Missense mutation at the C-terminus of PAX6 negatively modulates ...

© 2001 Oxford University Press

Human Molecular Genetics, 2001, Vol. 10, No. 9

911–918

Missense mutation at the C-terminus of PAX6 negatively modulates homeodomain function Sanjaya Singh, Lian Y. Chao, Rajnikant Mishra, Jonathan Davies and Grady F. Saunders+ Department of Biochemistry and Molecular Biology, The University of Texas M.D. Anderson Cancer Center, Box 117, 1515 Holcombe Boulevard, Houston, TX 77030, USA Received 12 December 2000; Revised and Accepted 20 February 2001

PAX6 is essential for ocular morphogenesis. Mutations in the PAX6 gene produce various phenotypes, including aniridia, Peters’ anomaly, foveal hypoplasia, autosomal dominant keratitis and congenital cataracts. PAX6 functions as a transcription factor and has two DNA binding domains (a paired domain and a homeodomain) which are joined by a linker, and a transactivation domain enriched in proline, serine and threonine (PST) at the C-terminus. The mechanism of PAX6 function is not clearly understood, and few target genes in vertebrates have been identified. We examined disease-causing missense mutations in the PST domain to understand how they affect the function of PAX6. Upon examining the DNA samples of aniridia patients, we identified three missense mutations in the PST domain: P375Q (a novel mutation) and the previously reported Q422R and X423L mutations. On the basis of functional analysis, the P375Q mutant appears to have a normal transactivation activity but lower DNA binding through the paired domain than the wild-type. The Q422R mutation resulted in the loss of DNA binding ability of the PAX6 homeodomain. Substitution analyses of the C-terminal amino acid (codon 422) indicated that an amino acid at codon 422 is required for DNA binding of the homeodomain of intact PAX6 and that the polarity and charge of the side-chain of the terminal amino acid influence this binding. INTRODUCTION PAX6 is an evolutionarily conserved gene in both vertebrates and invertebrates. It is considered the master control gene for morphogenesis of the eye (1–3). Complete loss of PAX6 function leads to anophthalmia and nasal hypoplasia, as well as to central nervous system defects that cause postnatal death (4,5). The PAX6 protein functions as a transcription factor. In PAX6, two DNA binding domains, the paired domain (PD) at the N-terminus and a paired-like homeodomain (HD) at the center of the protein, are linked by a glycine-rich domain. The transactivation domain, enriched in proline, serine and threonine (PST) and called the PST domain, occupies the 152 amino acids at the C-terminus of the PAX6 protein (Fig. 1). Our +To

earlier studies showed that the transactivation function is distributed throughout the PST domain (6). Although the involvement of PAX6 in ocular development is well known (7), biochemical identification of target genes of PAX6 has been hindered by two factors. Firstly, the PD of PAX6 can bind to a broad range of DNA sequences. Secondly, HDs in other proteins recognize sequences similar to those recognized by the PAX6 HD (8). In humans, heterozygous mutations in the PAX6 gene are responsible for several phenotypes, including aniridia, foveal hypoplasia, Peters’ anomaly, ectopia pupillae and autosomal dominant keratitis (9–14). PAX6 mutation in rodents results in small eye (15), but exhibits highly variable phenotypes in humans. The molecular basis of variable expressivity of phenotypes due to PAX6 mutation is not clear. Most of the mutations identified in PAX6 cause deletion or truncation of the PAX6 protein, resulting in haploinsufficiency. However, as more patient samples are analyzed, the number of missense mutations identified continues to increase (16,17). A list of PAX6 mutations is available on the web site: http:// www.hgu.mrc.ac.uk/Softdata/PAX6. Functional studies using PAX6 missense mutations with observed developmental defects are useful in understanding the mechanism of PAX6 function. Using mutations found in aniridia patients, we demonstrated that the two subdomains of the PD differentially influence the transactivation activity of the HD (18). This finding, in combination with our earlier studies (19), provides an explanation for the differential phenotypes seen with missense mutations located in the two different subdomains of the PD. Missense mutations in the DNA binding domains of PAX6 may cause failure to properly recognize binding sites in the target genes, resulting in partial or complete loss of protein function (6,18,20,21). Recently, several laboratories have identified missense mutations in the PST domain of PAX6 (as listed in the web site cited above) (9,22–24) which resulted in eye phenotypes. Some of the missense mutations in the C-terminal amino acid cause severe eye phenotypes (22,23). In this report, we describe three missense mutations in the PAX6 gene which were identified in the DNA of aniridia patients (Fig. 1). These mutations are located in the transactivation region (PST domain) of PAX6. We have analyzed the functional significance of these missense mutations. Our studies demonstrate that the polarity and charge of the side-chain of

whom correspondence should be addressed. Tel: +1 713 792 2690; Fax: +1 713 791 9478; Email: [email protected]

912


Figure 1. Schematic diagram of the PAX6 protein with missense mutations described in the report.

the C-terminal amino acid of PAX6 influence the DNA binding function of its HD but not of its PD.

More than 80 samples were examined in our laboratory and none showed mutations at this position.

RESULTS

The Q422R mutation in PAX6 results in loss of the HD-mediated transactivation function

Missense mutations in the C-terminal transactivation domain of PAX6 DNA samples from various aniridia patients were tested for the presence of mutations in the PAX6 gene. In this report, we describe one novel and two previously documented missense mutations in the transactivation domain of PAX6. An A→G transition that causes missense mutation Q422R was identified in patient 1. This patient, a 3-year-old girl, has bilateral aniridia. In addition to a small iris remnant, the patient had cataracts, mild lens subluxation, foveal hypoplasia and horizontal nystagmus in each eye. This mutation was reported earlier from Japan (22). The DNA sample of patient 2 revealed an A→T change either due to transversion or insertion resulting in the X423L mutation. The poly-dA sequence 3′ of the mutation site makes sequencing difficult. Polymerase chain reaction (PCR) products and several cloned fragments were analyzed to determine the number of A nucleotides in the poly-dA stretch following the mutation, but the results were inconclusive. The number of As varied between 21 and 24. This mutation was reported earlier by Baum et al. (23), who were also unable to determine the number of As in the poly-dA stretch following the mutation. Genomic sequencing of this site has shown 21 As (GenBank accession no. GI: 6966809) or 23 As (EMBL, Z95332). In the absence of an accurate determination of the number of As, it is difficult to determine whether this mutation was due to a transversion or insertion. This patient, a 3-month-old Hispanic girl, had poor vision, nystagmus, absence of the inner margin of the iris, a moth-eaten appearance of the pupillary margin and 2 mm of plaque on the anterior surface of the lens. The findings were bilaterally symmetrical. Analysis of the patient’s DNA on heteroduplex gel and comparison with the parents’ DNA revealed that this case was sporadic, since the parents did not have this mutation (data not shown). A third aniridia patient was analyzed by heteroduplex analysis followed by sequencing. A mutation was detected that had a tranversion of C→A, resulting in a P375Q missense mutation. P375 is a highly conserved amino acid in PAX6 (6).

To assess the functional significance of the Q422R mutation and to understand the role of the C-terminal amino acid, we compared the mutant and wild-type PAX6 proteins in in vivo transcriptional activation assays. The Q422R and Q422X mutant cDNAs were created by site-directed mutagenesis. The Q422X mutant was constructed to evaluate the role of amino acid 422 of PAX6. NIH/3T3 cells were cotransfected with PAX6, PAX6 (Q422R) or PAX6 (Q422X) expression plasmids and a luciferase reporter plasmid bearing either two copies of the PD-binding site CD19-2 (A-ins) or two copies of the HD-binding site P2 (18). When the PD DNA binding was used in the reporter plasmid, the ability of the mutant PAX6 proteins to activate transcription of the reporter gene was only moderately lower than that of wild-type PAX6 (Fig. 2A). However, when the HD DNA binding was used in the reporter plasmid, activation of the reporter gene was significantly lower with the mutant PAX6 protein than in the wild-type (Fig. 2B). Protein expression of the wild-type and mutant constructs was confirmed by western blot analysis using anti-PAX6 antibodies. The proteins were expressed in equal amounts from the wild-type and mutant constructs (Fig. 2C). Loss of transactivation function upon replacing glutamine at codon position 422 with an amino acid having a non-polar side-chain In the Q422R mutation results in change from a glutamine residue, which has an uncharged polar side-chain, to arginine residue, which has a positively charged side-chain. We wanted to determine whether the polarity of the side-chain of the C-terminal amino acid influences the transactivation function mediated through the HD DNA binding site. We generated expression constructs in which glutamine at codon 422 was replaced with amino acids that have polar but uncharged side-chains (threonine and asparagine), or with an amino acid that has a nonpolar aliphatic side-chain (leucine), or with an amino acid that has a negatively charged side-chain (glutamate). NIH/3T3 cells were cotransfected with PAX6 or various mutant PAX6 expression plasmids with a luciferase reporter plasmid bearing either two copies of the PD binding site, (CD19-2 (A-ins) or


913

two copies of the HD binding site), P2. Compared with wildtype PAX6, the PAX6 (Q422R) and PAX6 (Q422L) mutants demonstrated slightly reduced transactivation activity and PAX6 (Q422N) had slightly higher activity when the PD binding site CD19-2 (A-ins) was used in the reporter plasmid (Fig. 3A). However, these values were not significantly different. When the HD binding site P2 was used in the reporter plasmid, PAX6 (Q422R) PAX6 (Q422L) and PAX6 (Q422X) resulted in significantly lower transactivation than wild-type PAX6 (Fig. 3B and C) whereas other replacement mutants did not have a significant effect on transactivation (Fig. 3B and C). These results demonstrate that when the C-terminal amino acid has a side-chain that is either positively charged or non-polar, the transactivation function mediated through the HD is lost. Similar results are found when the C-terminal amino acid is removed (Fig. 3C). However, when the C-terminal amino acid has a side-chain that is either polar or negatively charged, the transactivation function appears to be normal (Fig. 3B and C). All the constructs gave similar levels of protein expression as confirmed by western blot analysis (Fig. 3D). DNA binding of PAX6 missense mutants The mutant PAX6 (Q422R) provided transactivation comparable to that of the wild-type when a PD DNA binding site was used, indicating that the structure required for the transactivation function was most likely intact in this mutant. Because failure to bind at a specific DNA binding site is another possible reason for the loss of function of a transcription factor, we conducted electrophoretic mobility shift assays (EMSAs). In our previous reports (6,18,25), we extensively characterized the specificity of DNA binding of PAX6 through both of its DNA binding domains. When a PD binding site was used as a probe, the level of binding was slightly lower with all the mutant PAX6 proteins examined than with wild-type. Interestingly, the PAX6 (Q422X) protein–DNA complex seemed to migrate faster on the gel than the other mutant proteins (Fig. 4A). When a HD DNA binding site was used as a probe, the PAX6 (Q422X) and PAX6 (Q422R) mutants failed to bind (Fig. 4B). The abundance of translated protein was determined on SDS-polyacrylamide gels (Fig. 4C). We also tested other PAX6 mutants used in previous transactivation studies for DNA binding. Although DNA binding through both DNA binding domains (PD and HD) was intact in the other mutants (Fig. 5A and B), PAX6 (Q422L) did not bind to the HD binding site P2 (Fig. 5B). Functional analysis of mutation P375Q

Figure 2. Differential transactivation of the luciferase reporter gene by mutant PAX6 (Q422R) from a PD and HD binding site. NIH/3T3 cells were transiently transfected with 0.1 µg of PAX6 and PAX6 mutant constructs, as indicated, together with CD19-2 (A-ins)-luc (A) or P2-luc reporter plasmid (B). The luciferase activity of the reporter construct is shown as the mean + SD of three separate experiments. Western blot analysis was performed to evaluate the amount of protein expressed by the expression constructs (C). Extracts (10 and 20 µl) from cells transfected with 0.1 µg of expression plasmids were resolved on a 10% SDS-polyacrylamide gel and immunoblotted using anti-PAX6 antibodies. The results showed that all the expression constructs expressed almost equal amounts of protein.

The effect of the P375Q mutation on PAX6 function was evaluated by DNA binding and transactivation studies. The amount of protein synthesized by in vitro translation using a reticulocyte lysate was determined by western blot analysis. Both proteins were synthesized at an equal level (Fig. 6A). To compare the DNA binding ability of the mutant with that of wild-type PAX6, we performed gel-shift analyses with the HD and PD DNA binding sites. No significant difference was seen in the DNA binding abilities of wild-type PAX6 and the P375Q mutant protein when the HD DNA binding P2 was used (Fig. 6B). Similarly, transactivation studies with the P2-luc reporter plasmid did not show significant difference in transactivation

914


Figure 3. The side-chain of the C-terminal amino acid of PAX6 influences the transcriptional activation of a luciferase reporter gene with a HD binding site. NIH/3T3 cells were transiently transfected with 0.1 µg of PAX6 and PAX6 mutant constructs, as indicated, together with CD19-2 (A-ins)-luc (A) or P2–luc reporter plasmids (B and C). The luciferase activity of the reporter construct is shown as the mean + SD of three separate experiments. Western blot analysis was performed to assess the amount of protein expressed by the expression constructs (D). Extracts (20 µl) from cells transfected with 0.1 µg of expression plasmids were resolved on a 10% SDS-polyacrylamide gel and immunoblotted using anti-PAX6 antibodies.

between these two proteins (data not shown). However, preliminary data indicated that when a PD DNA binding site CD19-2 (A-ins) was used, DNA binding was significantly lower in the P375Q mutant than in the wild-type PAX6. Therefore, in an effort to compare the differences in affinity between wild-type and mutant proteins, we used increasing concentrations of wild-type and mutant protein individually with a constant amount of the CD19-2 (A-ins) probe. Compared with the mutant, the wild-type protein showed consistently stronger binding (Fig. 6C). The bands were quantitated using a PhosphorImager and the amount of bound probe was plotted versus total probe. The data showed that the binding affinity for the PD DNA binding site was at least 2.5 times lower for the P375Q mutant (Fig. 6D). We also examined whether the lower DNA binding of the P375Q mutant affected its transactivation function. Increasing amounts of either the PAX6 or PAX6 (P375Q) mutant plasmids were cotransfected with the luci-

ferase reporter plasmid containing the CD19-2 (A-ins) PD DNA binding site. At lower dosages of the PAX6 plasmids (12.5 and 25 ng) transactivation by the wild-type PAX6 was about 2 times higher than the P375Q mutant, which is consistent with the DNA binding data (Fig. 6D). However, at the higher dosage of plasmids only slight differences were found between wild-type and mutant PAX6 (Fig. 6E). Again the similar protein expression levels of PAX6 and mutant constructs were confirmed by western blot analysis (Fig. 6F). DISCUSSION In this study, we describe three missense mutations in the PST domain of the PAX6 gene identified in three different aniridia patients. Functional studies with expression constructs producing various mutant proteins revealed that polarity and


915

Figure 5. DNA binding of various mutants of PAX6 to a PD binding site and a HD binding site. Wild-type and mutant PAX6 proteins synthesized in vitro were used in an EMSA to bind paired domain binding site CD19-2 A-ins) (A) and HD binding site P2 (B).

Figure 4. Mutant PAX6 (Q422R) fails to bind HD binding site P2. Two and four microlitres of the wild-type and mutant PAX6 proteins synthesized in vitro were used in an EMSA to bind the paired domain binding site CD19-2 (A-ins) (A) and HD binding site P2 (B). Because of the bipartite nature of P2, the analysis detected both monomeric and dimeric bands (marked by arrows). (C) Autoradiogram of the SDS–PAGE of in vitro translated wild-type and mutant PAX6 protein showing that all the proteins were used in similar quantities.

charge of the R side-chain of the C-terminal amino acid of PAX6 modulates the function of its HD. DNA binding domains of PAX6 are highly conserved among animal species and missense mutations in the DNA binding domains of PAX6 may result in the failure to properly recognize the binding sites in the target genes, thus causing partial or complete loss or gain of function. Unlike the DNA binding domains the PST domain shows a great degree of sequence

divergence at the amino acid level among animal species (6). Despite this divergence, the C-terminal amino acid glutamine is conserved in all known vertebrate PAX6 proteins and also in invertebrates including twin of eyeless (toy) (Fig. 7). toy is upstream of eyeless (ey) and regulates the expression of ey in Drosophila (26). As this amino acid is so highly conserved, it is not surprising that the mutation Q422 results in a disease phenotype. It is interesting to note that unlike toy there is little homology in the six C-terminal amino acids of vertebrate Pax6 and ey. Our data demonstrate that the Q422R mutation negatively modulates the function of the HD of PAX6. This modulation depends on the charge and polarity of the sidechain of the C-terminal amino acid. Mutations at this position appear to have little effect on the function of the PD. Most of the mutations identified in the PAX6 gene are unique (17), although some mutations have been identified 2–6 times. Mutations in codon 240 of PAX6 are more frequent than for other PAX6 codons (17). Most of the mutations are either located in the DNA binding regions of the PAX6 protein or result in early termination of the protein (17). Disproportionately fewer mutations have been identified in the PST region than would be expected (17; http://www.hgu.mrc.ac.uk/Softdata/PAX6). It is possible that some alterations in the PST region lead to slightly altered activity of the PAX6 gene and that the resulting milder phenotypes may have been clinically less remarkable than those found in aniridia patients. Although it is also possible that some mutations have no phenotypic effect, the extreme sequence conservation of the PAX6 gene would predict otherwise. Mutations Q422R and X423L are both located in exon 13 and have been previously identified by other investigators (22,23). Like the patient identified by Azuma and Yamada (22), the patient who we identified with a Q422R mutation had iris abnormalities but there are remarkable phenotypic differences between the two patients. For example, our patient had cataracts at the age of 3 years whereas the other patient (22), when examined at the age of 10 years, did not have cataracts but had corneal opacities. Unlike these two patients with Q422R mutations, different patients with the

916


Figure 7. C-terminal amino acid glutamine (Q) is highly conserved during evolution. The last 10 amino acids of PAX6 protein from various animal species are aligned. The Drosophila sequence is from toy (twin of eyeless). Relative to ey, the overall homology of toy is closer to PAX6.

Figure 6. Effect of P375Q mutation on DNA binding and transactivation function of PAX6. (A) Western analysis of in vitro translated wild-type and mutant PAX6 protein showing the proteins were used in equal quantities. (B) Two and four microliters of the wild-type PAX6 and mutant PAX6 (P375Q) proteins synthesized in vitro were used in an EMSA to bind HD binding site P2. (C) Increasing amounts (0.5–3.0 µl) of in vitro synthesized wild-type PAX6 and mutant PAX6 (P375Q) proteins were used in an EMSA to bind PD binding site CD19-2 (A-ins). (D) Bands were quantitated in gels as described in (C) by a PhosphorImager and bound probe was plotted as percentage of total probe. (E) NIH/3T3 cells were transiently transfected with various amounts of PAX6 and PAX6 mutant constructs, as indicated, together with CD19-2 (A-ins)-luc. The luciferase activity of the reporter construct is shown as the mean + SD of three separate experiments. (E) Extracts (20 µl) from cells transfected with 0.1 µg of expression plasmids were resolved on a 10% SDS-polyacrylamide gel and immunoblotted using anti-PAX6 antibodies.

X423L mutation have very similar eye phenotypes. There is no clear explanation as to why the same missense mutation

resulted in variation of phenotype. It is possible that modifier alleles are responsible for causing variation in the phenotype. DNA binding of the P375Q mutant was at least 2 times weaker than wild-type PAX6. When a higher dosage of expression plasmids of PAX6 and mutant protein was used, the difference in the transactivation level was not substantial. This result indicated that the transactivation function of this mutant might be intact. However, consistent with the findings on DNA binding, a lower dosage of expression vectors demonstrated a significant difference in the transactivation level using wildtype PAX6 and the P375Q mutant. These findings indicate that, in vivo, the P375Q allele may result in PAX6 haploinsufficiency. Proline has a non-polar side-chain, whereas glutamine has a polar side-chain that may contribute to structural changes affecting DNA binding domains, causing PAX6 (P375Q) to bind DNA with lower affinity than that observed in the wild-type or in other mutants. Simultaneously, in addition to the difference in polarity between proline and glutamine, the unique cyclic structure of proline constrains the structure of polypeptide backbone more than glutamine would. Our earlier reports (6,25) demonstrate that truncated PAX6 proteins bind to DNA through both the PD and HD binding domains with considerably higher affinity compared with the wild-type. Chow et al. (27) demonstrated that ectopic expression of truncated Pax6 protein inhibits eye formation in Xenopus, a finding consistent with our results. Although mRNA degradation may prevent truncated proteins from being expressed in vivo (28), these studies (6,25,27) demonstrate that the PST domain participates in modulating the DNA binding function of PAX6. Similarly, there is interdomain interaction(s) between PD and HD. In a recent study (18), we reported evidence of a differential interaction between the two subdomains of the PD and HD. Structural changes in the α-helix region of the N-terminal subdomain of PD enhance HD-mediated DNA binding, whereas mutation in the C-terminal subdomain of the PD negatively modulates the DNA binding function of the HD. Studies by Underhill et al. (29,30) with Pax3 demonstrate that structural changes in the conserved β-sheet in the N-terminal region of the PD not only affect the DNA binding function of the PD but


negatively modulate the function of the HD. Another study with Pax3 showed that the HD can influence PD sequence specificity (31). The data presented in this paper together with the findings discussed above suggest that the C-terminal transactivation region of PAX6 likely interacts with DNA binding domains which may in turn facilitate the interaction between the two DNA binding domains. This interaction may occur either by direct contact of the PST domain with the PD or through conformational alterations of the PAX6 protein. Mutation at the most C-terminal amino acid may influence the conformation of PAX6 preventing HD DNA binding. Enhanced DNA binding affinity observed with the truncated PAX6 proteins (6,18,25) provide further support in favor of this possibility. Truncation of the PST domain could free the DNA binding domains from the constraint of this interaction, thus enhancing the DNA binding of the truncated proteins.

917

Western blotting and EMSAs

MATERIALS AND METHODS

Crude nuclear extracts were prepared from transfected NIH/ 3T3 cells as described previously (34). In vitro-transcribed and -translated proteins were generated using the TNT coupled reticulocyte lysate system (Promega) with the pRc–CMV constructs described above. The amount of PAX6 protein expressed or translated was quantitated by western blot analysis with antibodies against the linker region of PAX6 as described previously (18). The bands on the western blots were quantitated using a Personal Densitometer Scanner 1.30 and Image Quant 3.3 software (Molecular Dynamics). Occasionally, quantitation of in vitro synthesized 35S-labeled proteins was further confirmed by drying the SDS-PAGE gels and analyzing using a Storm 840 PhosphorImager (Molecular Dynamics) and Image Quant 5.0 software (Molecular Dynamics). EMSAs for PD DNA binding and for HD DNA binding were carried out as described previously (18). The gel was dried and analyzed with a PhosphorImager followed by autoradiography on Kodak X-Omat AR films.

Mutation detection

ACKNOWLEDGEMENTS

To identify mutations in the PAX6 gene, genomic DNA samples were isolated from peripheral blood lymphocytes of aniridia patients. Each exon and its immediate flanking sequence was amplified by PCR with the primer sets described previously (10), except for the reverse primer of exon 13, which was 5′-CAC CAA AAT GAA TAA AAG TTT G-3′ (9, 32). Gel analysis and sequence confirmation were performed as described by Chao et al. (9).

We are grateful to Ms Ellen Taub, Ms Melody White and Drs Sylvia Crago and Louise Strong for clinical material. We also acknowledge Drs R. Mass and M. Busslinger for the recombinant vectors. We thank Rama Grenda for preparing the figures. The DNA sequencing work was done by the core sequencing facility at M. D. Anderson Cancer Center. This research was supported by National Institutes of Health grants EY09675, EY10608 and CA 16672.

Plasmid constructs Rc–CMV–PAX6 expression plasmids were constructed using a PCR cloning strategy. The forward primer used contains a HindIII site and a Kozak site from the PAX6 gene. The reverse primers contain a XbaI site and a stop codon. The sequence of the reverse primer was manipulated to introduce various mutations to substitute the C-terminal amino acid glutamine for arginine, stop codon, threonine, leucine, aspargine or glutamic acid. The P375Q mutation was introduced by a recombinant PCR method using two internal mutant oligonucleotides as described by Saiki et al. (33). All constructs were assessed by automated sequencing to ensure that no random mutations were introduced during PCR. Construction of CD19-2 (A-ins)-luc, P2-luc plasmids was described previously (18). Cell culture and transfections NIH/3T3, a murine fibroblast cell line, was maintained in Dulbecco’s minimal essential medium supplemented with 10% fetal calf serum. Transfections were performed with plasmid DNA coated with the polycationic lipid lipofectamine (Life Technologies) according to the manufacturer’s instructions. For transfections with the pRc–CMV–PAX6 expression vectors, each well in a 12-well plate was transfected with 0.1 µg of P2-luc or CD19-2 (A-ins)-luc, 0.1 µg of pRc-effector plasmid, and 0.02 µg of pSV2 β-gal plasmid (Promega) as an internal control. Luciferase and β-galactosidase assays were performed as described previously (18). Luciferase activities were normalized relative to β-galactosidase activity.

REFERENCES 1. Halder, G., Callaerts, P. and Gehring, W.J. (1995) Induction of ectopic eyes by targeted expression of the eyeless gene in Drosophila. Science, 267, 1788–1792. 2. Halder, G., Callaerts, P. and Gehring, W.J. (1995) New perspectives on eye evolution. Curr. Opin. Genet. Dev., 5, 602–609. 3. Tomarev, S.I. (1997) Pax-6, eyes absent, and Prox 1 in eye development. Int. J. Dev. Biol., 41, 835–842. 4. Matsuo, T., Osumi-Yamashita, N., Noji, S., Ohuchi, H., Koyama, E., Myokai, F., Matsuo, N., Taniguchi, S., Doi, H. and Iseki, S. (1993) A mutation in the Pax-6 gene in rat small eye is associated with impaired migration of midbrain crest cells. Nature Genet., 3, 299–304. 5. Glaser, T., Jepeal, L., Edwards, J.G., Young, S.R., Favor, J. and Maas, R.L. (1994) PAX6 gene dosage effect in a family with congenital cataracts, aniridia, anophthalmia and central nervous system defects. Nature Genet., 7, 463–471. 6. Tang, H.K., Singh, S. and Saunders, G.F. (1998) Dissection of the transactivation function of the transcription factor encoded by the eye developmental gene PAX6. J. Biol. Chem., 273, 7210–7221. 7. Wawersik, S. and Maas, R.L. (2000) Vertebrate eye development as modeled in Drosophila. Hum. Mol. Genet., 9, 917–925. 8. Wilson, D., Sheng, G., Lecuit, T., Dostatni, N. and Desplan, C. (1993) Cooperative dimerization of paired class homeo domains on DNA. Genes Dev., 7, 2120–2134. 9. Chao, L.Y., Huff, V., Strong, L.C. and Saunders, G.F. (2000) Mutation in the PAX6 gene in twenty patients with aniridia. Hum. Mutat., 15, 332–339. 10. Glaser, T., Walton, D.S. and Maas, R.L. (1992) Genomic structure, evolutionary conservation and aniridia mutations in the human PAX6 gene. Nature Genet., 2, 232–239. 11. Azuma, N., Nishina, S., Yanagisawa, H., Okuyama, T. and Yamada, M. (1996) PAX6 missense mutation in isolated foveal hypoplasia. Nature Genet., 13, 141–142. 12. Hanson, I.M., Fletcher, J.M., Jordan, T., Brown, A., Taylor, D., Adams, R.J., Punnett, H.H. and van Heyningen, V. (1994) Mutations at

918

13.

14. 15.

16. 17. 18. 19. 20. 21. 22. 23.


the PAX6 locus are found in heterogeneous anterior segment malformations including Peters’ anomaly. Nature Genet., 6, 168–173. Hanson, I., Churchill, A., Love, J., Axton, R., Moore, T., Clarke, M., Meire, F. and van Heyningen, V. (1999) Missense mutations in the most ancient residues of the PAX6 paired domain underlie a spectrum of human congenital eye malformations. Hum. Mol. Genet., 8, 165–172. Mirzayans, F., Pearce, W.G., MacDonald, I.M. and Walter, M.A. (1995) Mutation of the PAX6 gene in patients with autosomal dominant keratitis. Am. J. Hum. Genet., 57, 539–548. Hill, R.E., Favor, J., Hogan, B.L., Ton, C.C., Saunders, G.F., Hanson, I.M., Prosser, J., Jordan, T., Hastie, N.D. and van Heyningen, V. (1991) Mouse small eye results from mutations in a paired-like homeobox- containing gene. Nature, 354, 522–525. Brown, A., McKie, M., van Heyningen, V. and Prosser, J. (1998) The Human PAX6 Mutation Database. Nucleic Acids Res., 26, 259–264. Prosser, J. and van Heyningen, V. (1998) PAX6 mutations reviewed. Hum. Mutat., 11, 93–108. Singh, S., Stellrecht, C.M., Tang, H.K. and Saunders, G.F. (2000) Modulation of PAX6 homeodomain function by the paired domain. J. Biol. Chem., 275, 17306–17313. Tang, H.K., Chao, L.Y. and Saunders, G.F. (1997) Functional analysis of paired box missense mutations in the PAX6 gene. Hum. Mol. Genet., 6, 381–386. Azuma, N., Yamaguchi, Y., Handa, H., Hayakawa, M., Kanai, A. and Yamada, M. (1999) Missense mutation in the alternative splice region of the PAX6 gene in eye anomalies. Am. J. Hum. Genet., 65, 656–663. Yamaguchi, Y., Sawada, J., Yamada, M., Handa, H. and Azuma, N. (1997) Autoregulation of Pax6 transcriptional activation by two distinct DNA-binding subdomains of the paired domain. Genes Cells, 2, 255–261. Azuma, N. and Yamada, M. (1998) Missense mutation at the C terminus of the PAX6 gene in ocular anterior segment anomalies. Invest. Ophthalmol. Vis. Sci., 39, 828–830. Baum, L., Pang, C.P., Fan, D.S., Poon, P.M., Leung, Y.F., Chua, J.K. and Lam, D.S. (1999) Run-on mutation and three novel nonsense mutations identified in the PAX6 gene in patients with aniridia. Hum. Mutat. (Online), 14, 272–273.

24. Hanson, I., Brown, A. and van Heyningen, V. (1995) A new PAX6 mutation in familial aniridia. J. Med. Genet., 32, 488–489. 25. Singh, S., Tang, H.K., Lee, J.Y. and Saunders, G.F. (1998) Truncation mutations in the transactivation region of PAX6 result in dominantnegative mutants. J. Biol. Chem., 273, 21531–21541. 26. Czerny, T., Halder, G., Kloter, U., Souabni, A., Gehring, W.J. and Busslinger, M. (1999) twin of eyeless, a second Pax-6 gene of Drosophila, acts upstream of eyeless in the control of eye development. Mol. Cell, 3, 297–307. 27. Chow, R.L., Altmann, C.R., Lang, R.A. and Hemmati Brivanlou, A. (1999) Pax6 induces ectopic eyes in a vertebrate. Development, 126, 4213–4222. 28. Jacobson, A. and Peltz, S.W. (1996) Interrelationships of the pathways of mRNA decay and translation in eukaryotic cells. Annu. Rev. Biochem., 65, 693–739. 29. Underhill, D.A., Vogan, K.J. and Gros, P. (1995) Analysis of the mouse Splotch-delayed mutation indicates that the Pax-3 paired domain can influence homeodomain DNA-binding activity. Proc. Natl Acad. Sci. USA, 92, 3692–3696. 30. Underhill, D.A. and Gros, P. (1997) The paired-domain regulates DNA binding by the homeodomain within the intact Pax-3 protein. J. Biol. Chem., 272, 14175–14182. 31. Chalepakis, G., Goulding, M., Read, A., Strachan, T. and Gruss, P. (1994) Molecular basis of splotch and Waardenburg Pax-3 mutations. Proc. Natl Acad. Sci. USA, 91, 3685–3689. 32. Gronskov, K., Rosenberg, T., Sand, A. and Brondum-Nielsen, K. (1999) Mutational analysis of PAX6: 16 novel mutations including 5 missense mutations with a mild aniridia phenotype. Eur. J. Hum. Genet., 7, 274–286. 33. Saiki, R.K., Gelfand, D.H., Stoffel, S., Scharf, S.J., Higuchi, R., Horn, G.T., Mullis, K.B. and Erlich, H.A. (1988) Primer-directed enzymatic amplification of DNA with a therostable DNA polymerase. Science, 239, 487–491. 34. Singh, S. and Saunders, G.F. (1998) A method for obtaining reporter gene activity and nuclear extracts simultaneously from transiently transfected cells. Anal. Biochem., 265, 185–187.