oskar
Oskar protein directs the deployment of Nanos, the posterior body-patterning
morphogen in Drosophila. To avoid inappropriate activation of nos, osk activity must appear only at the posterior pole of the oocyte, where the OSK mRNA becomes localized during oogenesis. Translation of OSK mRNA is, and must be, repressed prior to its localization; absence of repression allows OSK protein to accumulate throughout the oocyte, specifying posterior body patterning throughout the embryo. Translational repression is mediated by an ovarian protein, Bruno, that binds specifically to Bruno response elements (BREs), present in multiple copies in the OSK mRNA 3'UTR. Addition of BREs to a heterologous mRNA renders it sensitive to translational repression in the ovary (Kim-Ha, 1995).
Bruno physically interacts with Vasa. Repression of translation by Bruno is alleviated once OSK mRNA is localized to the posterior pole of the oocyte. The mechanism of this process is unknown; however, it seems likely that the RNA helicase Vasa is involved, since it is localized to the posterior pole of the oocyte and is required for efficient activation of OSK translation. Immune sera, reactive to Bruno protein, cause a shift in the electrophoretic mobility of Osk protein (Webster, 1997).
Genetic analysis of the role of bruno in oogenesis is made difficult by the lack of ovaries in bruno mutants. Flies sensitized to changes in the level of Osk protein were examined for the affects of reducing wild-type Bru protein levels. If Bru acts to repress OSK translation, a reduction in Bru protein might lead to a partial derepression of OSK translation and, subsequently, to elevated Osk activity. A transgene was used which encodes a form of OSK mRNA that retains Bruno response sequences but is mislocalized to the anterior of the oocyte. Flies bearing this genotype produce embryos with modest head defects caused by the misexpressed OSK mRNA. Reduction in Bru level enhances this phenotype, resulting in progeny with extensive anterior deletions, often accompanied by duplication of posterior pattern elements (Webster, 1997).
The product of the oskar gene directs posterior patterning
in the Drosophila oocyte, where it must be deployed
specifically at the posterior pole. Proper expression relies
on the coordinated localization and translational control of
the Oskar mRNA. Translational repression prior to
localization of the transcript is mediated, in part, by the
Bruno protein, which binds to discrete sites in the 3'
untranslated region of the Oskar mRNA. To begin to
understand how Bruno acts in translational repression, a yeast two-hybrid screen was performed to identify Bruno-interacting proteins. One interactor, described here, is the
product of the apontic gene. Expression occurs in both the somatic follicle cells and
the germline nurse cells and oocyte. APT transcripts are detected
as early as stage 2A at low levels in the germarium and at
higher levels in the follicle cells. The amount of APT mRNA in
the soma decreases during the remainder of oogenesis, while
the level in the germline increases. APT mRNA becomes
concentrated in the oocyte and also accumulates in the nurse
cells at about stage 6. APT transcripts continue to be found in
both the oocyte and nurse cells throughout oogenesis.
To determine when and where Apt protein is expressed
during oogenesis, antisera directed against a recombinant Apt
protein were prepared and used for protein detection in whole-mount
ovaries by confocal microscopy. Apt protein appears in both the germline and somatic cells of the ovary
throughout all stages of oogenesis. In the germline, Apt protein
is present in both cytoplasm and nuclei. Within the nurse cells
the protein is more concentrated in the cytoplasm, while in the
oocyte more protein is found in the nucleus. The protein,
however, is not localized to any subdomain within the
cytoplasm of either the nurse cells or the oocyte. Although Apt protein is not strictly nuclear or cytoplasmic
in cells of the female germline, the protein is highly concentrated in nuclei of the ovarian follicle cells
and in post-cellularization-stage embryos. The developmental differences in subcellular
location suggest that Apt may have functions, perhaps different, in both nuclei and cytoplasm. Nuclear proteins expressed from maternal mRNAs are sometimes present at
high levels in the cytoplasm of early embryos. Examples
include the Bicoid, Caudal and Hunchback proteins, which
appear in both nuclei and cytoplasm shortly after egg laying.
As nuclear divisions progress and the density of nuclei increases, nuclear localization of these proteins remains strong while the fraction of protein in the cytoplasm diminishes. Thus there
appears to be no early impediment to nuclear localization, simply a paucity of nuclei. In contrast, the subcellular distribution of Apt protein appears to be actively controlled in early development. Apt protein was monitored in early embryos. Even after migration of nuclei to the surface of the embryo, Apt protein remains evenly distributed between nuclei and cytoplasm, unlike any of the examples described above. This unusual persistence of Apt protein in the cytoplasm suggests the existence of a mechanism to control its distribution, reinforcing the notion of roles for Apt in both cytoplasm and nuclei (Lie, 1999a).
Apt is an RNA binding protein. Remarkably, the regions of the OSK 3' UTR bound by Apt, the
AB and C regions, are precisely those bound by Bru. A test of Apt binding was performed to determine if Bru and Apt have the same RNA binding specificity: a series of RNAs was used to map the Bru binding sites, called BREs, within the OSK C region. Three of these RNAs retain the BREs and are bound by Bru, while a fourth RNA, C4, lacks the BREs and fails to bind Bru. Apt binds all four RNAs, including C4, indicating that Apt can bind to sites other than BREs (Lie, 1999a).
Coimmunoprecipitation experiments lend biochemical support to the idea that
Bruno and Apontic proteins physically interact in Drosophila. Genetic experiments using mutants defective in
apontic and bruno reveal a functional interaction between
these genes. arrest (aret) mutants are defective for Bru (i.e. aret and bru are the same gene and lead to a developmental arrest early in oogenesis. Mutants in apt are
zygotic lethal, and some alleles also cause arrested oogenesis.
Several different apt alleles were used for all analyses, since the genetics of apt are complex and
different alleles have different effects. Testing for a genetic interaction between the aret
and apt mutants, dosages of both genes were reduced to see if this would provide a distinct phenotype. Females heterozygous for aret,
heterozygous for any of the five apt alleles, or transheterozygous for both aret and an apt allele, were crossed to wild-type males, and the progeny embryos were then
examined for cuticular defects. Females transheterozygous for aret and apt produce a fraction of embryos with head defects. Head defects can result from ectopic or excessive posterior
body patterning activity, because this activity interferes with expression of the anterior body patterning morphogen, Bicoid. Consequently, the observed head defects could be explained if both Bru and Apt contribute to
repression of OSK mRNA translation. Given this interaction, Apontic is likely to act
together with Bruno in translational repression of Oskar mRNA (Lie, 1999a).
Earlier genetic analyses of apt have concentrated on the zygotic
phenotype. To define more completely the role of apt in the female
germline, females were created with apt minus germline clones using
the FLP/DFS method. Ovaries containing germline clones were dissected, stained with DAPI
to highlight nuclei, and examined for phenotype. Different apt
mutants display dramatically different ovarian phenotypes.
One allele is indistinguishable from wild type, because females with germline clones of this allele have phenotypically wild-type ovaries and lay eggs that develop into
fertile adults. Females with germline clones of another allele also have
phenotypically wild-type ovaries, but a small fraction of the
eggs laid developed into embryos with head defects. In contrast, ovaries from females with germline clones from two other mutants have phenotypes that are similar to one another and severe: development is arrested in early oogenesis, and the oocyte fails to differentiate, with all nuclei becoming polyploid. In addition, some of the egg chambers have an abnormal number of nuclei. It is concluded that apt
is necessary for oogenesis and that loss of apt activity leads to
a developmental arrest during oogenesis. Just as for aret mutants, the arrest occurs too early to allow the ovaries to be examined for defects in OSK mRNA translation (Lie, 1999a).
Interestingly, Apontic, like Bruno, is an RNA-binding protein and specifically binds certain regions of the
Oskar mRNA 3' untranslated region. A sequence shared by all of the bound RNAs could not be identified. Thus, despite its ability to efficiently discriminate between different parts of the OSK mRNA, Apt appears to be relatively promiscuous in its binding and may recognize many sites or
perhaps a structural feature common to many RNAs (Lie, 1999a).
Translational regulation plays a prominent role in
Drosophila body patterning. Progress in elucidating the
underlying mechanisms has been limited by the lack of a
homologous in vitro system that supports regulation.
Global control over all transcripts can be achieved through changes in the activity of the translation machinery. More specific controls often rely on cis-acting
regulatory elements present within mRNAs, in many cases in
their 3' untranslated regions (3' UTRs).
Extracts prepared from Drosophila tissues are
competent for translation. Ovarian extracts, but not
embryonic extracts, support the Bruno response element-dependent and
Bruno-dependent repression of Oskar mRNA
translation, which acts in vivo to prevent protein synthesis
from transcripts not localized to the posterior pole of the
oocyte. Consistent with suggestive evidence from in vivo
experiments, regulation in vitro does not involve changes in
poly(A) tail length. Moreover, inhibition studies strongly
suggest that repression does not interfere with the process
of 5' cap recognition. Translational regulation mediated
through the Bruno response element is thus likely to occur
via a novel mechanism (Lie, 1999b).
Two types
of experiments were performed to explore the possibility that
OSK translational regulation occurs by changes in the poly(A) tail.
First, a direct assay was used to determine the length of the
poly(A) tail on OSK ovarian transcripts. The poly(A)
tail is short, suggesting that polyadenylation is unlikely to
function in regulating translation of OSK mRNA. Second, the distribution of poly(A)-binding protein (PABP) in the Drosophila ovary was determined.
Although PABP is present at high levels in the germline cells
early in oogenesis, it is noticeably depleted from the oocyte at
the stage when OSK mRNA is localized to the posterior pole and
translationally activated. Although these experiments
are suggestive, neither observation provides a compelling
argument against a role for polyadenylation in translational
activation. For example, the poly(A) tail could act in a manner
not requiring PABP. Furthermore, the fraction of translationally
active OSK mRNA may be small, as observed for NOS mRNA, making it difficult to detect the
presence of transcripts with longer poly(A) tails.
To facilitate biochemical analysis of translational control in
Drosophila and definitively address the role of polyadenylation
in regulation of OSK mRNA translation, in vitro
translation systems were developed from Drosophila tissues. Although similar
systems have been prepared from other sources, they are
unlikely to contain the factors necessary for specific
translational regulation of Drosophila mRNAs. Translational activity of
the extracts was monitored using reporter mRNAs encoding
luciferase (Lie, 1999b).
The ability of Drosophila extracts to recapitulate regulated
translation was tested using the 3' UTR of the OSK mRNA,
which contains the BRE control elements that mediate Bru-dependent
translational repression. The
luciferase (luc) reporter mRNA was modified by addition of
the following sequences to its 3' end: either the wild-type OSK
3' UTR (BRE+) or a point-mutated version of the OSK 3' UTR
(BRE-) that is unable to bind Bru protein in vitro and fails to
support translational repression in vivo.
In embryo extracts, as well as reticulocyte lysates, both
mRNAs were translated with similar efficiencies,
revealing no inherent differences in their abilities to be
translated. In contrast, translation of the two mRNAs is
markedly different in ovary extracts: the luc BRE+ message
is translationally repressed approximately 9-fold, on average,
in comparison to the luc BRE- message. These results indicate that the ovary extract supports
specific regulation of translation (Lie, 1999b).
To determine whether the dependence of regulated
translation on BREs reflects a requirement for Bru, two related
experiments were performed. First, ovarian extracts were
immunodepleted using anti-Bru antibodies or antibodies
purified from normal rat serum and were then tested for
translational activity. Notably, the control antibodies have no
effect on the relative levels of BRE+ and BRE- translation, but
depletion with anti-Bru antibodies largely eliminates BRE-dependent
repression. Purified recombinant Bru
protein was added to the immunodepleted extracts but was not
sufficient to restore translational regulation activity. This result suggests that,
in addition to removing Bru, immunodepletion using anti-Bru
antibodies may remove other proteins required for BRE-mediated
translational regulation. Indeed, a significant fraction
of Bru protein is present in a large macromolecular complex. In a second type of
experiment, purified Bru was added to embryonic extracts that
lack Bru and do not support BRE-mediated
regulation of translation. Thus BRE-dependent
translational repression in vitro requires Bru.
Although addition of Bru to embryo extract promotes BRE-dependent
repression, the relative translation of BRE- versus
BRE+ RNAs is less (average 3.2-fold) than that measured in
ovary extracts (average 9.2-fold). There are at least two likely explanations
for this difference: the recombinant Bru protein may not be
fully active, although the protein displays RNA-binding
activity indistinguishable from that of ovarian Bru; alternatively, additional
protein(s) present in ovaries but not in embryos may also
contribute to repression and may be required for wild-type
levels of activity (Lie, 1999b).
The use of BRE+ and BRE- mRNAs reveals the importance
of BREs for translational regulation, but does not address the
possible role of other sequences in the OSK mRNA 3' UTR. To
determine if BREs alone are sufficient, luc reporter mRNAs
bearing multimerized consensus BREs (8 copies), either wild-type
or containing point mutations that abrogate Bru binding, were generated and their translation
measured in vitro. The BRE+ and BRE- RNAs were translated
with only a modest difference in efficiency,
suggesting that binding of Bru alone is insufficient for
complete repression. The 8x BRE
supports translational repression in vivo, but repression is not
efficient, consistent with the results of
the more quantitative in vitro assay. In contrast, RNAs
containing the full OSK AB region (a 124 nt region of the OSK
mRNA 3' UTR containing at least four consensus BREs
interspersed among other sequences) were translationally
regulated with a much higher efficiency. Thus it
appears that other cis-acting sequences and presumably other
factors contribute to Bru-mediated translational repression, a result
consistent with the incomplete repression conferred by the
addition of Bru to embryonic extracts (Lie, 1999b).
To begin to explore the mechanism of Bru-mediated
repression, the in vitro system was used to rigorously test the
role of the poly(A) tail in
translational regulation, as well as the role of the 5' cap. A number of maternal mRNAs in
Xenopus have been shown to be translationally regulated in a
manner involving changes in the length of the poly(A) tail (e.g.
cyclin mRNAs and c-mos mRNA), and
RNA control elements and proteins that contribute to this
process have been identified (e.g. the cytoplasmic
polyadenylation element (CPE) and the CPE-binding protein
(CPEB). Similar forms of regulation have been reported for
maternal mRNAs from other animals such as Bicoid and
Hunchback mRNAs in Drosophila and tPA and c-mos mRNAs in mouse, and for mRNAs from somatic
tissues such as alpha-CaMKII mRNA. To
specifically study the requirement for the poly(A) tail in OSK
mRNA translational regulation, BRE+ and BRE- reporter
RNAs that lacked poly(A) tails were tested in the in vitro
system. The translation of these RNAs was regulated in a
manner similar to that of transcripts with a poly(A) tail. Though the overall level of translation from the poly(A)-messages was slightly reduced, BRE-dependent
repression remains strong. To address the
possibility that reporter mRNAs might be polyadenylated
during the course of the reaction, two types of experiments
were performed using poly(A)- reporter RNAs bearing only
the 5' portion of the OSK 3' UTR, including the AB region. This
part of the OSK 3' UTR lacks all of the normal polyadenylation
elements, including sequences that resemble the binding site
for Xenopus CPEB, a factor involved in cytoplasmic
polyadenylation. Initially, the translational regulation of these reporter
mRNAs was tested and found to be similar to that of RNAs
containing the intact 3' UTR. Then the lengths of
reporter mRNAs were monitored over the course of incubation
in the in vitro translation extract: no changes in size were
observed, confirming that polyadenylation does not
occur in vitro. It is concluded from these data that translational
regulation of OSK mRNA in vitro is independent of the poly(A)
tail and that regulation occurs via a novel mechanism (Lie, 1999b).
Another mRNA feature often implicated in control events is
the 5' cap. Cap-dependent translational initiation requires
recognition of the cap by the eIF4F complex (or its constituent
components), a process that is modulated by several types of
regulation. A simple test of
the importance of the cap in Bru-mediated regulation is
to monitor relative translation efficiencies of BRE+ and BRE-
mRNAs lacking cap structures. However, uncapped mRNAs
are quite unstable in the extracts. An alternative approach is to inhibit cap-dependent
initiation by the addition of excess free cap analog (7-methyl-GpppG).
In a preliminary experiment, it was shown that
translation in the extracts can be substantially inhibited by free
cap analog and is thus largely cap-dependent. To test
for cap-dependence of Bru regulation, translation
of BRE+ and BRE- mRNAs was tested in the presence of free cap. If Bru
interferes with recognition or use of the cap, the levels of BRE-
and BRE+ translation would be expected to equalize under
conditions where cap-dependent initiation is inhibited.
Notably, the ratio of BRE-/BRE+ translation remains similar
under all conditions tested, ranging from 0 to 2 mM cap
competitor. These results very strongly suggest
that Bru interferes with a step in translation that is distinct
from cap recognition. Additional evidence supporting this
conclusion could come from examination of dicistronic
mRNAs, in which translation of one encoded protein is
initiated in a cap-independent pathway through the use of an
internal ribosome entry site (IRES). However, such an experiment must await the
identification of an IRES that is active in Drosophila ovaries (Lie, 1999b).
Prior analysis of OSK mRNA translational regulation has
provided strong but indirect evidence that Bru acts as a
repressor. Molecular genetic data reveals the essential role for
BREs, while genetic evidence
demonstrates that Bru acts in controlling the level of OSK
activity. However, the complex
phenotype of mutants defective in Bru prevented a direct
demonstration that the absence of Bru leads to a derepression of
OSK translation. The results of the in vitro
studies now provide compelling evidence that Bru is in fact
required for translational repression mediated through the
BREs. How the binding of Bru to the 3' UTR of OSK mRNA
leads to translational repression remains uncertain, although
the availability of the in vitro system defined here is likely to
prove useful in addressing that question. Indeed, a definitive
demonstration that the 5' cap and changes in poly(A) tail length
are not involved in regulation were only made possible through
use of this system (Lie, 1999b).
Bicaudal-C (Bic-C) is required during Drosophila melanogaster oogenesis for several processes, including anterior-posterior patterning. The gene encodes a protein with five copies of the KH domain, a motif found in a number of RNA-binding proteins. Using antibodies raised against the Bic-C protein, it is shown that multiple isoforms of the protein exist in ovaries and that the protein, like the RNA, accumulates in the developing oocyte early in oogenesis. The pattern of Bic-C expression is similar to that of the Bic-C RNA, except that the RNA is detectable in a single cell (the presumptive oocyte) as early as germarium region 2A. In contrast, specific accumulation of BIC-C protein in the oocyte is first detectable at stages 3 and 4 of oogenesis but remains very faint until stage 5, when the protein level increases substantially. During stages 4 to 6, Bic-C protein is visible throughout the oocyte cytoplasm but is enriched at the posterior pole of the oocyte. During stages 7 to 9, Bic-C protein is abundant in the oocyte cytoplasm, with some enrichment at the anterior of the oocyte and around the oocyte cortex. In stage 10 and in later stages, the protein is expressed at high levels in the nurse cells. Bic-C protein expressed in mammalian cells can bind RNA in vitro, and a point mutation in one of the KH domains that causes a strong Bic-C phenotype weakens this binding. In addition, Oskar translation commences prior to posterior localization of Oskar RNA in Bic-C- oocytes, indicating that Bic-C may regulate Oskar translation during oogenesis (Saffman, 1998).
OSK mRNA is ectopically localized in part to the anterior of eggs produced by Bic-C/+ or Bic-C/Bic-C females. However, a substantial amount of OSK mRNA still localizes normally to the oocyte posterior, even in homozygous Bic-C oocytes. To determine whether Osk protein expression is affected in Bic-C mutants, immunohistochemistry was used to analyze Osk protein expression in ovaries from Bic-C/Bic-C females. In wild-type oocytes, translation of OSK is repressed until posterior localization of the RNA at stage 9, resulting in restriction of Osk protein to the posterior tip of the oocyte. In contrast, in Bic-C ovaries, OSK mRNA is prematurely translated, beginning in stages 7 and 8. Through stages 7 to 10, Osk protein remains diffuse and is most concentrated near the center of the oocyte. Surprisingly, despite its precocious translation in Bic-C mutants and the substantial posterior concentration of its RNA, Osk does not accumulate at the posterior pole as late as stage 10. It is possible that pole plasm-specific activation of OSK translation is also compromised by Bic-C mutations. However, as many developmental defects become apparent in Bic-C egg chambers beyond stage 9, and oogenesis fails to progress beyond stage 10, it cannot be ascertained that this failure to activate OSK translation is a specific consequence of Bic-C mutations. These results indicate that the RNA-binding activity of Bic-C is necessary for the correct repression of OSK translation. These results suggest that Bic-C may function directly to regulate the translation of target RNAs such as OSK (Saffman, 1998).
Smaug
acts as a translational repressor; it binds NOS mRNA at a stem loop structure found within the Nos translational control element. Signals that mediate regulation of NOS mRNA reside in its 3' UTR. In particular, the 184 nt translational control element (TCE) contains
all of the 3' UTR signals that are necessary and sufficient for NOS function.
The key components of the TCE consist of a pair of
redundant hairpins, each bearing the loop sequence CUGGC. These mediate both repression of NOS mRNA in the bulk cytoplasm as well as Oskar-dependent activation in the pole plasm. Thus, the TCE hairpins constitute the
essential cis-acting elements of a translational switch responsible for generating a polarized distribution of Nos protein in the early embryo (Dahanukar, 1999 and references).
Using a set of mutant TCE hairpins, it was asked whether binding to Smaug in vitro correlates with TCE-mediated repression of NOS mRNA in vivo. Binding was monitored in gel mobility shift experiments, and TCE activity was monitored using transgenic flies that express appropriately altered NOS mRNAs. In brief, derepression of NOS mRNA in the bulk cytoplasm, where NOS mRNA is usually held in a translationally silent repressive complex, results in a reduction in the levels of anterior Bcd and Hb proteins, which in turn results in the development of lethal head defects. Binding of Smaug to the G12U mutant TCE, which is strongly defective in vivo, is reduced by a factor of at least 50 relative to wild type; binding to the moderately defective A15G mutant TCE is reduced by a factor of at least 5; and binding to the U18C mutant TCE, which regulates nos normally, is indistinguishable from binding to the wild-type hairpin. Thus, the variation in degree of binding of these mutant hairpins to Smaug indeed correlates with the mutant's capacity to repress translation of NOS mRNA in the bulk cytoplasm of the embryo (Dahanukar, 1999).
Overproduction of Smaug represses NOS mRNA in the pole plasm. Smaug protein is distributed throughout the preblastoderm embryo, with no detectable difference between its concentration in the bulk cytoplasm and the pole plasm. Why, then, does Smaug not repress the translation of NOS mRNA in the pole plasm? One explanation is that, in fact, Smaug-dependent repression competes with Osk-dependent activation, with Osk prevailing in wild-type embryos. To investigate this idea, Smaug was overproduced by introducing up to four extra copies of a smg+ transgene, thereby generating '6× smg+' embryos. The extent of Smaug overproduction appears to be approximately proportional to the gene dose. Otherwise, wild-type 6× smg+ embryos are completely viable and exhibit no segmentation defects. Moreover, the distribution of Nos protein during early development appears normal in such embryos, suggesting that this level of Smaug does not significantly interfere with Osk-dependent activation of wild-type NOS mRNA. However, excess Smaug clearly interferes with NOS mRNA translation in two different sensitized genetic backgrounds. While it is not understood why overproduction of Smaug is without apparent consequence in the pole plasm of wild-type embryos, one simple possibility is that a 3-fold increase in Smaug concentration is insufficient to repress translation in wild-type embryos, but that higher levels of Smaug would do so (Dahanukar, 1999).
Oskar interacts with the RNA-binding domain of Smaug. Smaug and Osk compete in the pole plasm, the former repressing and the latter activating translation of NOS mRNA. Smaug evidently acts by binding to the TCE hairpins of NOS mRNA. The molecular mechanisms by which Osk acts are not yet clear, although it plays a central role in both pole plasm assembly and activation of NOS translation.
In particular, two lines of evidence suggest that Osk is the limiting component in the embryo for translational activation of NOS: (1) unlike other gene products required for pole plasm assembly, which are also present throughout the bulk cytoplasm, Osk is found only in the pole plasm; (2) overexpression of Osk is sufficient to activate NOS translation throughout the embryo. The mutually antagonistic activities of Osk and Smaug might be the result of a direct interaction between the two. To test this possibility, plasmids that direct the synthesis of various fragments of Osk and Smaug in yeast were constructed, and protein-protein interactions were assessed using the two-hybrid technique. Smaug interacts specifically with Osk in yeast. The region of Smaug that mediates this interaction corresponds to a 31 kDa fragment that contains the minimal RNA-binding domain. Further mutational analysis of this domain suggests that its TCE- and Osk-binding activities are not readily separable (Dahanukar, 1999). The region of Osk that mediates binding to Smaug consists of residues 290-418 (Dahanukar, 1999), a domain of the protein that may also mediate interactions with the pole plasm constituents Vas and Staufen (Breitwieser, 1996).
Taken with earlier work, these results support a simple model for the operation of a translational switch that governs expression from NOS mRNA. In the bulk cytoplasm, repression of NOS mRNA is dependent on the activity of Smaug, which binds to the essential targets in the 3' UTR. In the pole plasm, Smaug-mediated repression is antagonized by Osk, which interacts with the RNA-binding domain of Smaug. Currently, it is not know whether Osk interacts with Smaug bound to the TCE or whether Osk competes with the RNA for binding to Smaug. In either case, Smaug-dependent repression is overcome, and Nos protein accumulates in the posterior of the embryo. Osk also activates translation via other signals in the NOS 3' UTR. However, unlike the translational switch governed by Smaug, these signals are dispensable for NOS function in the embryo (Dahanukar, 1999 and references).
The gene aubergine is required to enhance Oskar translation. While aubergine-dependence is conferred upon Oskar mRNA by sequences in the Oskar 3' UTR, Aubergine may influence Oskar translation through an interaction with sequences upstream of the Oskar 3' UTR (Wilson, 1996).
The establishment of polarity axes in the Drosophila egg and embryo depends upon the localization and on-site expression of maternal mRNAs. The critical step in
the targeting of posterior determinants is the localization of Oskar (OSK) mRNA to the pole and its on-site translation. Osk protein then recruits other posterior group
gene products involved in the formation of pole plasm and in the localization and regulation of the posterior determinant, Nanos. The role of
the Drosophila CPEB homolog, the orb gene, in the OSK mRNA localization pathway has been analyzed. The expression of Osk protein is dependent upon the orb
gene. In strong orb mutants, Osk protein expression is undetectable, while in the hypomorphic mutant, orbmel, little or no on-site expression of Osk protein at the
posterior pole is observed. The defects in Osk protein accumulation in orb mutant ovaries are correlated with a reduction in the length of the OSK poly(A) tails. OSK mRNA is found in immunoprecipitable complexes with Orb protein in ovaries. The OSK 3' UTR can be UV cross-linked to Orb protein in ovarian
extracts. These data suggest that Orb is required to activate the translation of OSK mRNA and that this may be accomplished by a mechanism similar to that used
by the Xenopus CPEB protein to control translation of 'masked' mRNAs (Chang, 1999).
What role does Orb protein play in the translation of OSK mRNA localized at the posterior pole in vitellogenic oocytes? One model that could potentially account for the current findings posits that translationally masked OSK mRNA would be deposited from the nurse cells into the anterior of the oocyte and then transported to the posterior pole by the Stauffen protein. Good candidates for ensuring that OSK mRNA is not translated in the nurse cells or at the anterior of the oocyte would be the Bruno protein and p50 protein. When the masked OSK mRNA arrives at the posterior pole, it would be bound by an Orb protein complex that is already anchored in the posterior cytoskeleton. Orb protein would then activate OSK mRNA translation, most probably by promoting polyadenylation (Chang, 1999 and references therein).
The coupled regulation of Oskar mRNA localization and
translation in time and space is critical for correct
anteroposterior patterning of the Drosophila embryo.
Localization-dependent translation of Oskar mRNA, a
mechanism whereby Oskar RNA localized at the posterior
of the oocyte is selectively translated and the unlocalized
RNA remains in a translationally repressed state, ensures
that Oskar activity is present exclusively at the posterior
pole. Genetic experiments indicate that translational
repression involves the binding of Bruno protein to
multiple sites, the Bruno Response Elements (BRE), in the
3' untranslated region (UTR) of Oskar mRNA. A cell-free translation system, derived from
Drosophila ovaries, has been established that faithfully reproduces critical
features of mRNA translation in vivo, namely cap structure
and poly(A) tail dependence. This ovary
extract, containing endogenous Bruno, is able to
recapitulate Oskar mRNA regulation in a BRE-dependent
way. Thus, the assembly of a ribonucleoprotein (RNP)
complex leading to the translationally repressed state
occurs in vitro. Moreover, a Drosophila
embryo extract lacking Bruno efficiently translates Oskar
mRNA. Addition of recombinant Bruno to this extract
establishes the repressed state in a BRE-dependent manner,
providing a direct biochemical demonstration of the critical
role of Bruno in Oskar mRNA translation. This approach
opens new avenues to investigate
translational regulation in Drosophila oogenesis at a
biochemical level (Castagnetti, 2000).
The hiiragi (hrg) gene encodes a poly(A) polymerase (PAP), an enzyme that attaches adenylyl residues to the 3' untranslated region of mRNAs. The single Drosophila PAP is active in specific polyadenylation in vitro and is involved in both nuclear and cytoplasmic polyadenylation in vivo (Juge, 2002). Therefore, the same PAP can be responsible for both processes. In addition, in vivo overexpression of PAP during embryogenesis does not affect poly(A) tail length during nuclear polyadenylation, but leads to a dramatic elongation of poly(A) tails and a loss of specificity during cytoplasmic polyadenylation, resulting in embryonic lethality. Thus regulation of the PAP level is essential for controlled cytoplasmic polyadenylation and early development. hrg is also probably essential to cell viability since strong hrg mutant germline clones do not survive. The PAP encoded by this gene is involved in polyadenylation of oskar mRNA in oocytes. This indicates that although the reactions of nuclear and cytoplasmic polyadenylation are not identical, a single PAP is responsible for both in Drosophila (Juge, 2002).
The molecular mechanism of cytoplasmic polyadenylation has been analysed extensively in Xenopus oocytes, and some aspects of the reaction are similar to that of nuclear polyadenylation. Nuclear polyadenylation consists of endonucleolytic cleavage of pre-mRNAs followed by the synthesis of a poly(A) tail onto the upstream cleavage product. Poly(A) addition can be reconstituted in vitro from three purified mammalian factors: poly(A) polymerase (PAP), cleavage and polyadenylation specificity factor (CPSF) and poly(A)-binding protein II [PABP2, the nuclear poly(A)-binding protein]. CPSF is a complex of four proteins that binds the polyadenylation signal AAUAAA located upstream of the cleavage site. Recognition of the poly(A) site also requires cleavage stimulation factor (CstF) that binds to a GU/U-rich element downstream of the cleavage site and interacts with CPSF. PAP catalyses the polyadenylation reaction, but is also required for efficient cleavage of pre-mRNAs in vitro. PAP by itself does not recognize pre-mRNAs specifically. Specificity requires the AAUAAA element and CPSF that binds PAP through its 160 kDa subunit. Even in the presence of CPSF, PAP activity remains weak; it is again stimulated by binding of PABP2 to the poly(A) tail. Together, CPSF and PABP2 stimulate PAP activity by holding PAP on the RNA such that a full-length poly(A) tail is synthesized in a single processive event. When the poly(A) tail has reached its complete length, elongation is no longer processive and becomes slow and distributive. PABP2 is required for this poly(A) tail length control (Juge, 2002 and references therein).
In Drosophila, the role of CPEs has not been addressed, and the polyadenylation signal is dispensable in some cases, since embryonic cytoplasmic polyadenylation occurs on a bicoid engineered mRNA deleted for this element. Although genes encoding the four subunits of CPSF are present in the Drosophila genome, their role in cytoplasmic polyadenylation has not been determined. The Drosophila homolog of CPEB is the Orb protein. orb encodes germline-specific proteins different in male and female, and its function has been determined in the female germline. Strong orb mutants arrest oogenesis early, before the formation of the 16-cell cyst that would normally differentiate into nurse cells and one oocyte. Using a weaker allele, orbmel, Orb was shown to be required for anchoring of oskar mRNA at the posterior pole of the oocyte. However, this could result from a failure in oskar mRNA translation since Oskar protein is required for anchoring its own mRNA at the posterior pole. A recent study suggests that Orb could have a function analogous to that of CPEB in cytoplasmic polyadenylation. In orb mutant egg chambers, the level of Oskar protein is decreased and poly(A) tails of oskar mRNAs are shortened (Juge, 2002 and references therein).
To address a possible role for hrg in cytoplasmic polyadenylation during early development, germline clones homozygous for hrgPAP45, hrgPAP21 or hrgPAP12 were induced. No germline clones were obtained for any of these mutants, possibly as a result of a requirement of PAP for cell viability. Therefore genetic interactions were studied between hrg and orb, which is known to be involved in cytoplasmic polyadenylation. Females homozygous for the weak orbmel allele produce egg chambers at all stages and lay eggs, 30% of which show a ventralized phenotype. hrg lethal mutants act as dominant enhancers of the orbmel phenotype, since hrgPAP45/+; orbmel and hrgPAP21/+; orbmel females lay almost no eggs. In these females, oogenesis stops most frequently at stage 7/8, after which egg chambers degenerate, even though one or two stage 14 oocytes per ovary can be observed. Poly(A) tails of oskar mRNA are shortened in orb mutant ovaries. The defect of these poly(A) tails were analyzed in hrg- /+; orbmel mutants by PAT assays. oskar mRNA poly(A) tails were measured to be up to 135 residues in wild-type ovaries. These poly(A) tails are weakly reduced in orbmel, but severely reduced in hrg- /+;orbmel double mutant ovaries, their maximal length reaching 40- 50 residues. These short poly(A) tails do not result from the oogenesis defect in hrg- /+; orbmel females, since unrelated mutants that stop oogenesis early show wild-type poly(A) tails of oskar mRNA. These poly(A) tails were also found to be of wild-type length in hrgPAP21/+ and hrgPAP45/+ ovaries. This shows that the strong shortening of oskar poly(A) tails in hrg- /+; orbmel mutants does not result from an additive effect of two phenotypes, but from a synergistic effect of the two mutants due to a simultaneous decrease in PAP and Orb protein levels. This strongly suggests that hrg and orb are involved together in cytoplasmic polyadenylation. This was confirmed by measurements of poly(A) tails of a control mRNA, sop, which is thought not to be regulated by cytoplasmic polyadenylation. Poly(A) tails of sop mRNAs are unaffected in orbmel as well as in hrg- /+; orbmel mutant ovaries. It was verified that shortening of oskar mRNA poly(A) tails in hrg- /+; orbmel mutants leads to a reduction of Oskar protein level, by immunostaining of ovaries with anti-Oskar. Oskar accumulates at the posterior of the oocyte from stage 9 onwards. The amount of Oskar decreases in orbmel oocytes. This amount decreases again in hrg- /+; orbmel oocytes to a barely detectable level (Juge, 2002).
Taken together, these results show that hrg and orb cooperate in poly(A) tail lengthening during cytoplasmic polyadenylation and that alteration of this process affects protein accumulation and oogenesis (Juge, 2002).
During Drosophila oogenesis, the posterior determinant, Oskar, is tightly localized at the posterior pole of the oocyte. The exclusive accumulation of Oskar at this site is ensured by localization-dependent translation of oskar mRNA: translation of oskar mRNA is repressed during transport and activated upon localization at the posterior cortex. Previous studies have suggested that oskar translation is poly(A)-independent. This study shows that a long poly(A) tail is required for efficient oskar translation, both in vivo and in vitro, but is not sufficient to overcome Bruno response element-mediated repression. Moreover, accumulation of Oskar activity requires the Drosophila homolog of Cytoplasmic Polyadenylation Element Binding protein (CPEB), Orb. Since posterior localization of oskar mRNA is an essential prerequisite for its translation, it was critical to identify an allele of orb that does localize oskar mRNA to the posterior pole of the oocyte. Flies bearing the weak mutation orbmel localize oskar transcripts with a shortened poly(A) that fails to enhance oskar translation, resulting in reduced Oskar levels and posterior patterning defects. It is concluded that Orb-mediated cytoplasmic polyadenylation stimulates oskar translation to achieve the high levels of Oskar protein necessary for posterior patterning and germline differentiation (Castagnetti, 2003).
Cytoplasmic polyadenylation of mRNA requires CPEB to recruit the enzyme
poly(A) polymerase (PAP) on the regulated mRNA. Until recently, only one
family of PAP, containing both a catalytic domain and an RRM-like domain, was
known. A novel family of PAP has now been identified (Wang, 2002) that differs from
the canonical PAP for the absence of the RRM-like domain. The prototype of
this family is represented by C. elegans GLD-2 whose binding to the
RNA is mediated by GLD-3, a KH domain containing protein of the BicC family.
Interestingly, Drosophila BicC interacts physically
with Orb in co-immunoprecipitation experiments. Since the phenotype of
BicC mutants implicates BicC protein as a negative regulator of
osk translation (Saffman,
1998), tests were performed to see whether Orb interacts with the translational
repressor Bru. Indeed, a physical interaction between Bru and
Orb has been detected, as revealed by
the co-immunoprecipitation of Bru with Orb. By contrast, no direct
interaction has been detected between Bru and BicC (Castagnetti, 2003).
The relevance of these interactions in vivo is further confirmed by the
genetic interactions between the BicC locus and the orb and
aret loci -- the latter encoding Bru protein. Females
heterozygous for BicC show a number of AP patterning defects, ranging
from head defects to bicaudal embryos. The
BicC phenotype is suppressed when the mutation is combined with an
orb allele or the strong aret allele,
aretQB72. 85% of embryos produced by
Bic-CYC33/+ females fail to hatch and of those 60% are
bicaudal. The null allele orbF343 efficiently suppresses
the Bic-C phenotype and only 13% of the embryos produced by
BicCYC33/+; orbF343/+ fail to hatch,
none of which show a bicaudal phenotype. The strength of the
phenotype and the extent of the suppression depends on the orb
allele. Embryonic viability is also improved, up to 72%, in embryos produced
by BicCYC33/ aretQB72 females (Castagnetti, 2003).
Thus, combining measurement of osk poly(A) tail length in vivo with
quantification of the translation activity of the corresponding mRNAs in vivo
and in vitro, polyadenylation has been shown to corrolate with the
translational status of the mRNA. A mutation in Orb, the Drosophila
CPEB, leads to shortening of the osk poly(A) tail and to a reduction
in Osk accumulation. The fact that, at least in vitro, a long poly(A) tail
neither overcomes BRE-mediated repression nor is necessary for repression of
osk reporter transcripts, suggests that cytoplasmic polyadenylation
is not the decisive event in translational activation of osk at the
posterior pole. Rather, it appears that the presence of a 200 A long tail on
osk mRNA promotes its efficient translation, allowing accumulation of
Osk to levels sufficient for both abdomen and germline formation to
proceed (Castagnetti, 2003).
The prevailing model, which is based on studies of translational control in
the Xenopus oocyte, suggests that the polyadenylation status of a
transcript correlates with its translational status: a short poly(A) tail
corresponding to a silenced mRNA and poly(A) tail elongation triggering
translational activation. In Xenopus, upon progesterone treatment a
wave of cytoplasmic polyadenylation activates translation of deadenylated and
silenced maternally derived mRNAs. In
Drosophila, translation of bicoid mRNA, which encodes the
anterior determinant of the embryo, is repressed until egg activation, when
poly(A) tail elongation triggers translation initiation.
Although the correlation between adenylation and translation still holds for
several transcripts, a growing body of evidence suggests that the two events
may be coincidental but not directly connected. Interestingly, deadenylation
and translational repression of Drosophila hunchback (hb)
and mouse tPA mRNAs can occur independently of each other. The transcripts
are deadenylated concomitant with translational repression, yet repression can
occur in the absence of ongoing deadenylation. In arrested primary mouse
oocytes, polyadenylation of the tPA mRNA is necessary to counteract the
default deadenylation that affects most other oocyte mRNAs, thus preventing
its degradation (Castagnetti, 2003).
Observations made in this study suggest that silencing and awakening of osk mRNA
translation can occur in the absence of changes in poly(A) tail length and, in
fact, osk mRNA bears a long poly(A) tail at all stages of oogenesis,
including when it is unlocalized and translationally silent. However, it is
still formally possible that at intermediate stages of oogenesis osk
mRNA undergoes a deadenylation that goes undetected in measurements on
bulk RNA, and that elongation of the poly(A) tail causes displacement of the
repressor complex, leading to translational derepression. This hypothesis is
supported by the fact that the repressor protein Bru shares a 50% sequence
identity with the Xenopus deadenylation promoting factor EDEN-BP.
However, no obvious pattern of deadenylation has been detected in vitro when
Bruno is added to the embryonic extract, nor could a shortening be detected of
the poly(A) tail of translationally silenced osk transcript recovered
from ovarian extract. Nevertheless, the results show that
BRE-mediated repression is effective independent of the length of the
poly(A) tail on osk transcripts, and that a silenced mRNP can be
assembled on a naked osk transcript, whether or not it bears a
poly(A) tail. These results suggest that polyadenylation is not the sole
determining event leading to translational derepression of osk mRNA
at the posterior pole, but that the maintenance of a long poly(A) tail, by
cytoplasmic polyadenylation, accounts for the enhancement of osk
translation and is required for efficient osk translation, to ensure
sufficient accumulation of Osk at the posterior pole of the
Drosophila oocyte to promote abdominal patterning and germline
differentiation (Castagnetti, 2003).
Furthermore, the physical interaction detected between Orb and Bru, and Orb
and BicC suggests the existence of a multi-protein complex containing both
positive and negative regulators of osk translation. In this
scenario, translational silencing and polyadenylation are linked through Bru
protein, offering a possible explanation as to how CPEB might be recruited to
mRNAs in Drosophila, where no canonical CPE has so far been
identified. Transcripts properly repressed by Bru, upon localization, could be
adenylated by the recruitment of Orb by Bru itself. Loss of Bru repression
would, therefore, result in loss of Orb binding with consequent deadenylation
and translational silencing. In this model, modulation of the poly(A) tail
would be part of the mechanism that regulates translation, ensuring a second
level of control over ectopic expression while localizing all the components
necessary for efficient translation. Remarkably, mutations in the BRE sites do
not result in ectopic osk translation,
suggesting the existence of a second layer of translational control. Moreover,
during embryonic development when osk translation is no longer
required, both Orb and Bru proteins are depleted in the embryo and
osk mRNA undergoes complete deadenylation (Castagnetti, 2003).
A DEAD-box protein, Me31B, forms a cytoplasmic RNP complex with oocyte-localizing RNAs. During early oogenesis, loss of Me31B causes premature
translation of oocyte-localizing RNAs within nurse cells, without affecting their transport to the oocyte. In early egg chambers that lack Me31B, at least two mRNAs in particles, OSK and Bicaudal-D mRNAs, are prematurely translated in nurse cells, though the transport of these RNAs to the oocyte is Me31B independent. These results suggest that Me31B mediates
translational silencing of RNAs during their transport to the oocyte. These data provide evidence that RNA transport and translational control are
linked through the assembly of RNP complex (Nakamura, 2001).
A visual screen was conducted with an ovarian GFP-cDNA library, in which fusion genes are expressed in germline cells during oogenesis. Transgenic flies were generated with this library and proteins were identified that distribute in a granular pattern during oogenesis. Screening ~3000 independent lines, one was isolated in which GFP signals were detected as cytoplasmic particles during oogenesis. The particles were dispersed in the cytoplasm of both nurse cells and oocytes but never detected within nuclei. The particles were frequently observed passing through ring canals, suggesting that the particles are assembled in nurse cell cytoplasm and transported to the oocyte (Nakamura, 2001).
The cDNA from this line was identified as me31B. In the cDNA fusion, almost the entire coding region of me31B, which lacks only the first four codons, was fused in frame with that of gfp. Me31B, a DEAD-box protein and therefore a putative ATP-dependent RNA helicase, was isolated as a gene expressed extensively during oogenesis. Me31B is a part of an evolutionally conserved DEAD-box protein group, which includes human RCK/p54 (71% identical), Xenopus Xp54 (73%), Caenorhabditis elegans C07H6.5 (76%), Schizosaccharomyces pombe Ste13 (68%) and Saccharomyces cerevisiae Dhh1 (68%). Furthermore, Me31B is phylogenetically close to two evolutionally conserved proteins, eIF4A and Dbp5/Rat8p but far from Vasa, which functions in germline development (Nakamura, 2001).
To examine distribution of the endogenous Me31B, antibodies were generated that specifically recognized Me31B. The distribution pattern of endogenous Me31B is identical to that of GFP-Me31B. No detectable signal in somatic follicle cells is observed at any stage of oogenesis. Me31B is first detected at a low level in germarium region 2B, where the signal is concentrated in the pro-oocytes. The signal remains concentrated in the oocyte until mid-oogenesis. In early egg chambers, a low level Me31B signal is detected in nurse cell cytoplasm. In both nurse cells and oocytes, the signal appears to be granular. Me31B signals in nurse cell cytoplasm become more evident from stage 5-6, when Me31B expression is drastically increased. In addition, Me31B is frequently enriched around nurse cell nuclei. Later, Me31B accumulates at the posterior pole of stage 10 oocytes. However, this posterior accumulation is transient, as revealed by uniform distribution of the signal in cleavage embryos. By cellular blastoderm stage, Me31B becomes undetectable in the entire embryonic region. No zygotic expression of Me31B was detected during embryogenesis (Nakamura, 2001).
Because Me31B is probably an RNA-binding protein that is transported to the oocyte, it was asked whether Me31B forms a complex with oocyte-localizing RNAs. Colocalization of OSK mRNA with Me31B was examined. OSK mRNA starts to accumulate in oocytes from germarium region 2B, with the concentration of OSK increasing over time. Posterior accumulation of OSK mRNA in the oocyte begins from stage 8 onwards. By fluorescent in situ hybridization, OSK mRNA exhibits particulate signals in the cytoplasm of both nurse cells and oocytes, and is frequently concentrated around nurse cell nuclei. This distribution pattern of OSK mRNA is essentially identical to that of Me31B, with colocalization present until OSK mRNA localizes to the posterior pole of stage 10 oocytes (Nakamura, 2001).
Colocalization of Me31B with other RNAs was also examined. Ovaries were doublestained for Me31B and BicD mRNA. In early egg chambers, BicD mRNA also produces particulate signals, and appears to localize in Me31B-containing particles. This colocalization becomes apparent from stage 5-6, when BicD mRNA expression is elevated. The oocyte-localizing RNAs examined [BCD, NOS, Oo18 RNA-binding (ORB), Polar granule component (PGC) and Germ cell-less (GCL)] all produce particulate signals in the cytoplasm of both nurse cells and oocytes, and colocalize with Me31B. In contrast, Vasa mRNA, which is not specifically transported to the oocyte, does not appear to be colocalized with Me31B. These results indicate that Me31B forms cytoplasmic particles that contain oocyte-localizing RNAs (Nakamura, 2001).
The complicated and redundant phenotypes observed in me31B- egg chambers in mid-oogenesis are unlikely to be the primary effect of loss of me31B function. Earlier phenotypes of me31B- egg chambers were examined using a FLP/FRT system to generate homozygous germline clones that are marked by the loss of Vas-GFP fusion protein. Based on Hoechst and phalloidin staining, me31B- egg chambers are morphologically normal until stage 4-5. From stage 6 onwards, oocytes in me31B- egg chambers fail to grow normally. At this stage, these egg chambers begin to degenerate. In early me31B- egg chambers, Exu signal is concentrated to the oocytes. Distributions of OSK and BicD mRNAs in me31B- egg chambers were examined. Both OSK and BicD mRNAs also accumulate in the oocytes of me31B- egg chambers until the chambers degenerate. Particulate signals for these RNAs are detectable in nurse cell cytoplasm in these egg chambers. These results indicate that Exu, OSK and BicD mRNAs can be transported to the oocyte even in the absence of Me31B. It is concluded that in early egg chambers, Me31B is dispensable for the transport of the molecules that form a complex with Me31B (Nakamura, 2001).
Whether loss of Me31B affects translation of OSK and BicD mRNAs was examined. Ovaries were immunostained with an anti-Osk antiserum. Although OSK mRNA is expressed during almost all stages of oogenesis, its translation is repressed to keep Osk protein level very low during early oogenesis. In me31B- egg chambers, Osk signal is significantly increased compared with that in the neighboring me31B+ egg chambers (Nakamura, 2001).
A similar increase of BicD signal in me31B- egg chambers is more evident. In wild-type egg chambers, BicD protein, like BicD mRNA, is highly concentrated in the oocytes starting from germarium region 2B. In the egg chambers lacking me31B, increased BicD signal is detected in nurse cell cytoplasm. These results suggest that loss of Me31B in germline cells causes derepression of OSK and BicD mRNA translation during their transport to the oocyte (Nakamura, 2001).
Cup has been identified as a component of an eight-protein complex that contains oskar mRNA (Wilhelm, 2000). cup is also required for oskar mRNA localization and is necessary to recruit the plus end-directed microtubule transport factor Barentsz to the complex. eIF4E is localized within the oocyte in a cup-dependent manner and binds directly to Cup in vitro. Thus, Cup is a translational repressor of oskar that is required to assemble the oskar mRNA localization machinery. It is proposed that Cup coordinates localization with translation (Wilhelm, 2003).
Because btz mutants display a late stage oskar mRNA localization defect similar to that of cup mutants (van Eeden, 2001), the effect of cup mutants on the distribution of Btz was examined. Normally, Btz protein is present on the nuclear envelope in nurse cells and colocalizes with oskar mRNA in the oocyte. However, in cup1/cup4506 egg chambers, the accumulation of Btz protein within in the oocyte is greatly reduced from stage 1 onward, whereas the Btz present on the nuclear envelope in the nurse cells is unaffected. The failure in the transport of Btz to the oocyte is not due to a general defect in assembly of the oskar RNP since cup1/cup4506 egg chambers localize Yps and oskar mRNA normally during early oogenesis. Thus, Cup is specifically required to localize Btz to the oocyte. This result, together with the findings that Cup and Btz colocalize as well as share similar oskar mRNA localization defects, argues that cup mutants fail to localize oskar mRNA because Cup is required to recruit Btz to the complex (Wilhelm, 2003).
Since all mutations isolated to date that disrupt oskar mRNA localization also block oskar translation, the role of cup in oskar translation was examined. Surprisingly, Oskar protein accumulated prematurely in the oocyte during stages 6 and 7 in cup1/cup4506 egg chambers, indicating that cup is required to translationally repress oskar mRNA during these stages. It is also worth noting that in cup mutants accumulation of Oskar protein was observed at only those sites where oskar mRNA is most enriched. This may be due to the fact that the cup alleles used in this study are hypomorphic alleles. The effects of cup are specific for oskar mRNA since the localized translation of gurken mRNA at the dorsal anterior region of the oocyte during stage 9 is unaffected in a cup1/cup4506 mutant background. Thus, cup is not a general translational regulator of localized messages (Wilhelm, 2003).
To better understand the role of Cup in maintaining the translational repression of oskar mRNA, attempts were made identify components of the translation machinery that were present in the complex by testing likely candidates. Immunoprecipitation of GFP-Exu and Yps show that eIF4E, the 5' cap binding component of the translation initiation complex, is specifically associated with these components of the oskar RNP complex. eIF4E and other components of the translation initiation machinery are generally thought of as being homogenously distributed due to their critical role in translation throughout the cell. Surprisingly, eIF4E is localized in a dynamic pattern within the oocyte. eIF4E is localized to the posterior of the oocyte early in oogenesis during stages 1-6. At stages 7 and 8, eIF4E redistributed to the anterior of the oocyte, and during stages 9 and 10, eIF4E accumulated at the posterior of the oocyte. This pattern of localization was also observed with a GFP-eIF4E protein trap line. Thus, eIF4E localizes in a temporal-spatial pattern identical to that of Cup, suggesting that it is a component of the complex in vivo (Wilhelm, 2003).
Since Cup is required for the correct localization of Btz to the oocyte, whether Cup is required for eIF4E localization was investigated. Immunostaining of cup1/cup4506 mutant egg chambers reveals that Cup is required for localization of eIF4E to the posterior of the oocyte from stage 1 onward. Disruption of cup function does not significantly affect the level of unlocalized eIF4E, indicating that the defect is primarily in the recruitment of eIF4E to the complex (Wilhelm, 2003).
Because Cup shares limited homology with 4E-T, a known eIF4E binding protein and a translational repressor in mammals, whether Cup binds to eIF4E was tested using a two-hybrid interaction assay. This assay showed a direct interaction between Cup and eIF4E. Cup interacted equally with both isoforms of eIF4E. Deletion analysis of Cup using the two-hybrid assay identified an eIF4E interaction domain that contains a canonical eIF4E binding motif. This motif is found in eIF4G as well as translational repressors (e.g., 4E-T) that block translation by preventing the eIF4E-eIF4G interaction. Thus, Cup is an eIF4E binding protein that acts directly to repress oskar translation (Wilhelm, 2003).
Thus, the assignment of Cup as a novel component of the oskar RNP complex is based on a number of findings: (1) Cup copurifies with both Exu and Yps, which have both been shown to be in a biochemical complex with oskar mRNA; (2) Cup protein exhibits the same dynamic localization pattern as that seen for oskar mRNA as well as other components of the complex; (3) Cup colocalizes with Yps and Btz particles, indicating that this these proteins form a complex in vivo; (4) the relevance of the biochemical association is supported by genetic studies of cup function, demonstrating a role for cup in translational repression of oskar mRNA as well as recruitment of Btz and eIF4E to the RNP complex (Wilhelm, 2003).
Because Cup is a translational repressor that is also required to assemble the oskar mRNA localization machinery, it is proposed that the coupling between localization and translation occurs by regulating these two functions of Cup. In this model, Cup is required early in the assembly of the transport complex in order to recruit components, such as Btz, that will later be used to dock to kinesin. This is consistent with the results that cup is required to localize Btz to the posterior pole and that cup mutants exhibit oskar mRNA localization defects comparable to those observed in btz mutants. The fact that mammalian Btz and 4E-T are nucleocytoplasmic shuttling proteins suggests that the defect in particle assembly in cup mutants may occur in the nucleus rather than in the cytoplasm. However, further studies will be necessary to determine the site of assembly (Wilhelm, 2003).
Because Btz is normally part of the transport complex throughout oogenesis even though it is only required for the kinesin-mediated transport step during stages 9 and 10, it is further proposed that the complex undergoes rearrangement in order to activate Btz and switch from minus end-directed transport to kinesin-mediated transport. Since the direct binding of Cup to Btz or Btz to kinesin has not yet been established it is unclear how many components of the complex may be involved in this reorganization (Wilhelm, 2003).
Once the complex reaches the posterior pole, it is argued that the localization machinery is disassembled and the interaction between Cup and eIF4E is broken to allow translational activation. Because Cup is stably maintained at the posterior pole after stage 9, whereas Btz is not, it is proposed that the trigger that disrupts the binding of Cup to eIF4E also leads to partial disassembly of the localization machinery via Cup. The molecular trigger for such rearrangements is unknown, however, the ability of 4E-T to bind eIF4E is regulated by phosophorylation (Pyronnet, 2001). Studies directed at identifying regulators of the Cup-eIF4E interaction might lead to greater mechanistic insights into the coupling mechanism (Wilhelm, 2003).
One of the attractive features of this model is that it suggests how coupling might be accomplished in other systems. Recent work in neurons on the translational regulator CPEB suggests that it can promote the transport of mRNA into dendrites. Since CPEB represses translation by recruiting the eIF4E binding protein, maskin, to transcripts, it is possible that the observed transport effect is due to a requirement for maskin to assemble the localization machinery. Thus, Cup may be representative of a general class of eIF4E binding proteins whose role is to couple mRNA localization to translational activation (Wilhelm, 2003).
Translational control is a critical process in the spatio-temporal restriction of protein production. In Drosophila oogenesis, translational repression of oskar1 (osk) RNA during its localization to the posterior pole of the oocyte is essential for embryonic patterning and germ cell formation. This repression is mediated by the osk 3' UTR binding protein Bruno (Bru), but the underlying mechanism has remained elusive. An ovarian protein, Cup, is required to repress precocious osk translation. Cup binds the 5'-cap binding translation initiation factor eIF4E through a sequence conserved among eIF4E binding proteins. A mutant Cup protein lacking this sequence fails to repress osk translation in vivo. Furthermore, Cup interacts with Bru in a yeast two-hybrid assay, and the Cup-eIF4E complex associates with Bru in an RNA-independent manner. These results suggest that translational repression of osk RNA is achieved through a 5'/3' interaction mediated by an eIF4E-Cup-Bru complex (Nakamura, 2004).
In a search for new components of the oskar RNP complex, this study identified the 147-kD protein of this complex as the product of the female sterile gene cup. Surprisingly, cup is required both for translational repression and localization of oskar mRNA. Cup was found to bind to eukaryotic initiation factor 4E (eIF4E) and is necessary to recruit the localization factor Barentsz to the complex. Thus, Cup is a translational repressor of oskar that is required to assemble the oskar mRNA localization machinery. Because of its interactions with both the localization and translational control complexes, it is proposed that Cup is a likely regulatory target for the coupling machinery (Nakamura, 2004).
During localization, osk RNA forms cytoplasmic granules in both nurse cells and the oocyte. The granules contain several proteins, including the DEAD-box protein Maternal expression at 31B (Me31B), the Y-box protein Ypsilon schachtel (Yps), and Exuperantia (Exu). Genetic evidence has shown that Exu is involved in the proper localization of bcd and osk RNAs in oogenesis, although the molecular function of Exu remains unknown. Both Yps and Me31B are involved, directly or indirectly, in the translational silencing of osk RNA in oogenesis. Yps antagonizes Orb, a positive regulator of osk RNA localization and translation. In egg chambers lacking me31B, osk RNA is prematurely translated in early oogenesis (Nakamura, 2001). These data indicate that the granules are maternal ribonucleoprotein (RNP) complexes containing proteins required for both RNA localization and translational control. The complex is highly enriched in eIF4E and a germline protein, Cup. Cup is required to repress osk translation. Evidence is provided that Cup-mediated translational repression is achieved by preventing the assembly of the eIF4F complex at the 5' end of osk RNA, and that Cup acts together with Bru to repress osk translation (Nakamura, 2004).
To identify new proteins in the Me31B complex, ovarian extracts from wild-type females were immunoprecipitated on a preparative scale using an affinity-purified anti-Me31B antibody (α-Me31B). α-Me31B specifically coprecipitatesmany proteins from the extracts. Mass spectrometric analyses of these proteins revealed that both Exu and Yps, the known components in the Me31B complex (Nakamura, 2001), are present in the immunoprecipitates. The analyses also revealed that the 35 kDa protein was eIF4E and the 150 kDa protein is Cup, a germline-specific protein required for oogenesis. Cup is expressed from early oogenesis and present until the blastoderm stage of embryogenesis. Numerous cup alleles have been isolated as female sterile mutants, which show a wide range of phenotypes. However, the biochemical function of Cup has remained elusive (Nakamura, 2004).
To examine the association among Me31B, eIF4E and Cup in vivo, ovaries expressing a GFP-Me31B fusion protein were stained for eIF4E and Cup. The GFP-Me31B form cytoplasmic particles in the germline, and the distribution patterns of the fusion protein are indistinguishable from those of endogenous Me31B (Nakamura, 2001). α-eIF4E stains cytoplasmic particles that are positive for GFP-Me31B. This colocalization is observed throughout oogenesis. Cup colocalized with GFP-Me31B is also found throughout oogenesis. Thus, eIF4E, Cup, and Me31B all form a complex during oogenesis (Nakamura, 2004).
To better understand the interactions between Me31B, eIF4E, and Cup, ovarian extracts were immunoprecipitated by α-Me31B and α-eIF4E, and the precipitates were analyzed by Western blotting. α-Me31B coprecipitates eIF4E and Cup, and α-eIF4E coprecipitates Me31B and Cup, indicating that they all form a complex. However, in the presence of RNase during immunoprecipitation, α-Me31B fails to coprecipitate eIF4E or Cup. Thus, the Me31B-eIF4E and the Me31B-Cup interactions are indirect and probably mediated through RNA in the complex. In contrast, α-eIF4E coprecipitates Cup even in the presence of RNases, suggesting a direct interaction between eIF4E and Cup in vivo (Nakamura, 2004).
The interaction of Cup and eIF4E in vitro was studied using a GST pull-down assay. GST-eIF4E pulls down Cup synthesized in vitro. The association is unaffected by RNase. These results indicate that Cup associates with eIF4E in vitro and that the interaction is RNA independent (Nakamura, 2004).
The results show that Cup is an eIF4E binding protein that is involved in translational repression of osk RNA during oogenesis. The conserved YxxxxLφ motif in Cup is important for eIF4E binding and Cup and eIF4G are likely to bind the same surface of eIF4E. These results suggest that Cup competes with eIF4G for eIF4E binding, and hence inhibits translation initiation. CupΔ212 protein, which lacks the conserved eIF4E binding sequence, is unable to bind eIF4E in vivo, and fails to repress osk translation. These results strongly suggest that the Cup-eIF4E interaction is essential for the Cup-mediated repression of osk translation, although it is possible that other of Cup's functions are also affected in the cupΔ212 mutant. Furthermore, Cup was found to interact with Bru in a yeast two-hybrid assay and that the Cup-eIF4E complex associates with Bru in an RNA-independent manner. Based on these results, it is speculated that the Bru-mediated repression of osk translation is operated, at least in part, through the interaction with Cup, which binds eIF4E and prevents the eIF4E-eIF4G interaction at the 5′ end of osk RNA (Nakamura, 2004).
The Staufen-dependent localization of oskar mRNA to the posterior of the Drosophila oocyte induces the formation of the pole plasm, which contains the abdominal and germline determinants. In a germline clone screen for mutations that disrupt the posterior localization of GFP-Staufen, three missense alleles were isolated in the hnRNPA/B homolog, Hrp48 (termed Hrb27C in FlyBase) These mutants specifically abolish osk mRNA localization, without affecting its translational control or splicing, or the localization of bicoid and gurken mRNAs and the organization of the microtubule cytoskeleton. Hrp48 colocalizes with osk mRNA throughout oogenesis, and interacts with its 5′ and 3′ regulatory regions, suggesting that it binds directly to oskar mRNA to mediate its posterior transport. The hrp48 alleles cause a different oskar mRNA localization defect from other mutants, and disrupt the formation of GFP-Staufen particles. This suggests a new step in the localization pathway, which may correspond to the assembly of Staufen/oskar mRNA transport particles (Huynh, 2004).
Previous genetic approaches have shown that Stau, Barentsz, Mago Nashi, and Y14 are required for osk mRNA localization, and colocalize with it to the posterior pole, strongly suggesting that these proteins are components of the mRNA localization complex. However, none of these proteins have been shown to bind directly to osk mRNA, although this seems likely to be the case for the RNA binding proteins, Stau and Y14.
Mago, Y14, Barentsz, and eIF4AIII are part of the exon-exon junction complex (EJC) that marks where introns have been excised. This suggests that the EJC is loaded onto osk mRNA during splicing in the nucleus to control its localization in the oocyte cytoplasm. Although Stau and the EJC are thought to play a direct role in coupling osk RNA to the factors that localize it, none of these proteins has been shown to bind specifically to any sequence elements within osk mRNA, and it is unclear how the RNA is recognized (Huynh, 2004 and references therein).
Biochemical strategies have led to the identification of Bruno, Apontic, p68, and p50 as proteins that bind to specific sequences in osk RNA, all of which have been implicated in the regulation of osk mRNA translation but not in the localization of the mRNA. The results described in this study provides a link between these two approaches, by demonstrating that p50 corresponds to Hrp48 and that it is specifically required for the transport of the mRNA to the posterior of the oocyte. Germline clones of the three missense alleles of hrp48 have no effect on the polarity of the oocyte, the organization of the microtubules, or the localization of bicoid or gurken mRNAs, but abolish the posterior localization of osk mRNA (Huynh, 2004).
Since Hrp48 has been shown to regulate alternative splicing, it is possible that the oskar mRNA localization phenotype of the hrp48linha mutations is an indirect consequence of a defect in the splicing of another mRNA that encodes a factor that is directly involved in osk mRNA transport. However, this is unlikely to be the case for two reasons. (1) These missense alleles have no effect on the alternative splicing of Ubx transcripts. Since the insertion alleles of Hrp48 do disrupt the Ubx splicing pattern, the missense alleles do not appear to impair the function of Hrp48 in splicing regulation. (2) Hrp48 binds to sequences in the 5′ region and the 3′UTR of osk mRNA, and colocalizes with the mRNA at the posterior. This strongly suggests that the requirement for Hrp48 in osk mRNA localization is direct, and that it functions as an essential trans-acting factor that recognizes the RNA and plays a role in coupling it to the localization machinery (Huynh, 2004).
The phenotype of the hrp48 missense alleles differs from that of the other mutants that disrupt osk mRNA localization, suggesting that it acts at a distinct step in the localization pathway. In stau, barentsz, mago, Y14, and tropomyosinII mutants, osk mRNA also fails to reach the posterior, but most of the RNA remains at the anterior cortex. Since osk mRNA shows a transient accumulation at the anterior in wild-type before it localizes to the posterior, these proteins may be required to release osk mRNA from the anterior, and to couple it to the posterior transport pathway. In contrast, no accumulation of osk mRNA at the anterior of the oocyte is detected in the hrp48 missense mutants, and the mRNA shows a uniform distribution throughout the oocyte cytoplasm. This raises the possibility that Hrp48 is required for the transient anterior accumulation of osk mRNA, and acts upstream of the other proteins required for posterior localization, such as Stau. One argument against this interpretation is that the localization to the anterior of the oocyte is thought to be a by-product of the transport from the nurse cells into the oocyte, which occurs normally in the hrp48 missense alleles. In support of this view, all mRNAs that are transported into the oocyte also localize, at least transiently, to the anterior cortex at stage 9. Therefore an alternative model is favored in which Hrp48 acts downstream of Stau, Mago, etc. In this case, the stau class of mutants might block the release of osk mRNA from the anterior cortex, whereas the hrp48 alleles might stimulate this release, but prevent the subsequent association of the mRNA with the factors that transport it to the posterior pole (Huynh, 2004).
Many localized mRNAs appear to move as large cytoplasmic particles, leading to the suggestion that they must be packaged into transport granules in order to be localized. For example, when pair rule, wingless, or bicoid mRNAs are injected into Drosophila embryos, they assemble into particles that move in a microtubule-dependent manner, and fluorescent MBP mRNA shows a similar behavior when introduced into cultured oligodendrocytes. The formation of these transport particles has also be visualized by labeling proteins that bind to localized mRNAs. For example, injected bicoid mRNA recruits Stau into motile particles in Drosophila syncytial blastoderm embryos, while GFP-tagged mouse Stau1 has been observed to form particles that move along microtubules in neuronal processes. In this context, it is very striking that two of the hrp48 missense mutants strongly reduce the formation of GFP-Stau particles in the nurse cell and oocyte cytoplasm. This suggests that Hrp48 plays a role in the formation of Stau-containing osk mRNA transport particles, which could account for the failure to localize the mRNA to the posterior pole in these mutants. If these particles sequester the mRNA and prevent its diffusion, this may also explain why osk mRNA remains at the anterior in the stau class of mutants, but not in the hrp48 missense alleles (Huynh, 2004).
Hrp48 is one of the three most abundant HnRNPs in Drosophila, along with Hrp40 (Squid) and Hrp38, and is thought to bind most, if not all, nascent transcripts in the nucleus. It is therefore surprising to recover hrp48 mutations that have such a specific effect on the localization of osk mRNA. P element insertions in hrp48 produce a distinct phenotype. Although osk mRNA is often not localized to the posterior, it is sometimes found in the center of the oocyte, which is indicative of a defect in oocyte polarity. Consistent with this, Kin-ß-gal also localizes to the center of the oocyte in these mutants, whereas it always shows a wild-type posterior localization in the missense alleles. Thus, the P element insertions presumably disrupt the regulation of another mRNA that is required for the polarization of the oocyte microtubule network. A second important difference is that the P element alleles cause the premature translation of osk mRNA, which is consistent with the identification of Hrp48 as p50, which binds to sites in the osk BREs. In contrast, the missense alleles have no discernable effect on translational repression, since osk mRNA is completely unlocalized, but no Osk protein is produced. P alleles of Hrp48 also affect gurken mRNA localization and translation, whereas no defects are observed in the distribution of either gurken mRNA or protein in germline clones of the missense alleles. Finally, the P alleles, but not the missense alleles, disrupt the alternative splicing of Ubx RNAs. These differences presumably reflect the distinct nature of the molecular lesions in the two types of allele. All of the P element insertions fall in the promoter or an intron in the 5′UTR of the hrp48 gene, and produce reduced levels of wild-type Hrp48 protein. In contrast, each of the missense alleles expresses approximately normal levels of mutant Hrp48, in which a single amino acid has been changed (Huynh, 2004).
The comparison between the phenotypes of the two classes of hrp48 mutations indicates that the missense alleles affect regions of the protein that are specifically required for osk mRNA localization. This is particularly interesting in the case of the two Trp to Asn mutations in the Glycine-rich domain (GRD), since this domain is not involved in RNA binding, and neither mutant affects the interaction of Hrp48 with osk mRNA. Evidence from other HnRNPA/B family members suggests that this region functions as an oligomerization domain. For example, the GRD of Human HnRNP A1 has been shown to self-associate in vitro, and this interaction depends on large aromatic residues embedded with the Glycine-rich region. Thus, it is possible that Hrp48 also oligomerizes, and that this is disrupted by mutating the aromatic Tryptophans in the GRD, which could explain why both mutations impair the formation of GFP-Stau particles. These mutations may therefore abolish RNA localization because Hrp48 oligomerization is required to form high-order osk RNP complexes, which are the substrate for posterior transport (Huynh, 2004).
It is more difficult to explain why the 5B2-6 allele disrupts osk mRNA localization. Although one would expect a mutation in a conserved residue in the RNA recognition motif to disrupt RNA binding, the mutant protein binds to osk mRNA in vitro as well as the wild-type protein, in both UV-crosslinking and pull-down assays. Furthermore, it presumably also associates with osk mRNA in vivo, because it mediates normal translational repression, which requires binding to the sites in the BREs. The Hrp48 binding sites in osk mRNA that are necessary for localization have not been mapped, and these sites could be distinct from those involved in translational repression. Thus, this mutation may only disrupt Hrp48 binding to a specific subset of its target sites, including unidentified sites in osk mRNA that are required for localization. Since the glycine that is mutated is not directly involved in RNA binding, another possibility is that the mutation disrupts the interaction of Hrp48 with another protein that is required to couple osk mRNA to the localization machinery (Huynh, 2004).
Two other members of the hnRNP A/B family have also been implicated in mRNA localization. Drosophila Squid, which is most closely related to hnRNPA1, binds directly to gurken mRNA, and is required for its localization to the anterior-dorsal corner of the oocyte. Squid is also required to repress the translation of unlocalized grk mRNA, and interacts with Bruno, a translational repressor of both grk and osk mRNAs. Since Hrp48 binds to the Bruno Response elements in osk mRNA, and is required to repress its premature translation, these two Drosophila hnRNP A/B proteins perform remarkably similar functions in the regulation of osk and gurken mRNAs. Furthermore, hnRNP A2, which is one of the closest mammalian homologs of Drosophila Hrp48, plays a comparable role in the localization of MBP mRNA in oligodendrocytes. When MBP mRNA is injected into oligodendrocytes, it forms large particles that move along microtubules into the distal processes, and, like osk, the mRNA is probably transported by kinesin. This localization requires the specific binding of hnRNP A2 to a 21 nt A2RE element in the MBP 3′UTR that is necessary and sufficient for localization and efficient translation. In addition, recent work has shown that the microtubule-dependent localization into oligodendrocyte processes is mediated by the second RRM of hnRNP A2, and it is intriguing that the hrp48 5B2-6 mutation falls in the equivalent domain of the Drosophila protein. These clear parallels between the functions of Hrp48, Squid, and hnRNP A2 suggest that these hnRNP A/B proteins play a conserved role in mRNA localization and translational control in flies and mammals (Huynh, 2004 and references therein).
Hrp48 is a predominantly nuclear protein that associates with nascent transcripts, and regulates alternative pre-mRNA splicing. It therefore seems very likely that Hrp48 binds to osk mRNA in the nucleus, and is exported into the cytoplasm with the RNA, where it regulates localization and translation. This adds to the growing body of evidence that the cytoplasmic fate of mRNAs is determined in part by their nuclear history. In the case of osk mRNA, cytoplasmic localization seems to require the binding of at least four distinct proteins in the nucleus; Hrp48, which probably associates with the RNA cotranscriptionally, and Mago, Y14, and eIF4AIII, which are presumably recruited to the RNA during splicing as part of the exon-exon junction complex. The mRNA must then recruit additional essential localization factors in the cytoplasm, such as Stau and Barentsz, before it is competent to localize to the posterior pole of the oocyte. Thus, the assembly of a functional osk RNA localization complex requires the stepwise recruitment of multiple nuclear and cytoplasmic proteins, all of which are essential for posterior localization. One of the main challenges for the future will be to determine how this complex is assembled, and how these proteins act together to link the mRNA to the motor that mediates its transport to the posterior pole (Huynh, 2004).
Establishment of the Drosophila embryonic axes provides a striking example of RNA localization as an efficient mechanism for protein targeting within a cell. oskar mRNA encodes the posterior determinant and is essential for germline and abdominal development in the embryo. Tight restriction of Oskar activity to the posterior is achieved by mRNA localization-dependent translational control, whereby unlocalized mRNA is translationally repressed and repression is overcome upon mRNA localization. The oskar RNA binding protein p50 has been identified as Hrp48, an abundant Drosophila hnRNP. Analysis of three hrp48 mutant alleles reveals that Hrp48 levels are crucial for polarization of the oocyte during mid-oogenesis. These data also show that Hrp48, which binds to the 5' and 3' regions of oskar mRNA, plays an important role in restricting Oskar activity to the posterior of the oocyte, by repressing oskar mRNA translation during transport (Yano, 2004).
HnRNPs are abundant RNA binding proteins involved in many aspects of mRNA regulation. HnRNPs associate with transcripts at their site of synthesis in the nucleus, and can remain associated with the RNA during processing and export, as well as during cytoplasmic processes such as translation and localization. Mammalian hnRNP A2 binds to specific sequences in the 3′UTR of myelin basic protein mRNA and mediates its localization and translational control in rat oligodendrocytes. In Xenopus oocytes, VgRBP60, an hnRNP I-related protein, colocalizes with and binds sequence elements in Vg1 mRNA that are critical for its localization at the vegetal pole. In the Drosophila oocyte, two isoforms of Squid/hrp40 have essential roles in localization and translational control of gurken mRNA, the dorso-ventral polarity determinant (Yano, 2004).
Drosophila Hrp48, a member of the hnRNPA/B family of proteins, is an abundant hnRNP, bearing two N-terminal RRM-type RNA binding domains and a C-terminal Glycine-rich domain. Hrp48 is expressed in somatic and germline cells of the ovary, where it is detected at low levels in the nucleus and predominantly in the cytoplasm. Mammalian hnRNP A1, a putative homolog of hrp48, has been shown to function in splice-site selection and to shuttle between the nucleus and the cytoplasm. Hrp48 is a cofactor in regulation of alternative splicing . Mutations reducing hrp48 expression cause developmental defects and lethality, indicating that hrp48 is essential in the fly (Yano, 2004).
Although hrp48 affects oskar mRNA regulation, it does not appear to regulate oskar mRNA processing, because the mRNA is accurately spliced in the mutants. The hrp48 mutants analyzed also show defects in the organization of the oocyte microtubules. These defects are due to low levels of Hrp48 in the mutants, since they are suppressed by expression of hrp48 from a transgene. Given the demonstrated involvement of Hrp48 in RNA splicing, localization, and translational control, it is likely that polarity defects in hrp48 mutants are caused by deregulation of RNAs whose products are required for oocyte polarization. The dumping defects of some hrp48 mutant egg chambers suggest that Hrp48 may also regulate the function of the actin cytoskeleton. However, the oskar-related defects observed in the mutants do not resemble anchoring defects: their onset occurs before stage 10, when cytoplasmic streaming commences, and they decrease in severity as streaming proceeds. The difference in phenotype of the mutants analyzed, in which the level of expression of wild-type Hrp48 protein is reduced, causing polarity defects, and those identified by Huynh (2004), which express mutant Hrp48 proteins at wild-type levels but show normal oocyte polarity, highlights the importance of the threshold of expression of Hrp48 and the involvement of the protein in multiple RNA regulatory events (Yano, 2004).
The only protein shown previously to bind to oskar mRNA directly and to repress its translation during transport is Bruno. Null alleles of arrest fail to develop beyond the early stages of oogenesis, and hypomorphic alleles that allow development until stage 9 show no precocious translation of oskar mRNA. The evidence that Bruno is an oskar translational repressor in vivo is extensive and comes from analysis of transgenes in which Bruno binding sites were mutated, from the cuticle phenotype of embryos in which arrest gene dosage was varied in a sensitized genetic background, and from overexpression of Bruno in the germline, which causes the development of embryos with a posterior group phenotype (Yano, 2004).
Several lines of evidence support a role of Hrp48 in oskar translational repression. (1) A substantial portion of embryos developing from hrp48K16203 germline clones show A/P patterning defects consistent with misregulation of posterior patterning activity: reduced abdominal segmentation, presumably due to defects in oskar mRNA localization, and head defects. (2) An enhancement of the head defects caused by overexpression of the oskar 3′UTR is observed when hrp48 levels are reduced, and in extreme cases embryos with a bicaudal phenotype develop. (3) Translational derepression of the lacZ-oskar translational reporter is observed in hrp48 heterozygous females. The absence of detectable Oskar protein in the cytoplasm of heterozygous or homozygous hrp48 mutant oocytes during the early stages of oogenesis suggests that the lacZ reporter is more sensitive than the antibody, and thus the extent to which oskar is derepressed in the mutants is unclear (Yano, 2004).
Bruno binds to the A, B, and C regions of the oskar 3′UTR, each of which contains a pair of U(G/A)U(A/G)U(G/A)U sequence elements defining the BRE consensus. Mutations reducing binding of Bruno to BRE region A in vitro cause precocious reporter translation in the nurse cells and in the oocyte, indicating a role of Bruno in oskar repression in both cell types. Hrp48 binds the A, B, and C regions in vitro and represses oskar translation in vivo. Sequences with homology to the characterized Hrp48 binding site in P element transposase mRNA, the F2 site (UAGGUUAAG), are located in the oskar 5′ translation regulatory region, and within 10 nucleotides of the BREs in regions A and B, and overlapping with the BRE in region C in the 3′UTR. Deletion of these F2-like sequences from an AB region probe selectively disrupts Hrp48 binding in vitro, without affecting Bruno binding. An oskar-lacZ reporter transgene bearing these deletions shows precocious translation in the oocyte, but not in the nurse cells. Hence, the F2-like elements in regions A and B of the 3′UTR may mediate the repressive effect of Hrp48 on translation of unlocalized oskar mRNA in the oocyte. Proof that the repressive effect of Hrp48 on oskar translation is mediated by binding of the protein to these elements will require a complete mutational analysis, as has been performed for Bruno, resulting in the definition of the BREs (Yano, 2004).
Interestingly, when the region to which Hrp48 binds at the 5′ end of oskar mRNA is deleted in the context of the BRE A LS5 mutation, an increase in precocious translation both in the nurse cells and in the oocyte is observed, indicating a function of the 5′ region in translational repression. The recent demonstration that mutations in Cup, an eIF4E binding protein that coprecipitates with Bruno in ovarian extracts, cause precocious oskar translation suggests that translational repression may occur at the initiation step. The fact that sequences at the 5′ and 3′ ends of oskar mRNA regulate localization-dependent translation, together with the observation that Hrp48 can homodimerize in a yeast two-hybrid assay raises the possibility that Hrp48 may promote circularization of oskar mRNA, and thus facilitate Cup-mediated repression (Yano, 2004).
Hrp48 binds oskar mRNA directly and regulates both its localization and translation, suggesting that these processes might be coupled and mediated by a bifunctional mRNA localization/repression complex comprising Hrp48. It is not known at what stage Hrp48 first associates with oskar mRNA. However, the similar distributions of Hrp48 and oskar mRNA in the oocyte and the abundance of Hrp48 in the cytoplasm suggest their association in the cytoplasm (Yano, 2004).
How mRNA localization/translation complexes are assembled and associate with the transport machinery are central questions. The involvement in oskar mRNA localization of the EJC components Mago-Nashi and Y14, which associate with RNAs at exon-exon junctions upon splicing, indicates that assembly of the oskar mRNA localization complex begins in the nucleus. The fact that all the exon-exon junctions in oskar mRNA are located in the coding region and that distinct regions of the oskar 3′UTR are required for mRNA localization indicates the involvement of a higher order complex in oskar mRNA localization. In this context, it is interesting to speculate that Hrp48 molecules bound to distinct regions of oskar mRNA may promote the association of several mRNAs and associated proteins into large transport-competent particles. The perinuclear localization of Hrp48, Barentsz, Bruno, and Cup may indicate that they associate with the mRNA at the time of its nuclear-cytoplasmic transport (Yano, 2004).
During mid-oogenesis, oskar mRNA and associated proteins are localized in a microtubule-dependent manner to the posterior pole of the oocyte, where the plus ends of the oocyte microtubule are focused. In Kinesin heavy chain mutants, oskar mRNA localization fails and the RNA remains all along the cell cortex, indicating a role of Kinesin heavy chain in polarized transport of the oskar mRNP. However, it is still unclear how the oskar mRNP associates with the motor, as Kinesin light chain is not required for oskar mRNA localization in the oocyte. Thus, future work will be required to determine the molecules and cellular mechanisms whereby nuclear-associated proteins, such as the EJC components, Barentsz, and Hrp48, assemble an oskar mRNA complex competent for transport to the posterior pole of the Drosophila oocyte (Yano, 2004).
The appearance of Oskar protein occurs coincident with localization of oskar mRNA to the posterior pole of the Drosophila oocyte, and earlier accumulation of the protein is prevented by translational repression. The nascent polypeptide-associated complex (NAC) is required for correct localization of oskar mRNA. The timing of the defects suggests that, if NAC acts directly via an interaction with nascent Oskar protein, oskar mRNA should be undergoing translation prior to its localization. Polysome analysis confirms that oskar mRNA is associated with polysomes even in the absence of localization of the mRNA or accumulation of Oskar protein. Thus, the mechanisms that prevent accumulation of Oskar protein until it can be secured at the posterior pole of the oocyte include regulated degradation or inhibition of translational elongation (Braat, 2004).
Multiple RNA binding proteins are required for oskar repression, including Bruno (Bru), Bicaudal C (BicC), p50, Apontic (Apt), and Me31B, and at least a subset of these bind to a part of the osk 3' untranslated region (UTR) that contains Bruno Response Elements (BREs). Additional factors that are candidates to act in translational regulation of osk mRNA are identified by mutants with a bicaudal phenotype, since this phenotype is frequently associated with inappropriate expression of Osk. The original bicaudal (bic) locus encodes the ß subunit of the nascent polypeptide-associated complex (NAC; Markesich, 2000). Intact NAC, which contains an additional subunit, alpha, acts to prevent inappropriate interactions of the nascent peptide with signal recognition particle (SRP) and to ensure that nonsecretory proteins are not targeted to the endoplasmic reticulum (ER) (Braat, 2004).
Reduction of either alpha or ß NAC expression has an identical effect on localization of osk mRNA and accumulation of Osk protein during oogenesis. If the NAC action is direct and entails an interaction with nascent Osk peptide, osk mRNA should be associated with ribosomes at a time and place when it is thought to be translationally repressed. Indeed, by polysome analysis it has been found that a significant fraction of osk mRNA is associated with polysomes, even when Osk protein accumulation is dramatically reduced. Thus the apparent repression of osk mRNA translation includes, at least in part, a regulatory event that occurs after the initiation of translation (Braat, 2004).
Data is presented implicating NAC in the proper localization of osk mRNA. Although each NAC subunit has been suggested to have additional but separate roles, the similar phenotypes of the two mutants argues that the reduction of NAC activity is the primary defect. How does NAC contribute to osk mRNA localization? The possibilities are most usefully divided into direct and indirect models (Braat, 2004).
If NAC acts directly, then the complex would interact with the nascent Osk polypeptide as it exits from the ribosome. The one strong prediction of any direct model is that osk mRNA be associated with ribosomes at the stage when the localization defects appear. It is not possible to directly test that prediction, but it was shown that osk mRNA is polysomal in a mutant that accumulates very little Osk protein. Thus the phenomenon of translational repression of osk mRNA need not act exclusively at the level of initiation of translation. Consequently, it is probable that osk mRNA is indeed associated with ribosomes at the developmental stage when reduction of NAC activity causes defects in osk mRNA localization (Braat, 2004).
Given the known activity of NAC in mammalian cells, where its interaction with the nascent peptide prevents inappropriate binding of SRP and introduction into the secretory pathway, a direct role for NAC in osk mRNA localization could be to prevent Osk protein from association with ER. No association of osk mRNA with ER was detected in the NAC mutants, but the binding of NAC to nascent polypeptides may govern or restrict a wider range of interactions than has been demonstrated. If so, the absence of NAC might allow the nascent Osk peptide to make a different type of inappropriate interaction, which could direct the osk mRNA on polysomes to the large discrete bodies observed in the ooplasm of the mutants. The bodies appear to migrate slowly to the posterior pole, suggesting that the osk mRNA contained in the bodies can still interact with the localization machinery (Braat, 2004).
An indirect role for NAC in osk mRNA localization would involve an effect on the expression or processing of a factor required for osk mRNA localization. This indirect role could require either known or novel activities of NAC. It is noteworthy that the mRNA localization defects observed are unlike any that have been described previously, making it unlikely that the phenotype arises from indirect effects via the known localization factors (Braat, 2004).
Translation of osk mRNA has been thought to begin only after it is localized to the posterior pole of the oocyte, but the results now indicate that translation can be initiated prior to localization. Thus there appears to be a system to degrade the nascent protein, and translational elongation may be dramatically inhibited or blocked. Evidence for control after initiation does not challenge the conclusion that Bru-dependent repression of osk translation occurs at the level of initiation, but instead reveals that regulation can occur at multiple stages. Since disruption of the Bru-dependent mechanism leads to excessive Osk accumulation (although the effect can be modest), the postinitiation form of control does not appear to be sufficient for repression. The use of two forms of control may facilitate the highly efficient restriction of Osk protein accumulation prior to the posterior localization of osk mRNA (Braat, 2004).
Any model for regulated Osk protein stability must address the fact that removal of 3′UTR regulatory elements allows the protein to accumulate at any position in the oocyte or the nurse cells. Thus the simplest model, that a stabilizing condition exists exclusively at the posterior pole of the oocyte, is quite unlikely as a sole mechanism (Braat, 2004).
An effect on Osk protein stability mediated by the 3′UTR could best be explained by two models. In the targeting model, a regulatory element in the 3′UTR would recruit a factor that targets the growing polypeptide for proteolysis: recruitment of the targeting factor to the mRNA would be allowed prior to localization, and the binding protein or degradation factor would be displaced from the mRNA or inactivated upon posterior localization. In the pausing model, a protein bound to a 3′UTR regulatory element would inhibit the elongation phase of translation, increasing the time during which incomplete forms of Osk protein are exposed during synthesis. The pausing model of course demands that incomplete Osk proteins be unstable, either because of the inherent instability of an Osk polypeptide that has not assumed the folding of the mature protein or because of the action of factors that specifically recognize the partial protein and target it for degradation. Thus both models rely on regulatory elements in the osk 3′UTR, and the pausing model also requires a specific property -- instability when incomplete -- of the protein encoded by the regulated protein. Similar models have been proposed for regulation of the nanos mRNA, which is also polysome-associated under conditions in which the Nanos protein does not accumulate (Braat, 2004).
In addition to osk mRNA, there are several other examples in which protein accumulation is prevented by a regulatory event that does not disrupt the assembly of polysomes on the transcript. In Drosophila, the nos mRNA displays this behavior, as do the lin-14 and lin-28 mRNAs of C. elegans. In no case has the mechanism responsible for preventing protein accumulation been elucidated. However, the C. elegans mRNAs are notable in that a trans-acting RNA, the lin-4 microRNA, mediates the effect. Recently, it has been shown that mutants defective in RNA interference lead to premature accumulation of Osk protein, as well as a variety of other defects. RNA interference could have an indirect action on translational control of osk mRNA. Alternatively, microRNAs might act directly, as in the translational control of lin-14 and lin-28. The demonstration that regulation of Osk accumulation occurs independent of polysome association strongly suggests a direct role. Proof will require demonstration of direct interactions of microRNAs with osk, and evidence that their selective disruption interferes with translational regulation (Braat, 2004).
Prior to reaching the posterior pole of the Drosophila oocyte, oskar mRNA is translationally silenced by Bruno binding to BREs in the 3' untranslated region. The eIF4E binding protein Cup interacts with Bruno and inhibits oskar translation. Validating current models, the mechanism proposed for Cup-mediated repression has been directly demonstrated: inhibition of small ribosomal subunit recruitment to oskar mRNA. However, 43S complex recruitment remains inhibited in the absence of functional Cup, uncovering a second Bruno-dependent silencing mechanism. This mechanism involves mRNA oligomerization and formation of large (50S-80S) silencing particles that cannot be accessed by ribosomes. Bruno-dependent mRNA oligomerization into silencing particles emerges as a mode of translational control that may be particularly suited to coupling with mRNA transport (Chekulaeva, 2006).
Tight restriction of Oskar protein to the posterior pole of the Drosophila oocyte is crucial for development of the future embryo and is largely achieved by posterior localization of oskar mRNA and its translational inhibition prior to localization. This molecular analysis of oskar mRNA translational repression and of the relative roles of Bruno and Cup in this process has demonstrated the existence of two distinct modes of repression by Bruno and their mechanistic basis. This study has demonstrated directly the mechanism hypothesized for Bruno/Cup function, whereby cap-dependent 43S complex recruitment is inhibited. It has also been discovered that Bruno exerts its function through a second mechanism that does not require functional Cup and its interaction with eIF4E. This mode of repression involves Bruno-dependent oskar mRNA oligomerization and assembly into silencing particles, unusually large RNPs in which oskar remains inaccessible to the translation machinery (Chekulaeva, 2006).
This analysis of ribosomal complexes assembled on oskar reporter mRNA in vitro revealed that 48S initiation complex formation is inhibited both in the presence and in the absence of Cup-eIF4E interaction. This result is compatible with either of two possible mechanisms: (1) inhibition of small ribosomal subunit recruitment and (2) blocking of the following step: scanning of the 5'UTR by the small ribosomal subunit. Indeed, such scanning complexes in which the 43S subunit moves along the mRNA searching for the initiation codon are not stable and can easily dissociate during centrifugation in the sucrose density gradient. Therefore, as with a failure in recruitment of the small ribosomal subunit, interfering with scanning would also result in a reduction of the 48S peak (Chekulaeva, 2006).
The first of the two oskar repression mechanisms requires the interaction of Cup and eIF4E. This Cup-dependent repression process also requires a m7GpppN cap on the mRNA. Since binding of the small ribosomal subunit represents the cap-dependent step in translation initiation, these results provide a direct demonstration of the hypothesized mechanism for Cup regulation of oskar mRNA: a block of cap-dependent 43S recruitment mediated by a functional interaction between Cup-eIF4E and Bruno. Interestingly, it was observed that Cup recruits eIF4E to the mRNA in a cap-independent manner, suggesting an unexpected role for Cup, over and beyond its role in translational repression. Recruitment of eIF4E to oskar mRNA complexes by Cup might ensure colocalization and local enrichment of this otherwise limiting translation factor at the posterior pole, where oskar mRNA is translationally activated (Chekulaeva, 2006).
The second mechanism of oskar regulation revealed by this analysis also involves Bruno but requires neither Cup-eIF4E interaction nor a m7GpppN cap. It is therefore unlikely that this mechanism directly interferes with cap-dependent recruitment of the 43S complex (Chekulaeva, 2006).
This analysis shows that repressed oskar reporter mRNA forms unusually heavy complexes sedimenting between the 48S and 80S peaks. Importantly, these complexes form in the absence of the Cup-eIF4E interaction and of ribosomal subunit binding, as revealed by their persistence upon addition of cap analog. It is therefore proposed that oskar mRNA is sequestered in such large RNP complexes and hence inaccessible to the 43S preinitiation complex. Consistent with such a sequestration hypothesis, the repressed mRNA is selectively protected from the degradation machinery. Interestingly, a model of 'masked' (translationally inactive, stable) mRNAs was put forward 40 years ago. Masking factors were proposed to bind to mRNA and promote aggregation into higher-order condensed particles, protected from any processive events, including translation, degradation and polyadenylation/deadenylation (Chekulaeva, 2006).
The current experiments reveal that assembly of oskar mRNA into RNP complexes as large as monoribosomes can occur without any involvement of the RNA with the ribosomal subunits. These findings shed an unexpected light on the published literature, where complexes of 80S and larger can be intuitively taken as an indication of ribosomal association and translation elongation. Based on the co-sedimentation of oskar mRNA with polysomes and experiments involving the polysome-disrupting agent puromycin, it has been concluded that in the ovary, repressed oskar mRNA is associated with translating ribosomes (Braat, 2004). The current data challenge this conclusion, because it is shown directly that heavy RNPs (up to 80S in vitro) can form on oskar reporter mRNA without ribosomal subunit binding. The Braat study employed experimental conditions in which more than one variable was simultaneously changed. Specifically, the Mg2+ concentration, which can affect both polysome and RNP stability, differed by an order of magnitude between the puromycin-treated samples (2.5 mM Mg2+) and the cycloheximide control (25 mM Mg2+). This experiment was repeated, altering only one variable (puromycin). When the Mg2+ concentration is kept constant, puromycin does not affect the heavy RNPs that were previously interpreted as being 'polysomal'. It is suggested that oskar mRNA is engaged in puromycin-insensitive, heavy silencing particles that are sequestered from ribosomal engagement and that cosediment with polysomes (Chekulaeva, 2006).
Remarkably, oskar silencing particles comprise not single mRNA molecules but mRNA oligomers, whose formation is dependent on the specific association of Bruno with the BREs. The fact that the same components, Bruno and BREs, are responsible for both translational repression and mRNA oligomerization into silencing particles suggests a causal relationship between oligomerization and translational silencing (Chekulaeva, 2006).
The interesting finding that Cup is present in the heavy but not in the light RNP peak highlights the role of silencing particles in oskar repression. The sucrose gradient analysis of repressed complexes in cupΔ212 extract demonstrates that Cup-4E interaction is not required for silencing particle formation. However, the fact that Cup is exclusively associated with the silencing particles but not with the light RNP peak of repressed mRNA suggests that particle formation may contribute not only to Cup-independent repression but also to Cup-dependent repression (Chekulaeva, 2006).
Consistent with the in vitro demonstration of oskar mRNA multimerization in silencing particles, it has been demonstrated that oskar mRNA molecules can self-associate through the 3′UTR for localization to the posterior pole of the oocyte. Since oskar mRNA is translationally repressed prior to posterior localization, it is tempting to speculate that the large silencing complexes that have been identified as containing oskar mRNA multimers are related to oskar mRNA localization complexes. It should be noted, however, that at present, there is no evidence for a role of the translational repressor Bruno in oskar mRNA localization. It is also possible that direct intermolecular RNA-RNA interactions might contribute to oskar oligomerization, as in the case of bicoid mRNA (Chekulaeva, 2006).
The current work suggests that silencing particles in Drosophila ovary extracts form by Bruno-mediated mRNA oligomerization from lower complexity precursors. Recent reports have described the presence of large particles, P bodies, in both yeast and mammalian cells. From these P bodies, silenced mRNAs may either return to the translating pool or be targeted for degradation. Furthermore, this work has suggested that RNP particles may aggregate from precursors into higher-order structures. In this regard, it is notable that both Cup and Me31B [Me31B has been implicated in translational regulation of Oskar mRNA during early oogenesis (Nakamura, 2001) and is a homolog of the S. cerevisiae P body component and translational repressor Dhh1p (Coller, 2005)] are present in silencing particles -- it has recently been shown that the mammalian eIF4E binding protein, 4E-T, and Dhh1p, the S. cerevisiae homolog of Me31B, are P body components. While the factors that promote P body aggregation in mammals and yeast are currently unknown, Bruno has been identified as a critical factor for silencing particle formation. Interestingly, this analysis shows that while Bruno is associated with the repressed mRNA both in silencing particles and lighter RNPs, Cup associates only with the mRNA in silencing particles. The fact that Bruno does not recruit Cup in the light RNP peak suggests that effectors may exist that regulate this interaction and cause RNP transition to silencing particles by addition/modification of factors and/or conformational change. It will be interesting to further explore the relationship between silencing particles and P bodies (Chekulaeva, 2006).
The exciting finding that oskar silencing particles comprise not single mRNA molecules, but mRNA multimers, suggests a mode of mRNA translational control that seems particularly suited to coupling of translational repression with mRNA transport within the cell. Such a repression mechanism would also allow coordinate repression of multiple oskar mRNAs, as well as coordinate derepression of the mRNAs within the silencing mRNP, upon its localization at the oocyte posterior pole. The particles could in principle contain other RNAs regulated and assembled into RNPs by common components. It will be interesting to determine if gurken mRNA, which is translationally repressed by Bruno (but not Cup) and colocalizes with oskar mRNA during the early stages of oogenesis, is coassembled with oskar mRNA in silencing particles (Chekulaeva, 2006).
Zip code-binding protein 1 (ZBP-1) and its Xenopus laevis homologue, Vg1 RNA and endoplasmic reticulum-associated protein (VERA)/Vg1 RNA-binding protein (RBP), bind repeated motifs in the 3' untranslated regions (UTRs) of localized mRNAs. Although these motifs are required for RNA localization, the necessity of ZBP-1/VERA remains unresolved. The role of ZBP-1/VERA was addressed through analysis of the Drosophila melanogaster homologue Insulin growth factor II mRNA-binding protein (IMP). Using systematic evolution of ligands by exponential enrichment, the IMP-binding element (IBE) UUUAY, was identified as a motif that occurs 13 times in the oskar 3'UTR. IMP colocalizes with oskar mRNA at the oocyte posterior, and this depends on the IBEs. Furthermore, mutation of all, or subsets of, the IBEs prevents oskar mRNA translation and anchoring at the posterior. However, oocytes lacking IMP localize and translate oskar mRNA normally, illustrating that one cannot necessarily infer the function of an RBP from mutations in its binding sites. Thus, the translational activation of oskar mRNA must depend on the binding of another factor to the IBEs, and IMP may serve a different purpose, such as masking IBEs in RNAs where they occur by chance. These findings establish a parallel requirement for IBEs in the regulation of localized maternal mRNAs in D. melanogaster and X. laevis (Monro, 2006).
IMP contains the four signature KH-type RNA-binding domains and the glutamine-rich COOH terminus that are present in the vertebrate orthologues. Affinity-purified antibodies against IMP reveal the protein in nurse cells and the oocyte early in oogenesis. However, the high concentration of IMP in the follicle cells blocks the penetration of the antibody into the oocyte after stage 4. Therefore, IMP localization was evaluated in a homozygous, viable, and fertile GFP-IMP protein trap line. GFP-IMP is enriched around the nurse cell nuclei and accumulates in the oocyte as soon as it is specified in the germarium, where it shows a uniform distribution until stage 7. IMP accumulates transiently at the anterior of the oocyte during stages 8-9 and then localizes in a crescent at the posterior pole at stage 9, where it remains for the duration of oogenesis. This pattern of localization is very similar to that observed for osk mRNA and Stau protein, which colocalize with IMP throughout oogenesis (Monro, 2006).
To ascertain whether IMP localization depends on osk, whether it is perturbed in various mutants that affect the posterior accumulation of osk mRNA and protein was examined. IMP does not localize to the posterior of the oocyte in staufen, barentsz, and hrp48 mutants, which block the transport of oskar mRNA to the posterior pole. Furthermore, IMP colocalizes with osk RNA to an ectopic dot in the center of the oocyte in a par-1 mutant that disrupts microtubule polarity. Together, these results demonstrate that the localization of IMP to the oocyte posterior pole requires the localization of osk mRNA (Monro, 2006).
IMP could localize to the posterior through a direct interaction with osk mRNA or protein or could be recruited to the posterior by a downstream component of the pole plasm. To distinguish between these possibilities, IMP localization was examined in a strong vasa hypomorph (vasaPD/Df(2L) TW2), which prevents the posterior recruitment of Vasa by Osk and disrupts all subsequent steps in pole plasm assembly. IMP localizes normally to the posterior of these oocytes, suggesting that its posterior accumulation depends on osk directly. Finally, whether IMP localization depends on Osk protein rather than osk mRNA was addressed by examining a nonsense mutation (osk54/Df) that disrupts the anchoring, but not the initial localization, of osk mRNA. IMP still localizes to the posterior of these oocytes at stage 9, but the posterior crescent is weaker than in wild type (WT) and disappears at stage 10. Thus, IMP behaves like osk mRNA in every mutant combination examined, suggesting that it localizes to the posterior in association with the mRNA (Monro, 2006).
The object of this study was to address whether D. melanogaster IMP is required for mRNA localization, because previous studies of its vertebrate homologues, ZBP-1 and VERA/Vg1RBP, had not resolved this question definitively. This study demonstrated that IMP binds directly to osk mRNA at well defined sites that are required for osk translation and anchoring. The best evidence that these sites are bona fide IBEs is that IMP is not recruited to the posterior by osk mRNA in which all 13 IBEs have been mutated with a single base change. Indeed, this is one of the only cases where it has been possible to demonstrate that an RBP interacts in vivo with well defined elements identified biochemically in vitro. In vitro, mutant RNA still competes for binding of IMP, albeit less effectively than the WT osk RNA, suggesting that the 3'UTR may contain other lower affinity sites. However, these sites are not involved in the recruitment of IMP to the posterior in vivo, nor are they sufficient for translational activation. Although the IBEs are thus bona fide in vivo IMP-binding sites, their role in osk RNA translation and anchoring is independent of IMP, which