(Received for publication, July 9, 1996, and in revised form, August 30, 1996)
From the Department of Applied Physiology and
Molecular Biology, Tohoku University School of Medicine, Aoba-ku,
Sendai, Miyagi 980-77 and the § Department of Dermatology,
Akita University School of Medicine, Akita 010, Japan
Tyrosinase, tyrosinase-related protein-1 (TRP-1), and TRP-2 are the enzymes involved in melanin biosynthesis and are preferentially expressed in pigment cells. Their human gene promoters share the 11-base pair M box containing a CATGTG motif, which was shown here to be bound in vitro by microphthalmia-associated transcription factor (MITF). Transient cotransfection analysis showed that MITF overexpression increased the expression of a reporter gene under the control of the human tyrosinase or TRP-1 gene promoter but not the TRP-2 promoter. The promoter activation caused by MITF is dependent on each CATGTG motif of the distal enhancer element, the M box, and the initiator E box of the tyrosinase gene and the TRP-1 M box. Furthermore, a truncated MITF lacking the carboxyl-terminal 125 amino acid residues transactivated the tyrosinase promoter less efficiently than did MITF, suggesting that MITF's carboxyl terminus contains a transcriptional activation domain, but unexpectedly such a truncated MITF remarkably transactivated the TRP-2 gene promoter. These results suggest that MITF is sufficient to direct pigment cell-specific transcription of the tyrosinase and TRP-1 genes but not the TRP-2 gene.
Melanin production is specifically observed in the differentiated melanocytes of neural crest origin and in the retinal pigment epithelium derived from the optic cup of the brain. Tyrosinase (EC 1.14.18.1) is a rate-limiting enzyme in melanin biosynthesis that catalyzes the conversion of tyrosine to 3,4-dihydroxyphenylalanine (DOPA),1 DOPA to DOPA-quinone, and possibly 5,6-dihydroxyindole to indole-5,6-quinone (1). Two other enzymes were recently shown to be involved in melanin production: tyrosinase-related protein-1 (TRP-1) (2, 3) and TRP-2 (4, 5). There is about 40% amino acid identity among human tyrosinase (6, 7, 8), TRP-1 (9, 10), and TRP-2 (11, 12), all of which constitute the tyrosinase family and are specifically expressed in melanin-producing cells. Thus, the tyrosinase family genes provide a useful experimental model for the study of a network of transcription factor interactions involved in cell type-specific gene expression.
The tyrosinase family gene promoters share an M box of 11 bp, containing a CATGTG motif in its center whose motif is a consensus sequence of the binding site for a large family of transcription factors with a basic helix loop helix (bHLH) structure (13, 14). The bHLH structure is required for DNA binding of transcription factors (15). Microphthalmia-associated transcription factor (MITF) is the human homolog of the mouse microphthalmia (mi) gene product, which was recently identified as a novel factor with a bHLH-leucine zipper structure (16, 17, 18). The mi phenotype is associated with the mutant mi locus and is characterized by small nonpigmented eyes, a lack of melanocytes in the skin and inner ear, a deficiency in mast cells, and osteopetrosis caused by dysfunction of osteoclasts. Mi protein (mi gene product) shares the significant amino acid sequence similarity with several transcription factors, such as TFE3, TFEB, and TFEC (19, 20, 21), and was shown to bind in vitro to DNA as a heterodimer formed with TFE3, TFEB, or TFEC (22). Recently, the carboxyl-terminal region of TFE3, sharing a significant amino acid similarity with that of Mi or TFEB, was shown to contain an activation domain (23).
The structural gene coding for MITF is organized into nine exons (24),
although at least two types of exon 1 of the mouse mi gene
have been reported (25). In this context, we have recently identified a
melanocyte-type promoter of the MITF gene, which was able to direct
preferential expression of a reporter gene in melanoma cells (26). The
mutations in the MITF gene were identified in two families affected
with Waardenburg syndrome type 2, which is a dominantly inherited
syndrome of sensorineural hearing loss and patchy abnormal pigmentation
of the eyes, hairs, and skin (24). It is therefore conceivable that
MITF is involved in the differentiation and maintenance of melanocytes.
Recently, we reported that MITF is able to transactivate the human
tyrosinase gene promoter through the tyrosinase distal element (TDE)
containing a CATGTG motif, located 1842 bp upstream from its
transcriptional initiation site (27). TDE is sufficient for
melanocyte-specific expression of a reporter gene linked to either the
homologous or heterologous promoter (27, 28). In contrast,
melanocyte-specific expression of the human TRP-2 gene is directed by
the two separate regulatory regions, the 32-bp element and the proximal
region containing an M box (29). On the other hand, melanocyte-specific promoter function was not detectable within the 3.5-kilobase pair 5-flanking region of the human TRP-1 gene by transient expression assays (29), although the TRP-1 promoter contains an M box (30).
Here we explore the possible involvement of MITF in directing pigment cell-specific transcription of the human tyrosinase family genes and show the functional heterogeneity of the M box shared by these genes. We also show that the 133-bp TRP-1 promoter region, containing an M box, is able to direct preferential expression of a reporter gene in pigment cells. In addition, we suggest that the carboxyl terminus of MITF contains a transcriptional activation domain and also plays an important role in defining the target gene.
All constructs used contain
the firefly luciferase gene as a reporter linked to each 5-flanking
region of the human tyrosinase family genes. The luciferase gene was
derived from pSV2/L, containing the luciferase gene under the control
of SV40 early promoter (31), kindly provided by Dr. S. Subramani
(University of California at San Diego, La Jolla, CA). A promoterless
plasmid pL1 and the tyrosinase-luciferase fusion plasmids pHTL6,
pHTL10, and pHTL12 were constructed as described previously (28). The
constructs pHTRPL16 (29) and pHTRPL4 (30) contain the human TRP-1 gene segments spanning from positions about
3600 to +82 and from positions
839 to +82, respectively. The constructs pHTRPL18, pHTRPL19, and
pHTRPL20 contain the SpeI-Fnu4HI fragment
(positions
133 to +82), the BlnI-Fnu4HI
fragment (positions
484 to +82), and the
BglII-Fnu4HI fragment (
291 to +82), derived
from pHTRPL4, respectively. The TRP-2-luciferase fusion plasmids,
pHDTL1, pHDTL2, pHDTL8, pHDTL10, and pHDTL12, were described previously
(29). pHHOSVL4 (27) contains the MscI-XhoI
fragment (positions
176 to +19) of the human heme oxygenase-1 gene
(32).
The nucleotide substitutions were
introduced into each CATGTG motif of the tyrosinase family gene by
using a TransformerTM site-directed mutagenesis kit (Clontech)
according to the method of Deng and Nickoloff (33). The mutagenic
primers used were the oligonucleotide M01
(5-GATCAGCACA
ATTCCCTTCC-3
, complementary to positions
55 to
34) for the TRP-1 M box (30), M02
(5
-CATTAGCACA
ACCCAAAGTG-3
, complementary to positions
124 to
145) for the TRP-2 M box (29), M03
(5
-AGTCATCACA
ATCTCCACAA-3
, complementary to positions
1844 to
1865) for TDE (27), M04
(5
-TGAAAAGCACA
ACTGATTTTC-3
, complementary to positions
115 to
92) for the tyrosinase proximal element (TPE) (27), and M05
(5
-GTGATTATCACA
TCTTGGCTGAG-3
, complementary to
positions
23 to +2) for the initiator E box (27). The native TG
dinucleotides were converted to the GT dinucleotides (underlined). The
selection primer used for these constructions was a Trans Oligo
NdeI/NcoI (Clontech), and experimental procedures were according to the manufacturer's instructions. The introduced mutations were confirmed by determining the nucleotide sequences. The
constructs, pHTRPL4M and pHDTL12M, thus obtained contain the ACTGTG
motif instead of the CATGTG motif at each M box carried by the original
constructs, pHTRPL4 and pHDTL12. Each CATGTG motif of TDE, TPE, and
initiator E box of pHTL12 was changed to the ACTGTG motif, yielding
pHTL12M, pHTL12M4, and pHTL12M5, respectively.
MeWo human
melanoma cells and HeLa human cervical cancer cells were transfected
with each fusion plasmid and a -galactosidase expression plasmid by
the calcium phosphate-precipitation method as described previously (28,
34). HMV-II human melanoma cells (35, 36) were maintained and
transfected as described previously (27, 34). Luciferase activity was
assayed by using a PicaGene luciferase assay system (Toyo Ink) and
was measured with a Lumat LB9501 (Berthold). The obtained
luciferase activity was normalized with each
-galactosidase activity
that represents an internal control. An expression plasmid for a human
MITF cDNA, pRc/CMV-MITF, was constructed as described previously
(27).
The entire region of MITF was
fused to the carboxyl terminus of glutathione S-transferase
(GST), and the recombinant MITF was produced in Escherichia
coli as described previously (34). GST-MITF fusion protein was
prepared by affinity chromatography with glutathione-sepharose 4B
column (Pharmacia Biotech Inc.). Synthetic TDE was end-labeled with
[-32P]ATP (NEN Research) (specific radioactivity,
2 × 108 cpm/µg) and was used for the gel mobility
shift assay as described previously (34). The oligonucleotides used as
a probe or competitors are shown in the figures (see Figs. 3 and
5).
Construction of Expression Plasmids for an MITF Isoform and a Mutant MITF
An HMV-II human melanoma cDNA library was
screened with an MITF cDNA probe as described previously (27). The
cDNA insert of an isolated clone was subcloned into the pRc/CMV
expression vector, yielding pRc/CMV-MITF. A deletion mutant MITF
cDNA, pRc/CMV-MITF
2, was constructed as follows. The
HindIII-XhoI fragment coding for the
amino-terminal 294 amino acids of MITF was isolated from a subclone
phMI9 containing a full-length MITF cDNA (27) and was inserted
between the HindIII and XhoI sites of pRc/CMV
expression vector (Invitrogen). Both restriction sites are located in
the polylinker sequence. pRc/CMV-MITF
2 thus codes for a mutant MITF lacking the carboxyl-terminal 125 amino acids but instead containing the Cys-Ile dipeptide, encoded by the polylinker sequence.
The constructs, containing each 5-flanking region
of the human tyrosinase family genes, are schematically shown in Fig.
1. There are three cis-regulatory elements in
the human tyrosinase gene, TDE (positions
1861 to
1842), the M box
(positions
104 to
94) located in TPE (positions
112 to
93), and
the initiator E box (CATGTG, positions
12 to
7) (27). The two
regulatory regions of the TRP-2 gene, conferring pigment cell-specific
transcription, are the 32-bp element (positions
447 to
416) and the
proximal region (positions
268 to
56) containing an M box
(positions
138 to
128) (29). The TRP-1 promoter also contains
an M box (positions
48 to
38), although its pigment cell-specific
promoter function was not detectable in our previous work (29, 30).
To investigate the effects of MITF overexpression on the tyrosinase family gene promoters, we performed transient cotransfection assays in MeWo melanoma cells and nonpigmented HeLa cells, in which MITF mRNA expression was not detectable (26). pSV2/L was highly expressed in both cell lines and was included as a control for the cotransfection assay. As evident in Fig. 1, MITF transactivated the TRP-1 gene promoter as well as the tyrosinase promoter, although the magnitude of activation of the TRP-1 gene promoter is smaller than that of the tyrosinase gene promoter. In contrast, the effect of MITF seen on the TRP-2 promoter was similar to the effect on the expression of pL1 or pSV2/L. We were thus unable to detect the transactivation by MITF of the TRP-2 gene promoter in both cell lines.
Functional Analysis of the CATGTG Motifs of the Tyrosinase Gene PromoterTo explore the functional significance of each CATGTG
motif in the pigment cell-specific expression of the human tyrosinase gene, we constructed pHTL12M, pHTL12M4, and pHTL12M5, each containing an ACTGTG motif instead of the CATGTG motif at TDE, the M box, and the
initiator E box, respectively (Fig. 2A).
Transfection experiments demonstrated that the introduced modification
at TDE reduced the relative luciferase activity by about 4-fold in MeWo melanoma cells, and the relative luciferase activity of pHTL12M was
comparable with that of pHTL12M5, containing the substitution at the
initiator E box (Fig. 2A). In contrast, the luciferase activity of pHTL12M4 was slightly decreased (about 82% of the luciferase activity of pHTL12), but this difference is not
statistically significant. Expression levels of pHTL6 or pHTL10,
containing the M box and the initiator E box, are comparable with those
of pHTL12M containing the mutation at TDE. The expression levels of all
the constructs used were similarly lower in HeLa cells. These results
suggest that TDE and initiator E box are required for the efficient
expression of the human tyrosinase gene in pigment cells, and TPE may
have only weak promoter activity. The similar results of the
transfection experiments were also obtained in another human melanoma
cell line, HMV-II (data not shown).
Subsequently, we performed cotransfection assays with an MITF expression plasmid in HeLa cells not expressing the endogenous MITF. The expression levels of any fusion genes were low in HeLa cells (Fig. 2A) but were remarkably increased by MITF overexpression (Fig. 2B). The induction ratio of pHTL12 is about four times higher than that of pHTL12M, pHTL12M4, pHTL6, or pHTL10. The magnitude of transactivation of pHTL12M5 was lowest, suggesting that initiator E box may play an important role in the transactivation by MITF. These results were reproducible in MeWo melanoma cells (data not shown), except that the magnitude of transactivation by MITF was much lower in MeWo cells than in HeLa cells as shown in Fig. 1.
Binding of MITF to the CATGTG Motifs of the Human Tyrosinase GeneOur initial attempt to detect the DNA binding activity of
MITF produced by in vitro translation was unsuccessful (27).
Consequently, in the present study, using GST-MITF fusion protein and
synthetic TDE as a probe, we analyzed the DNA binding properties of
MITF by gel mobility shift assay (Fig. 3). The specific
DNA-protein complexes were detected as at least three bands, suggesting
that the fusion protein may be partially degraded during experimental procedures. The formation of these bands was inhibited by excess amounts of synthetic TDE (Fig. 3, lanes 2-4) but not by the
oligonucleotide TDEM containing an ACTGTG motif (Fig. 3, lanes
5 and 6), indicating that GST-MITF fusion protein
specifically binds to the CATGTG motif of TDE but not to the ACTGTG
motif. The formation of these specific bands was also inhibited by
either synthetic TPE containing the M box or the oligonucleotide Inr-E
containing the initiator E box (Fig. 3, lanes 7 and
8 and lanes 13 and 14, respectively). The oligonucleotide TPEM1 also competed for the formation of the protein-DNA complexes (Fig. 3, lanes 9 and 10).
It should be noted that TPEM1 contains a base change at the 5 end of
the tyrosinase M box and thus TPEM1 contains the sequence identical to
the TRP-2 M box, GGTCATGTGCT (29). In contrast, any competitors
containing the base changes at each CATGTG motif did not affect the
specific protein bindings (Fig. 3, lanes 5 and 6,
lanes 11 and 12, and lanes 15 and
16). These results suggest that MITF is able to bind each
CATGTG motif of TDE, the tyrosinase M box (TPE), the initiator E box,
and the TRP-2 M box.
The apparent activation of the TRP-1
promoter by MITF (see Fig. 1) prompted us to re-evaluate its pigment
cell-specific promoter function. Indeed, in the course of
cotransfection assays (Fig. 1), we noticed that expression levels of
the TRP-1-reporter construct, pHTRPL16, are slightly higher in MeWo
cells than those in HeLa cells. Transfection experiments in MeWo
melanoma cells revealed the similar relative luciferase activity
obtained with either pHTRPL16, pHTRPL4, pHTRPL19, pHTRPL20, or pHTRPL18
(Fig. 4A). These expression levels were
always higher in MeWo cells than those in HeLa cells by about
2-3-fold, but the differences were smaller than those seen for the
tyrosinase or TRP-2 gene by more than 20-fold (see Fig. 2; data not
shown), as reported previously (29). However, the differences in the
expression levels of pHTRPL16 between the two cell lines were not
statistically significant. It is also noteworthy that expression levels
of a construct pHTRPL4M containing a mutated M box were lower than
those of pHTRPL4 by about 5-fold in MeWo cells, but expression of both
constructs was similar in HeLa cells (Fig. 4A). These
results suggest that the segment (positions 133 to +82) containing
the M box, carried by pHTRPL18, is able to direct preferential
expression of the TRP-1 gene in pigment cells.
We then searched for the cis-acting element of the TRP-1 gene, which is responsible for the transactivation by MITF. The relative luciferase activities of any TRP-1 constructs were increased about 3-fold by coexpression with MITF cDNA in MeWo melanoma cells, whereas no activation was detected with pHTRPL4M (Fig. 4B). Because all of the constructs used in this experiment contain an intact M box, except for pHTRPL4M, the M box seems to be responsible for the transactivation of the TRP-1 promoter by MITF.
MITF Does Bind to the TRP-1 M Box but Not to the 32-bp Element of the TRP-2 GeneTo examine whether MITF binds to the regulatory
elements of the TRP-1 or TRP-2 gene, we performed gel mobility shift
assay using the synthetic TDE as a binding probe. As evident in Fig. 5, the binding of GST-MITF fusion protein to TDE was
completely inhibited by the oligonucleotide TRP1E containing the TRP-1
M box (Fig. 5, lanes 2 and 3) but was not by the
32-bp element (447 to
416) containing a CAATTG motif of the TRP-2
gene (Fig. 5, lanes 4-9) (29). These results suggest that
MITF binds the TRP-1 M box but not the 32-bp element of the TRP-2 gene,
indicating the significance of a CATGTG motif for the DNA binding by
MITF.
To confirm the observations that the TRP-2 gene promoter is
not activated by MITF as shown in Fig. 1, we explored the effect of
MITF on the TRP-2 promoter activity using several other constructs (Fig. 6). pHDTL12, containing the 32-bp element and the
proximal region, was preferentially expressed in MeWo cells (29). The expression level of pHDTL12M, containing the base changes at the M box,
was similar to that of pHDTL1 (Fig. 6), suggesting that the TRP-2 M box
is essential for efficient transcription in pigment cells and thus
represents the function of the proximal region (29). On the other hand,
we could not detect a significant increase in the luciferase activities
by the cotransfection with MITF cDNA as compared with the
transactivation seen in the tyrosinase or TRP-1 construct (see Fig. 1).
The relative luciferase activities of pHDTL12M, pHDTL10, pHDTL2, and
pHDTL1 were not noticeably influenced by MITF. The induction ratio
obtained with pHDTL12 was slightly higher than that with pHDTL10, but
the difference was not statistically significant.
Functional Analysis of an MITF Isoform and a Truncated MITF
In the course of cDNA cloning for MITF (27), one type of
cDNA lacking the 84-bp DNA segment representing the exon 3 was isolated (Fig. 7). This type of mRNA may be
generated by alternative splicing and codes for an MITF isoform
containing an internal in-frame deletion of 28 amino acids, termed
MITF. Expression of this mRNA species in melanoma cells was
confirmed by reverse transcription and polymerase chain reaction (data
not shown). In addition, we constructed an expression cDNA coding
for a mutant MITF lacking the carboxyl-terminal 125 amino acids, termed
MITF
2. This deleted region is enriched in serine residues and shares a significant amino acid similarity with the carboxyl-terminal region
of TFE3, which was shown to contain an activation domain (23). The
effects of MITF
or MITF
2 on the tyrosinase and TRP-2 gene
promoters were examined, because both 5
-flanking regions possess the
strong enhancer(s) that directs melanocyte-specific expression on the
reporter gene. MITF
transactivated the tyrosinase promoter as
efficiently as MITF but not the TRP-2 promoter. Thus, no functional
differences were detected between MITF and MITF
. On the other hand,
MITF
2 transactivated the tyrosinase promoter 2-fold, but this
activation level is 2-3-fold less than that obtained with MITF. The
reduced activation of the tyrosinase promoter by MITF
2 suggests that
the carboxyl terminus of MITF may contain an activation domain.
Unexpectedly, MITF
2 transactivated the TRP-2 promoter by about
4-fold, whereas MITF
2 did not transactivate the heme oxygenase-1
gene promoter containing the two cis-acting elements, each
containing an E box motif (37, 38).
The present study indicates that MITF transactivates the
tyrosinase promoter mainly by binding to the CATGTG motifs of TDE and
the initiator E box. It should be noted that even a single copy of TDE
is able to confer the pigment cell-specific expression on a tyrosinase
promoter (27) and on a heterologous SV40 promoter (28) and that TDE
deviates from the M box sequence except for a CATGTG motif (Fig.
8). By introducing base changes into a CATGTG motif, we
have established the significance of this motif of TDE in its enhancer
function (Fig. 2) and in the MITF binding (Fig. 3). Furthermore, we
suggested that the initiator E box rather than the M box plays an
important role in transcription of the tyrosinase gene (Fig. 2),
although MITF overexpression transactivated the promoter activity
through the M box (Figs. 2B and 3). These results suggest
that the affinity of MITF to the M box is lower in vivo than
to TDE and the initiator E box. Other investigators also reported by
using a construct containing the 300-bp human tyrosinase promoter that
the initiator E box is responsible for the transactivation of the human
tyrosinase promoter by Mi protein, a mouse homolog of MITF (39).
Considering all these results, we suggest that MITF is sufficient to
direct the pigment cell-specific transcription of the tyrosinase gene
by interacting with TDE and the initiator E box.
The present study also indicates that the M box, shared by the human tyrosinase family genes, shows the functional heterogeneity as summarized in Fig. 8. For example, in contrast to the tyrosinase M box, each of the TRP-1 and TRP-2 M boxes is essential for the respective basal promoter activity in pigment cells (Figs. 4 and 6) but shows different responses to MITF overexpression. MITF transactivated the TRP-1 gene promoter by binding to the CATGTG motif of the M box (Fig. 4) but not the TRP-2 gene promoter (Fig. 6), even though MITF bound in vitro to the TRP-2 M box (Fig. 3). In this context, it is noteworthy that the sequences flanking each M box are entirely different among the tyrosinase family genes, and even the M box is not completely identical (10 of 11 are identical to each other) (Fig. 8). It is therefore conceivable that each M box or its overlapping sequence may be bound in vivo by certain other factors, which are involved in the basal promoter function for the TRP-1 and TRP-2 genes but not for the tyrosinase gene.
Using a sensitive assay method for luciferase activity that was not
available in our previous studies (29, 30), we could detect the
noticeable promoter function of the TRP-1 gene, directing preferential
expression of a luciferase activity in pigment cells. However, such an
activity is lower than the pigment cell-specific promoter activity of
the tyrosinase or TRP-2 gene by more than 20-fold (Fig. 2 and data not
shown) as reported previously (29). These results suggest that
additional regulatory element(s) equivalent to TDE or the initiator E
box for the tyrosinase gene may be present either in a region further
upstream or somewhere downstream of the human TRP-1 gene, which is
however not represented in the fusion genes used in this study. It is
therefore conceivable that MITF may regulate pigment cell-specific
transcription of the TRP-1 gene by binding to such an unrecognized
element and to the M box. Alternatively, MITF bound to a hitherto
unrecognized element may interact with a certain factor bound to the M
box or its overlapping sequence, leading to pigment cell-specific
expression of the TRP-1 gene. In this context, only 133-bp 5-flanking
region of the mouse TRP-1 gene is sufficient for pigment cell-specific
promoter activity, and its M box may be responsible for the
transactivation by MITF (34). A repressor sequence, MSE, has been
reported in the mouse TRP-1 gene (40), but the sequence homologous to
MSE is not present within the 839-bp 5
-flanking region of the human
TRP-1 gene.
The pigment cell-specific transcription of the TRP-2 promoter is directed by the two separate regulatory elements, the 32-bp element and the proximal region containing an M box (29). Here we show that the M box is essential for efficient expression of the TRP-2 promoter in pigment cells (Fig. 6), but this M box by itself is not sufficient for the transactivation by MITF. The presence of the 32-bp element containing a CAATTG motif in addition to the TRP-2 M box caused no significant transactivation by MITF (Fig. 6), which is consistent with the results that the 32-bp element was not bound by MITF (Fig. 5). These results suggest that MITF is not indispensable for the melanocyte-specific transcription of the TRP-2 gene, and thus the mechanism for melanocyte-specific transcription of the TRP-2 gene is different from that of the tyrosinase gene. Such a difference may reflect a character of TRP-2, whose expression is detected in melanoblasts of developing mouse embryo and precedes the expression of the tyrosinase and TRP-1 genes (41).
We also identified an MITF isoform lacking the region encoded by exon
3, whose function was, however, indistinguishable from that of MITF.
Thus, the deleted region of 28 amino acid residues exerted no
noticeable effects on the function of MITF. On the other hand,
MITF2, which lacks the carboxyl-terminal 125 residues of MITF,
showed remarkable functional differences compared with MITF. As
expected from the functional analysis of TFE3 (23), the activation
levels of the tyrosinase promoter by MITF
2 were noticeably reduced,
suggesting that MITF may contain an activation domain in its
carboxyl-terminal region. On the other hand, MITF
2 unexpectedly
transactivated the TRP-2 gene promoter,but not the heme oxygenase-1 and
SV40 early promoters. Aberrant activation of the TRP-2 promoter by
MITF
2 suggests that the carboxyl-terminal portion of MITF may be
involved in the specific MITF-protein interactions, which in turn
define the binding sequence in the tyrosinase family genes. It should
be noted that the heme oxygenase-1 gene promoter, carried by pHHOSVL4,
contains the CATATG and CACGTG motifs, each of which is bound by USF
and other factors (37, 38). Furthermore, expression of pHDTL12M
containing the mutated M box was not increased by MITF
2 (data not
shown). These results suggest that MITF
2 still retains the ability
to recognize the CATGTG motif. The carboxyl terminus deleted in
MITF
2 is encoded by the 5
portion of exon 9, which was shown to
represent the 3
end of the MITF gene (24). Thus, it remains to be
investigated whether a naturally occurring MITF isoform, functionally
equivalent to MITF
2, is present in pigment cells. Alternatively, it
is a hitherto unknown factor rather than an MITF isoform that may
function as a transcription factor for the TRP-2 gene, which seems to
be consistent in part with the unique expression patterns of the TRP-2
gene among the tyrosinase family genes (41).
In addition to transcription factors, such as TFE3, TFEB, and TFEC
(22), the retinoblastoma protein was recently shown to interact with Mi
protein in vitro (42). The retinoblastoma protein has been
characterized as a regulator for gene transcription or cell cycles
(43). Thus, MITF may play an important role in cell differentiation and
maintenance of not only melanocytes but also certain cell types of
other lineage, as evident from the phenotype of mi mice.
These observations together with a function of MITF2 are of
particular interest in view of the pathogenesis of a dominantly inherited Waardenburg syndrome type 2. It is conceivable that certain
mutations in or near the MITF's carboxyl-terminal portion may lead to
produce a mutant MITF with an aberrant function.
The regulation of the tyrosinase family gene expression will provide a model system to study a regulatory network of multiple transcription factors involved in the cell-specific gene transcription. The present study will facilitate the research on searching for the factors involved in the regulation of pigment cell-specific transcription of each human tyrosinase family gene as well as on the pathogenesis of Waardenburg syndrome type 2.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBankTM/EMBL Data Bank with accession number(s) D10243[GenBank].
We thank S. Subramani for providing pSV2/L.