P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL

September 17, 1997 12:4 Annual Reviews AR041-03

Annu. Rev. Cell Dev. Biol. 1997. 13:5382

Copyright c 1997 by Annual Reviews Inc. All rights reserved

LEFT-RIGHT ASYMMETRY
IN ANIMAL DEVELOPMENT
William B. Wood
Department of Molecular, Cellular, and Developmental Biology, University of
Colorado, Boulder CO 80309-0347; e-mail: [email protected]

KEY WORDS: chirality, embryonic axes, handedness, laterality, polarity, situs inversus

ABSTRACT
Most animal species exhibit left-right asymmetry in their body plans and show
a strong bias for one handedness over the other. The mechanism of handedness
choice, recognized as an intriguing problem over a century ago, is still a mys-
tery. However, from recent advances in understanding when and how asymmetry
arises in both invertebrates and vertebrates, developmental pathways for estab-
lishment and maintenance of left-right differences are beginning to take shape,
and speculations can be made on the initial choice mechanism.

CONTENTS
THE PUZZLE OF HANDEDNESS BIAS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
Axes and Asymmetries in Animal Body Plans . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54
Establishing and Maintaining Handed Asymmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55
Control of Embryonic Handedness . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
L/R Asymmetry versus Bilateral Symmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
CONTROL OF HANDEDNESS IN INVERTEBRATE EMBRYOS . . . . . . . . . . . . . . . . . . . . 56
Handedness in Snails and Its Reversal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 56
Handedness in Nematodes and Its Reversal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 58
Drosophila: True Bilateral Symmetry? . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 62
CONTROL OF HANDEDNESS IN VERTEBRATE EMBRYOS . . . . . . . . . . . . . . . . . . . . . . 64
Development of L/R Asymmetries in Vertebrates . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 64
Twinning and L/R Polarity Reversal . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 71
CONCLUSIONS AND SPECULATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Mechanisms of Initial Handedness Choice . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 73
Bilateral Symmetry versus L/R Asymmetry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 78

53
1081-0706/97/1115-0053$08.00
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

54 WOOD

Only asymmetry can beget asymmetry.

F. R. Japp, 1898

. . . the contrast of left and right is connected with the deepest problems concerning the
phylogenesis as well as the ontogenesis of organisms.

Hermann Weyl, 19521

THE PUZZLE OF HANDEDNESS BIAS

Axes and Asymmetries in Animal Body Plans
Animal body plans can be described in terms of three orthogonal axes: anterior-
posterior (A/P), dorsal-ventral (D/V), and left-right (L/R). Although animals
with clearly defined A/P and D/V axes are generally bilaterally symmetric, most
also exhibit some internal L/R asymmetries. Familiar examples in humans
include the placement of the heart, liver, and other viscera.
A consistent L/R asymmetry presents the possibility of two alternative mirror-
image forms of the body plan, which are of opposite handedness but otherwise
identical.2 L/R asymmetry itself is no more mysterious than A/P or D/V asym-
metries, but how one handedness can be chosen over the other poses an in-
triguing problem. Most L/R-asymmetric species exhibit a bias, usually nearly
absolute, for one enantiomeric form of the body plan over the other. Among
humans, for example, the heart is (almost) always to the left of the midline and
the liver to the right. How is the choice made?
To understand why this question is puzzling, consider sequential establish-
ment of the three axes in a hypothetical embryo, starting with a spherically
symmetric egg (Figure 1). The first axis to be determined, generally A/P, can
pass through any point on the surface and its antipode; which of these becomes
A and which becomes P does not affect the final geometry of the embryo (which
now has A and P poles and an equatorial plane between). Orientation of the
D/V axis is now constrained to be parallel to the equatorial plane, but it can
still be established by choosing any point on the equator, and whether this point
defines D or V also does not affect the final embryonic geometry. Orientation
of the L/R axis is now further constrained to be parallel to only one diameter of
the sphere, orthogonal to the other two axes. Which side will be called L and
1 First quotation from a lecture on Stereochemistry and Vitalism before the British Association,

quoted by Weyl (1952, p. 31); second quotation from Weyl (1952, p. 38).
2 Some definitions: An asymmetric object and its mirror image, e.g. left and right hands, have

the same pattern of asymmetry along their L/R axes but are of opposite handedness; they are called
enantiomeric. A form that has a clear handedness, like a helix, screw, or spiral, is said to be chiral;
such a form and its mirror image have opposite chirality (handedness).
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

LEFT-RIGHT ASYMMETRY 55

Figure 1 Sequential establishment of the A/P, D/V, and L/R axes in a hypothetical embryo. See
text for further explanation.

which R is dictated by convention, but the polarity of the axis is not. Choice
of this polarity determines the handedness of the embryo; it is the only axis for
which orientation is completely predetermined, and selection of one polarity or
the other affects the overall body plan. This choice is clearly not random. How
and when is it made?

Establishing and Maintaining Handed Asymmetry

Many embryos exhibit handed L/R asymmetry, in morphology or differential
gene expression or both, from early stages onward, indicating that polarity of
the L/R axis is an early decision. There is direct evidence that this decision is
made after A/P and D/V polarities have been established in nematodes (Priess
& Thomson 1987), sea urchins (McCain & McClay 1994), and Xenopus (Danos
& Yost 1995). Properties of conjoined twins, discussed further below, suggest
that the same is true for birds and mammals.3
In some invertebrate embryos, L/R polarity is clearly determined as early
as second cleavage. In vertebrates, the first L/R asymmetry so far detected
occurs later. Regardless of when L/R symmetry is broken, handedness must be
dictated either (a) by external influences (in the parental uterus or the egg shell,
for example) or (b) by the intrinsic handedness of some internal embryonic
component. Once established, L/R polarity must be maintained throughout
embryogenesis. The consequences of later ambiguities regarding L/R deci-
sions, as occur in some pathologies of human visceral development, are severe
dysfunction or death of the embryo.

3 In the discussion in the previous section, this order of axis determination was assumed; note,

however, that the properties described for L/R apply to the third orthogonal axis established,
whatever it may be.
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

56 WOOD

Control of Embryonic Handedness

Embryonic handedness and its control were recognized as an intriguing puzzle
more than a century ago (e.g. Crampton 1894, zur Strassen 1896). Early work
on this problem is reviewed in an excellent monograph (Ludwig 1932), which
was reprinted in 1970. During the past several years, partly stimulated by a
wide-ranging Ciba Symposium on this topic (Bock & Marsh 1991), there has
been a resurgence of interest in L/R asymmetry and considerable progress in
documenting its cellular and molecular basis in several organisms, although
the puzzle of handedness bias remains unsolved. This review discusses cur-
rent understanding of L/R asymmetry in embryogenesis, with regard to some
general questions: How does handed L/R polarity arise initially? How is it
controlled genetically? Is it (unlike A/P and D/V polarity) dictated by a uni-
versal mechanism in diverse organisms? How and when is the initial cellular
L/R asymmetry manifested in differential gene expression patterns on the two
sides of the embryo? How are different patterns of gene expression and cell
fates on the left and right maintained as complexity and cell number increase
during development?

L/R Asymmetry versus Bilateral Symmetry

A related intriguing question, on which there is little evidence, regards the
primacy of bilateral symmetry or L/R asymmetry; that is, which derives from
the other? Is bilateral symmetry the ground state for embryosexpected to
result from the ordered modulation of cleavage planes during cell prolifera-
tion in embryogenesisso that L/R asymmetry must be imposed by special
mechanisms? If so, why has such asymmetry evolved? Or is L/R-asymmetric
embryogenesis the norm on which bilateral symmetry must be imposed during
development? Extensive discussion of these questions is beyond the scope of
this review, but they come up in several contexts below and are dealt with again
briefly in conclusion (for more comprehensive treatment, see Jefferies 1991).

CONTROL OF HANDEDNESS IN INVERTEBRATE

EMBRYOS
Handedness in Snails and Its Reversal
SPIRAL CLEAVAGE Gastropods such as the fresh-water snail Limnea have heli-
cal shells that usually coil with a right-handed (dextral) screw axis. Their early
embryonic cell divisions show a pattern of spiral cleavage, in which the spindle
axes of sister cells are tilted alternately clockwise and counterclockwise, so that
successive cleavages give rise to tiers of cells that adopt a close packing with
minimal surface area. The handedness of the tilt is normally dextral at second
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

LEFT-RIGHT ASYMMETRY 57

a b

Limnea - dextral

c d e

Physa - sinistral

Figure 2 Orientation of spindle axes and cleavage planes in early snail embryos. All embryos are
oriented with the animal pole toward the viewer. Panels (a) and (c) are schematic representations of
2-cell embryos showing tilts of the spindles at mitosis: the broad end of each arrowhead represents
the spindle pole that is tilted upward, i.e. toward the viewer. Panels (b, d, e) show 4-cell and 8-cell
embryos of Limnea and Physa as drawn by Crampton (1894).

cleavage (Figure 2); this handedness, in turn, is reflected in the subsequent

handedness of the body plan, which (usually, see Robertson 1993) corresponds
to the direction of shell coiling. Other similar genera, such as Physa, and
rare populations or individuals of Limnea have left-handed (sinistrally) coiled
shells. Crampton (1894) observed that in Physa embryos, the handedness of
spiral cleavage was also reversed relative to dextral cleavage, beginning at the
second division (Figure 2).

SINISTRAL MUTATIONS IN LIMNEA Later breeding experiments between the

common dextral and rare sinistral individuals of Limnea showed that sinis-
trality is inherited as a recessive maternal-effect mutation at a locus named
sinistral (Sturtevant 1923), the wild-type function of which is required mater-
nally for normal dextral spindle orientation at second cleavage and subsequent
shell coiling. The alteration in this function resulting from sinistral mutant
alleles causes not randomization of spindle orientation, but rather 100% rever-
sal of handedness. Further genetic analysis of this phenomenon (Freeman &
Lundelius 1982) showed that the so-called sinistral mutations arise and revert
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

58 WOOD

at a frequency higher than expected for point mutations and that the sinis-
tral locus is probably complex. These authors presented genetic evidence for
a model in which anomalous dextral and sinistral animals arise by mitotic or
meiotic crossover events in the maternal germ line between two closely linked
sites at this locus, both of which must be + on one chromosomal homologue
during oogenesis for production of dextral offspring. They also reported the
surprising finding that injection of ooplasm from dextral eggs into precleavage
eggs from a sinistral parent caused the latter to cleave and develop dextrally!
Unfortunately, neither the molecular genetics nor the biochemical nature of
the dextralizing factor have been further investigated. However, these results
suggest that the default mode may be sinistral cleavage, which is reversed by
action of the normal sinistral gene product.

Handedness in Nematodes and Its Reversal

ASYMMETRY IN ADULTS AND EMBRYOS Nematodes are predominantly bilater-
ally symmetric, but they exhibit several internal L/R asymmetries (zur Strassen
1896, 1951, 1959). In the free-living species Caenorhabditis elegans these
include positioning of the gonad and intestine in the pseudocoelom, placement
of the coelomocytes and a few other cells, and the path and anatomy of the
ventral nerve cord, in which cell bodies are on one side and processes on the
other (Sulston & Horvitz 1977, Sulston et al 1983; reviewed in White 1988).
These asymmetries have the normally invariant handedness shown in Figure 3a,
which is arbitrarily referred to as dextral.
The origins of handed asymmetry, which first appears early in nematode
embryogenesis (zur Strassen 1896), have been described in detail for C. elegans.
The sperm entry point at fertilization defines the future posterior pole of the
embryo and establishes its A/P polarity (Goldstein & Hird 1996). The 2-cell
embryo appears to be radially symmetric around the A/P axis. D/V polarity
is established at second cleavage as the elongating AB-cell spindle, originally
orthogonal to the A/P axis, is forced by constraints of the egg shell to assume
an oblique angle, with one spindle pole posterior to the other. The posterior
spindle pole gives rise to the nucleus of the ABp cell, whose position in the
resulting 4-cell embryo (Figure 3b; see legend for cell nomenclature) defines
the future dorsal side of the embryo (Deppe et al 1978, Sulston et al 1983). D/V
polarity can be experimentally reversed during this process by reorienting the
AB spindle with a microneedle (Priess & Thomson 1987).
L/R asymmetry and polarity become apparent during the next cleavage, in
which ABa and ABp divide in the L/R direction. The two spindles originally
form parallel to the L/R axis, orthogonally to both the A/P and D/V axes, but
then skew clockwise (as viewed from the dorsal side) in parallel just prior to
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

LEFT-RIGHT ASYMMETRY 59

a b
D
A

A P

XC

CC'S
V
G

I R
R L
VNC
CC'S A P
V

D R

A P
A

V L

Figure 3 Asymmetries in normal (dextral) C. elegans adult hermaphrodite and early embryos.
(a) Adult hermaphrodite, ventral view: A, anterior; P, posterior; D, dorsal; V, ventral; L, (animals)
left; R, (animals) right; XC, excretory cell; CCs, coelomocytes; G, gonad; VNC, ventral nerve
cord; V, vulva. (b) Early embryos: stages and views as indicated. Sister cell pairs are indicated
where convenient by a hash mark across the cleavage plane that separates them. Diagrams show
four of the five somatic founder cells (AB, MS, E, C) and the germ line cells P2 and P3. Descendants
of founder cells (e.g. AB) are named according to their relative a/p or l/r positions: e.g. ABa is the
anterior daughter of AB, and ABal is the left daughter of ABa. For clarity in the diagrams of 6-
and 8-cell embryos, the AB prefix has been omitted from the names of AB descendants (modified
from Wood et al 1996).
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

60 WOOD

cytokinesis, so that both spindle poles on the left are anterior to their respective
opposite poles on the right. As a result, in the 6-cell embryo (Figure 3b), the
two daughter cells on the left side, ABal and ABpl, lie anterior to their sister
cells on the right, ABar and ABpr, respectively (Sulston et al 1983, Wood
1991). The subsequent EMS and P2 cell divisions, primarily A/P in direction,
are constrained by the positions of the AB-derived cells to skew as well, so
that the resulting 8-cell embryo is markedly L/R asymmetric (Figure 3b). This
asymmetry persists throughout the cleavage stage but becomes less pronounced
during later embryogenesis, as the embryo establishes bilateral symmetry (see
below).
Does the early asymmetry of the cleavage stage dictate the handed asym-
metries seen in adults? In his classic cell lineage studies on the large parasitic
nematode Ascaris megalocephala, zur Strassen (1896), using fixed prepara-
tions, observed that 2.5% of the embryos had reversed (sinistral) handedness,
corresponding to a similar percentage of sinistral animals in adult populations.
However, because embryos of this obligate parasite do not develop to adulthood
in the laboratory, he was unable to demonstrate that reversed embryos give rise
to reversed adults.
In C. elegans, such cause and effect was established with the demonstra-
tion that reversal of embryonic handedness at the 6-cell stage by microma-
nipulation (Wood 1991) results in apparently complete mirror-image devel-
opment. Following a successful operation, all the normal L/R asymmetries
examined were reversed during embryogenesis and larval growth to produce
viable, healthy sinistral adults. Sinistral hermaphrodites could mate with dextral
males to yield outcross progeny and were also self-fertile, producing normal
numbers of progeny with normal dextral handedness. These results demonstrate
that dextral handedness is genetically, rather than epigenetically determined,
that it is not essential for viability, and that embryonic handedness at the 6-cell
stage dictates the handedness of most if not all subsequent L/R asymmetries
during development.
The mechanism by which the normal handedness bias might be established
remains enigmatic. There is no apparent L/R asymmetry in the C. elegans
4-cell embryo. Moreover, reversal of the D/V axis by micromanipulation at the
3- to 4-cell stage does not result in L/R reversed embryos (Priess & Thomson
1987), indicating that L/R polarity is not established independently of D/V
polarity prior to this time. Presumably, therefore, some internal component of
the embryo with intrinsic chirality is responsible for the biased skewing of the
ABa and ABp spindles during their cleavage to produce the observed invariant
handedness. One such component could be the centrioles, but there is presently
no evidence bearing on this possibility (Wood et al 1996; see Conclusions and
Speculations).
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

LEFT-RIGHT ASYMMETRY 61

Other nematodes differ from C. elegans in their degree of handedness in-

variance. Late in his career, zur Strassen (1951, 1959) examined embryos
from eight diverse free-living and parasitic nematodes and found characteris-
tic but very different frequencies of sinistral embryos, ranging from <0.1% in
one species to 50% for another (the dung beetle parasite Bradynema rigidum).
However, he found the same dextral embryonic handedness preference for all
seven of the species that exhibited a bias.
A possible explanation for the frequency differences could be that constraints
exerted by the egg shell can differ in the extent to which they reinforce the
handedness bias. Among animals developing from early C. elegans embryos
treated with chitinase to remove the egg shell, the frequency of reversed animals
increases to about 5% (Wood & Kershaw 1991, Wood et al 1996). In these
embryos, the configuration of the 4-cell stage appears quite similar to that seen
normally in 4-cell Ascaris embryos (zur Strassen 1896). Other early workers
claimed to have observed rotation of the AB quadrant of cells at the 6-cell stage
in living Ascaris embryos, resulting in handedness reversal (e.g. Dunschen
1929). It may be that a less constraining egg shell in Ascaris allows occasional
cell rearrangements in the early embryo that give rise to reversed handedness,
and that the different reversal frequencies seen in other species could be at least
partially explained by differences in egg shell constraints during early cleavage.
A recent finding that low-temperature growth increases the reversal frequency
in C. elegans can be interpreted as an effect on egg shell formation (Wood et al
1996).

SINISTRAL NEMATODE SPECIES AND SINISTRAL MUTANTS Although most ne-

matodes appear to be primarily or exclusively dextral, there is at least one
species, Acrobeloides bodenheimeri, in which non-interbreeding sinistral as
well as dextral isolates or subspecies have been reported (Steiner 1936, Siddiqi
et al 1992, Felix et al 1996). This situation is reminiscent of the snail genera
Limnea and Physa, or dextral and sinistral subpopulations of Limnea (which
can interbreed). It raises the question of whether similar influences on early
cleavage patterns could be operating in both snails and worms (see Conclusions
and Speculations).
Genetic analysis of handedness bias in C. elegans would obviously be highly
desirable. Ongoing screens for mutants exhibiting either randomization or
complete reversal of handedness without significant lethality have so far been
unsuccessful, although mutation in at least one gene that affects spindle ori-
entations primarily in the cleavages of ABa and ABp, where handedness is
established, does lead to a much higher than normal percentage of reversed
animals among survivors (D Bergmann, L Rose & WB Wood, unpublished
results).
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

62 WOOD

IMPOSITION OF BILATERAL SYMMETRY: MAINTAINING L/R DEVELOPMENTAL

DIFFERENCES TO ELIMINATE L/R MORPHOLOGICAL DIFFERENCES Except for the
few L/R asymmetries in the adult body plan shown in Figure 3, nematodes are
essentially bilaterally symmetric, with similar cells and structures at equivalent
positions on both sides along the body axis. Given the striking asymmetry of
the early embryo, bilateral symmetry must somehow be superimposed during
subsequent embryogenesis. This means that the body plan cannot arise from
identically programmed bilaterally symmetric cell lineages on the two sides of
the embryo, with lineally homologous cells adopting equivalent fates (which
is generally the rule in C. elegans development; Sulston 1983, Sulston et al
1983). Instead, there must be L/R differences in the specification of many of
these homologues to compensate for the asymmetric arrangement of the early
blastomeres and produce a bilaterally symmetric animal, as first pointed out by
Sulston et al (1983). The extent of these differences can be seen by tracing the
lineal origins of bilaterally equivalent cells on the two sides of the animal. As
shown in Figure 4, the origins of many of these cells are quite different on the
left and the right.
The embryonic reversal experiments cited above showed that these differ-
ences must be dictated by cell interactions after the 6-cell stage, rather than
by segregation of cell-autonomous left and right determinants, which would
have caused physically reversed embryos to develop abnormally (Wood 1991,
Wood & Kershaw 1991). Subsequent cell ablation and genetic studies have
indeed demonstrated that interactions between specific AB-derived and non-
AB-derived blastomeres establish cell fate differences between lineally homol-
ogous cells on the two sides of the embryo, and the molecular bases for a few
of these interactions have been identified (Hutter & Schnabel 1995). For exam-
ple, at the 12-cell stage, because of the embryonic asymmetry, the MS founder
cell contacts the hypodermal precursor cell ABalp on the left, but not its lineal
homologue ABarp on the right. MS passes an inductive signal, mediated by
the Notch homologue GLP-1 to ABalp, causing its fate to be different from that
of ABarp. If the MS cell is ablated, or if the embryo is mutant for the glp-1
gene, this signal is not passed, and both ABalp and ABarp adopt the ABarp fate
(Hutter & Schnabel 1994, Gendreau et al 1994). Several later inductions have
been shown to specify subsequent more minor L/R differences in the AB lineage
(Bowerman et al 1992, Hutter & Schnabel 1995, Moskowitz & Rothman 1996).

Drosophila: True Bilateral Symmetry?

Another potentially valuable organism for genetic analysis of handedness de-
termination would, of course, be Drosophila. The fruit fly, however, is almost
entirely bilaterally symmetric throughout development. The only handed asym-
metries reported so far are a twist in the gut and a rotation of the male sex organ,
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

LEFT-RIGHT ASYMMETRY 63

Figure 4 Lineal relationships among 18 pairs of contralaterally analogous cells in the AB lineage
of C. elegans. Pairs of cells shown opposite each other in the center of the diagram are analogous
in both position and fate: the members of each pair occupy approximately equivalent positions
on the left and right sides of the embryo at the time they are born (32-AB-cell to 128-AB-cell
stages), and they give rise to nearly or completely identical numbers and types of progeny cells
during subsequent development. Their ancestors, back to the 4-AB-cell stage, are represented
in their approximate relative A/P positions, with cells on the left more anterior than their lineal
homologues on the right. Note that analogues among ABp descendants in the seven most posterior
pairs are also lineal homologues, whereas the remaining more anterior analogues all have different
lineal ancestries on the two sides of the embryo (as can be seen from their cell names as well as
from the lineage diagram). Four analogues are generated by divisions that cross the midline (based
on Sulston et al 1983, Sulston 1983; reproduced with permission from Wood & Kershaw 1991).
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

64 WOOD

so that phenotypic identification of reversed animals in mutant screens would

be difficult. Based on failed attempts to select for L/R asymmetry in popula-
tions, it has been suggested that there is no L/R polarity in Drosophila embryos
(Tuinstra et al 1990).

CONTROL OF HANDEDNESS IN VERTEBRATE

EMBRYOS
Development of L/R Asymmetries in Vertebrates
MORPHOLOGICAL L/R ASYMMETRIES Like nematodes, vertebrates are gener-
ally bilaterally symmetric with a few L/R asymmetries of nearly invariant hand-
edness, most notably the placement of the visceral organs. Unlike nematodes,
however, morphological asymmetry is not apparent during cleavage stages. In
chick embryos, a transient slight asymmetry can be seen in Hensens node
during gastrulation (stages 5-7; Hara 1978, Cooke 1995), but the first gross
morphological L/R asymmetry in vertebrates arises only during organogenesis,
when cardiac primordia migrating from either side of the embryo fuse at the
midline to form an initially symmetrical heart tube, which then develops a dex-
tral loop. Subsequently, mouse and chick embryos undergo an axial rotation
toward the right, and further asymmetries become apparent in the placement of
the gut and other viscera.
GENETIC CONTROL OF HANDEDNESS Several mutations have been described
in mice and humans that affect the establishment of embryonic L/R polarity
(laterality), causing either partial or complete reversal of the normal handed
asymmetry of the viscera. Complete reversal (situs inversus viscerum) occurs
at a frequency of at least 1 in 20,000 live human births, often, though not al-
ways, as a heritable condition. Complete situs inversus is often entirely benign
and, therefore, not recognized until later in life (if at all), but partial reversal
(heterotaxia) can lead to severe cardiac abnormalities and other visceral plumb-
ing problems. The best-characterized human example of a heritable laterality
defect is Kartageners syndrome, an autosomal recessive condition resulting in
male sterility, bronchial problems, and randomization of laterality; that is, com-
plete situs inversus in about 50% of afflicted individuals and normal situs (situs
solitus) in the remainder (McKusick 1988). The basis for these phenotypes has
been identified as lack of the outer dynein arms in cilia and flagella (Afzelius
1976). However, the ultrastructural defects differ among affected families,
suggesting that the syndrome may represent a family of mutations affecting
the outer dynein arms, which consist of many proteins. (For convenience, this
syndrome is nevertheless referred to subsequently as if it were a single gene
defect.) The defect causes lack of motility (dyskinesia) in sperm flagella and
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

LEFT-RIGHT ASYMMETRY 65

bronchial cilia and, in an unknown manner, randomization of embryonic hand-

edness, but the mutant gene(s) have not been identified. Other human heritable
laterality defects are even less well understood (e.g. Burn 1991, Brueckner et al
1991, Casey et al 1993).
In mice, the recessive spontaneous iv (inversus viscerum) mutation on chro-
mosome 12 also leads to randomization of handedness; about 50% of surviv-
ing homozygous animals have normal situs, and the rest exhibit situs inversus
(Hummel & Chapman 1959, Layton 1976, Brueckner et al 1989, McGrath et al
1992). The randomization depends on embryonic and not maternal genotype,
as shown by transfers of iv/iv embryos into +/+ mothers and vice versa (Brown
et al 1990). In contrast, the inv (inversion of embryonic turning) mutation, re-
sulting from a transgene insertion on chromosome 4, causes almost 100% of
homozygous embryos to develop with reversed situs, although these embryos
do not survive and may have other defects (Yokoyama et al 1993). inv is unique
so far among mammalian mutations in causing 100% reversal of the L/R axis,
reminiscent of the sinistral mutation in snails. Legless, also a mouse insertional
mutation on chromosome 12, causes multiple asymmetric limb malformations
and situs inversus with correlated variable expressivity (Schreiner et al 1993).

MOLECULAR EVIDENCE FOR EARLY L/R ASYMMETRY IN MOUSE AND CHICK

EMBRYOS Recent molecular analyses have not only identified several more
genes involved in the establishment of embryonic handedness, but have also
demonstrated that L/R asymmetry, although not evident from morphology, is
established in vertebrate embryos well before heart looping. The stage was set
for this work by experiments showing that the handedness of heart looping and
subsequent situs could be influenced by early surgical or pharmacological treat-
ments. In the rat, adrenergic agonists applied during but not after gastrulation
were shown to cause situs reversal (Fujinaga & Baden 1991a,b), and a similar
effect on heart looping was demonstrated in chick embryos (Hoyle et al 1992).
In Xenopus, perturbation of the blastocoel roof extracellular matrix (ECM) dur-
ing early gastrulation was found to cause random orientation of the heart and
other viscera (Yost 1992; see also Ludwig 1932, p. 381). These indications that
handedness of the body plan is established during gastrulation led to a search
for asymmetric expression of genes known to be involved in axial patterning
at this stage. By in situ hybridization with molecular probes in chick embryos,
Levin and coworkers (1995) demonstrated asymmetric expression, in a tempo-
ral sequence beginning at stage 4, of activin receptor IIa, Sonic hedgehog (Shh),
and cNR-1, a homologue of mouse nodal. The winged-helix class transcription
factor HNF-3 also showed a brief, transient asymmetry of expression during
stage 4. Several other patterning genes tested were expressed symmetrically.
Further, these authors showed that ectopic application of activin or expression
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

66 WOOD

Figure 5 Proposed developmental pathways for establishment of normal laterality in mouse, chick,
and Xenopus. Question marks indicate possible unidentified components based on the likelihood
that these pathways are homologous. Parentheses enclose factors whose involvement or position
is still tentative. See text for further explanation.

of Shh on the wrong side could alter heart situs. The results suggested a pathway
(Figure 5b) in which activation of activin receptor IIa on the right side of the
primitive streak leads to repression of Shh on the right, thereby restricting its
expression to the left side of the streak, where it causes asymmetric induction
of nodal. Subsequent work has shown that the activin B gene is expressed
asymmetrically on the right side at stage 3, strongly suggesting that it is the
ligand that triggers this sequence of signals, which demonstrates the earliest
L/R asymmetry so far in the chick embryo (Levin et al 1997). The products of
all these genes are of known importance in axial patterning. Activin and Nodal,
both members of the TGF superfamily of growth factors, and Shh, a vertebrate
homologue of the Drosophila hedgehog gene product, are all secreted proteins
likely to act non-cell-autonomously as signaling ligands.
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

LEFT-RIGHT ASYMMETRY 67

Another possible signaling molecule in this process could be retinoic acid

(RA). Application of RA to the right precardiac field, but not to the left, caused
randomization of heart looping, suggesting a normal role for RA signaling
in specification of the left side (Smith et al 1997). These investigators also
observed statistically insignificant effects of RA application on the laterality
of nodal (though not Shh) expression, consistent with the possibility that RA
could act early in the pathway of Figure 5b.
Recent studies in the mouse have suggested the presence of a similar pathway
and related it to the action of the iv and inv genes. A possible new component,
the lefty gene on chromosome 1, encodes a divergent TGF superfamily mem-
ber. It is first expressed symmetrically, but by 8 days, in embryos with a few
pairs of somites, the protein is detected asymmetrically on only the left side of
the primitive streak and in the left lateral plate mesoderm (Meno et al 1996).
In about half of iv homozygous embryos and about one fourth of the em-
bryos produced by inv/+ x inv/+ matings (presumably the inv homozygotes),
the Lefty protein is detected only on the right. A nodal::lacZ reporter allele
has been used to show that, as in chick, nodal is expressed only on the left
side of the node (Collignon et al 1996). Moreover, in embryos from inv/+ x
inv/+ matings, the sidedness of nodal expression corresponds to the direction
of heart looping and embryonic turning: on the left in normal embryos and
on the right in the approximately 20% of reversed, presumed inv homozygous
embryos. In contrast to the chick studies, no asymmetry in either transcript
or protein localization was seen for Shh or HNF-3 gene products. Lack of
asymmetry for Shh expression has been reported for other vertebrates as well
(Ekker et al 1995, Collignon et al 1996). Moreover, Shh/ mice produced
by targeted gene disruption exhibited no laterality defects (Chiang et al 1996),
despite major abnormalities in axial patterning as well as in establishment or
maintenance of the notochord (see following section). In contrast, doubly het-
erozygous HNF-3 +/ nodallacZ/+ embryos exhibited LacZ staining on both
left and right sides and frequent defects in visceral situs. Embryos with more
staining on the left developed with normal situs; in embryos with apparently
equal staining on both sides, situs was random. These results indicate a ge-
netic interaction between HNF-3 and nodal, supporting a role for the former
in establishment of L/R asymmetry. Similar results on the sidedness of nodal
expression in mice and the nodal homologue Xnr-1 in Xenopus were obtained
by in situ hybridization, which also was used to demonstrate that sidedness of
mouse nodal expression is randomized in iv/iv and reversed in inv/inv embryos
(Lowe et al 1996). Together these studies show that the signaling events lead-
ing to asymmetric nodal expression are downstream of the iv and inv genes
(Figure 5a), which could therefore be involved in the initial choice of L/R
polarity.
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

68 WOOD

EARLY L/R DETERMINANTS AND THE ROLE OF THE NOTOCHORD IN XENOPUS AND
ZEBRAFISH The first evident morphological asymmetry in both Xenopus and
zebrafish embryos is the dextral looping of the heart. Danos & Yost (1995)
noted that when Xenopus anterior dorsal development is perturbed, either by
UV irradiation during the first cell cycle or by injection of Xwnt-8 DNA into
dorsal blastomeres between the 4- and 32-cell stages, the resulting embryos not
only fail to develop dorsal anterior structures (e.g. head, notochord) but also
exhibit cardiac left-right reversals at a frequency that correlates with the extent
of dorsal anterior defects, approaching 50% for severely defective embryos. The
best correlation was observed with extent of anterior notochord development,
which in turn reflects the amount of organizer activity (Stewart & Gerhart
1990). These results are therefore consistent with those from chick and mouse
described above, which implicate gene expression in the organizer (node) region
in establishment of L/R polarity.
Further evidence for notochord involvement has come from the finding that
two zebrafish mutants, no tail (ntl) and floating head ( flh), both of which lack
notochord (Halpern et al 1993, 1995, Odenthal et al 1996, Stemple et al 1996),
exhibit randomized laterality of cardiac looping (Danos & Yost 1996). Interest-
ingly, both genes encode putative transcription factors: homologues of mouse
Brachyury (Schulte-Merker et al 1994) and Xenopus Xnot (Halpern et al 1995),
respectively. Moreover, experiments in Xenopus involving extirpation of noto-
chord and explantation of cardiac primordia show that dextral cardiac looping
requires the presence of dorsoanterior structures, including the notochord, dur-
ing neural fold stages (Danos & Yost 1996).
The first molecular asymmetry in Xenopus embryos is found in the expres-
sion of the nodal homologue Xnr-1 (Lowe et al 1996, Lohr et al 1997), which
is seen to the left of the organizer region during gastrulation, as in mouse
and chick embryos. However, an earlier L/R asymmetry had been noted in
the mesoderm-inducing potential of the Nieuwkoop center (Boterenbrood &
Nieuwkoop 1973), a group of dorsal blastomeres near the vegetal pole of the
32-cell embryo (Figure 6) that produces maternally encoded signals for meso-
derm induction and formation of the organizer before embryonic transcription
begins. To identify the molecular basis for this asymmetry, Hyatt and coworkers
(1996) injected mRNAs (for proteins implicated in dorsoanterior axis forma-
tion) into these blastomeres on either the left or the right. They found that
injection of mRNA for an activatable form of the TGF-related Vg1 (BVg1)
on the right caused randomization of cardiac and visceral L/R asymmetry as
well as Xnr-1 expression; injection on the left did not. Injection of several
other mRNAs had no effect on L/R asymmetry, including Vg1 and a mutant
form of BVg1 (BVg1-2cx), neither of which can be processed to the mature
active molecule. However, an RNA encoding a dominant-negative truncated
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

LEFT-RIGHT ASYMMETRY 69

a b

Nieuwkoop
center
Figure 6 Dorsal views of (a) 16-cell and (b) 32-cell Xenopus embryos, showing vegetal blas-
tomeres (L and R) into which mRNAs were injected to assay for randomization of cardiac asym-
metry. Large black dot represents Nile blue stain marking the cleavage plane that bisects the left
and right sides at approximately the position of the future organizer region. See text for further
explanation (modified from Hyatt et al 1996).

activin receptor (tAR), known to block Vg1 signaling in mesoderm induction

assays (Kessler & Melton 1995), caused rates of cardiac inversion and Xnr-1
mislocalization that were considerably higher when injected on the left than on
the right. The BVg1- and tAR-injected embryos exhibited normal dorsoanterior
and notochord development, arguing against a secondary effect of Vg1 through
effects on notochord formation. These results suggest that establishment of
L/R asymmetry in Xenopus is mediated by asymmetric differential processing
of maternally derived Vg1 precursor protein to the mature form, and that this
event is upstream of asymmetric Xnr-1 expression. The findings, therefore, are
similar to those obtained in chick and mouse, in that an activin-like signal-
ing molecule appears responsible for controlling establishment of L/R polarity
mediated by asymmetric expression of a nodal homologue (Figure 5c). An ap-
parent difference is that in Xenopus, Xnr-1 expression is normally induced on
the left by Vg1 activation on the same side, whereas in chick, nodal expression
is likely to be restricted to the left by expression of activin B on the right
(Levin et al 1997).
In Xenopus, there is evidence that the activin signal may act through the
notochord itself to control nodal expression and subsequent cardiac looping.
Experiments cited above (Danos & Yost 1996) implicated the notochord in spec-
ifying the direction of heart looping. Subsequent timed extirpation and explan-
tation experiments showed that dorsal midline structures, probably including
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

70 WOOD

the notochord, repress Xnr-1 expression in the right lateral plate mesoderm dur-
ing closure of the neural tube, thereby restricting its expression to the left side
prior to the specification of cardiac orientation (Lohr et al 1997). These authors
also observed good correlation between the rates of symmetric or reversed Xnr-
1 expression patterns and cardiac reversal, both among normal embryos and
those treated with UV during the first cell cycle to inhibit formation of dorsoan-
terior structures to varying degrees. Their results suggest that midline cells in
Xenopus may be responsible for the asymmetric inhibitory role postulated for
Shh expression in chick on nodal expression and that Xnr-1 expression in turn
controls cardiac looping (Figure 5c).
Because signaling molecules such as Shh and Nodal are able to move through
considerable distances in developing tissues, their asymmetric restriction to one
side or the other near the dorsal midline and their failure to cross the midline
when ectopically expressed (Levin et al 1995) suggest the existence of a barrier,
which seems likely to be required for generation and maintenance of laterality.
Danos & Yost (1996) note that the notochord could be serving three functions in
these processess: providing a source of asymmetric signals, acting as a midline
barrier or sink for signaling molecules, and perhaps mediating alignment of
ECM fibrils with which the cardiac primordia must interact for correct looping
(Yost 1992). The above findings predict that mutations preventing normal no-
tochord formation are likely to cause randomization of laterality as a secondary
defect. Two examples are the zebrafish mutants flh and ntl described above;
the mouse mutant no turning (Ewart et al 1996) could be another. It is curious
that the Shh mutation in mouse does not appear to cause abnormal laterality
(Chiang et al 1996).

CONTROL OF LATERALITY IN ORGANOGENESIS The asymmetric expression of

nodal appears to dictate the handedness of heart looping and other subsequent
asymmetries in organogenesis. The mechanisms are not yet clear, but compo-
nents in this process are beginning to be identified. In Xenopus, as mentioned
above, Yost (1992) showed that perturbation of the ECM lining the blasto-
coel roof during early gastrulation (by microsurgery, administration of RGD
peptides, or treatment with heparinase) causes random orientation of the heart
and other viscera. A similar effect was seen after application of proteoglycan
synthesis inhibitors during early neurula stages when mesodermal cardiac pro-
genitor cells are moving across the matrix to the ventral midline (Yost 1990).
These results suggest that asymmetric Xnr-1 expression could affect down-
stream signals communicated to organ primordial cells through ECM.
What are these downstream signals? In chick embryos, Isaac et al (1997)
have shown that the cSnR gene, encoding a zinc-finger protein of the Snail
family, is expressed asymmetrically in the right lateral mesoderm. Antisense
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

LEFT-RIGHT ASYMMETRY 71

inhibition of this expression caused randomization of heart looping and sub-

sequent embryo rotation, but did not perturb the normal asymmetry of cNR-1
(nodal) expression. Ectopic expression of activin or Shh decreased the asymme-
try of both cSnR and nodal expression. Taken together, these results suggest that
cSnR acts downstream of or in parallel to nodal in determining laterality. Smith
et al (1997) have reported asymmetric expression of two ECM proteinsone
on the left, the other on the right, in the precardiac mesodermand have shown
that these asymmetries are abolished by prior application of RA to the right pre-
cardiac field. Also in chick, two novel related basic helix-loop-helix (bHLH)
proteins designated dHAND and eHAND have been identified as required for
cardiac looping (Srivastava et al 1995). The mouse dHAND and eHAND ho-
mologues are expressed primarily on the right and the left sides of the looping
heart, respectively. A dHAND null mutation results in complete absence of a
right ventricle (D Srivastava & E Olson, personal communication).
Control of laterality in the three so far best-characterized vertebrates is sum-
marized and compared in Figure 5. The pathways are fragmentary but similar
enough to suggest that they will turn out to be homologous. All three clearly
involve expression during gastrulation of a nodal-like gene to the left of the mid-
line, which controls later laterality decisions. Other features that seem likely to
be common are involvement of activin signaling upstream and of ECM com-
ponents and the bHLH proteins eHAND and dHAND downstream of nodal.
Ordering of the early inv and iv functions in the mouse and their roles in the
initial handedness choice are discussed under Conclusions and Speculations.

Twinning and L/R Polarity Reversal

Laterality defects including complete situs inversus occur in some monozygotic
and conjoined human twins, as well as in spontaneously or induced conjoined
twin embryos in experimental animals. The findings described above provide
a plausible explanation for this phenomenon and its correlation with different
types of embryonic conjunction. For example, Spemann & Falkenberg (1919)
observed that in conjoined amphibian twins induced by medial constriction of
early newt embryos (Triton), laterality was normal in the twin on the left and
randomized in the twin on the right. This result is consistent with the prediction
from Xenopus that the asymmetrically (left-side) processed Vg1 required for
correct lateralization would be present in the left-half embryo and deficient in
the right-half following constriction (Hyatt et al 1996).
In spontaneously arising or induced twin chick embryos, the presence of
laterality defects depends on the orientation of the two primitive streaks (see
Figure 7). In head-to-head twins, where the two streaks oppose each other, no
laterality defects are seen, even if the twinning is induced by transverse cutting
of a normal blastodisc (Levin et al 1997). This result argues that laterality
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

72 WOOD

a c

b d

Figure 7 Orientation of primitive streaks in development of conjoined twins and proposed molec-
ular interactions leading to laterality defects in one twin of a conjoined pair. (a) Various orientations
of a secondary streak relative to the primary streak (heavier line) give rise to twins conjoined as in
(b) for humans or their chick counterparts. Dots at one end of each streak indicate the node. (b)
Names and configurations of the four classes of human conjoined twins. (c) Postulated interactions
of signaling molecules giving rise to laterality defects in the right twin, as often seen for dicephalic
twins arising from parallel twin streaks. (d ) Postulated interactions of signaling molecules giving
rise to laterality defects in the left twin, as sometimes seen for thoracophagus twins arising from
obliquely oriented twin streaks. See text for further description (modified from Levin et al 1996).

is streak-autonomous, determined only after establishment of A/P and D/V

polarity. If left and right external cues were already present at the blastodisc
stage, the secondary embryo with its A/P polarity reversed would be expected
to also exhibit L/R reversal. When twins develop with parallel streaks, the
embryo on the left develops normally, while the embryo on the right exhibits
randomized laterality. This is consistent with the view that activin from the left
embryo would likely inhibit the proposed Shh activation of nodal on the left side
of the adjacent right embryo, which would then express nodal on neither side
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

LEFT-RIGHT ASYMMETRY 73

(Figure 7c). Levin and coworkers (1996) showed that this prediction was borne
out experimentally when Shh and nodal expression patterns were analyzed
in parallel twin streaks by in situ hybridization. In obliquely oriented twins,
where the two streaks are initially farther apart but grow toward each other,
the opposite result was observed: randomization in the left twin and normal
laterality in the right twin. In situ analysis showed that nodal was expressed on
both sides of the left twin, but only on the left side (i.e. normally) in the right
twin (Figure 7d ). This result could be explained by induction of ectopic nodal
expression in the left embryo by Shh protein produced in the left side of the
right embryo as the streaks grow toward each other.
These results appear to explain the types of laterality defects observed in
human conjoined twins, which are grouped into four classes on the basis of their
orientation (Figure 7b). In the sample of 167 pairs of conjoined twins studied by
Levin and coworkers (Levin et al 1996), none of those that would have originated
from primitive streaks oriented end-to-end (craniopagi, joined only at the head;
or ischiopagi, joined at the pelvis) exhibited laterality defects. In contrast, almost
half of those that would have originated from parallel streaks (dicephali, joined
laterally at the chest) or from obliquely oriented streaks (thoracophagi, joined
obliquely at the chest or abdomen) exhibited reversal of heart situs in one of
the two twins, usually the twin on the right. However, as also predicted by the
model, laterality defects in the left twin were more frequent in thoracophagi
(29%) than in dicephali (14%).

CONCLUSIONS AND SPECULATIONS

Mechanisms of Initial Handedness Choice
INTRINSIC OR EXTRINSIC? Is determination of handedness intrinsic to the em-
bryo, or can it be imposed by asymmetries in the embryos external environ-
ment? In all cases where there is evidence bearing on this question, handedness
determinants appear to be intrinsic and somehow dependent on the D/V axis.
In nematodes (Priess & Thomson 1987) and sea urchins (McCain & McClay
1994), experimental reversal of the D/V axis also reverses the L/R axis. In
Xenopus, establishment of laterality is dependent on generation of a complete
dorsoanterior axial structure (Danos & Yost 1995), and in chick (Levin et al
1997) and human (Levin et al 1996), the absence of laterality defects in head-to-
head twins argues against extrinsic determinants (Levin et al 1997), as discussed
above.

SPECULATIONS ON INTRINSIC MECHANISMS The enigma remains: how is L/R

polarity initially established? Is it established by similar mechanisms in all
embryos, or is it, like the A/P and D/V axes, established differently in different
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

74 WOOD

phyla? From the conclusions above, we must begin with the assumption that
laterality is initially dictated by the handed asymmetry of an internal component
or process. A workable model for establishment of L/R differences in devel-
opment must (a) identify this component or process, (b) explain how it can
function to generate observed handed asymmetries, and (c) rationalize in terms
of this component or process the known mutations that result in randomization
or reversal of laterality.
Klars intriguing proposal to explain the inv mutation (Klar 1994), suggesting
that asymmetry in the mouse embryo could be initially dictated by nonrandom
segregation of nonequivalent sister chromatids to specific daughter cells early
in development, fails to satisfy the first criterion: the direction of asymmetric
segregation would have to be specified by an earlier decision dependent on a
pre-existing asymmetry, which the model does not identify.4
Brown & Wolpert (1990) proposed that a chiral molecule or molecular com-
plex could somehow be fixed in a particular orientation relative to the A/P and
D/V axes (implying chirality of a midline structure) so that it could mediate
asymmetric distribution of other components to provide initial specification of
handedness. These authors also pointed out that choice of handedness is sepa-
rable from generation of L/R asymmetry, which will still occur randomly in the
absence of handedness choice, as appears to result from the Kartageners and
iv mutations. This model, postulating handedness determination after many
cells are present, is easier to visualize for vertebrate than for invertebrate em-
bryos in which handedness already can be specified at the second (snails) or
third (nematodes) cleavages. The model also has difficulty explaining the 100%
asymmetry reversal resulting from the inv mutation, although there is precedent
for a single mutation causing chirality reversal of a macromolecular structure
in the case of bacterial flagella (Shimada et al 1975).
As a third model, I propose that centrioles or centrosomes, independently
of their chirality, could specify the initial handedness choice in at least some
organisms. The two centrosomes in each cell at mitosis are temporally distin-
guishable based on the histories of their constituent centriolar pairs (in most
animal species). These are comprised of three generations of centrioles, which
can be called grandmother (G), mother (M), and daughter (D). The centrosome
in each cell following cytokinesis contains a centriolar pair, designated GM;
during G, and S phases G and M separate and replicate to produce a GD and an
MD centriolar pair. The two differ structurally (Rieder & Borisy 1981, Vorobjev
4 Even
if there is nonrandom segration at each cleavage, as suggested below for centrioles, this
process would place the appropriate chromosome in the appropriate cell at a particular location only
in embryos with invariant cell lineages, like those of C. elegans, and unlike those of mammals. In
C. elegans there is evidence that the chromosomes introduced by the sperm at fertilization segregate
randomly (Ito & McGhee 1987).
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

LEFT-RIGHT ASYMMETRY 75

& Chentsov 1982) as well as functionally, as demonstrated most clearly in algal

basal bodies (Ruffer & Nultsch 1987, Beech et al 1988). During subsequent
prophase, the GD and MD centrosomes migrate about 90 away from each other
to opposite points on the nuclear membrane prior to the onset of metaphase. The
plane of this migration, specification of which is not understood, dictates the
orientation of the subsequent spindle that will be formed between the two cen-
trosomes. This in turn determines the cleavage direction of the following cell
division, except in cases where the spindle is specifically reoriented prior to di-
vision by interaction with the cell cortex (Hyman & White 1987, Hyman 1989).
Centrosomes seem especially attractive as possible handedness determinants
in the snail embryo, in which the behavior of the two spindle poles in each cell
during second cleavage is sufficient to determine the entire L/R asymmetry of
the body plan. Following first cleavage to produce an anterior and a posterior
blastomere, the spindles in each cell form perpendicularly to the preceding
spindle orientation, but they become tilted before division, so that in the anterior
cell of the more common dextrally cleaving embryo, the spindle pole on the
embryos right is below the L/R axial plane, whereas in the posterior cell, the
spindle pole on the left is below the plane (Figure 2a). In a sinistral Physa or
Limnea sinistral mutant embryo (Figure 2c), the tilts are reversed (Crampton
1894). Nonrandom segregation of mature and immature centrioles to the two
spindle poles could account for the handed rotations of the two spindles, which
could be moved by attachment at either the mature or the immature end to the
cell cortex in each cell. The 100% reversal resulting from sinistral mutations,
which are recessive, suggests that sinistral cleavage is the ground state and that
the dextral phenotype requires function of the wild-type sinistral gene product
in order for the spindles to rotate appropriately.
To explain how such a mechanism might operate, let us consider a more ex-
plicit model. In the C. elegans 2-cell embryo, there is evidence that the spindle
in the posterior P1 cell is pulled from a D/V to an A/P orientation by fibers that
initially attach both spindle poles to a site on the anterior cortex (Hyman 1989).
A tug-of-war appears to ensue, which one fiber wins, pulling the corresponding
pole around toward the anterior. Suppose, in Limnea at the two-cell stage that
(a) such a cortical attachment site forms in both cells on the vegetal midline;
that (b) the GD centrosome has a stronger affinity than MD for this site; and that
(c) GD centrioles have been segregated nonrandomly to the right spindle pole in
the anterior (upper) cell and the left spindle pole in the posterior (pointed ends
of spindles in Figure 2a).5 These spindle poles will be pulled ventrally in both
5 Horwich & Brueckner (1993) have also proposed involvement of cortical attachment sites
in handedness determination. There is a precedent for nonrandom segration of centrosome-like
organelles in budding yeast, where the newer spindle pole body goes preferentially into the bud
(Vallen et al 1992).
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

76 WOOD

cells, giving rise to the common dextral orientation. The sinistral gene could
encode a maternally supplied component, associating with GD centrosomes
only, which confers higher affinity to the attachment site. Suppose, further, that
if this component is lacking, as in a sinistral mutant, then the MD centrosomes
have a higher affinity than the incomplete GD centrosomes, and sinistral ori-
entation of the spindles will result. One can also imagine a different kind of
mutation, for example, causing a defect in the cortical attachment site. The
resulting predicted phenotype would be randomization of handedness (as for
the iv or Kartageners mutations), with perhaps some inviable embryos. Thus
these two genes would be required for separate processes: sinistral for what I
will call L/R polarity assignment, and the product of the second hypothetical
gene for what I will call choice execution. The first process specifies the in-
trinsic handedness that will be chosen if a choice is made; the second executes
the choice. Together these two processes could constitute the mechanism of
handedness choice that actually establishes a preferred or unique polarity for
the L/R axis during development in a given species.
Similar informational mechanics could be operating to dictate handedness
in nematodes. Laufer et al (1980) noted that an isolated C. elegans AB cell in
culture divides with a spiral cleavage pattern, similar to that of Limnea, except
that succeeding rounds of division spiral in the same direction to give a helical
array of cells, rather than alternating as in the snail. Also, the handedness of this
spiral and the tilt direction of the spindles beginning at the second AB cleavage
are opposite to that of the corresponding features in dextral snails. Therefore,
in a C. elegans embryo, after the anterior cell (ABa) divides, the daughter cell on
the embryos left (ABal) is located below the plane of the L/R axis, and the right
daughter cell ABar is located above. It seems likely that this cleavage pattern is
the ground state for AB division and that in the intact embryo, AB cells follow
this pattern to the extent that interactions with the egg shell and P1-derived
blastomeres allow.6 In the 6-cell embryo, ABal is in fact ventral to ABar. The
clockwise skewing of the ABa and ABp spindles during cleavage that causes
ABal also to be anterior to ABar (Wood 1991) could perhaps result from down-
ward force on the left spindle pole of ABa against the inclined plane provided
by the underlying EMS cell (see Figure 3b), causing this pole to slide forward.
Could initial handedness choice be dictated by a similar mechanism in em-
bryos with other cleavage patterns? In sea urchin and Xenopus, cleavage is
equatorial and no cellular L/R asymmetry is apparent during the cleavage stage.
In most species of both organisms, the first cleavage bisects the embryo at the
dorsal midline, dividing it into left and right halves. At least in the sea urchin,
however, L/R polarity is not yet established at this stage; embryos bisected into

6 Note that the normal handedness of this pattern has been defined as dextral in C. elegans.
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

LEFT-RIGHT ASYMMETRY 77

right and left halves at this or several subsequent cleavage stages can still give
rise to a pair of embryos with normal handedness (McCain & McClay 1994).
The first handed structural asymmetry is seen only at the pluteus stage. Non-
random centriolar segregation at first cleavage therefore seems unlikely as the
mechanism for handedness determination.
In mammalian embryos, which undergo rotational cleavage, handedness also
seems unlikely to be determined in early cleavage because chimeras made at
later stages do not show the heterotaxia that might be expected if cells commit-
ted to left- or right-side fates were mixed. Moreover, among 21 aggregation
chimeras made by combining 8-cell iv/iv and +/+ embryos, the majority de-
veloped with normal situs (Brown et al 1990). In mammals, however, we can
also look to mutant phenotypes for clues to early events. The Kartagener and
iv mutations do not eliminate but simply randomize laterality; therefore, these
genes function in choice execution. The 100% reversal seen in mouse embryos
homozygous for the recessive inv mutation suggests that this gene, like the
sinistral gene in snails, may normally function to reverse a ground-state L/R
asymmetry that is opposite to the normal state in terms of absolute left and
right. Therefore, like sinistral, it functions in polarity assignment.
Can the inv, iv, and Kartageners gene(s) functions be ordered and placed in
the pathway for establishment and maintenance of laterality in mammalian em-
bryos (Figure 5a)? The functions of activin and nodal are clearly downstream
of the inv and iv genes in the mouse. Horwich & Brueckner (1993) point out that
the degree of heterotaxia associated with laterality mutations should correspond
to their position in such a pathway; a defect in the initial choice should result
in only complete reversals or no reversal of the entire body plan, while defects
in later steps would be more likely to affect some organs or parts of organs but
not others. By this criterion, the Kartageners gene would be furthest upstream,
followed by iv and then inv. However, if inv is required for polarity assignment
and iv for choice execution as postulated, it would be predicted that in doubly
homozygous mouse embryos, mutant for both genes, the iv mutation would be
epistatic; that is, the embryos would exhibit randomized laterality. This predic-
tion is borne out by experiment (P Overbeek, personal communication; cited
in Horwich & Brueckner 1993, Klar 1994). By the usual conventions of de-
velopmental genetics, this result indicates that iv acts downstream of inv as the
model proposes.
Given that the defect in Kartageners syndrome may be in a motor protein,
it is tempting to speculate that the normal function of the gene(s) could be
involved in positioning of a chiral intracellular component, and we are back to
the possibility of a role for the centriole (Horwich & Brueckner 1993). Perhaps,
after all, it is also involved in vertebrate handedness choice through control of
spindle orientations, although a mechanism is difficult to imagine at this point.
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

78 WOOD

It is small comfort that there remains a cellular organelle about which so little
is still known that it can serve as the basis for such speculation! We hope
that this situation will change in the near future, allowing a general centriole-
based model for handedness determination to be more rigorously evaluated.
In addition, identification of the iv, inv, and Kartageners gene products could
provide important insight into how the initial handedness choice is made (see
note added in proof).
Bilateral Symmetry versus L/R Asymmetry
It seems likely that external bilateral symmetry, in particular for aquatic organ-
isms, could have adaptive value for efficient locomotion and, therefore, would
have been selected for in evolution. In some embryos, such as those of the ne-
matodes, early embryonic stages are highly asymmetric, and bilateral symmetry
must be superimposed by later compensating adjustments of cell fates for lineal
homologues on the two sides. However, in embryos with radial cleavage (e.g.
sea urchins) or bilaterally symmetric cleavage (e.g. Ascidians), bilateral sym-
metry is present from the outset, and morphological L/R asymmetries appear to
be added at later stages. As more molecular markers are discovered, it will be
of interest to learn how early in these L/R-symmetric embryos asymmetry of
gene expression can be detected. It will also be of interest to determine whether
invertebrate homologues of vertebrate asymmetry markers can be found. If so,
we can ask whether they are expressed with the same absolute handedness as
in vertebrates, or whether the L/R axis, like the D/V axis (Holley et al 1995,
Hogan 1995), could have become inverted in the course of vertebrate evolution.
ACKNOWLEDGMENTS
I am grateful to S Dutcher, R McIntosh, and J Yost for helpful discussions, to
members of my research group for suggestions and comments on the manuscript,
and to M Brueckner, E Olson, P Overbeek, R Schnabel, C Tabin, and J Yost for
communication of unpublished results. Research from the authors laboratory
was supported by a National Institutes of Health grant (HD-29397).

Visit the Annual Reviews home page at

http://www.annurev.org.

Literature Cited

Afzelius BA. 1976. A human syndrome caused rochrysis carterae. Protoplasma 145:47
by immotile cilia. Science 193:3l7 54
Beech PL, Wetherbee R, Pickett-Heaps JD. Bock GR, Marsh J, eds. 1991. Biological Asym-
1988. Transformation of the flagella and metry and Handedness. Ciba Found. Symp.
associated flagellar components during cell Chichester, UK: Wiley. 327 pp. Vol. 162
division in the Coccolithophorid Pleu- Boterenbrood EC, Nieuwkoop PD. 1973. The
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

LEFT-RIGHT ASYMMETRY 79

formation of the mesoderm in urodelan am- Ekker SC, McGrew LL, Lai CJ, Lee JJ, von
phibians. V. Its regional induction by en- Kessler DP, et al. 1995. Distinct expres-
doderm. Wilhelm Roux Arch. Dev. Biol. sion and shared activities of members of the
173:31932 hedgehog gene family of Xenopus laevis. De-
Bowerman B, Tax F, Thomas J, Priess J. 1992. velopment 121:233747
Cell interactions involved in development of Ewart JL, Melloy PG, Cohen MF, Desmond
the bilaterally symmetrical intestinal valve ME, Kuehn MR, Lo CW. 1996. No turning, a
cells during embryogenesis in C. elegans. De- mutation causing laterality defects in mouse
velopment 116:111320 embryos. Mol. Biol. Cell 7:3743 (Abstr.)
Brown N, McCarthy A, Wolpert L. 1990. The Felix M-A, Sternberg PW, De Ley P.
development of handed asymmetry in ag- 1996. Sinistral nematode population. Nature
gregation chimeras of situs inversus mutant 381:122
and wild-type mouse embryos. Development Freeman G, Lundelius JW. 1982. The develop-
110:94954 mental genetics of dextrality and sinistrality
Brown N, Wolpert L. 1990. The development in the gastropod Lymnaea peregra. Wilhelm
of handedness in left/right asymmetry. De- Roux Arch. Entwicklungsmech. Org. 191:69
velopment 109:19 83
Brueckner M, DEustachio P, Horwich A. 1989. Fujinaga M, Baden J. 1991a. Critical period of
Linkage mapping of a mouse gene, iv, that rat development when sidedness of asymmet-
controls left-right asymmetry of the heart and ric body structures is determined. Teratology
viscera. Proc. Natl. Acad. Sci. USA 86:5035 44:45362
38 Fujinaga M, Baden J. 1991b. Evidence for an
Brueckner M, McGrath J, DEustachio P, Hor- adrenergic mechanism in the control of body
wich AL. 1991. Establishment of left-right asymmetry. Dev. Biol. 143:2035
asymmetry in vertebrates: genetically dis- Gendreau S, Moskowitz I, Terns R, Rothman
tinct steps are involved. See Bock & Marsh J. 1994. The potential to differentiate epider-
1991, pp. 20218 mis is unequally distributed in the AB lineage
Burn J. 1991. Disturbance of morphological lat- during early embryonic development in C. el-
erality in humans. See Bock & Marsh 1991, egans. Dev. Biol. 166:77081
pp. 28299 Goldstein B, Hird SN. 1996. Specification of
Casey B, Devoto M, Jones KL, Ballabio A. the anteroposterior axis in Caenorhabditis el-
1993. Mapping a gene for familial situs ab- egans. Development 122:146774
normalities to human chromosome Xq24- Halpern ME, Ho RK, Walker C, Kimmel CB.
q27.1. Nat. Genet. 5:4037 1993. Induction of muscle pioneers and floor
Chiang C, Litingtung Y, Lee E, Young K, Cor- plate is distinguished by the zebrafish no tail
dent J, et al. 1996. Cyclopia and defective mutation. Cell 75:99111
axial patterning in mice lacking Sonic hedge- Halpern ME, Thisse C, Ho RK, Thisse B,
hog gene function. Nature 383:40713 Riggleman R, et al. 1995. Cell autonomous
Collignon J, Varlet I, Robertson EJ. 1996. Rela- switch from axial to paraxial development
tionship between asymmetric nodal expres- in zebrafish floating head mutants. Develop-
sion and the direction of embryonic turning. ment 121:425764
Nature 381:15558 Hara K. 1978. Spemanns organizer in birds. In
Cooke J. 1995. Vertebrate embryo handedness. Organizer: A Milestone of a Half Century
Nature 374:681 Since Spemann, ed. A Nakamura, S Toivo-
Crampton H. 1894. Reversal of cleavage in nen, pp. 22165. Amsterdam: Elsevier/North
a sinistral gasteropod. Ann. NY Acad. Sci. Holland
8:16770 Hogan B. 1995. Upside-down ideas vindicated.
Danos M, Yost J. 1995. Linkage of cardiac left- Nature 376:21011
right asymmetry and dorsal-anterior develop- Holley S, Jackson P, Sasai Y, Lu B, De Robertis
ment in Xenopus. Development 121:146774 E, et al. 1995. A conserved system for dorsal-
Danos MC, Yost HJ. 1996. Role of notochord in ventral patterning in insects and vertebrates
specification of cardiac left-right orientation involving sog and chordin. Nature 376:249
in Zebrafish and Xenopus. Dev. Biol. 177:96 53
103 Horwich A, Brueckner M. 1993. Left, right, and
Deppe U, Schierenberg E, Cole T, Krieg C, without a cue. Nat. Genet. 5:32122
Schmitt D, et al. 1978. Cell lineages of the Hoyle C, Brown N, Wolpert L. 1992. Devel-
embryo for the nematode C. elegans. Proc. opment of left/right handedness in the chick
Natl. Acad. Sci. USA 75:37680 heart. Development 115:107178
Dunschen F. 1929. Inversentwicklung und Mo- Hummel KP, Chapman DB. 1959. Visceral
saikfrage bei Ascaris megalocephala cloq. inversion and associated anomalies in the
Arch. Entw. Mech. 115:237335 mouse. J. Hered. 50:913
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

80 WOOD

Hutter H, Schnabel R. 1994. glp-1 and induc- T, Wright CVE, et al. 1996. Conserved left-
tions establishing embryonic axes in C. ele- right asymmetry of nodal expression and al-
gans. Development 120:205164 terations in murine situs inversus. Nature
Hutter H, Schnabel R. 1995. Establishment of 381:15861
left-right asymmetry in the Caenorhabditis Ludwig W. 1932. Das Rechts-Links Problem
elegans embryo: a multistep process involv- im Tierreich und beim Menschen. Berlin:
ing a series of inductive events. Development Springer-Verlag. 496 pp. Reprinted 1970
121:341724 McCain ER, McClay DR. 1994. The establish-
Hyatt B, Lohr J, Yost H. 1996. Initiation of ver- ment of bilateral asymmetry in sea urchin em-
tebrate left-right axis formation by maternal bryos. Development 120:395404
Vg1. Nature 384:6265 McGrath J, Horwich AL, Brueckner M. 1992.
Hyman A. 1989. Centrosome movement in the Duplication/deficiency mapping of situs in-
early divisions of Caenorhabditis elegans: versus viscerum (iv), a gene that determines
a cortical site determining centrosome posi- left-right asymmetry in the mouse. Genomics
tion. J. Cell Biol. 109:118593 14:64348
Hyman A, White J. 1987. Determination of McKusick VA, ed. 1988. Mendelian Inheritance
cell division axes in the early embryogene- of Man, pp. 102325 (Entry 24440). Balti-
sis of Caenorhabditis elegans. J. Cell Biol. more: Johns Hopkins Univ. Press
105:212335 Meno C, Saijoh Y, Fugii H, Ikeda M, Yokoyama
Isaac A, Sargent MG, Cooke J. 1997. Control T, et al. 1996. Left-right asymmetric expres-
of vertebrate left-right asymmetry by a snail- sion of the TGF-family member lefty in
related zinc finger gene. Science 275:13014 mouse embryos. Nature 381:15155
Ito K, McGhee J. 1987. Parental DNA strands Moskowitz IPG, Rothman JH. 1996. lin-
segregate randomly during embryonic de- 12 and glp-1 are required zygotically for
velopment of Caenorhabditis elegans. Cell early embryonic cellular interactions and
49:32936 are regulated by maternal GLP-1 signal-
Jefferies RPS. 1991. Two types of bilateral sym- ing in Caenorhabditis elegans. Development
metry in the Metazoa: chordate and bilate- 122:410517
rian. See Bock & Marsh 1991, pp. 94127 Odenthal J, Haffter P, Vogelsang E, Brand M,
Kessler DS, Melton DA. 1995. Induction of dor- van Eeden FJM, et al. 1996. Mutations affect-
sal mesoderm by soluble, mature Vg1 pro- ing the formation of the notochord in the ze-
tein. Development 121:215564 brafish, Danio rerio. Development 123:103
Klar AJS. 1994. A model for specification of the 15
left-right axis in vertebrates. Trends Genet. Priess J, Thomson JN. 1987. Cellular interac-
10:39296 tions in early Caenorhabditis elegans em-
Laufer J, Bazzicalupo P, Wood W. 1980. Seg- bryos. Cell 48:24150
regation of developmental potential in early Rieder CL, Borisy GG. 1981. The centrosome
embryos of Caenorhabditis elegans. Cell cycle in PTK2 cells: asymmetric distribution
19:56977 and structure changes in the pericentriolar
Layton WM. 1976. Random determination of material. Biol. Cell 44:11732
a developmental process. J. Hered. 67:336 Robertson R. 1993. Snail handedness. Nat.
38 Geog. Res. Expl. 9:10419
Levin M, Johnson R, Stern C, Tabin C. 1995. Ruffer U, Nultsch W. 1987. Comparison of
A molecular pathway determining left-right the beating of the cis- and trans-flagella of
asymmetry in chick embryogenesis. Cell Chlamydomonas cells held on a micropipette:
82:80314 II. changes in flagellar beat patterns. Cell
Levin M, Pagan S, Roberts DJ, Cooke J, Kuehn Motil. 18:26978
MR, Tabin CJ. 1997. Different aspects of lat- Schreiner CM, Scott WJ Jr, Supp DM, Potter
erality are independently controlled by an ap- SS. 1993. Correlation of forelimb malforma-
parently streak-autonomous signaling path- tion asymmetries with visceral organ situs
way initiated by activin. Development. In in the transgenic mouse insertional mutation
press legless. Dev. Biol. 158:56062
Levin M, Roberts DJ, Holmes LB, Tabin C. Schulte-Merker S, van Eeden FJM, Halpern
1996. Laterality defects in conjoined twins. ME, Kimmel CB, Nusslein-Volhard C. 1994.
Nature 384:321 no tail (ntl) is the zebrafish homologue of
Lohr JL, Danos MC, Yost HJ. 1997. Left-right the mouse T (Brachyury) gene. Development
asymmetry of a nodal-related gene is regu- 120:100915
lated by dorsoanterior midline structures dur- Shimada K, Kamiya R, Asakura S. 1975.
ing Xenopus development. Development. In Left-handed to right-handed helix conver-
press sion in Salmonella flagella. Nature 254:332
Lowe LA, Supp DM, Sampath K, Yokoyama 34
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

LEFT-RIGHT ASYMMETRY 81

Siddiqi M, De Ley P, Khan H. 1992. Acro- velopment. Proc. R. Soc. London. Ser. B 241:
beloides saeedi spn. from Pakistan and re- 14652
description of A. bodenheimeri (Steiner) and Vallen EA, Scherson TY, Roberts T, van Zee
Placodira lobata Throne (Nematoda: Cep- K, Rose MD. 1992. Asymmetric mitotic seg-
halobidae). Afro-Asian J. Nematol. 2:5 regation of the yeast spindle pole body. Cell
16 69:50515
Smith SM, Dickman ED, Thompson RP, Sin- Vorobjev IA, Chentsov YS. 1982. Centrioles in
ning AR, Wunsch AM, Markwald RR. 1997. the cell cycle. I. epithelial cells. J. Cell Biol.
Retinoic acid directs pre-cardiac laterality 98:93849
and the expression of early markers of car- Weyl H. 1952. Symmetry. Princeton, NJ: Prince-
diac asymmetry. Dev. Biol. 182:16271 ton Univ. Press. 168 pp.
Spemann H, Falkenberg H. 1919. Uber asym- White J. 1988. The anatomy. In The Nema-
metrische Entwicklung und Situs inversus tode Caenorhabditis elegans, ed. WB Wood,
viscerum bei Zwillingen and Doppelbildun- pp. 81122. Cold Spring Harbor, NY: Cold
gen. Arch. Entwicklungsmech. Org. 45:371 Spring Harbor Press
422 Wood WB. 1991. Evidence from reversal of
Srivastava D, Cserjesi P, Olson E. 1995. A sub- handedness in C. elegans embryos for early
class of bHLH proteins required for cardiac cell interactions determining cell fates. Na-
morphogenesis. Science 270:199599 ture 349:53638
Steiner G. 1936. Opuscula miscellanea nema- Wood WB, Bergmann D, Florance A. 1996. Ma-
tologica, IV. Proc. Helmintholog. Soc. Wash. ternal effect of low temperature on handed-
2:7480 ness determination in C. elegans embryos.
Stemple DL, Solnica-Krezel L, Zwartkruis F, Dev. Genet. 19:22230
Neuhaus SCF, Schier AF, et al. 1996. Muta- Wood WB, Kershaw D. 1991. Handed asymme-
tions affecting development of the notochord try, handedness reversal and mechanisms of
in zebrafish. Development 123:11728 cell fate determination in nematode embryos.
Stewart RM, Gerhart JC. 1990. The anterior ex- See Bock & Marsh 1991, pp. 14364
tent of dorsal development of the Xenopus Yokoyama T, Copeland N, Jenkins N, Mont-
embryonic axis depends on the quantity of gomery C, Elder F, Overbeek P. 1993. Rever-
organizer in the late blastula. Development sal of left-right asymmetry: a situs inversus
109:36372 mutation. Science 260:67982
Sturtevant AH. 1923. Inheritance of direction of Yost H. 1990. Inhibition of proteoglycan syn-
coiling in Limnea. Science 58:26970 thesis eliminates left-right asymmetry in
Sulston J. 1983. Neuronal cell lineages in the Xenopus laevis cardiac looping. Develop-
nematode C. elegans. Cold Spring Harbor ment 110:86574
Symp. Quant. Biol. 48:44352 Yost H. 1992. Regulation of vertebrate left-right
Sulston J, Horvitz H. 1977. Post-embryonic cell asymmetries by extracellular matrix. Nature
lineage of the nematode Caenorhabditis ele- 357:15861
gans. Dev. Biol. 56:11056 zur Strassen O. 1896. Embryonalentwickelung
Sulston J, Schierenberg E, White J, Thomson J. der Ascaris megalocephala. Arch. Entw.
1983. The embryonic cell lineage of the ne- Mech. 3:27105
matode Caenorhabditis elegans. Dev. Biol. zur Strassen O. 1951. Der Erbgang der
100:64119 Nematoden-Asymmetrie. Verh. Deutsche
Tuinstra EJ, De Jong G, Scharloo W. 1990. Lack Zool. Ges. Wilhelmshaven 1951:7781
of response to family selection for directional zur Strassen O. 1959. Neue Beitraege zur En-
asymmetry in Drosophila melanogaster: wicklungsmechanik der Nematoden. Zoolog-
Left and right are not distinguished in de- ica 107:1142

NOTE ADDED IN PROOF

Review of the literature for this article was concluded in March 1997. Since
that time, the gene for an apparently cytoplasmic dynein heavy chain, with
similarity to axonemal dyneins in the motor domain, has been positionally
cloned as tightly linked to the mouse iv mutation and designated left/right
dynein (lrd ). The allelic iv and legless (lgl) mutations both cause lrd defects:
In lgl mutants lrd is deleted, and in iv mutants a conserved glutamate in the
P1: MBL/ary P2: MBL/plb QC: MBL/tkj T1: MBL
September 17, 1997 12:4 Annual Reviews AR041-03

82 WOOD

Lrd motor domain is replaced with lysine. The lrd mRNA is present in oocytes
and early embryos (M Brueckner, personal communication; presented by DM
Supp et al at the 62nd Cold Spring Harbor Symposium, Pattern Formation Dur-
ing Development, May, 1997). In addition, a YAC clone that rescues both the
lethal and reversed laterality phenotypes resulting from the mouse inv mutation
has been identified, giving promise that the gene(s) responsible will also soon
be molecularly characterized (P Overbeek, personal communication).