Intracellular mRNA localization and translation are ways to achieve asymmetric protein sorting in polarized cells, and they play fundamental roles in cell-fate decisions and body patterning during animal development. These processes are regulated by the interplay between cis-acting elements and trans-acting RNA-binding proteins that form and occur within a ribonucleoprotein (RNP) complex. Recent studies in the Drosophila oocyte have revealed that RNP complex assembly in the nucleus is critical for the regulation of cytoplasmic mRNA localization and translation. Furthermore, several trans-acting factors promote the reorganization of target mRNAs in the cytoplasm into higher-order RNP granules, which are often visible by light microscopy. Therefore, RNA localization and translation are likely to be coupled within these RNP granules. Notably, diverse cytoplasmic RNP granules observed in different cell types share conserved sets of proteins, suggesting they have fundamental and common cellular functions.
Proper mRNA localization and translation are governed by the interactions between cis-acting RNA elements and trans-acting RNA-binding proteins. These cis-acting elements are often found in, albeit not absolutely restricted to, the 3′ untranslated region (UTR) of the mRNA. Specific trans-acting factors recognize and bind these elements, forming a ribonucleoprotein (RNP) complex. In many cases, RNP complexes congregate to form large granules that are often visible by light microscopy. Therefore, mRNA localization and translation are probably coupled, at least in part, through interactions between trans-acting factors within the RNP granules.
Drosophila oogenesis has been used as a tractable model system for studying the mechanisms of mRNA localization and translational control. The localizations of several maternal mRNAs in the Drosophila oocyte are key events for formation of the embryonic body axes and germ cells (Johnstone & Lasko 2001; Kuersten & Goodwin 2003). Taking advantage of the powerful genetics available in Drosophila, combined with biochemistry and live-cell imaging techniques, the cis-acting elements and their trans-acting factors have been functionally analyzed in detail for a number of localized mRNAs, including bicoid (bcd), gurken (grk), nanos (nos), and oskar (osk). In this review, we summarize recent advances in the mechanisms of mRNA localization and translational control during Drosophila oogenesis. We focus particularly on the regulation of osk mRNA localization and translation, which are crucial for the assembly of the germ plasm during Drosophila oogenesis (Lehmann & Nüsslein-Volhard 1986; Ephrussi et al. 1991; Kim-Ha et al. 1991; Ephrussi & Lehmann 1992).
Drosophila oogenesis proceeds in an egg chamber, which is composed of a cluster of 16 germline cells surrounded by a single epithelial layer of somatic follicle cells (Fig. 1) (Spradling 1993; Bastock & St. Johnson 2008). The 16-cell cyst is produced from four rounds of synchronous division of a germline stem cell daughter (cystoblast). Because of incomplete cytokinesis during each division, the cells remain interconnected through specialized cytoplasmic bridges called ring canals (Fig. 1). During egg-chamber formation, one of the 16 cells is determined to be the oocyte, while the remaining 15 develop into polyploid nurse cells. Maternal mRNAs are synthesized in the nurse cells and pass through the ring canals into the oocyte, where several key maternal mRNAs are localized to particular regions for specification of the embryonic body axes and pole cells, the germ-cell precursors (St Johnston 2005).
Maternal mRNA localization in embryonic axis and germ-cell formation
The first embryonic axis to be established is the anterior-posterior axis, which happens during early oogenesis. Prior to the establishment of this axis in the oocyte, grk mRNA is transported to the oocyte, translated, and Grk protein (a transforming growth factor-α [TGF-α]-like secretory molecule) signals the adjacent follicle cells to specify their posterior fate (Neuman-Silberberg & Shüpbach 1993; González-Reyes et al. 1995; Steinhauer & Kalderon 2006). The posterior-fated follicle cells in turn signal back to the oocyte to promote the reorganization of microtubule (MT) arrays, which mediate the localization of several maternal mRNAs, including bcd, nos, and osk, within the oocyte (Fig. 2A).
The localization of bcd mRNA to the anterior and nos mRNA to the posterior pole of the oocyte initiates the formation of the embryonic anterior-posterior axis (Fig. 2C). Translation of the anteriorly localized bcd mRNA after egg deposition leads to an anterior-to-posterior gradient of Bcd protein in the embryo (St. Jonhston & Nüsslein-Volhard 1992) (Fig. 2E). Bcd is a morphogen that promotes the concentration-dependent, zygotic transcription of several genes required for head and thorax formation.
On the other hand, nos mRNA is localized to and translated at the posterior pole of the embryo, generating a posterior-to-anterior gradient of Nos protein (Fig. 2F). Nos is a translational repressor that prevents the translation of maternally supplied hunchback (hb) mRNA in the posterior half of the embryo (Irish et al. 1989; Sonoda & Wharton 1999). Disruption of the Nos-dependent repression of maternal hb translation results in the uniform distribution of Hb protein and failed specification of the embryonic abdomen.
The dorsal-ventral axis of the oocyte is specified by the signaling between the oocyte and follicle cells during mid-oogenesis (Van Eeden & St Johnston 1999; Steinhauer & Kalderon 2006). The localization of grk mRNA to the anterior-dorsal corner of the oocyte promotes the localized production of Grk, which signals to the overlaying follicle cells (Fig. 2B). The multiple signaling events launched by localized Grk in the oocyte ultimately lead to the specification of embryonic dorsal-ventral patterning (Van Eeden & St Johnston 1999).
Another important event during Drosophila oocyte development is the establishment of a specialized cytoplasm, the germ (pole) plasm, which resides at the posterior pole of oocytes and embryos (Mahowald 2001). It is assembled during oogenesis through the localization of a number of maternal mRNAs and proteins, such as nos, germ cell-less (gcl) and polar granule component (pgc) mRNAs, and Aubergine, Vasa and Tudor proteins, to the posterior pole of the oocyte. These germ plasm factors play essential roles in the formation of the embryonic abdomen and pole cells (Mahowald 2001).
The first germ plasm mRNA that localizes to the oocyte posterior is osk (Ephrussi et al. 1991; Kim-Ha et al. 1991) (Fig. 2D). Localization and translation of osk mRNA at the oocyte posterior initiate the assembly of the germ plasm. Therefore, embryos lacking maternal Osk activity lack the germ plasm and fail to form the abdomen or pole cells (Lehmann & Nüsslein-Volhard 1986). Moreover, mislocalization of osk mRNA to the oocyte anterior pole causes ectopic germ-plasm assembly, further indicating that the localization of osk mRNA directs the assembly of the germ plasm (Ephrussi & Lehmann 1992).
Microtubule-mediated transport of osk mRNA
The transport of mRNAs and proteins from the nurse cells to the oocyte is a selective process that depends on the proper alignment of MT networks (St Johnston 2005; Steinhauer & Kalderon 2006). When the oocyte is determined in the egg chamber during early oogenesis, the microtubule-organizing center (MTOC) forms in the oocyte’s distal side, such that the MT plus ends extend across the ring canals into the nurse cells (Steinhauer & Kalderon 2006) (Fig. 1B). The MT minus end-directed motor, the Dynein–Dynactin complex, mediates the nurse cell-to-oocyte transport of maternal mRNAs and proteins (Clark et al. 2007; Mische et al. 2007). The Dynein–Dynactin motor complex interacts with the adaptor protein Bicaudal-D (BicD), which binds to several distinct cargo adaptors and mediates cargo transport. For example, BicD can bind to the small GTPase Rab6 to mediate vesicle transport, or to Egalitarian (Egl) for mRNA transport (Bullock & Ish-Horowicz 2001; Coutelis & Ephrussi 2007; Januschke et al. 2007; Mach & Lehmann 1997; Matanis et al. 2002). Although Egl has no canonical RNA-binding motif, it was recently revealed that it contains a novel RNA-binding domain that recognizes specific localization elements in a variety of RNAs for Dynein-dependent transport (Dienstbier et al. 2009). Egl also interacts with the Dynein light chain to facilitate the transport of cargo RNAs (Navarro et al. 2004). Loss-of-function mutants for either BicD or egl, as well as germline clones for dynein heavy chain mutants, fail to transport maternal molecules, such as osk mRNA, into a future oocyte, resulting in the development of 16 nurse cells (and no oocyte) in the egg chamber (Suter et al. 1989; Wharton & Struhl 1989; Suter & Steward 1991; Carpenter 1994; Mach & Lehmann 1997). Therefore, the Dynein-mediated mechanism is essential for the nurse cell-to-oocyte transport and oocyte determination (Fig. 3).
During stages 6–7, in response to the Grk signaling between the oocyte and adjacent follicle cells, the MT cytoskeleton is reorganized extensively. The MT minus ends are lost from the oocyte posterior and accumulate along the lateral and anterior cortex (González-Reyes et al. 1995; Roth et al. 1995; Cha et al. 2002). As a result of this MT reorganization, an MT plus-end marker, Kinesin-β-galactosidase fusion protein (Kin-βgal), and the component of the endogenous MT plus-end-directed motor, Kinesin heavy chain, accumulate at the posterior pole of the oocyte, during stages 8–10 (Clark et al. 1994; Palacios & St Johnston 2002).
Consistent with the posterior enrichment of the MT plus-end markers, the localization of osk mRNA to the posterior pole of the oocyte is carried out by MTs and the MT plus-end-directed motor, Kinesin I. In oocytes lacking kinesin heavy chain function, osk mRNA fails to localize to the posterior pole (Brendza et al. 2000; Cha et al. 2002). A recent study using time-lapse imaging of live oocytes with green fluorescent protein (GFP)-tagged osk mRNA showed that the osk mRNA forms cytoplasmic particles that exhibit Kinesin-dependent directional movements throughout the oocyte (Zimyanin et al. 2008) (Fig. 3).
Surprisingly, in stage 9 oocytes, when osk mRNA starts to accumulate at the posterior pole, the osk particles actually move in all directions with a small bias toward the posterior (Zimyanin et al. 2008). Therefore, the MT array is not absolutely polarized to the posterior pole, and osk RNPs localize to the oocyte posterior by a biased random walk along a weakly polarized cytoskeleton. It is worth noting that directional transport by biased random walk appears to be a general system for establishing cell polarity, as a similar phenomenon is observed in the vesicle transport occurring during the establishment of planar polarity in an epithelial cell layer (Shimada et al. 2006).
Localization of osk mRNA is coupled with its translational control
Cis-elements and trans-acting factors for osk mRNA localization and translation
The osk mRNA localization signals were originally mapped to its 3′ UTR, because osk mRNA that lacks its 3′ UTR does not localize to the oocyte posterior, while a lacZ-osk 3′ UTR hybrid RNA, in which the β-galactosidase (lacZ) coding sequence is fused to the osk 3′ UTR, does localize to the posterior pole (Kim-Ha et al. 1993).
The localization of osk mRNA to the oocyte posterior requires Staufen (Stau). Stau was identified through a genetic screen as a regulator of osk mRNA localization (St. Jonhston et al. 1991; Ephrussi & Lehmann 1992). Structural and biochemical analyses revealed that Stau contains five copies of double-stranded RNA (dsRNA) binding motifs (Bycroft et al. 1995; Micklem et al. 2000). The posterior localization of Stau and osk mRNA in the oocyte is interdependent, supporting the idea that they interact directly in vivo (St. Jonhston et al. 1991). Although the specific interaction of Stau’s dsRNA-binding domains with osk mRNA has not been demonstrated in vitro, the osk 3′ UTR is required for the transport of Stau to the oocyte (Jenny et al. 2006). Thus, Stau might directly bind the osk 3′ UTR to assemble the transport-competent osk RNP complex, and promote its posterior localization. However, recent studies have revealed that the localization of osk mRNA is determined by far more complicated mechanisms than initially thought (Fig. 3).
The nuclear history of osk mRNA determines its cytoplasmic localization
The characterization of additional genes involved in osk mRNA localization led to the surprising conclusion that osk mRNA localization to the oocyte posterior is not simply owing to the presence of a localization element in the mature mRNA. Indeed, it is now clear that the cytoplasmic fate of osk mRNA is dictated by nuclear events, including the splicing of osk pre-mRNA and the subsequent association of the mature osk mRNA with exon–exon junction complex (EJC) components, such as Mago-nashi, Y14/Tsunagi, Barentsz, and eIF4AIII (van Eeden et al. 2001; Hachet & Ephrussi 2001, 2004; Mohr et al. 2001; Palacios et al. 2004). The EJC was discovered as a complex loaded approximately 20–24 nucleotides upstream of the exon–exon boundary of the processed mRNA upon the splicing reaction in the nucleus (Le Hir & Séraphin 2008). The EJC shuttles between the nucleus and the cytoplasm, and facilitates nuclear export and translation of spliced mRNAs. In the Drosophila oocyte, its components localize to the posterior pole along with osk mRNA (Hachet & Ephrussi 2001; Mohr et al. 2001). The splicing of the first intron in the osk pre-mRNA is essential for the posterior localization of mature osk mRNA in the oocyte (Hachet & Ephrussi 2004). The exact mechanism by which the splicing event and subsequent EJC loading close to the first exon–exon boundary mediate osk mRNA localization remains unknown. Nevertheless, these studies indicate that RNP complex assembly in the nucleus is critical for the regulation of cytoplasmic mRNA localization (Fig. 3).
The localization of osk mRNA also requires several RNA-binding proteins that shuttle between the nucleus and the cytoplasm (Fig. 3). Heterogeneous nuclear RNP (hnRNP) proteins are a family of RNA-binding proteins that are highly enriched in the nucleus. Many of them are nucleo-cytoplasmic shuttling proteins with various functions in both the cytoplasm and nucleus (Dreyfuss et al. 2002). Among them, Hrp48 and Squid/Hrp40, members of the hnRNP A/B protein family, are involved in osk mRNA regulation (Huynh et al. 2004; Yano et al. 2004; Norvell et al. 2005). In several missense hrp48 mutants, osk mRNA fails to localize to the oocyte posterior, although neither the MT array polarization nor the osk splicing is affected (Huynh et al. 2004). Oocytes lacking sqd function also show defects in osk mRNA localization (Norvell et al. 2005). Hrp48 and Sqd localize to the posterior pole of the oocyte in an osk mRNA-dependent manner. Hrp48 binds to both the osk 5′ UTR and 3′ UTR (Huynh et al. 2004; Yano et al. 2004), and Sqd interacts directly with both Hrp48 and osk mRNA (Norvell et al. 2005). Notably, Hrp48 also regulates osk translation. In hrp48 mutant flies where Hrp48 expression is reduced, ectopic Osk is detected in the ooplasm. Therefore, Hrp48 appears to be involved in the coupling of osk mRNA localization with its translational control.
Translational control of osk mRNA through its 3′ and 5′ interaction
Bruno (Bru), an RNA-binding protein containing three RNA-recognition motif (RRM) domains, was the first protein discovered to be a translational repressor for osk mRNA. It binds to specific repeated sequences in the osk 3′ UTR, called Bruno response elements (BREs) (Kim-Ha et al. 1995). Mutations in BREs cause ectopic osk translation throughout the oocyte (Kim-Ha et al. 1995; Webster et al. 1997). Direct evidence for Bru’s role in the specific translational repression of osk mRNA was obtained using in vitro translation systems from Drosophila embryonic and ovarian extracts (Lie & Macdonald 1999; Castagnetti et al. 2000). However, the mechanism by which Bru’s binding to the osk 3′ UTR inhibits translation, which occurs from the 5′ end of the mRNA, remained a mystery.
The translation of most mRNAs requires the 7-methyl guanosine cap at the 5′ end. The cap-binding eukaryotic initiation factor 4E (eIF4E) is essential for cap-dependent translational initiation, during which eIF4E interacts with the scaffold protein eIF4G, which recruits the 43S ribosomal pre-initiation complex to the mRNA via eIF3 (Gebauer & Hentze 2004). Interestingly, eIF4G also promotes the intra-molecular circularization of mRNAs through an interaction with poly(A)-binding protein (PABP). Translational initiation is often targeted for regulation. For example, the eIF4E-binding proteins (4E-BPs) negatively regulate translation by competing with eIF4G for the interaction with eIF4E (Richter & Sonenberg 2005; Jackson et al. 2010).
The ovarian protein Cup and eIF4E were identified through biochemical analyses as components of the osk mRNA-containing complex in Drosophila ovaries (Wilhelm et al. 2003; Nakamura et al. 2004). Cup binds eIF4E through the conserved sequence also found in eIF4G and 4E-BP, and acts as a translational repressor of osk mRNA (Wilhelm et al. 2003; Nakamura et al. 2004). Therefore, it is likely that Cup binds eIF4E in the same manner as eIF4G to inhibit the translational initiation of osk mRNA. Intriguingly, Cup also associates with Bru (Nakamura et al. 2004). These observations led to a model in which the specific repression of osk translation is achieved, at least in part, through 5′ and 3′ interactions mediated by eIF4E–Cup–Bru interactions.
Cup is also required for normal osk mRNA distribution. In hypomorphic cup mutants, osk mRNA forms aberrant granules that are dispersed in the cytoplasm of both nurse cells and the oocyte, although osk eventually accumulates highly at the posterior pole of late-stage oocytes (Nakamura et al. 2004). Therefore, Cup appears to be involved in both osk mRNA localization and translational control, and might be important for coupling these processes.
Intracellular mRNA localization and translation might be coupled through RNP oligomerization
In an early study, lacZ-osk 3′UTR mRNA was shown to localize to the posterior pole of the oocyte, which appeared to provide convincing evidence that the 3′ UTR of the osk mRNA is sufficient to promote its posterior localization (Kim-Ha et al. 1993). However, in an apparent contradiction, another group reported that the splicing of osk pre-mRNA is essential for the posterior localization of osk mRNA (Hachet & Ephrussi 2004). Surprisingly, when intronless lacZ-osk 3′ UTR mRNA is expressed during oogenesis without endogenous osk mRNA expression, the lacZ mRNA cannot be detected at the oocyte posterior, suggesting that the intronless osk 3′ UTR reporter is carried to the posterior pole of the oocyte by hitchhiking on the osk mRNA that has experienced the splicing in the nucleus (Hachet & Ephrussi 2004). These results further suggest that one of critical roles of the osk 3′ UTR is to promote the oligomerization of osk RNP complexes into granules containing multiple osk mRNA molecules (Fig. 3).
The Cup–Bru complex sequesters eIF4E and inhibits osk translation (Nakamura et al. 2004). In a cell-free translation assay using ovarian lysates, however, Bru can repress osk translation by another mechanism that requires neither the Cup–eIF4E interaction nor the cap structure on the mRNA (Chekulaeva et al. 2006). Intriguingly, in the cell-free extracts, the reporter mRNAs are oligomerized in a BRE-dependent manner, such that they are sequestered into particles and are inaccessible to translational apparatuses such as ribosomes. Therefore, a second repression mechanism appears to involve Bru-dependent mRNA oligomerization into silencing particles (Chekulaeva et al. 2006).
Polypyrimidine tract-binding protein (PTB)/hnRNP I also promotes osk mRNA oligomerization (Besse et al. 2009). In ptb mutant oocytes, the osk mRNA-positive granules are smaller than normal, and osk translation is not properly repressed. PTB leads to the formation of large multimers of osk 3′ UTR RNA in vitro, suggesting that the osk 3′ UTR alone is sufficient for PTB-dependent osk mRNA oligomerization. Although PTB is a nucleo-cytoplasmic shuttling protein, a cytoplasmic form of PTB is sufficient to repress osk translation. Therefore, PTB is a key structural component of the osk RNP complex that controls the formation of high-order RNP particles and translational silencing in the cytoplasm. In the ptb mutant, endogeneous osk mRNA is almost properly localized to the posterior pole in stage 9 oocytes. In contrast, the posterior accumulation of osk 3′ UTR-containg reporters that would hitchhike on the endogeneous of osk mRNA during localization is severly reduced (Besse et al. 2009). These findings suggest that the localization and translational control of osk mRNA may be coupled through PTB-mediated oligomerization.
osk RNP forms granules that are visible by light microscopy
Interestingly, many factors known to be involved in the decay of mRNAs are also concentrated in maternal RNP granules and regulate osk mRNA localization and translation. The degradation of mRNAs in eukaryotes is a regulated process, and the factors involved are highly conserved among species (Eulalio et al. 2007a; Parker & Sheth 2007). mRNA degradation usually begins with deadenylation, the removal of the poly(A) tail. Deadenylated mRNAs are in turn degraded by exonucleases from both the 5′ and 3′ ends. Degradation from the 3′ end is catalyzed by a multi-protein complex containing the exosome and the Ski complex. In contrast, the 5′–3′ mRNA decay pathway requires removal of the cap structure by the decapping enzyme complex, which consists of catalytic Dcp2 and regulatory Dcp1 subunits. The decapped mRNAs are then degraded by the 5′–3′ exonuclease, Xrn1 (Parker & Song 2004). In a dcp1 hypomorphic mutant, osk mRNA fails to localize to the posterior pole, and remains in the anterior oocyte (Lin et al. 2006). Although Dcp1 and Dcp2 colocalize to form cytoplasmic granules in the nurse cells, only Dcp1, and not Dcp2, localizes to the posterior pole of the oocyte (Lin et al. 2006, 2008). Therefore, Dcp1 may regulate osk mRNA localization independently of its decapping function.
Genetic analyses in budding yeasts have suggested that the Dcp1/Dcp2 complex is controlled by many regulatory proteins (Eulalio et al. 2007a; Parker & Sheth 2007). Although these proteins, known as enhancers of decapping (Edc) or decapping activators, are highly conserved in eukaryotes, their molecular functions in vivo remain largely unknown. Me31B is an evolutionarily conserved DEAD-box RNA helicase. The yeast homologue of Me31B, called Dhh1p, is a decapping activator that also functions, together with the protein Pat1p, in the general translation repression caused by gluose deprivation (Coller & Parker 2005). Furthermore, the overexpression of either Dhh1p or Pat1p is sufficient to block translation grobally. In Drosophila ovaries lacking me31B, a low level of Osk is ectopically detected in the nurse cell cytoplasm during early oogenesis, suggesting that Me31B represses osk translation (Nakamura et al. 2001). Interestingly, the Pat1p homologue in Drosophila (Patr-1/HPat) colocalizes with Me31B during oogenesis (Y. Kato and A. Nakamura unpubl. data, 2009.).
Ge-1 is a metazoan protein that binds to Dcp1 and enhances its decapping activity in vitro (Fengar-Grøn et al. 2005). Drosophila Ge-1 (also known as Hedls/RCD-8) associates with osk mRNA in vitro, and colocalizes with osk mRNA during oogenesis (Fan et al. 2011). Drosophila ovaries with a ge-1 hypomorphic allele shows defects in osk mRNA localization and strong genetic interactions with components of maternal RNP granules. Interestingly, maternal RNP granules become undetectable in ge-1 mutant egg chambers, suggesting that Ge-1 plays a direct role in the assembly of maternal RNP granules. Furthermore, in sucrose density gradients, the osk RNP complex from ge-1 mutant egg chambers shifts to lighter density fractions. Therefore, Ge-1 appears to regulate the osk mRNA localization, and possibly translation, by promoting proper assembly of the osk RNP.
The 5′–3′ exonuclease Xrn1, called Packman (Pcm) in Drosophila, also localizes to maternal RNP granules in nurse cells (Lin et al. 2008). The depletion of Pcm from ovaries results in maternal RNP granules of increased size and number. An enlargement of maternal RNP granules is also observed when Dcp1 or Dcp2 is mutated (Lin et al. 2008). Given that a lack of Dcp1/Dcp2 or Pcm can stabilize mRNAs that are destined for degradation, these molecules may upregulate the assembly of cytoplasmic RNP granules indirectly through mRNA stabilization. Whether Pcm is involved in osk mRNA regulation has not yet been reported.
Maternal RNP granules are related, albeit not identical, to somatic P bodies
Intriguingly, many proteins found in maternal RNP granules are also expressed in somatic cells and form cytoplasmic granules called processing bodies (P bodies) (Fig. 5B). P bodies have been observed in eukaryotic cells from budding yeasts to humans (Sheth & Parker 2003; Cougot et al. 2004; Eulalio et al. 2007a). In budding yeasts, P bodies typically contain Dcp1p, Dcp2p, Xrn1p (Pcm in Drosophila), Dhh1p (Me31B in Drosophila), Pat1p, the Lsm1-7 complex, Sdc6p/LSm15 [Trailer hitch (Tral) in Drosophila], and Edc3p/LSm16 (Table 1). The P bodies of metazoan cells, such as Drosophila and human, are also highly enriched in Ge-1 (Fengar-Grøn et al. 2005; Yu et al. 2005; Eulalio et al. 2007b). Therefore, P bodies share many protein components with maternal RNP granules (Fig. 5).
Although maternal RNP granules are known to play important roles in osk mRNA regulation, there remain many unanswered questions. For example, what are the exact mechanisms by which the splicing-dependent loading of the EJC onto the mature osk mRNA in the nucleus promotes its posterior localization in the oocyte. The EJC proteins might interact directly with the motor protein complex to steer RNP transport, or the EJC proteins might modulate the secondary structure of osk mRNA, thereby facilitating the recruitment of trans-acting factors required for osk mRNA localization in the oocyte. Another important issue is understanding how the dynein-mediated transport of osk mRNA from nurse cells to the oocyte is transferred to a kinesin-dependent process in the oocyte. Of interest is the factors involved that promote this switch. A final intriguing question is determining the similarities and differences of the biochemical composition and function between P bodies, maternal RNPs, and neural RNA granules. The findings obtained from any of these studies on RNPs will likely broaden our general understanding on the roles of cytoplasmic RNP granules. Therefore, the continued analyses of osk regulation in Drosophila oocytes should provide further insights into the general mechanisms of mRNA regulation occurring in RNP granules.
We apologize to those whose primary work could not be cited because of space constraints. We thank Drs Tsubasa Tanaka and Kazuko Hanyu-Nakamura for comments, and Ms Hazuki Hiraga for proofreading of the manuscript. Research in the Nakamura lab was supported in part by the RIKEN President Discretionary Fund, and a Grant-in-Aid from the Ministry of Education, Culture, Sports, Science and Technology, Japan, and the Japan Society of the Promotion of Science.