the authors investigated whether this truncation allele encodes a hypomorph. Surprisingly, they found two proteins translatedfrom the truncated transcript — NBS1p27, the predictedtruncated protein, and the larger NBS1p70, which lacks the NBS1 amino-terminus. NBS1p70 is produced from an internaltranslation initiation site in NBS1 messenger RNA that allows anopen reading frame from the deletion transcript to be translated.Unlike NBS1p27, NBS1p70 still associates with the MRE11 DNA-repair complex, possibly diminishing the NBS phenotype.

DNA shuffling method for generating highly recombinedgenes and evolved genomes.Coco, W. M. et al. Nature Biotechnol. 19, 354–359 (2001)

DNA shuffling, or chimeragenesis, has been used to evolve gene families in vitro. These techniques create gene libraries in which original sequence polymorphisms are recombined at random to generate new sequence combinations. Moststrategies achieve approximately four crossover events per gene. Coco et al. now report a new approach called RACHITT(random chimeragenesis on transient template), in whichsingle-stranded DNA fragments are hybridized to a transientDNA template, generating 14 crossovers per gene and highrecombination frequencies, even between polymorphisms that are less than 10 bases apart in regions of high and lowhomology. Using RACHITT, the authors created new enzymecombinations for the desulphurization of fossil fuels.

T E C H N I Q U E S

H U M A N G E N E T I C S

D E V E LO P M E N TA L B I O LO G Y

NATURE REVIEWS | GENETICS VOLUME 2 | MAY 2001 | 329

embryonic tissue. The authorstackle this by proposing that Amnmodulates Bmp2 signalling withinthe VE itself, controlling targets thatmight themselves directly interactwith embryonic tissue.Amn mightalso act in concert with a Bmpreceptor, Alk2, which is present onthe VE’s apical surface, to modifythe activity of other extra-embryonically expressed Bmpmolecules, such as Bmp7. Futurefunctional studies should tell ifAmn is to join this hotly pursuedfamily of signalling antagonists.

Jane Alfred

References and linksORIGINAL RESEARCH PAPER Kalantry, S.et al. The amnionless gene, essential formouse gastrulation, encodes a visceral-endoderm-specific protein with anextracellular cysteine-rich domain. NatureGenet. 27, 411–416 (2001) FURTHER READING Dunn, N. R & Hogan, B. L. M. How does the mouse get its trunk?Nature Genet. 27, 351–352 (2001) WEB SITE Elizabeth Lacy’s lab

A wild-type, E7.5 mouse embryo stained with an anti-Amn antibody. Amn is present only onthe apical surface of the visceral endoderm.Courtesy of Sundeep Kalantry and KatiaManova, Memorial Sloan-Kettering CancerCenter, USA.

© 2001 Macmillan Magazines Ltd