Mutation in a Novel Connexin-like Gene (Gjf1) in theMouse Affects Early Lens Development and Causes aVariable Small-Eye Phenotype

Oliver Puk,1 Jana Loster,1 Claudia Dalke,1 Dian Soewarto,2 Helmut Fuchs,2

Birgit Budde,3,4 Peter Nurnberg,3,4 Eckhard Wolf,5 Martin Hrabe de Angelis,2,6 andJochen Graw1,7

PURPOSE. The purpose of the study was the characterization ofthe novel small-eye mutant Aey12 in the mouse.

METHODS. The eyes of the mutants were described morpholog-ically and histologically and by in situ hybridization.

RESULTS. The homozygotes were viable and fully fertile, whichidentifies Aey12 as a new microphthalmia phenotype in themouse, different from Maf or Pax6 mutants. Histologic analysisindicated the presence of the lens vesicle; however, the pri-mary fiber cells did not elongate properly. Genome-wide link-age analysis mapped the mutation to the proximal region ofchromosome 10 between the markers D10Mit206 andD10Mit189. Among the positional candidate genes, one EST(expressed sequence tag), D230044M03Rik, encodes a con-nexin-like protein. A G3T point mutation was identified atcDNA position 96, resulting in an R32Q amino acid exchangein a transmembrane domain. The mutation leads to a loss of anSsiI restriction site, which is present in five wild-type mousestrains (102, C3H, C57BL/6, DBA, and JF1). The gene is ex-pressed in the posterior part of the lens vesicle, where theprimary fiber elongation starts. In the mutants, the expressionpattern of Pax6, Prox1, Six3, and Crygd are modified, but notthe pattern of Pax2.

CONCLUSIONS. The mutated mouse gene belongs to the family ofconnexin-encoding genes (gene symbols Gja–Gje). Togetherwith its rat and human homologues, it defines a new subgroup,referred to as Gjf1. The mouse mutant described herein offersa new functional candidate gene for microphthalmia-relateddisorders at the corresponding locus on human chromosome

6, area q24. (Invest Ophthalmol Vis Sci. 2008;49:1525–1532)DOI:10.1167/iovs.07-1033

Microphthalmia in humans and small eye in the mouse arefrequent, but genetically diverse phenotypes caused by

severe defects in early eye development (for a recent review,see Graw1). The major players resulting in small-eye pheno-types in the mouse are Mitf (encoding the microphthalmia-associated transcription factor2–5) or Pax6 (paired-box gene66–8). In addition to these major players characterized by ahigh number of affected alleles, the Pitx3, Maf, Sox2, and Foxgenes have been shown to lead to similar phenotypes; how-ever, only a few alleles of these genes have been characterizedup to now (for a recent review, see Graw9).

In humans, severe microphthalmia frequently shows someevidence of familial recurrence but usually no clear Mendeliantransmission pattern.10 Nevertheless, the molecular basis ofseveral congenital diseases, including microphthalmia, havebeen identified. The causative mutations affect PAX2, PAX6,RAX, CHX10, MAF, SHH, or SOX2 (for a recent overview, seeRagge et al.11).

We present a novel dominant mouse model for micro-phthalmia, which was identified in a currently running ENU(N-ethyl-N-nitrosourea) screen with C3H mice.12 In contrast tomice that carry mutations in most of the genes just mentioned,homozygous Aey12 mutants are viable and fully fertile. At thegenetic level, the mutation is characterized by a point mutationin a novel gene that codes for a connexin-like protein. Thephenotype is characterized by an arrest of the outgrowth of theprimary lens fibers. This new mouse model, Aey12, points tothe outgrowth of the primary lens fiber cells as a crucial step inlens and eye development. The phenotype is different fromother mutants that carry connexin-encoding genes.

MATERIAL AND METHODS

Mice

Mice were kept in specific pathogen-free conditions at the NationalResearch Center for Environment and Health (GSF) and monitoredwithin the ENU mouse mutagenesis project.12,13 The use of animalswas in accordance with the German Law on Animal Protection and theARVO Statement for the Use of Animals in Ophthalmic and VisionResearch. Male C3HeB/FeJ mice were treated with ENU (3 � 100mg/kg body weight) at the age of 10 weeks according to Ehling et al.14

and mated to untreated female C3HeB/FeJ mice. The offspring of theENU-treated mice were screened at the age of 3 weeks for abnormal-ities of the eye (Aey). Mice with phenotypic deviation were tested fora dominant mode of inheritance.

Phenotypic Characterization

The eyes of the mutants were examined by slit lamp at the age of 3weeks. Histologic analysis of the eyes was performed during embryo-genesis only. The embryos or their heads were fixed for 3 hours in

From the Institutes of 1Developmental Genetics and 2Experimen-tal Genetics, Helmholtz Center Munich, German Research Center forEnvironmental Health, Neuherberg, Germany; the 3Institute of Molec-ular Genetics, Max-Delbruck-Center for Molecular Medicine, Berlin,Germany; 5Lehrstuhl fur Molekulare Tierzucht und Biotechnologie,Ludwig-Maximilians-Universitat, Munchen, Germany; and the Institutesof 6Experimental Genetics and 7Genetics, Technical University Mu-nich, Center of Life and Food Sciences, Weihenstephan, Germany.

4Present affiliation: Cologne Center for Genomics, University ofCologne, Koln, Germany.

Supported in part by a grant from the German Human GenomeProject (EW, MHdA).

Submitted for publication August 9, 2007; revised September 21and November 14, 2007; accepted February 22, 2008.

Disclosure: O. Puk, None; J. Loster, None; C. Dalke, None; D.Soewarto, None; H. Fuchs, None; B. Budde, None; P. Nurnberg,None; E. Wolf, None; M.H. de Angelis, None; J. Graw, None

The publication costs of this article were defrayed in part by pagecharge payment. This article must therefore be marked “advertise-ment” in accordance with 18 U.S.C. §1734 solely to indicate this fact.

Corresponding author: Jochen Graw, Helmholtz Center Munich,German Research Center for Environmental Health, Institute of Devel-opmental Genetics, Ingolstadter Landstrasse 1, D-85764 Neuherberg,Germany; graw@helmholtz-muenchen.de.

Investigative Ophthalmology & Visual Science, April 2008, Vol. 49, No. 4Copyright © Association for Research in Vision and Ophthalmology 1525

Carnoy’s solution and embedded in JB-4 plastic medium (PolysciencesInc., Eppelheim, Germany), according to the manufacturer’s protocol.Sectioning was performed with an ultramicrotome (OMU3; Reichert-Jung, Walldorf, Germany). Serial transverse 3-�m sections were cutwith a glass knife and stained with methylene blue and basic fuchsin.

In Situ Hybridization

In situ hybridization of sections from embryonic day (E)12.5 wasperformed, essentially as described by Grimm et al.15 Briefly, theembryos were fixed in paraformaldehyde and embedded (Jung Histo-wax; Cambridge Instruments, Nussloch, Germany). Sections (7–10�m) were cut with a microtome (model RM-2065; Leica, Nussloch,Germany) and mounted on slides. The sections were evaluated with alight microscope (Axioplan; Carl Zeiss Meditec, Oberkochen, Ger-many). The images were acquired by means of a scanning camera(Progress 3008; Jenoptik, Jena, Germany) and imported into an image-processing program (Photoshop ver. 6.0, Adobe Illustrator. ver. 9.0;Adobe Systems, Unterschleissheim, Germany).

Mapping

Heterozygous carriers (first generation) were mated to wild-typeC57BL/6J mice, and the offspring (second generation) were back-crossed to wild-type C57BL/6J mice. DNA was prepared from tail tipsof affected offspring of the third generation (G3) according to standardprocedures. DNA was adjusted to a concentration of 50 ng/�L. For agenome-wide linkage analysis of the Aey12 mutation, several microsat-ellite markers were used for each autosome (Graw et al.16).

PCR and Sequencing

For the molecular analysis, RNA was isolated at E12.5 from the headsof the Aey12�/� or wild-type embryos (C3HeB/FeJ). The RNA sampleswere reverse transcribed to cDNA (Ready-to-Go kit; GE Healthcare,Freiburg, Germany), and genomic DNA was isolated from the tail tipsor spleens of wild-type C3HeB/FeJ and C57BL/6J mice or homozygousmutants, according to standard procedures. For RT-PCR amplification,primers were selected from the EMBL-GenBank databases (smart.embl-heidelberg.de/ provided in the public domain by the European Molec-ular Biology Laboratory, Heidelberg, Germany; http://www.ncbi.nlm.nih.gov/Genbank; provided in the public domain by the NationalCenter for Biotechnology Information, Bethesda, MD).

Total RNA was prepared from the analyzed tissues (peqGOLDRNAPure extraction kit; Peqlab, Erlangen, Germany) in accordancewith the protocols supplied by the manufacturer. Total RNA (5 �g)was transcribed to the first strand of cDNA (Superscript II RT kit;Invitrogen, Karlsruhe, Germany). In particular, RNA from the retinawas isolated by placing the eye globe into ice-cold PBS. After thecornea was dissected and the lens removed, the retina was pulled outand immediately frozen on dry ice. RNA was isolated (RNA-Bee Kit;

AMS Biotechnology, Abingdon, UK) and reverse transcribed (Ready-To-Go T-Primed First-Strand Kit; GE Healthcare, Freiburg, Germany).

PCR cycling after the RT step was performed with the primersD230044M03Rik-1 (5� -GCCAGGATAACACCTCCTGA-3�) andD230044M03Rik-2 (5�-TGACTGATAGGAACTATGGTATGG-3�),which were designed to flank both intron regions ofD230044M03Rik to assure specific amplification of the cDNA. Toverify that cDNA synthesis had been successful, we amplified afragment of the housekeeping gene gapdh by use of the forwardgapdh primer 5�-TGAAGGTCGGTGTGAACGGA-3� and the reverseprimer 5�-GATGGCATGGACTGTGGTCA-3�). The products were an-alyzed by electrophoresis on a 1.5% agarose gel.

PCR was performed with a thermocycler (BD-Clontech, Heidel-berg, Germany, or ABI, Weiterstadt, Germany). PCR products wereanalyzed on a 1% agarose gel. Sequencing was performed commercially(SequiServe, Vaterstetten, Germany), after direct purification of thePCR products (Nucleospin Extract II; Machery-Nagel, Duren, Ger-many), after isolation of the DNA from the gel with an extraction kit(Qia-Quick; Qiagen, Hilden, Germany), or after cloning into the pCR-2.1 cloning vector (Invitrogen).

Phylogenetic Analysis

Degrees of similarities among the connexin proteins of Homo sapiens,Mus musculus, and Rattus norvegicus were calculated by an align-ment program (DiAlign; Genomatix, Munich, Germany) based on theunweighted pair–group method with arithmetic mean (UPGMA). Theunrooted phylogenetic tree diagram was plotted by the Drawtreesoftware tool (developed by Joe Felsenstein, Departments of GenomicScience and Biology, University of Washington, Seattle, WA; availableat www.phylodiversity.net/rree/drawtree).

General

Chemicals were from Merck (Darmstadt, Germany) or Sigma-AldrichChemicals (Deisenhofen, Germany). The enzymes used for cloning andreverse transcription were from Roche (Mannheim, Germany).

RESULTS

Phenotype

The Aey12 mutant was detected during an ongoing ENU mu-tagenesis screening because of its dominant small-eye pheno-type; subsequently, it was established as a homozygous line. Inaddition to the small-eye phenotype, slit lamp analysis identi-fied opacities in the cornea and the lens of variable intensity.Figure 1 demonstrates the reduced size of the lenses at the ageof 6 months in hetero- and homozygous mutants comparedwith the wild-type control. Besides the size difference, the

FIGURE 1. Eyes of Aey12 mutants.The lenses of the heterozygousAey12 mutants (middle; age: 6months) were smaller than in thewild-type (left) and showed a poste-rior polar cataract. The lens of thehomozygous mutant was alsosmaller, empty, and fused with otherocular tissues.

1526 Puk et al. IOVS, April 2008, Vol. 49, No. 4

heterozygotes have a polar lens opacity. In the homozygotes,the lens is empty and consists only of an envelope. It is growntogether with several other ocular tissues.

Because the small eyes were obvious at weaning, we inves-tigated eye development in the mutants (homozygous andheterozygous) during embryogenesis (Fig. 2) starting at E10.5.This stage represents the classic placode stage where the lensplacode just begins to invaginate, forming the lens vesicle. Atthis particular stage, no major differences were obvious be-tween the wild-types and the mutants. One day later, the lensvesicle formed and the primary lens fibers started to elongatein the wild-type embryos. Similarly, in the heterozygotes, theposterior part of the lens vesicle seemed to be thicker, but inthe homozygotes no preparation for elongation was obvious.At E12.5, the primary lens fibers of the wild-type elongatedfrom the posterior side of the lens vesicle and finally filled theentire lens vesicle. This process did not take place in homozy-gous mutants; the lens vesicle remained empty. In the het-erozygotes, the primary fibers elongated but did not reach theanterior epithelium. In heterozygous mutants, the lens fiber

elongation was not complete at E15.5 and E17.5. Similar to thesituation in homozygotes, the lens was smaller than in thewild-type. In homozygotes, some further tissue was present inthe small lens, and the cornea was thickened (which led laterto the frequently observed cornea opacity). The retina wasmisfolded because of the smaller size of the lens and waspoorly differentiated. The optic stalk was present.

For a first screening of potentially affected pathways, at E12.5we investigated the expression of genes well-known to be essen-tial for early eye development, including Pax6, Pax2, Prox1,Six3, and Crygd. These results are summarized in Figure 3. Thespatial expression pattern of Pax6 was restricted to the centralarea of the central anterior epithelium of the lens vesicle, whereasPax2 expression in the posterior part of the eye and in thedeveloping optic nerve was not altered. In contrast, transcriptionfactors involved in the regulation of Cryg genes (such as Prox1and Six3) were more strongly expressed over a larger region.Consequently, Crygd was also strongly expressed, but, because ofthe missing lens fibers, it was present in the posterior part of thelens vesicle only.

Aey12-/-C3H Aey12-/+

E12

.5E

15.5

E17

.5

LP

LVPLF

R

OS

PLF

R

R

C

C

R

E11

.5E

10.5

FIGURE 2. Histology of the Aey12mutant eyes. Histologic data are givenfrom E10.5, to which point the eyesdeveloped normally. At E11.5, the lensvesicle was formed, which was usuallyclosed at E12.5 in the wild-type, butnot in Aey12�/� mutants. Aey12�/�

mutants revealed retarded lens fiberelongation and disrupted connectionsbetween lens fibers and lens epithe-lium. At E17.5, the lenses of Aey12�/�

and Aey12�/� remained very small.The retina of Aey12�/� was not welldifferentiated and was misfolded. C,cornea; L, lens; LP, lens placode; LV,lens vesicle; OS, optic stalk; PLF, pri-mary lens fibers; R, retina. Bar, 100 �m

IOVS, April 2008, Vol. 49, No. 4 Novel Small-Eye Mouse Mutant 1527

Mapping of the Aey12 Mutation and CandidateGene Analysis

For linkage analysis, homozygous Aey12 mutants on a C3Heb/FeJ background (G1) were crossed to wild-type C57BL/6J mice;heterozygous mutants (G2) were backcrossed to C57BL/6Jmice. The analysis of 70 G3 offspring revealed a clear linkage ofthe small-eye phenotype to chromosome 10, close to the cen-tromere (between the markers D10Mit123 and D10Mit168)spanning more than 8 Mb. This mapping result excluded othergenes that lead to a similar phenotype (like Maf or Pparbp),because these genes are located on different chromosomes.Because a further fine mapping did not reveal conclusive re-sults, a straightforward positional candidate gene approachwas not possible. Therefore, we made a second linkage analysisand collected, in 230 G3 mice, much more than in the firstapproach. Four animals (0.02%) did not show any linkage to

chromosome 10 and were therefore not considered in furtheranalyses. (These four animals were not homogenous for sex,phenotype, or coat color.)

Based on the 226 G3 mice, the genetic order was calculated(genetic distance � SD in parenthesis): D10Mit206 (0.9 � 0.6cM), Aey12; D10Mit188–D10Mit189 (0.9 � 0.6 cM). Thecritical interval can be calculated to be 4.3 Mb, including 17genes and a pseudogene. Several genes in the critical intervalare predicted only because of the existence of ESTs, but are notyet fully annotated. Among the genes in the critical interval,none is known to be responsible for ocular disorders.

Because the marker D10Mit188 did not show any recombina-tion among the 226 mutant offspring, it indicates a very closelinkage to the Aey12 mutation. Since the EST D230044M03Rikencoding a connexin-like protein is very close to this marker andbecause mutations in the genes Gja3 and Gja8 (coding for con-nexin46 or connexin50, respectively) have been shown fre-quently to cause cataracts in mouse and human (for a review, seeGraw1), we considered this EST to be a good candidate to beinvolved in this particular mutation. Indeed, RT-PCR demon-strated that it is strongly expressed in the embryonic lens, andsequence analysis detected a G3A exchange at position 95 of thecDNA (counting the starting ATG as position 1; Figs. 4a, b).Moreover, it leads to a loss of an SsiI restriction site, which wasconfirmed in the genomic DNA of all tested homozygous Aey12mutants (n � 5); all wild-type strains tested (C3H, C57BL/6, JF1,102, and DBA/2) kept this restriction site, indicating that it is nota polymorphic site (Fig. 4c).

a)

b)

C3H

Aey12-/-

SsiI

D230044M03Rik-Exon 2

PCR product(596 bp)

293 bp 303 bp

- - - - -+ + + + +C3H B6 JF1 102 DBA MM

600 bp

300 bp

596 bp (PCR)

293 bp/303 bp (RS)

- - - - -+ + + + +600 bp

300 bp

596 bp (PCR)

MM

AEY12-/-

#1 #2 #3 #4 #5

c)

FIGURE 4. Sequence analysis and restriction digests. (a, b) Sequenceanalysis detected a G3A mutation in homozygous Aey12 mutants. (c)Restriction digest in five different wild-type strains and five homozy-gous Aey12 mutants. M, marker; �, without digestion; �, digestedwith SsiI; RS, PCR product after cleavage by SsiI.

C3H Aey12-/-

b)a)

OS

R

L

C

k)

i)

g)

e)

c) d)

f)

h)

j)

l)

Histology

Pax6

Pax2

Prox1

Six3

Crygd

FIGURE 3. In situ hybridization of genes known to be involved in eyedevelopment. In situ hybridizations were performed at E12.5 withprobes specific for Pax2, Pax6, Prox1, Six3, and Crygd. It is obviousthat the expression pattern of Pax2 was not altered, but the pattern ofPax6 was restricted to very central areas. In contrast, the expressionareas of Prox1 and Six3 were broader than in the wild-type, and Crygdwas present in the posterior part of the empty lens vesicle. C, cornea;L, lens; OS, optic stalk; R, retina.

1528 Puk et al. IOVS, April 2008, Vol. 49, No. 4

The 95G3A exchange is predicted to cause an alteration ofArg to Gln at amino acid position 32. According to secondarystructure prediction programs, it is suggested that this aminoacid exchange is located within the first transmembrane do-main of the new connexin-like gene. (This new protein hasfour transmembrane domains like other connexins.) Compari-son of the sequence surrounding the Aey12 mutation withother connexins (data not shown) indicated that the ex-changed Arg residue at position 32 is highly conserved amongthe family of connexins and is present in all members de-scribed so far.

Phenotypic Heterogeneity of Aey12 Mutantsin G3 Offspring

Surprisingly, the phenotype of the G3 offspring in both sets ofexperiments was quite heterogenous. In the first mappingapproach, 41 mutants were characterized by small eyes withcataracts, 25 had small eyes with cornea opacities, and 4 hadsmall eyes with small lenses. In the second mapping approach4 years later, the phenotypic heterogeneity was similar. Toexclude the possibility that other genes on the C3H back-ground are responsible for the variation of the phenotype, wetried to establish independent lines from the heterogenous G3offspring. However, this approach failed, and subsequentbreeding of these heterogenous phenotypes resulted again(G4) in a rather homogenous phenotype, mainly with corneaopacities. Therefore, a genetic modifier may act on theC57BL/6 background and interfere somehow with the con-nexin-like gene, most likely with the regulation of its spatialand temporal expression pattern. Additional experiments arecurrently undertaken to characterize this phenotypic hetero-geneity in more detail.

Expression Pattern of D230044M03Rik

To confirm that the new connexin-like gene is expressed in thedeveloping eye, we performed in situ hybridization analysisduring early embryonic stages. As demonstrated in Figure 5a,

D230044M03Rik was expressed at E11.5 in the posterior lensvesicle in a very restricted central area. It fits the regionexactly, where the primary fiber cells start. Therefore, it canbe concluded from this study, that the mutation inD230044M03Rik is very likely to be causative of the pheno-type observed in the Aey12�/� mutants.

Moreover, the expression pattern of D230044M03Rik sug-gests that it is involved in the transition of the cells during theirterminal differentiation process. In the elongating primary fibercells, it is expressed at the anterior tip of the cells, but not atthe posterior end. When the differentiation of the secondaryfibers begins, it is present in the lens epithelial cells and in theelongating epithelial cells at the lens equator (but no longer inthe primary lens fiber cells). Since the identity of the lens cellsin the Aey12 mutant is different, it is not surprising thatD230044M03Rik is expressed mainly in the cells surroundingthe empty lens vesicle. However, in addition to the lens, it isalso present in the developing retina—at weak intensities inthe wild-type, but strongly expressed in the homozygous mu-tants. Since the retina in the homozygous mutants is not welldifferentiated and is also heavily folded, it is tempting to spec-ulate whether D230044M03Rik in the tip of the retina inter-feres with the proper differentiation of the retina. However,further analysis demonstrated that D230044M0Rik mRNA can-not be detected in the retinas of 6-week-old wild-type mice.RT-PCR with cDNA from other organs of these mice demon-strated the specific expression of D230044M03Rik in the lens;no expression was found in brain, lung, heart, liver, or kidney(Fig. 5b).

DISCUSSION

We have described a novel dominant mouse mutant, Aey12,that is characterized by small eyes, homozygous viability,full penetrance, and fertility. The phenotypic features in-clude cornea opacities and small lenses with cataracts andvacuoles. Analysis of eye development during embryogene-

E11.5 E12.5 E15.5

C3H

Aey12

-/-

50 µm 100 µm

100 µm50 µm50 µm

50 µm

LVL L

RRR

OS

CC

M

692 bp

Lens

Bra

in

Lung

Hea

rtLi

ver

Kid

ney

D230044M03Rik

501 bp/489 bp

Ret

ina

GAPDH

FIGURE 5. Expression analysis ofD230044M03Rik in the develop-ing eye. (a) Expression analysis ofD230044M03Rik in the eye at vari-ous stages of embryonic develop-ment. At E11.5, D230044M03Rikwas expressed at the very posteriorpart of the lens vesicle. One daylater, it was present at the tip of theelongating primary fiber cells and atE15.5 in the lens epithelial cells, in-cluding the elongating cells at theequatorial zone. In the homozygousAey12 mutants, its expression wasrestricted to the cells surroundingthe open lens vesicle (E11.5; E12.5).However, at E15.5 it was also stronglyexpressed in the retina, where in thewild-type embryo only a slight ex-pression was found. C, cornea; Llens; LV, lens vesicle; OS, optic stalk;R, retina. (b) Expression analysis ofD230044M03Rik by RT-PCR in lens,retina, brain, lung, heart, liver, andkidney. D230044M03Rik is exclu-sively expressed in the lens. M,marker.

IOVS, April 2008, Vol. 49, No. 4 Novel Small-Eye Mouse Mutant 1529

sis showed a characteristic feature of the Aey12 mutant tobe retarded lens fiber elongation in addition to partly dis-rupted fiber– epithelium connections (heterozygotes) andthe arrest of eye development at the lens vesicle stage(homozygotes). Irregular cell-to-cell appositions betweenepithelial cells and fibers (however, in a weaker form) havebeen described previously in lenses of mice lacking theactivity of the gap junction protein connexin 43.17 Concern-ing the major observations for Aey12�/� (the empty lensvesicle in the homozygous mutants), two similar phenotypeshave been reported previously. First of all, in the Maf knock-out18 and in the Maf mutant mouse Ofl,19 the primary lensfiber cells do not grow out, and therefore the lens vesicleremains empty. The second phenotype similar to our Aey12mutant is caused by the knockout of the Pparbp gene(coding for the peroxisome proliferator activator receptor[PPAR] binding protein20). However, linkage analysis of theAey12 mutation to the proximal part of mouse chromosome10 excludes these genes as candidates, since Gja1 (con-nexin43) is located on a significantly different region ofmouse chromosome 10, whereas Maf and Pparbp are lo-cated on mouse chromosomes 8 and 11, respectively. More-over, another dominant cataract mutation, Cat5, has beenmapped to the same region (5 cM distal to the centro-mere).21 This mutant is not yet characterized.

Fine mapping placed the Aey12 mutation within a 4.3-Mbinterval containing 17 coding genes (some of them are not yetfully characterized in the database). Even if none of the genesin the candidate region has been reported to be involved in eyedevelopment or in (congenital) ocular disorders, we demon-strated in the current study that the EST D230044M03Rik isexpressed in the developing lens and a mutation affecting thefirst (predicted) transmembrane domain of the connexin-likeprotein is very likely to be causative of the mutation. Thephenotypic heterogeneity in the heterozygous Aey12 mutantson the C57BL/6 background could suggest an additional mod-ifier gene that may be involved in the spatial and temporalexpression of D230044M03Rik.

The EST D230044M03Rik was identified as a member ofthe connexin-encoding genes a few years ago (referred to asPC17005) and was characterized to be expressed in theeye22; however, no functional data have been reported up tonow. The gene was found within a representative transcriptand protein set (RTPS; ftp://fantom2.gsc.riken.go.jp/RTPS/);the encoded connexin was designated as Cx23. Because ofits difference from the other connexin-encoding genes, itwas grouped into the � group of divergent connexins andtherefore referred to as Gjd5.23,24 However, to characterizethe relationship of the EST D230044M03Rik to the otherconnexin-encoding genes, we compared the entire family ofthe mouse genes to those from rat and human, where alsoall the sequences are available. The result (Fig. 6) demon-strates clearly that D230044M03Rik is significantly differ-ent from the members of the Gjd gene group. The ESTD230044M04Rik forms together with the also unknowngene LOC684664 (rat) and ENSG00000203733 (human) agroup of its own within the connexin-encoding gene family.Based on this relationship, we propose Gjf1 as the corre-sponding new gene symbol (gap junction protein f1; Gjf1 isthe next one in the alphabetical series: the mouse Gje1 geneencodes Cx2925; see Appendix).

As for the other connexins, the same general feature of fourtransmembrane domains is predicted for the new memberGjf1. In addition, the cytosolic N- and C-terminal parts are alsopresent. In contrast to many other connexins, the C-terminalregion is very short, leading to a calculated molecular mass of23.8 kDa and an isoelectric point at pH 8.7. Moreover, the

protein has only two, rather than three, highly conservedcysteine residues in each of the presumed extracellularloops,23 which also explains the relative large evolutionarydifference from the other Gj genes.

The new connexin-encoding gene Gjf1 is expressed inthe posterior part of the lens vesicle, later in the anteriorpart of the primary lens fibers, and finally in the lens epi-thelium and the lens equator, suggesting that it has animportant function during the process of lens fiber celldifferentiation. It is not expressed in the cornea; therefore,the observed cornea opacities have to be considered to besecondary effects. In the mutant lenses, Gjf1 expression inthe early embryonic stages remains restricted to the poste-rior area of the lens vesicle, but is strongly expressed in theanterior part of the developing retina at later stages as well.However, RT-PCR data in 6-week-old wild-type mice demon-strated clearly that the expression in the retina is onlytransient (Fig. 5). This finding points to a crucial role of Gjf1

Mm-D230044M03Rik

Hs-ENSP00000356620Rn-LOC684664

Rn-Gje1

Mm-Gje1

Hs-Gje1

Rn-Gjd4

Mm-Gjd4

Hs-Gjd4

Rn-Gjc1

Mm-Gjc1Hs-Gjc1

-Rn

-Mm

-Hs

Gja

12

-Mm-H

sG

ja7-R

n-M

nG

ja10-H

s

Cx62

-Hs

Gja10

-Mm-Hs

Gja8

-Rn-Mm-Hs

Gja3

-Rn-Mm-Hs

Gja5

-MmGja6-Rn

-Mm-Hs

Gja1

-Rn-Mm

-HsGja4

-Rn

-Mm-Hs

Gja9

-Rn-Mm

-HsGjb1-Rn

-Mm-Hs

Gjb6-R

n

-Mm-H

s

Gjb2 -Rn

-Mm

-Hs

Gjb

3 -Rn

-Mm

-Hs

Gjb

5

-Rn

-Mm

-Hs

Gjb

4-H

sG

jb7

0.01

FIGURE 6. Phylogenetic tree diagram of connexin proteins. The con-nexin proteins of the mouse were compared with those from the ratand human. Obviously, the D230044M03 gene product and its ortho-logues in rat and human define a new subgroup of the connexin family.Hs: Homo sapiens; Mm: Mus musculus; Rn: Rattus norvegicus. Scalebar, branch length representing 0.01 amino acid substitutions per site.

1530 Puk et al. IOVS, April 2008, Vol. 49, No. 4

during early development and differentiation of the retina,but shows that it is not needed at later stages.

In accordance with the consequences of the loss of the Gjf1function in early lens development, important lens-related tran-scription factors are also changed in their expression patterns.While Pax6 is restricted to a more central area, Prox1 (in thelateral parts of the lens vesicle) and Six3 (in the anterior part ofthe lens vesicle and in the prospective retina) are markedlyenhanced; Crygd (as a marker for lens fiber cells) is expressedin the very posterior part of the lens vesicle. This aspect is ofparticular surprise, since Crygd expression in the wild-typelens at this age is found at the anterior part of the primary lensfibers—the same position in which Gjf1 is expressed in thewild-type.

Besides Gja3 and Gja8, the newly identified Gjf1 gene isthe third connexin-encoding gene that affects early eye andlens development. For Gja3 and Gja8, several knockout andknockin mutations have been described in the mouse. Theseinvestigations led to the concept that Gja3 is important forthe functional integrity of the lens (its knockout leads to adense nuclear cataract in a lens of regular size), whereasGja8 is responsible for the size of the lens (Gja8�/� aresmaller, but have only a mild nuclear cataract; for recentreviews, see Gong et al.26 or Gerido and White27). In addi-tion, in mice, two-point mutations in Gja8 have been de-scribed to be involved in congenital cataracts28,29; there isno further mutation known in the mouse Gja3 gene. Incontrast, several point mutations in both genes are reportedin humans (for a review, see Graw1). To this list Gjf1 mustnow be added, and it would be interesting to see the clinicalphenotype of mutations in the corresponding human gene.Up to now, there is no entry in the OMIM database for anuncharacterized eye disorder at the homologous position onhuman chromosome 6, area q24.

CONCLUSION

We identified a new mouse mutant that has arrested early lensdevelopment. Molecular analysis characterized a causative mu-tation in a novel gene, which is now referred to as Gjf1. Thecorresponding protein obviously has an important role in theinitiation of various steps of lens fiber cell differentiation, butalso in the early development and differentiation of the anteriorretina.

Acknowledgments

The authors thank Erika Burkle, Sabrina Hauser, Michaela List, andMonika Stadler for expert technical assistance; and Utz Linzner (Helm-holtz Center Munich, Institute of Experimental Genetics), and MWGBiotech (Ebersberg, Germany) for providing the oligonucleotides.

APPENDIX

Because of a change in the connexin nomenclature system bythe connexin nomenclature committee, the D230044M03Rikgene has recently been referred to as Gje1. The new nomen-clature was based on that suggested by Sohl and Willecke23 andthe phylogenetic tree from Cruciani and Michalsen.30 As isobvious from the comparison of the old and the new system(see Supplementary Table S1 online at http://www.iovs.org/cgi/content/full/49/4/1525/DC1), two genes have the samesymbol in the old and new system, however, with differentmeanings (Gjc1old is Cx30.2 and Gjc1new is Cx45; Gje1old isCx29 and Gje1new is Cx23). To avoid confusion, at least forGje1, we still prefer Gjf1 as the appropriate gene symbol forD230044M03Rik (Cx23).

References

1. Graw J. The genetic and molecular basis of congenital eye defects.Nat Rev Genet. 2003;4:877–888.

2. Steingrımsson E, Moore KJ, Lamoreux ML, et al. Molecular basis ofmouse microphthalmia (mi) mutations helps explain their devel-opmental and phenotypic consequences. Nat Genet. 1994;8:256–263.

3. Steingrımsson E, Arnheiter H, Hallsson JH, Lamoreux ML, Cope-land NG, Jenkins NA. Interallelic complementation at the mouseMitf locus. Genetics. 2003;163:267–276.

4. Hallson JH, Favor J, Hodgkinson C, et al. Genomic, transcriptionaland mutational analysis of the mouse microphthalmia locus. Ge-netics. 2000;155:291–300.

5. Hansdottir AG, Palsdottir K, Favor J, et al. The novel mouse mi-crophthalmia mutations Mitfmi-enu5 and Mitfmi-bcc2 produce domi-nant negative Mitf proteins. Genomics. 2004;83:932–935.

6. Hill R, Favor J, Hogan BLM, et al. Mouse Small eye results frommutations in a paired-like homeobox-containing gene. Nature.1991;354:522–525.

7. Favor J, Peters H, Hermann T, et al. Molecular characterization ofPax62Neu through Pax610Neu: an extension of the Pax6 allelicseries and the identification of two possible hypomorph alleles inthe mouse Mus musculus. Genetics. 2001;159:1689–1700.

8. Thaung C, West K, Clark BJ, et al. Novel ENU-induced eye muta-tions in the mouse: models for human eye disease. Hum MolGenet. 2002;11:755–767.

9. Graw J, Loster J. Congenital hereditary cataracts. Int J Dev Biol.2004;48:1031–1044.

10. Morrison D, FitzPatrick D, Hanson I, et al. National study ofmicrophthalmia, anophthalmia, and coloboma (MAC) in Scotland:investigation of genetic aetiology. J Med Genet. 2002;39:16–22.

11. Ragge NK, Lorenz B, Schneider A, et al. SOX2 anophthalmia syn-drome. Am J Med Genet. 2005;135:1–7.

12. Hrabe de Angelis M, Flaswinkel H, Fuchs H, et al. Genome-wide,large-scale production of mutant mice by ENU mutagenesis. NatGenet. 2000;25:444–447.

13. Hrabe de Angelis M, Balling R. Large scale ENU screens in themouse: genetics meets genomics. Mutat Res. 1998;400:25–32.

14. Ehling UH, Charles DJ, Favor J, et al. Induction of gene mutationsin mice: the multiple endpoint approach. Mutat Res. 1985;150:393–401.

15. Grimm C, Chatterjee B, Favor J, et al. Aphakia (ak), a mousemutation affecting early eye development: fine mapping, consid-eration of candidate genes, and altered Pax6 and Six3 expressionpattern. Dev Genet. 1998;23:299–316.

16. Graw J, Jung M, Loster J, et al. Mutation in the �A3/A1-crystallinencoding gene Cryba1 causes a dominant cataract in the mouse.Genomics. 1999;62:67–73.

17. Gao Y, Spray DC. Structural changes in lenses of mice lacking thegap junction protein connexin43. Invest Ophthalmol Vis Sci.1998;39:1198–1209.

18. Jamieson RV, Perveen R, Kerr B, et al. Domain disruption andmutation of the bZIP transcription factor, MAF, associated withcataract, ocular anterior segment dysgenesis and coloboma. HumMol Genet. 2002;11:33–42.

19. Lyon MF, Jamieson RV, Perveen R, et al. A dominant mutationwithin the DNA-binding domain of the bZIP transcription factorMaf causes murine cataract and results in selective alteration inDNA binding. Hum Mol Genet. 2003;12:585–594.

20. Crawford SE, Qi C, Misra P, et al. Defects of the heart, eye, andmegakaryocytes in peroxisome proliferator activator receptor-binding protein (PBP) null embryos implicate GATA family oftranscription factors. J Biol Chem. 2002;277:3585–3592.

21. Everett CA, Glenister PH, Taylor DM, Lyon MF, Kratochvilova-Loster J, Favor J. Mapping of six dominant cataract genes in themouse. Genomics. 1994;20:429–434.

22. Gustincich S, Batalov S, Beisel KW, et al. Analysis of the mousetranscriptome for genes involved in the function of the nervoussystem. Genome Res. 2003;13:1395–1401.

23. Sohl G, Willecke K. An update on connexin genes and theirnomenclature in mouse and man. Cell Commun Adhes. 2003;10:173–180.

IOVS, April 2008, Vol. 49, No. 4 Novel Small-Eye Mouse Mutant 1531

24. Sohl G, Willecke K. Gap junctions and the connexin proteinfamily. Cardiovasc Res. 2004;62:228–232.

25. Sohl G, Eiberger J, Jung YT, Kozak CA, Willecke K. The mouse gapjunction gene connexin29 is highly expressed in sciatic nerve andregulated during brain development. Biol Chem. 2001;382:973–978.

26. Gong X, Cheng C, Xia CH. Connexins in lens development andcataractogenesis. J Membr Biol. 2007;218:9–12.

27. Gerido DA, White TW. Connexin disorders of the ear, skin, andlens. Biochim Biophys Acta. 2004;1662:159–170.

28. Steele EC, Lyon MF, Favor J, Guillot PV, Boyd Y, Church RL. Amutation in the connexin50 (Cx50) gene is a candidate for the No2mouse cataract. Curr Eye Res. 1998;17:883–889.

29. Graw J, Loster J, Soewarto D, et al. Characterization of a mutationin the lens-specific MP70 encoding gene of the mouse leading to adominant cataract. Exp Eye Res. 2001;73:867–876.

30. Cruciani V, Mikalsen SO. Evolutionary selection pressure and fam-ily relationships among connexin genes. Biol Chem. 2007;388:253–264.

1532 Puk et al. IOVS, April 2008, Vol. 49, No. 4