Organizer Induction Determines Left-Right Asymmetry in Xenopus · 2017. 2. 25. · In the simplest view, this hypothesis would suggest Harland, 1991; Sokol et al., 1991), suggesting

DEVELOPMENTAL BIOLOGY 189, 68–78 (1997)ARTICLE NO. DB978635

Organizer Induction Determines Left–RightAsymmetry in Xenopus

Nanette Nascone and Mark MercolaDepartment of Cell Biology, Harvard Medical School, 240 Longwood Avenue,Boston, Massachusetts 02115

Vertebrates appear bilaterally symmetrical but have considerable left–right (LR) asymmetry in the anatomy and placementof internal organs such as the heart. Although a number of asymmetrically expressed genes are known to affect LRpatterning, both the initial source of asymmetry and the mechanism that correctly orients the LR axis remain controversial.In this study, we show that the induction of dorsal organizing centers in the embryo can orient LR asymmetry. Ectopicorganizing centers were induced by microinjection of mRNA encoding a variety of body axis duplicating proteins, includingmembers of the Wnt signal transduction pathway. The ectopic and primary body axes form side-by-side conjoined twins,with the secondary axis developing as either the left or right sibling. In all cases, correct LR asymmetry was observed inthe left twin, regardless of whether it was derived from the primary axis or induced de novo by injection of Xwnt-8, b-catenin, or Siamois mRNA. In contrast, the right twin was generally unbiased, regardless of the origin of the left bodyaxis, as seen in many instances of experimentally induced and spontaneous conjoined twins. An unanticipated exceptionwas that right twins induced by b-catenin and Siamois, two downstream effectors of Wnt signaling, exhibited predominatelynormal heart looping, even when they formed the right twin. Taken together, these results indicate that LR asymmetryis locally oriented as a consequence of Wnt signaling through b-catenin and Siamois. We discuss the possibility that signalsupstream of b-catenin and Siamois might be required in order for a right sibling to be randomized. q 1997 Academic Press

INTRODUCTION 1996). Although these studies begin to provide a geneticbasis for asymmetry, the regulatory mechanisms that orient

Despite its bilaterally symmetric appearance, the verte- side-specific gene expression or asymmetrically process ma-brate body has invariant left–right (LR) asymmetries exem- ternal mRNA are unclear. Moreover, the nature of the ini-plified by the position and shape of visceral organs such as tial asymmetry in the embryo is unknown.the heart, gall bladder, and spleen. The developmental ori- Many models for the developmental origin of LR asym-gin of LR asymmetry and how it is correctly oriented is not metry have been influenced by observations of experimen-well understood, although a cascade of secreted signals has tally induced conjoined twins. In 1919, Spemann and Fal-recently been implicated in its regulation in chick embryos kenberg used a longitudinal ligature to bifurcate an amphib-(Levin et al., 1995). Side-specific signaling molecules are ian embryo such that the left and right halves of the originalthought to relay distinct positional information to tissues axis developed into separate twins joined side-by-side at theon the left and right of Hensen’s node during gastrulation trunk (Spemann and Falkenberg, 1919). While the left twinto bias the final morphology of asymmetric organs. While developed normally, the right twin appeared to have lostnot all of the molecular components in the chick have been its normal asymmetric bias and showed an increased pro-shown to function in other vertebrates, the left-sided ex- pensity for inverted organogenesis. In the right twin, indi-pression of genes homologous to the chick nodal-related vidual organs such as the heart are thought to have a roughlyprotein-1 (Cnr-1) appears conserved in mouse and frog em- equal probability of orienting normally or as a mirror imagebryos and may influence handed morphogenesis in these of the normal anatomy. Similar randomization of organ si-species as well (Collignon et al., 1996; Lowe et al., 1996; tus also occurs most frequently in the right sibling of humanMeno et al., 1996). In addition, asymmetric processing of a twins conjoined at the chest and/or abdomen (Burn, 1991;maternally encoded protein in Xenopus, Vg-1, has recently M. Levin and C. Tabin, personal communication).been proposed to affect cardiac situs and the expression of The absence of a consistent polarity in the right twin

has been interpreted to indicate that the right half of thethe Xenopus nodal related protein-1 (Xnr-1) (Hyatt et al.,

68

0012-1606/97 $25.00Copyright q 1997 by Academic Press

All rights of reproduction in any form reserved.

AID DB 8635 / 6x29$$$161 08-04-97 21:31:33 dba

69Determination of Left–Right Asymmetry

vertebrate embryo lacks information critical for correctly tives. First, left or right secondary axes were analyzed fortheir ability to influence the LR asymmetry of the primaryorienting asymmetry. This model suggests that LR posi-

tional information must have been patterned prior to twin- axis. We find that the asymmetric development of the pri-mary axis can be perturbed during gastrulation by a con-ning, possibly by a determinant localized to the left side

(Wilhelmi, 1921; Hyatt et al., 1996), a maternal gradient joined secondary axis on its left (but not right) side, demon-strating that the influence of the left body axis may begenerated from the left (Morgan, 1977; Corballis and Mor-

gan, 1978; Morgan and Corballis, 1978; Corballis, 1983; responsible for the anomalous laterality observed in rightconjoined twins. Second, and most importantly, the second-Morgan, 1991; Almirantis, 1995), or an irreversible modifi-

cation of the right hemisphere (Brown and Wolpert, 1990). ary axes were themselves evaluated for their ability to de-velop normal left–right asymmetry. We show that ectopicSimilarly, predetermination has also been postulated to re-

sult from an asymmetric segregation of imprinted DNA dur- organizers initiated by molecules in the Wnt signaling path-way can correctly determine and orient LR asymmetry ining early cleavage stages (Klar, 1994).

An alternative hypothesis was originally proposed by von these secondary axes, by both morphological and molecularcriteria. Components of this pathway have been implicatedWoellwarth (1950), based on the observation that damage to

tissues on the left, but not the right, side of postgastrulation in the activation of endogenous organizer activity in thedorsovegetal blastomeres of the early embryo (Smith andstage amphibian embryos caused transpositions of organ

situs. In the simplest view, this hypothesis would suggest Harland, 1991; Sokol et al., 1991), suggesting that Wnt-me-diated organizer induction may be sufficient to determinethat mesoderm on the left side of the right twin is compro-

mised by being joined to its sibling, thus randomizing its the LR axis in Xenopus.LR polarity. A more detailed model is based on the recentlyidentified signaling cascade in the chick and argues that afactor from the right side of the left twin would suppress MATERIALS AND METHODSleft-side-specific gene expression in the right sibling (Levinet al., 1996). These models imply that the anomalous later- Embryo Culture and Microsurgeryality in the right sibling is the result of cross-signaling be-

Xenopus laevis embryos were obtained by in vitro fertilization,tween the twins rather than a lack of pattern information.dejellied at the 2- to 4-cell stage in 2% cysteine–HCl, and main-Thus, each axis in a twinned embryo may have an intrinsictained in 3% Ficoll 400 (Sigma) in 11 MMR (Peng, 1991) until thecapacity to generate and orient LR asymmetry.time of microinjection. For microsurgery, embryos were main-

The activity of Spemann’s organizer (or its amniote equiv- tained in 0.11 MMR and dejellied at late blastula stages. Stage 10alent, Hensen’s node) is critical for the induction of dorso- embryos were manually devitellinized and dissected in 0.751ventral (DV) and anteroposterior (AP) pattern in both indi- MMR on agarose-coated operating dishes with an eyebrow knifevidual embryos and twins (Sive, 1993; Slack, 1993; Smith, and sharpened watchmakers’ forceps. A small block of the dorsal

lip (or lateral marginal zone) was grafted to the ventral marginal1993; M. Levin and C. Tabin, personal communication). Aszone (VMZ) of a host gastrula, approximately 1707 to the left orsuch, the organizer is a likely candidate to align the LRthe right of the dorsal midline. Operated embryos were allowed toaxis with these coordinates and may be the source of LRheal in agarose wells, 15 min in 0.751 MMR and §1 hr in 0.11asymmetry. Microsurgical or genetic manipulations thatMMR, before being transferred to individual wells of an agarose-disrupt the integrity of the organizer or its derivatives, thecoated 24-well plate for long-term culture in 0.11MMR plus genta-notochord and the prechordal mesoderm, will perturb bothmycin (25 mg/ml).

dorsoanterior character and heart situs in Xenopus and ze-brafish embryos (Danos and Yost, 1995, 1996). In addition,recent experiments in the chick show that correct asym- mRNA Synthesis and in Situ Hybridizationmetric expression of sonic hedgehog (shh) and HNF-3b is

Synthetic capped mRNA was transcribed in vitro from linearizedrestored when node tissue is regenerated following surgicalpSP64T or pCS2 plasmids containing cDNA encoding axis-induc-ablation (Psychoyos and Stern, 1996). It is uncertain froming molecules. In situ hybridization protocol was performed onthese studies if the organizer is itself capable of orientingstage 23–24 embryos with digoxigenin-labeled probe as described

this LR polarity or whether it maintains or propagates a (Harland, 1991).preexisting pattern. Nonetheless, these results are consis-tent with the hypothesis that the organizer locally specifiesLR asymmetry and that it (or its induction) coordinates LR Production of Twinspolarity with DV and AP pattern.

Secondary body axes were induced by co-injection of syntheticIn this paper, we directly assess the capacity of the Xeno-mRNA encoding Xwnt-8 (1–5 pg), b-catenin (1–5 pg), Siamois (10–pus organizer to determine LR asymmetry in conjoined100 pg), noggin (100–200 pg), chordin (100–200 pg), Xnr-1 (50–

twins. Secondary body axes were induced by microinjection 100 pg), Xnr-4 (30–50 pg), dishevelled (0.5–1 ng), or a dominant-of mRNAs thought to mimic organizer activity and/or the negative version of rat GSK-3 (dnGSK-3; 0.5–1 ng) with 5–10 ngnormal process by which the organizer is induced. The po- of rhodamine-conjugated lysinated dextran (RLDx, 10,000 MW;larity of heart looping and Xnr-1 expression was then ana- Molecular Probes). A total volume of 1–5 nl was injected into a

ventrovegetal blastomere of a 4- to 32-cell stage blastula. Injectedlyzed in these twinned embryos with two distinct objec-

Copyright q 1997 by Academic Press. All rights of reproduction in any form reserved.

AID DB 8635 / 6x29$$$162 08-04-97 21:31:33 dba

70 Nascone and Mercola

embryos were allowed to recover for §30 min in 0.75–11 MMR each mRNA injected. This indicates that perturbation ofbefore being transferred to 0.11 MMR plus gentamycin (25 mg/ml) heart situs is likely due to the presence of a left body axisfor long-term culture. per se, rather than the property of a particular inducing

molecule. Importantly, left secondary axes induced bytransplanting a second blastopore lip to the ventral side of

Imaging and Fluorescence Microscopy a gastrula stage recipient also randomize the direction ofheart looping, indicating that the perturbation can occurAnesthetized (0.01% tricaine in 0.11 MMR) embryos were

viewed on a dissecting microscope for direction of heart looping past the onset of gastrulation (stage 10). Neither injectionat stage 40/. A Zeiss Axiophot microscope was used to visualize of control mRNA encoding b-galactosidase nor transplanta-fluorescent label and for digital imaging. tion of lateral mesoderm to the ventral side of a host embryo

affected LR polarity.The normal asymmetric expression of Xnr-1 mRNA was

Statistical Methods also perturbed in these experiments. In situ hybridizationof stage 23–24 embryos shows that the normally left-sidedThe probability that the various treatments yielded an equalXnr-1 expression in the primary body axis was suppressedincidence of normally and inversely looped hearts as predicted for

unbiased looping was calculated using the x2 algorithm with a by the presence of a left partial axis (Fig. 2A), but was unaf-Pearson continuity correction for sample sizes between 25 and 200, fected by a nearly complete right axis (Fig. 2C). Taken to-as described (Sokal and Rohlf, 1995). Thus, the value (P) listed in gether, these results indicate that organ situs anomalies inTables 1–3 indicates the probability that the empirically derived right twins are caused by a negative influence from theirincidence deviates from a theoretical frequency of 50%. A Põ 0.05 left sibling.was interpreted to indicate that the null hypothesis (that thesefrequencies reflect a random determination of LR polarity) couldbe rejected. Conversely, P ú 0.05 was interpreted to indicate that Secondary Axes Have Normal LR AsymmetryLR polarity is statistically random. P values were not calculated

The above experiments demonstrated that ectopic orga-where the incidence of normal looping is 100%.nizer activity can alter LR polarity when positioned on theleft side of an embryo. We then used this system to addresswhether induction of an organizing center is sufficient toRESULTSproperly generate and orient LR asymmetry de novo in thesecondary body axes. As before, injections were targeted toLeft Secondary Axes Can Perturb Asymmetricprospective ventral blastomeres such that the primary andDevelopmentectopic organizing centers would be positioned 1807 apart.Should the induced twin exhibit bilateral asymmetry, theBefore assessing the ability of the Xenopus organizer to

autonomously specify LR pattern in the induced axes, it polarity of this asymmetry would reveal the ability of theectopic organizer to specify correct LR orientation. For ex-was important to first characterize the influence that an

induced twin may exert on the asymmetric development ample, an inverted or random orientation would mean thatthe induced organizing center was incapable of specifyingof a neighboring axis. We induced a second organizer by

microinjecting synthetic mRNAs into a ventrovegetal blas- LR polarity (Figs. 3A and 3B). An inverted orientation, inparticular, would suggest that the entire embryo is pat-tomere at the 4- to 32-cell stage. As diagrammed in Fig. 1,

a second blastopore lip forms opposite that of the primary terned according to the LR polarity of the primary body axis(Fig. 3A). In contrast, a correct LR orientation (relative tobody axis at stage 10 (onset of gastrulation). However, dis-

placement of the site of injection slightly off the original the DV and AP axes of the induced twin) would mean thatthe ectopic organizer may have an innate asymmetry thatventral midline causes the twins to develop side-by-side

rather than belly-to-belly. Thus, the secondary axis can be is sufficient to determine LR polarity independent of theprimary body axis (Fig. 3C).positioned to either the left or the right side of the original

embryonic axis. We initially scored heart looping and Xnr-1 expressiononly in secondary axes that give rise to left twins becauseTable 1 shows that ventrally induced secondary axes can

randomize the cardiac situs of the primary body axis, caus- the preceding experiments showed that asymmetry in righttwins is perturbed by the presence of a conjoined left siblinging an approximately equal incidence of hearts that looped

in normal and inverted orientations. Importantly, however, (Table 1). Moreover, we analyzed the asymmetric develop-ment of secondary axes that were dorsoanteriorly completeheart looping is affected only when the ectopic axis is on

the left side of the embryo. For these embryos, the incidence and indistinguishable from the primary axis. Examiningonly complete secondary axes ensured that a normal orga-of normally and inversely looped hearts did not differ statis-

tically from random (P ú 0.05) for each mRNA except for nizer had been induced, minimizing the possibility thatheart looping would be randomized by organizer deficiencyXwnt-8 and Xnr-1, which yielded a slightly greater inci-

dence of inversions. In contrast, nearly 100% of embryos (Danos and Yost, 1995). Such ideal axes are best producedby molecules in the Wnt signaling cascade. Microinjectionwith ectopic body axes on the right side had normally

looped hearts. The results were essentially the same for of mRNA encoding Xwnt-8, or downstream mediators such


AID DB 8635 / 6x29$$$162 08-04-97 21:31:33 dba


TABLE 1Effect of Left or Right Secondary Body Axes on Heart Looping in Primary Axes

Incidence of normal heart looping in embryos with:

Left secondary axes Right secondary axes

Inducer Incidence (n) P Incidence (n)

Xwnt-8 35% (51) 0.050 100% (ú50)dnGSK-3 53% (30) 0.855 94% (31)dishevelled 36% (47) 0.080 100% (40)b-catenin 58% (33) 0.486 100% (ú30)Siamois 50% (30) 0.855 100% (35)noggin 57% (59) 0.298 100% (ú50)chordin 64% (39) 0.109 100% (ú30)Xnr-1 40% (108) 0.043 100% (62)Xnr-4 39% (82) 0.060 100% (52)b-Galactosidase (control mRNA) 100% (ú30) 100% (ú30)

Organizer graft 54% (28) 0.850 100% (17)Lateral MZ graft (control graft) 100% (24) Not determined

Note. Ectopic body axes were induced by injection of mRNA encoding the indicated molecules or by grafts of organizer tissue to theventral marginal zone at the onset of gastrulation (Materials and Methods). Left or right lateral mesoderm (MZ) was grafted as an operationcontrol. The incidence of normal heart looping in the primary body axis, in the presence of either a left or a right secondary axis, isindicated. The balance of the embryos developed with inverted looping. Probability (P) of a statistical deviation from a theoretical 50%incidence of normal heart looping (e.g., unbiased) was calculated by the x2 test with a Pearson continuity correction for low sample size(Materials and Methods). Note that when the ectopic axis is on the left, the incidence of looping was not statistically different from 50%(P ú 0.05) for any inducer except Xwnt-8 and Xnr-1, which caused fewer normal than inverted hearts. In contrast, normal heart orientationis observed in nearly 100% of all cases with the ectopic body axis on the right side.

as b-catenin and Siamois, is thought to mimic a signal trans- slightly greater than that of normal heart situs, but thisdifference was statistically significant only for embryos in-duction pathway that normally establishes the Nieuwkoop

center, an early signaling source in the dorsovegetal blasto- jected with Xwnt-8 (P õ 0.01) but not for embryos injectedwith b-catenin or Siamois mRNA (P ú 0.05). The reasonmeres (Smith and Harland, 1991). The Nieuwkoop center,

in turn, establishes the Spemann’s organizer at the equator for a tendency toward reversals is not known but mightreflect the dose of injected mRNA used to create the twins.(marginal zone), which orchestrates normal axis formation

(Gimlich, 1986). Consistent with this view, lower doses of these molecules(which yield only partial left twins) gave a greater incidenceOur results showed that induction of ectopic organizing

centers can correctly specify LR asymmetry. Xwnt-8, b-ca- of normal heart situs in the right primary axis (Table 1). Atendency toward reversals in the primary (right) sibling intenin, or Siamois mRNAs were injected into individual ven-

tral blastomeres. In each case, RLDx (MW 10,000) was co- these experiments does not alter the conclusion that ectopicactivation of a Wnt signaling cascade leads to normal mor-injected with the mRNA to distinguish the induced twin

from the primary body axis. The second hearts looped nor- phological and molecular asymmetry in the induced (left)body axes.mally in over 85% of Xwnt-8-, b-catenin-, or Siamois-in-

duced siblings (Table 2 and Fig. 4). This deviation from 50%(e.g., equal incidence of normal and inverted hearts) was

Asymmetric Development in Right Secondary Axesstatistically significant (P õ 0.01) for each mRNA, sug-Is Inducer-Dependentgesting that LR asymmetry was normally biased. In addi-

tion, in situ hybridization of stage 23–24 embryos showed The above results indicated that induced organizers arecapable of specifying normal LR asymmetry in left second-that Xnr-1 was correctly expressed on the left flank of the

induced twin, despite the fact that this tissue is derived ary axes. We next asked whether complete secondary axeswould also suffer the laterality defects expected when posi-from the original right side of the embryo (Fig. 5).

As expected from the results in the first section, the direc- tioned on the right. As above, the direction of heart loopingwas scored in secondary right twins that were dorsoanteri-tion of heart looping was frequently inverted in the primary

axes that formed the right twin in these experiments (Table orly complete and overtly indistinguishable from the pri-mary body axis. As shown in Table 3, Xwnt-8-induced right2 and Fig. 4). The incidence of inverted heart looping was


AID DB 8635 / 6x29$$$163 08-04-97 21:31:33 dba


We attempted to determine whether the direction of heartlooping in these induced right twins correlated with theretention of left-sided Xnr-1 expression. As expected, thepattern of Xnr-1 expression in the primary body axes (leftsibling) was unaffected in Xwnt-8-, b-catenin-, or Siamois-injected embryos (see Fig. 2C for example with Xwnt-8).However, Xnr-1 mRNA was undetectable in the majorityof right secondary axes created with any of these inducers.It should be noted, however, that unlike twins containingone partial axis, most with two complete axes (independentof whether they were induced by injection of Xwnt-8, b-catenin, or Siamois mRNA) are joined along the length oftheir flanks up to, and sometimes including, the visceralarches. Thus, although complete conjoined twins fre-quently have separate neural tubes, their intervening flankswere commonly fused. In rare twins that were well sepa-rated, we occasionally detected weak or diffuse Xnr-1 hy-bridization behind the left arches of right twins inducedwith b-catenin or Siamois (not shown), which develop nor-mal cardiac situs. Earlier asymmetric markers, expressedprior to neurula stages, have not yet been identified in Xeno-pus but would be necessary to conclude that aspects of LR

FIG. 1. Creation of conjoined twins derived from the dorsal (D)and ventral (V) hemispheres of the embryo. Axis-duplicatingmRNA is co-injected with a lineage tracer (RLDx, red) at the 4- to8- or 16- to 32-cell stage into the left (L) or right (R) ventral vegetalblastomere. This manipulation induces a second organizer approxi-mately 1807 opposite the original dorsal blastopore lip. Later mor-phogenetic movements cause the developing secondary axis (27) tobecome aligned to either the left or the right of the primary bodyaxis (17). The resultant tadpole stage twins are oriented side-by-side, possessing either a left or right secondary axis as indicated bythe position of the RLDx label.

twins had a nearly equal incidence of normal and invertedhearts. The incidence of normal and inverted looping didnot deviate significantly from 50% (P ú 0.05); thus, ectopic

FIG. 2. In situ hybridization showing Xnr-1 expression in con-Xwnt-8-induced axes, like primary body axes, become unbi-joined twins with secondary axes induced by ectopic Xwnt-8. (A)ased when they form the right sibling. Surprisingly, the ma-A partial left secondary axis suppresses Xnr-1 expression in thejority of right twins induced by b-catenin (90%) and Sia-adjoining left flank of the stage 23–24 primary axis. (B) Xnr-1 ex-mois (89%) developed with correctly oriented heart looping.pression in an uninjected control embryo, stage 23–24. (C) TheThe deviation from random (Põ 0.001) was unexpected andpresence of a complete right secondary axis does not affect Xnr-1

indicates that the orientation of LR asymmetry in body axes expression (arrow) in the left twin’s left lateral mesoderm (stageinduced by microinjection of b-catenin or Siamois is not 23–24). (D) Fluorescent micrograph of embryo in C, showing thataffected by the presence of a left sibling. As expected, the the right twin is the induced axis. Faint fluorescence in the flankleft primary body axes in all cases exhibited consistently is due to autofluorescence. Positions of left (L) and right (R) twins

are indicated.normal heart looping (Table 3).


AID DB 8635 / 6x29$$$163 08-04-97 21:31:33 dba


FIG. 3. Three predicted outcomes for the asymmetric development of ventrally induced secondary axes. If LR orientation is patternedglobally (depicted as solid shading on left side) and only with respect to the primary axis, then the asymmetry of the secondary axis willbe inverted (A). Alternatively, if the embryo’s LR information is restricted to the vicinity of the primary axis, and the induced organizeris unable to establish its own LR pattern, then the secondary axis will develop an unbiased, random asymmetry (B). In contrast, if eachorganizer independently determines LR pattern, then the secondary axis will develop with normal handedness (C). D, V, L, R, 17, and 27

are as in Fig. 1.

pattern other than heart situs are properly specified in the DISCUSSIONb-catenin- or Siamois-induced right twins. Although it isdifficult to conclusively define the state of Xnr-1 expression Rotation of the cortex following fertilization breaks thein our induced right twins, the morphological data nonethe- radial symmetry of the egg and establishes a dorsovegetalless suggest that LR asymmetry is differentially influenced signaling center known as the Nieuwkoop center (for re-

view, see Gerhart et al., 1991). Ventral injection of mRNAsby different members of the Wnt signaling pathway.

TABLE 2Incidence of Normal Heart Looping in Left Secondary Axes

Left secondary axis Right primary axis

Inducer Incidence (n) P Incidence (n) P

Xwnt-8 96% (27) õ0.001 22% (27) 0.007b-Catenin 85% (22) 0.001 32% (22) 0.136Siamois 86% (30) õ0.001 36% (30) 0.201

Note. Complete left secondary axes were induced by co-injection of synthetic mRNA encoding Xwnt-8, b-catenin, or Siamois withRLDx (see Materials and Methods). Each twinned embryo was examined for the presence of two complete axes with normal dorsoanteriorcharacter and the location of the RLDx label to confirm that the secondary axis gave rise to the left sibling. The incidence of normalheart looping in each axis is indicated. The balance developed with inverted looping. Probability (P) of a statistical deviation from atheoretical 50% incidence of normal heart looping was calculated as for Table 1 (see Materials and Methods). Note that left secondarybody axes induced by microinjected mRNA developed with predominantly normally looped hearts. A x2 test of independence comparingresults with Xwnt-8-, b-catenin-, and Siamois-induced twins indicated that the three inducers do not act differently (P § 0.39) in theseexperiments.


AID DB 8635 / 6x29$$$163 08-04-97 21:31:33 dba


FIG. 4. Cardiac situs in conjoined twins with a complete left sec-ondary axis. (A) Ventral view of heart region of Xwnt-8-inducedtwinned embryo at tadpole stage; the hearts of the right and theleft twins are indicated. The arrows trace the direction of loopingin each heart. In this example, the right twin (primary axis) exhibits FIG. 5. Xnr-1 expression in conjoined twins with complete leftinverted looping, while the left twin (induced axis) has a normally secondary axes. Secondary axes were induced ventrally with syn-looped heart (i.e., toward the embryo’s right). (B) Fluorescent micro- thetic mRNA encoding Xwnt-8 (A, B), b-catenin (C, D), or Siamoisgraph of embryo shown in A. The RLDx fluorescence identifies the (E, F) and fixed at stage 23–25. In situ hybridization was performedleft twin as the induced axis. with an antisense probe for Xnr-1. (A, C, E) Xnr-1 expression (pur-

ple) in the left twin’s lateral mesoderm is indicated by the arrow(anterior is toward top of figure, dorsal is to right). The more ante-rior Xnr-1 expression in embryos shown in C and E is normal for

encoding molecules normally expressed in the Nieuwkoop their slightly later stage (Lowe et al., 1996). (B, D, F) Fluorescentcenter, or in the subsequent blastula-stage Spemann orga- micrographs of dorsal views of the same embryos as in A, C, andnizer, duplicates the inductive properties of these regions E showing that the left twin is the induced axis. Position of leftand creates conjoined twins. We have shown that the sec- (L) and right (R) twins is indicated.


AID DB 8635 / 6x29$$$163 08-04-97 21:31:33 dba


TABLE 3 each other but shift during gastrulation and neurulationIncidence of Normal Heart Looping in Right Secondary Axes such that the conjoined twins lie side-by-side and are joined

at the trunk. Thus, while physically resembling SpemannLeft primary Right secondary and Falkenberg’s embryos, our twins differ in that the pri-

axis axis mary body axes are undivided and retain their original leftand right flanks. Nonetheless, the effect of twinning onInducer Incidence (n) Incidence (n) Porgan situs is the same as that of Spemann and Falkenberg:

Xwnt-8 100% (29) 45% (29) 0.700 a twin positioned on the left side leads to randomization ofb-Catenin 100% (31) 90% (31) õ0.001 heart looping in the right twin (Table 1). This result indi-Siamois 100% (19) 89% (19) õ0.001 cates that the loss of LR bias in the right twin is not caused

by a lack of positional information but, more likely, byNote. Complete right secondary axes were induced and scoredcross-signaling from the left body axis.as for Table 2, except that the secondary axes formed the right

The molecular character of the interfering signal fromsibling. The incidence of normal heart looping in each axis is indi-cated. The balance developed with inverted looping. Probability (P) the left twin in Xenopus is unknown, although it may beof a statistical deviation from a theoretical 50% incidence of nor- mediated by an activin-like factor. Activin is normally lo-mal heart looping was calculated as for Table 1 (see Materials and calized to the right side in chicks, and activin beads placedMethods). Note that polarity of heart situs in right secondary body on the left side are capable of suppressing Cnr-1 expressionaxes induced by microinjected Xwnt-8 was unbiased, whereas it (Levin et al., 1997). The stage when the interfering signalwas normally oriented when induced by b-catenin and Siamois. A may be transmitted in Xenopus is also not entirely clear.x2 test of independence confirmed that the incidence of inversion

However, a dorsal blastopore lip (Spemann’s organizer) graftin the Xwnt-8-induced axes was significantly different from thatat stage 10 can randomize the LR asymmetry of the recipi-seen with the other inducers (P õ 0.01).ent embryo (Table 1), indicating that susceptibility to LRperturbation extends into gastrulation.

It is important to note that randomization of heart loop-ing in our right primary twins correlates with loss of Xnr-ondary body axes induced by ectopic Xwnt-8 and down-

stream effectors b-catenin and Siamois develop appropri- 1 expression in the left lateral mesoderm. Unbiased heartlooping has previously been correlated with both bilateralately polarized LR asymmetry as measured by left-sided

expression of Xnr-1 and heart looping. These data suggest expression and absence of Xnr-1 in chick and mouse em-bryos (Levin et al., 1995; Lowe et al., 1996). Thus, Xnr-1that the normal orientation of the LR axis is locally deter-

mined as a consequence of the establishment of the Nieuw- does not seem to be necessary for heart looping to occur, butleft-sided expression is prognostic for a consistently normalkoop center and is likely to reflect endogenous Wnt activity

within the dorsovegetal region of the postfertilization em- orientation.bryo. Implications for the LR asymmetry in conjoined twinsand the potential role of Wnts in coordinating the LR, DV

Wnt Signaling and Patterning of LR Asymmetryand AP axes are discussed below.In amphibians, sperm entry triggers a rotation of the egg

cortex during the first cell cycle that breaks symmetry andAsymmetry in Conjoined Twins activates a determinant in the dorsal side of the zygote.

Members of the Xenopus Wnt family have been proposedLaterality defects in induced Xenopus conjoined twinsmost frequently occurred in the right twin, reminiscent of to be the dorsal determinants based on their ability to rescue

dorsal development of embryos ventralized by UV irradia-cases of organ reversals in spontaneous human and chickconjoined twins (Burn, 1991; M. Levin and C. Tabin, per- tion (Sokol et al., 1991). A maternally expressed Xenopus

Wnt, Xwnt-8b, may be the endogenous dorsovegetal activitysonal communication). The same phenomenon was notedby Spemann and Falkenberg in embryos that had been par- and is closely related to Xwnt-8 used in this work (Cui

et al., 1995). Several cytoplasmic components of the Wnttially bisected longitudinally (Spemann and Falkenberg,1919). Wilhelmi originally ascribed randomization of the pathway in Xenopus have been described, including dishev-

elled (Xdsh) (Sokol et al., 1995), b-catenin and plakoglobinright twin to a prior asymmetrical segregation of LR deter-minants such that the right twin lacked positional informa- (homologous to Drosophila armadillo) (Funayama et al.,

1995; Karnovsky and Klymkowsky, 1995), and glycogention (Wilhelmi, 1921). More recently, Hyatt et al. postulatedthat Vg-1, a member of the TGF-b growth factor superfam- synthase kinase 3 (GSK-3, homologous to shaggy/zeste-

white 3) (He et al., 1995; Pierce and Kimelman, 1995) asily, is selectively processed on the left side of the embryoand that only the left twin has adequate amounts of Vg- well as Siamois, which is a downstream target (Brannon and

Kimelman, 1996; Carnac et al., 1996). Ectopic activation1 signaling to initiate proper LR asymmetry (Hyatt et al.,1996). of the Wnt signal transduction cascade by expression of

dishevelled (Sokol et al., 1995), dominant-negative GSK-3In our experiments, the conjoined twins were producedby the induction of second body axes on the ventral side of (He et al., 1995; Pierce and Kimelman, 1995), b-catenin

(Guger and Gumbiner, 1995), or Siamois (Lemaire et al.,the embryos. The ectopic and primary axes initially face


AID DB 8635 / 6x29$$$163 08-04-97 21:31:33 dba


1995) in Xenopus vegetal cells mimics the normal induction the injected b-catenin and Siamois mRNA may overridethe normal susceptibility to randomization.of the Nieuwkoop center, which in turn activates the Spem-

ann organizer at the equator of the blastula stage embryo.Moreover, depletion of maternally encoded b-catenin pre-

The Developmental Origin of LR Asymmetryvents dorsal development and cannot be rescued by injec-tion of Wnt mRNA (Heasman et al., 1994). These results Our experiments suggest that activation of the Wntconstitute compelling evidence that signaling through b- pathway is sufficient to pattern LR asymmetry and orientcatenin and Siamois induces the Nieuwkoop center. it relative to the DV and AP axes. A key unanswered

We find that injection of Xwnt-8, b-catenin, or Siamois question is how Wnt activity is translated into LR pat-mRNA induces secondary axes with left-sided Xnr-1 expres- tern. Although cortical rotation creates DV asymmetrysion and hearts that loop normally (if the axes develop on during the first cell cycle and is thought to activate Wntthe left side of a conjoined twin). This implies that both signaling, we do not believe that cortical rotation itselfthe generation of asymmetry (leftward or rightward looping disrupts the intrinsic bilateral symmetry of the zygote.rather than an ambiguous or unlooped morphology) and its We find this possibility unlikely because the secondaryconsistent orientation (normal rightward looping rather body axes in our twins can develop normal LR asymmetrythan the inverse) are correctly specified following activation (Table 2) despite their initial orientation approximatelyof the Wnt signaling cascade. 1807 opposite the primary body axis.

Left-sided expression of Xnr-1 and normal heart looping Maternally encoded Wnts and Vg-1 are both candidatesare indicative of correct patterning of the left side of the for dorsal determinants activated by cortical rotation andembryo. For example, ablation of the left heart primordium, have also been implicated in early axial patterning in thebut not the right, causes frequent reversals in heart looping chick embryo (for instance, see Vize and Thomsen, 1994;(our unpublished observations and Hoyle et al., 1992). We Cui et al., 1995). It has recently been suggested that thesuspect, however, that the development of normal heart selective processing of Vg-1 on the left side of the Nieuw-situs also depends on correct patterning of the right side. koop center is an early step in patterning LR asymmetryTherefore, we predict that right side development (exempli- in Xenopus (Hyatt et al., 1996). The ability of opposingfied by liver and gall bladder situs) would also occur nor- secondary axes to develop with correct asymmetry in ourmally following activation of the Wnt signaling cascade in twins suggests that, if Vg-1 is necessary for this process,induced twins. However, the lack of right side markers in injected Wnt pathway molecules (including Siamois, whichXenopus and the common fusion of the twins in the gut is not normally expressed until after MBT) must also biasregion preclude direct verification of this prediction by our Vg-1 processing in ventral blastomeres. It has recently beentwinning experiments. suggested that synergy between Wnt activity and Vg-1 (or

We did not expect b-catenin- or Siamois-induced axes to a similar TGF-b molecule) is needed to specify dorsal meso-exhibit consistently normal heart looping when they give derm and endoderm, including Spemann’s organizer. There-rise to the right twin because both primary axes and Xwnt- fore, one possibility is that the induction of the Nieuwkoop8-induced twins are randomized by the presence of a left center leads to a left-sided bias in the processing of Vg-1sibling. This suggests that members of the Wnt signal trans- and that processed Vg-1 interacts with Wnt pathway mole-duction pathway differentially regulate LR asymmetry in cules in the organizer to correctly orient the LR axis.our induced axes. One possible explanation is that suscepti-bility to randomization is conferred by a factor or receptorpresent in primary and Xwnt-8-induced axes but lacking in ACKNOWLEDGMENTSaxes induced by b-catenin or Siamois. Consistent with thispossibility is the observation that the Wnt signal transduc- We thank Michael Raffin and Louise Evans for help with in sitution pathway diverges upstream of armadillo (b-catenin) in hybridization. In addition, we thank Melissa Grandeau, Michael

Levin, Kelly McLaughlin, and Jennifer Payne for helpful discussionsDrosophila. For example, dishevelled and shaggy/zeste-and for critical comments on the manuscript. M.M. gratefully ac-white 3 (GSK-3) antagonize Notch (Ruel et al., 1993; Axel-knowledges the support of the National Institutes of Health (RO1rod et al., 1996). In addition, Frizzled (a wingless receptor)HD28460), American Heart Association (94014850), and March of(Bhanot et al., 1996) and dishevelled, but not armadillo (b-Dimes Birth Defects Foundation (FY96-1071).catenin), are required for tissue polarity in imaginal disks

(Theisen et al., 1994; Krasnow et al., 1995). It is difficult todistinguish this model from two alternatives. First, down-

REFERENCESstream factors may establish LR polarity sooner than Xwnt-8, which must be secreted and interact with cell surface

Almirantis, Y. (1995). Left-right asymmetry in vertebrates. BioEs-receptors. In this scheme, LR polarity in the b-catenin- and says 17, 79–83.Siamois-induced axes must become irreversibly determined Axelrod, J. D., Matsuno, K., Artavanis-Tsakonas, S., and Perrimon,before the left axis is competent to perturb its pattern, thus N. (1996). Interaction between Wingless and Notch signalingrescuing the right twin from the normally deleterious ef- pathways mediated by dishevelled. Science 271, 1826–1832.

Bhanot, P., Brink, M., Samos, C. H., Hseieh, J. C., Wang, Y., andfects of being joined to a left sibling. Second, persistence of


AID DB 8635 / 6x29$$$163 08-04-97 21:31:33 dba


Macke, J. P. (1996). A new member of the frizzled family from Karnovsky, A., and Klymkowsky, M. W. (1995). Anterior axis dupli-cation in Xenopus induced by the overexpression of the cadherin-Drosophila functions as a Wingless receptor. Nature 382, 225–

230. binding protein plakoglobin. Proc. Natl. Acad. Sci. 92, 4522–4526.Brannon, M., and Kimelman, D. (1996). Activation of Siamois by

the Wnt pathway. Dev. Biol. 180, 344–347. Klar, A. J. S. (1994). A model for specification of the left-right axisin vertebrates. Trends Genet. 10, 292–295.Brown, N. A., and Wolpert, L. (1990). The development of handed-

ness in left/right asymmetry. Development 109, 1–9. Krasnow, R. E., Wong, L. L., and Adler, P. N. (1995). dishevelled isa component of the frizzled signalling pathway in Drosophila.Burn, J. (1991). Disturbance of morphological laterality in humans.Development 121, 4095–4102.Ciba Found. Symp. 162, 282–296.

Lemaire, P., Garrett, N., and Gurdon, J. B. (1995). Expression clon-Carnac, G., Kodjabachian, L., Gurdon, J. B., and Lemaire, P. (1996).ing of Siamois, a Xenopus homeobox gene expressed in dorsal-The homeobox gene Siamois is a target of the Wnt dorsalizationvegetal cells of blastulae and able to induce a complete secondarypathway and triggers organizer activity in the absence of meso-axis. Cell 81, 85–94.derm. Development 122, 3055–3065.

Levin, M., Johnson, R. L., Stern, C. D., Kuehn, M., and Tabin, C.Collignon, J., Varlet, I., and Robertson, E. J. (1996). Relationship(1995). A molecular pathway determining left-right asymmetrybetween asymmetric nodal expression and the direction of em-in chick embryogenesis. Cell 82, 803–814.bryonic turning. Nature 381, 155–158.

Levin, M., Pagan, S., Roberts, D. J., Cooke, J., Kuehn, M., and Tabin,Corballis, M. C. (1983). ‘‘Human Laterality. Perspectives in Neuro-C. J. (1997). Different aspects of laterality are independently con-linguistics, Neuropsychology, and Psycholinguistics.’’ Academictrolled by an apparently streak-autonomous signaling pathwayPress, New York.initiated by activin. Dev. Biol., in press.Corballis, M. C., and Morgan, M. (1978). On the biological basis of

Levin, M., Roberts, D. J., Holmes, L. B., and Tabin, C. (1996). Later-human laterality: I. Evidence for a maturational left-right gradi-ality defects in conjoined twins. Nature 384, 321.ent. Behav. Brain Sci. 2, 261–336.

Lowe, L. A., Supp, D. M., Sampath, K., Yokoyama, T., Wright,Cui, Y., Brown, J., Moon, R. T., and Christian, J. L. (1995). Xwnt-C. V. E., Potter, S. S., Overbeek, P., and Kuehn, M. R. (1996). Con-8b: A maternally expressed Xenopus Wnt gene with a potentialserved left-right asymmetry of nodal expression and alterationsrole in establishing the dorsoventral axis. Development 121,in murine situs inversus. Nature 381, 158–161.2177–2186.

Meno, C., Saijoh, Y., Fujii, H., Ikeda, M., Yokoyama, T., Yokoyam,Danos, M. C., and Yost, H. J. (1995). Linkage of cardiac left-rightM., Toyoda, Y., and Hamada, H. (1996). Left-right asymmetricasymmetry and dorsal-anterior development in Xenopus. Devel-expression of the TGFb-family member lefty in mouse embryos.opment 121, 1467–1474.Nature 381, 151–155.Danos, M. C., and Yost, H. J. (1996). Role of notochord in specifica-

Morgan, M. (1977). Embryology and Inheritance of Asymmetry. Intion of cardiac left-right orientation in zebrafish and Xenopus.‘‘Lateralization in the Nervous System (S. Harnad, Ed.), pp. 173–Dev. Biol. 177, 96–103.194. Academic Press, New York.Funayama, N., Fagotto, F., McCrea, P., and Gumbiner, B. (1995).

Morgan, M., and Corballis, M. C. (1978). On the biological basis ofEmbryonic axis induction by the armadillo repeat domain of b-human laterality: II. The mechanisms of inheritance. Behav.catenin: Evidence for intracellular signaling. J. Cell Biol. 128,Brain Sci. 2, 270–277.959–968.

Morgan, M. J. (1991). The asymmetrical genetic determination ofGerhart, J. C., Stewart, R., and Doniach, T. (1991). Organizing thelaterality: Flatfish, frogs and human handedness. Ciba Found.Xenopus organizer. In ‘‘Gastrulation: Movements, Patterns andSymp. 162, 234–247.Molecules’’ (R. Keller, W. Clark, Jr., and F. Griffin, Eds.), pp. 57–

Peng, H. B. (1991). Solutions and Protocols. In ‘‘Xenopus laevis:77. Plenum, New York.Practical Uses in Cell and Molecular Biology (B. K. Kay and H. B.Gimlich, R. L. (1986). Early cellular interactions promote embry-Peng, Eds.), pp. 657–662. Academic Press, San Diego.onic axis formation in Xenopus laevis. Dev. Biol. 115, 340–352.

Pierce, S., and Kimelman, D. (1995). Regulation of Spemann orga-Guger, K. A., and Gumbiner, B. M. (1995). b-Catenin has Wnt-likenizer formation by the intracellular kinase Xgsk-3. Developmentactivity and mimics the Nieuwkoop signaling center in Xenopus121, 755–765.dorsal–ventral patterning. Dev. Biol. 172, 115–125.

Psychoyos, D., and Stern, C. D. (1996). Restoration of the organizerHarland, R. M. (1991). In situ hybridization: An improved wholeafter radical ablation of Hensen’s node and the anterior primitivemount method for Xenopus embryos. In ‘‘Xenopus laevis: Practi-streak in the chick embryo. Development 122, 3263–3273.cal Uses in Cell and Molecular Biology’’ (B. K. Kay and H. B.

Ruel, L., Bourouis, M., Heitzler, P., Pantesco, V., and Simpson, P.Peng, Eds.), pp. 685–695. Academic Press, San Diego.(1993). Drosophila shaggy kinase and rat glycogen synthase ki-He, X., Saint-Jeannet, J.-P., Woodgett, J., Varmus, H., and Dawid,nase-3 have conserved activities and act downstream of Notch.I. (1995). Glycogen synthase kinase-3 and dorsoventral patterningNature 362, 557–560.in Xenopus embryos. Nature 374, 617–622.

Sive, H. L. (1993). The frog prince-ss: A molecular formula for dorso-Heasman, J., Crawford, A., Goldstone, K., Garner-Hamrick, P.,ventral patterning in Xenopus. Genes Dev. 7, 1–12.Gumbiner, B. M., Kintner, C., Yoshido-Noro, C., and Wylie, C.

Slack, J. M. W. (1993). Embryonic induction. Mech. Dev. 41, 91–(1994). Overexpression of cadherins, and underexpression of b-107.catenin inhibit dorsal mesoderm induction in early Xenopus em-

bryos. Cell 79. Smith, J. C. (1993). Mesoderm-inducing factors in early vertebratedevelopment. EMBO J. 12, 4463–4470.Hoyle, C., Brown, N. A., and Wolpert, L. (1992). Development of

left/right handedness in the chick heart. Development 115, Smith, W. C., and Harland, R. M. (1991). Injected Xwnt-8 RNA actsearly in Xenopus embryos to promote formation of a vegetal1071–1078.dorsalizing center. Cell 67, 753–765.Hyatt, B. A., Lohr, J. L., and Yost, H. J. (1996). Initiation of left-right

axis formation by maternal Vg1. Nature 384, 62–65. Sokal, R. R., and Rohlf, F. J. (1995). ‘‘Biometry: The Principles and


AID DB 8635 / 6x29$$$164 08-04-97 21:31:33 dba


Practice of Statistics in Biological Research.’’ Freeman, New nalling to establish both cell polarity and cell identity. Develop-ment 120, 347–360.York.

Sokol, S., Christian, J. L., Moon, R. T., and Melton, D. A. (1991). Vize, P. D., and Thomsen, G. H. (1994). Vg-1 and regional specifica-tion in vertebrates: A new role for an old molecule. Trends Genet.Injected Wnt RNA induces a complete body axis in Xenopus

embryos. Cell 67, 741–752. 10, 371–376.Sokol, S. Y., Klingensmith, J., Perrimon, N., and Itoh, K. (1995). von Woellwarth, C. (1950). Experimentelle Untersuchungen uber

Dorsalizing and neuralizing properties of Xdsh, a maternally ex- den Situs inversus der Eingeweide und der Habenula despressed Xenopus homolog of dishevelled. Development 121, Zwischenhirns bei Amphibien. Wilhelm Roux’ Arch. En-3487–3498. twicklungsmech. Org. 144, 178–256.

Spemann, H., and Falkenberg, H. (1919). Uber Asymmetrische Wilhelmi, H. (1921). Experimentelle Untersuchungen uber SitusEntwicklung und Situs inversus viscerum bei Zwillingen und inversus viscerum. Wilhelm Roux’ Arch. Entwicklungsmech.Doppelbildungen. Wilhelm Roux’ Arch. Entwicklungsmech. Org. 48, 517–532.Org. 45, 371–422.

Received for publication April 7, 1997Theisen, H., Purcell, J., Bennet, M., Kansagara, D., Syed, A., andMarsh, J. L. (1994). Dishevelled is required during Wingless sig- Accepted May 20, 1997


AID DB 8635 / 6x29$$$164 08-04-97 21:31:33 dba

Documents

Organizer Induction Determines Left-Right Asymmetry in Xenopus · 2017. 2. 25. · In the simplest view, this hypothesis would suggest Harland, 1991; Sokol et al., 1991), suggesting