Saadi et al.

1

Identification of a Dominant Negative Homeodomain Mutation

in Rieger Syndrome

Irfan Saadi‡, Elena V. Semina§, Brad A. Amendt¶,||, David J. Harris**, Kenneth P. Murphy‡‡, Jeffrey C.

Murray‡,§,§§, and Andrew F. Russo‡,¶,¶¶

‡Genetics Program, Departments of §Pediatrics, ‡‡Biochemistry, §§Biological Sciences, and ¶Physiology

and Biophysics, University of Iowa; Iowa City, IA 52242. **Children’s Mercy Hospital, Kansas City, MO

64108.

|| Present address: Department of Biological Sciences, The University of Tulsa, Tulsa, OK 74104

¶¶ Correspondence should be addressed to A.F.R.

e-mail: andrew-russo@uiowa.edu

Tel: (319) 335-7872

Fax: (319) 335-7330

Copyright 2001 by The American Society for Biochemistry and Molecular Biology, Inc.

JBC Papers in Press. Published on April 11, 2001 as Manuscript M008592200 by guest on M

arch 27, 2020http://w

ww

.jbc.org/D

ownloaded from

Saadi et al.

2

Running title

“Dominant negative Rieger syndrome mutation”

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

3

Abstract

Mutations in the PITX2 bicoid-like homeobox gene cause Rieger syndrome. Rieger syndrome

is an autosomal-dominant human disorder characterized by glaucoma, as well as dental

hypoplasia, mild craniofacial dysmorphism and umbilical stump abnormalities. PITX2 has also

been implicated in the development of multiple organs and left-right asymmetry in the body plan.

The PITX2 homeodomain has a lysine at position 50, which has been shown to impart the bicoid-

type (TAATCC) DNA-binding specificity to other homeodomain proteins. A mutation (K88E),

found in a Rieger syndrome patient, changes this lysine to glutamic acid. We were intrigued by

the relatively pronounced phenotypic consequences of this K88E mutation. In the initial analyses,

the mutant protein appeared to simply be inactive, with essentially no DNA binding and

transactivation activities and, unlike the wildtype protein, with an inability to synergise with

another transcription factor, Pit-1. However, when the K88E DNA was cotransfected with wildtype

PITX2, analogous to the patient genotype, the K88E mutant suppressed the synergism of wildtype

PITX2 with Pit-1. In contrast, a different PITX2 homeodomain mutant, T68P, which is also

defective in DNA binding, transactivation, and Pit-1 synergism activities, did not suppress the

wildtype synergism with Pit-1. These results describe the first dominant negative missense

mutation in a homeodomain and support a model that may partially explain the phenotypic

variation within Rieger syndrome.

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

4

Introduction

The PITX2 gene was cloned based on its linkage in Rieger syndrome (1). Rieger syndrome is an

autosomal-dominant human disorder characterized by ocular anterior chamber anomalies causing

glaucoma in more than 50% of affected individuals, as well as dental hypoplasia, mild craniofacial

dysmorphism and umbilical stump abnormalities (1,2). Other features associated with this syndrome

include abnormal cardiac, limb and pituitary development. PITX2 has also been shown to be a

downstream target in left-right asymmetry pathways and organogenesis (3,4). It is expressed

asymmetrically to the left side during early development in the lateral plate mesoderm (LPM) and later in

most of the organs derived from LPM, including the heart and the gut (4,5).

The PITX2 protein is a 33 kD protein that has a homeodomain similar in sequence to other bicoid

binding homeodomain proteins, i.e. a characteristic lysine at residue 50 of the homeodomain (1). Bicoid-

type homeodomains have a TAATCC consensus binding site and PITX2 has been shown to specifically

bind to this consensus sequence (6,7,8). In addition, PITX2 transactivates a reporter gene with multimers

of the bicoid binding site (6). Amendt et al. (6) have shown that PITX2 interacts with Pit-1, an important

transcription factor regulating expression of thyroid-stimulating hormone, growth hormone, and prolactin,

and the terminal differentiation of the cell types in which they are expressed (9,10). Transient transfection

assays with the prolactin promoter and both PITX2 and Pit-1 showed a strong synergistic effect on

transactivation (6).

It has been demonstrated that residue 50 of a given homeodomain interacts with base pairs 5 and 6

of the hexanucleotide consensus (11,12,13,14). Interestingly, a mutation found in a Rieger syndrome

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

5

patient changes the lysine (K) of position 50 in PITX2 homeodomain to a glutamic acid (E), a mutation

that has marked phenotypic consequences. The mutation is referred to as K88E (position 88 is the

location of K50 in PITX2 isoform “a”). In this study we have examined the functional properties of the

K88E mutant PITX2 protein and compared them to a previously described Rieger syndrome mutation that

changes the threonine at position 30 of the homeodomain to a proline (T68P) (6). The T68P mutant has

been shown to be defective in DNA binding and transactivation properties (6,15). T68P mutant also failed

to show synergistic transactivation of the prolactin promoter when cotransfected with Pit-1 (6). Similarly,

the K88E mutant protein had reduced DNA binding and transactivation activities. K88E was also unable

to synergise with the transcription factor, Pit-1. Unexpectedly, in contrast to T68P, K88E suppressed the

Pit-1 synergism of wildtype PITX2. These results support a model in which a dominant negative mutant

protein can contribute to phenotypic variation within the haploinsufficiency Rieger syndrome.

Materials and Methods

Clinical Data

The proband was the six-pound, six-ounce infant born to a G1 P0 30-year old mother following a

pregnancy complicated by asthma and gestational diabetes. At birth, the proband was noted to have

midline abdominal defects, hyperflexive joints and an abnormal ocular exam. The initial eye exam was

interpreted as possible bilateral Peter’s anomaly with bilateral iris hypoplasia and posterior embryotoxin, a

normal-appearing lens and macula and several small corneal opacities with radial striae. Abdominal

exam showed diastasis recti, a protuberant umbilicus and a small amount of bowel present in the

underlying protuberant area. There is a bifid uvula and submucous cleft palate. Renal ultrasound

showed no evidence of abnormalities. Subsequent multiple exams under anesthesia show a spectrum

consistent with anterior segment dysgenesis bilaterally that was more severe than those seen in many

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

6

Rieger patients. The abdominal abnormalities gradually resolved with time, and the primary teeth were

noted to be small and dysplastic. Physical parameters to age five years include normal growth

parameters for height, weight, and head circumference and normal cognitive, gross and fine motor

milestones. The parents are not affected with any ocular, umbilical, or dental anomalies, and one

subsequent sibling is also unaffected. The combination of ocular, umbilical, and dental anomalies are

consistent with the spectrum of abnormalities seen in Rieger syndrome, and there are no clinical

abnormalities yet that suggest growth hormone deficiency, hydrocephalus, or congenital heart disease.

Electrophoretic mobility shift assays

The human PITX2 cDNA (6), PITX2 T68P cDNA (6) and PITX2 K88E cDNA were cloned into

pGex6p-2 (Amersham Pharmacia Biotech). The wildtype and mutant cDNAs correspond to PITX2

isoform “a”. The K88E point mutation was created by PCR-based site-directed mutagenesis (16) using

Pfu DNA polymerase (Stratagene) and confirmed by sequencing. Protein preparations were done similar

to as described (6). The GST moiety was removed by PreScission protease (Amersham Pharmacia

Biotech). Proteins were quantitated by Bradford assays (BioRad) and visualized on SDS-polyacrylamide

gels.

The wildtype PITX2 (6), T68P (6), and K88E cDNAs were also cloned into pcDNA3.1 MycHisC

expression vector (Invitrogen) containing the T7 promoter and an in-frame C-terminal c-Myc epitope.

These vectors were used in in vitro transcription and translation (TNT) reactions. One microgram of DNA

was used in TNT reactions that were carried out according to the protocol supplied with the TNT kit

(Invitrogen).

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

7

The 32P-labeled probe for EMSAs was generated by annealing complementary oligonucleotides

containing the sequence gatccGCACGGCCCATCTAATCCCGTGggatc and end-filled using Klenow

polymerase with 32P-dATP as described (17). Standard binding assays were carried out by incubating the

oligonucleotide probe (1.0 pmol) in a 20 µL reaction containing binding buffer (20 mM Hepes, pH 7.5, 5%

glycerol, 50 mM NaCl, 1 mM dithiothreitol), 0.1 µg of poly(dI-dC), and 100-300 ng protein on ice for 15

minutes. The samples were electrophoresed for 3.5 hours at 280 V on 8% polyacrylamide gels as

described (6).

Computer modeling

The structure of Antp homeodomain-DNA complex was obtained from the Protein Database1

[accession no. 1AHD] (14). The computer program SYBYL (Tripos, Inc.) was used to model the wildtype

and K88E PITX2-DNA complexes by substituting all the necessary amino acid residues in the

homeodomains and nucleotide base pairs. The structures were minimized (dielectric constant = 80;

∆energy for termination = 0.05 kcal/mol), and the complex was reannealed around the residue change

(dielectric constant = 80; ∆energy for termination = 0.05 kcal/mol; reannealing ‘hot’ radius = 6 Å,

‘interesting’ radius = 12 Å) and the entire complex minimized (dielectric constant = 80; ∆energy for

termination = 0.005 kcal/mol). Total energy, E, of the entire complex was calculated using the Tripos

force field, represented by the equation: E = ΣEstr + ΣEbend + ΣEoop + ΣEtors + ΣEvdw + [ΣEele + ΣEdist_c +

ΣEang_c + ΣEtors_c + ΣErange_c + ΣEmulti + ΣE field fit]. Where str = bond stretched or compressed from its

equilibrium bond length; bend = bending bond angles from their equilibrium values; oop = bending planar

atoms out of the plane; tors = torsion or twisting about bonds; vdw = van der Waals non-bonded

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

8

interactions; ele = electrostatic interactions; dist_c = distance constraints; tors_c = torsion angle

constraints; range_c = range constraints; multi = multifit.

Cell culture and reporter gene assays

The wildtype PITX2 (6), T68P (6), and K88E cDNAs cloned into pcDNA3.1 MycHisC expression

vector (Invitrogen) contain the cytomegalovirus (CMV) promoter and an in-frame C-terminal c-Myc

epitope. The TK-Bic reporter contains the herpes simplex virus thymidine kinase minimal promoter with

four bicoid elements (6). The prolactin-luciferase reporter contains 2500 base pairs of the rat prolactin

enhancer/promoter (19). A CMV β-galactosidase (6) or an SV40 β-galactosidase (Promega) reporter

plasmid was used as a control for transfection efficiency. The rat Pit-1 expression plasmid is under the

control of the Rous sarcoma virus promoter (6). The rat Pit-1 ∆8-128 (20), Pit-1 ∆POU (21), and Pit-1

W261C (22) expression plasmids were provided by S.J. Rhodes. The MEKK (MEK Kinase amino acids

380 to 672) expression vector was obtained from Stratagene.

COS7 and CHO cells for Figures 4, 5, 6 and 7 were cultured in 60 mm dishes and electroporated as

described (23). Transfection assays used 5-10 µg of the Myc-tagged PITX2 plasmids along with either

the TK-Bic or prolactin reporter in COS7 (260 mV; 960 µF) or CHO (340 mV; 960 µF) cells. 1 µg of CMV-

βgal (COS7) or pSV-βgal (CHO) was also added to each sample. For Figures 8, 9, 10 and 11, 2x105

COS7, CHO, or GH3 cells were cultured in 12-well plates and transfected by lipofection in the absence of

antibiotics. FuGene 6 (Roche) was used for COS7 (9 µL) and CHO cells (3 µL), whereas, Lipofectamine

2000 (Gibco BRL) was used for GH3 cells (4 µL). Assays used 0.5 µg of each construct for CHO and

1 The atomic coordinates for the NMR structure of this protein are available in Research Collaboratory for

Structural Bioinformatics Protein Databank (http://www.rcsb.org/pdb/) under PDB # 1AHD (18).

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

9

GH3 cells, while 1.5 µg of each construct were used for COS7 cells. The DNA was incubated with the

corresponding transfection reagent in the absence of serum according to the manufacturer’s protocol.

The DNA-transfection reagent mixture was added to cells growing in antibiotic-free media in the presence

of serum and incubated for 16-20 hrs. Following incubation, the cells were scraped, lysed and assayed

using luciferase assay reagents (Promega) and β-galactosidase assay reagents (Tropix Inc.). Luciferase

activities were corrected for both transfection efficiency and protein levels (Bradford assays) (BioRad).

For western blots, equal amounts of cell extracts from transfection assays were electrophoresed on a

12.5% SDS-polyacrylamide gel, transferred to polyvinylidine difluoride filters (Millipore), and

immunoblotted using a c-Myc antibody (Santa Cruz Labs) and ECL reagents (Amersham Pharmacia

Biotech). Statistical significance was calculated using a simple t-test for means of two samples.

Results

Identification of the K88E mutation

The pedigree of the affected proband and family with a de novo mutation is shown in Figure 1a. The

proband presented the characteristic Rieger syndrome features including small and dysplastic primary

teeth (Fig. 1b), protuberant umbilicus, and ocular anomalies. The initial eye exam was interpreted as

possible bilateral Peter’s anomaly, however, subsequent exams revealed a spectrum consistent with

bilateral anterior segment dysgenesis more severe than those seen in many patients with Rieger

syndrome. Single strand conformation polymorphism analysis of PITX2 exons in proband genomic DNA

identified a heterozygous A→G change in codon 88 (Fig. 1c) resulting in a lysine (AAG) to glutamic acid

(GAG) substitution.

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

10

K88E is defective in DNA binding

We evaluated the DNA binding activity of the mutant PITX2 protein. Wildtype PITX2 has been shown

to bind TAATCC bicoid consensus sequence in an electrophoretic mobility shift assay (EMSA) (Fig. 2a)

(6). In contrast, the K88E mutant showed little or no binding to bicoid probe (Fig. 2a). Similar amounts of

bacterial-synthesized wildtype PITX2 and K88E mutant proteins were used (Fig. 2b). EMSA was also

done with similar amounts of wildtype and mutant PITX2 proteins obtained from in vitro transcription and

translation (TNT) reactions (Fig. 2c-d). As seen with the bacterial-generated proteins, the K88E mutant

protein had little or no binding to the bicoid probe. For comparison, the T68P mutant protein also showed

greatly reduced binding under these conditions, in agreement with previous reports (6, 15).

In order to get a better understanding of how the change from a lysine to a glutamic acid can result in

significantly reduced DNA binding, we used a computer-modeling program to predict the PITX2

homeodomain-DNA complex structure (Fig. 3). There have been only three homeodomain-DNA complex

structures determined through crystallization or NMR spectroscopy2. Of the three, we decided to model

the PITX2 homeodomain after the Antennapedia (Antp) homeodomain-DNA complex primarily because it

is one of the best characterized homeodomain proteins, the homeodomain is 43% identical to PITX2, and

its DNA binding consensus sequence (TAATGG) closely matches that of PITX2 (TAATCC) (7,14). The

Antp homeodomain contains a glutamine at position 50 (Q50), which was shown to positively interact with

positions 5 and 6 of its hexanucleotide consensus sequence (GG) (14). We used the computer program

SYBYL to substitute all the mismatched residues in the Antp homeodomain-DNA complex to those of the

2 The atomic coordinates for the NMR structure of these proteins are available in Research Collaboratory

for Structural Bioinformatics Protein Databank (http://www.rcsb.org/pdb/) under PDB # 1AHD, 1APL,

1HDD (18).

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

11

PITX2 homeodomain-DNA complex, including the substitution of Q50 to K50 and TAATGG to TAATCC,

with recalculated local annealing and energy minimization. In the case of wildtype PITX2, K50 does not

clash with positions 5 and 6 (Fig. 3). However, the glutamic acid at position 50 causes an observable

change in the spatial orientation of GG bases on the antisense strand of the TAATCC consensus

sequence (Fig. 3). The overall calculated energy for the K88E mutant homeodomain-DNA complex is 38

kcal/mole higher than wildtype PITX2 complex. The greatest contributions to the difference were from

angle bending energy (∆Ebend = 24.5 kcal/mol) and torsional energy (∆Etor = 21.4 kcal/mol).

K88E is defective in transactivation

Given the decreased DNA binding activity, we predicted that the K88E mutant protein would also

have reduced transactivation activity. The transactivation activity of K88E protein was measured by

transient transfection assays in COS7 and CHO cells, which do not express the endogenous PITX2 gene.

Wildtype PITX2 transactivated a luciferase reporter gene containing the thymidine kinase promoter with

four TAATCC binding sites (TK-Bic) approximately 5-fold (Fig. 4a). Wildtype PITX2 specifically activates

this reporter and does not transactivate the parental TK-luciferase or the CMV-βgal reporter genes (6).

The K88E mutant did not appreciably transactivate the TK-Bic reporter (~1.3-fold) (Fig. 4a). This is

consistent with the reduced binding to the bicoid element in vitro. We then asked whether this defective

transactivation was also manifested with a natural target, the prolactin promoter (Pro-Luc). Wildtype

PITX2 activated the prolactin promoter approximately 8-fold in COS7 (6) and 13-fold in CHO (Fig. 4b)

cells. In contrast, the K88E only weakly activated the prolactin promoter (~2-fold) in CHO cells (Fig. 4b).

There were approximately equivalent amounts of wildtype and K88E mutant PITX2 proteins expressed in

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

12

the transfected cells (Fig. 4c). These results demonstrate that the K88E mutant protein has defective

transactivation activity that is similar to its reduced binding activity.

K88E does not synergise with Pit-1

We also looked at the transactivation activity of the K88E mutant in the presence and absence of the

POU homeodomain transcription factor, Pit-1. It has previously been shown that Pit-1 and PITX2 can

synergistically activate the prolactin promoter (6). The significance of the PITX2 interaction with Pit-1 is

supported by their coexpression in early pituitary development and the phenotype of Pitx2 knockout mice

(24,25). Coexpression of Pit-1 with wildtype PITX2 yielded about 350-fold transactivation of the prolactin

reporter plasmid in CHO cells (Fig. 5a, lane 4). In contrast, the K88E mutant did not synergize with Pit-1

(Fig. 5a, lane 7). This is similar to the T68P mutant, which also failed to synergise with Pit-1 (Fig. 5a, lane

10) (6).

K88E suppresses wildtype PITX2 synergism with Pit-1

In order to explain the marked phenotype of the patient carrying the K88E mutation, we tested the

possibility that the mutant protein might interfere with the wildtype protein activity. Indeed, the presence

of one wildtype and one mutant allele in Rieger syndrome can be viewed as an autosomal dominant as

well as haploinsufficiency disorder. To simulate this situation, we co-transfected equal amounts of

wildtype PITX2 and K88E mutant DNA with either the TK-Bic (Fig. 4a) or prolactin (Fig. 4b) reporter

plasmids and found no significant effect on wildtype activity. We then repeated the same experiments in

the presence of Pit-1. Surprisingly, the K88E mutant protein suppressed the synergistic effect of wildtype

PITX2 with Pit-1 significantly (Fig. 5, lane 8). In contrast, the T68P mutant did not suppress the

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

13

synergism of wildtype PITX2 with Pit-1 (Fig. 5, lane 11). As a control for the increased amount of

transfected DNA, transfection with additional wildtype PITX2 DNA yielded similar activity as seen with 5

µg wildtype PITX2 (Fig. 5, lane 5). As additional controls, K88E did not suppress the co-transfected CMV

β-galactosidase or an SV40 β-galactosidase reporters. Titration of the K88E mutant DNA in the co-

transfection assay indicated that approximately equimolar amounts of wildtype and K88E DNAs are

required for significant inhibition of wildtype PITX2 activity (Fig. 6). Again, titration of the T68P mutant

DNA did not result in any inhibition of wildtype synergism (Fig. 6). In order to confirm the dose-dependent

nature of the dominant-negative effect, we transfected additional wildtype PITX2 DNA to the equimolar

amounts of wildtype and K88E (Fig. 7). The additional amount of wildtype DNA was able to significantly

rescue the dominant-negative effect (Fig. 7, lane 3; p < 0.015) to the same level (70 %) as predicted by

titration curve in Fig. 6 (lane 3).

We wanted to know if K88E could suppress Pit-1 independent activation of the prolactin promoter.

We used a constitutively active MEK kinase (MEKK), which stimulates the MAP (mitogen activated

protein) kinase signal transduction pathway, to activate the prolactin promoter in CHO cells (Fig. 8, lane

2). The K88E mutant was not able to suppress the MEKK stimulated activation of the prolactin promoter

(Fig. 8, lane 3). Thus, the dominant negative effect of K88E requires Pit-1 DNA binding or its interaction

with PITX2 or both.

In order to determine the dependence of this dominant-negative effect on Pit-1 DNA binding,

cotransfections of wildtype and mutant PITX2 were done in the presence of Pit-1 with TK-Bic reporter

construct which lacks the Pit-1 binding site in COS7 cells (Fig. 9). Wildtype PITX2 activated the TK-Bic

reporter approximately 5-fold (Fig. 9, lane 2) as shown in Fig. 4, while both K88E (Fig. 9, lane 5) and

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

14

T68P (Fig. 9, lane 7) showed little or no transactivation. As expected, the presence of Pit-1 did not affect

the transactivation of wildtype PITX2 (Fig. 9, lane 4), K88E (Fig. 9, lane 6), or T68P (Fig. 9, lane 8).

Cotransfection of the wildtype and the K88E mutant in the absence of Pit-1 (Fig. 9, lane 9) or in the

presence of Pit-1 (Fig. 9, lane 10) did not show any dominant negative effect. Likewise, the T68P mutant

was not dominant negative as expected (Fig. 9, lanes 11, 12). This implies that the Pit-1 binding site is

required for the dominant negative effect.

In order to map the protein domains required for the K88E suppression of wildtype activity, we tested

several Pit-1 and PITX2 mutants (Fig. 10). PITX2 ∆C39 lacks the C-terminal 39 amino acids that have

been shown to interact with Pit-1 (6). Pit-1 ∆8-128 lacks the N-terminal 120 amino acid residues that

constitute the major transactivation domain (20). These N-terminal residues have been shown to interact

with the closely related PITX1 (10). Pit-1 ∆POU lacks the POU-specific domain of Pit-1 required for DNA

binding (21). Finally, Pit-1 W261C contains the mutation identified in the Snell dwarf mice that disables

the POU DNA domain (22). Figure 10 shows the cotransfection of wildtype PITX2 and K88E mutant in

the presence of Pit-1 mutants with the prolactin reporter in CHO cells. Wildtype PITX2 (Fig. 10, lane 2)

and wildtype Pit-1 (Fig. 10, lane 3) show approximately 15-fold activation of prolactin promoter.

Cotransfection of both PITX2 and Pit-1 showed a synergistic response of approximately 60-fold (Fig. 10,

lane 4). The lower fold activation seen here compared to that seen in Fig. 5 (lane 4) could be due to

differences in methods of transfection, electroporation for Fig. 5 versus lipofection for Fig. 10. The

coexpression of K88E mutant with wildtype PITX2 and wildtype Pit-1 (Fig. 10, lane 5) showed significantly

reduced activation (p < 0.009) as expected. The PITX2 ∆C39 mutant that lacks the C-terminal 39 amino

acids shown to interact with Pit-1, has been reported to synergize poorly with Pit-1 in COS cells (20).

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

15

Surprisingly, in CHO cells, PITX2 ∆C39 showed a strong synergism with Pit-1 (Fig. 10, lane7) which was

suppressed by cotransfection of K88E (Fig. 10, lane 8, p < 0.001). The Pit-1 ∆8-128 mutant (Fig. 10, lane

9) that lacks the major transactivation domain showed significantly reduced activity than wildtype Pit-1

(Fig. 10, lane 3, p < 0.0004). Interestingly, this mutant which lacks the reported PITX1 interaction domain

was also able to synergize with wildtype PITX2 (Fig. 10, lane 10). This synergism was again suppressed

by K88E mutant coexpression (Fig. 10, lane 11). The PITX2 C∆39 and Pit-1 ∆8-128 results imply that, in

CHO cells, physical interaction of Pit-1 with PITX2 is not required for synergism. Both Pit-1 ∆POU (Fig.

10, lane 12) and Pit-1 W261C (Fig. 10, lane 15) mutants that are defective in Pit-1 DNA binding showed

little or no transactivation. Cotransfection of Pit-1 ∆POU (Fig. 10, lane 13) and Pit-1 W261C (Fig. 10, lane

16) mutants with wildtype PITX2 did not show synergistic activation underscoring the need for Pit-1 DNA

binding for synergism. Coexpression of K88E resulted in some suppression of the activity, although to a

lesser degree (about 3-fold) than seen with wildtype Pit-1 (6-fold) (Fig. 10, lanes 14 and 17 respectively).

These results are consistent with the trend seen in Fig. 4b where coexpression of wildtype PITX2 and

K88E mutant resulted in lower, albeit statistically non-significant (p < 0.14) transactivation. Results from

Figures 9 and 10 thus show that both Pit-1 binding site and Pit-1 DNA binding activity are required for a

pronounced dominant negative effect. Furthermore, the K88E suppression of PITX2-Pit-1 synergism

does not require the respective C-terminal or N-terminal domains of these proteins.

K88E suppresses prolactin promoter activity in GH3 cells

We wanted to see if K88E mutant would suppress the activity of endogenous Pitx2 and Pit-1. For this

experiment, we used the GH3 pituitary cell line. GH3 cells are known to express the endogenous Pitx2

and Pit-1 genes (10). Under our conditions, overexpression of wildtype PITX2 (Fig. 11a, lane 2), or Pit-1

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

16

(Fig. 11a, lane 3) did not significantly change prolactin promoter activity in GH3 cells, suggesting that the

endogenous Pitx2 and Pit-1 proteins were not rate-limiting. Interestingly, overexpression of the K88E

mutant (Fig. 11a, lane 4) was able to significantly suppress the prolactin promoter activity in GH3 cells (p

< 0.005). In contrast, overexpression of the T68P mutant did not affect promoter activity (Fig. 11a, lane

5). As a control, we used the TK-Bic reporter (Fig. 11b) which lacks Pit-1 binding site. Neither the K88E

or T68P mutants were able to suppress the activity of TK-Bic reporter in GH3 cells (Fig. 11b). This

control underscores the need for the Pit-1 binding site for the dominant negative effect of K88E mutant.

Hence the K88E mutant can repress wildtype PITX2 activity in both heterologous cells and in cells that

normally express the Pitx2 gene. Furthermore, in both situations, Pit-1 interaction with DNA is required

for repression by the K88E mutant.

Discussion

In this study we have focused on a point mutation, K88E, that changes a single amino acid in the

recognition helix of the homeodomain from a lysine to a glutamic acid. We show that the K88E mutant is

defective in its capacity to bind DNA and transactivate synthetic and natural promoters in transfection

assays. Our computer modeling suggests that the defect in DNA binding of the K88E mutant

homeodomain is due to the steric constraints which require significant unfavorable bending and torsional

angles in the DNA. This energetic penalty results in a very large decrease in the binding constant.

Although the exact value of the energy difference (∆E = 38 kcal/mol) must be considered more qualitative

than quantitative, a ∆E of this size would result in a decrease in the binding constant by about 30 orders

of magnitude. Interestingly, the only documented homeodomain proteins with a glutamic acid at position

50 are Zmhox1a and Zmhox1b maize proteins that appear to bind very divergent sequences (26). The

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

17

T68P mutant protein is also defective in DNA binding. Amendt et al. (6) used bacterially expressed T68P

protein in EMSA to show that it could still bind the bicoid probe, but with reduced binding capacity.

Kozlowski et al. (15) used total extracts from cells transfected with T68P DNA in their EMSA to show that

the mutant protein cannot bind DNA. We have used T68P protein obtained from in vitro TNT reactions in

EMSA to show greatly diminished binding to the bicoid DNA probe. The quantitative differences

observed are likely due to different types of protein preparations and binding conditions.

Most surprisingly, the K88E protein acts in a dominant negative manner to repress wildtype PITX2

activity in the presence of Pit-1. In contrast, another PITX2 missense mutation, T68P, that abolishes

transactivation properties (6), does not repress wildtype PITX2 synergism with Pit-1. As the K88E mutant

protein does not appear to have any intrinsic repressor activity, the dominant negative effect is likely due

to the formation of nonfunctional complexes between wildtype PITX2 and the K88E mutant proteins.

Previous studies have indicated that PITX2 can form homodimers (6,27). One mechanism for the

dominant negative inhibition would be that the K88E protein could heterodimerize with wildtype PITX2

and that the wildtype PITX2-K88E heterodimer would be unable to synergise with Pit-1. Simply

interpreted, our results imply that wildtype PITX2-K88E mutant dimers are nonfunctional. At face value,

this would explain the dominant-negative effect of K88E. However, the repression cannot be that simple

since the suppression of wildtype PITX2 in CHO cells was significant only in the presence of Pit-1 DNA

binding activity. Hence, it seems likely that the wildtype PITX2-K88E heterodimer is still able to bind

DNA. Furthermore, the T68P mutant has decreased DNA binding, defective transactivation and

synergism (6,15), yet it is not a dominant-negative mutant protein.

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

18

We propose that the dominant negative nature of the PITX2-K88E wildtype-mutant heterodimer is

dictated by its inability to synergise with other transcription factors, such as Pit-1. In support of this

hypothesis, K88E was unable to suppress Pit-1 independent activation of the prolactin promoter by a

constitutively active MEK kinase. Interestingly, the suppression by K88E requires a functional, but not

physical, interaction between PITX2 and Pit-1. The C-terminal domain of PITX2 that binds Pit-1 is not

required for repression by K88E. In this regard, we found that the requirement for the C-terminal 39

residues for synergism is cell-specific. It appears that these residues are not required for synergism in at

least 3 cell lines, CHO, LS8, and N2A, but are required in COS7 and HeLa cells. The basis for this

specificity is not known, but is suggestive that other factors are involved in PITX2 transactivation, as we

had previously speculated based on data from the PITX2 ∆C39 mutant (6). We have now shown that

both PITX2 ∆C39, which lacks the Pit-1 interacting residues (6), and Pit1 ∆8-128, which lacks the

reported PITX1 interacting residues (10), are capable of synergism in CHO cells and that this synergism

can be suppressed by K88E.

In a further test of the role that Pit-1 plays in K88E suppression, we have shown that Pit-1 DNA

binding activity is required. This was shown by complementary experiments using a reporter lacking a

Pit-1 binding site (TK-bicoid) and by using mutant Pit-1 proteins that lack DNA binding activity (Pit-1

∆POU and Pit-1 W261C). Taken together, these data imply that in CHO cells, Pit-1 and PITX2 DNA

binding activities may be sufficient for the dominant negative effect of K88E. Finally, we showed that

K88E can suppress the prolactin promoter in the pituitary adenoma GH3 cell line which expresses the

endogenous Pit-1 and Pitx2 genes. The ability of K88E to suppress the prolactin promoter in GH3 cells

supports the potential physiological relevance of K88E activity in the pituitary.

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

19

Our results describe the first molecular characterization of a dominant negative missense mutation in

a homeodomain. The other examples of dominant negative mutations in homeodomain proteins have

been truncations that remove the transactivation domain yet retain the DNA binding homeodomain

(28,29,30). Hence, the K88E mutation represents a new type of homeodomain mutation in human

disorders. We propose that the phenotypic variation of Rieger syndrome can be determined by a

continuum of mutations in PITX2 ranging from null deletions (1) to dominant negative proteins. This

proposal is in agreement with recent studies by Kozlowski et al. (15) who showed that two mutations with

mild defects in PITX2 activities were correlated with milder anterior segment aberrations (iris hypoplasia

and iridogoniodysgenesis syndrome) than seen in most Rieger patients. This perspective suggests an

approach to understanding genotype-phenotype correlations for other transcription factor disorders, such

as Saethre-Chotzen syndrome (TWIST) (31,32), Waardenburg syndrome (PAX3) (33,34), Aniridia, and

Peter’s Anomaly (PAX6) (35,36).

Acknowledgements

We thank the family for their cooperation and willingness to participate in the study and for providing

photos and dental X-rays; L.B. Sutherland, S.D. Hirsch for reagents and assistance; and S.J. Rhodes

(Indiana University-Purdue University Indianapolis, IN) for kindly providing the Pit-1 mutants. This work

was supported by National Institutes of Health grants DE13076 and EY12384.

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

20

References

1. Semina, E.V., Reiter, R., Leysens, N. J., Alward, L. M., Small, K. W., Datson, N. A., Siegel-

Bartelt, J., Bierke-Nelson, D., Bitoun, P., Zabel, B. U., Carey, J. C., and Murray, J. C. (1996)

Nature Genet. 14, 392-399.

2. Rieger, H. (1935) Z. Augenheilk. 86, 333.

3. Yoshioka, H., Meno, C., Koshiba, K., Sugihara, M., Itoh, H., Ishimaru, Y., Inoue, T., Ohuchi, H.,

Semina, E.V., Murray, J. C., Hamada, H., and Noji, S. (1998) Cell 94, 299-305.

4. Lu, M., Pressman, C., Dyer, R., Johnson, R.L., and Martin, J.F. (1999) Nature 401, 276-278.

5. Ryan, A.K., Blumberg, B., Rodriguez-Esteban, C., Yonei-Tamura, S., Tamura, K., Tsukui, T., de

la Pena, J., Sabbagh, W., Greenwald, J., Choe, S., Norris, D.P., Robertson, E.J., Evans, R.M.,

Rosenfeld, M.G., and Izpisua Belmonte, J.C. (1998) Nature 394, 545-551.

6. Amendt, B.A., Sutherland, L.B., Semina, E.V., and Russo, A.F. (1998) J. Biol. Chem. 273, 20066-

20072.

7. Duboule, D. (1994) Guidebook to the Homeobox Genes. Oxford University Press, New York, NY.

8. Dave, V., Zhao, C., Yang, F., Tung, C., and Ma, J. (2000) Mol. Cell. Biol. (In Press).

9. Lamonerie, T., Tremblay, J. J., Lanctot, C., Therrien, M., Gauthier, Y., and Drouin, J. (1996)

Genes Dev. 10, 1284-1295.

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

21

10. Szeto, D. P., Ryan, A. K., O'Connell, S. M., and Rosenfeld, M. G. (1996) Proc. Natl. Acad. Sci.

USA 93, 7706-7710.

11. Hanes, S.D., and Brent, R. (1989) Cell 57, 1275-1283.

12. Treisman, J., Gonczy, P., Vashishtha, M., Harris, E., and Desplan, C. (1989) Cell 59, 553-562.

13. Hanes, S.D., and Brent, R. (1991) Science 251, 426-430.

14. Qian, Y.Q., Otting, G., Billeter, M., Muller, M., Gehring, W., and Wuthrich, K. (1993) J. Mol. Biol.

234, 1070-1083.

15. Kozlowski, K., and Walter, M.A. (2000) Hum. Mol. Genet. 9(14), 2131-2139.

16. Horton, R.M., Ho, S. N., Pullen, J. K., Hunt, H. D., Cai, Z., and Pease, L. T. (1993) Meth. Enzymol

217, 270-279.

17. Tverberg, L.A., and Russo, A.F. (1993) J. Biol. Chem. 268, 15965-15973.

18. Berman, H.M., Westbrook, J., Feng, Z., Gilliland, G., Bhat, T.N., Weissig, H., Shindyalov, I.N.,

and Bourne, P.E. (2000) Nucl. Acids. Res. 28(1), 235-242.

19. Maurer, R.A., and Notides, A.C. (1987) Mol. Cell. Biol. 7, 4247-4254.

20. Rhodes, S.J., and Rosenfeld, M.G. (1996) J. Animal Sci. 74, 94-106.

21. Jacobson, E.M., Li, P., Leon-del-Rio, A., Rosenfeld, M.G., and Agarwal, A.K. (1997) Genes Dev.

11, 198-212.

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

22

22. Li, S., Crenshaw, E.B. 3rd, Rawson, E.J., Simmons, D.M., Swanson, L.W., and Rosenfeld, M.G.

(1990) Nature 347, 528-533.

23. Lanigan, T.M., and Russo, A.F. (1997) J. Biol. Chem. 272, 18316-18324.

24. Lin, C.R., Kioussi, C., O'Connell, S., Briata, P., Szeto, D., Liu, F., Izpisua-Belmonte, J.C.,

Rosenfeld, M.G. (1999) Nature 401, 279-282.

25. Gage, P.J., Suh, H., and Camper, S.A. (1999) Development 126, 4643-4651.

26. Bellmann, R., and Werr, W. (1992) EMBO J. 11, 3367-3374.

27. Amendt, B.A., Sutherland, L.B., and Russo, A.F. (1999) Mol. Cell. Biol. 19, 7001-7010.

28. Singh, S., Tang, H.K., and Saunders, G.F. (1998) J. Biol. Chem. 273, 21531-21541.

29. Yamagata, K., Yang, Q., Yamamoto, K., Iwahashi, H., Miyagawa, J., Okita, K., Yoshiuchi, I.,

Miyazaki, J., Noguchi, T., Nakajima, H., Namba, M., Hanafusa, T., and Matsuzawa, Y. (1998)

Diabetes 47, 1231-1235.

30. Stoffers, D.A., Stanojevic, V., and Habener, J.F. (1998) J. Clin. Invest. 102, 232-241.

31. El Ghouzzi, V., Lajeunie, E., Le Merrer, M., Cormier-Daire, V., Renier, D., Munnich, A., and

Bonaventure, J. (1999) Europ. J. Hum. Genet. 7, 27-33.

32. Howard, T.D., Paznekas, W.A., Green, E.D., Chiang, L.C., Ma, N., Ortiz De Luna, R.I., Delgado,

C.G., Gonzalez-Ramos, M., Kline, A.D., and Jabs, E.W. (1997) Nature Genet. 15, 36-41.

33. Baldwin, C.T., Hoth, C.F., Macina, R.A., and Milunsky, A. (1995) Am. J. Med. Genet. 58, 115-122.

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

23

34. DeStefano, A.L., Cupples, L.A., Arnos, K.S., Asher, J.H., Jr., Baldwin, C.T., Blanton, S., Carey,

M.L., da Silva, E.O., Friedman, T.B., Greenberg, J., Lalwani, A.K., Milunsky, A., Nance, W.E.,

Pandya, A., Ramesar, R.S., Read, A.P., Tassabejhi, M., Wilcox, E.R., and Farrer, L.A. (1998)

Hum. Genet. 102, 499-506.

35. Prosser, J., and van Heyningen, V. (1998) Hum. Mutat. 11, 93-108.

36. Hanson, I., Churchill, A., Love, J., Axton, R., Moore, T., Clarke, M., Meire, F., and van Heyningen,

V. (1999) Hum. Molec. Genet. 8, 165-172.

Figure legends

Figure 1. Identification of the K88E mutation in a Rieger patient. a) The pedigree of the affected

proband and family showing a de novo mutation. b) The maxillary occlusal radiograph on the left shows

conically shaped primary central and lateral incisors. The radiodensity of the enamel suggests either

hypoplastic or hypomineralized enamel. The primary canines have bulbous crowns. There is no

evidence of the developing permanent incisor crowns, suggesting that they are congenitally missing. The

mandibular occlusal radiograph on the right shows small primary central and lateral incisors with

hypoplastic or hypomineralized enamel. The developing permanent incisor crowns are conical in shape.

The lower right permanent central incisor is congenitally missing. c) Sequence of the PITX2 fragment

from genomic DNA of unaffected (1) and affected (2) members of the family. Nucleotide substitution of A

to G in one allele from affected individual results in a change of a codon for lysine (AAG) to a codon for

glutamic acid (GAG) at position 88.

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

24

Figure 2. K88E mutant protein does not bind the bicoid DNA element. a) Increasing amounts (100,

200, and 300 ng) of wildtype PITX2 and K88E mutant proteins were incubated with a probe containing the

bicoid consensus sequence (TAATCC) in an electrophoretic mobility shift assay (EMSA). The bound

complex is indicated. Competitor lanes indicate incubation of 200 ng protein with 33-fold molar excess of

unlabeled probe prior to incubation with labeled probe. The autoradiograph was purposefully

overexposed to maximize potential detection of a K88E-DNA complex. b) Coomassie-stained gel of the

PITX2 and K88E protein preparations used in panel a (following PreScission protease cleavage to

remove the GST moiety), with prestained molecular weight standards (Gibco BRL) indicated. The

wildtype and K88E proteins are the same size. c) 5µl of wildtype PITX2, K88E, and T68P in vitro

transcription and translation (TNT) products were incubated with the bicoid probe in an EMSA. The

bound complex is indicated. d) An autoradiograph of a gel electrophoresed with 2 µl of wildtype PITX2,

K88E, and T68P TNT products showing comparable expression. The molecular weight markers (Gibco

BRL) are indicated. The wildtype, K88E, and T68P TNT products are the same size.

Figure 3. Computer model of the K88E mutant homeodomain-DNA complex. The PITX2

homeodomain-DNA complex was modeled upon the NMR solution structure of an antennapedia

homeodomain-DNA complex using the SYBYL program. The homeodomain (yellow) and DNA (dark

green) are displayed as a ribbon-tube except for lysine (wild-type PITX2) or glutamic acid (K88E mutant)

at position 50 of the homeodomain and GG dinucleotide representing the antisense bases to positions 5

and 6 of the hexanucleotide DNA binding sequence (TAATCC).

Figure 4. K88E mutant protein has little or no transactivation activity. a) COS7 cells were

transfected with a luciferase reporter plasmid containing the thymidine kinase promoter with four

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

25

upstream bicoid consensus sites (TK-Bic) alone or in combination with PITX2 or K88E. The luciferase

activity is shown as the mean fold-activation compared to reporter alone ± S.E.M. from 5 independent

experiments. *, significantly different from wildtype PITX2 (p < 0.018). b) The transfections in (a) were

repeated in CHO cells with a luciferase reporter plasmid containing the prolactin promoter. The data from

4 independent experiments are shown as the mean fold-activation compared to reporter alone ± S.E.M.

*, significantly different from wildtype PITX2 (p < 0.001). c) Western blot using anti-c-Myc antibody of

equal amounts of cell extracts (20 µg) from one of the cotransfection experiments in CHO cells. The cells

were transfected with the indicated amounts of either wildtype and/or K88E PITX2 expression vector

DNAs. The prestained molecular weight standards (Gibco BRL) and PITX2 bands are indicated. The WT

PITX2 and K88E proteins are the same size.

Figure 5. K88E mutant protein prevents wildtype PITX2 synergism with Pit-1. CHO cells were

transfected with the prolactin-luciferase reporter DNA alone or with PITX2, K88E and/or T68P in the

presence and absence of Pit-1 (10 µg) as indicated. Cotransfection of equal amounts of K88E mutant

with wildtype (WT) PITX2 (5 µg each) yielded decreased activation relative to 5 µg of WT PITX2 alone (*,

p < 0.001) or 10 µg of PITX2 (p < 0.006). The activity is shown as mean fold-activation compared to

reporter alone ± S.E.M. from 5 independent experiments.

Figure 6. K88E mutant protein suppresses wildtype PITX2 synergism with Pit-1 in a dose-

dependent manner. CHO cells were cotransfected with prolactin-luciferase reporter (5µg), Pit-1 (10 µg),

WT PITX2 (5 µg), and the indicated amounts (µg) of additional WT PITX2, K88E or T68P DNA.

Cotransfection of increasing amounts of K88E mutant DNA with WT PITX2 shows a significant decrease

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

26

in activation relative to 5 µg of WT PITX2 alone (*, p < 0.028; **, p < 0.006). The data represent 4

independent experiments with the mean fold-activation ± S.E.M.

Figure 7. Wildtype PITX2 can rescue the dominant negative effect of the K88E mutant. CHO cells

were cotransfected with prolactin-luciferase reporter (5µg), Pit-1 (5 µg), and the indicated amounts (µg) of

WT PITX2, or K88E DNA. Cotransfection of equal amounts of K88E mutant DNA with WT PITX2 in the

presence of Pit-1 (5 µg each) (lane 2) shows a significant decrease in activation relative to 5 µg of WT

PITX2 alone with Pit-1 (lane 1) (*, p < 0.006). Cotransfection of 10 µg of WT PITX2 with 5µg of K88E in

the presence of Pit-1 (lane 3) increased the activation significantly higher than the activation with equal

amounts of WT PITX2 and K88E (lane 2) (**, p < 0.015). The data represent 3 independent experiments

with the mean fold-activation ± S.E.M.

Figure 8. K88E mutant cannot suppress MEKK activation of the prolactin promoter. CHO cells

were transfected by lipofection with prolactin-luciferase reporter plasmid (0.5 µg) alone or in combination

with MEKK, and/or K88E mutant (0.5 µg each). The luciferase activity is shown as the mean fold-

activation compared to reporter alone ± S.E.M. from 4 independent experiments.

Figure 9. Pit-1 binding site is required for the dominant negative effect of the K88E mutant. COS7

cells were transfected by lipofection with TK-Bic reporter plasmid (1.5 µg) alone or in combination with

Pit-1, PITX2, K88E, and/or T68P (1.5 µg each). The luciferase activity is shown as the mean fold-

activation compared to reporter alone ± S.E.M. from 4 independent experiments.

Figure 10. Pit-1 DNA binding is required for the dominant negative effect of K88E mutant. CHO

cells were transfected by lipofection with the prolactin-luciferase reporter DNA alone (0.5 µg) or with

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Saadi et al.

27

PITX2, PITX2 ∆C39, K88E, Pit-1, Pit-1 ∆8-128, Pit-1 ∆POU, and/or Pit-1 W261C (0.5 µg each) as

indicated. Cotransfection of equal amounts of K88E mutant and wildtype (WT) Pit-1 with WT PITX2 (0.5

µg each; lane 5; **, p < 0.009) or PITX2 ∆C39 (0.5 µg each; lane 8; **, p < 0.001) yielded decreased

activation relative to WT Pit-1 and WT PITX2 (lane 4) or PITX2 ∆C39 (lane 7), respectively. Similarly,

cotransfection of wildtype PITX2, K88E, and Pit-1 ∆8-128 (lane 11; **, p < 0.017) yielded decreased

activation than wildtype PITX2 and Pit-1 ∆8-128 (lane 10). Cotransfection of wildtype PITX2, and K88E

with Pit-1 ∆POU (lane 14; *, p < 0.04) or Pit-1 W261C (lane 17; *, p < 0.04) yielded marginally decreased

activation than wildtype PITX2 and Pit-1 ∆POU (lane 13) or PITX2 and Pit-1 W261C (lane 16),

respectively. The activity is shown as mean fold-activation compared to reporter alone ± S.E.M. from 4

independent experiments.

Figure 11. K88E mutant protein suppresses prolactin promoter activity in GH3 cells. a) GH3 cells

were transfected by lipofection with the prolactin-luciferase reporter DNA alone or with Pit-1, PITX2,

K88E, or T68P (1 µg each) as indicated. Transfection of K88E mutant yielded decreased activation

relative to WT PITX2 alone (*, p < 0.005). The activity is shown as mean fold-activation compared to

reporter alone ± S.E.M. from 18 independent experiments. b) GH3 cells were transfected by lipofection

with the TK-Bic reporter DNA alone or with Pit-1, PITX2, K88E, or T68P (1 µg each) as indicated. The

activity is shown as mean fold-activation compared to reporter alone ± S.E.M. from 4 independent

experiments.

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

c)

a)

b)

Figure 1

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Bound

WTPITX2 K88EC

omp

Pro

be O

nly

Com

p

30

60

40

20

WT P

ITX2

K88E

PITX2

a) b)

Bound

WT

PITX2

K88E

T68P

Probe

Onl

y

c)

WT

PITX2

K88E

T68Pd)

70

60

40PITX2

Figure 2

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Wildtype PITX2

K50

G5G6

K88E

E50

G5G6

PITX2 homeodomain

DNA

Sense: TAATC5C6

Antisense: ATTAG5G6

CarbonOxygenHydrogenNitrogenPhosphorus

Figure 3

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

0

1

2

3

4

5

6

1 2 3 4

Fo

ld A

ctiv

atio

n

PITX2K88E

--

+- +

-

a)

*

++

LucTK-Bic

b)

0

2

4

6

8

10

12

14

16

18

1 2 3 4

Fo

ld A

ctiv

atio

n

PITX2K88E

--

+- +

-

LucProlactin 2.5

*

++

*

Figure 4

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

60

40

30

K88E - - 55

PITX2 5 10 5-

-

-

PITX2

c)

Figure 4

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

0

50

100

150

200

250

300

350

400

450

*

Fo

ld A

ctiv

atio

n

K88E - 55- 5-PITX2 5

-

-10 55

Pit-1 +

-

++ +- -

-

-

+

-

-

T68P 5

5

+

- - - - - 5

-

- - -

-

-

-

5

-

-

+

LucProlactin 2.5

1 2 3 4 5 6 7 8 119 10

Figure 5

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

0

20

40

60

80

100

120

140

160

1 2 3 4 5 6

WT PITX2 55 55 5

WT, K88E,or T68P

1-

5

WT

K88E

Act

ivat

ion

(% c

on

tro

l)

*

**

*LucProlactin 2.5

T68P

32 4 5

Figure 6

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

PITX2+

10Pit-1

K88E5

-5+

5 5

+

0

20

40

60

80

100

120

1 2 3

*

LucProlactin 2.5

Figure 7

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

MEKK +K88E

+--

- +

0

1

2

3

4

5

6

7

1 2 3

LucProlactin 2.5

Figure 8

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

0.0

1.0

2.0

3.0

4.0

5.0

6.0

7.0

8.0

1 2 3 4 5 6 7 8 9 10 11 12

Fo

ld A

ctiv

atio

n

K88E - -++ --PITX2 +

+

--+

Pit-1 +

-

-- +- -

-

-

+

-

-

T68P +

+

-

- - - + + -

-

- - -

+

+

-

-

+

+

+

+

+

+

-

-

LucTK-Bic

Figure 9

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

0.0

20.0

40.0

60.0

80.0

100.0

120.0

140.0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17

Pit-1 + --+ --K88E -

-

-+-

PITX2 +

-

++ ++ -

-

+

-

-

-

Pit-1 ∆8-128 -

-

-

- - + + + -

-

- - -

-

-

+

-

-

+

+

-

-

+

-

+

Fo

ld A

ctiv

atio

n

Pit-1 ∆POU

Pit-1 W261C

+

+

-

-

-

-

-

-

-- - - - - +- - - + - -+

+- - - - - -- - - - + +-

LucProlactin 2.5

****

* *

-- - - - - -- - - - - --PITX2 ∆C39

+ +

- +

- --

-

-

- --

- --

- --

+ ++

**

Figure 10

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1.4

1.6

1.8

1 2 3 4 5

Fo

ld A

ctiv

atio

n

b)

PITX2

++

++

Pit-1

T68P

K88E-

----

- -- -

---

-

- --

LucTK-Bic

0.0

0.2

0.4

0.6

0.8

1.0

1.2

1 2 3 4 5

Fo

ld A

ctiv

atio

n

a)

PITX2

++

++

Pit-1

T68P

K88E-

----

- -- -

---

-

- --

LucProlactin 2.5

*

Figure 11

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from

Jeffrey C. Murray and Andrew F. RussoIrfan Saadi, Elena V. Semina, Brad A. Amendt, David J. Harris, Kenneth P. Murphy,

Identification of a dominant negative homeodomain mutation in Rieger syndrome

published online April 11, 2001J. Biol. Chem. 

  10.1074/jbc.M008592200Access the most updated version of this article at doi:

 Alerts:

  When a correction for this article is posted• 

When this article is cited• 

to choose from all of JBC's e-mail alertsClick here

by guest on March 27, 2020

http://ww

w.jbc.org/

Dow

nloaded from