Aims Holt–Oram syndrome (HOS) is a heart/hand syndrome clinically characterized by upper limb and cardiac malformations. Mutations in T-box transcription factor 5 (TBX5) underlie this syndrome, the majority of which lead to premature stops. In this study, we present our functional analyses of five (novel) missense TBX5 mutations identified in HOS patients, most of whom presented with severe cardiac malformations.
Methods and results Functional characterization of mutant proteins shows a dramatic loss of DNA-binding capacity, as well as diminished binding to known cardiac interaction partners NKX2-5 and GATA4. The disturbance of these interactions leads to a loss of function, as measured by the reduced activation of Nppa and FGF10 in rat heart derived cells, although with variable severity. Two out of the five mutations are peculiar: one, p.H220del, is associated with additional extra-cardiac defects, perhaps by interfering with other T-box dependant pathways, and another, p.I106V, leads to limb defects only, which is supported by its normal interaction with cardiac-specific interaction partners.
Conclusion Overall, our data are consistent with the hypothesis that these novel missense mutations in TBX5 lead to functional haploinsufficiency and result in a reduced transcriptional activation of target genes, which is likely central to the pathogenesis of HOS.
Holt–Oram syndrome (HOS) is a hand–heart syndrome that segregates in an autosomal dominant fashion and is characterized by upper limb anomalies and congenital heart defects (CHD).1,2 Upper limb anomalies involve the preaxial radial ray and are often bilateral and asymmetric. The most common forms of CHD associated with HOS are atrial septal defect (ASD), usually of the ostium secundum variety (ASD II), and ventricular septal defects (VSDs). Cardiac conduction disease and atrial fibrillation can also occur, regardless of the presence or absence of CHD.2–4 HOS patients may suffer from limb malformations, heart defects or both, the severity of which varies substantially, even within families.5 No obvious correlation exists between the severity of the cardiac and skeletal abnormalities of the patient.6
HOS is linked to single-gene mutations in the transcription factor T-box transcription factor 5 (TBX5), a member of the evolutionary conserved T-box family of transcription factors.7,8 Mutations in several other T-box genes are associated with malformations such as cardiac septal defects and dilated cardiomyopathy (TBX20),9 and syndromes such as DiGeorge syndrome (TBX1) and Ulnar–Mammary syndrome (TBX3).10 Seventy per cent of the identified TBX5 mutations lead to a premature stop codon and in these patients HOS is presumably caused by haploinsufficiency. Although HOS-associated mutations are distributed across all exons of TBX5, the majority are found within the T-box DNA-binding domain.11 Nevertheless, there is no evidence that either the type of mutation or the location of a mutation is predictive for the severity of heart or limb malformations in HOS patients.12
The T-box is involved in DNA-binding and protein–protein interactions.13,14 Several HOS mutations within the T-box of TBX5, including G80R, R237Q, R237W affect DNA-binding and interaction with NKX2-5 and GATA4, resulting in reduced activation of downstream targets in cardiac development like NPPA and CX40.15–17 Recently, a gain of function mutation in TBX5 was described, G125R, that lead to enhanced DNA-binding and induction of NPPA and CX40 expression.4 Tbx5 also synergizes with Sall4 in the regulation of Fgf10 expression in the forelimbs via direct effects on the Fgf10 promoter.18
In this study, we present five mutations in TBX5 identified in patients with HOS. Four of these are novel mutations, which were identified in patients with multiple cardiac malformations. Functional analysis shows that four mutations encode proteins which fail to bind DNA, have reduced binding to NKX2-5 and GATA4 and are functionally deficit, as measured by the failure to activate the Nppa and Fgf10 genes. In addition to this, we identified a previously published TBX5 mutation (p.Ile106Val)19 in a patient with upper limb defects only, which was also present in two family members without a history of limb or heart defects. In line with this, the interaction of this mutant protein with cardiac-specific protein partners and binding to an essential cardiac NPPA promoter fragment were similar to wildtype (WT) TBX5. We propose that this mutation leads to a loss of limb-specific function, the expressivity of which may be very mild depending on the genetic background.
2.1 Patients and molecular analysis
Patients with a clinical diagnosis of HOS were referred to the Department of Clinical Genetics of Erasmus MC, Rotterdam or AMC Amsterdam, the Netherlands. Family studies were initiated for all TBX5 mutation patients and included genetic counselling, physical, and cardiologic evaluation of first and second-degree relatives when appropriate. Further functional molecular characterization was performed on mutations leading to small deletions or single amino acid substitutions. This study was approved by the Medical Ethical Committees of Academic Medical Center Amsterdam and Erasmus Medical Center Rotterdam, written informed consent was obtained from all participants. The investigation conforms with the principles outlined in the Declaration of Helsinki.
2.2 Genetic testing
Genomic DNA was isolated from blood samples using standard procedures. Coding regions and intron–exon boundaries of TBX5 (NM_000192.3) were analysed using direct sequence analysis. M13-tagged PCR products were sequenced on an ABI3730xl capillary sequencer using Big-Dye Terminator v3.1 (Applied Biosystems). Data were analysed using SeqScape analysis software (v2.5, Applied Biosystems). To exclude genomic rearrangements in TBX5, MLPA analysis was performed (MRC Holland kit P180). Identified mutations were absent from 600 chromosomes of ethnically matched control individuals.
2.3 Plasmid constructs and transfections
Constructs encoding TBX5 fused to maltose-binding protein (MBP; pMAL2C-TBX5-T-box) and NKX2-5 fused to GST (pRP265nb-NKX2.5) have been described before.20 Constructs encoding MBP-TBX5-mutants were constructed by PCR using pcDNA-based expression plasmids mentioned below as a template. Eukaryotic expression vectors pcDNA-flag-TBX5, pcDNA-flag-NKX2-5, pcDNA-myc-NKX2-5, and pcDNA-myc-GATA4, as well as the Nppa luciferase reporter construct have been described before.4,17,21pcDNA-Flag-TBX5-mutants were constructed using site-directed mutagenesis. PCR generated constructs were fully verified by sequencing. Transfections were performed using polyethylenimine (25 kDa, linear, Brunschwick).
2.4 Nuclear localization
Immortalized rat neonatal heart derived cells (H10 cells22) were seeded on cover slips in standard 12-wells plates and transfected with 500 ng WT or mutant pcDNA-flag-TBX5. 24 h post-transfection, cells were fixed in 2% paraformaldehyde, permeabilized using 0.3% Triton X-100, and incubated with mouse anti-flag (stratagene, 1:500 for 24 h) and Alexa488-conjugated goat anti-mouse antibodies (Molecular Probes, 1:500 for 1 h), as described before.4
2.5 Electrophoretic mobility shift assay
Non-radioactive electrophoretic mobility shift assay (EMSA) was performed using bacterially expressed, purified WT and mutant MBP-TBX5-T-box, as described before,20 with an oligonucleotide containing the T-box binding element (TBE) at position −250 of the Nppa promoter (5′-TCTGCTCTTCTCACACCTTTGAAGTGGGGGCCTCTTG).
2.6 Binding assays
Escherichia coli BL21 cells were transformed with bacterial expression constructs. Cells were induced with 1 mM isopropyl-β-d-thiogalactopyranoside (IPTG) (Gibco-BRL) and after 2 h growth at 30°C, harvested by centrifugation and resuspended in 5 mL of ice-cold phosphate-buffered saline containing 0.05% v/v Triton X-100 (Sigma) (PBS-Tr). Cell suspensions were lysed by sonication and centrifuged to pellet cell debris. GST containing fusion constructs were purified on glutathione 4B-Sepharose following the manufacturer's instructions (Pharmacia). Binding assays were set-up as described previously23 except that a total 2 μg of target GST-fusion was passed over the amylose column in 1 mL PBS-Tr.
HEK cells were transfected with a combination of pcDNA-based expression vectors, using empty pcDNA3.1 vector as input correction or negative control. Cells were harvested 48 h post-transfection and lysed in lysis buffer (150 mM NaCl, 10 mM NaPO4, pH 7.4, 0.2% Triton X-100, 1 mM EDTA, 10% glycerol) for 30 min followed by two short sonification pulses. Cell lysates were cleared by centrifugation (16 000 g, 15 min), incubated 2 h with M2-anti-flag-beads (Sigma) and washed three times with lysis buffer. Proteins were eluted by addition of two times sample buffer (120 mM Tris–HCl pH 6.8, 4% SDS, 20% glycerol, 10% 2-mercaptoethanol, bromophenol blue). Samples were run on 12% polyacrylamide gel (SDS–PAGE) and blotted onto 0.45 μm polyvinylidene fluoride membrane (PVDF; Immobilon P, Millipore). Incubations were performed in blocking buffer [2% protifar plus (Nutricia), 50 mM Tris pH7.5, 150 mM NaCl, 0.1% Tween-20 (Sigma)]. Immunodetection was performed using mouse anti-FLAG (M2, stratagene), rabbit anti-myc (sigma) or goat anti-Gata4 (E-20; Santa-Cruz), and appropriate horse-radish peroxidase conjugated secondary antibodies (GAM-HRP, DAR-HRP or DAG-HRP). Blots were visualized using enhanced chemiluminescence (Amersham), recorded with a LAS-3000 imager (FujiFilm) and analysed using image analysis software (AIDA v3.44, Raytest). Results of three independent experiments were subjected to statistical analysis using two-tailed t-test.
2.8 Luciferase assay
Neonatal rat heart myocytes, immortalized with a temperature-sensitive SV40 antigen (H10 cells22), grown in standard 12-wells plates in DMEM supplemented with 10% FCS (Gibco-BRL) and glutamine, were transfected in triplicate. Standard 700 ng Nppa-luciferase construct was co-transfected with 3 ng of phRG-TK vector, as normalization control (Promega), together with appropriate combinations of pcDNA3.1 constructs. Measurements were performed on a Glomax E9031 luminometer. Duplo transfection experiments were repeated at least three times for each condition, data were corrected for intersession variation as described.24 Statistical analysis was performed using two-tailed t-test, P < 0.05 was considered significant.
2.9 Quantitative PCR
H10 and H9C225 cells were plated on standard 6-well plate the day prior to transfection with 5 μg of expression construct. Forty-eight hour post-transfection total-RNA was isolated using Nucleospin RNA II kit (Clontech). Isolated RNA was reverse transcribed using oligo(dT) primer and Superscript II RT–PCR kit (Invitrogen). Expression of target genes was quantified using quantitative PCR (qPCR) on a LightCycler480 (Roche). Samples were measured in triplicate, experiments were repeated at least three times. Quantification was performed using LinReg software26 and factor correction for intersession variation24 followed by statistical analysis using two-tailed t-test.
Molecular analysis of TBX5 in suspected HOS patients resulted in the identification of five missense mutations, leading to p.Met74Ile, p.Leu94Arg, p.Ile106Val, p.His220del, and p.Arg237Pro mutant proteins. Genomic rearrangements were investigated using MLPA, however, none were found. The clinical features in the probands and their affected relatives are consistent with previous descriptions of HOS (Table 1).1,2 All had upper limb malformations, and apart from the p.Ile106Val mutation, also congenital cardiac anomalies. However, significant differences were observed in the expression of limb and cardiac phenotypes resulting from distinct TBX5 mutations.
aDNA reference sequence NM_000192.3, protein reference sequence NP_852259.
3.1 p.Met74Ile (M74I)
This patient presented with a small VSD, left ventricular hypertrophy, and symptomatic ventricular tachycardia for which an implantable cardioverter defibrillator was implanted at the age of 60. Congenital defects of upper extremities were present. On the right side the index finger and thumb were fused and bones were absent from the thumb. On both hands the patient is unable to flex his fingers. Supination of both the forearms is limited and the patient has sloping shoulders. The mutation was absent in DNA from both parents (Table 2).
Molecular characterization of HOS-associated mutations from this and other studies
Loss of function
Loss of function
Nuclear and cytoplasmic
Loss of function
Loss of function
Loss of function
Loss of function
Loss of function
Nuclear and cytoplasmic
Loss of function
Gain of function
−, reduced; +, enhanced; =, no change.
aUsing GST or MBP pulldown and CoIP (this study and Fan et al.16) or using in vitro translated proteins17.
bIn H10 cells (this study) or NIH-3T3 cells.16
cLoss of function: activation (with or without Nkx2.5) is reduced compared with WT TBX5.
3.2 p.Leu94Arg (L94R)
The index presented with classical clinical signs of HOS. The patient has an ASD and a small VSD. On both hands she has a triphalangeal non-opposable finger-like thumb. The mutation was absent in DNA from both parents.
3.3 p.His220del (H220del)
The index presented with an aplasia of the radius on the right side, a shortened radius and ulna on the left side, bilateral tri-phalangeal thumbs and 11 pairs of ribs. Cardiac defects were severe including a complete atrioventricular septal defect (AVSD), multiple muscular VSD's, hypoplastic right ventricle and an atrioventricular valve insufficiency. Additional vascular defects included pulmonary artery stenosis and hypoplasia of pulmonary veins. The cardiac malformations were inoperable and the patient died at the age of 11 months. She had a normal female karyotype (46,XX). The mutation was absent in both parental DNA samples.
3.4 p.Arg237Pro (R237P)
The index presented prenatally with a suspected ASD. Postnatal an ASD II and multiple VSDs were confirmed for which patient was operated at the age of 4.5 months. After surgery there was a complete AV-block requiring an external pace-maker. Her hand phenotype was very mild with somewhat longer thumbs, X-rays of the upper limbs showed no additional abnormalities. The father of the patient also shows characteristics of HOS and carries the mutation. He has a large ASD II and persistent left superior vena cava ending in coronary sinus. The ASD was surgically closed at the age of 7 years and the left caval vein was reconstructed. Subsequently he developed tachycardias and an atrium flutter. At the age of 11, he presented with sinus node dysfunction for which an epicardial pacemaker was implanted. Upper limb abnormalities were limited to sloping shoulders. The mutation was absent in DNA samples from both parents of the father of the index patient.
3.5 p.Ile106Val (I106V)
The index presented with severe upper limb malformations. On the left side there was monodactyly and an absent ulna and radius. On the right side there was a short radius and ulna, with three digits. Because the nature of these defects (ulnar, radial, or central reduction defect) was inconclusive, TBX3 (Ulnar mammary syndrome27), and p63 (central reduction defects28) were included in genetic testing, however, no mutations were found. Mother and grandfather did not show limb defects or CHD, although they both carry the p.Ile106Val mutation. The same mutation (c.316A>G, p.Ile106Val) was described as pathogenic in literature, in a HOS patient with upper limb deformities without cardiac anomalies.19
3.6 Mutations affect the T-box of TBX5
All the identified TBX5 mutations reside within the highly conserved T-box, the DNA-binding domain that is present in all T-box proteins (Figure 1). Furthermore, the M74 and R237 residues are highly conserved between orthologues and paralogues (Figure 1). Interestingly, arginine residue 237 mutations have been described before in patients suffering from HOS (p.Arg237Trp and p.Arg237Gln), indicating this area as a mutational hotspot.12,29 The L94 and H220 residues are conserved between orthologues, but paralogues show substitutions at these positions. The I106 residue (mutated to a valine), is the least conserved residue of the mutations. In zebrafish Tbx5 a valine residue is located at this position, indicating that an isoleucine to valine change in this location is likely to be tolerated in the T-box structure and function.
Conservation and location of mutated residues in the T-box of Tbx5. (A) Three-dimensional structure of the T-box of TBX5 bound to DNA indicating the location of amino acid residues M74, L94, I106, H220, and R237, mutated in missense mutations identified in this study. Alpha helices shown in red, beta-sheets in orange. (B) Alignment of protein regions flanking the sites of mutation in TBX5. Human TBX5 is aligned with orthologous proteins in multiple species, and in the lower part, TBX5 is aligned with paralogous human proteins, from the Tbx-subfamily 1 (TBX1, TBX20) and Tbx-subfamily 2 (TBX2, TBX3, and TBX4).
The positions of the mutations are shown in a protein model of the T-Box bound to the DNA in Figure 1. From this figure it can be appreciated that amino acid residues M74, R237, and H220 are located in close proximity to the DNA, whereas L94 and I106 are located at the surface of the T-box, and are more exposed.
3.7 Functional defects in DNA-binding of TBX5 mutations
The transcription factor TBX5 activates its target genes by binding to the DNA via its DNA-binding domain, the T-box. The mutations discussed in this article are all located within the T-box. Consequently, we tested the DNA-binding capacity of the mutant proteins, using EMSA. A fragment of the Nppa promoter was used, containing a well characterized functional TBE.13,14 The T-box of TBX5 can strongly bind to the DNA, whereas the M74I, L94R, R237P, and H220del mutant proteins fail to bind DNA (Figure 2). In contrast, the I106V mutation does not interfere with DNA-binding.
EMSA showing binding of TBX5-T-box to an Nppa-promoter fragment. WT TBX5-T-box binds to DNA, as does the I106V mutant. Mutant Tbx5 proteins M74I, L94R, R237P, and H220del do not bind DNA. Lower panel shows comparable amounts of protein have been loaded (Coomassie).
3.8 Normal subcellular distribution of TBX5 mutations
To be able to regulate transcription and exert its function, TBX5 needs to be present in the nucleus. The localization of the mutant TBX5 proteins was assessed by transfecting flag-tagged mutant proteins into rat heart-derived cells (H1022). The localization of mutant and WT proteins was visualized with anti-flag immunostaining. Figure 3 shows that WT TBX5 protein localizes exclusively inside the nucleus and all mutant TBX5 proteins show a remarkably similar distribution pattern, indicating that the process of nuclear import is not affected by the mutations.
Nuclear localization of WT and mutant TBX5 transfected to H10 cells as shown by immunofluorescence. In green, all nuclei (sybr green); in red, TBX5 (anti-flag).
3.9 TBX5 mutations impair interaction with NKX2.5 in vitro
TBX5 cooperates and physically interacts with the homeodomain transcription factor NKX2-5, an interaction that is essential for proper heart development.13,14 This interaction depends upon the T-box of TBX5 and the homeodomain of NKX2-5. To test the effect of the mutant TBX5 proteins on the interaction with NKX2-5 we used an MBP pull down assay. Using this approach we show that the WT T-box of TBX5 binds to the homeodomain of NKX2-5 as expected (Figure 4A). However, the interaction is severely diminished in the case of the M74I, L94R, H220del, and R237P mutations, whereas the binding of the I106V mutant to NKX2-5 is not significantly changed and similar to WT.
Mutant proteins show reduced binding to NKX2-5 and GATA4. (A) In vitro protein interaction assay. (A) Coomassie stained protein gel showing equivalent amounts of WT and mutant TBX5 has been loaded. Western blot (lower panel) shows the amount of NKX2-5 retained by WT and mutants TBX5. (B–E) CoIP of TBX5 mutants with NKX2-5 and GATA4. (B, D) Western blot showing equivalent inputs of NKX2-5 or GATA4 (anti-myc, top), equivalent IP of flag-tagged TBX5 (wt & mutants, anti-flag, middle panel) and different levels of CoIP of NKX2-5 or GATA4 (anti-myc/anti-Gata4, bottom panel). G125R is a previously described gain of function mutation that has retained its capacity to interact with NKX2-5,4 R237W is known to have lost (part of) its interaction with NKX2-5 and GATA4. (C, E) Graph showing quantification of CoIP results from three separate experiments, error bars represent standard errors, *P < 0.01, #P < 0.1.
3.10 TBX5 mutations affect binding to NKX2-5 and GATA4 in mammalian cells
To further explore binding of TBX5 mutants to NKX2-5, we tested the interaction between full length TBX5 and NKX2-5 in the context of mammalian cells using co-immunoprecipitation (CoIP). As shown in Figure 4B, NKX2-5 is precipitated specifically when co-expressed with WT TBX5 (lane1-2). The M74I, L94R, H220del, and R237P mutant proteins show a drastic reduction of NKX2-5 binding capacity. NKX2-5 binding of the previously published R237W mutant protein was also affected, although some residual binding can be observed. However, this effect is not significantly different from the R237P mutant protein. The I106V mutant retains the capacity to bind NKX2-5, although results varied between experiments, which are reflected by relatively large error bars (Figure 4C). This variation was also observed for the G125R gain-of-function mutant,4 which, in line with the previous study, behaves similar to the WT protein with respect to NKX2-5 binding.
GATA4, another essential partner of TBX5 during heart development, can also bind to TBX5.17 Therefore, we subsequently tested the ability of TBX5 mutants to bind GATA4 using CoIP. GATA4 and WT TBX5 interact in this human cell-line (Figure 4D). Mutant proteins M74I, L94R, H220del, and R237P have largely lost their capacity to interact with GATA4. The R237W mutant shows only mild reduction of its GATA4 interacting capacity, to an extent that it is not significantly different from the WT protein in our assay. The I106V and G125R mutants both show a tendency to enhanced binding of GATA4, although this is not significant.
3.11 Functional defects in target gene activation
To test the functional consequences of the mutations in a cellular context, we used a reporter assay in which the proximal NPPA promoter (−270 to +1) was fused to a luciferase reporter. This promoter contains functional TBEs essential for NPPA regulation during development and adult life.13 TBX5 synergizes with NKX2-5 in the activation of this promoter, and therefore we tested the synergetic effect between NKX2-5 and the mutant TBX5 proteins. Indeed, WT TBX5 and NKX2-5 synergistically activate the reporter construct (Figure 5A). The M74I, L94R, H220del, and R237P mutants all show a decreased activation of the reporter in the absence as well as in the presence of NKX2-5. The I106V mutant activates this promoter construct in a manner similar to the WT protein, whereas the gain of function mutant G125R significantly enhanced activation. Interestingly, the mutation replacing arginine 237 by a proline has a more drastic effect on the activation of this promoter than when replaced by a tryptophan.
Activation of Nppa and FGF10 promoters. (A) Relative activation of NPPA-luciferase construct by TBX5 mutants compared with WT TBX5, without (open bars, #P < 0.05) and with (filled bars, *P < 0.05) NKX2-5. (B) Relative activation of FGF10-luciferase construct by M74I and I106V mutants compared with WT TBX5 (*P < 0.05). (C) Quantification of endogenous Nppa activation in mammalian cells using qRT–PCR. *P < 0.05. Error bars represent standard error (SEM). Each condition has been tested at least in three independent triplicate experiments.
The I106V protein functions similar to the WT protein in these assays. To further characterize the functionality of this mutant we performed qPCR for known target genes in H10 and H9C2 cells. Expression of Nppa was not affected by the I106V mutation, whereas the G125R gain-of-function mutant shows the expected increase in Nppa expression (Figure 5C). Since the patient carrying the I106V mutation only shows limb defects, we also tested the functionality of the mutant protein in the activation of FGF10, a TBX5 target gene essential for limb development.30 In contrast to the p.Met74Ile mutant protein, which acts as a loss of function mutant in this assay, the I106V mutant protein induces luciferase expression from the FGF10 promoter construct in a manner similar to WT TBX5 (Figure 5B).
HOS is a developmental disorder characterized by upper limb abnormalities and cardiac defects, which is linked to single gene mutations in TBX5. In this study, we present four novel mutations which are associated with HOS in patients with multiple cardiac malformations. In order to investigate the effect of the mutations on the proteins function during heart development, the mutant proteins were biochemically characterized using in vitro assays and tested for functionality in a mammalian cellular context.
Most of the TBX5 mutations found in HOS patients result in a premature stop, and encode a non-functional abrogated protein, which is likely to be degraded due to nonsense mediated decay and thus represent cases with functional haploinsufficiency. Missense mutation and small deletions may be more informative on the role of specific residues or active domains within the protein, or uncover novel functions of the protein. In this study, we present the novel missense mutations M74I, L94R, and R237P, a small deletion H220del and a previously published I106V missense mutation. All mutations reside inside the T-box and all patients carrying novel mutations suffer from severe cardiac defects. When overexpressed in H10 cells, WT and mutant TBX5 were localized exclusively inside the nucleus, suggesting that nuclear localization signals are still intact. Although it can not be excluded that endogenous TBX5 may be mislocalized under physiological circumstances, overexpression in a cell-line has been used successfully to show mislocalization of flag-tagged TBX5 mutants.16
The M74 residue is located in the second helix of the T-box, which is in close proximity to the DNA. In line with this, mutation of this residue, as seen in the M74I patient, leads to a complete loss of DNA-binding. Furthermore, the mutant protein can not bind to NKX2-5 and GATA4 and consequently, this protein acts like a functional null when tested in the Nppa promoter activation assay. Haploinsufficiency of TBX5 is therefore likely to underlie the defects seen in this patient.
In contrast to the M74 residue, the L94 residue is located relatively distant from the DNA, but mutation of this residue also leads to a complete loss of DNA-binding. Similarly, a mutation of this conserved leucine residue in TBX3 (p.Leu143Pro) has been shown to underlie Ulnar Mammary Syndrome.31 The Tbx3-L143P mutant does not bind to the consensus T-box binding site32 and encodes a non-functional protein when expressed in the mouse heart.33 The overall phenotype of this L94R patient is relatively mild, which may indicate that this protein retains some of its function. This was also noted in the luciferase assay, in which activation of the Nppa promoter is reduced, but not completely abolished.
The small deletion H220del mutant was found in a patient with severe cardiac malformations including complete AVSD with additional multiple pulmonary venous and arterial stenoses. The H220del protein behaves like a functional null in our assays, and the deletion likely disrupts the third alpha-helix of the T-box, which is essential for DNA-binding. To the best of our knowledge, a complete AVSD in a HOS patient has only been described once before,34 describing a frame-shift mutation near the H220 residue (p.V214fsX225). In parallel, pulmonary venous stenoses are not a typical feature of HOS, although have been described before in HOS patients.6,35,36 Whereas a mutation in TBX5 (c.1390delC) leading to an extended protein was associated with right lung agenesis previously,37 we report on the first case of HOS associated with pulmonary venous stenoses confirmed by a de novo mutation in the TBX5 gene. During mouse and chicken development, Tbx5 is expressed in the developing lungs.38,39 In mice lacking the transcription factor FoxF1, transcript levels of Tbx4 and Tbx5 are diminished, and pulmonary vasculature development is disturbed.40 In addition, knock-down of both Tbx4 and Tbx5 completely abrogates lung outgrowth.41 Therefore, absence of defects in the lungs in most Holt–Oram patients, or Tbx5 null mice, may well correlate with the redundant action of the related factor Tbx4 in the lungs.42 The fact that two a-typical features are observed in a HOS patient with a small deletion in TBX5 may suggest that the mutant protein has not only lost its cardiac functionality, but also interferes with pathways in other organs depending on the interaction with additional T-box proteins. Still, we cannot rule out that the genetic background contributes to these defects.
Adding to the list of mutations of amino acid residue 237 is the R237P mutation. The coding sequence in this area of TBX5 is highly GC rich, which may have contributed to the numerous independent mutations in this location, i.e. a mutation hotspot. These mutations are predicted to perturb binding of TBX5 to the minor groove of DNA,29 and replacement of this residue to a proline is certainly expected to be disruptive. Whereas previous studies report that missense mutations of R237 (R237Q and R237W) mainly induce mild cardiac defects,12,29 the R237P mutant protein is found in a patient with more complex cardiac malformations. In previous protein interaction assays utilizing in vitro translated TBX5 mutants, the R237W mutation was shown to abolish interaction with GATA4.17 In contrast, our CoIP assays in which TBX5 and GATA4 were co-expressed in mammalian cells, show that the R237W mutation does not significantly affect binding to GATA4, whereas the R237P mutation leads to a loss of interaction with GATA4 (Figure 4). In line with this, the R237P mutation causes a more severe loss of Nppa activation when compared with R237W (Figure 5).13
The expressivity of HOS is remarkably variable; even within families the same mutation gives rise to malformations of variable severity.5 However, although defects range from very mild to severe, carriers of a mutation are always affected. A special case is represented by the mutation causing replacement of isoleucine 106 by a valine. Although this genotype has previously been presented as a disease causing mutation,19 in our study, two out of three carriers shown no signs of HOS, raising the question whether the I106V mutation is causative for the observed defects. In contrast, the index patient presented with limb defects typical of HOS, but without signs of heart defects. This is in line with the findings of McDermott et al., whose I106V patient meets the strict criteria for diagnosis of HOS, although no cardiac anomalies were detected, however, the family history of that patient was not described. The absence of cardiac defects in all reported carriers is consistent with our biochemical and functional studies, as these assays show no signs of impaired binding to cardiac-specific interaction partners, or misregulation of the established cardiac target gene Nppa. The variability of I106V binding NKX2-5 and GATA4, which was also observed for G125R, may reflect structural consequences of the mutation, expression of which depends on presence of other proteins or small differences in salt or detergent concentrations between experiments.
Since affected carriers of the I106V mutation suffer only from upper limb malformations, this may suggest that the interaction with limb-specific partner proteins is affected. The expressivity of the limb defects in our study does not seem to be fully penetrant, perhaps depending on the genetic background. However, based on its position in the T-box, I106 is available for interaction with other proteins and, importantly, the neighbouring V107 residue is essential for interaction with histone modifying complexes.43 Interestingly, in zebrafish two forms of tbx5 exists which differ at amino acid position 106.44 The tbx5a form, with an isoleucine at position 106 is expressed in both the heart and the limbs, in contrast, the tbx5b form, with a valine at position 106, is only expressed in the heart, but not in the limbs. This could suggest that this site, through an unknown mechanism, is somehow involved in limb expression. Although limb defects in this patient may have occurred independent of TBX5 perturbations, the I106V mutant may thus represent a unique mutation that causes solely a loss of limb-specific function, possibly only if co-segregating with additional (unknown) polymorphisms or mutations.
In summary, we have shown that four novel missense mutations in TBX5 investigated in this study are functional mutations that cause a spectrum of biochemical and cellular defects. Our results show loss of DNA-binding and loss of interaction with known protein partners leading to reduced transcription activation of known TBX5 target genes. Our data are consistent with the hypothesis that the missense mutations in TBX5 result in a functional haploinsufficiency. In addition, we have shown that the I106V mutant, which has been associated with HOS before, is also present in seemingly unaffected relatives. The I106V protein retains its DNA-binding capacity and interacts with known cardiac protein partners. Future research will have to show the exact nature of this variation.
Conflict of interest: none declared.
This work was supported by the European Community's Sixth Framework Programme contract (‘HeartRepair’ LSHM-CT-2005-018630) to C.J.J.B., A.F.M., and P.B.
↵† These two authors contributed equally to this work.
↵‡ Present address: Department of Medical Genetics, University Medical Center Utrecht, Utrecht, The Netherlands.
. Mutations in cardiac T-box factor gene TBX20 are associated with diverse cardiac pathologies, including defects of septation and valvulogenesis and cardiomyopathy. Am J Hum Genet 2007;81:280-291. doi:10.1086/519530.
. Expanding the spectrum of TBX5 mutations in Holt–Oram syndrome: detection of two intragenic deletions by quantitative real time PCR, and report of eight novel point mutations. Hum Mutat 2006;27:975-976. doi:10.1002/humu.9449.
. Coordinated but physically separable interaction with H3K27-demethylase and H3K4-methyltransferase activities are required for T-box protein-mediated activation of developmental gene expression. Genes Dev 2008;22:2980-2993. doi:10.1101/gad.1689708.