animal evolution a view from the genome

8/7/2019 ANIMAL EVOLUTION a view from the genome

1/10

42 E N G I N E E R I N G & S C I E N C E N O . 3 / 4

Chordates (fish, amphibians,

reptiles, birds, mammals)

Deuterostomes

Hemichordates

(acorn worms)

Echinoderms

(sea urchins)

Lophotrochozoans

Phoronids (horseshoe worms)

Brachiopods (lampshells)

Ectoprocts (bryozoans)

Cycliophorans

Rotifers (wheel animals)

Gnathostomulids

Platyhelminths (flatworms)

Entoprocts

Nemerteans (ribbon worms)

Annelids (earthworms)

Echiurans (spoon worms)

Molluscs

(squids, bivalves,snails)

Sipunculans (peanut worms)

EcdysozoansChaetognaths (arrowworms)

Gastrotrichs

Cephalorhynchs (priapulids)

Nematodes (roundworms)

Nematomorphs (horsehair worms)

Tardigrades (water bears)

Onychophorans (velvet worms)

Arthropods

(crustaceans, insects,

centipedes, spiders)

Cnidarians (jellyfish, corals, sea anemones)

Poriferans (sponges)

Protostomes

Bilaterians


2/10

43E N G I N E E R I N G & S C I E N C E N O . 3 / 4

How did animals evolve, in the words ofDarwin, from so simple a beginning endlessforms most beautiful and most wonderful? Howdid the simple bodies of the first truly multicellu-lar animals lead to six-legged insects, five-armedstarfish, four-legged mammals, and legless snakes?Where did novelties such as insect wings and birdfeathers come from? Theres now a way, withouthaving to rely on the patchy and often nonexistentfossil record, to trace back the origins of thedifferent body plans and anatomical structuresthat give the animal kingdom its rich diversity.The evolutionary history of the animal kingdom isembedded in the genomes of the animals alivetoday, and can be studied in the laboratory.

This breakthrough has come from developmen-tal biologythe study of how embryos developand its created a huge upsurge of interest in theevolution of development. In fact, a completelynew field of bioscience (colloquially referred to asevo-devo), in which evolutionary biologists,developmental biologists, paleontologists, andphylogeneticists share their expertise, is takingshape. Things are moving forward rapidly, andfascinating new insights into animal evolution arebeing published almost on a weekly basis.

To investigate the genetic changes that led tothe evolution of different body plans, its veryimportant to have an accurate idea of the evolu-tionary relationships (phylogeny) of the animalsalive todayto know who is descended from

whom. Molecular phylogeny, such as comparingribosomal DNA, has recently clarified a lot of thedoubtful relationships. The family tree systemof animal classification, with single-celled animalsat the base and humans at the top, is out of favornowadays, because it implies that evolution has adirection, and its going our way. The branchingdiagram shown opposite, called a cladogrambecause it links clades (groups of animals whohave all descended from a common ancestor), is amuch more accurate way of showing relationships.

This current picture of the

evolutionary relationships

among multicellular

animalsi.e. the order in

which they branched from

common ancestorsis

based mainly on ribosomal

DNA analysis. The watery

backdrop is the gene

pool of the Beckman

Institute.

Pairs of clades related by a common ancestor arelinked by straight lines, so every branch in a clado-gram is a Y (or a tuning-fork shape, as here).Most animals are now known to belong to a hugeclade called the Bilateria, a name that refers totheir unifying feature of bilateral symmetry; thebody plan has a right-left axis, and a front-backaxis and a top-bottom one, too. All bilateriananimals evolved from the same ancestral animal.The next most closely related clade to the Bilateriaincludes jellyfish, sea anemones, and corals(collectively called the Cnidaria), while sponges(the Porifera) are more distantly related.

Bilaterians are a step up in multicellular com-plexity from jellyfish and sponges, which have justtwo tissue layersthe ectoderm and the endo-derm. The Bilateria have a third tissue layer, themesoderm, between the ecto- and endoderm. Andalthough cnidarians have some functional differ-ences between layers of cells, only bilaterians havethe complex 3-D arrays of cell types called organs.

The bilaterian lineage divided early on into twoclades: the deuterostomes, which gave rise to thevertebrates and their cousins, and the protostomesThe latter divided again into the ecdysozoans andthe lophotrochozoans. Insects, spiders, crustaceansnematodes (roundworms) and the like are ecdyso-zoans, a name that derives from the fact that theyall molt (ecdysis), while the lophotrochozoanclade embraces molluscs, earthworms, flatworms,and many lesser-known phyla (one of which, the

Cycliophora, has only one member, discovered afew years ago on the mouthparts of the Norwegianlobster). Nowadays, each of these three greatclades has a set of characteristics unique to thatclade, but manythose inherited from theoriginal bilateriansare common to all three.By comparing whats unique and whats sharedbetween descendants of common ancestors, it isnow possible to work out the genetic changes thathave given us the wonderful anatomical varietythat we see all around us today.


Animal Evolution: A View fromthe Genomeby Barbara El l i s


3/10


In development it is as if the wall, once erected

must then turn around and talk to the ceiling in

order to place the windows in the right positions

and the ceiling must use the joint with the wall to

decide where its wires will go.

The first bilaterians evolved in the remotePrecambrian, perhaps as much as 600 to 1,200million years ago, from an ancestor shared with thecnidarians. Precambrian fossils are extremely rare,but with the current upsurge of interest inevolution, palaeontologists are searching world-wide for more, and theyll doubtless find them.Already, the 590- to 550-million-year-oldDoushantuo deposits in southwest China haveyielded some microscopic animal fossils bearing a

striking resemblance to the embryos of modernbilaterians. It looks pretty certain now that themajor evolutionary diversification of the bilateriansinto the three primary clades also occurred in thePrecambrian. The Cambrian period (545490million years ago) has an abundance of fossils, instriking contrast to the Precambrian, and theyreveal a flamboyant blossoming of body plans andnovel structures. During this era, almost all themajor animal groups on earth today made theirappearance, although one prominent group, thevertebrates, didnt appear until the Silurian, 100million years later.

What made the bilaterians so much moresuccessful than their cnidarian relatives, enablingthem to spread out across the planet; adapt to lifein seawater, freshwater, land and air; and grow aslarge as dinosaurs and whales? Their diversity andcomplexity are the result of having a largercomplement of genes or gene families, a moresophisticated system of gene regulation, and, inparticular, an abstract patterning mechanism forbuilding body parts during development. In hisnew book, Genomic Regulatory Systems: Development

and Evolution,Eric Davidson, the Norman ChandlerProfessor of Cell Biology, whose pioneering workon regulatory gene analysis contributed greatly tothe current progress in understanding evolution,calls this abstract patterning mechanism thesecret of the bilaterians.

Development is a difficult task for a multicellu-lar animal. It starts life as a single cell, whichdivides over and over again as quickly as it caninto many, initially almost identical, cells, which

then differentiate intospecialized tissues and

organs and anatomicalstructures such as limbs and

wings. And every member of a species mustdevelop correctly in the same way, each and everytime. So why does one small, round cell develop

into a sea urchin and another very similar one intoa mammal? An obvious answer would be that thegenes are different. But theyre not: data from thegenome sequencing projects, which have nowprovided full sequences for a variety of animals,has confirmed what was already becoming appar-ent from other researchbilaterians all have thesame basic set of developmental genes. Some haveduplicate copies and some have lost a few duringevolution, but theres an astonishing commonality.Which presents an intriguing paradox: if the genes

are the same, how come there are so many differenttypes of animals?

At the molecular level, development involvesthe execution of a remarkable genetic programthat regulates the construction of an organism. Ofthe thousands of genes in the genome, most areused at some time during development, and theirdeployment must be controlled accurately in spaceand time. The answer to the paradox lies not inthe genes, but in the gene regulatory program, aprogram unique to each species. In some ways,writes Davidson in his book, this genetic programcan be likened to an architects blueprint for alarge and complex building. Different buildingsperhaps a railway station and a cathedralcan bemade from the same set of stones. Its the blue-print that dictates the different arrangement of thestones to make the different buildings. Similarly,different animals can be made from the same set ofgenes by following different blueprints. Butanimal blueprints also have to be interactive. Indevelopment it is as if the wall, once erected, mustthen turn around and talk to the ceiling in orderto place the windows in the right positions, hewrites, and the ceiling must use the joint withthe wall to decide where its wires will go.Development also means a progressive increase incomplexity; new populations of cells are generatedeach of which reads out a genetic subprogram.And all the time, these populations are beinginstructed to expand to a given extent, through

cell growth.Theres no master gene that coordinatesdevelopment, the way a site foreman wouldoversee the implementation of an architecturalblueprint. Each cell of the embryo has the samecomplete set of genes, derived from the fertilizedegg cell. What makes one cell different from itsneighbor depends on which genes are expressed, orturned on, to make proteins that have somefunction in the cell or transmit signals betweencells, and which genes are blocked. Instead of a

A remarkably well-preserved upper Lower

Cambrian fossil of a

segmented worm, about

525 million years old, from

the Chengjiang deposits in

South China.


4/10


single site foreman, the genes in each cell arecontrolled by short sequences of DNA called cis-regulatory elements. (Cis- means theyre part ofthe DNA, as opposed to molecules that are trans-,not part of the DNA.)

With impressive foresight, way back in 1969,Davidson and colleague Roy Britten, now Distin-guished Carnegie Senior Research Associate inBiology, Emeritus, proposed a theoretical modelfor such a system of genetic regulation. Theunderlying logic turned out to be more or lesscorrect, but it wasnt possible to know for certainfor another 30 years; only now are the tools ofmolecular biology (many of them developed by theDavidson group) good enough to detect such verysmall sequences of DNA and to analyze theirfunction. A gene is thousands of base pairs long (abase pair is one rung of the DNA ladder), butthe cis-regulatory elements have sequences of onlya few hundred base pairs.

Cis-regulatory elements are usually adjacent tothe gene they control but, just to make thingsmore interesting, they can sometimes be severalthousands of base pairs away along the chromo-some. Although a few genes are controlled by justone cis-regulatory element, most are regulated bymore than one, and some have a whole chain ofthem strung out along the chromosome. Theyreessentially devices that make choices, saysDavidson. Each cis-regulatory element has, onaverage, four to eight regulatory proteins, called

transcription factors, associated with it. Theseproteins bring information to the cis-regulatoryelement from the world outside the cell nucleus,from other genes within it, and from neighboringcells (see diagram, above). When these transcrip-tion factors arrive (by diffusion) at the cis-regulatoryelement, they dock onto their own particularlanding bay, a very short sequence of DNAspecific just to that transcription factor. Whetheror not a transcription factor docks depends on itsconcentration and sometimes on its activation by

other molecules, the cofactors. The cis-regulatoryelement reads the multiple inputs from thedifferent transcription factors that docktherecould be one telling it what cell type its going tobecome (muscle, nerve, or bone, for example),another to say where it is in relation to the othercells around it, yet another announcing that thecell is going to divideand based on all thisinformation the element produces a single output,an instruction that activates the gene or, as oftenas not, blocks it.

In essence, each cis-regulatory element functionslike a tiny but very powerful biological computingdevice. The information-processing function ofthe cis-regulatory element is the link between thethings that are happening in each cell and theresponse of the genes to them. And as the cis-regulatory elements are part of the DNA sequencetheyre hardwired into the genome, and anychanges in their sequence (such as by mutation,insertion or deletion of bases) are passed on tofuture generationssomething that is of greatsignificance in evolution.

As transcription factors are proteins, theyre alsoencoded by genes, and these genes in turn arecontrolled by cis-regulatory elements. Duringdevelopment, certain genes that encode transcrip-tion factors play a very important role; theyreknown as the regulatory genes, and they choreo-graph the highly successful abstract patterningsystem of bilaterian development.

The first bilaterians probably developed in away still seen in the embryos of many moderninvertebrate marine animals. When a fertilizedegg cell from such an animal starts to divide, thecells of the embryo get to know the cell typetheyre going to be as soon as theyre born, andtheir differentiation gene batteriessets ofgenes that are all expressed at the same time in acoordinated way so that the proteins they encodedefine the cell typeare turned on straight away.Theyre coordinated because their cis-regulatory

An example of how a cis-regulatory element works. Signals

A and B, which can be intra- or extracellular, activate

transcription factors A and B along a signalling pathway

On reaching the cis-regulatory element, they help to

initiate the synthesis of mRNA by RNA polymerase situated

at the beginning of the target gene. The mRNA is trans

lated into a functional protein (perhaps another transcription factor), which also provides feedback loops into the

system

In essence, each cis-

regulatory element

functions like a tiny

but very powerful

biological computing

device.

Courtesy US Department of Energy Genomes To Life program, DOEGenomesToLife.org


5/10


In the pattern-formation system of development, a simple snapshot taken

during developmental time in the animal will not resemble any parts of the

structure that will finally emerge. Until the final stages, itll look like

abstract patterns.

elements all respond to the same regulatory-genetranscription factor. This direct cell-type specifi-cation way of developing is an effective way ofproducing a free-swimming, self-feeding larvalstage as quickly as possible, but it can only workwhen a small number of cells are involved, andseriously limits embryonic size to the product ofabout 10 cell-division cycles, or a few thousand cells

If a more sophisticated system of developmentabstract patterning or, to give it its full name,pattern formation by stepwise regional specifica-tionhadnt evolved, bilaterians would never havegrown any bigger or more complex than theirjellyfish cousins. But as increasingly complex cis-regulatory control subcircuits were set up overtime, linking regulatory genes encoding spatialtranscription factors with those responsible forsignalling pathways and growth control, largerembryos with new body structures could develop.Astonishingly, the subcircuits set up in those earlydays, more than 550 million years ago, are stillused today by all bilaterians. Most types ofinvertebrate animals still use direct cell-typespecification to get to the free-swimming larvalstage, with pattern formation taking over afterthat to remodel the larva into an often verydifferent adult form (as in the sea urchin, lowerleft). Interestingly, some groups that evolved afterthe appearance of the first bilaterian groups, suchas the insects and the vertebrates, escaped thisbasic mechanism and devised their own ways ofturning eggs into embryos.

In the pattern-formation system of developmenta simple snapshot taken during developmentaltime in the animal will not resemble any parts ofthe structure that will finally emerge, writesDavidson. Until the final stages, itll look likeabstract patterns. Early on, basic elements of thebody plan such as the anterior-posterior axis andleft-right symmetry are established. Later pattern-formation events define the spatial organization ofthe main parts of the body planhead, tail, fore-legs, hindlegsthen even later pattern-formationevents define the detailed and smaller elements,such as the arrangement of the limb digits. Eachstage involves the partitioning off of one group of

Davidson learned much

about the ordering of

complex perceptions from

his father, leading American

abstract expressionist

painter Morris Davidson.

In the sea urchin, pattern

formation remodels the

larva into an adult form.

From top (not to scale):

embryo at the blastula

stage; 8-arm larva, a small

early rudiment of the

adult body growing at the

left-hand side of the

stomach; metamorphosing

larva with arm tissue

contracting and tube feet

of the 5-sided adult

emerging from the side;

one-week-old sea urchin

juvenile. C.Arenas-Menaetal.,Development,2000,127,4631-4653.

CompanyofBiologistsLtd.


6/10


cells into subgroups by the expression of regulatorygenes encoding transcription factors. There can bea whole cascade of transcription factors, sometimeslinked through signaling pathways, each of whichcontrols the activity of other regulatory genesfurther downstream at the next level of regulation,which could again be genes encoding otherspatially expressed transcription factors, and so on.

Eventually, the gene batteries that make thedifferentiated tissues and organs are switched on,and its only at this stage that an observer wouldstart to see recognizable body parts emerging fromthe abstract picture. Thats why its called abstractpatterning.

So the key players in the complexgenetic program for patternformation are the genes that encodethe regulatory transcription factorsand their cis-regulatory elements.Lets look at the best-known set ofthese, the hox cluster. In 1978, eightlinked hox genes involved in the development ofbody segments inDrosophila melanogaster, aka thefruit fly, were discovered by Ed Lewis (PhD 42;now Morgan Professor of Biology, Emeritus).Lewis had been patiently working away on

Drosophila in Kerckhoff Lab since 1939, andDavidson, who joined Caltech in 1970, and wasalready thinking about how development andevolution could be interlinked, feels he wasfortunate to have been in the same department atthat time. Ed was always upstairs, and he used tosay hey, come and look at this, he recalls. Iimmediately realized that Eds genes were some ofthe most interesting genes being worked on in thebiology division. Lewis was awarded the NobelPrize for his work in 1995 (see E&S, 1996, No. 1).

In the fruit fly, hox genes play an important rolein the development of body segments. The keyfeature of these genes is order. They are ordered inthe genome as two clusters in a long segment ofthe DNA on one of the chromosomes, and inspace, the genes are expressed in the same generalorder along the body as that in which they liealong the chromosome. The first and second geneof one cluster is expressed in the head segments,

then the third gene comes on a little fartherposterior in the thorax, and so on. Hox genes havebeen found in every animal type looked at, and arealways involved in anterior-posterior patterning ofthe body. We were all surprised at that,Davidson recalls. No one had expected to findthat humans had the same developmental genes asflies (left).

Even more surprising was finding that regula-tory genes have been so highly conserved through-

out evolution that theyre sometimeseven inter-changeable between animals. Some fly hox geneshave functioned well when transplanted into mice,and some mouse hox genes can replace those offlies. Thepax6 gene is particularly interesting.One of the important transcription factor-encod-ing regulatory genes,pax6 is involved in develop-ment of the vertebrate eye. Its fruit-fly equivalenteyeless (having these different gene names isconfusing, but scientists had no idea, when theyfound and named them in their own particular labanimals, that they were dealing with the samegenes) regulates development of compound insecteyes, with their numerous eyelets. Vertebrate andinsect eyes are very different in construction,building materials, and the way they work. Sowhat would happen if the eyeless gene of a fly wastransferred into a mouse embryo? A mouse withfly eyes? Nothe mouse develops a normal mouse

eye, even using a fly gene. If a humanpax6gene was transplanted into a spider, spidereyes would develop. No spiders wouldlook out from their webs with six big, bluehuman eyes, unfortunately (or perhapsfortunately). Moreover, ifeyeless orpax6

genes are made to function in a differentpart of an embryo, the cells there form an extra,

ectopic eyefrogs have grown extra frog eyes ontheir backs, flies have developed fly eyes on theirlegs (lower left).

The explanation for these unexpected results isthatpax6 is a regulatory gene active in a growingfield of undifferentiated cells near the top of theembryonic patterning cascade mentioned earlier.It encodes a transcription factor that sets up a trainof events leading to the formation and patterningof an entire structure (the eye), but it doesntcontrol the actual construction of the eye (such asthe lens, the cornea, the optical pigment), which isdone by batteries of genes further along in thedevelopment program. Mice always grow mouse

eyes because the flys eyeless gene activates themouses own eye-differentiation gene batteries,which go on to make all the parts for a mouse eye.To explain the ectopic eyes, think of the undiffer-entiated field of cells as a clean slate, prepared torespond to any of a number of regulatory genesthat start a differentiation program. Inducing

pax6 to run in undifferentiated back cells of a frogembryo, or leg cells of a fly embryo, activates theeye-differentiation gene batteries in those cells. Infact,pax6 works at the terminal differentiation

Above: The hoxgenes are

aligned in the same order

on the chromosomes of

fruit flies and humans.

This diagram indicates

roughly which body parts

are patterned by which

hoxgene (courtesy Ed

Lewis).

Right: Spider from

Corcovado National Park,

Costa Rica (courtesy L. E.

Gilbert, Integrative Biology,

University of Texas at

Austin).

Ectopic eye induced on the

leg of a fruit fly by forcing

the expression of a fly

homolog of the pax6gene

in the embryonic leg.

Czerny, T. et al., Mol. Cell3, 1999, 297-

307. Elsevier Science


7/10


stages of eye formation as well, because over time,regulatory genes can gain extra functions in newareas and at several different levels of the cascade,and this can lead to the creation of new structuresthat, when preserved by natural selection, contrib-ute to new animal forms.

How do pattern-formation systems reinventthemselves, and animal forms change, duringdevelopment? One of the most important ways isby cooption, which is when a regulatory genegains control of a new target site downstream, orcontrols the same apparatus but in a new area ofthe developing animal. The hypothetical regula-tory gene followed through three different stagesof evolution in the box on the opposite page showshow this could happen. Over time, the systembecomes more and more complicated as thedownstream effects of the gene affect more genebatteries; but all the while, the gene is still usedfor its ancestral functionto start development ofthe structure in the embryo. Cooption is ratherlike walking, writes Davidson. One linkage,upstream or downstream, stays where it was lastput and bears functional weight, while the othermoves; and then, if its move is useful, it may serveas the functional anchor while the first changes.After a few such steps, all the linkages surround-ing a given phase of activity of a regulatory genemay be different from the ancestral stage.

Normally, mutations to any of the genes activein the early stages of embryogenesis are just too

disruptive to be survivable, and the mutation dieswith the embryo. But if the cooptive change wassuch that the gene carried on doing its old job inaddition to the new one, a viable, but somewhatdifferent, animal could result, one that mightsurvive to adulthood and pass this cooption on toits offspring. Small genetic changes in cis-regulatory control happen continuously in allanimals. When theyre at a downstream level ofdevelopment fairly close to the final differentiationstages, they cause small differences between

animals of the same species, the sort that breederstake advantage of. However, ifcis-regulatorycontrol of particularly significant upstream regula-tory genes changed, there could be far moresignificant changes. A duplicated subset of thevertebrate hox cluster, for instance, was coopted topatterning limb development in vertebrates about350 million years ago, a serendipitous evolution-ary change that resulted in paired limbsandenabled vertebrates to swim, walk, run, and flytheir way all over the world.

If a cis-regulatory module controlling the timingof cell division gained a downstream gene control-ling commitmentthat moment in a cells lifewhen it stops developing and resigns itself tobeing an adult cell type foreverareas of the bodycould grow bigger or smaller. Lets imagine thatthis happened in the nose of a developing tapir,and that the program that determines the numberof cell divisions changed from six to 10 cycles.There would be a 60-fold increase in size, and ababy tapir would be born with a bigger nose. Thi

would be inherited by its offspring, so eventuallylots of long-nosed tapirs would be runningaround. And if there was an evolutionary advan-tage in having such an extended nose, or if itincreased the tapirs breeding success, a newspecies might eventually arise. Could this explainhow the elephant got its trunk? Its far toosimplistic a way of looking at speciation, of course

The eyes of squids, flatworms, flies (top row) and verte

brates (bottom row: trumpet fish, human, heron) use

different optical principles and visual pigments and are

constructed from different materials, but their develop

ment is always initiated by the same pax6gene. (Anima

photos courtesy BioMEDIA ASSOCIATES

www.ebiomedia.com.


8/10


9/10


coastal waters, and have few natural enemiesexcept the occasional fish, sea otters and fishermensupplying Japanese restaurants. For moleculardevelopmental biologists, the sea urchin embryohas many virtues, including transparency, incred-ible fecundity, high tolerance for micromanipula-tion, easy gene transfer, and a simple embryology.Best of all, it grows into a larva that swims, feeds,and looks after itself in a matter of days. Seaurchins are great for cis-regulatory analysis,

Davidson says. You do something with the eggsone day, and you get results the next. Theres noneed to wait until they grow up and have off-spring.

To map an entire regulatory system, theDavidson lab has developed a new set of technolo-gies for finding genes expressed in the endodermor mesoderm at different timestheyve foundhundreds, which theyre now sorting through tofind the ones most central to the process. Thenthey intend to analyze the way in which thesegenes are regulated; that is, how theyre connectedin the network through their cis-regulatoryregions. The cis-regulatory elements are notori-ously difficult to find; the very short stretches ofbases for each one are hidden in millions of basepairs of no apparent function, the so-called junkDNA, but thats another story. To locate theregulatory elements, theyre enlisting the help ofevolutionary conservation: when the nucleotidesequences around these genes are compared amongdifferent species of sea urchins whose common

Purple sea urchins at Kerckhoff Marine Lab during a

winter harvesting campaign for rare nucleoproteins such as

transcription factors, left, showing some of the 1,500 males

and 1,500 females being spawned. A gravid female deposit

about 10 million eggs (top left), so a total of 15 billion

eggs can be collected, which are poured into 4-liter beaker

(top right) and mixed with sperm from the males. Growin

the fertilized embryos for 24 hours to the 200-cell stage

provides 3 trillion nuclei from which workable amounts o

nucleoprotein can be extracted

ancestor is millions of years old, only regions thathave a use remain unchanged. All parts of theDNA sequence that dont bind proteins canchange over this long a time, so the short, un-changed segments will stand out in these compari-sons. And these identical little patches that arethe same between the different species have turnedout to be the cis-regulatory elementsa veryinteresting finding. Cis-regulatory analysis comesnextthis is really what the Davidson group is

best known for. A short fragment of DNAcontaining the cis-regulatory element is isolated,attached to another piece of DNA that encodes atraceable protein (usually colored or fluorescent),and injected back into the embryo. The cells ofthe developing embryo in which the proteinappears are the ones in which the gene regulatedby that cis-regulatory element is active. This way,the network connections can be checked. The finastage is to knock out genes to find the effect onall the other genes in the system. All the resultsare fed into an impressive computational model,the wiring diagram (shown opposite), thatsupdated every week on the lab Web page(www.its.caltech.edu/~mirsky/endomeso.htm) asthe results come in. Eventually, it will showwhere and when each gene is expressed throughoutthe various stages of endomesoderm development.Its going to be a lot of work, but once completeregulatory systems have been mapped for keymembers of the different animal clades, theirsimilarities and differences will reveal the preciserole of development in the evolutionary history ofanimals, something no one would have believedpossible just a few years ago.

Soon, biologists will be able to work back, likecomparative linguists who reconstruct extinctprotolanguages from languages still spoken today,to the last common ancestor of all the Bilateria.As Davidson writes: Although the ancestors ofmodern animals are extinct, the evidence of howthey worked is still swimming, walking, flyingaround outside, in the form of the DNA of themodern bilaterians. And when groups of animalsbecome extinct, what the planet is actually losingis their specific developmental-gene regulatory


10/10

animal evolution a view from the genome

Documents