Genomic, Transcriptional and Mutational Analysis of the Mouse microphthalmia Locus

Corresponding author: Eiríkur Steingrímsson, Department of Biochemistry and Molecular Biology, School of Medicine, University of Iceland, Vatnsmyrarvegur 16, 101 Reykjavík, Iceland., eirikurs{at}hi.is (E-mail)

Communicating editor: C. KOZAK

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mouse microphthalmia transcription factor (Mitf) mutations affect the development of four cell types: melanocytes, mast cells, osteoclasts, and pigmented epithelial cells of the eye. The mutations are phenotypically diverse and can be arranged in an allelic series. In humans, MITF mutations cause Waardenburg syndrome type 2A (WS2A) and Tietz syndrome, autosomal dominant disorders resulting in deafness and hypopigmentation. Mitf mice thus represent an important model system for the study of human disease. Here we report the complete exon/intron structure of the mouse Mitf gene and show it to be similar to the human gene. We also found that the mouse gene is transcriptionally complex and is capable of generating at least 13 different Mitf isoforms. Some of these isoforms are missing important functional domains of the protein, suggesting that they might play an inhibitory role in Mitf function and signal transduction. In addition, we determined the molecular basis for six microphthalmia mutations. Two of the mutations are reported for the first time here (Mitf^mi-enu198 and Mitf^mi-x39), while the others (Mitf^mi-ws, Mitf^mi-bws, Mitf^mi-ew, and Mitf^mi-di) have been described but the molecular basis for the mutation not determined. When analyzed in terms of the genomic and transcriptional data presented here, it is apparent that these mutations result from RNA processing or transcriptional defects. Interestingly, three of the mutations (Mitf^mi-x39, Mitf^mi-bws, and Mitf^mi-ws) produce proteins that are missing important functional domains of the protein identified in in vitro studies, further confirming a biological role for these domains in the whole animal.

THE development of several cell types is affected by microphthalmia transcription factor (Mitf) mutations. Common to all the mutations are defects in neural-crest-derived melanocytes, manifested by a lack of pigment in the coat, eye, and inner ear. Most mutations also produce mast cell defects in addition to defects in pigmented epithelial cells, which results in abnormal eye development. A few of the mutations also induce osteoclast defects. Approximately half of the mutations are semidominantly inherited and show white spotting and/or coat color dilution when heterozygous with wild type. The remaining mutations are classified as recessive since they only exhibit a coat color phenotype in the homozygous condition or, like the Mitf^mi-sp mutation, only in combination with another mutation at the locus (reviewed by MOORE 1995 ).

Mitf is a member of the Tfe3, Tfeb and Tfec basic-helix-loop-helix-leucine zipper (bHLH-Zip) transcription factor subfamily (HODGKINSON et al. 1993 ; HUGHES et al. 1993 ). Like other Tfe subfamily members, Mitf can bind the canonical CACGTG E-box sequence as either a homodimer or as a heterodimer with one of the Tfe subfamily members (HEMESATH et al. 1994 ). Consistent with its role as a regulator of gene expression, Mitf is primarily located in the nucleus (TAKEBAYASHI et al. 1996 ) where it can activate expression from the pigment-cell-specific tyrosinase (Tyr) and tyrosinase-related protein 1 (Tyrp1) promoters (BENTLEY et al. 1994 ; YASUMOTO et al. 1994 ; YASUMOTO and SHIBAHARA 1997 ).

HEMESATH et al. 1998 have shown that Mitf may be a target of a signal transduction pathway involving mast cell growth factor (Mgf), the receptor tyrosine kinase Kit, and the mitogen-activated protein (MAP) kinase Erk2. Stimulation of Mitf-expressing cells by Mgf results in the Erk2-induced phosphorylation of Mitf on serine 73, leading to increased transcriptional activation potential of the Mitf protein (HEMESATH et al. 1998 ). Although the significance of the serine 73 phosphorylation has not been confirmed in vivo, the well-documented similarity in coat-color phenotypes between Mitf, Mgf (Steel), and Kit (W) mutant mice suggests an in vivo link between the function of these three proteins.

The molecular and biochemical defects associated with many different Mitf alleles have been described (HODGKINSON et al. 1993 ; HUGHES et al. 1993 ; STEINGRIMSSON et al. 1994 , STEINGRIMSSON et al. 1996 , STEINGRIMSSON et al. 1998 ; YAJIMA et al. 1999 ). The semidominant mutations are largely basic region mutations (i.e., DNA binding) or are missing important transcriptional activation domains of the protein (HODGKINSON et al. 1993 ; STEINGRIMSSON et al. 1994 , STEINGRIMSSON et al. 1996 ). These mutant proteins act as dominant-negative proteins by dimerizing with wild-type Mitf, or with one of the Tfe proteins, and inhibiting its DNA-binding or transcriptional activation ability (HEMESATH et al. 1994 ; STEINGRIMSSON et al. 1996 ). The dominant-negative behavior of these mutant proteins is thought to account for the phenotype seen in heterozygous mutant mice. Consistent with this hypothesis, the recessive mutations either have defects in the dimerization domains or cause reduced or aberrant transcriptional activity of the gene (HODGKINSON et al. 1993 ; HUGHES et al. 1993 ; STEINGRIMSSON et al. 1994 ).

Here we describe the genomic structure and transcriptional profile of the mouse Mitf gene as well as the molecular defects associated with six mutations at the locus. These studies have uncovered the existence of a number of previously unrecognized Mitf isoforms and helped to confirm the in vivo importance of functional domains of the protein identified by in vitro experiments. We also show that the Mitf^mi-di mutation and the previously characterized Mitf^mi-ce mutation (ZIMRING et al. 1996 ) are caused by identical single base changes, suggesting a mutation hot spot at Mitf.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mice:
The Mitf^mi-enu198 mutation was recovered in the offspring of a (102 x C3H) F₁ mouse exposed to 250 mg ethylnitrosourea (ENU)/kg body weight. Mitf^mi-x39 was found in the offspring of (102 x C3H) F₁ females exposed to 2 + 2 Gy gamma irradiation. Both mutations were shown to be allelic to Mitf by mating to previously characterized Mitf mutations [Mitf^mi-enu122 (STEINGRIMSSON et al. 1998 ) and Mitf^mi-enu305 (NEUHAUSER-KLAUS et al. 1987 ), respectively]. These mutant mice are maintained at the GSF-Institute of Mammalian Genetics, Neuherberg, Germany. The Mitf^mi-ce (ZIMRING et al. 1996 ) and Mitf^mi-di (WEST et al. 1985 ) mice are maintained at Texas A&M University while the Mitf^mi-ew, Mitf^mi-bws, and Mitf^mi-rw mice are maintained at the Mouse Cancer Genetics Program, Frederick, Maryland.

Northern and RNA RT-PCR analysis:
All RNA RT-PCR and sequence analysis was performed on poly(A)⁺ RNA isolated from heart. Total RNA was prepared from hearts of mutant and wild-type control mice using the RNAzol method (Tel-Test, Inc., Friendswood, TX). RNA was poly(A) selected once using the Pharmacia (Piscataway, NJ) mRNA purification kit. For Northern analysis, the RNA was electrophoresed and transferred to Zetabind membrane (Cuno, Meriden, CT) by standard methods. Hybridization and washes were performed as described (CHURCH and GILBERT 1984 ). For RNA RT-PCR analysis, RNA was reverse transcribed using SuperScript reverse transcriptase (Life Technologies, Bethesda, MD) and amplified by PCR. The resulting PCR products were electrophoresed on agarose gels, purified using GeneClean (BIO 101, Vista, CA), and sequenced using an ABI377 sequencing machine and BigDye chemistry (Perkin Elmer, Foster City, CA) or cloned into the pCR2.1 cloning vector (Invitrogen, San Diego) before sequencing. For each mutation analyzed, the entire coding region was sequenced and, when appropriate, splice regions suspected to carry mutations were amplified from genomic DNA and sequenced using the same technology. The existence of sequence alterations was confirmed by amplifying the affected region and sequencing from several other mutant animals as well as from animals of the genetic background on which the mutation arose.

Primers used for the amplification of suspected splice region mutations were the following. For the Mitf^mi-ew and Mitf^mi-enu198 mutations, primers 5' CCTGTGAAATTGCTGGAATCACC 3' and 5' GGCTCAAGTCTCTGGGATCTGATG 3' were used. For the Mitf^mi-bws mutation, primers 1-40, 5' GCTAGAATACAGTCACTACCAG 3', and 1-44, 5' CAGCAAGCTCAGAGGCACCAG 3', were used. For Mitf^mi-ce, DBA/2N, or Mitf^mi-di, primers 5' TCTGCACAATGTCTCCACCTTATG 3' and 5' TTGGGCAAAGAGCTGCAAGC 3' were used to amplify a 602-bp fragment from each genotype, which was then sequenced directly. Products containing exons 1a and 6a were amplified using primers kid3Arev, 5' CAAAAGTCAACCTCTGAAGAGC 3', and 1-23, 5' GGACAATCACAACTTGATTGAACGAAG 3'. Products containing exons 1h and 6a were amplified using primers 1-26, 5' GATGGAGGCGCTTAGATTTTGAG 3', and 1-28, 5' CGAAGAAGAAGATTTAACATAAAC 3'.

Genomic structure analysis:
To analyze the genomic structure of the Mitf locus, the ~100-kb BAC clone 369C11, which contains most of the Mitf gene (E. STEINGRÍMSSON, unpublished results), was sequenced using primers designed to span through exon-intron boundaries. Since the most 5' exons are not contained within this BAC clone, genome walking was used to generate fragments containing these exon-intron junctions. Briefly, mouse genomic DNA was digested with BstYI and then ligated to adaptors R12/R24, which can be ligated to the BstYI overhang. PCR was then performed using exon-specific primers against a primer specific for the adaptor. The resulting products were sequenced directly. For Southern analysis, genomic DNAs were isolated from the spleens of wild-type and mutant mice and the DNA analyzed by restriction enzyme digestion, gel electrophoresis, and Southern transfer as described (JENKINS et al. 1982 ). The probes used for hybridization are as described by HODGKINSON et al. 1993 and STEINGRIMSSON et al. 1994 . The Mitf^mi-x39 genomic deletion was amplified using primers 144-6, 5' GAGTTCACTGAGAATCCG 3', and x4/16b, 5' GGCACACCACACAAAAGTAACTAC 3', while the Mitf^mi-ws genomic deletion was amplified using primers 1-40, 5' GCTAGAATACAGTCACTACCAG 3' and x4/16b.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mitf genomic structure:
The exon/intron organization of the mouse Mitf gene was determined largely by sequencing across exon/intron junctions of BAC clone 369C11. To determine the structure of the 5' end of the Mitf gene, which is missing from the BAC clone, genomic walking and sequencing of genomic DNA was used (Fig 1A). Exon/intron junction sequences are listed in Table 1. The exon/intron organization of the mouse gene is very similar to the human gene and the exons are numbered accordingly (TASSABEHJI et al. 1994 ). While the intron size between exons 1a and 1h has not yet been determined, the relative positions of these two exons are known from deletion analysis (see below).

View larger version (168K): [in this window] [in a new window] [Download PPT slide]   Figure 1. (A) Genomic structure of the mouse Mitf gene. Exons are indicated as boxes and are numbered according to TASSABEHJI et al. 1994 . Intron sizes are drawn to scale. The size of the first intron, between exons 1a and 1h, is unknown. The deletions associated with the Mitfmi-x39, Mitfmi-ws, and Mitfmi-rw mutations are indicated by lines and proposed ATG translational initiators by arrows. The Mitfmi-x39 mutation results from a genomic deletion that includes exon 4 of Mitf and seems to also affect the splicing of the exon 3, leading to its exclusion in some cases. The Mitfmi-rw mutation deletes exons 1h and 1b. (B) Sequence of exon 9 and immediate downstream genomic region (GenBank accession no. AF222344). The known upstream intron sequence is shown in lowercase letters, the previously characterized cDNA sequence is underlined, and a tentative polyadenylation signal is shaded. (C) Alternative splice products as indicated by RT-PCR and sequencing. The exons are shown as boxes and are numbered as in A. On the left are clones made by primers specific for exon 1a and on the right clones made by primers specific for exon 1h. The numbers on the right of each set of clones represent the total number of clones sequenced vs. the total number of clones that contain the 18-bp alternative exon 6a.

The mouse Mitf gene is large and covers over 50 kb. Exon 9 contains the last coding nucleotides of Mitf. The published Mitf cDNA sequence is <2 kb long while the two major transcripts are 5.5 and 5.7 kb in size, as determined by Northern analysis (HODGKINSON et al. 1993 ). Since the 5' sequences have been fairly well characterized by rapid amplification of cDNA ends (RACE) and S1 nuclease experiments (STEINGRIMSSON et al. 1994 ; AMAE et al. 1998 ), the ~3.5 kb missing from the cDNA are likely to be 3' sequences, possibly encoded by a continuous exon 9, together with a poly(A) tail. Consistent with this notion, the last 935 bp of the known cDNA sequence are continuous and not interrupted by an intron in genomic DNA. The mouse EST clone AV327306, generated by 3'RACE, perfectly overlaps genomic DNA and ends 20 bp downstream from an ATTAAA sequence located 2617 bp from the end of the known cDNA (Fig 1B). This, together with the 2 kb of published Mitf sequence, accounts for most, but not all of the complete length of the Mitf transcripts as detected by Northern analysis. As we have been unable to amplify the 3' sequences of Mitf using RACE experiments, the complete analysis of 3'-untranslated region sequences must await further analysis.

Most of the exon/intron junctions conform to the consensus splice sequence (OHSHIMA and GOTOH 1987 ), except the splice donor of exon 7 where a rare C occurs in the 3'-most position (Table 1). In addition, the splice acceptors of exons 5 and 7 are unusual since a T occupies the 5'-most position, and exon 6 has the infrequent TGA triplet as splice donor.

Mitf^mi-rw deletional analysis positions exon 1h downstream of exon 1a:
Due to the large intronic distances in the 5' end of the Mitf gene, it was not possible to orient exons 1a and 1h relative to each other. However, this was achieved through deletional analysis of the Mitf^mi-rw mutation. Southern analysis has shown that Mitf^mi-rw results from a genomic deletion, which begins ~6 kb upstream of exon 1m and includes exons 1b and 1h (Fig 1A) (STEINGRIMSSON et al. 1994 ). RACE experiments have also shown that Mitf^mi-rw transcripts contain a novel 5' end (STEINGRIMSSON et al. 1994 ). BLAST searches show that this novel 5' end sequence is identical to the recently described alternative 5' exon 1a sequence (AMAE et al. 1998 ) fused directly to exon 2. Because exon 1a is not deleted by the Mitf^mi-rw deletion, it must lie upstream of exon 1h.

Alternative Mitf splice products:
Several alternate Mitf splice products have been described and characterized (STEINGRIMSSON et al. 1994 ; AMAE et al. 1998 ). An alternative splice product that includes an 18-bp alternate exon, here termed exon 6a (Fig 1), has been found in all tissues that express Mitf (STEINGRIMSSON et al. 1994 ; YASUMOTO et al. 1998 ). The incorporation of exon 6a into the Mitf message results from the use of an alternative exon 6 splice acceptor site (STEINGRIMSSON et al. 1994 ). Mice carrying the Mitf^mi-sp mutation are unable to incorporate exon 6a into the Mitf message due to a splice site mutation (STEINGRIMSSON et al. 1994 ). Failure to express exon 6a has little phenotypic effect in heterozygous or homozygous mutant mice: an effect is seen only when the Mitf^mi-sp mutation is in combination with another mutation at the locus. In vitro experiments suggest that exon 6a helps stabilize the Mitf-DNA interaction (HEMESATH et al. 1994 ). Similar observations have been made for the bHLH-Zip protein Max, which also contains an alternatively spliced exon at approximately the same position (BOUSSET et al. 1993 ; KRETZNER et al. 1993 ).

A number of alternatively used 5' exons have also been identified: exon 1m, which is melanocyte specific, and exons 1a and 1h, which are more widely expressed (STEINGRIMSSON et al. 1994 ; AMAE et al. 1998 ). Because Mitf transcripts that contain exons 1a, 1h, or 1m all differ at their 5' ends, and because our genomic sequencing studies have shown that the three exons are widely spaced in genomic DNA (Fig 1A), it is very likely that these transcripts are expressed from different promoters.

RNA RT-PCR amplification, cloning, and sequencing of Mitf cDNAs from heart, using upstream PCR primers from exons 1a or 1h and downstream primers from exon 7, identified several additional rare alternative Mitf splice products (Fig 1C). Among the splice products are clones containing exon 1a or 1h, but lacking exon 1b, showing that this exon is alternatively spliced. We also isolated clones that were missing all of exon 2 as well as clones missing only the last 168 nucleotides of exon 2 (exon 2b, Fig 1C). Since the genomic sequence is not interrupted in this region, the shorter product must be due to the use of an alternative splice donor site. The sequence in this region, ²²¹AG/GTAAA²²⁷ (numbering according to HODGKINSON et al. 1993 ), fits the splice donor consensus (OHSHIMA and GOTOH 1987 ). A few clones were isolated that are missing exons 2, 3, and 4 (Fig 1C) and, intriguingly, we also isolated a clone missing exon 6 (encoding the basic domain) in addition to exons 2, 3, and 4. All the clones, except those missing exon 1b, contain an open Mitf reading frame that can be conceptually translated, suggesting the existence of various alternate Mitf proteins in wild-type tissue. No alternative splice products were identified in the region between exons 7 and 9.

While the biochemical nature of these products has not been characterized, nor their presence in other tissues verified, the existence of these alternative splice products suggests that Mitf proteins are expressed in the cells that lack both the activation domain and serine73 phosphorylation site implicated in mitogen-activated protein kinase activation. Most interestingly, our studies also suggest the existence of a rare Mitf protein missing the basic region, in addition to the activation and serine73 phosphorylation domain. This form of Mitf is reminiscent of Id, a HLH protein lacking a basic domain and acting as a negative regulator of bHLH proteins (BENEZRA et al. 1990 ). This suggests the existence of dominant-negative Mitf proteins in normal tissue. From the mutant analysis (see below) it is clear, however, that the splicing of Mitf is tightly regulated and that a full-length protein is required for full Mitf biological activity.

Most of the alternative Mitf transcripts listed in Fig 1C, regardless of whether they were initiated in exon 1a or 1h, were isolated in two forms: one that contained exon 6a and one that did not. This is in contrast to a previous report by YASUMOTO et al. 1998 , who were unable to amplify any Mitf message using an exon 1a-specific forward primer against an exon 6a-specific reverse primer. Their failure to obtain an amplification product is therefore likely due to the nature of the exon 6a reverse primer used in their experiments.

Two new Mitf alleles:
The importance of Mitf functional domains identified in vitro to Mitf biological activity in vivo was further addressed by determining the molecular basis of six Mitf alleles. The molecular basis of these six mutations has not been previously reported. These results were correlated with the genomic and transcriptional data presented here as well as the phenotypic data available for each mutation. Two of the six Mitf alleles, Mitf^mi-enu198 and Mitf^mi-x39, are recent mutations and have not previously been reported. The phenotypes of these mutations, as well as the other Mitf mutations discussed in this article, are summarized in Table 2. The Mitf^mi-enu198 and Mitf^mi-x39 mutations were recovered in the offspring of mice exposed to ENU or -irradiation, respectively. Heterozygotes for both mutations have slightly reduced retinal pigmentation as measured by slit-lamp biomicroscopy (data not shown). In addition, Mitf^mi-x39 heterozygotes express a white belly spot. In the homozygous condition, the phenotypes also differ: Mitf^mi-enu198 homozygotes have a white coat and minor microphthalmia, while Mitf^mi-x39 homozygotes have a white coat with pigmented patches and near normal-sized eyes (Fig 2 and Table 2). The Mitf^mi-x39 mutation is classified as a semidominant mutation with respect to coat color while Mitf^mi-enu198 is recessive. This suggests that Mitf^mi-x39 encodes a dominant-negative protein, at least in melanocytes.

View larger version (82K): [in this window] [in a new window] [Download PPT slide]   Figure 2. The Mitfmi-enu198 and Mitfmi-x39 mutant phenotypes. Homozygous Mitfmi-x39 (foreground) and Mitfmi-enu198 (background) mice. These mice show only minor evidence of microphthalmia.

RT-PCR experiments generated two different-sized bands from Mitf^mi-x39 RNA. One band was 96 nucleotides shorter than the band amplified from wild-type RNA while the other was 180 nucleotides shorter (Fig 3A). Sequencing studies showed that the larger transcript is missing exon 4, while the shorter transcript is missing exons 3 and 4 of the Mitf message. Both are in-frame deletions and both encode proteins lacking a transactivation domain (Fig 1A) (ROMAN et al. 1991 ). Because Mitf^mi-x39 homozygotes still have pigmented patches in their coats and near normal-sized eyes, it is apparent that this activation domain is not essential for Mitf function, although it is required for full biological activity. Southern analysis and sequencing of PCR-amplified products from Mitf^mi-x39 genomic DNA showed that the mutation is a simple deletion that removes 1408 bp, including exon 4, 705 bp of the upstream intron, and 607 bp of the downstream intron (Fig 1A, data not shown). Although the deletion does not remove exon 3, it apparently affects splicing, resulting in exon 3 skipping in some cases.

View larger version (55K): [in this window] [in a new window] [Download PPT slide]   Figure 3. The molecular defects of Mitf mutations. (A) RT-PCR analysis of Mitf expression in wild-type (C57BL/6J), Mitfmi-ew, Mitfmi-enu198, Mitfmi-bws, and Mitfmi-x39 hearts. The product of the wild-type band (C57BL/6J) is 769 nucleotides, the Mitfmi-ew product is 75 nucleotides shorter, and the Mitfmi-enu198 band comes in both sizes. Mitfmi-bws and Mitfmi-x39 also come in two sizes: Mitfmi-bws in a normal band and a 168-bp shorter band and Mitfmi-x39 in a 96-nucleotide-shorter band and a band that is 180 nucleotides shorter than wild type. (B) The molecular defects of Mitfmi-ew and Mitfmi-enu198 mice. The Mitfmi-ew mutation is located in the splice donor and leads to the apparently complete exclusion of exon 6 from the Mitf message in Mitfmi-ew mutant mice (shaded area). The resulting product is missing the basic domain, the DNA-binding domain of the protein. The Mitfmi-enu198 mutation is an alteration in exon 6 that results in two different products. One product is missing exon 6 and results in a protein that lacks the basic domain (shaded area). The other product contains exons 6a and 6b but results in an Asp207Gly alteration in the basic domain. (C) The Mitfmi-bws mutation. The 3' splice acceptor sequences of exon 2 from control (C57BLKS/J-m+/+Leprdb) and the Mitfmi-bws mutant are compared. The A-to-G transition, 13 nucleotides upstream from the splice site, leads to a partial loss of exon 2b (shaded area) and an mRNA molecule that is 168 nucleotides shorter than the wild-type message.

RT-PCR experiments also generated two different-sized bands from Mitf^mi-enu198 RNA. One band was the same size as the wild-type band while the other was 75 bp shorter (Fig 3A). Sequencing studies showed that the wild-type-sized band contained an A-to-G transition at position 749 of the cDNA. This alteration is located 15 nucleotides upstream from the exon 6/exon 7 splice junction and results in an Asp207Gly substitution in the DNA-binding domain of the protein (Fig 3B). The shorter transcript is lacking the basic domain altogether. No alterations were found in the exon 6/exon7 splice sites in Mitf^mi-enu198 genomic DNA (data not shown), suggesting that the A-to-G transition is responsible for exon skipping. A number of examples of exon skipping caused by mutations in exon sequences have been reported (DEL GATTO et al. 1996 ), providing precedence for such an event.

The effects of the Asp207Gly substitution on Mitf function are difficult to predict. This position is not well conserved among bHLH-Zip proteins, although most proteins have charged residues at this position. Like Mitf, the Tfe3, Tfec, and Tfeb proteins all have aspartate at this position, suggesting a functional importance in this subfamily. In the crystal structures of MAX and upstream stimulatory factor (USF), the corresponding amino acid alanine faces away from the DNA and into solution (FERRE-D'AMARE et al. 1993 , FERRE-D'AMARE et al. 1994 ). However, since these crystal structures are not based on full-length products, it is possible that this position is involved in internal interactions not visible in the crystal.

The Mitf^mi-ew mutation affects splicing of exon 6:
Previous studies have shown that Mitf^mi-ew mice express normal levels of an Mitf message that is missing exon 6, just like the Mitf^mi-enu198 deletion (Fig 3A) (STEINGRIMSSON et al. 1994 ). To determine whether this deletion results from a splicing defect, the relevant splice donor and acceptor sites were amplified and sequenced from Mitf^mi-ew genomic DNA. This analysis revealed an A-to-T transversion four nucleotides downstream from the exon 6b splice donor site (Fig 3B). A comparison of the exon-side of this donor sequence to the consensus splice donor sequence revealed a TGA-triplet, which appears in low frequency in constitutively spliced exons (OHSHIMA and GOTOH 1987 ). This, in addition to the A (76% frequency in constitutively spliced exons) to T (9% frequency in constitutively spliced exons) transversion observed in Mitf^mi-ew DNA, is likely to be sufficient to abolish normal splicing, leading to exon skipping and an mRNA missing exons 6a/6b.

The Mitf^mi-x39 and Mitf^mi-ws deletions overlap:
Southern analysis has shown that the Mitf^mi-ws mutation results from a genomic deletion that removes exons 2, 3, and 4 (HODGKINSON et al. 1993 ) (Fig 1A). Sequencing of PCR products amplified from Mitf^mi-ws genomic DNA shows that this is a 4280-bp deletion that starts in an intron 170 bp upstream of the exon 2 splice acceptor and ends 767 bp downstream from the exon 4 splice donor site (Fig 1A). This deletion removes the serine 73 phosphorylation site present in exon 2b as well as the amphipathic helix located in exon 4 and results in a protein product that is identical to some of the rare alternative splice products expressed in heart (Fig 1C). Like the Mitf^mi-ws mutation, the Mitf^mi-x39 mutation also deletes the amphipathic helix located in exon 4.

Mitf^mi-bws mice also carry an exon 2 deletion:
Northern analysis failed to identify any Mitf expression differences between Mitf^mi-bws and wild-type mice (data not shown). However, RT-PCR amplification showed that Mitf^mi-bws mice express two different-sized Mitf transcripts: One is wild type in size while the other is 168 bp shorter (Fig 3A). Sequencing of the shorter transcript showed that it is missing exon 2b. A transition of G to A, not found in the strain of origin, was identified 12 nucleotides upstream of the exon 2a splice acceptor site (Fig 3C). Upstream mutations that result in partial exon skipping have been reported previously, including in the human autosomal dominant disorder congenital contractural arachnodactyly (MASLEN et al. 1997 ). The phenotype of Mitf^mi-bws mice is mild; heterozygotes are normal while homozygotes have white spots and black eyes (Table 2). This phenotype is milder than that of Mitf^mi-ws mice, which is not surprising given that Mitf^mi-ws mice also lack exon 3 as well as the amphipathic helix located in exon 4. While the phenotype of Mitf^mi-bws mice is mild, the fact that they have a phenotype argues that exon 2, and by inference Ser 73 phosphorylation, is required for Mitf biological activity in vivo.

The Mitf^mi-di mutation is identical to the previously reported Mitf^mi-ce mutation:
The Mitf^mi-di mutation arose in a cross between a PT female and an ENU-treated C3H/HeH male (WEST et al. 1985 ). Heterozygotes appear normal except for slightly reduced choroidal pigmentation, while homozygotes are white with reduced eye size (~60% of wild type) and eye pigment (WEST et al. 1985 ; Table 2). Skeletal abnormalities or atypical osteopetrotic bone in the upper tibia and lower femur were also reported for one of the six mice examined in the original studies.

Northern analysis showed that Mitf^mi-di mice express normal levels of Mitf message (data not shown). However, RT-PCR amplification and sequencing studies showed that this message carries a C-to-T transition in exon 8 at position 916 of the cDNA, which introduces a premature stop codon between the HLH and leucine zipper domains. Surprisingly, this mutation is identical to the mutation reported previously for Mitf^mi-ce (STEINGRIMSSON et al. 1994 ), which arose independently on the DBA/2N strain (ZIMRING et al. 1996 ). To confirm the independent origin of these two mutations, exon 8 and adjacent intron sequences were amplified from genomic DNA of DBA/2N, DBA/2N-Mitf^mi-ce, Mitf^mi-di, and 129/Sv mice. Both DBA/2N-Mitf^mi-ce and Mitf^mi-di genomic DNAs had C-to-T transitions at position 916, confirming the existence of these alterations in genomic DNA. However, these two DNAs differed in sequence at three intron locations. In all cases, the DBA/2N-Mitf^mi-ce DNA was identical to DBA/2N while Mitf^mi-di DNA was identical to 129/Sv (data not shown). Nucleotide 916 may therefore represent a mutational hot spot in the Mitf gene.

The similarity in molecular defect is consistent with the similar phenotypes described for these two mutations. Like Mitf^mi-di homozygotes, Mitf^mi-ce homozygotes are white with reduced eye size and no eye pigment (ZIMRING et al. 1996 ; M. L. LAMOREUX and E. S. RUSSELL, unpublished results) (Table 2). Detailed radiographic analysis has revealed no gross bone malformations in Mitf^mi-ce homozygotes (ZIMRING et al. 1996 ), whereas similar studies on Mitf^mi-di homozygotes suggested mild osteopetrosis in one of six animals studied (WEST et al. 1985 ). While this minor difference in phenotype needs to be confirmed, it may be due to differences in the genetic background of the two mutant strains.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

Mitf mutant mice serve as an important model system for the study of two human deafness and hypopigmentation disorders, WS2A and Tietz syndrome (TASSABEHJI et al. 1994 ; AMIEL et al. 1998 ). Genomic sequencing studies have shown that the human MITF gene is composed of nine coding exons (TASSABEHJI et al. 1994 ), and analysis of a large cohort of WS2A and a family of Tietz syndrome patients has identified many different mutations affecting six of these exons (READ and NEWTON 1997 ). Here we show that the mouse Mitf gene shares a similar exon/intron structure with the human gene, which should facilitate comparative studies between human and mouse. We also provide additional evidence indicating that the mouse gene is transcriptionally complex. The gene is composed of at least three different alternatively used 5' exons and a number of alternatively used internal exons, which together are capable of encoding numerous protein isoforms. The functional importance of these isoforms has not been determined. However, our analysis strongly suggests that splicing of Mitf is tightly regulated.

The genomic structure of Mitf is quite different from that reported for other bHLH-Zip genes. For example, while the bHLH-Zip domain of Mitf is interrupted by three introns, the bHLH-Zip domain of Myc is encoded by a single exon while the bHLH-Zip domains of Max and Usf1 are disrupted by one and two introns, respectively (KAYE et al. 1988 ; VASTRIK et al. 1993 ; HENRION et al. 1996 ). At present, it is not clear whether the bHLH-Zip proteins comprise a natural evolutionary group distinct from other bHLH proteins. ATCHLEY and FITCH 1997 have performed a phylogenetic analysis of 242 different bHLH proteins, including Mitf. According to their analysis, the bHLH-Zip genes do not represent a monophyletic group since they were unable to discriminate the bHLH proteins that possess a leucine zipper from other proteins of the family (ATCHLEY and FITCH 1997 ). The further characterization and phylogenetic comparison of the sequence and genomic structure of all bHLH and bHLH-Zip genes, as they become available, will test whether this idea is true.

While many functional domains of the Mitf protein have been identified by sequencing or in vitro studies, the biological function of a number of these domains has not been verified in vivo. Toward this goal, we have identified the functional domains affected by six mutations at the Mitf locus and then attempted to correlate these results with the phenotype of the mutations as well as the genomic and transcriptional data presented here.

Three of the mutations, Mitf^mi-x39, Mitf^mi-bws, and Mitf^mi-ws, produce protein deletions that reveal important information about the functional domains of the protein. The Mitf^mi-x39 mutation deletes the amphipathic helix located in exon 4, which is thought to be important for transcriptional activation by Mitf. The Mitf^mi-bws mutant protein is missing the domain containing serine 73, which is phosphorylated in response to Mgf-signaling and leads to increased transcriptional activation of the Mitf protein. The Mitf^mi-ws mutation deletes both the amphipathic helix and the serine 73 phosphorylation site. The heterozygous phenotypes of Mitf^mi-x39 and Mitf^mi-ws mice are similar in that both have white belly spots. Both mutations therefore appear to be acting as dominant-negative mutations as would be predicted from the nature of the two mutations. However, Mitf^mi-ws mice also have white tails and toes (MILLER 1963 ), indicating that Mitf^mi-ws is the stronger of the two alleles. This phenotypic difference is also seen in homozygous mutant animals where Mitf^mi-ws is the stronger of the two alleles (Table 2). This difference may result from the loss of the serine 73 phosphorylation site in Mitf^mi-ws mice, which is still expressed in Mitf^mi-x39 mice. Interestingly, although Mitf^mi-x39 and Mitf^mi-ws homozygotes are both missing the transcriptional activation domain encoded by exon 4, eye development is largely unaffected and Mitf^mi-x39 homozygotes even retain some pigmentation in the eye (Table 1 and Fig 2). This suggests that the transcriptional activation domain is not essential for Mitf function in these tissues and raises the possibility that other domains of the protein are able to partially substitute for the loss of this domain in the eye. Alternatively, transcriptional activation may not be the exclusive function of Mitf.

In contrast to the Mitf^mi-x39 and Mitf^mi-ws mutations, Mitf^mi-bws animals do not show a phenotype in the heterozygous condition, indicating that the serine 73 deletion found in Mitf^mi-bws mice does not produce a dominant-negative protein. While Mitf^mi-bws mice have a coat color phenotype in the homozygous condition, consistent with a role for serine 73 phosphorylation in Mitf function in neural crest-derived skin melanocytes, eye development is more or less normal. Therefore, like the activation domain, serine 73 does not seem to be essential for Mitf function in the eye. These results are also consistent with other mutational studies, which suggest that the requirement for Mitf is considerably higher in neural crest-derived melanocytes than in the retinal pigment epithelium (MOORE 1995 ).

Recently, YAJIMA et al. 1999 showed that the Mitf^{mi-black-eyed white} (Mitf^mi-bw) mutation is the result of an L1 element insertion into the intronic sequences located between exons 3 and 4. As a result of this insertion, Mitf transcripts initiating from exon 1m are missing, while the expression of transcripts containing exons 1a and 1h are reduced. This result raised the possibility that this intron contains long-range tissue-specific enhancers that are necessary for Mitf transcription in melanocytes. However, the Mitf^mi-x39 and Mitf^mi-ws mutations both delete this very region of the intron, yet the mice still retain some pigment-producing melanocytes in their coat. These results indicate that these intronic sequences are not essential for Mitf melanocyte expression and raise the possibility that the absence of exon 1m-initiated transcripts results from a long-range effect of L1 element sequences on the exon 1m promoter.

In vitro studies have shown that the Mitf^mi-ew protein cannot bind DNA, yet can dimerize with wild-type Mitf protein or with other family members (HEMESATH et al. 1994 ). This suggests that the Mitf^mi-ew mutation should behave in a dominant-negative fashion and show a phenotype in the heterozygous condition. This is inconsistent, however, with the normal phenotype of heterozygous Mitf^mi-ew mutant animals (Table 2). An explanation for this aberrant behavior has been provided by TAKEBAYASHI et al. 1996 , who showed that the nuclear localization potential of Mitf^mi-ew protein is impaired by the exon 6 deletion. It is interesting to note that Mitf^mi-enu198 animals, which also express the exon 6 deletion, have a heterozygous mutant phenotype (i.e., reduced retinal pigmentation, Table 2). Their homozygous phenotype, however, is less severe than that of Mitf^mi-ew mutant animals. This difference can be explained by the fact that Mitf^mi-enu198 mutant animals also express a mutant form of the protein that contains exon 6, albeit with a Asp207Gly substitution. These results strongly suggest that this mutant form of the protein acts in a dominant-negative fashion in heterozygotes, yet is less severe than the exon 6 deletion in the homozygous condition (i.e., retains some residual biological function).

A major difference between mouse and human MITF mutations is that loss-of-function mutations in the mouse are recessively inherited while loss-of-function mutations in humans are dominantly inherited (TASSABEHJI et al. 1994 ). MITF thus appears to be haploinsufficient in humans but not in mice. This haploinsufficiency also results in phenotypic differences between comparable human MITF mutations and semidominantly inherited mouse mutations. An excellent example is provided by the human mutation that causes Tietz syndrome (uniform dilution of pigmentation and profound deafness in heterozygotes; TASSABEHJI et al. 1995 ) and the semidominant mouse mutation Mitf^mi (HODGKINSON et al. 1993 ). Both mutations are caused by a 3-bp deletion mutation removing one arginine from the basic domain of the protein (HODGKINSON et al. 1993 ; AMIEL et al. 1998 ). Heterozygous Mitf^mi mice show minor white bellyspotting, at least on the C57BL/6J background (HODGKINSON et al. 1993 ). However, unlike with Tietz syndrome patients, their coat is not diluted and they are not deaf. These differences aside, the mouse is still an excellent model system for the study of WS2A and Tietz syndrome. Not only do human and mouse MITF share a similar sequence and genomic organization, they also affect the same primary cell types. Continued genomic, transcriptional, and mutational analysis of the type described here should provide further insights into the pathobiology of these interesting human disorders.

	ACKNOWLEDGMENTS

We thank Joann Dietz, Fran Dorsey, and Steinunn G. Ástrádsdóttir for expert technical assistance and Scott Wilson for help with isolating the BAC clone. We also thank Hans G. Thormar and Jon J. Jonsson for the R12/R24 adaptors. This work was supported by the National Cancer Institute, Department of Health and Human Services, by the Icelandic Research Council, by Adstodarmannasjodur (J.H.H.) and partially by National Institutes of Health grant EY-10223 to M.L.L.

Manuscript received November 3, 1999; Accepted for publication January 12, 2000.

	LITERATURE CITED

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION LITERATURE CITED

AMAE, S., N. FUSE, K. YASUMOTO, S. SATO, and I. YAJIMA et al., 1998  Identification of a novel isoform of microphthalmia-associated transcription factor that is enriched in retinal pigment epithelium. Biochem. Biophys. Res. Commun. 247:710-715[Medline].

AMIEL, J., P. M. WATKIN, M. TASSABEHJI, A. P. READ, and R. M. WINTER, 1998  Mutation of the MITF gene in albinism-deafness syndrome (Tietz syndrome). Clin. Dysmorphol. 7:17-20[Medline].

ATCHLEY, W. R. and W. M. FITCH, 1997  A natural classification of the basic helix-loop-helix class of transcription factors. Proc. Natl. Acad. Sci. USA 94:5172-5176[Abstract/Free Full Text].

BENEZRA, R., R. L. DAVIS, D. LOCKSHON, D. L. TURNER, and H. WEINTRAUB, 1990  The protein Id: a negative regulator of helix-loop-helix DNA binding proteins. Cell 61:49-59[Medline].

BENTLEY, N. J., T. EISEN, and C. R. GODING, 1994  Melanocyte-specific expression of the human tyrosinase promoter: activation by the microphthalmia gene product and role of the initiator. Mol. Cell. Biol. 14:7996-8006[Abstract/Free Full Text].

BOUSSET, K., M. HENRIKSSON, J. M. LÜSCHER-FIRZLAFF, D. W. LITCHFIELD, and B. LÜSCHER, 1993  Identification of casein kinase II phosphorylation sites in Max: effects on DNA-binding kinetics of Max homo- and Myc/Max heterodimers. Oncogene 8:3211-3220[Medline].

CHURCH, G. M. and W. GILBERT, 1984  Genomic sequencing. Proc. Natl. Acad. Sci. USA 81:1991-1995[Abstract/Free Full Text].

DEL GATTO, F., M. C. GESNEL, and R. BREATHNACH, 1996  The exon sequence TAGG can inhibit splicing. Nucleic Acids Res. 24:2017-2021[Abstract/Free Full Text].

FERRÉ-D'AMARÉ, A. R., G. C. PRENDERGAST, E. B. ZIFF, and S. K. BURLEY, 1993  Recognition by Max of its cognate DNA through a dimeric b/HLH/Z domain. Nature 363:38-45[Medline].

FERRÉ-D'AMARÉ, A. R., P. POGNONEC, R. G. ROEDER, and S. K. BURLEY, 1994  Structure and function of the b/HLH/Z domain of USF. EMBO J. 13:180-189[Medline].

HEMESATH, T. J., E. STEINGRÍMSSON, G. MCGILL, M. J. HANSEN, and J. VAUGHT et al., 1994  microphthalmia, a critical factor in melanocyte development, defines a discrete transcription factor family. Genes Dev. 8:2770-2780[Abstract/Free Full Text].

HEMESATH, T. J., E. R. PRICE, C. TAKEMOTO, T. BADALIAN, and D. E. FISHER, 1998  MAP kinase links the transcription factor Microphthalmia to c-Kit signalling in melanocytes. Nature 391:298-301[Medline].

HENRION, A. A., S. VAULONT, M. RAYMONDJEAN, and A. KAHN, 1996  Mouse USF1 gene cloning: comparative organization within the c-myc gene family. Mamm. Genome 7:803-809[Medline].

HODGKINSON, C. A., K. J. MOORE, A. NAKAYAMA, E. STEINGRÍMSSON, and N. G. COPELAND et al., 1993  Mutations at the mouse microphthalmia locus are associated with defects in a gene encoding a novel basic-helix-loop-helix-zipper protein. Cell 74:395-404[Medline].

HUGHES, M. J., J. B. LINGREL, J. M. KRAKOWSKY, and K. P. ANDERSON, 1993  A helix-loop-helix transcription factor-like gene is located at the mi locus. J. Biol. Chem. 268:20687-20690[Abstract/Free Full Text].

JENKINS, N. A., N. G. COPELAND, B. A. TAYLOR, and B. K. LEE, 1982  Organization, distribution and stability of endogenous ecotropic murine leukemia virus DNA sequences in chromosomes of Mus musculus.. J. Virol. 43:26-36[Abstract/Free Full Text].

KAYE, F., J. BATTEY, M. NAU, B. BROOKS, and E. SEIFTER et al., 1988  Structure and expression of the human L-myc gene reveal a complex pattern of alternative mRNA processing. Mol. Cell. Biol. 8:186-195[Abstract/Free Full Text].

KRETZNER, L., E. M. BLACKWOOD, J. MAC and R. N. EISENMANN, 1993 Transcriptional repression by Max proteins p21 and p22, pp. 75–82 in The Negative Regulation of Hematopoiesis, edited by M. GUIGON, F. LEMOINE and N. DAINIAK. Colloque INSERM/John Libbey Eurotext.

MASLEN, C., D. BABCOCK, M. RAGHUMATH, and B. STEINMANN, 1997  A rare branch-point mutation is associated with missplicing of fibrillin-2 in large family with congenital contractural arachnodactyly. Am. J. Hum. Genet. 60:1389-1398[Medline].

MILLER, D. S., 1963  Coat color and behavior mutations in inbred mice under chronic low-level g-irradiation. Radiat. Res. 19:184-185.

MINER, G., 1998  New microphthalmia mutant. Mouse News Lett. 38:25.

MOORE, K. J., 1995  Insight into the microphthalmia gene. Trends Genet. 11:442-448[Medline].

NEUHAUSER-KLAUS, A., A. SCHAFFER, and W. PRETSCH, 1987  Characterization of a newly recovered Mi mutation. Mouse News Letter 78:64.

OHSHIMA, Y. and Y. GOTOH, 1987  Signals for the selection of a splice site in pre-mRNA. Computer analysis of splice junction sequences and like sequences. J. Mol. Biol. 195:247-259[Medline].

READ, A. P. and V. E. NEWTON, 1997  Waardenburg Syndrome. J. Med. Genet. 34:656-665[Abstract].

ROMAN, C., L. COHN, and K. CALAME, 1991  A dominant negative form of transcription activator mTFE3 created by differential splicing. Science 254:94-97[Abstract/Free Full Text].

STEINGRÍMSSON, E., K. J. MOORE, M. L. LAMOREUX, A. R. FERRÉ-D'AMARÉ, and S. K. BURLEY et al., 1994  Molecular basis of mouse microphthalmia (mi) mutations helps explain their developmental and phenotypic consequences. Nat. Genet. 8:256-263[Medline].

STEINGRÍMSSON, E., A. NII, D. E. FISHER, A. R. FERRÉ-D'AMARÉ, and R. J. MCCORMICK et al., 1996  The semidominant Mi-b mutation identifies a role for the HLH domain in DNA binding in addition to its role in protein dimerization. EMBO J. 15:6280-6289[Medline].

STEINGRÍMSSON, E., J. FAVOR, A. F. FERRÉ-D'AMARÉ, N. G. COPELAND, and N. A. JENKINS, 1998  Mitf mi-enu122 is a missense mutation in the HLH dimerization domain. Mamm. Genome 9:250-252[Medline].

TAKEBAYASHI, K., K. CHIDA, I. TSUKUMOTO, E. MORII, and H. MUNAKATA et al., 1996  The recessive phenotype displayed by a dominant negative mirophthalmia-associated transcription factor mutant is a result of impaired nuclear localization potential. Mol. Cell. Biol. 16:1203-1211[Abstract].

TASSABEHJI, M., V. E. NEWTON, and A. P. READ, 1994  Waardenburg syndrome type 2 caused by mutations in the human microphthalmia (MITF) gene. Nat. Genet. 8:251-255[Medline].

TASSABEHJI, M., V. E. NEWTON, X.-Z. LIU, A. BRADY, and D. DONNAI et al., 1995  The mutational spectrum in Waardenburg Syndrome. Hum. Mol. Genet. 4:2131-2137[Abstract/Free Full Text].

VASTRIK, I., P. J. KOSKINEN, R. ALITALO, and T. P. MAKELA, 1993  Alternative mRNA forms and open reading frames of the max gene. Oncogene 8:503-507[Medline].

WEST, J. D., G. FISHER, J. F. LOUTIT, M. J. MARSHALL, and N. W. NISBET et al., 1985  A new allele of microphthalmia induced in the mouse: microphthalmia-defective iris (mi di). Genet. Res. 46:309-324[Medline].

WOOD, B. C. and G. MINER, 1969  Mutant report. Mouse News Lett. 40:32.

YAJIMA, I., S. SATO, T. KIMURA, K. YASUMOTO, and S. SHIBAHARA et al., 1999  An L1 element intronic insertion in the black-eyed white (Mitf^mi-bw) gene: the loss of a single Mitf isoform responsible for the pigmentary defect and inner ear deafness. Hum. Mol. Genet. 8:1431-1441[Abstract/Free Full Text].

YASUMOTO, K. and S. SHIBAHARA, 1997  Molecular cloning of a cDNA encoding a human TFEC isoform, a newly identified transcriptional regulator. Biochim. Biophys. Acta 1353:23-31[Medline].

YASUMOTO, K., K. YOKOYAMA, K. SHIBATA, Y. TOMITA, and S. SHIBAHARA, 1994  Microphthalmia-associated transcription factor as a regulator for melanocyte-specific transcription of the human tyrosinase gene. Mol. Cell. Biol. 14:8058-8070[Abstract/Free Full Text].

YASUMOTO, K., S. AMAE, T. UDONO, N. FUSE, and K. TAKEDA et al., 1998  A big gene linked to small eyes encodes multiple Mitf isoforms: many promoters make light work. Pigment Cell Res. 11:329-336[Medline].

ZIMRING, D. C., M. L. LAMOREUX, N. J. MILLICHAMP, and L. C. SKOW, 1996  Microphthalmia cloudy-eye (mi(ce)): a new murine allele. J. Hered. 87:334-338[Abstract/Free Full Text].

This article has been cited by other articles:

				O. Puk, J. Loster, C. Dalke, D. Soewarto, H. Fuchs, B. Budde, P. Nurnberg, E. Wolf, M. H. de Angelis, and J. Graw Mutation in a Novel Connexin-like Gene (Gjf1) in the Mouse Affects Early Lens Development and Causes a Variable Small-Eye Phenotype Invest. Ophthalmol. Vis. Sci., April 1, 2008; 49(4): 1525 - 1532. [Abstract] [Full Text] [PDF]

				K. Bharti, W. Liu, T. Csermely, S. Bertuzzi, and H. Arnheiter Alternative promoter use in eye development: the complex role and regulation of the transcription factor MITF Development, March 15, 2008; 135(6): 1169 - 1178. [Abstract] [Full Text] [PDF]

				K. Bismuth, S. Skuntz, J. H. Hallsson, E. Pak, A. S. Dutra, E. Steingrimsson, and H. Arnheiter An Unstable Targeted Allele of the Mouse Mitf Gene With a High Somatic and Germline Reversion Rate Genetics, January 1, 2008; 178(1): 259 - 272. [Abstract] [Full Text] [PDF]

				R. Hu, S. M. Sharma, A. Bronisz, R. Srinivasan, U. Sankar, and M. C. Ostrowski Eos, MITF, and PU.1 Recruit Corepressors to Osteoclast-Specific Genes in Committed Myeloid Progenitors Mol. Cell. Biol., June 1, 2007; 27(11): 4018 - 4027. [Abstract] [Full Text] [PDF]

				S. Tshori, A. Sonnenblick, N. Yannay-Cohen, G. Kay, H. Nechushtan, and E. Razin Microphthalmia Transcription Factor Isoforms in Mast Cells and the Heart Mol. Cell. Biol., June 1, 2007; 27(11): 3911 - 3919. [Abstract] [Full Text] [PDF]

				N. Esumi, S. Kachi, P. A. Campochiaro, and D. J. Zack VMD2 Promoter Requires Two Proximal E-box Sites for Its Activity in Vivo and Is Regulated by the MITF-TFE Family J. Biol. Chem., January 19, 2007; 282(3): 1838 - 1850. [Abstract] [Full Text] [PDF]

				A. Sonnenblick, C. Levy, and E. Razin Immunological Trigger of Mast Cells by Monomeric IgE: Effect on Microphthalmia Transcription Factor, STAT3 Network of Interactions J. Immunol., August 1, 2005; 175(3): 1450 - 1455. [Abstract] [Full Text] [PDF]

				A. Sonnenblick, C. Levy, and E. Razin Interplay between MITF, PIAS3, and STAT3 in Mast Cells and Melanocytes Mol. Cell. Biol., December 15, 2004; 24(24): 10584 - 10592. [Abstract] [Full Text] [PDF]

				F. Vetrini, A. Auricchio, J. Du, B. Angeletti, D. E. Fisher, A. Ballabio, and V. Marigo The Microphthalmia Transcription Factor (Mitf) Controls Expression of the Ocular Albinism Type 1 Gene: Link between Melanin Synthesis and Melanosome Biogenesis Mol. Cell. Biol., August 1, 2004; 24(15): 6550 - 6559. [Abstract] [Full Text] [PDF]

J. H. Hallsson, B. S. Haflidadottir, C. Stivers, W. Odenwald, H. Arnheiter, F. Pignoni, and E. Steingrimsson The Basic Helix-Loop-Helix Leucine Zipper Transcription Factor Mitf Is Conserved in Drosophila and Functions in Eye Development Genetics, May 1, 2004; 167(1): 233 - 241. [Abstract] [Full Text]