Cooperative and antagonistic interactions between Sall4 and ... - Nature

© 2006 Nature Publishing Group http://www.nature.com/naturegenetics

ARTICLES

Cooperative and antagonistic interactions between Sall4 and Tbx5 pattern the mouse limb and heart Kazuko Koshiba-Takeuchi1–3,7, Jun K Takeuchi1–3,7, Eric P Arruda1–4, Irfan S Kathiriya5, Rong Mo2, Chi-chung Hui2,4, Deepak Srivastava5,6 & Benoit G Bruneau1–3 Human mutations in TBX5, a gene encoding a T-box transcription factor, and SALL4, a gene encoding a zinc-finger transcription factor, cause similar upper limb and heart defects. Here we show that Tbx5 regulates Sall4 expression in the developing mouse forelimb and heart; mice heterozygous for a gene trap allele of Sall4 show limb and heart defects that model human disease. Tbx5 and Sall4 interact both positively and negatively to finely regulate patterning and morphogenesis of the anterior forelimb and heart. Thus, a positive and negative feed-forward circuit between Tbx5 and Sall4 ensures precise patterning of embryonic limb and heart and provides a unifying mechanism for heart/hand syndromes.

Patterning and morphogenesis of embryonic structures require precise interactions between signaling molecules and transcriptional regulators. Transcription factors regulate key aspects of patterning by activating and/or repressing downstream target genes. The complexity of transcriptional regulation of embryonic development is still poorly understood, but its importance is exemplified by the severe defects caused by dominant mutations in transcription factor genes in human syndromes1. In the developing heart, transcription factor interactions are an important mechanism to ensure robust gene activation, and they may form the basis for gene dosage sensitivity2. For example, interactions between the T-box transcription factor TBX5 and the cardiac transcription factors GATA4 and NKX2-5 are thought to be key interactions that, when disrupted, lead to congenital heart defects (CHDs)3–5. Despite these known interactions, several aspects of cardiac patterning remain poorly understood. TBX5 mutations in humans cause Holt-Oram syndrome (HOS)6, which is characterized by CHDs as well as upper limb defects of the radial ray that predominantly include thumb anomalies such as absent thumb or a triphalangeal digit, defects in carpal bones and the radius, or in more severe cases phocomelia7–9. Although a few dosagesensitive targets and interacting partners for mouse Tbx5 have been defined in the heart3–5, the molecular mechanism of limb defects in HOS has not been defined. Anterior-posterior patterning of the autopod of the limb, which gives rise to the hand, leads to the formation of digits with distinct morphologies, such as a short and mobile thumb (digit 1) anteriorly, and fingers of varying lengths (digits 2–5) more posteriorly. Limb patterning along the anterior-posterior axis is mediated largely

through the actions of the signaling molecule Sonic hedgehog (Shh) in a posterior signaling center called the zone of polarizing activity10. Shh acts to determine the numbers of digits and their identity11,12, partly via bone morphogenetic proteins (BMPs)13–15, and it regulates limb bud growth via fibroblast growth factors (FGFs)10. But Shh patterns only digits 2–5; it is not involved in regulating the formation of digit 1 (refs. 16–18). Downstream of Shh, BMPs act in part through the T-box transcriptional repressors Tbx2 and Tbx3, which are expressed in the posterior region of the limb, and which are thought to be important for posterior digit morphogenesis and identity, including the regulation of Shh15,19,20. These results suggest that signaling relays between Tbx2/3 and Shh/BMP are important in specifying posterior digit identity. Although much is known about the mechanisms that control autopod patterning and the establishment of the posterior digits, little is known about the mechanism that control the patterning and fate determination of the anterior structures of the limb, which are affected by loss of Tbx5. Limb defects identical to those in HOS are also found in Okihiro syndrome (OS), which is caused by mutations of the Spalt-family zinc-finger transcription factor gene SALL4 (refs. 21,22). OS is characterized by absent abducens neurons, congenital deafness, renal abnormalities, anal stenosis and CHDs, especially ventricular septal defects (VSDs)21–24. CHDs are found in approximately 13% of OSafflicted individuals with SALL4 mutations. Because of the similarity of limb phenotypes between OS and HOS, and because of their association with CHDs, the two syndromes are sometimes confused, and SALL4 mutations have been found in individuals who were initially diagnosed with HOS23,25. The roles of Sall4 in embryogenesis

1Programs in Cardiovascular Research and 2Developmental Biology, The Hospital for Sick Children, Toronto, Ontario M5G 1X8, Canada. 3The Heart and Stroke/Richard Lewar Centre of Excellence and 4Department of Molecular and Medical Genetics, University of Toronto, Toronto, Ontario M5S 1A8, Canada. 5Departments of Pediatrics and Molecular Biology, University of Texas Southwestern Medical Center at Dallas, Dallas, Texas 75390-9148, USA. 6Present address: Gladstone Institute of Cardiovascular Disease, University of California, San Francisco, California 94158, USA. 7These authors contributed equally to this work. Correspondence should be addressed to B.G.B. ([email protected]).

Received 22 August; accepted 10 November; published online 25 December 2005; doi:10.1038/ng1707

NATURE GENETICS VOLUME 38

[

NUMBER 2

[

FEBRUARY 2006

175

ARTICLES

c

Figure 1 Sall4 and Tbx5 expression in mouse development. (a) Sall4 mRNA expression in an E11.5 embryo. (b) lacZ expression (salmon-gal) of an E11.5 Sall4GT/+ embryo. (c,d) lacZ staining of E10.5 (c) and E12.5 (d) Sall4GT/+ embryos showing expression in the heart; section below shows predominant expression in the interventricular septum. (e–l) Dorsal views of Sall4 expression during limb development. The anterior side is top. (e,i) E9.5, (f,j) E10.5, (g,k) E11,5, (h,l) E12.5 forelimbs (FL) and hindlimbs (HL). (m–p) Dorsal views of Tbx5 expression during forelimb development. The anterior side is top. (m) E9.5, (n) E10.5, (o) E11,5, (p) E12.5. Red numbers indicate digits. (q–s) Expression in E11.5 forelimbs of the following: (q)Tbx5 (purple), (r) Sall4 lacZ (red), (s) Tbx5+Sall4, (t) Hoxd12 (purple), (u) Sall4 lacZ (red) and (v) Hoxd12+Sall4.

d

rv lv

lv

rv E10.5

rv

E9.5

FL

f

g

h

E10.5

E11.5

E12.5

l

HL

j

k

l

m

FL

n

o

p

Sall4

rv

lv

2

1

3

4 Tbx5

5

q

r

Tbx5

t

s

u

Hoxd12

Tbx5 + LacZ

LacZ

v

LacZ

Hoxd12 + LacZ

have yet to be elucidated, but the similarity in patterning defects in OS and HOS suggest that Sall4 and Tbx5 may have similar or intersecting roles in development. Here we show that a feed-forward circuit between Tbx5 and Sall4 has a key role in the formation of the heart and anterior structures of the forelimb. RESULTS Sall4 expression in the developing mouse Sall4 expression was observed at embryonic day 7 (E7.0) in extra embryonic tissues, including chorion, allantois and amnion. Between E8.5 and 9.5, Sall4 expression was mainly restricted to neural tissues, as well as the invaginating hindgut pocket and lateral plate mesoderm (data not shown). By E11.5, Sall4 expression was observed in the midbrain, the rostral edge of the forebrain, maxillary arch, genital

tubercle, limb and tail (Fig. 1a), as previously reported26. Expression of b-galactosidase from a lacZ gene inserted into the endogenous Sall4 locus (from a Sall4 gene trap allele, referred to as Sall4GT; see Supplementary Fig. 1 online for details) showed expression patterns identical to the in situ hybridization pattern (Fig. 1b). We did not detect expression of Sall4 in the heart by in situ hybridization, but lacZ expression from the gene trap allele was clearly detected in left ventricular myocardium at E10.5, with higher levels in the interventricular septum (Fig. 1c). Cardiac Sall4 expression was maintained at E12.5 (Fig. 1d). Sall4 and Tbx5 expression overlaps in the forelimb Anterior limb defects are typical in both OS and HOS, so we examined the expression of Sall4 (mRNA or lacZ from the genetrap allele) during limb development and compared it to that of Tbx5 (Fig. 1). At E9.5, when forelimbs are more developed than hindlimbs, Sall4 was expressed in the peripheral mesenchyme of forelimbs, whereas it was expressed in the entire limb mesenchyme of hindlimbs (Fig. 1e,i). At E10.5, forelimb expression of Sall4 was detected in distal mesenchyme, whereas it was widely expressed in the hindlimb (Fig. 1f,j). From E11.5 onwards, the most intense forelimb signal was detected on the anterior side of the distal limb, whereas in hindlimbs, Sall4 was expressed throughout the distal region (Fig. 1g,k). At E12.5, when digit formation has started, Sall4 expression was markedly different: the signal was detected in the entire distal-most region of both forelimbs and hindlimbs (Fig. 1h,l).

a

+/+

Tbx5

+/–

b

+/+

+/+

GT/+ GT/+

Sall4 E11.0

Sall4-lacZ GAPDH

c

160%

Heart (E10.5)

140% 120% 100%

176

Forelimb

Figure 2 Tbx5 regulates Sall4 expression in forelimb and in heart. (a) Decreased expression of Sall4 in Tbx5+/– forelimbs at E11 and E12; expression was not altered in hindlimb. Red arrowheads show anteriormost expression of Sall4, which is most affected in Tbx5+/– mice. (b) RT-PCR of Sall4 from E10.5 heart ventricle RNA showing Sall4 expression in the heart and expression of a Sall4-lacZ fusion transcript in Sall4GT/+ heart. (c) Quantitative RT-PCR showing decreased Sall4 mRNA levels in Tbx5+/– cardiac ventricles at E10.5. *P ¼ 0.007 by unpaired t-test.

*

80% 60% 40% 20%

E12.0

0%

+/+

Tbx5

+/–

Hindlimb


e

lv

LacZ

Sall4

Forelimb

b

Hindlimb

a

VOLUME 38

[

NUMBER 2

[

FEBRUARY 2006 NATURE GENETICS

ARTICLES Sall4 GT/+

+/+

a 3

2

4 5

FL

* * *

1

2 3

4 5

1


HL

* * Sall4 GT/+

+/+

Sall4 GT/+

+/+

b

Fgf10

e

Hoxd12

c

Tbx4

f

EphA5

d

Pitx1

g

EphA3

At early stages (E9.5 and E10.5) Tbx5 was expressed in the entire forelimb bud (Fig. 1m,n). At E11.5, the expression of Tbx5 in the posterior-distal region was lower, restricting Tbx5 to the anterior portion of the forelimb bud (Fig. 1o). At this stage in the anterior distal region, Sall4 and Tbx5 were coexpressed (Fig. 1q–s). The Tbx5positive boundary was very clear at E12.5 and its posterior boundary was between digits 1 and 2 (Fig. 1p). Sall4 and Tbx5 were coexpressed anterior to the Hoxd12 domain (Fig. 1t–v), which is known to be excluded from the digit-1 primordium27. These results indicate that Sall4 and Tbx5 become progressively colocalized to the anterior distal forelimb, which gives rise to digit 1. Sall4 is regulated by Tbx5 Given the coinciding expression patterns of Tbx5 and Sall4 and the similarity in limb phenotypes in HOS and OS, we investigated a potential link between Tbx5 and Sall4 in the limb. Tbx5+/–mice have a longer digit 1 than wild-type, as well as abnormalities involving the carpal bones4,28. These abnormalities are due to early defects, as we could see clear alterations in digit 1 primordium at E13.5 and E16.5 (Supplementary Fig. 2 online) and changes in Hoxd12 expression at E12.5. Sall4 expression was unchanged in Tbx5+/– mice at E10.5 (data not shown), but by E11, Sall4 expression was specifically decreased in the forelimb of Tbx5+/– mice (Fig. 2a; 30% of Tbx5+/– mice, n ¼ 20); at E12, this decrease was restricted to the anteriormost domain of Sall4 expression. Expression of Sall4 in the hindlimb was identical in wildtype and Tbx5+/– mice (Fig. 2a). Sall4 expression could be readily detected in embryonic heart by RT-PCR (Fig. 2b). We assessed Sall4 expression in wild-type and Tbx5+/– cardiac ventricles at E10.5 by quantitative real-time RT-PCR (Fig. 2c): the amount of Sall4 mRNA in Tbx5+/– hearts was 55 ± 20% of that in wild-type controls (n ¼ 6, P ¼ 0.007). Thus, Sall4 is regulated by Tbx5 in the embryonic forelimb and heart.


[

NUMBER 2

[

Figure 3 Limb defects in Sall4GT/+ mice. (a) Skeletal preparations of postnatal day 1 (P1) wild-type (WT) and Sall4GT/+ mice, showing in Sall4GT/+ mice slight elongation of digit 1 of the forelimb, abnormalities in the shape and size of forelimb metacarpal bones, triphalangeal hindlimb digit 1 and shortened digit 1 (C, capitate; TD, trapezoid; TM, trapezium). (b–g) Altered gene expression in Sall4GT/+ hindlimbs. Unilateral reduction in Fgf10 expression at E9.5 accompanies a smaller hindlimb bud (b). Decreased Tbx4 expression at E9.5 was also observed (c), whereas Pitx1 expression at E10.5 was normal despite a smaller limb bud (d). Anterior expansion of Hoxd12 was observed at E11.5 (e), and decreased expression of EphA5 and EphA3 in digit 1 primordium at E12.5 (f,g).

FEBRUARY 2006

Limb defects in Sall4-deficient mice To study the effect of Sall4 mutations on mouse development, we obtained a gene trap mouse line that interrupts the Sall4 gene. These B6;129P2-Sall4Gt(pGT1Lxf)1Ucd mice (referred to here as Sall4GT/+ mice) have an insertion in exon 2, which is predicted to result in a fusion between the Sall4 protein after amino acid residue 383 and the bgeo protein (Supplementary Fig. 1). Sall4GT/+ mice were born, but were underrepresented at weaning (30% of littermates, rather than 50% as expected). We observed two major classes of phenotypes in Sall4GT/+ mice: type 1 was quite severe and results in 20% of embryos dying by E11.5 of heart defects (although potential placental defects may also contribute), and type 2 was milder, resulting in live-born mice, of which many survived to adulthood. When crossed into the Black Swiss background strain, increased severity of defects was observed. This indicates that the dosage of Sall4 is critical, and that in Sall4GT/+ mice a precarious balance exists between early lethality and survival. Homozygous Sall4GT/GT embryos died before E8.0 and are not discussed further in the present paper. Skeletal preparations of live-born Sall4GT/+ mice (Fig. 3) showed mildly elongated forelimb digit 1 and digit 4 metacarpals and digit 1 proximal phalange, as well as defects in forelimb carpal bone morphology, consistent with those reported in mild cases of OS21. The carpal bone defects included reduced size of the trapezoid and capitate, fusion of the trapezoid with the adjacent digit 1 metacarpal, and abnormal junction of the metacarpal bone of digit 2 with both the trapezoid and trapezium (Fig. 3a). Somewhat surprisingly, more severe defects in hindlimb morphology were observed, including triphalangeal digit 1 or shortened digit 1 (Fig. 3a). Sall4 dosage is therefore critical for patterning or growth of digit 1. In more severe cases, defects in hindlimb bud outgrowth at E9.5 were also observed; hindlimb bud defects were often unilateral, with no preference for laterality of the affected limb (Fig. 3b,d). Analysis of limb gene expression (Fig. 3b–g) showed markedly less expression of Fgf10 and Tbx4, genes required for limb bud outgrowth29,30, in severely affected Sall4GT/+ hindlimb buds. Expression of Pitx1, which lies upstream of Tbx4 (refs. 31–33), was unaffected. In mildly affected hindlimbs, decreased expression of the digit 1 markers EphA3 and EphA5 was observed, and expression of Hoxd12, which is normally restricted to the primordial digits 2–5 (ref. 27), was expanded anteriorly, indicating defects in limb anteriorposterior patterning. Sall4 interacts with Tbx5 in the anterior forelimb We generated Sall4GT/+Tbx5+/– mice and assessed the effect of compound haploinsufficiency on limb morphology (Fig. 4). Sall4GT/+Tbx5+/– mice rarely survived until weaning, and most (80%) died between E11.5 and the perinatal period. Live-born Sall4GT/+Tbx5+/– mice had significantly longer forelimb digit 1

177

ARTICLES

a

+/+ 3

Sall4

GT/+

Tbx5

Gene expression was affected by compound loss of Tbx5 and Sall4 (Fig. 4e–t): at E10.5, Fgf10 expression was lower in Sall4GT/+Tbx5+/– embryos, and at E12, expression of the digit 1 markers Epha3 and Epha5 was reduced in Sall4GT/+Tbx5+/– forelimbs. We conclude that Tbx5 and Sall4 genetically interact to regulate patterning and morphogenesis of digit 1 in the forelimb.

+/–

Sall4

GT/+

Tbx5

Heart patterning defects in Sall4GT/+ mice Sall4GT/+ embryos derived from crosses of Sall4GT/+ mice and Black Swiss wild-type mice had defects in heart formation (Fig. 5). These consisted of a less pronounced or absent interventricular groove, which in the most severe cases resulted in a univentricular heart with thin myocardium (Fig. 5b,e,h), but most often resulted in abnormal formation of the interventricular groove and disorganized myocardium (Fig. 5c,f,i and Fig. 6). Sall4GT/+ mice that survived past E11.5 (type-2 mutants) often had muscular VSDs (Fig. 5j). Sall4 and Tbx5 interacted significantly in regulating heart morphology. We observed severe defects in atrial and ventricular septation in Sall4GT/+ Tbx5+/– mice at E17.5, including perimembranous VSDs (Fig. 5m), multiple muscular VSDs (Fig. 5m,o), defects in atrioventricular valve morphology (Fig. 5n) and complete lack of the atrial septum (Fig. 5p). In one case, the mitral valve was displaced downward toward the apex of the heart (Fig. 5n). These defects cannot be accounted for only by the loss of Tbx5, as Tbx5+/– mice rarely have VSDs, and the severity of Tbx5+/– CHDs is much less pronounced overall4. In our current cohort of Tbx5+/– mice, which have been outbred into the Black Swiss strain for over 15 generations,

+/–

4

2 5 1 PP m

25

c

P < 0.001 P < 0.05

60

WT

T5

S4

S4T5

P < 0.001 P < 0.05

20

WT (n = 14)

15

S4 (n = 20) T5 (n = 12) S4T5 (n = 10)

10

5

0

MD1

3 2

50

40

30

20

10

0

PP1

MD2

Sall4 GT/+

+/+

d

Segment length *100 / femur length

b Segment length *100 / femur length


elements than either Sall4GT/+ or Tbx5+/– mice (Fig. 4a,b). Although other elements were also longer in Tbx5+/– mice (Fig. 4a,c), only the length of digit 1 elements was further altered by loss of one allele of Sall4. Digit 1 elongation in Sall4GT/+Tbx5+/– mice was apparent as early as E12.5 (Fig. 4d), and we occasionally observed severe truncations of the anterior forelimb in Sall4GT/+Tbx5+/– embryos (Fig. 4l).

MD3

Tbx5 +/–

MD4

MD5

Sall4 GT/+Tbx5 +/–

4

5 1

Sall4 GT/+

+/+

Tbx5 +/–


e

f

g

h

i

j

k

l

m

n

o

p

q

r

s

t

Figure 4 Genetic interaction between Tbx5 and Sall4 in patterning the forelimb. (a) Skeletal preparations of forelimbs from wild-type (WT), Sall4GT/+, Tbx5+/– and Sall4GT/+Tbx5+/– mice show increased digit 1 length in Sall4GT/+Tbx5+/– mice. (b) Quantitation of digit 1 element length (segment length 100 / femur length) for all genotypes (MD1, digit 1 metacarpal; PP1, digit 1 proximal phalange). (c) Quantitation of metacarpal length for all digits for all genotypes. (b,c) Bars indicate statistical significance between groups at P o 0.05 (black) or P o 0.01 (red), as determined by ANOVA followed by Tukey’s pairwise comparison. (d) Early digit-1 elongation defects in E12.5 Sall4GT/+Tbx5+/– forelimbs. Forming digits were labeled by in situ hybridization for Sox9 mRNA. Red line shows length of digit 1. (e–t) Altered gene expression in Sall4GT/+, Tbx5+/– and Sall4GT/+Tbx5+/– forelimbs. Decreased Fgf10 expression was detected in Sall4GT/+Tbx5+/– E10.5 forelimbs (h), as well as decreased EphA3 and EphA5 in Sall4GT/+Tbx5+/– digit 1 primordium at E12.5 (red arrowheads in p,t). Severe anterior forelimb bud truncations were occasionally seen in Sall4GT/+Tbx5+/– E12 forelimbs (l, yellow arrowhead).

Fgf10

Hoxd12

EphA5

EphA3

178

VOLUME 38

[

NUMBER 2

[


ARTICLES Sall4 GT/+

+/+

a

b

c

g

h rv

i

Iv

Iv

rv

Iv E10.5

j

Tbx5 +/–

+/+

Sall4 GT/+


E10.5

d

e rv

f rv

Iv

Iv

rv

Iv

k


m

+/+


o

rv

Figure 5 Sall4 and Tbx5 regulate heart patterning. (a–i) WT embryos are shown in a, d and g. Severe (b,e,h) and mild (c,f,i) heart patterning defect in Sall4GT/+ embryos at E10.5. Note pericardial effusion in b, abnormal interventricular groove in both Sall4GT/+ embryos (orange arrowhead), and very thin myocardium in the severely affected embryos. (j) Ventricular septal defects (VSD, red arrowhead) in E16.0 Sall4GT/+ hearts compared to WT or Tbx5+/– hearts, which typically only have atrial septal defects. (k–p) Severe heart defects in E17.5 Sall4GT/+Tbx5+/– mice (m–p), including multiple VSDs (arrowhead and arrows in m,o), larger perimembranous VSD (red asterisk in m), complete loss of the atrial septum (p, blue asterisk), downwards displacement of the mitral valve (n, blue arrowhead).

no VSDs have been observed, and indeed no VSDs were seen in Tbx5+/– littermates of the affected Sall4GT/+Tbx5+/– mice. Molecularly, cardiac patterning defects were apparent at E9.5 in type 1 (severely affected) Sall4GT/+ embryos, when decreased expression of Gja5 was observed (Fig. 6e–g), as well as expansion of the leftventricle marker Nppa across the interventricular groove, from which it is normally excluded (Fig. 6n–p). This was not secondary to gross defects in cardiac gene expression or morphology, as other markers of cardiac differentiation were not affected (Fig. 6a–d), and as evidenced by histology (Fig. 6). This is similar to what is observed in Tbx5+/– Tbx5 a embryos4 (Fig. 6q), except that in Sall4GT/+ embryos, increased left-ventricle Nppa expression accompanies the expanded field of expression, whereas in Tbx5+/– mice, left+/+ Sall4 GT/+ ventricle Nppa expression is reduced. In Gja5 compound heterozygous Sall4GT/+Tbx5+/– e mice, Gja5 expression was more affected than in single heterozygous mice, reflecting an interaction between Tbx5 and Sall4 (Fig. 6i,m), whereas Nppa expression +/+ Sall4 GT/+ resembled that of Sall4GT/+ embryos (Fig. 6r,v). Therefore, Sall4 and Tbx5 activate

rv


[

NUMBER 2

n

Nppa

+/+

p

la

ra la

lv

Gja5 in the heart, and additionally, as exemplified by regulation of Nppa, Sall4 patterns the left ventricle/interventricular septum boundary by repressing gene expression at the interventricular groove. Sall4 activates Fgf10 synergistically with Tbx5 and Tbx4 We used the Fgf10 promoter to examine the transactivation potential of Sall4 and to explore the possibility that Sall4 might associate with Tbx5 to activate transcription (Fig. 7). When a Sall4 expression

b

c

Nkx2-5

Gata4

Sall4 GT/+

+/+

f

k

l

Tbx5 +/–

u rv

Iv

Iv

r

Sall4 GT/+

t Iv

rv

Iv

q

+/+

rv

m rv

Iv

p


Tbx5 +/–

rv

o

Sall4 GT/+

+/+

i

Sall4 GT/+

Iv

Actc

h

+/+

j

d

Sall4 GT/+

+/+

g

Sall4 GT/+

FEBRUARY 2006

n

la

ra

s

[

Iv

Iv

l

rv

Figure 6 Altered gene expression in mutant hearts. (a–d) Expression of most cardiac genes is normal in Sall4GT/+ hearts at E9.5. (e–i) Decreased Gja5 expression in Sall4GT/+, Tbx5+/– and Sall4GT/+Tbx5+/– heart at E9.5 as shown by whole-mount in situ hybridization. (j–m) Gja5 section in situ hybridization; genotypes as in f–i. (n–r) Expanded Nppa expression in Sall4GT/+, Tbx5+/– and Sall4GT/+Tbx5+/– heart at E9.5 (arrowhead). (s–v) Histology of the Nppa-stained embryos shown in o–r (lv, left ventricle; rv, right ventricle).

rv

Iv


v rv

Iv

rv

Iv

179

ARTICLES

80

Fgf10mut1a-luc

80

Relative activation

Relative activation

90

70 60 50 40 30 20

40 30 20

0

80

70

50 40 30 20

Tbx5

Tbx4

–

Tbx1

f

Fgf10-luc Tbx5 (0 ng)

70

Relative activation

Tbx5 (250 ng)

S4-WT

Q

S4-R1 S4-R2

Q

S4-F1

g

Fgf10 -luc

40

122 kDa

40

IB: α-myc

30 20

–

h

WT

R1

Myc-Sall4:

+

Flag-Tbx5:

WT

+

R2

+

F1

+

+

+

+

i Myc-Sall4: HA-Tbx4: IP:

Q49K I54T G80R G169R R237W R252I

Myc

**

10

**

0 –

WT

Q49K

I54T

+ +

Myc

Input 5% (HA-Tbx4)

IB: α-Myc

20

+ +

IB: α-HA

IB: α-Flag

*

30

74 kDa 33.3 kDa

IP:

*

50

50

18.5 kDa

70

F1

IB: α-Flag (Tbx5)

60

Input 5% (Flag-Tbx5)

Sall4 (250 ng)

Sall4 + Tbx5

Myc-Sall4: WT R1 R2 IP: Myc

0 Sall4:

Sall4 (0 ng)

60

Sall4

10

90 80

– Tbx2 (250 ng) Tbx2 (500 ng) Tbx3 (250 ng) Tbx3 (500 ng)

60

0 –

e

Fgf10-luc

80

10

+ ++ + ++ + ++ + ++

d

90

+Sall4 (250 ng)

50

0 Tbx5 Sall4

+ ++ + ++

c

+Sall4 (0 ng)

60

10 + ++ + ++

Fgf10-luc

70

10

Relative activation


b

Fgf10-luc

90

Relative activation

100

a

Input 5% (myc-Sall4)

Input 5% (Flag-Tbx5)

G80R G169R R237W R252I

Figure 7 Sall4 and Tbx5 interact to synergistically activate the Fgf10 promoter. (a) Transactivation by Sall4 and Tbx5 of the Fgf10-luciferase (Fgf10-luc) or Fgf10mut1a-luciferase (Fgf10mut1a-luc) reporter constructs in 10T1/2 cells. The reporter construct was cotransfected with 250 ng (+) or 500 ng (++) of Sall4 or Tbx5 expression constructs. (b) Sall4 and Tbx5 or Tbx4 activated the Fgf10 promoter synergistically; Sall4 and Tbx1 activated it additively. ‘–’ represents ‘no expression construct added’. (c) Tbx2 and Tbx3 reduced the activation of the Fgf10 promoter by Sall4 in a dose-dependent manner and prevented the synergism between Sall4 and Tbx5. (d) Schematic diagram of wild-type and mutant forms of Sall4. Ovals show C2H2 or C2HC zinc-finger motifs, and box labeled Q shows a glutamine-rich region. (e) Mutant forms of Sall4 can activate the Fgf10 promoter, except for R1. Synergistic activation with Tbx5 does not occur for R1 and is reduced for R2 and F1. We used 250 ng of wild-type and mutant forms of Sall4 for transfection. (f) Sall4 can physically interact with Tbx5, as demonstrated by coimmunoprecipitation. IB, immunoblot. (g) When wild-type human TBX5 or TBX5 harboring diseasecausing mutations was transfected (250 ng), the Q49K, G80R and R237W substitutions showed decreased activation. When mutated TBX5 was cotransfected with Sall4, synergistic activation was reduced by the Q49K and I54T mutations, and eliminated by the G80R and R237W mutations. *P o 0.05, **P o 0.01 versus wild-type TBX5. (h) Coimmunoprecipitation of transfected Flag-tagged wild-type TBX5 or mutated TBX5 with the Sall4 expression construct shows that G80R and R237W cannot interact with Sall4. (i) Sall4 can physically interact with Tbx4, as demonstrated by coimmunoprecipitation. HA, hemagglutinin.

construct and a 7-kb construct containing an Fgf10 promoter linked to a luciferase reporter (Fgf10-luciferase) were cotransfected into 10T1/2 cells, activation of Fgf10-luciferase was increased in a Sall4 dose-dependent manner (Fig. 7a). The potential of Sall4 to activate the Fgf10 promoter was similar to that of Tbx5 (Fig. 7a). As expression of Sall4 and Tbx5 overlap in the anterior forelimb region, we thought there was a possibility that Sall4 and Tbx5 might associate during limb patterning. Indeed, cotransfection of Sall4 and Tbx5 expression constructs markedly increased Fgf10-luciferase activation in a synergistic way (Fig. 7a). Mutation of the T-box element responsible for response to Tbx5 (ref. 34) slightly affected activation by Sall4, but eliminated the Tbx5/Sall4 synergism (Fig. 7a). Other Tbx genes are also expressed in the limb: Tbx4 in the hindlimb, and Tbx2 and Tbx3 in the posterior region15,35,36.Tbx4 synergized potently with Sall4, whereas cotransfection of Sall4 with Tbx1 resulted in only additive activation of the Fgf10 promoter (Fig. 7b). Notably, cotransfection of expression constructs for the Tbx2 or Tbx3 repressor

180

transcription factors (refs. 37,38) with the Sall4 expression construct resulted in a lack of activation of Fgf10-luciferase (Fig. 7c). Furthermore, when Tbx2 or Tbx3 expression constructs were cotransfected with both Tbx5 and Sall4 expression constructs, the synergistic activation of Fgf10-luciferase by Sall4 and Tbx5 was completely prevented (Fig. 7c). Our results show that Sall4 and Tbx5 (or Tbx4) synergistically activate transcription, and that Tbx2 and Tbx3 counter this activity. We constructed several mutated forms of Sall4: R1 (K170X) and R2 (K294X) are identical to predicted protein products potentially resulting from mutations that cause OS21,22, and F1, which lacks the N-terminal zinc-finger motif and glutamine-rich region (Fig. 7d). All mutant forms translocated normally to the nucleus (Supplementary Fig. 3 online). When wild-type or mutant forms of Sall4 alone were transfected, R2 and F1 activated the Fgf10 promoter normally, but R1 did not (Fig. 7e). These mutant forms had an impaired ability to activate the Fgf10 promoter synergistically with Tbx5, although R2

VOLUME 38

[

NUMBER 2

[


ARTICLES

40

Nppa-luc

b

Gja5-luc

40

30 25 20 15 10 5

30 25 20 15 10

0 50 0 1, 00 0

0

0 Sall4 Tbx5 Nkx2–5 Gata4

25

–

10

0 00 0

0

50

25

10

0

5

1,


0 Sall4 (ng): –

c

Nppa-luc

35 Relative activation

Relative activation

35

+ +

Figure 8 Cooperative and counteracting interactions between Sall4 and Tbx5 on cardiac promoters. (a) Sall4 can activate Gja5luciferase, but not Nppa-luciferase. ‘–’ represents ‘no expression construct added’. (b) Sall4 interferes with Tbx5-dependent activation of Nppa-luciferase, but does not affect Gata4- or Nkx2-5dependent activation. (c) Sall4 synergizes with Tbx5 but not Gata4 or Nkx2-5 to activate Gja5-luciferase. (d) Coimmunoprecipitation in COS-7 cells shows that Sall4 physically interacts with Tbx5 but not Nkx2-5 or Gata4. White asterisk in third lane indicates nonspecific background signal. (e) Model for Tbx5/Sall4 feed-forward network in the developing heart (created with Biotapestry).

++ + + + + + +

NUMBER 2

[

FEBRUARY 2006

+ ++ ++

d

+ ++ ++

+ ++ ++ ++

+

+

+ ++ ++

+ ++ ++

Gja5

50 kDa 36 kDa 57 kDa 48 kDa 35 kDa

5% input α-myc

+ ++ ++

+ 60 kDa

α-Flag

++ + + + +

e

f-Tbx5 f-Gata4 f-Nkx2–5

IP:α-myc IB:α-Flag

Sall4 counteracts activation of Nppa by Tbx5 We examined the response of the Nppa and Gja5 promoters to Sall4. The well-defined Nppa promoter responds to several cardiac transcription factors, including Tbx5 and its interacting partners Gata4 and Nkx2-5 (refs. 3–5,41). The Gja5 promoter also responds to Tbx5 and Nkx2-5 (ref. 4). Sall4 did not activate, and at higher doses suppressed, activity of Nppa-luciferase, whereas Gja5 transcription

[

+ ++ ++

mys-Sall4:

could still synergize with Tbx5 albeit less efficiently (Fig. 7e). These results show that several domains of Sall4 are important for transcriptional activation, and the most N-terminal zinc-finger motif is not sufficient to do so. We investigated the physical interactions between Tbx5 and Sall4 by coimmunoprecipitation of epitope-tagged Tbx5 and Sall4 coexpressed in COS-7 cells (Fig. 7f). The Sall4 R1 and F1 mutants interacted only weakly with Tbx5, whereas R2 interacted well with Tbx5 (Fig. 7f). This correlates well with the inability of R1 and F1 to synergize with Tbx5 (only additive activation was observed), whereas R2 could still synergize with Tbx5 (Fig. 7e). These results indicate that the Tbx5interacting domain of Sall4 corresponds primarily to the region between the first two zinc finger domains. Tbx4 could also physically interact with Sall4 (Fig. 7i). Several human TBX5 missense mutations have been reported to be involved in HOS6. We examined whether human TBX5 proteins harboring these mutations maintain the potential to synergize with Sall4 on the Fgf10 promoter. Most missense mutations did not prevent the synergistic activation of Fgf10, but the G80R and R237W mutations completely prevented synergism with Sall4, and synergism was also significantly reduced by the Q49K and I54T mutations (Fig. 7g). The TBX5 mutants G80R and R237W have reduced, but not absent, DNA-binding ability39,40, and all mutant proteins tested localize to the nucleus and can interact with other transcription factors (data not shown). We then examined whether these Tbx5 mutants can physically interact with Sall4. The Tbx5 mutants Q49K and I54T could associate with Sall4, but the G80R and R237W mutations abolished this interaction (Fig. 7h). The G80R and R237W mutations are both located in the T-box domain of TBX5, which suggests that interactions between Tbx5 and Sall4 occur via the Tbx5 T-box.


45 Gja5-luc 40 35 30 25 20 15 10 5 0 Sall4 + Tbx5 + Nkx2–5 + Gata4 + Relative activation

a

Tbx5

Sall4

122 kDa

Nkx2–5/Gata4

Nppa

was activated by Sall4 in a dose-dependent manner (Fig. 8a). Consistent with our in vivo data, Sall4 counteracted the activation of the Nppa promoter by Tbx5; synergistic activation of Nppa by Tbx5 with Gata4 and/or Nkx2-5 was also markedly suppressed by Sall4 (Fig. 8b). Sall4 did not, however, affect activation of Nppa-luciferase by Nkx2-5 and Gata4, indicating that its repressive effects are specific to activation by Tbx5 (Fig. 8b). Furthermore, synergistic activation of Gja5 by Sall4 and Tbx5 was achieved, but only additive interactions of Sall4 with Nkx2-5 or Gata4 could be observed (Fig. 8c). Indeed, coimmunoprecipitation showed that although Sall4 interacts with Tbx5, it does not interact with Gata4 or Nkx2-5 (Fig. 8d). DISCUSSION The interaction between Sall4 and Tbx5 in patterning and morphogenesis of the anterior limb provides an important missing component in our understanding of limb development. Our data support a mechanism by which patterning and growth of digit 1 is accomplished by an interaction between Tbx5 and Sall4. Indeed, not only do the two coexpressed proteins show physical interaction to activate transcription, but Sall4 and Tbx5 also genetically interact to influence anterior forelimb gene expression and morphogenesis. As formation of digit 1 is independent of Shh signaling11,16–18 and may require inhibition of BMP signaling14,15, the combined actions of Tbx5 and Sall4 may therefore be important to refine the patterning of the anterior forelimb in combination with Shh and BMP signaling and to counteract the influence of Tbx2 and Tbx3. Our data thus provide a framework for understanding digit 1 specification and growth in the forelimb. Significant hindlimb defects were seen in Sall4GT/+ mice. Because Tbx5 is specifically expressed in forelimbs, whereas Tbx4 is expressed in hindlimbs35,36, it is likely that Sall4 interacts in the hindlimb with Tbx4. The T-box domains of Tbx4 and Tbx5 are well conserved, and here we show that Tbx4 also has the potential to activate Fgf10 with Sall4. These facts support the idea that Tbx4 and Tbx5 are functionally equivalent, and that the different molecular environments of the forelimb and hindlimb are responsible for their identity and morphology42. In OS, malformations of the hindlimb are also reported, but the phenotypes are different from that of the forelimb23,24. Unlike Tbx5, Tbx4 does not show a digit-1-specific expression pattern (data

181


ARTICLES not shown). Such differences of expression of T-box genes might be important for limb patterning. Importantly, the decreased expression of Tbx4 in Sall4-deficient hindlimbs results in a compounded loss of Sall4- and Tbx4-dependent activation of target genes, resulting in more profound defects than in the forelimb, where Sall4 loss is not accompanied by a decrease in Tbx5 expression. This suggests that Sall4 is a primary regulator of hindlimb outgrowth, in part through regulation of Tbx4 (directly or indirectly) as well as direct activation of Fgf10 in conjunction with Tbx4. This hindlimb-specific, Sall4dependent, feed-forward interaction with Tbx4 may explain the different requirements for Tbx4 and Tbx5 in limb bud formation30,34,43: as Sall4 acts upstream of Tbx4 in the hindlimb but is downstream of Tbx5 in the forelimb, more pronounced effects on Fgf10 expression are seen in the forelimb due to loss of Tbx5 because of the concomitant loss of Sall4. In the heart, Sall4 activates genes such as Gja5 through its interaction with Tbx5, but it also takes on an additional role. We find that Sall4 also acts as a repressor of Tbx5-dependent activation of specific genes (such as Nppa), and its predominant expression in the interventricular septum (IVS) results in more pronounced repressor effects in the IVS than in the working myocardium. This creates a boundary of gene expression between the left ventricle and the IVS. As in the limb, Sall4 is positively regulated by Tbx5 in the heart, but in the case of the heart, this relay and interaction sets up a clear boundary definition, with Tbx5 increasing expression of a corepressor at the border of its expression domain. Therefore, decreased expression of Tbx5 in Tbx5+/– hearts reduces Sall4 levels, resulting in de-repression of Nppa in the IVS; conversely, decreased Sall4 in Sall4GT/+ hearts allows Tbx5 (or other activating factors) to act unhindered on Nppa, thus resulting in its increased expression in the IVS. Nppa expression in the left ventricle working myocardium is increased in Sall4GT/+ hearts, while it is decreased in Tbx5+/– hearts, again consistent with the proposed roles for each factor. Thus, Sall4 assists Tbx5 in refining cardiac morphogenetic boundaries by placing strict constraints on the location of Tbx5-dependent transcription. In the disease state, relief of Sall4-dependent transcription by Sall4 haploinsufficiency (OS) or by reduction in Sall4 expression due to Tbx5 haploinsufficiency (HOS) may results in a patterning defect that is manifest as ventricular septal defects. Indeed, Sall4GT/+ mice had muscular VSDs, similar to some cases of OS with SALL4 mutations23–25. The expression of Sall4 in endocardial cushions and the cushion defects in Sall4GT/+Tbx5+/– mice may also provide an explanation for more severe defects such as tetralogy of Fallot and truncus arteriosus, which have been reported in individuals with SALL4 mutations24,25. Our mutational analysis of Sall4 indicates that inefficient activation of target genes in concert with Tbx5 contributes to the limb malformations in OS. Sall4 R2, which truncates after the glutamine-rich domain, activates the Fgf10 promoter by itself, but does not fully synergize with Tbx5. Similarly, several TBX5 missense mutations abrogated the activation of the Fgf10 promoter or the interaction with Sall4. Thus, the lack of physical interaction between Tbx5 and Sall4 is likely to be critical for digit 1 and interventricular septum formation. The TBX5 Q49K and I54T mutations reduced the potential of Tbx5 to synergize with Sall4, but these mutants could still interact with Sall4. It is likely that these mutations induce a conformational change that prevents the synergism between Tbx5 and Sall4, while still maintaining their physical association. Regardless of the mode of action of the mutations, it is clear that proper stoichiometry between Tbx5 and Sall4 is key to patterning the anterior forelimb. As TBX5 mutations also prevent interaction with GATA4, a cardiac zinc-finger transcription factor3, our results provide a unifying mechanism for

182

limb and heart defects in HOS via impaired interactions with tissuespecific transcription factors. In conclusion, we have found a complex interplay between Sall4 and T-box genes in the patterning and morphogenesis of the mouse limb and heart, in which Sall4 is regulated by Tbx5, and both Tbx5 and Sall4 interact both positively and negatively to modulate transcription. This has the characteristics of a feed-forward circuit (Fig. 8e), which has been shown to be critical for transcriptional networks in lower eukaryotes and in human embryonic stem cells44–46. This network thus provides a powerful regulatory pathway that ensures precise patterning of embryonic structures. METHODS Mouse Sall4 gene. We searched mouse GenBank sequences by BLAST using human SALL4 sequence and one National Institute of Aging EST clone (H3023E05) showed the highest homology. As this gene also has homology to other Sall genes and its positioning region on the chromosome (2H3) shared conserved synteny with that of human SALL4 (20q13.13-2), we concluded that it was mouse Sall4. A BLAST search of GenBank using the human SALL4 amino acid sequence provided a full sequence of mouse Sall4 (BC33598). Mice. Tbx5+/–mice were used for experiments, and breeding and genotyping were performed as described previously4. B6;129P2-Sall4Gt(pGT1Lxf)1Ucd (Sall4GT/+) mice were obtained from Baygenomics. Embryo stages were defined according to Kaufman47, using morphological landmarks to estimate the embryonic day of development. Skeletal preparations were performed as previously described4; length of skeletal elements in P0 or P1 mice was calculated as segment length times 100, divided by femur length. Analysis of gene expression. In situ hybridization was performed according to standard protocols, and for series of in situ hybridizations, multiple littermates were used. For in situ hybridization using sections, we used 10 mm paraffin sections. Quantitative real-time RT-PCR was performed using a Sall4 Assay-ondemand Taqman probe (Applied Biosystems) and a rodent GAPDH control Taqman probe (Applied Biosystems). Primer sequences are available on request. Luciferase assays and coimmunoprecipitation. Luciferase assays and coimmunoprecipitation experiments were performed essentially as described previously4,34,48. The pCAGGS-Tbx4 construct was provided by T. Ogura (Tohoku University, Sendai, Japan), pcDNA3.1-Tbx1 was from A. Baldini (Baylor College of Medicine, Houston, Texas) and pcDNA3.1-hTBX2 and pcDNA3.1-hTBX3 by V. Christoffels (University of Amsterdam, The Netherlands). Flag-tagged mutant TBX5 constructs were prepared as previously described3. Myc-tagged full length and mutant Sall4 constructs, as well as HA-tagged Tbx4, were generated by PCR and cloned into a pcDNA3 vector. Statistical analysis. Data were analyzed by one-way ANOVA followed by Tukey’s pairwise post-hoc test, or by unpaired t-test. The threshold for significance was P o 0.05. Accession codes. Online Mendelian Inheritance in Man (OMIM): Holt-Oram syndrome, 142900; Okihiro syndrome, 607323. URLs. Biotapestry: http://www.biotapestry.org; baygenomics.ucsf.edu/.

Baygenomics:

http://

Note: Supplementary information is available on the Nature Genetics website. ACKNOWLEDGMENTS We are grateful to A. Mori for help with statistics and quantitative RT-PCR. We also thank V. Christoffels, M. Nemer and T. Ogura for expression vectors and reporter constructs, and A. Brown, D. Duboule, M. Logan, G. Martin and C. Oka for in situ probes. This work was supported by grants from the Canadian Institutes of Health Research (B.G.B., C.c.H), the Heart and Stroke Foundation of Ontario (B.G.B.) and the March of Dimes Birth Defects Foundation (B.G.B.). K.K.-T. was supported by the Uehara Memorial Foundation. J.K.T holds a longterm fellowship from the Human Frontiers Science Program and was partly supported by the Mochida Memorial Foundation for Medical and Pharmaceutical Research. B.G.B. holds a Canada Research Chair in Developmental Cardiology.

VOLUME 38

[

NUMBER 2

[


ARTICLES COMPETING INTERESTS STATEMENT The authors declare that they have no competing financial interests. 24.

Published online at http://www.nature.com/naturegenetics Reprints and permissions information is available online at http://npg.nature.com/ reprintsandpermissions/

25.


26. 1. Seidman, J.G. & Seidman, C. Transcription factor haploinsufficiency: when half a loaf is not enough. J. Clin. Invest. 109, 451–455 (2002). 2. Bruneau, B.G. Transcriptional regulation of vertebrate cardiac morphogenesis. Circ. Res. 90, 509–519 (2002). 3. Garg, V. et al. GATA4 mutations cause human congenital heart defects and reveal an interaction with TBX5. Nature 424, 443–447 (2003). 4. Bruneau, B.G. et al. A murine model of Holt-Oram syndrome defines roles of the T-box transcription factor Tbx5 in cardiogenesis and disease. Cell 106, 709–721 (2001). 5. Hiroi, Y. et al. Tbx5 associates with Nkx2–5 and synergistically promotes cardiomyocyte differentiation. Nat. Genet. 28, 276–280 (2001). 6. Mori, A.D. & Bruneau, B.G. TBX5 mutations and congenital heart disease: Holt-Oram syndrome revealed. Curr. Opin. Cardiol. 19, 211–215 (2004). 7. Poznanski, A.K., Gall, J.C. Jr. & Stern, A.M. Skeletal manifestations of the Holt-Oram syndrome. Radiology 94, 45–53 (1970). 8. Basson, C.T. et al. The clinical and genetic spectrum of the Holt-Oram syndrome (heart-hand syndrome). N. Engl. J. Med. 330, 885–891 (1994). 9. Newbury-Ecob, R.A., Leanage, R., Raeburn, J.A. & Young, I.D. Holt-Oram syndrome: a clinical genetic study. J. Med. Genet. 33, 300–307 (1996). 10. Tickle, C. Patterning systems–from one end of the limb to the other. Dev. Cell 4, 449– 458 (2003). 11. Litingtung, Y., Dahn, R.D., Li, Y., Fallon, J.F. & Chiang, C. Shh and Gli3 are dispensable for limb skeleton formation but regulate digit number and identity. Nature 418, 979– 983 (2002). 12. te Welscher, P. et al. Progression of vertebrate limb development through SHHmediated counteraction of GLI3. Science 298, 827–830 (2002). 13. Drossopoulou, G. et al. A model for anteroposterior patterning of the vertebrate limb based on sequential long- and short-range Shh signalling and Bmp signalling. Development 127, 1337–1348 (2000). 14. Dahn, R.D. & Fallon, J.F. Interdigital regulation of digit identity and homeotic transformation by modulated BMP signaling. Science 289, 438–441 (2000). 15. Suzuki, T., Takeuchi, J., Koshiba-Takeuchi, K. & Ogura, T. Tbx genes specify posterior digit identity through Shh and BMP signaling. Dev. Cell 6, 43–53 (2004). 16. Chiang, C. et al. Manifestation of the limb prepattern: limb development in the absence of sonic hedgehog function. Dev. Biol. 236, 421–435 (2001). 17. Ahn, S. & Joyner, A.L. Dynamic changes in the response of cells to positive hedgehog signaling during mouse limb patterning. Cell 118, 505–516 (2004). 18. Harfe, B.D. et al. Evidence for an expansion-based temporal Shh gradient in specifying vertebrate digit identities. Cell 118, 517–528 (2004). 19. Bamshad, M. et al. Mutations in human TBX3 alter limb, apocrine and genital development in ulnar-mammary syndrome. Nat. Genet. 16, 311–315 (1997). 20. Davenport, T.G., Jerome-Majewska, L.A. & Papaioannou, V.E. Mammary gland, limb and yolk sac defects in mice lacking Tbx3, the gene mutated in human ulnar mammary syndrome. Development 130, 2263–2273 (2003). 21. Kohlhase, J. et al. Okihiro syndrome is caused by SALL4 mutations. Hum. Mol. Genet. 11, 2979–2987 (2002). 22. Al-Baradie, R. et al. Duane Radial Ray syndrome (Okihiro syndrome) maps to 20q13 and results from mutations in SALL4, a new member of the SAL family. Am. J. Hum. Genet. 71, 1195–1199 (2002). 23. Kohlhase, J. et al. Mutations at the SALL4 locus on chromosome 20 result in a range of clinically overlapping phenotypes, including Okihiro syndrome, Holt-Oram syndrome,


[

NUMBER 2

[

FEBRUARY 2006

27. 28. 29. 30.

31. 32. 33.

34. 35.

36. 37.

38. 39. 40. 41.

42.

43. 44. 45. 46. 47. 48.

acro-renal-ocular syndrome, and patients previously reported to represent thalidomide embryopathy. J. Med. Genet. 40, 473–478 (2003). Borozdin, W. et al. Novel mutations in the gene SALL4 provide further evidence for acro-renal-ocular and Okihiro syndromes being allelic entities, and extend the phenotypic spectrum. J. Med. Genet. 41, e102 (2004). Brassington, A.M. et al. Expressivity of Holt-Oram syndrome is not predicted by TBX5 genotype. Am. J. Hum. Genet. 73, 74–85 (2003). Kohlhase, J. et al. Cloning and expression analysis of SALL4, the murine homologue of the gene mutated in Okihiro syndrome. Cytogenet. Genome Res. 98, 274–277 (2002). Spitz, F., Gonzalez, F. & Duboule, D. A global control region defines a chromosomal regulatory landscape containing the HoxD cluster. Cell 113, 405–417 (2003). Pizard, A. et al. Connexin 40, a target of transcription factor Tbx5, patterns wrist, digits and sternum. Mol. Cell. Biol. 25, 5073–5083 (2005). Sekine, K. et al. Fgf10 is essential for limb and lung formation. Nat. Genet. 21, 138–141 (1999). Naiche, L.A. & Papaioannou, V.E. Loss of Tbx4 blocks hindlimb development and affects vascularization and fusion of the allantois. Development 130, 2681–2693 (2003). Szeto, D.P. et al. Role of the Bicoid-related homeodomain factor Pitx1 in specifying hindlimb morphogenesis and pituitary development. Genes Dev. 13, 484–494 (1999). Logan, M. & Tabin, C.J. Role of Pitx1 upstream of Tbx4 in specification of hindlimb identity. Science 283, 1736–1739 (1999). Lanctot, C., Moreau, A., Chamberland, M., Tremblay, M.L. & Drouin, J. Hindlimb patterning and mandible development require the Ptx1 gene. Development 126, 1805–1810 (1999). Agarwal, P. et al. Tbx5 is essential for forelimb bud initiation following patterning of the limb field in the mouse embryo. Development 130, 623–633 (2003). Gibson-Brown, J.J. et al. 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 (1996). Logan, M. Finger or toe: the molecular basis of limb identity. Development 130, 6401–6410 (2003). Habets, P.E. et al. Cooperative action of Tbx2 and Nkx2.5 inhibits ANF expression in the atrioventricular canal: implications for cardiac chamber formation. Genes Dev. 16, 1234–1246 (2002). Hoogaars, W.M. et al. The transcriptional repressor Tbx3 delineates the developing central conduction system of the heart. Cardiovasc. Res. 62, 489–499 (2004). Ghosh, T.K. et al. Characterization of the TBX5 binding site and analysis of mutations that cause Holt-Oram syndrome. Hum. Mol. Genet. 10, 1983–1994 (2001). Fan, C., Liu, M. & Wang, Q. Functional analysis of TBX5 missense mutations associated with Holt-Oram syndrome. J. Biol. Chem. 278, 8780–8785 (2003). Durocher, D., Charron, F., Warren, R., Schwartz, R.J. & Nemer, M. The cardiac transcription factors Nkx2–5 and GATA-4 are mutual cofactors. EMBO J. 16, 5687– 5696 (1997). Minguillon, C., Del Buono, J. & Logan, M.P. Tbx5 and Tbx4 are not sufficient to determine limb-specific morphologies but have common roles in initiating limb outgrowth. Dev. Cell 8, 75–84 (2005). Rallis, C. et al. Tbx5 is required for forelimb bud formation and continued outgrowth. Development 130, 2741–2751 (2003). Shen-Orr, S.S., Milo, R., Mangan, S. & Alon, U. Network motifs in the transcriptional regulation network of Escherichia coli. Nat. Genet. 31, 64–68 (2002). Mangan, S. & Alon, U. Structure and function of the feed-forward loop network motif. Proc. Natl. Acad. Sci. USA 100, 11980–11985 (2003). Boyer, L.A. et al. Core transcriptional regulatory circuitry in human embryonic stem cells. Cell 122, 947–956 (2005). Kaufman, M.H. The Atlas of Mouse Development (Academic, London, 1992). Lickert, H. et al. Baf60c is essential for function of BAF chromatin remodelling complexes in heart development. Nature 432, 107–112 (2004).

183