Musashi
(Msi), the Drosophila neural RNA-binding protein, plays a part in eye development. Msi expression is observed in
the nuclei of all photoreceptor cells (R1-R8). Although a msi
loss-of-function mutation results in only weak abnormalities in
photoreceptor differentiation, the msi eye phenotype is
significantly enhanced in a seven in absentia (sina) background.
sina is known to be involved in the degradation of the Tramtrack
(Ttk) protein, leading to the specification of the R7 fate. Msi also functions
to regulate Ttk expression. The sina msi mutants show significantly
high ectopic expression of Ttk69 and failure in the determination of the R1,
R6, and R7 fates. Other photoreceptor cells also fail to differentiate, with
abnormalities occurring late in the differentiation process. These results
suggest that Msi and Sina function redundantly to downregulate Ttk in
developing photoreceptor cells (Hirota, 1999).
To examine the pattern of Msi expression in eye development, a
monoclonal antibody was generated against the Msi protein. The N-terminal
region of Msi (210 amino acids) was used as the antigen for immunization.
The antibody 3A5 recognizes three protein species with relative molecular
masses of 60-70 kDa in immunoblots of protein extracts from wild-type eye
discs. These sizes are compatible with the molecular weight predicted from
the cDNA sequence (63 kDa, 606 amino acids). Msi may receive three
different types of modification, resulting in the three bands observed. In the
immunoblot analysis, the intensities of these bands were reduced in
msi1/+ extracts and undetectable in msi1/msi1extracts.
Immunohistochemistry with this antibody does not detect any signals in
msi1/msi1 discs. Taken together, these results suggest that
the antibody 3A5 recognizes the Msi protein specifically. To determine the
types of cells expressing Msi in developing eye discs, double staining with
antibodies against Msi and a neuronal marker, Elav (Robinow and White,
1991), was performed. In eye imaginal discs from third-instar larvae, a low
level of Msi expression is observed in cells forming a stripe immediately
anterior to the morphogenetic furrow (MF); a more intense signal is also observed in the
developing ommatidia posterior to the MF. Msi expression is restricted to
nuclei in cells posterior to the MF, although it appears not to be restricted to
nuclei anterior to it. Compared with Elav staining, some cells anterior and
just posterior to the MF are immunopositive for Msi, but negative for Elav.
To compare Msi and Elav expression in detail, discs were stained with both
Msi and Elav antibodies using HRP labeled secondary antibodies. The
ommatidial clusters located 4 to 6 rows posterior to the MF contain three
Elav-positive neurons: R8, R2, and R5. These three cells are also positive for
Msi, and Msi expression is also detectable in the presumptive R3 and R4 cells
prior to their Elav expression. At 7 to 10 rows posterior to the MF, the
ommatidial clusters contain five Elav-positive neurons: R8, R2, R5, R3, and
R4. In addition to these five cells, three Msi-positive cells are located basally
adjacent to R2, R8, and R5; these locations correspond to the presumptive R1, R6, and
R7 cells. This result suggests that expression of Msi begins earlier than that
of Elav during the neuronal differentiation of photoreceptor cells. Staining of
pupal retinas 40 h after puparium formation (40 h APF) with the anti-Msi
antibody shows that all photoreceptor cells continue to express Msi protein
during pupal eye development. Msi is also expressed in the photoreceptor
cells of adult flies (Hirota, 1999).
To investigate the functions of Msi during eye development, the eye
phenotype of the msi1/msi1mutants was examined.
msi1/msi1eyes contain abnormal ommatidia, with deformed
rhabdomeres and/or irregular orientation, at a low frequency (3.05%).
Staining developing msi1/msi1-eye discs with antibodies
against several neuronal markers reveals that the number of photoreceptor
cells is not affected, suggesting that msi is involved in late processes
of photoreceptor cells differentiation, including the formation of
rhabdomeres. However, the penetrance of the
msi1/msi1phenotype is so low that it was difficult to
investigate the Msi function. Also examined were the genetic interactions of msi with
ttk, a possible target gene of Msi, and sina, a factor involved
in the degradation of Ttk. Strong genetic interactions occur
between msi and sina mutations in eye development. In
wild-type flies, the compound eye has a regular array of ommatidia, each of
which contains eight photoreceptor cells (R1-R8). At the R7 level, the
rhabdomeres of seven photoreceptor cells (R1-R7) are arranged in a
characteristic asymmetrical trapezoid. Since Sina is essential for the
differentiation of R7, R7 is missing in 90% of the ommatidia of
sina2/sina3 mutants (in which little if any functional gene products
are produced) and the external morphology of the eyes shows a slight
roughness. Notably, the eye phenotype of the double homozygous mutants
of msi and sina (sina2msi1/sina3msi1) show
synergistic, but not additive, enhancement. The external morphology of the
double homozygous mutants shows strikingly disturbed ommatidial arrays.
Most of the rhabdomeres are severely deformed and no ommatidia contain
more than five rhabdomeres, suggesting that msi and sina
have important and distinct roles in the differentiation of photoreceptor
cells. To confirm the function of Msi, rescue experiments of sina msi
double mutants by Msi were performed. A genomic region containing the
wild-type msi gene (referred to as P[msi1+]) was introduced
into the sina msi mutant flies by P element-mediated germline
transformation. The P[msi1+] fragment significantly rescues the
defects in the sina msi eyes. The external morphology of the eye and
the formation of rhabdomeres are nearly normal. This result indicates that
the sina msi eye phenotype is partially due to the loss of msi
function. Since Msi contains two RNA recognition motifs (RRMs), RRM-A and
RRM-B, the RNA-binding activity of Msi is likely to be essential for its
rescuing activity. To test this possibility, a transgenic rescue experiment was
performed in which mutant Msi proteins whose RNA-binding activities were
designed to have been abolished were expressed. The RRM domains contain a
consensus sequence that is composed of two highly conserved short
segments, referred to as RNP1 (ribonucleoprotein octamer consensus) and
RNP2. Since the aromatic side chains of RNP1 are known to be crucial for
RNA binding, mutations that change phenylalanine to alanine in three places
in the RNP1 were induced into both of the two RRMs of Msi (referred to as
P[msiA*B*]). The P[msiA*B*] fragment does not rescue the
sina msi eye phenotype. Additionally, P[msi1+] fully rescues
the weak defects in the msi1eyes described above, while
P[msiA*B*] had no effect. Taken together, it is concluded that the
RNA-binding ability of Msi is involved in normal eye development (Hirota,
1999).
To examine how the sina and msi mutations causes the
severe eye defects described above, the neuronal differentiation of this
double mutant was examined by staining with anti-Elav antibody. All the
ommatidia posterior to the eighth row from the MF contain eight
photoreceptor cells in wild-type eye discs. In the same region, 90% of the
ommatidia of sina2/sina3 lack R7. In the same region, all the
ommatidia of sina2msi1/sina3msi1contain only five
cells, consistent with the appearance of the phenotype in adult eyes. In the
sina2msi1/sina3msi1eye discs, a nearly normal
pattern of Elav staining was observed in developing ommatidia up to the
five-cell precluster stage, which is composed of R2, R3, R4, R5, and R8.
Subsequently, however, R1, R6, and R7 are never added to the ommatidia.
These results suggest that Msi and Sina have redundant functions in the cell
fate determination of R1, R6, and R7. Furthermore, in the posterior region of
the eye discs, the spatial arrangement of the five cells in each ommatidium
changes and overlaps abnormally. In the most posterior region of the eye
discs, many ommatidia with reduced numbers of Elav-positive cells are
observed. These results indicate that R2, R3, R4, R5, and R8, which had once
expressed Elav, gradually have their spatial arrangement in the ommatidia
disrupted, and that some of the photoreceptor cells fail to maintain Elav
expression, resulting in a strikingly deformed eye (Hirota, 1999).
To confirm the requirement of Sina and Msi for the cell fate
determination of R1 and R6, the sina msi eye discs were stained with
antibody against Bar, an R1/R6-specific marker. Bar is expressed in R1 and
R6 in wild-type eye discs. In 40% of the ommatidia of sina2/sina3 eye
discs, the number of Bar-positive cells is reduced to one or zero,
suggesting that Sina has some roles in the cell fate determination of R1 and
R6, where Sina is known to be expressed. Consistent with anti-Elav staining,
Bar-positive cells completely disappear in the
sina2msi1/sina3msi1eye discs. In combination with
sina1, a hypomorphic allele, instead of sina2, one or two
Bar-positive cells per ommatidium are detected in 50% of the
sina2msi1/sina3msi1ommatidia, confirming that Msi
and Sina have redundant functions required for the cell fate determination
of R1 and R6 (Hirota, 1999).
Sina has been suggested to be involved in the degradation of Ttk, a
general inhibitor of neuronal differentiation. Since sina msi double
mutants show defects in neuronal differentiation, the possibility that the
expression pattern of Ttk69 is different in these animals was tested by
examining the Ttk expression pattern in eye discs stained with anti-Ttk69
antibody. In wild-type eye discs, Ttk69 is detected in four cone cells per
ommatidium. The expression pattern of Ttk69 in the
msi1/msi1eye discs is indistinguishable from that of
wild-type. In sina2/sina3 eye discs, five cells were labeled in 5% of
the ommatidia, suggesting that degradation of Ttk69 is reduced. Notably,
50% of the ommatidia in the sina2msi1/sina3msi1eye
discs contain additional Ttk69-expressing cells, indicating that the average
number of cells per ommatidium that express Ttk69 ectopically is larger in
sina2msi1/sina3msi1 than in sina2/sina3 eye discs.
Since both photoreceptor and cone cells are deformed at later stages of
development the cell types ectopically expressing Ttk69 could not be
identified. This result suggests that the Ttk69 expression is negatively
regulated by both Msi and Sina. Consistent with this idea, the morphology of
sina msi double mutants is significantly recovered in a heterozygous
background for ttkosn (sina2msi1ttk osn/sina3msi1), a
mutation that disrupts expression of both the Ttk69 and Ttk88 proteins.
The external morphology of the eye and the formation of the rhabdomeres
are nearly normal. Thirty percent of the
sina2msi1ttkosn/sina3msi1ommatidia show the normal
number and arrangement of photoreceptor cells and are indistinguishable
from wild-type ommatidia. These results suggest that the severe eye
phenotype of the sina2msi1/sina3msi1 double mutant may result
from an elevation in Ttk expression levels (Hirota, 1999).
These results demonstrate that Msi and Sina redundantly function as
factors required for the downregulation of Ttk69; however, the mechanism
remains unknown. A recent study, exploring the target RNA for Msi,
indicates that TTK69 mRNA contains multiple sites that could
potentially be recognized by Msi. Therefore, it is likely that Msi binds to the
TTK69 mRNA to inhibit its translation or reduce its stability. In
contrast, Sina has been shown to function with Phyl to target Ttk protein for
degradation. Thus, Msi and Sina are likely to function to down-regulate Ttk
at different levels in independent manners. If one functions via the other, the
phenotype of the sina msi double null mutants would be identical to
the phenotype of single null mutants for the gene that functions downstream
of the other. Instead, the sina msi mutants show a synergistically
enhanced eye phenotype that is much more severe than that of the single
mutants. Furthermore, the finding that a half-reduction in the gene dosage
of ttk suppresses the sina msi double null mutants indicates
that ttk is downstream of msi and sina. Taken
together, the expression of Ttk is likely to be regulated posttranscriptionally
by factors including Msi, and posttranslationally by Sina, Phyl, and Ebi (a WD
repeats protein). The negative regulation of Ttk by Msi and Sina is required
for both the early processes of R1, R6, and R7 differentiation and the late
processes of the differentiation of other photoreceptor cells in eye
development. Further studies will extend knowledge of how the
posttranscriptional regulation of gene expression by Msi controls ommatidial
development (Hirota, 1999).
Asymmetric cell division is a fundamental strategy for generating cellular diversity during animal development. Daughter cells manifest asymmetry in their differential gene expression. Transcriptional regulation of this process has been the focus of many studies, whereas cell-type-specific 'translational' regulation has been considered to have a more minor role. During sensory organ development in Drosophila, Notch signaling directs the asymmetry between neuronal and non-neuronal lineages, and a zinc-finger transcriptional repressor Tramtrack69 (Ttk69) acts downstream of Notch as a determinant of non-neuronal identity. Repression of Ttk69 protein expression in the neuronal lineage occurs translationally rather than transcriptionally. This translational repression is achieved by a direct interaction between cis-acting sequences in the 3' untranslated region of ttk69 messenger RNA and its trans-acting repressor, the RNA-binding protein Musashi (Msi). Although msi can act downstream of Notch, Notch signaling does not affect Msi expression. Thus, Notch signaling is likely to regulate Msi activity rather than its expression. These results define cell-type-specific translational control of ttk69 by Msi as a downstream event of Notch signaling in asymmetric cell division (Okabe, 2001).
Mechanosensory bristle development in Drosophila is an excellent model system in which to address the molecular mechanisms of asymmetric cell division. Four successive asymmetric cell divisions from a common precursor cell, called the sensory organ precursor (SOP) generate a sensory bristle comprising four different non-neuronal support cells and one neuron. The first asymmetric cell division of SOP into IIa non-neuronal and IIb neuronal precursors is regulated by Notch signaling; specific activation of Notch occurs in IIa owing to inhibition in IIb by Numb, an intracellular negative regulator of Notch. Activation of Notch signaling results in the appearance of Ttk69 protein in the IIa precursor, but not in IIb. The expression pattern of Ttk69, phenotypes of ttk69 loss-of-function mutants, and the overexpression of Ttk69 suggest that Ttk69 is necessary and sufficient to specify the IIa non-neuronal lineage; however, the mechanism through which Notch signaling regulates Ttk69 expression has remained elusive (Okabe, 2001).
Although Ttk69 protein is not detected in the IIb neuronal precursor in which Notch signaling is inactive, similar levels of ttk69 mRNA are present in the IIb precursor and its non-neuronal sibling. Using a lacZ reporter in a ttk enhancer trap line, it has been confirmed that there are indistinguishable levels of expression between IIa and IIb. Selective repression of Ttk69 protein expression in the IIb neuronal precursor thus must occur post-transcriptionally, either by post-translational degradation or translational repression (Okabe, 2001).
Ttk69 protein also regulates neural development in developing adult photoreceptor cells, where Ttk69 expression is controlled by selective degradation in neuronal cells dependent on seven in absentia (sina) and phyllopod functions. These two genes are also involved in the cell-fate decision of mechanosensory bristle lineage; loss of either gene function causes a 'double-bristle phenotype', suggesting that the IIb precursor is transformed into IIa. However, as the penetrance of the sina-null mutant phenotype (12.1% transformation of IIb precursor into IIa) is much less than that of the phenotype of overexpression of ttk69, other mechanisms must exist for post-transcriptionally regulating ttk69 (Okabe, 2001).
A good candidate for a repressor of ttk69 is the msi gene product. Loss of msi function causes a double-bristle phenotype similar to that of overexpression of Ttk69 (36.3%). Furthermore, it was found that in msi mutants Ttk69 protein is ectopically expressed in IIb precursors, and that a reduction of one copy of the ttk69 gene dominantly suppresses the msi double-bristle phenotype (14.3%). Loss of ttk transforms IIa precursor into IIb precursor even in the absence of msi, confirming that ttk69 acts downstream of msi in the first asymmetric cell division (Okabe, 2001).
These results indicate that the msi phenotype is caused by a derepression of Ttk69 protein expression in the IIb neuronal precursor. This msi-dependent pathway acts parallel to the sina-dependent pathway, because a sina;msi double-null mutation shows a much more severe phenotype than either one alone (the production of extra outer support cells in 95.3% of bristles. Since msi encodes an RNA-binding protein containing two RNA recognition motifs, Msi may act directly on cis-acting sequences of the ttk69 mRNA and regulate expression of Ttk69 protein at the translational level (Okabe, 2001).
To identify cis-element(s) crucial for the post-transcriptional regulation of ttk69 mRNA, advantage was taken of ttk1, an allele of ttk with a transposable element (P-element) insertion in the 3' untranslated region (UTR) of ttk69 mRNA. Although this insertion causes the loss of another isoform of TTK (Ttk88), it also behaves as a gain-of-function allele of ttk69; Ttk69 protein is ectopically expressed in the IIb precursor in ttk1 pupal nota. In fact, it was found that ttk1 mutants display a semi-dominant double-bristle phenotype (5.3% in heterozygous mutants; 26.2% in homozygous mutants), a phenotype identical to overexpression of either Ttk69 or Ttk88. Thus, the bristle phenotype in ttk is caused by ectopic expression of Ttk69 rather than the loss of Ttk88 (Okabe, 2001).
Analysis of the ttk69 cDNA from ttk1 reveals that its ttk69 mRNA lacks the distal 80% of the 3' UTR sequence, owing to a read-through into P-element sequences. These results indicate the importance of the 3' UTR of ttk69 mRNA in vivo for preventing Ttk69 protein expression in the IIb precursor. Thus, the absence of msi or the ttk69 mRNA 3' UTR similarly results in ectopic expression of Ttk69 protein in the IIb precursor and the double-bristle phenotype (Okabe, 2001).
To test whether Msi binds ttk69 mRNA directly, Msi target sequences in vitro were analyzed in vitro, and these sequences were sought in the ttk69 mRNA. Random 30- or 50-base-pair RNA polymers were selected by using recombinant gluthathione S-transferase (GST)-Msi fusion protein immobilized to glutathione beads, and candidate RNAs were concentrated by five cycles of polymerase chain reaction with reverse transcription (RT-PCR) and in vitro transcription. 48 independent RNA polymers were isolated and sequenced, and it was found that all contained a common motif of GU3-6(G/AG) (Okabe, 2001).
Fifteen such sequences are present in the 3' UTR of the ttk69 mRNA, and ten of these are located in the region that is absent in the ttk69 mRNA produced in the ttk1 mutant. By gel mobility shift assays, it has been shown that recombinant Msi protein binds specifically to these sequences in ttk69 mRNA. The RNA ß-region, containing three Msi target sequences (corresponding to position 3,871-3,880 of ttk69 cDNA), was bound by Msi but not by a mutant Msi protein with substitutions in the RNA recognition motifs (MsiA*B*). Binding of the RNA ß-region can be competed by a homologous sequence, but not by a region of ttk69 mRNA lacking Msi-binding motifs. These results show that Msi is a sequence-specific RNA-binding protein that binds the 3' UTR of ttk69 mRNA in vitro (Okabe, 2001).
To address the functional significance of the binding of Msi to ttk69 mRNA 3' UTR, a translational reporter assay was performed in Drosophila S2 cells. Luciferase reporter genes that contained the entire 3' UTR of ttk69 (luc-ttkUTR), the proximal 20% of the 3' UTR (luc-ttkUTRdelta) or that of alcohol dehydrogenease gene (luc-AdhUTR) were introduced in S2 cells that express Msi, and translational and transcriptional outputs were quantified by measuring the enzymatic activities and mRNA levels (Okabe, 2001).
In cells expressing Msi, these three reporter genes produce roughly the same amount of mRNA. When the translational output was measured by luciferase activity, however, the luc-ttkUTR reporter produced only 24.7% activity as compared with the Adh 3'UTR. The luc-ttkUTRdelta, which has 5 of the 15 Msi-binding sites present in the ttk69 3' UTR and mimics ttk69 mRNA produced in ttk1, exhibits only a modest decrease in the translational output (69.8% activity as compared with luc-AdhUTR). This is consistent with the sharp reduction in Msi-dependent translational repression in the ttk1 mutant, because the penetrance of ttk1 homozygous is less than that of msi null homozygous. The effect of ttk69 3' UTR is dependent on the RNA-binding activity of Msi; in cells expressing Msi lacking RNA-binding activity (MsiA*B*), translational repression of the ttk69 3' UTR is not observed. The 3' UTR of ttk69 mRNA therefore confers Msi-dependent translational repression in this assay system. Together, these results indicate that Msi inhibits translation of ttk69 mRNA in IIb precursors by binding to its 3' UTR (Okabe, 2001).
Although the translational inhibitory effect of Msi on ttk69 mRNA is specific to the IIb precursors, Msi protein is present in both IIa and IIb precursors. Thus, IIa precursors must somehow be able to escape the action of Msi as a translational repressor of Ttk69, probably through the effect of Notch signaling. Loss of Notch function in the SOP lineage causes the transformation of IIa precursor into IIb, with mutants showing a balding phenotype owing to loss of socket and shaft cells. This phenotype in IIa precursor is dependent on Msi activity; loss of both Notch and msi function results in a dense double-bristle phenotype with no neurons in the subepidermal layer, indicating that the IIb precursor took the non-neuronal fate. msi is thus epistatic to Notch during the asymmetric cell division giving rise to IIa and IIb precursors. Taken together, it is proposed that Notch inhibits Msi activity in the IIa precursor (non-neuronal cell), thus allowing translation of ttk69 mRNA, whereas in the IIb precursor (neuronal cell), where Notch is inactivated by Numb, Msi prevents ttk69 translation (Okabe, 2001).