Developmental Cell, Vol. 8, 75–84, January, 2005, Copyright 2005 by Elsevier Inc.DOI 10.1016/j.devcel.2004.11.013
Tbx5 and Tbx4 Are Not Sufficient to Determine
Limb-Specific Morphologies but Have
Common Roles in Initiating Limb Outgrowth
prospective hindlimb region but not in the developing
forelimb (Lamonerie et al., 1996; Logan et al., 1998;
Shang et al., 1997). The limb-type restricted expression
opment. Misexpression experiments in the chick have
of limb-specific morphologies. Ectopic expression of
Tbx5 in the chick leg bud can induce a partial leg-to-
in the chick wing bud is able to cause a partial wing-
to-leg transformation (Rodriguez-Esteban et al., 1999;
Takeuchi et al., 1999). Similarly, misexpression of Pitx1
in the wing bud leads to the development of limbs with
leg-like characteristics (Logan and Tabin, 1999; Szeto
et al., 1999; Takeuchi et al., 1999). Accordingly, in Pitx1
mutant mice, hindlimbs show a loss of hindlimb charac-
teristics (Lanctot et al., 1999; Szeto et al., 1999).
Gene deletion and knockdown experiments have
shown that Tbx5, Tbx4, and Pitx1 are required for the
initiation and/or maintenance of limb bud outgrowth.
Functional knockdown of zebrafish tbx5 results in a fail-
ure to initiate pectoral fin bud formation (Ahn et al.,
2002). Similarly, all skeletal elements of the forelimb are
missing in a limb-restricted Tbx5 knockout (Rallis et al.,
2003). One of the earliest molecular read-outs of limb
initiation is the expression of Fgf10 in the prospective
limb fields (Min et al., 1998; Sekine et al., 1999). When
Tbx5 is inactivated, Fgf10 is never expressed in the
prospective forelimb region (Agarwal et al., 2003). Tbx5
is therefore required for the induction of Fgf10 in the
LPM at prelimb bud stages, which leads to forelimb
bud initiation. In Tbx4?/?embryos, induction and initial
patterning of the hindlimb appears normal, but it fails
to develop further, and Fgf10 expression is not main-
tained in the hindlimb bud mesenchyme (Naiche and
Papaioannou,2003). Pitx1mutantmouse hindlimbsalso
display an outgrowth defect, although much less severe
than that observed in either Tbx5 or Tbx4 mutants.
Pitx1?/?hindlimbs are smaller than wild-type, yet the
skeletal elements, with the exception of the ilium, are
present (Lanctot et al., 1999; Szeto et al., 1999).
We have used loxP/Cre technology in combination
with transgenic methods in the mouse to disrupt and
attempting to rescue the no-forelimb phenotype of the
Tbx5 limb-restricted knockout (Rallis et al., 2003) by
expressing either Tbx4 or Pitx1, or both genes simulta-
tion has been specifically deleted. This genetic assay,
which we refer to as the limb-rescue assay, allows us
to test the properties of these factors in two processes
of limb development: (1) initiation of limb outgrowth and
(2) specification of limb-specific morphologies. We
show that Tbx4 can replace the function of Tbx5 and
rescue limb outgrowth, whereas Pitx1 cannot. In contrast
to previous chick misexpression studies, Tbx4-rescued
limbs have a forelimb-like phenotype, suggesting that
Tbx4 alone is not able to dictate hindlimb-specific mor-
phology and that forelimb characteristics can develop
in the absence of Tbx5. To determine whether Pitx1 can
Carolina Minguillon, Jo Del Buono,
and Malcolm P. Logan*
Division of Developmental Biology
National Institute for Medical Research
London NW7 1AA
Morphological differences between forelimbs and
hindlimbs are thought to be regulated by Tbx5 ex-
pressed in the forelimb and Tbx4 and Pitx1 expressed
in the hindlimb. Gene deletion and misexpression ex-
periments havesuggested thatthese factorshave two
distinct functions during limb development: the initia-
tion and/or maintenance of limb outgrowth and the
netic methods in the mouse, we have investigated the
roles of Tbx5, Tbx4, and Pitx1 in both processes. Our
results support a role for Tbx5 and Tbx4, but not for
Pitx1, in initiation of limb outgrowth. In contrast to
conclusions from gene misexpression experiments in
the chick, our results demonstrate that Tbx5 and Tbx4
do not determine limb-specific morphologies. How-
ever, our results support a role for Pitx1 in the specifi-
cation of hindlimb-specific morphology. We propose
a model in which positional codes, such as Pitx1 and
Hox genes in the lateral plate mesoderm, dictate limb-
Vertebrate forelimbs and hindlimbs are serially homolo-
are derived are patterned by common signals during
embryonic development, they ultimately form morpho-
logically distinct structures. A question that arises is
how cells exposed to common signals can respond dif-
ferentially and give rise to distinct morphologies.
Vertebrate forelimbs and hindlimbs arise from regions
of the lateral plate mesoderm (LPM) at defined locations
along the rostral-caudal axis of the embryo. Trans-
plantation experiments in the chick have demonstrated
that limb-type specification, the process by which cells
to form either forelimb or hindlimb, occurs prior to the
initiation limb bud outgrowth (reviewed in Logan, 2003;
Saito et al., 2002; Stephens et al., 1989).
Three genes have been identified that fulfill many of
the criteria to be candidates to specify limb-type iden-
tity. Two T-box transcription factors, Tbx5 and Tbx4,
are expressed in the LPM of either the prospective fore-
limb or hindlimb region, respectively (Chapman et al.,
1996; Gibson-Brown et al., 1996). In addition, a paired-
related homeodomain factor, Pitx1, is expressed in the
Figure 1. Transgenic Mouse Lines Express-
ing Tbx4 or Pitx1 in the Limbs
(A) Schematic of the Prx1-Tbx4 and Prx1-
Pitx1 transgenic constructs. The ORF frag-
ments (blue box) contain the HA epitope
(black box) as a 3? fusion. The numbers in
brackets denote the size of each fragment
(B–E) Whole-mount in situ hybridization for
transgene-derived expression in the devel-
oping limbs at E10.5. Expression is also de-
tected in the head and flank in all lines. Tbx4
expression in Prx1-Tbx4(4.1) (B) and Prx1-
Tbx4(4.48) (C). Pitx1 expression in Prx1-
Pitx1(P1.1) (D) and Prx1-Pitx1(P1.2) (E). Lat-
eral views are shown.
(F) Western blot analysis of protein extracts
from transgenic lines. Transgene-derived
protein is detected by virtue of the HA tag.
In this exposure, protein levels in the Prx1-
Tbx4(4.1) are not detectable. Robust levels of
protein are detected in Prx1-Tbx4(4.48) and
Prx1-Pitx1(P1.1) lines. Lower levels of protein
are detected in the Prx1-Pitx1(P1.2) line.
transform a forelimb to a more hindlimb-like character,
we analyzed the morphology of Pitx1 transgenic fore-
limbs (expressing endogenous Tbx5), and we rescued
the no-forelimb phenotype of the Tbx5 limb-restricted
knockout by supplying both Tbx4 and Pitx1 simultane-
ously. In both cases, we observed a partial forelimb-to-
gies but that other factors may also be required.
Tbx4, but Not Pitx1, Can Rescue Limb Outgrowth
in the Absence of Tbx5
Tbx5 is required for initiation and outgrowth of the fore-
limb (Agarwal et al., 2003; Ahn et al., 2002; Rallis et al.,
2003). Using a conditional allele of Tbx5 and a Prx1-Cre
deleter transgenic line, we have previously shown that
in the absence of Tbx5 function, the forelimb fails to
form (Figure 2B; Rallis et al., 2003). To investigate
whether Tbx4 or Pitx1 are capable of replacing the func-
tion of Tbx5 in the forelimb, we crossed the Tbx4 and
Pitx1 transgenic lines into the genetic background of
the conditional deletion of Tbx5 in the limb (Tbx5lox/lox;
Prx1-Cre). Heterozygote Tbx5 mice (Tbx5lox/?;Prx1-Cre),
which form normal forelimbs (Rallis et al., 2003), serve
as controls (Figure 2A). One of the Tbx4 lines (4.48) was
capable of rescuingthe
Tbx5lox/lox;Prx1-Cre mice (Figure 2C), and a limb formed
in the forelimb region. This demonstrates that Tbx4 can
replace the function of Tbx5 in limb outgrowth. The
Tbx4(4.1) line, which expresses much lower levels of
Tbx4protein (Figure1F),was notableto rescueforelimb
development (Figure 2D), suggesting that insufficient
of the Pitx1-expressing lines were able to rescue limb
outgrowth in the absence of Tbx5 (Figures 2E and 2F),
than Tbx4 in the Tbx4(4.48) line (Figure 1F). These data
demonstrate that Tbx4, but not Pitx1, can replace the
function of Tbx5 in controlling limb outgrowth.
Generation and Characterization
of Transgenic Lines
Conditional deletion of Tbx5 in the developing limbs
leads to the complete absence of all forelimb elements
(Rallis et al., 2003). We have exploited this genetic back-
ground and developed an assay to test the ability of
Tbx4 and Pitx1 to rescue the forelimb defect that results
from the absence of Tbx5 function.
To distinguish transgene-derived expression of Tbx4
and Pitx1 from endogenous expression, the chick
cDNAs of each gene were placed under the regulation
of the Prx1 regulatory element (Martin and Olson, 2000).
To enable detection of transgene-derived protein, the
cDNAs were tagged with the HA epitope (Figure 1A).
Two independent lines were generated for both Tbx4
and Pitx1 and denoted 4.1 and 4.48 and P1.1 and P1.2,
respectively. In all four lines, the transgene is expressed
in the hindlimb and forelimbs as well as the cranial mes-
et al., 2002). Transgene expression in the limbs corre-
sponds to overexpression of either Tbx4 or Pitx1 in the
hindlimbs and ectopic expression of these genes in the
forelimbs. Western blot analyses and detection with an
anti-HA antibody showed differences in the levels of
protein expression. The Prx1-Tbx4(4.48) and Prx1-
Pitx1(P1.1) lines express higher levels of protein than
The Tbx4-Rescued Limb Buds
Are Normally Patterned
During limb development, a positive feedback loop be-
tween Fgf8, expressed in the apical ectodermal ridge
(AER), and Fgf10, expressed in the mesenchyme, is es-
sential for proximodistal outgrowth of the limb bud (re-
viewed in Martin, 1998). Shh expression in cells of the
zone of polarizing activity (ZPA) in the posterior limb
mesenchyme is required for the precise anterior-poste-
rior patterning of the limb (Riddle et al., 1993). Expres-
sion of Fgf8, Shh, and Fgf10 in Prx1-Tbx4(4.48)-rescued
Outgrowth and Identity of the Vertebrate Limb
model using genetic techniques in the mouse, we ana-
lyzed the limb-type identity of the Prx1-Tbx4(4.48)-res-
cued limb buds and limbs. In addition to Tbx5, Tbx4,
and Pitx1, certain Hox genes belonging to the HoxC
cluster have a limb-type restricted expression pattern
and have been implicated in limb-type specification
(Nelson et al., 1996; Peterson et al., 1994). Hoxc4 and
Hoxc5 are expressed in the forelimb, while Hoxc9,
Hoxc10, Hoxc11, Hoxc12, and Hoxc13 are expressed in
the hindlimb. Surprisingly, the Tbx4-rescued limb has
a forelimb-type pattern of gene expression. Following
deletion of Tbx5 and replacement with Tbx4, transcripts
from the endogenous Tbx5 conditional allele that has
been disrupted by Cre recombinase activity are still de-
tected using a probe that recognizes sequences still
ure 4G). Hence, signals that normally restrict Tbx5 ex-
pression to the forelimb (Figure 4A) are still functioning.
limb buds (Figures 4B and 4C), are also expressed in
Prx1-Tbx4(4.48)-rescued limb buds (Figures 4H and 4I).
Conversely, genes normally restricted to the hindlimb
are not ectopically expressed in the Prx1-Tbx4(4.48)-
rescued limbs. Tbx4, Pitx1, and Hoxc10 are expressed
in the hindlimbs but not in the Prx1-Tbx4(4.48)-rescued
limb buds (Figures 4J–4L), as seen in control littermates
patterns characteristic of a forelimb in the rescued limb.
To further analyze the identity of the Prx1-Tbx4(4.48)-
rescued limbs, we examined the skeletal morphology in
newborn pups and compared them to forelimbs and
hindlimbs from control littermates. The forelimb-type
character of the rescued limb was evident (compare
features are noticeable: the presence of a scapula, the
relative length of the stylopod and zeugopod bones,
and the joint between stylopod and zeugopodal ele-
ments. The scapula of Tbx4-rescued newborn pups is
indistinguishable from that found in control littermates
pod articulate to form an elbow-like joint with the distal
end of the humerus-like bone sitting in an apparent
trochlear notch at the proximal end of the ulna-like bone
(Figure 4N, arrowhead). In addition, the relative size of
the stylopod and zeugopodal bones is similar. This is
most comparable to the arrangement in the forelimb,
where the humerus is of equivalent length to the radius
andulnabones whileinthehindlimbthe femurissmaller
than the tibia. Furthermore, both zeugopodal bones in
the rescued limb are of similar length, resembling the
radius and ulna in forelimbs and not the tibia and fibula
in hindlimbs. Although overall the skeletal morphology
of the Tbx4-rescued limb is remarkably similar to that
of the forelimb, it is not completely identical. The deltoid
tuberosity of the humerus is absent, and the flexure of
the wrist is altered such that the hand extends directly
from the wrist and fails to turn inward. In summary, both
analyses of Tbx4-rescued limbs with limb-type re-
stricted markers at limb bud stages and examination of
the limb skeletal morphology in newborn pups demon-
strate their forelimb-like phenotype, refuting the postu-
lated role for Tbx4 in specifying hindlimb identity.
To determine whether these contradictory mouse/
chick results were due to differences in the expression
Figure 2. Transgene-Derived Tbx4, but Not Pitx1, Can Rescue the
Limb Defect in the Conditional Deletion of Tbx5
All panels show lateral views of E17.5 mouse embryos with the
exception of (E), which is P0.
(A) No limb defect is observed following deletion of one copy of
Tbx5 in the limbs.
(B) No forelimbs form following conditional deletion of Tbx5 in the
(C) In the absence of Tbx5, limb formation is rescued by transgene-
derived Tbx4 using the Prx1-Tbx4(4.48) line (arrow).
(D) Limb formation is not rescued by the Prx1-Tbx4(4.1) trans-
(E and F) In the absence of Tbx5, limb formation is not rescued by
Pitx1 from either the Prx1-Pitx1(P1.1) (E) or the Prx1-Pitx1(P1.2)
limb buds is identical to that in control littermates
(Tbx5lox/?;Prx1-Cre) (compare Figures 3A with 3D for
Fgf8, 3B with 3E for Shh, and 3C with 3F for Fgf10).
This demonstrates that in Prx1-Tbx4(4.48)-rescued limb
buds, key signaling centers essential for normal limb
development are established and appear to function
normally, consistent with our observation that limb out-
growth is rescued at E17.5 (Figure 2C). The failure of
Pitx1 and low doses of Tbx4 protein to sustain limb
development in Tbx5lox/lox;Prx1-Cre mice was confirmed
in E10.5 embryos. Fgf8 expression is absent in the fore-
limb-forming region following attempted rescue with ei-
Pitx1(P1.2) lines (Figures 3G–3I). Similarly, although ex-
pression of Fgf10 can be detected in the hindlimbs in
it is absent from the forelimb region at late E9.5 stage.
This demonstrates that Tbx4, but not Pitx1, can initiate
and maintain limb outgrowth in the absence of Tbx5.
Forelimb-like Identity of the Tbx4-Rescued Limbs
suggested a role for Tbx5 and Tbx4 in specification of
forelimb and hindlimb identity, respectively (Rodriguez-
Esteban et al., 1999; Takeuchi et al., 1999). To test this
Figure 3. Signaling Centers in the Limb Are
Established Normally in Tbx4-Rescued Limb
(A–F) Tbx5lox/?;Prx1-Cre control littermates
(A–C) and Tbx5lox/lox;Prx1-Cre;Prx1-Tbx4(4.48)
embryos (D–F) at E10.5. All are lateral views,
except (C) and (F), which are dorsal views.
Fgf8 is expressed in the AER (A), Shh in cells
of the ZPA (B), and Fgf10 in the limb mesen-
chyme (C). Fgf8 (D), Shh (E), and Fgf10 (F)
are all expressed normally in Tbx5lox/lox;Prx1-
(G–I) Fgf8 is not expressed in the forelimb-
forming region (indicated with an asterisk) in
embryos in which transgene-derived Tbx4 or
Pitx1 has failed to rescue the limb defect fol-
lowing deletion of Tbx5: (G) Prx1-Tbx4(4.1),
(H) Prx1-Pitx1(P1.1), (I) Prx1-Pitx1(P1.2).
(J) Fgf10 is expressed in the hindlimb bud
but is not expressed in the forelimb-forming
hasfailed torescuelimb outgrowth(asterisk).
FL, forelimb; HL, hindlimb; RL, rescued limb.
levels of Tbx4, we doubled the dose of the Tbx4 trans-
gene by crossing the line to homozygosity in the
normal limb morphology is severely affected, forelimb-
like characteristics can be still detected. A trochlear
notch and an olecranon process in the ulna-like bone
are clearly identifiable (Figures 4P and 4Q).
bryos show that Tbx4 is still able to rescue limb out-
growth even when the cells in the prospective forelimb
field have never expressed endogenous Tbx5 (Figure
5A). Consistent with our results using the Prx1-Cre de-
like gene expression pattern. The normally forelimb-
restricted genes Tbx5, Hoxc4, and Hoxc5 are expressed
in the Tbx4-rescued limb in a pattern indistinguishable
from wild-type (Figures 5B–5D), while the normally hind-
limb-restricted markers Tbx4, Pitx1, and Hoxc10 (Fig-
demonstrate that Tbx4 is able to rescue limb outgrowth
in cells in the forelimb-forming region that have never
been exposed to Tbx5 activity and that the resultant
limb expresses genes normally restricted to the fore-
limb. Furthermore, genes normally restricted to the
that an initial pulse of Tbx5 expression is not able to
determine forelimb-specific morphologies.
Forelimb-like Identity of Tbx4-Rescued Limbs
after Ubiquitous Deletion of Tbx5
In the Prx1-Cre deleter line, Cre activity is first detected
at E9.0–E9.5 (Logan et al., 2002). However, Tbx5 tran-
scripts are first detected at E8.5 (Agarwal et al., 2003).
From our previous work (Rallis et al., 2003) and the
results presented here for the 4.1, P1.1, and P1.2 lines
that were unable to rescue limb outgrowth (Figure 2),
we have demonstrated that this transient expression of
Tbx5 is not sufficient to initiate forelimb development.
However, this observation raises the interesting possi-
bility that a short pulse of endogenous Tbx5 transcript
is sufficient to determine forelimb-specific morphology
such that, following deletion of Tbx5 and replacement
with Tbx4, a forelimb develops. To address this issue,
we used the ?-actin-Cre transgenic line (Lewandoski
and Martin, 1997) to disrupt Tbx5 gene function ubiqui-
tously in the early embryo and tested the ability of the
Prx1-Tbx4(4.48) line to rescue limb outgrowth. Although
these embryos die around E10.0 due to heart defects
(Bruneau et al., 2001), they survive long enough to de-
termine whether the appropriate set of normally limb-
type restricted genes are expressed in these Prx1-
Pitx1 Can Partially Transform Forelimb
to Hindlimb-like Morphologies
Pitx1 is also expressed in a hindlimb-restricted manner
and has been implicated in specifying hindlimb-specific
morphologies (Lanctot et al., 1999; Logan and Tabin,
1999; Szeto et al., 1999; Takeuchi et al., 1999). Pitx1-
expressing transgeniclines were notable torescue limb
outgrowth followingdeletion ofTbx5. Wewere therefore
unable to address the ability of this gene to influence
limb-type identity in the absence of Tbx5. Instead, we
have used our transgenic reagents to test the ability of
Pitx1 to transform the forelimb-like morphology of the
Outgrowth and Identity of the Vertebrate Limb
Figure 4. Tbx4-Rescued Limbs Have Forelimb Characteristics
Tbx5, Hoxc4, and Hoxc5 are normally expressed in the forelimb but not the hindlimb of control littermates (A–C). They are also expressed in
Tbx4-rescued limbs (G–I). Tbx4, Pitx1, and Hoxc10 are normally expressed in the hindlimb but not the forelimb of E10.5 control embryos (D–F).
These genes are not ectopically expressed in the rescued limb at E10.5 (J–L). All the embryos are E10.5 and shown in lateral views except
(D), (F), (J), and (L), which are ventral views, and (C) and (I), which are dorsal views. Alcian blue/Alizarin red staining of the control forelimb
(M), Tbx4-rescued limb (N), and control hindlimb (O) skeletal elements at P0. Alcian blue/Alizarin red staining of a rescued limb (E16.5) in
which the Tbx4 transgene is present to homozygosity (P). An outline of the skeletal elements of the limb shown in (P) illustrates that although
abnormal in shape, the articulation is elbow-like (Q). au, autopod; Dt, deltoid tuberosity; F, femur; Fi, fibula; FL, forelimb; H, humerus; HL,
hindlimb; op, olecranon process; R, radius; RL, rescued limb; Sc, scapula; st, stylopod; T, tibia; Tn, trochlear notch; U, ulna; zg, zeugopod.
Tbx4-rescued limb and the wild-type forelimb express-
ing endogenous Tbx5.
We generated double-rescued embryos (Tbx5lox/lox;
Prx1-Cre;Prx1-Tbx4(4.48);Prx1-Pitx1(P1.1)) and Prx1-
elements to forelimbs and hindlimbs of control lit-
termates (Figure6) and tothe Tbx4-rescuedlimb (Figure
4N). Following deletion of Tbx5 in cells of the forelimb-
forming region and replacement with Tbx4 and Pitx1,
the limb element that forms shares many morphological
characteristics with a normal hindlimb. The articulation
between the stylopod and zeugopod skeletal elements
is strikingly knee-like (Figure 6B, arrowhead), while be-
tween the zeugopod and autopod it is ankle-like (Figure
6B,double arrowhead)comparedtothe normalforelimb
(Figure 6A). The heads of the stylopod and zeugopod
bones in the double rescue limb have a head-to-head
apposition and the heads of each bone are larger and
broadened, as is found in the knee (Figure 6F, dashed
line). Moreover, the double-rescued limb zeugopodal
similar to that observed in the tibia (Figure 6H, arrow)
head-to-head appositionand an extendedtuberosity on
the zeugopodal bone is present in the Prx1-Pitx1(P1.1)
transgenic forelimb (Figure 6G, dashed line and arrow,
respectively). In addition, at this articulation in the Prx1-
Pitx1(P1.1) transgenic forelimb, the olecranon process
and trochlear notch normally present in the ulna (Figure
in the hindlimb are longer than digits in the forelimbs. In
the double-rescued and the Prx1-Pitx1(P1.1) transgenic
limbs, digit lengths are increased and therefore more
similar to those in the hindlimb than the forelimb. To
quantify these differences, we compared the ratio of the
lengths of the second digit metacarpal/metatarsal and
first phalange (yellow continuous versus dashed line in
Figures 6I–6L). The ratio of metacarpal:phalange length
is lower than 2 in the forelimb (Figure 6I, ratio 1.6), while
in the hindlimb the second metatarsal is longer than
twice the length of the first phalange (Figure 6L, ratio
2.5). In the double-rescued limb, this ratio is also above
2 (Figure 6J, ratio 2.6). Similarly, in the Prx1-Pitx1(P1.1)
transgenic autopod, the ratio is also above 2 (Figure 6K,
ratio 2.25). The length of the anterior zeugopodal bone
in the double-rescued and Prx1-Pitx1(P1.1) transgenic
limbs is longer than the stylopodal bone, resembling the
difference in femur/tibia length in the hindlimb rather
than the similar length of the humerus/radius of the
There are also differences between the Tbx4/Pitx1-
rescued limb and the control hindlimb as well as the
Prx1-Pitx1(P1.1) transgenic limb. The most obvious is
the lack of a posterior zeugopodal bone, with the con-
comitant loss of the two posterior-most digits. This
could be the result of gene dosage effects and may
ectopic limbs when similarly misexpressed (M.P.L., un-
published), consistent with our observation that Pitx1
alone is not able to rescue limb formation in the Tbx5
mutant. Furthermore, Pitx1 null mice hindlimbs, al-
though smaller, possess all the skeletal elements with
the exception of the ilium (Lanctot et al., 1999; Szeto et
al., 1999), suggesting it has only a minor role in limb out-
A role for Tbx5 in initiation of limb outgrowth has been
demonstrated in a range of species (reviewed in Logan,
2003). However, a similar role for Tbx4 in hindlimb initia-
tionhas notbeen demonstrated.In Tbx4null mice,hind-
limb bud formation is initiated normally, although after
E10 further outgrowth is disrupted (Naiche and Papai-
oannou, 2003). One explanation for the normal initiation
of hindlimb development in the Tbx4 null could be that
other factor(s) compensate for the loss of Tbx4. We
predict that Pitx1 is not such a compensatory factor
because it is not capable of rescuing the Tbx5 null
One of the earliest defects in Tbx5 null mice is the
absence of Fgf10 expression in the LPM, which is re-
quired for outgrowth of the limb bud. Moreover, it has
been shown that the Fgf10 promoter contains T-box
binding sites and that Tbx5 is able to directly upregulate
Fgf10 expression (Agarwal et al., 2003). The failure to
maintain Fgf10 expression in Tbx4 mutant hindlimbs
suggests that Tbx4 may also recognize these Tbx bind-
ing sites and activate Fgf10 expression (Naiche and Pa-
paioannou, 2003). Consistent with the idea that Fgf10
may be a common target of Tbx5 and Tbx4, both genes
are necessary to activate the expression of Fgf10 in
the lung mesenchyme (Cebra-Thomas et al., 2003). We
predict that in Tbx4-rescued limbs, Tbx4 binds to T-box
binding sites in the Fgf10 promoter to activate expres-
sion and initiate limb outgrowth.
Figure 5. Transgene-Derived Tbx4 Can Rescue Limb Outgrowth
Following Constitutive Deletion of Tbx5, and the Rescued Limb
Maintains Its Forelimb Identity
All embryos are Tbx5lox/lox;?-actin-Cre;Prx1-Tbx4(4.48) at E10 and
(A) Rescued limb.
(B–D)Endogenous Tbx5(B), Hoxc4(C),and Hoxc5(D) areexpressed
in the rescued limb buds.
(E–G) Hindlimb-restricted Tbx4 (E), Pitx1 (F), and Hoxc10 (G) are
expressed in the hindlimb (arrows) but not ectopically expressed in
the rescued limb.
reveal a role for Pitx1 and Tbx4 in regulating cellular
rior patterning of the limb. In summary, the addition of
Pitx1 into Tbx4-rescued limbs or wild-type forelimbs
causes a partial forelimb-to-hindlimb transformation,
confirming a role for Pitx1 in specification of hindlimb-
specific morphologies. However, these limbs are not
factors are required to dictate complete hindlimb char-
acteristics and/or that factors in the forelimb region
transformation. Interestingly, Hoxc10 is not ectopically
cating that, in this assay, Tbx4 and Pitx1 are not suffi-
cient to induce Hoxc10 expression and that the partial
transformation in limb morphology we observe occurs
independently of Hoxc10.
Tbx5 and Tbx4 Are Not Sufficient to Determine
The limb-type restricted expression patterns of Tbx5
and Tbx4 in a range of vertebrate species have sug-
ary conserved mechanism to specify limb-specific mor-
Logan, 2003). Moreover, misexpression experiments in
the chick led to the conclusion that Tbx5 and Tbx4 are
sufficient to specify forelimb- and hindlimb-specific
morphologies, respectively (Rodriguez-Esteban et al.,
1999; Takeuchi et al., 1999).
Our results force a reexamination of the roles of Tbx5
and Tbx4 in the specification of limb-specific morphol-
ogy. In our limb-rescue experiments, although Tbx4 is
able to replace the function of Tbx5 so that limb out-
growth is maintained, Tbx4 does not produce a limb
with hindlimb-likemorphology, and instead thelimb ele-
ments resemble those of a forelimb. Forelimb-specific
genes are expressed in the rescued limb, whereas hind-
limb-specific genes are not expressed at any stages
analyzed (Figures 4G–4L). Significantly, our results also
demonstrate that forelimb morphologies do form in the
absence of Tbx5 and reveal that Tbx5 is not required
for the specification of forelimb-specific morphology.
We therefore conclude that Tbx5 and Tbx4 do not play
Common Roles for Tbx5 and Tbx4 but Not for Pitx1
during Limb Bud Outgrowth
Our results demonstrate that Tbx4, but not Pitx1, is able
to rescue limb outgrowth in the absence of Tbx5 (Figure
2), suggesting that Tbx5 and Tbx4 play identical bio-
chemical roles in forelimb and hindlimb outgrowth, re-
ated and the limb bud generated is normally patterned
(Figure 3). Consistent with our observations, Tbx4 can
induce the formation of an additional limb when misex-
pressed in the interlimb flank of a chick embryo (Take-
uchi et al., 2003). In contrast, Pitx1 is not able to induce
Outgrowth and Identity of the Vertebrate Limb
Figure 6. Pitx1 Transforms the Forelimb-like Morphology of a Normal Forelimb or a Tbx4-Rescued Limb to a More Hindlimb-like Morphology
(A–D) Alcian blue/Alizarin red staining of E17.5 control forelimb (A), Tbx4/Pitx1-rescued limb (B), Pitx1 transgenic forelimb (C), and control
hindlimb (D) skeletal elements.
(E–L) Higher magnifications of the stylopod/zeugopod joint are shown for the control forelimb (E), Tbx4/Pitx1-rescued limb (F), Pitx1 transgenic
limb (G), and control hindlimb (H) and of the autopod region of the control forelimb (I), Tbx4/Pitx1-rescued limb (J), Pitx1 transgenic limb (K),
and control hindlimb (L). au, autopod; F, femur; Fi, fibula; H, humerus; ish, ishium; il, ilium; R, radius; Sc, scapula; st, stylopod; T, tibia; U,
ulna; zg, zeugopod.
(M) A Tbx4/Pitx1-rescued embryo at E10.5. Hoxc10 is expressed in the hindlimb but not ectopically induced in the double-rescued limb
a role in the specification of limb-specific morphologies
but instead have common roles in the initiation and
maintenance of limb outgrowth. Reasons for the dis-
crepancies between our results in the mouse and those
from misexpression experiments in the chick are un-
clear. However, we do not believe this is due to differ-
We have introduced the Tbx4 transgene to homozygos-
ity in the Tbx5lox/lox;Prx1-Cre background, and these res-
cued limbs, although morphologically abnormal, are
clearly forelimb-like (Figures 4P and 4Q). We suggest
that in our transgenic model, we have expressed Tbx4
at levels appropriate for limb formation and at levels
sufficiently above physiological levels such that normal
limb formation is disrupted. Even at these higher levels
of Tbx4 expression, we do not detect any apparent
transformation of limb-specific morphology.
The conclusions we have drawn from our observa-
ments. In Tbx4 mutant embryos, Pitx1 continues to be
retain their hindlimb-like characteristics (Naiche and
Papaioannou, 2003). Similarly, a forelimb-to-hindlimb
transformation is not observed in Tbx5?/?embryos
(Agarwal et al., 2003; Rallis et al., 2003). In these em-
bryos, Tbx5 transcripts are still expressed in the pro-
spective forelimb region, and neither Tbx4 nor Pitx1 are
ectopically expressed. Inaddition, transplantation stud-
ies in the chick suggest that limb-type identity is speci-
fied at stages 9–12, before the induction of Tbx5 and
Tbx4 in their respective limb fields (Saito et al., 2002;
Stephens et al., 1989).
Pitx1 Is a Candidate Axial Cue Required for
Specification of Hindlimb-Specific Morphology
Pitx1 is expressed in a broad, caudal domain of the
embryo prior to expression of Tbx4 in the presumptive
hindlimb-forming region (Lamonerie et al., 1996; Logan
et al., 1998). This appears to be an ancient arrangement
that has been conserved during evolution, since the
single, ancestral Pitx gene is also expressed in a caudal
domain in amphioxus (Yasui et al., 2000). In our double-
rescued embryos, in which Tbx5 is replaced with both
Tbx4 and Pitx1, and in our Pitx1 transgenics, the mor-
phology of the resultant limb is significantly more hind-
limb-like than the limb element that forms following
rescue with Tbx4 alone. The morphology of the limb
elements that form following replacement of Tbx5 with
ing ectopic expression of Pitx1 in a forelimb expressing
endogenous Tbx5. However, we cannot conclude that
differences in levels of transgene-derived Tbx4 in com-
parison to endogenous Tbx5 are apparent since limb
elements are absent in Tbx4/Pitx1-rescued limbs that
do form in the Prx1-Pitx1 transgenic limb. Our results
indicate thatPitx1 is a keydownstream factor regulating
hindlimb identity. This conclusion is consistent with ex-
periments in the chick that demonstrated that misex-
pression of Pitx1 is capable of transforming the wing to
a more hindlimb-like structure (Logan and Tabin, 1999;
Szeto et al., 1999; Takeuchi et al., 1999) and gene dele-
teristics are lost in Pitx1 mutants (Lanctot et al., 1999;
Szeto et al., 1999).
Our data also suggest that Pitx1 may not be the single
positional cue involved in the regionalization of rostral
versus caudal domains in the embryo, subterritories of
which will ultimately develop into either a forelimb or a
hindlimb, respectively. Replacement of Tbx5 with both
Tbx4 and Pitx1 or misexpression of Pitx1 in a Tbx5 wild-
to-hindlimb transformation, presumably due to the ab-
sence of other necessary factors or due to the presence
support to this model, the limbs that form following
replacement of Tbx5 with Tbx4 (Tbx4 rescue) and in
Pitx1?/?embryos have different morphologies despite
both expressing only Tbx4 and not Tbx5 or Pitx1. In the
Tbx4 rescue, a limb is formed in a rostral (forelimb)
the limb that forms in a caudal (hindlimb) domain in
the Pitx1?/?mutant lacks any forelimb characteristics.
These results indicate that Pitx1 requires additional fac-
tors to distinguish limb-specific morphologies.
Other candidates to specify limb-specific morpholog-
ies are the Hox genes. Specific combinations of Hox
genes expressed in the embryonic LPM correlate well
with the type of limb that will develop (Cohn et al., 1997),
lated with the absence of forelimbs in snakes (Cohn and
Tickle, 1999). We propose a model (Figure 7) in which
limb-specific identity and ultimately limb-specific mor-
phology is specified by different combinatorial codes of
factors in the LPM at rostral versus caudal positions.
These factors may include a particular combination of
Hox proteins and Pitx1. In response to an axial cue that
triggers Tbx-mediated limb initiation at the prospective
of a combinatorial code of “rostral” Hox genes, whereas
Tbx4 expression is initiated by a combinatorial “caudal”
Hox code (Ruvinsky and Gibson-Brown, 2000). The acti-
and hindlimb level, respectively, is responsible for the
initiation and outgrowth of the limb bud. We conclude
that Tbx5 and Tbx4 are accurate markers of forelimb
and hindlimb identity, respectively, but they do not
themselves play a significant role in the specification of
Figure 7. A Model of the Mechanisms that Control Limb-Type Iden-
tity and Outgrowth of the Vertebrate Limbs
Broad territories in the flank of the embryo capable of forming either
a forelimb or a hindlimb are specified by distinct rostral and caudal
a combinationof Hoxgenes inthe rostraldomain and acombination
of Hox genes and Pitx1 in the caudal domain. In response to a
putative axial cue, which may be common to both forelimb and
hindlimb, cells in the domain that express a rostral code activate
Tbx5, which is required for forelimb outgrowth. Cells in the caudal
domain, which express a different positional code including Pitx1,
respond to this axial cue and activate Tbx4, which is required for
hindlimb outgrowth. IM, intermediate mesoderm; LPM, lateral plate
mesoderm; NT, neural tube; so, somite.
to form a two-gene cluster that was later duplicated and
dispersed in the genome. The evolutionary history of
these genes argues for a high degree of functional over-
lap between cognate genes (Agulnik et al., 1996). Fur-
thermore, the residues that have been shown to be
important for DNA binding and dimerization of the
T-domain of Xenopus Brachyury (the prototype of the
T-box family) are identical in mouse and chick Tbx4 and
Tbx5 (data not shown). Therefore, the simplest model
would predict that the Tbx4 and Tbx5 proteins share
the vast majority of their target genes rather than having
Several models have been proposed to explain the
maintenance of duplicate genes in the genome after an
initial phase of redundancy. Classical models propose
that one of the duplicates will normally degenerate due
to the accumulation of deleterious mutations. However,
on rare occasions one of the copies may acquire a novel
function, endowing the organism with a favorable, se-
lected new function (Ohno, 1970). The duplication-
degeneration-complementation (DDC) model was pro-
posed to explain the higher number of duplicate gene
copies maintained in duplicated genomes (such as the
vertebrate genome) than classical evolutionary models
would predict. The DDC model predicts that degenera-
tive mutations in the regulatory elements can increase
the probability of duplicate gene preservation and, im-
portantly in our model for Tbx4 and Tbx5 function, that
the usual mechanism of duplicate gene preservation is
the partitioning of ancestral functions rather than the
of the Tbx4/Tbx5 gene pair has been partitioned to initi-
ate outgrowth of two sets of serially homologous ap-
pendages. Both copies of the gene are then retained
Evolution of Tbx5 and Tbx4
Two cognate gene pairs compose the Tbx2/3/4/5 sub-
family of T-box transcription factors: the Tbx2/Tbx3 pair
and the Tbx4/Tbx5 pair. These gene pairs have evolved
from a single ancestral gene by unequal crossing over
Outgrowth and Identity of the Vertebrate Limb
described previously: cTbx4, cPitx1, mTbx4, mTbx5 (Logan et al.,
1998), mFgf8 (Crossley and Martin, 1995), mShh (Echelard et al.,
1993), mFgf10 (Bellusci et al., 1997), mHoxc4, mHoxc5, mHoxc10
(Burke et al., 1995), and mPitx1 (Logan et al., 1998).
functions that were previously accomplished by the an-
cestral gene (Force et al., 1999). In this example, both
for formation of the hindlimbs and forelimbs, respec-
tively. Regulatory changes, rather than structural changes
in the coding region (which tend to be deleterious), have
beeninvolved innonlethal andrapid morphologicalvari-
ation andare therefore candidatesto beimportant com-
ponents of evolutionary changes. Interestingly, Pitx1
and Tbx4 have been involved in macroevolutionary
changes related to the morphology of pelvic structures
of natural occurring population of three spine stickle-
backs (Shapiro et al., 2004).
The cartilage and bone elements of E17.5 mouse embryos and new-
born pups were stained with alcian blue and alizarin red, respec-
tively, essentially as described (Hogan et al., 1994).
C.M. is funded by an EMBO long-term fellowship. M.P.L. has re-
ceived funding from the EMBO young investigator program and the
Medical Research Council. We thank Benoit Bruneau, John Seid-
man, and Christine Seidman for their generosity in providing the
Tbx5 conditional mouse line. We thank S. Bogni for supplying the
?-actin-Cre transgenic line. We are deeply indebted to the staff of
the Biological Services and Procedural Services sections, NIMR, for
assistance with the animal work. We thank F. Johnson, Photograph-
ics section, NIMR, for producing the artwork. We thank members
of the lab, T. Heanue, and three anonymous reviewers for their
Generation of Transgenic Lines
The conditional allele of Tbx5 and Prx1-Cre transgenic have been
described previously (Bruneau et al., 2001; Logan et al., 2002). Prx1-
Pitx1 transgenic lines were generated by the NICHD Transgenic
Mouse Development Facility operated by the University of Alabama
at Birmingham. Prx1-Tbx4 transgenic lines were generated by the
Procedural Services section, NIMR. The cDNAs for Tbx4 (AF033670)
and Pitx1 (AF069397) were HA-tagged at their 3? end. An insulator
element (5?HS4) (Chung et al., 1993) was placed at the 5? end of
the construct. The SV40 polyadenylation signal and artificial intron
sequence was placed at the 3? end of the construct.
Received: June 23, 2004
Revised: September 29, 2004
Accepted: November 3, 2004
Published: January 3, 2005
Agarwal, P., Wylie, J.N., Galceran, J., Arkhitko, O., Li, C., Deng,
C., Grosschedl, R., and Bruneau, B.G. (2003). Tbx5 is essential for
forelimb bud initiation following patterning of the limb field in the
mouse embryo. Development 130, 623–633.
Agulnik, S.I., Garvey, N., Hancock, S., Ruvinsky, I., Chapman, D.L.,
tion of mouse T-box genes by tandem duplication and cluster dis-
persion. Genetics 144, 249–254.
Ahn, D.G., Kourakis, M.J., Rohde, L.A., Silver, L.M., and Ho, R.K.
(2002). T-box gene tbx5 is essential for formation of the pectoral
limb bud. Nature 417, 754–758.
Bellusci, S., Grindley, J., Emoto, H., Itoh, N., and Hogan, B.L. (1997).
Fibroblast growth factor 10 (FGF10) and branching morphogenesis
in the embryonic mouse lung. Development 124, 4867–4878.
tion of microgram quantities of protein utilizing the principle of pro-
tein-dye binding. Anal. Biochem. 72, 248–254.
Bruneau, B.G., Nemer, G., Schmitt, J.P., Charron, F., Robitaille, L.,
Caron, S., Conner, D.A., Gessler, M., Nemer, M., Seidman, C.E.,
and Seidman, J.G. (2001). A murine model of Holt-Oram syndrome
defines roles of the T-box transcription factor Tbx5 in cardiogenesis
and disease. Cell 106, 709–721.
Burke, A.C., Nelson, C.E., Morgan, B.A., and Tabin, C. (1995). Hox
genes and the evolution of vertebrate axial morphology. Develop-
ment 121, 333–346.
Cebra-Thomas, J.A., Bromer, J., Gardner, R., Lam, G.K., Sheipe,
H., and Gilbert, S.F. (2003). T-box gene products are required for
mesenchymal induction of epithelial branching in the embryonic
mouse lung. Dev. Dyn. 226, 82–90.
Chapman, D.L., Garvey, N., Hancock, S., Alexiou, M., Agulnik, S.I.,
Papaioannou, V.E. (1996). Expression of the T-box family genes,
Tbx1-Tbx5, during early mouse development. Dev. Dyn. 206,
Chung, J.H., Whiteley, M., and Felsenfeld, G. (1993). A 5? element
of the chicken beta-globin domain serves as an insulator in human
erythroid cells and protects against position effect in Drosophila.
Cell 74, 505–514.
Cohn, M.J., and Tickle, C. (1999). Developmental basis of limbless-
ness and axial patterning in snakes. Nature 399, 474–479.
Mouse embryos were staged according to Kaufman (2001). Noon
on the day a vaginal plug was observed was taken to be E0.5 days
of development. Mice carrying the conditional Tbx5 allele, Tbx5lox/lox,
were identified as previously described (Bruneau et al., 2001). The
Prx1-Cre transgene was identified as previously described (Logan
et al., 2002). The Prx1-Tbx4 and Prx1-Pitx1 transgene was identified
using a common reverse primer SV40rev and specific forward prim-
ers Tb4fwd, Pitxfwd, respectively. The ?-actin-Cre transgene was
identified using Cre forward and reverse primers. Details available
on request. To generate rescued embryos (Tbx5lox/lox;Prx1-Cre;Prx1-
Tbx4orPitx1), Tbx5lox/?;Prx1-Cre;Prx1-Tbx4orPitx1 mice were crossed
to homozygote Tbx5lox/lox. To obtain ?-actin-Cre-deleted Tbx4-res-
cued embryos, adult heterozygote Tbx5 mice (Tbx5lox/?;?-actin-Cre)
were crossed to Tbx5lox/lox;Prx1-Tbx4(4.48) mice. Heterozygote Tbx5
mice (Tbx5lox/?;Prx1-Cre or Tbx5lox/?;?-actin-Cre) were used as con-
Western Blot Analysis
Proteins from both forelimb and hindlimb buds of E10.5 transgenic
embryos were obtained using a protein extraction buffer (50 mM
determined by Bradford method (Bradford, 1976) with a Coomassie
protein assay reagent kit (Pierce) with bovine serum albumin as
standard protein following manufacturer’s instructions. 4 ?g of total
protein was subjected to standard SDS-PAGE. Proteins were elec-
a XCell II Blot module (Invitrogen). The membrane was treated with
5% nonfat dry milk in a solution containing 50 mM Tris-HCl (pH 7.4),
0.5 M NaCl, and 0.1% Tween 20 (TBS-T) for 1 hr and then incubated
overnight at 4?C with anti-HA rat monoclonal antibody (Roche) at a
1:5000 dilution. Unbound antibodies were washed out with TBS-T.
Incubation with peroxidase-conjugated goat anti-rat IgG (Calbio-
chem) at a 1:5000 dilution was performed at room temperature for
1 hr. Unbound antibodies were washed out with TBS-T. The Su-
perSignalWest PicoChemiluminescent Substrate(Pierce) wasused
for HRP detection. The membrane was subsequently exposed to
CL-X posure Film (Pierce).
Whole-Mount In Situ Hybridization
Whole mount in situ hybridizations were carried out essentially as
previously described (Riddle et al., 1993). All probes have been
Developmental Cell Download full-text
Cohn, M.J., Patel, K., Krumlauf, R., Wilkinson, D.G., Clarke, J.D.,
and Tickle, C. (1997). Hox9 genes and vertebrate limb specification.
Nature 387, 97–101.
Crossley, P.H., and Martin, G.R. (1995). The mouse Fgf8 gene en-
codes a family of polypeptides and is expressed in regions that
direct outgrowthand patterningin the developingembryo. Develop-
ment 121, 439–451.
Echelard, Y., Epstein, D.J., St-Jacques, B., Shen, L., Mohler, J.,
McMahon, J.A., and McMahon, A.P. (1993). Sonic hedgehog, a
member of a family of putative signaling molecules, is implicated
in the regulation of CNS polarity. Cell 75, 1417–1430.
Force, A., Lynch, M., Pickett, F.B., Amores, A., Yan, Y.L., and Post-
lethwait, J. (1999). Preservation of duplicate genes by complemen-
tary, degenerative mutations. Genetics 151, 1531–1545.
Gibson-Brown, J.J., Agulnik, S.I., Chapman, D.L., Alexiou, M., Gar-
vey, N., Silver, L.M., and Papaioannou, V.E. (1996). Evidence of a
role for T-box genes in the evolution of limb morphogenesis and the
specification of forelimb/hindlimb identity. Mech. Dev. 56, 93–101.
Hogan, B., Beddington, R., Constantini, F., and Lacy, E. (1994).
Manipulating the Mouse Embryo, Second Edition (Cold Spring Har-
bor, NY: Cold Spring Harbor Press).
Kaufman, M.H. (2001). The Atlas of Mouse Development, Second
Edition (Cambridge, UK: Academic Press).
Lamonerie, T., Tremblay, J.J., Lanctot, C., Therrien, M., Gauthier, Y.,
factor involved in transcription of the pro-opiomelanocortin gene.
Genes Dev. 10, 1284–1295.
Lanctot, C., Moreau, A., Chamberland, M., Tremblay, M.L., and
Drouin, J. (1999). Hindlimb patterning and mandible development
require the Ptx1 gene. Development 126, 1805–1810.
Lewandoski, M., and Martin, G.R. (1997). Cre-mediated chromo-
some loss in mice. Nat. Genet. 17, 223–225.
Logan, M. (2003). Finger or toe: the molecular basis of limb identity.
Development 130, 6401–6410.
Logan, M., and Tabin, C.J. (1999). Role of Pitx1 upstream of Tbx4
in specification of hindlimb identity. Science 283, 1736–1739.
Logan, M., Simon, H.G., and Tabin, C. (1998). Differential regulation
of T-box and homeobox transcription factors suggests roles in con-
trolling chick limb-type identity. Development 125, 2825–2835.
Logan, M., Martin, J.F., Nagy, A., Lobe, C., Olson, E.N., and Tabin,
C.J. (2002). Expression of Cre Recombinase in the developing
mouse limb bud driven by a Prxl enhancer. Genesis 33, 77–80.
Martin, G.R. (1998). The roles of FGFs in the early development of
vertebrate limbs. Genes Dev. 12, 1571–1586.
Martin, J.F., and Olson, E.N. (2000). Identification of a prx1 limb
enhancer. Genesis 26, 225–229.
Min, H., Danilenko, D.M., Scully, S.A., Bolon, B., Ring, B.D., Tarpley,
J.E., DeRose, M., and Simonet, W.S. (1998). Fgf-10 is required for
both limb and lung development and exhibits striking functional
similarity to Drosophila branchless. Genes Dev. 12, 3156–3161.
Naiche, L.A., and Papaioannou, V.E. (2003). Loss of Tbx4 blocks
hindlimb development and affects vascularization and fusion of the
allantois. Development 130, 2681–2693.
Nelson, C.E., Morgan, B.A., Burke, A.C., Laufer, E., DiMambro, E.,
Murtaugh, L.C., Gonzales,E., Tessarollo, L., Parada,L.F., and Tabin,
C. (1996). Analysis of Hox gene expression in the chick limb bud.
Development 122, 1449–1466.
Ohno, S. (1970). Evolution by Gene Duplication (Heidelberg, Ger-
Peterson, R.L., Papenbrock, T., Davda, M.M., and Awgulewitsch, A.
(1994). The murine Hoxc cluster contains five neighboring AbdB-
related Hox genes that show unique spatially coordinated expres-
sion in posterior embryonic subregions. Mech. Dev. 47, 253–260.
Rallis, C., Bruneau, B.G., Del Buono, J., Seidman, C.E., Seidman,
for forelimb bud formation and continued outgrowth. Development
Riddle, R.D., Johnson, R.L., Laufer, E., and Tabin, C. (1993). Sonic
hedgehog mediates the polarizing activity of the ZPA. Cell 75, 1401–
Rodriguez-Esteban, C., Tsukui, T., Yonei, S., Magallon, J., Tamura,
K., and Izpisua Belmonte, J.C. (1999). The T-box genes Tbx4 and
Tbx5 regulate limb outgrowth and identity. Nature 398, 814–818.
Ruvinsky, L., and Gibson-Brown, J.J. (2000). Genetic and develop-
mental bases of serial homology in vertebrate limb evolution. Devel-
opment 127, 5233–5244.
Saito, D., Yonei-Tamura, S., Kano, K., Ide, H., and Tamura, K. (2002).
Specification and determination of limb identity: evidence for inhibi-
tory regulation of Tbx gene expression. Development 129, 211–220.
Sekine, K., Ohuchi, H., Fujiwara, M., Yamasaki, M., Yoshizawa, T.,
Sato, T., Yagishita, N., Matsui, D., Koga, Y., Itoh, N., and Kato, S.
(1999). Fgf10 is essential for limb and lung formation. Nat. Genet.
Shang, J., Luo, Y., and Clayton, D.A. (1997). Backfoot is a novel
homeobox gene expressed in the mesenchyme of developing hind
limb. Dev. Dyn. 209, 242–253.
Shapiro, M.D., Marks, M.E., Peichel, C.L., Blackman, B.K., Nereng,
K.S., Jonsson, B., Schluter, D., and Kingsley, D.M. (2004). Genetic
and developmental basis of evolutionary pelvic reduction in threes-
pine sticklebacks. Nature 428, 717–723.
Stephens, T.D., Beier, R.L., Bringhurst, D.C., Hiatt, S.R., Prestridge,
M., Pugmire, D.E.,and Willis, H.J. (1989). Limbness inthe early chick
embryo lateral plate. Dev. Biol. 133, 1–7.
Szeto, D.P., Rodriguez-Esteban, C., Ryan, A.K., O’Connell, S.M.,
Liu, F., Kioussi, C., Gleiberman, A.S., Izpisua-Belmonte, J.C., and
Rosenfeld, M.G. (1999). Role of the Bicoid-related homeodomain
factor Pitx1 in specifying hindlimb morphogenesis and pituitary de-
velopment. Genes Dev. 13, 484–494.
A., Naitoh-Matsuo, M., Ogura, K., Takahashi, N., Yasuda, K., and
Ogura, T. (1999). Tbx5 and Tbx4 genes determine the wing/leg iden-
tity of limb buds. Nature 398, 810–814.
Takeuchi, J.K., Koshiba-Takeuchi, K., Suzuki, T., Kamimura, M.,
Ogura, K., and Ogura, T. (2003). Tbx5 and Tbx4 trigger limb initiation
through activation of the Wnt/Fgf signaling cascade. Development
Yasui, K., Zhang, S., Uemura, M., and Saiga, H. (2000). Left-right
asymmetric expression of BbPtx, a Ptx-related gene, in a lancelet
species and the developmental left-sidedness in deuterostomes.
Development 127, 187–195.