The evolutionary origins andsignificance of vertebrateleft–right organisationJonathan Cooke

SummaryIn the last few years, an understanding has emerged ofthe developmental mechanism for the consistent internalleft–right structure, termed situs, that characterisesvertebrate anatomy. This involves largely vertebrate-conserved (i.e. ‘phylotypic’) gene expression cascadesthat encode ‘leftness’ and ‘rightness’ in appropriatetissues either side of the embryo’s midline soon aftergastrulation. Recent evidence indicates that the initial,directional symmetry breaking that initiates these cas-cades utilises mechanisms that are conserved or at leastclosely related in different vertebrate types. I describe ascenario whereby the capacity for directional modi-fication of an otherwise bilateral body plan can be viewedas an adaptive innovation rather closely connected withvertebrate origins, enabling optimal ‘design’ for veryactive lifestyles. But an alternative scenario, while re-taining the view that situs and indeed other vertebratefunctional lateralisations are deeply adaptive, pro-poses that they originated in the co-optation of left–rightdevelopmental information inherited from a very earlystage in metazoan diversification. It is proposed thata remote chordate ancestor lost its original or ‘ur-bilaterian’ symmetry to pass through an altogether non-symmetrical stage, and that the vertebrate dorsoventralmidline plane is not descended from that original one.I review the considerable evidence in favour of thisscenario, and discuss its wider implications for direc-tional asymmetries across the Metazoa. BioEssays26:413–421, 2004. � 2004 Wiley Periodicals, Inc.

Outline; levels of interest in study of

vertebrate left–right organisation

Most of us are aware that our outwardly bilaterally symmetrical

form hides a systematic left–right disposition or ‘packing’ of

our major visceral organs, coordinated among subsystems

suchasheart,major bloodvessels, gut andassociated glands.

The directionality of this asymmetry, known as situs, is clearly

conserved among our vertebrate relatives from a latest com-

mon ancestor, despite their varying degrees of elaboration of

visceral structure. Coordinated directionality as between the

major organ subsystems is clearly functionally important,

since instances of human developmental situs incoordination

(order of one per 103 births) show variably but often severely

reduced viability. But fascinatingly, the direction of the whole

asymmetry when coordinated appears to be of no functional

relevance. Much rarer (1 in 104) individuals have internal

anatomy that is the complete mirror-image of normal, yet their

incidence was not ascertainable until recently since most of

them live normal lives unaware of their singularity.

It would not have been a safe a priori assumption that

a macroscopically left–right reversed vertebrate anatomy

would function every bit as well as the normal. At the level of

cell structure, biology is intrinsically ‘handed’ i.e chiral. In

particular, various key protein assemblies such as cilia and

their basal bodies cannot be aligned so as to have equivalent

right and left sides, if they are orientated or anchored in the

cells of an embryo with respect to its dorsoventral or antero-

posterior axes. One or more aspects of bodily function might

thus have been compromised and inefficient unless gross-

anatomical and cell-structural ‘handedness’ were in the parti-

cular relation that had evolved as ‘normal’. While, in fact, this

relationship turns out not to matter at all for normal vertebrate

function, chiral molecular activity is somehow undoubtedly

central to normally directional symmetry breaking in the

embryo for gross anatomical pattern.(1) This provides a strong

explanation for the evolutionary, as well as within-species con-

servation of normal situs direction.

Vertebrate development involves the formation of various,

frequently paired, organ rudiments across considerable tissue

distances and over an appreciable time period. Subsequently,

many of these undergo left–right differential morphogenesis,

growing in coordination with each other to give the packing

arrangement seen.Amechanismmaximising reliability of situs

development might thus incorporate one initial ‘symmetry-

breaking’ signal at an early stage, triggering distinctive,

mutually exclusive left and right cascades of gene expression

that propagate through development. Then at any subsequent

time and anatomical location, tissue in at least one germ-layer

of the embryo would be labelled with its lateralised identity

before beginning its contribution to morphogenesis, thus

coordinating the directionality with which the individual organ

rudiments develop their left–right structure. This turns out to

BioEssays 26:413–421, � 2004 Wiley Periodicals, Inc. BioEssays 26.4 413

Museum of Zoology, The University of Cambridge, Cambridge, UK.

E-mail: ac@mole.bio.cam.ac.uk.

DOI 10.1002/bies.20015

Published online in Wiley InterScience (www.interscience.wiley.com).

Hypothesis

be indeed how situs is controlled developmentally. A variety

of pathological anatomies (heterotaxias—internal left–right

confusions, or isomerisms—partial or complete double-‘right’

or double-‘left’ structure) occur when the symmetry-breaking

signal has failed, or the right or left tissue-labelling cascade

has later failed at particular times or locations. Only much

more rarely does symmetry breaking fail or get subverted at

the earliest time, leading to entire, coordinate reversal, situs

inversus.

One might conclude that, viewed as a vertebrate evolu-

tionary invention (or a chordate one—see later), situs ‘could

have turnedout’ to showeither directionality, or indeed random

directionality, provided only that a mechanism ensured its

coherence within each individual. In human terms, the latter

could havemeant that left and right were nomore than apurely

abstract social convention—perhaps arising recently from the

need to communicate about human navigation, for instance.

It is a fascinating realisation that we owe to the robust, chiral-

molecule-driven mechanism of our developmental symmetry-

breaker, the fact that these termsmeanmuchmore than that to

our societies.(2)

Below, I outline the developmental arrangements whereby

right and left cascades of tissue-labelling gene expression are

stabilised in order to show that, in fact, these form a typical

example of an embryo axial patterning system. Then, I explore

a tension that often accompanies ‘explanation’ of biological

organisation. This tension is between, on one hand, the belief

that, mechanisms simulating optimal design can always arise

de novo in response to new selective pressures on function

and, on the other hand, the idea that historical contingency

constrains and shapes evolutionary opportunities. I outline a

purely adaptationist scenario, according to which situs could

have evolved, in a fundamentally symmetrical vertebrate-like

ancestor, as an optimal engineering response to demands of

a newly active, predatory lifestyle. However, I then review the

evidence indicating that, even while adaptive, the vertebrate

mode of encoding right and left identities within tissues

represents a co-optation of a deep ancestral feature of the

clade of animals to which we belong. This feature, not shared

with other lineages of bilaterian metazoans, was a morpholo-

gical transformation whereby our ancestors had become

fundamentally non-symmetrical as adults.

The molecular developmental biology

of vertebrate situs: loose ends

Many excellent, detailed reviews of the vertebrate left–right

molecular-geneticmechanismarenowavailable (for example,

Refs. 3–6), and the list of genes and signal pathways involved

continues to expand, some possibly acting in parallel rather

than in one cascade. There is nevertheless vertebrate con-

servation of particular ‘phylotypic’ gene expressions, certainly

for left and probably for right identity, centred in mesoderms

lateral to the gastrular midline and commencing some hours

into gastrulation. At this stage, axial midline structure is already

present to divide the embryo, and particular gene products

activated in this midline additionally serve to block inappropri-

ate cross-invasion of the ‘left-phylotypic’ intercellular signal,

the nodal protein. A further, earlier-acting set of right and left

intercellular signals, more locally expressed and near the

midline, have to date been studied only in bird (chick and quail)

embryos. These are not evident in either the yolky embryos of

zebrafish and frog, or the highly morphologically ‘derived’

mouse version of vertebrate development. In such embryos,

nodal seems to be the first easily recognisable lateralisedgene

expression,(7) whereas the earlier cascade in birds precedes

and seems to trigger the left nodal expression leading into the

phylotypic postgastrular sector.(8–10) The possibility remains,

however, that this difference of organisation tracks with

gastrular morphology rather than truly with phylogeny. Thus

the ‘blastoderm’—type rabbit embryo resembles chick more

than mouse anatomically, and perhaps in its left–right gene

expressions, with deployment of Sonic hedgehog (Shh) and

FGF-8 signal geneexpressions to left and right at its equivalent

of the avian Hensen’s node.(11)

Left-lateral expression of the homeobox-containing Pitx-2

transcription factor, once triggered by the nodal signalling

pathway, exhibits developmentally prolonged expression

consistent with a role in maintenance of ‘left’ character in

morphogenesis for a variety of organ primordia (viscera, lungs

etc.) as well as for left-specific cardiac wall structure. A

reciprocal right-sided role, but one more restricted in space

and time, may be played by the zinc finger-containing trans-

cription factor Snr (related to the Drosophila Snail gene;

sometimes known just as ‘vertebrate snail ’). This appears to

control the direction of heart looping and of embryo torsion, but

also to participate in the transmission of ‘rightness’ by initially

confining ‘leftness’; i.e.Pitx-2 expression, to its proper side.(12)

SnR itself is repressed by nodal signal. Indeed, the nexus of

positive feed-forward and negative feed-back regulatory

relationships among the ‘left–right’ gene expressions, ensur-

ing their appropriate spatial exclusivity and temporal ordering,

is strongly reminiscent of other primary pattern-forming de-

velopmental gene cascades that are understood, e.g. the

Drosophila gap-to-segmentation or the vertebrate gastrular

dorsoventral systems. A forthcoming review(13) deals with

these aspects among others.

It has long been realised that, if the early embryo is

genuinely bilaterally symmetrical in cellular anatomy, then

reliably directional breaking of that initial symmetry can only

derive, ultimately, from the orientated chiral structure of one

or more macromolecular assemblies within its cells.(14–16) If

somehow tethered and alignedwith one ormore other embryo

axes (anteroposterior, or apicobasal within an epithelium),

such molecular assemblies could initiate handed asymmetry

across a midline, for instance by a net left–right component of

intracellular or intercellular transport of an initiating signal

Hypothesis

414 BioEssays 26.4

molecule. Cilia and their basal bodies had attracted specula-

tion because of their obviously chiral or ‘handed’ structure and

mechanical activity (a spiral component to the beat), and their

defined mode of tethering in cells. It has now been reported

that the normal beating of a densely distributed group of

atypically structured cilia (‘protocilia’) produces a flow to the

mouse embryo’s left within the medium bathing the archen-

teric epithelial surface, and that this flow is necessary and

determinative for normal directionality of the lateralised nodal

and Pitx-2 expressions that stabilise situs development.(17,18)

The authors propose that, at least in mouse, developmental

symmetry is broken by ciliary driven leftward extracellular

flow of a ‘morphogen’, possibly nodal protein itself or a related

protein, whereby phylotypic left nodal gene expression is

activated in the gastrula.

This discovery has seemed to rationalise a diversity of data

from knockout mice, in which absence of gene functions

associated with the building or activity of cilia correlates with

heterotaxia or with random ‘flipping’ to situs inversus, as well

as a human familial syndrome inwhichmalfunction of all cilia in

the body correlates with random situs direction.(14) Ciliary

structures that are clear homologues in molecular terms

with the mouse nodal protocilia are observable in several

vertebrate embryo types, at times just preceding their first

‘phylotypic’ left nodal expressions.(19) Nevertheless, as

recently reviewed,(1) extensive data now suggest that these

observations may not have defined the truly initiating, con-

served vertebrate symmetry-breaking mechanism. Left–right

directional net transport processes get underway within the

joint intracellular space, considerably before gastrulation

begins. The data hint at a plurality of other (pleiotropic) roles

for gene products involved in building cilia/basal bodies, and

such roles also provide opportunity for the derivation of left–

right directionality from the chiral structure of an anchored

molecule through intracellular transport processes. In parti-

cular, the community awaits elucidation of the function of the

intracellular protein Inversin, the remarkable inv gene product

whose malfunction reverses, rather than randomising situs,

and which is now reported to be associated with protocilia.(20)

inv mutant mouse embryos are reported to show detectably

abnormal, but clearly not reversed nodal ciliary flow!(21)

Evolutionary origins of vertebrate left–right;

optimal design versus historical constraint

and co-optation

Most ‘major body plan’ features are less transparently

adaptive in origin than are, say, the design features of wings

or skin, though such lack of present transparency means little.

But among body plans, the vertebrate one, including situs,

perhaps has some of the best adaptive credentials. Onset of

an active searching lifestyle, probably accompanied by pro-

gressive size increase, probably marked vertebrate origins.

In addition to the extra demands for metabolic rate accom-

panying such a lifestyle change, the ‘square-cube’ law dictates

that, for bodies of identical form, totalmetabolic demandwould

increase more rapidly with linear size than would available

internal area for absorptive and exchange processes. There

would thus have been a selective premium on relative in-

crease in the length and complexity of formerly simple tubular

viscera and circulatory vessels, within a symmetrically stream-

lined outer ‘locomotory’ body wall. For processes like nutritive

absorption and circulatory metabolite exchange, the relation-

ship is intuitively clear, but recent analysis of vertebrate heart

function, for instance, has demonstrated strikingly how, for a

given mass of muscle, the spirally coiled form has much

greater pumping efficacy than would a simple valved tube.(22)

Development guided by body-wide cascades of left- and right-

specific gene expression has obvious advantages, ensuring

that near-identical packing arrangements among the complex

viscera can occur in every ‘normal’ individual, rather than a

random series of mostly suboptimal ones. Complexity in the

organ systems can then further co-evolve. Such intimate,

cooperative fine tuning between separate organ systems is

manifest when the first symptom of acute human heart

malfunction resembles indigestion—or when acute indiges-

tion forces arrhythmias in the otherwise healthy heart!

Incontestably, in view of molecular phylogenetic data, ver-

tebrates belong to the great animal clade known as Bilateria.

If the assumption is made that the proto-vertebrate had

descended from its ur-bilaterian ancestors with the original

dorsoventral plane of bilateral symmetry intact, directional

left–right organisation could then only occur by de novo

‘invention’ of a symmetry-breaking mechanism. One form of

this scenario, namely that the whole situs mechanism

including the derivation of symmetry breaking from molecular

chirality was a vertebrate adaptive innovation, has often

been assumed by molecularly trained workers who lack

acquaintance with the fields of comparative anatomy and

phylogenetics.

Recent comparisons of developmental expression pat-

terns for nodal and Pitx orthologues in the protochordates

Amphioxus (Cephalochordata) andCiona (Urochordata) have

already forced a revision of this view. There are clear and

striking left-lateral expressions of these genes during larval

stages most equivalent to the vertebrate neurula/pharyngula.

At the least, the asymmetry innovationmust haveoccurred in a

more remote chordate-like ancestor rather than just a proto-

vertebrate (reviewed in 23). But even this scenario runs into

trouble when a wider range of data is considered. The striking

developmental asymmetries of anatomy in the head and gill

region of Amphioxus (reviewed in 24), and of gene expression

even without associated anatomical features (also in Am-

phioxus but especially in the urochordate tadpole larva(23)),

themselves seem functionally gratuitous. Nothing currently

indicates that these forms had ancestors that led active

vertebrate-like lives, with the functional demands for organ

Hypothesis

BioEssays 26.4 415

packing that the adaptationist scenario for the origin of situs-

like asymmetry requires (though such arguments are inher-

ently insecure). Amphioxus is small and inactive enough to

need no functional heart.

Amphioxus shares with adult urochordates a simple direc-

tional looping and asymmetrical structure of the viscera that

could plausibly be homologous with that in extant echino-

derms. The simple gut looping in extant echinoderm classes

bears a conserved directionality in relation to the oral–aboral

axis, within their pentaradially symmetrical adult body plan.

Echinoderms, hemichordates (acorn worms) and the above-

mentioned ‘protochordate’ types along with vertebrates are

now accepted on molecular phylogenetic grounds as repre-

senting a clade, the Deuterostomia.(25) Comparative anato-

mists recognise a set of structural features that links these

diverse forms (radial cleavage, lack of correspondence be-

tween mouth and blastopore at gastrulation, larval forms,

notochord, pharyngeal gill slits), even though each group only

manifests a subset of them. As bilaterians, deuterostomes

share an ancestry with all those other animal groups in which

the Hox/Hom cluster of homeobox genes is deployed as the

basis for a longitudinal (anteroposterior) axial organisation.

But earlywithin their ownphylogeny there occurred, either as a

single founding transformation or as a repeating tendency to

such transformation, a radical departure from ur-bilaterian

organisation in adult forms. In extant echinoderms, a sub-

stantial trace of properly bilaterian, co-linear deployment for

Hox orthologues occurs along the bilateral coelomic pouches

of the larva.(26) But in the internal reorganisation of metamor-

phosis, the adult ‘oral–aboral’ axis arises and is not aligned

with either the larval dorsoventral or longitudinal ones. The

remaining deuterostome forms show, to varying extents, the

appearance of bilaterality in parts of their bodies coupled

with (gratuitous-seeming or functional) left–right structure in

others.

We must then ask, does being bilaterian by descent

necessitate that left–right organisation, when it occurs within

deuterostome bodies, directly corresponds to ‘ur-bilaterian’

left and right? The ancestral ur-bilaterian organism presum-

ably had a dorsoventral cross-sectional organisation, which

with its anteroposterior axis defined right and left sides. But

this would not have provided a molecular basis for encoding

its mirror-image sides with ‘leftness’ and ‘rightness’ as such.

An alternative idea is that, during bilaterian diversification, a

transformation occurredwithin the deuterostome lineage.One

original ‘ur-bilaterian’ side progressively became a new

‘ventral’ side and the other, the new ‘dorsal’ side, in essence,

a 908 rotation.(24,27,28) This would most probably have oc-

curred to give rise to a less axially organised, slow-moving

or sessile, detritus- or filter-feeding adult. But, in reponse to

evolving lifestyle, descendants of such an organism might

reacquire to varying extents a symmetrical bilateral outer

structure. They might nevertheless retain the developmental

organisation of what had once been a dorsoventral dimension,

now co-opted as the capacity to encode genetically regiona-

lised and thus independently evolvable right and left internal

structure. At this point, it should be mentioned that there

are plausible alternative explanations for the appearance of a

literal 1808dorsoventral inversion having occurred as betweenprotostome and deuterostome (arthropod versus vertebrate)

body plans.(29–31)

The above type of 908 transformation was originally pro-

posed to have occurred after the origin of hemichordates,

making the latter the only deuterostomes retaining ‘ur-

bilaterality’. But molecular phylogeny now suggests that they

and the (as adult) radially non-bilateral echinoderms are sister

groups, and together form a sister group to all other, more

chordate-like forms.(25) Thus the transition might be a deuter-

ostome-defining character, with hemichordates the group that

has most profoundly reacquired symmetry. The adults of just

some hemichordate species retain a feature shared with

echinodermmetamorphosing larvae, which pre-figure the adult

axial transformation in that only the left protocoel (anterior

coelomic cavity) is linkedwith the outside via a canal. If, in fact,

hemichordates have obliterated a real original transformation

very completely from their life history, they may turn out not to

show any lateralised gene expressions orthologous to the

‘vertebrate-phylotypic’ ones of situs.

The settling of, for instance, a crinoid echinoderm larvamay

be recapitulating an ancestral 908 transformation when its

right-hand coelomic cavities degenerate on turning to face

the substratum, while the left ones collaborate with the left

ectoderm to surround the gut and generate the radially

symmetrical adult system.(32) Definitive information on the

gene expression signals that orientate formation of the adult

oral–aboral axis within the bilateral echinoderm larva will be of

great interest in relation to the vertebrate situs cascade. The

direction of torsion at larval settling, while interesting, is less

germane to the issue, because among extant echinoderm

groups, either the oral or aboral pole can be applied to the

substratum dependent upon the adult form and lifestyle.(33)

Amphioxus may represent a form which, while having

reacquired a largely fusiform adult body, still recapitulates

dramatically some steps in its morphological derivation from

thenon-symmetrical ancestor. Theascidian ‘tadpole’maybea

dispersal form, but is not larval in the sameprimordial sense as

the ciliated echinodermpluteus. It thus shows the anatomically

gratuitous left expressions of nodal and pitx as remnants of its

axially transformed origin. Vertebrates, according to this view,

havevery largely redevelopedananatomical symmetry in their

locomotory/ neural structure in adaptation to lifestyle. But, in

their visceral structure, as a heritage from the deep past, they

have retained the left–right (one-time dorsoventral) devel-

opmental organisation that is now positively co-opted for

the adaptive arrangements of situs. Fig. 1 outlines, in a way

divorced from any particular interpretation of fossil forms,

Hypothesis

416 BioEssays 26.4

such derivation of contemporary left–right structure via an

axial transformation from an ‘ur-bilaterian’ ancestor.

What are we to make of the contemporary vertebrate

mechanism of developmental symmetry breaking within an

ostensibly symmetrical embryo? It could be a de novo

invention, utilising molecular chirality reliably to directionalise

activation of the one-time dorsoventral, now to be left–right

dimension of organisation. This might be necessary because

the massive outer-resymmetrisation of the vertebrate body

has reached backwards into development, dominating gas-

trular structure and threatening to remove original cues for the

left–right axis. Alternatively, it may be more directly derived

from a directional settling mechanism whereby the still ur-

bilateral, ciliated larva of ancestral deuterostomes underwent

its 908 axial re-organisation to give the benthic adult. Intere-

stingly, the locomotion of some contemporary deuterostome

larval forms is spiral because of ‘off-axis’ ciliary beat, rather

than solely due to their hydrodynamic shape.(34)

Additionally, there is now striking evidence from a uro-

chordate(35) and a vertebrate(36) for a left–right differential

distribution of gene products, immediately postfertilisation,

that could be functionally relevant to polarisation of the phylo-

typic left–right gene expressions in the otherwise symmetrical

embryo. Such direct asymmetrical placement of proteinsmust

reflect a utilisation of molecular chirality in the mechanical re-

organisations that follow fertilisation in these forms. It is of

great interest that urochordates and the vertebrates each

exhibit modes of development that might be considered to

have by-passed entirely the ‘ur-bilateral’ truly larval form

primordial to the group.

The transient but clear, directional non-symmetry in bird

gastrular anatomymay be yet a further clue to the evolutionary

Figure 1. Schematic of a 908 transition within vertebrate ancestry. Three body forms are shown, each in dorsal plan and in composite

transverse sectional views (representing axial levels opposite dashed lines). The original bilaterian ancestor (left) has paired special sense

organs (black), filter- or detritus-feeding apparatus anteriorly, paired and probably segmented mesoderm structures (stippled), and a

tubulargutwithout left–right but possiblywithdorsoventral complexity.Shortlyafter theproposedmorphological transition (centre), thecase

is shown in which an original right side becomes ‘ventral’, and substratum-applied, with concomitant invasion of ventral surface

specialisations and loss of some subset of the original right members of paired structures. The result is a non-symmetrical body plan.

On the right is shown an arbitrary stage in the secondary re-acquisition of bilaterality in the locomotory and neural (sensory) structures, by

organ-pairing in the head, but perhaps by progressive re-rotation of the locomotory tail structure and its forward invasion along the dorsal

body wall, nowwith a notochord (see Ref. 24). The result is essential symmetry of the outer body wall, together with a developmental basis

for left–right structure centred in the viscera. This scenario takes account of evidence from comparative developmental anatomy of extant

forms, comparative gene expression and molecular taxonomy, but does not follow any detailed phylogeny proposed from fossil evidence.

Hypothesis

BioEssays 26.4 417

depthofvertebrate left–right structure.Noticedat least75years

ago and mentioned at intervals since, this has recently

received renewed attention.(37,38) As shown schematically in

Fig. 2, during several hours of development, the region fated

to give rise to the axial structure of the anterior body does not

give the appearance of a bilateral animal’s embryo. At these

stages, the notochord—traditionally mid-dorsal mesodermal—

is being derived entirely from deep tissue at the right of the

node, while the floorplate—traditionally mid-dorsal in the over-

lying neural plate—originates entirely from the more super-

ficial layer at the node’s left. This structure becomes apparent

well before ‘phylotypic’ lateralised expressions of nodal, Pitx

and Snail-related genes are in place, but late enough to be

developmentally downstream of early left–right ion transport

processes as discussed by Levin.(1) It is reversed in the right-

hand member when parallel twin axes form within one blasto-

derm, in keeping with the classical observation that situs itself

is often disturbed or reversed in thismember of experimentally

produced twin patterns.(39) In single embryos, the node

asymmetry is directionally robust, however, even though the

later asymmetries and gene expressions of situs can be con-

fused or reversed by subsequent experimental treatments.(37)

Due to lack of observations during the restricted relevant

stages, we cannot currently say how many other ‘blastoderm-

type’ vertebrate embryos, e.g. reptiles or even some mam-

mals, show such transient handed asymmetry.

The possibility of an ur-bilaterian dorsal–ventral to left–

right axial transformation deep within vertebrate ancestry has

not found wide favour, mainly because the fossil evidence

on which it was originally proposed seems impossibly fragile

and subjective to cellular and molecular developmental-

ists.(24,27,28,40) The transformation itself has been presented

as if it had occurred within one ontogeny to create a ‘hopeful

monster’ ancestor. While such a scenario is not unthinkable in

a remote earlymetazoanworld, themajority of current thinkers

about even macro-evolutionary change are of gradualist

persuasions. But any such transformation relevant to chordate

origins must have occurred back in the explosive Cambrian

phase of metazoan diversification, where fossil traces will

remain problematic. Its palaeontological originators should

be given full credit for the idea of a 908 transformation, but

evolutionary developmentalists at large should not be dis-

tracted from assessing the other independent evidence for

its occurrence, outlined above. The chordate, or extremely

chordate-like, vetulicolians (lower Cambrian, China) are a

plausible source of vertebrate ancestry,(41) but it is currently

unclear whether they display situs-like directional asymmetry.

908 axial transformations could even have occurred sepa-

rately in the ancestry of echinoderms (including ‘calcichor-

dates’) and of chordates, rather than in a common ancestor.

Summary and conclusions:

a differential evolutionary constraint?

Situs, the directional left–right structure within the vertebrate

body, is most obvious in the packing of the viscera but is in fact

more pervasive, (e.g. embryo torsion, and also form of the

human cerebral hemispheres(42)). It develops through initia-

tion of a left–right difference, across the midline of the early

Figure 2. The asymmetrical anatomy of the

avian (chick) gastrula. At left, the whole embryo-

forming area (area pellucida) of the blastoderm is

shown in dorsal view, anterior at top. Dashed box

indicates the central region, embracing the centre

of gastrulation activity (Hensen’s node) that is cur-

rently giving rise to anterior parts of the postgastr-

ular midline (head process). At right, detail of this

boxed region is shown, together with its transverse

section (below). The axial level of section is

indicated by the horizontal dashed line, and the

observer of the section faces anteriorly. Arrows

indicate indentations in the neurectodermal sur-

face at the edges of the thickened node. The left

indentation is much deeper, and continuous both

with the central pit of the node via a gap in the node

wall, and with the gutter-like groove that passes

anteriorly to swing into the midline as the future

floorplate of the nervous system (shown cross-

hatched). The right-hand indentation flanks the

deep-lying mass of the nascent notochord, which

is positioned eccentrically to the right. The emer-

ging notochord (shown stippled) swings progres-

sively into place in the midline beneath the

floorplate as the node regresses from each level.

Hypothesis

418 BioEssays 26.4

embryo, in a signal that triggers mutually exclusive, propagat-

ing cascades of ‘left’ and ‘right’ gene expressions within

relevant tissue layers as formation of the body plan proceeds.

Left and right tissue characters are subsequently expressed

throughdistinctive contributions to growth andmorphogenesis

in asymmetrical organ systems. Understanding is currently

most advanced for the middle, ‘propagative’ sector, which is

essentially vertebrate-conserved, although the list of gene

interactions involved in the cascades continues to grow and

some of these may act independently in parallel. Our relative

ignorance concerning the ‘downstream execution’ sector

only tracks with current ignorance regarding ‘downstream

executive’ form-shaping mechanisms in organogenesis

generally (though see the Theoretical Morphology article in

this issue on pages 405–412(43)).

Understanding of the ‘initial symmetry-breaking’ sector is

currently at a particularly interesting andcontroversial stage.(1)

Evidence has been accumulating in several vertebrates for

the basic role, at stages preceding gastrulation, of vectorial

transport processes within the junction-linked intracellular

space, leading to ion-channel-driven electrochemical gradi-

ents across the embryo midline. In at least one vertebrate and

one protochordate (related outgroup) form, molecular compo-

nents of this intracellular symmetry ‘breaking’ system achieve

left–right differential distribution during the very first cleavage

divisions when the fate map for the axial body plan is being

orientated within the egg material, suggesting at least that

chordate left–right structure is evolutionarily deep. Together

with a wealth of less direct evidence from comparative devel-

opmental anatomy, comparative expression of gene ortholo-

gues and molecular phylogeny, this indicates that biologists

should take seriously the central (and non-original, for

example, Ref. 28) proposal of this article. This is that while

adaptive, in enabling optimal design for new lifestyle, verte-

brate directional left–right structure was not a latter-day

innovation. Rather, it was a preadaptation stemming from a

remote ancestral dorsal–ventral to right–left axial transforma-

tion, around which a newer, secondary bilaterality of the outer

locomotory/neural body has been built.

In addition to the structure of situs, complementarities of

function between anatomically equivalent regions in right and

left halves of the brain may also be uniquely advantageous,

and distinctive to vertebrates. These could enable an increase

in the diversity of parallel subprocesses that feed into the final

behavioural outcome of advanced cognition. The devel-

opmental information enabling such neural lateralisations

appears in some cases to be independent of that underlying

situs, but could also stem ultimately from vertebrates’ unique

deep history. Developmental relationships between neural

functional ‘handednesses’ and situs are reviewed in depth

elsewhere.(13,44,45)

An implication of this ‘axial transformation’ proposal is that

left–right encoding, in vertebrate development, may be of a

kind not found elsewhere in the animal kingdom. Thus neither

vertebrate-type left–right neural-functional complementarity

nor profound left–right differential anatomy may be attainable

among otherwise similarly complex protostome animals.

Being untransformedly bilaterian, these animals lack the

developmental basis for genetic tissue labelling of ‘leftness’

and ‘rightness’ as such, and with it the possibility for highly

controlled left–right differential expression of the geno-

me. A non-trivial example of differential evolutionary con-

straint,(see e.g. 46) deriving from historical contingency, may

thus have been in operation since the Cambrian metazoan

diversification. If conjectures regarding the adaptiveness of

left–right structure for vertebrate rates of energy expenditure

and lateralisations of higher cognitive function have any sub-

stance, that constraint has been quite consequential.

Certainly, in development of gastropod molluscs or nema-

todes, right–left differential organisation is propagated struc-

turally via a cleavage pattern of defined ‘handedness’. It is

probable that chiral information from a protein assembly

ultimately underlies such cases, and acts via relative cell posi-

tioning, as left–right adult structure can be reversed experi-

mentally by micromanipulation of the latter in the embryo.(47)

But no such forms have confronted, as successfully as have

vertebrates, the problemsposedbyhighly energetic lifestyle at

large body mass, or by complex cognition.

Genetic selection experiments on the ectodermal pattern-

ing of Drosophila have consistently failed to demonstrate that

any left–right encoding information exists to be harnessed for

controlled asymmetry.(e.g. 48) A very slight but real directional

asymmetry of wing size/shape, conserved among several

dipteran fly species, hasneverthelessbeen recently observed,

and claimedasevidence for an arthropod ‘left–right axis’ in the

sense that term has acquired for vertebrates.(49) But such

asymmetry must surely be functionally gratuitous, rather than

in any sense adaptive, and it seemsmore plausible to propose

some intrinsic directional biass in the accuracy with which

mirror-image shape can be generated, for certain kinds of

multicellular structure. The morphogenetic program of wing-

disc development through metamorphosis is very demanding

mechanically, involving collaboration among thousands of

ultimately flattened and extended cells, whose cytoskele-

tons are after all built of ‘handed’ or chiral self-assembling

molecules. Such a phenomenon, a small, biophysically driven

biass to shape-control originating in protein chirality, is quite

distinct froma left–right ‘axis’ of regionalised geneactivation in

the vertebrate sense. The contrast is illustrated abstractly in

Fig. 3. Striking but circumscribed directional asymmetries

occur sporadically among protostome forms with indefinite

cleavage patterns,(reviewed in 50) e.g. the directionality of the

cheliceral size asymmetry in certain free-living decapod

crustaceans. These may represent amplifications of such

minute biophysically driven shape inequalities, through their

subsequent genetic linkage to growth patterns. But without a

Hypothesis

BioEssays 26.4 419

general developmental ‘left and right’ encoding through gene

expression, such mechanisms appear to have been of limited

evolutionary potential.

Further test of this idea of differential phylogenetic con-

straint would come from systematic search for directional

neural-functional left–right complementarity in the brains of

cephalopod molluscs and eusocial insects, neither of which

utilise determinate cleavage plans in development. A pan-

adaptationist view might suggest that, if this can and has

been achieved in vertebrates, it should be present among

such animals, as indeed should a situs-like visceral packing in

fast-swimming squids. I currently know of no examples.

Figure 3. Two contrasting modes of developing directional left–right asymmetry in bilaterian animal forms. Each column of diagrams

represents, from top to bottom, aspects of the progression of development in an axial animal seen in plan view. Left column: thearrangement seen in vertebrates. At early stages, probably while the midline of patterning is itself being established, vectorial transport

processes establish a right–left differential situation across this midline with respect to some signal that initiates distinctive cascades of

gene activity (horizontal arrow concentrating an initially evenly distributed molecule). These cascades (solid black arrows) ‘label’

widespread components of the tissue structure with ‘left’ and ‘right’ identity in normal development (represented as filled or hollow nuclei),

and are stabilised by negative regulatory inter-relationships between their component genes and by a barrier to signal invasion within the

now anatomically discrete midline. This directly enables rudiments of paired organ structures to be assigned different numbers of founder

cells, and/or to undergo differential schedules of subsequent growth and morphogenesis, in effect, potentially separately evolvable left nd

right structures.Rightcolumn:a possible,more evolutionarily restrictedmechanism in protostomeanimals directly inheriting the original

bilaterian dorsoventral axis, and lacking a stereotyped cleavage plan that directly propagates molecule-based chirality. There can be no

differential left and right labelling of early tissue, so both founder-cell numbers and early growth schedules for paired organ primordia can

show only fluctuating, i.e. non-directional asymmetry dictated by the intrinsic limits of developmental control (represented here by those of

the author’s freehanddrawing control). Butwithin cells of an epithelial sheet, whichmight undergoa demandingprogramof vectorial growth

and force production in morphogenesis of complex shapes, an apicobasal structural polarity becomes defined. An intrinsic, molecularly

defined chirality of the cytoskeletal assembly will then bear an opposite relation to the forming supracellular structure on either side of the

midline. This could be sufficient to provide an unavoidable directional asymmetry to the results of mirror-image shape production (shown

exaggerated in the diagram, in relation to that empirically observed in, for example, the insect wing, Ref. 49). Downstream gene control

linkages, harnessed to this local differential result, could conceivably be utilised to achieve further differential development within the

structures concerned.

Hypothesis

420 BioEssays 26.4

Acknowledgments

I am grateful to Michael Akam, Max Telford, Simon Conway-

Morris, AdamWilkins and anonymous referees for comments

during the preparation of this article, though the views ex-

pressed are my own.

References1. Levin M. 2003. Motor protein control of ion flux is an early step in

embryonic left–right asymmetry. Bioessays 25:1002–1010.

2. McManus C. 2002. Right hand, left hand: the origins of asymmetries in

brains, bodies, atoms and cultures. Harvard University Press.

3. Hamada H, Meno C, Watanabe D, Saijoh Y. 2002. Establishment of

vertebrate left–right asymmetry. Nature Reviews: Genetics 3:103–113.

4. Burdine RD, Schier AF. 2000. Conserved and divergent mechanisms in

left–right axis formation. Genes Dev 14:763–776.

5. Yost HJ. 2001. Establishment of left–right asymmetry. Int Rev Cytol

203:357–381.

6. Wright CV. 2001. Mechanisms of left–right asymmetry: what’s right and

what’s left? Developmental Cell 1:179–186.

7. Collignon J, Varlet I, Robertson EJ. 1996. Relationship between asym-

metric nodal expression and the direction of embryonic turning. Nature

381:155–158.

8. Levin M, Johnson RL, Stern CD, Kuehn M, Tabin C. A molecular pathway

determining left–right asymmetry in chick embryogenesis. Cell 82:803–814.

9. Boettger T, Wittler L, Kessel M. 1999. FGF8 functions in specification of

the right body side of the chick. Curr Biol 9:277–280.

10. Rodriguez-Esteban C, Capdevilla J, Economides AN, Pascual J, Ortiz A,

Izpisua-Belmonte J-C. 1999. The novel Cer-like protein Caronte mediates

establishment of embryonic left–right asymmetry. Nature 401:243–251.

11. Fischer A, Viebahn C, Blum M. 2002. FGF8 acts as a right determinant

during establishment of the left–right axis in the rabbit. Curr Biol 12:

1807–1816.

12. Patel K, Isaac A, Cooke J. 1999. Nodal signalling and the roles of

the transcription factor factors SnR and Pitx2 in vertebrate left right

asymmetry. Curr Biol 9:609–612.

13. Cooke J. 2004. Developmental mechanisms and evolutionary origin of

vertebrate left/right asymmetries. Biol Rev Cambridge Phil Soc in press

(vol. 79 (3)).

14. Afzelius BA. 1976. A human syndrome caused by immotile cilia. Science

193:317–319.

15. Brown NA, Wolpert L. 1990. The development of handedness in left/right

asymmetry. Development 109:1–9.

16. Brown NA, McCarthy A, Wolpert L. 1991. Development of handed body

asymmetry in mammals. In: Bock GR, Marsh J, editors. CIBA Symposium

162: Biological Asymmetry and Handedness. Chichester: Wiley & Sons;

p 182–195.

17. Nonaka S, Tanaka Y, Okada Y, Takeda S, Harada K, Harai Y, Kido M,

Hirokawa N. 1998. Randomisation of left–right asymmetry due to loss of

nodal cilia generating leftward flow of extraembryonic fluid in mice

lacking KIF3b motor protein. Cell 95:829–837.

18. Nonaka S, Shiratori H, Saijoh Y, Hamada H. 2002. Determination of left–

right patterning of the mouse embryo by artificial nodal flow. Nature

418:96–99.

19. Essner JJ, Vogan KJ, Wagner MK, Tabin CJ, Yost HJ, Bruekner M. 2002.

Conserved function for embryonic nodal cilia. Nature 418:37–38.

20. Watanabe D, et al. 2003. The left–right determinant Inversin is a com-

ponent of node monocilia and other 9þ0 cilia. Development 130:1725–

1734.

21. Okada Y, Nonaka S, Tanaka Y, Saijoh Y, Hamada N, Hirokawa N. 1999.

Abnormal nodal flow precedes situs inversus in iv and inv mice.

Molecular Cell 4:459–468.

22. Kilner PJ, Yang G-Z, Wilkes AJ, Mohiaddin RH, Firmin DN, Yacoub MH.

2000. Asymmetric redirection of flow through the heart. Nature 404:759–761.

23. Boorman CJ, Shimeld SM. 2002. The evolution of left–right asymmetry in

chordates. Bioessays 24:1004–1011.

24. Jefferies RPS, Brown NA, Daley PEJ. 1996. The early phylogeny of

chordates and echinoderms and the origin of chordate left–right asym-

metry and bilateral symmetry. Acta Zoologica Stockholm 77:101–122.

25. Cameron CB, Garey JR, Swalla BJ. 2000. Evolution of the chordate body

plan: insight from phylogenetic analysis of deuterostome phyla. Proc Nat

Acad Sci USA 97:4469–4474.

26. Popodi E, Raff RA. 2001. Hox genes in a pentameral animal. Bioessays

23:211–214.

27. Gislen TRE. 1930. Affinities between the Echinodermata, Enteropneusta

and Chordonia. Zoologiske Bidrag Uppsala 12:99–304.

28. Jefferies RPS. 1991. Two types of bilateral symmetry in the Metazoa;

chordate and bilaterian. In: Bock GR, Marsh J, editors, CIBA Symposium

162: Biological Asymmetry and Handedness Chichester: Wiley & Sons;

p 94–127.

29. Gerhart J. 2000. Inversion of the chordate body axis: are there

alternatives? Proc Nat Acad Sci USA 97:4445–4448.

30. Nielsen C. 1999. Origin of the chordate central nervous system—and

the origin of chordates. Dev Genes Evol 209:198–205.

31. Holley SA, Jackson PD, Sasai Y, Lu B, De Robertis EM, Hoffman FM,

Ferguson EL. 1995. A conserved system for dorso-ventral patterning in

insects and vertebrates involving sog and chordin. Nature 376:249–253.

32. Ruppert EE, Barnes RD. 1996. Invertebrate Zoology. USA Harcourt

College Publishers. 6th Edn; p 923–924.

33. Holland ND. 1988. In: Paul CRC, Smith AB, editors. Echinoderm

Phylogeny and Evolutionary Biology. Oxford: Clarendon Press; p 13–25.

34. Chia ES, Buckland-Nicks J. 1984. Locomotion of marine invertebrate

larvae: a review. Can J Zool 62:1205–1222.

35. Albrieux M, Villaz M. 2000. Bilateral asymmetry of the inositol tris-

phosphate-mediated calcium signalling in two-cell ascidian embryos.

Biol Cell 92:277–284.

36. Levin M, Thorlin T, Robinson KR, Nogi T, Mercola M. 2002. Asymmetries

of Hþ/Kþ ATPase and cell membrane potentials comprise a very early

step in left–right patterning. Cell 111:77–90.

37. Cooke J. 1995. Vertebrate embryo handedness. Nature 374:681.

38. Dathe V, Gamel A, Manner J, Brand-Saberi B, Christ B. 2002. Morpho-

logical left–right asymmetry of Hensen’s node precedes the asym-

metric expression of Shh and Fgf8 in the chick embryo. Anatomy and

Embryology (Berlin) 205:343–354.

39. Spemann H, Falkenberg H. 1919. Uber asymmetrische entwicklung

und Situs inversus bei Zwillingen und Doppelbildungen. Roux Arch.

EntwMech 45:371–422.

40. Dominguez P, Jacobson AG, Jefferies RPS. 2002. Paired gill slits in a

fossil with a calcite skeleton. Nature 417:841–844.

41. Shu DG, Conway Morris S, Han J, Chen L, Zhang X-L, Zhang Z-F, Liu HQ,

Li Y, Liu J-N. 2001. Primitive deuterostomes from the Chengjiang

Lagerstatte (Lower Cambrian, China). Nature 414:419–424.

42. Kennedy DN, O’Craven KM, Ticho BS, Goldstein AM, Makris N, Henson

JM. 1999. Structural and functional brain asymmetries in human situs

inversus totalis. Neurol 53:1260–1265.

43. Rasskin-Gutman D, Izpisua-Belmonte J-C. 2004. Theoretical Morphology

of Developmental Asymmetries. Bioessays 26:405–412.

44. Rogers LJ, Andrew RJ, editors. Comparative Vertebrate Lateralisation.

Cambridge University Press.

45. Bisazza A, Rogers LJ, Vallortigara G. 1998. The origins of cerebral

asymmetry: a review of the evidence of behavioural and brain later-

alisation in fishes, reptiles and amphibians. Neuroscience and Neuro-

behavioral Reviews 22:411–426.

46. Maynard Smith J, Burian R, Kauffman S, Alberch P, Campbell J,

Goodwin B, Land R, Raup D, Wolpert L. 1985. Developmental constraint

and evolution. Q Rev Biol 60:265–287.

47. Wood WB. 1991. Evidence from reversal of handedness in C. elegans

embryos for early cell interactions determining cell fates. Nature 349:

536–538.

48. Tuinstra EJ, DeJong G, Scharloo W. 1990. Lack of response to family

selection for directional asymmetry in Drosophila melanogaster: left

and right are not distinguished in development. Proc R Soc London B

241:146–152.

49. Klingenberg CP, McIntyre GS, Zaklan SD. 1998. Left–right asymmetry

of fly wings and the evolution of body axes. Proc R Soc London B

265:1255–1259.

50. Palmer RA. 1996. From symmetry to asymmetry: phylogenetic patterns

of asymmetry variation in animals and their evolutionary significance.

Proc Nat Acad Sci USA 93:14279–14286.

Hypothesis

BioEssays 26.4 421