|Home | About | Journals | Submit | Contact Us | Français|
Localization of the germ plasm to the posterior of the Drosophila oocyte is required for anteroposterior patterning and germ cell development during embryogenesis. While mechanisms governing the localization of individual germ plasm components have been elucidated, the process by which germ plasm assembly is restricted to the posterior pole is poorly understood. In this study, we identify a novel allele of baz, the Drosophila homolog of Par-3, that has allowed the analysis of baz function throughout oogenesis. We demonstrate that baz is required for spatial restriction of the germ plasm and axis patterning and we uncover multiple requirements for baz in regulating the organization of the oocyte microtubule cytoskeleton. Our results suggest that distinct cortical domains established by Par proteins polarize the oocyte through differential effects on microtubule organization. We further show that microtubule plus-end enrichment is sufficient to drive germ plasm assembly even at a distance from the oocyte cortex, suggesting that control of microtubule organization is critical not only for the localization of germ plasm components to the posterior of the oocyte but also for the restriction of germ plasm assembly to the posterior pole.
Axial patterning during embryonic development often relies on molecular asymmetries that are established during oogenesis and propagated in the early embryo. Anteroposterior (A-P) patterning of the Drosophila embryo requires the asymmetric localization of bicoid (bcd) and nanos (nos) mRNAs during oogenesis, with bcd targeted to the anterior and nos to the posterior (reviewed in Berleth et al., 1988; Wang et al., 1994). After fertilization, opposing protein gradients produced by translation of these localized maternal mRNAs specify cell fates along the A-P axis (Driever and Nusslein-Volhard, 1988; Gavis and Lehmann, 1992). Consequently, mutations that disrupt bcd function or mRNA localization affect development of head and thoracic segments whereas mutations that disrupt nos function or mRNA localization produce embryos lacking abdominal segments (Driever and Nusslein-Volhard, 1988; Frohnhofer et al., 1986; Lehmann and Nusslein-Volhard, 1991; Wang et al., 1994).
nos is localized to the germ plasm, a specialized cytoplasm at the posterior of the oocyte. In addition to containing nos mRNA, and, consequently directing abdominal segmentation, this assemblage of localized RNAs and proteins is necessary and sufficient for the formation of the germ cells at the posterior of the embryo (reviewed in Mahowald, 2001). Germ plasm assembly occurs by a hierarchical pathway that begins with the transport of oskar (osk) mRNA to the posterior of the oocyte (Ephrussi et al., 1991; Kim-Ha et al., 1991). In turn, osk localization relies on the polarization of the A-P axis of the oocyte, a process initiated earlier in oogenesis with local signaling by Gurken (Grk), a TGFα-like ligand (Gonzalez-Reyes et al., 1995). Drosophila oogenesis proceeds through 14 morphologically distinct stages (reviewed in Spradling, 1993) during which the oocyte is supplied with maternal mRNAs, proteins and organelles by 15 accessory nurse cells. Early in oogenesis, microtubules nucleated by a microtubule-organizing center (MTOC) at the posterior of the oocyte mediate transport of maternal mRNAs from the nurse cells into the oocyte (Theurkauf et al., 1992). Using this network, grk mRNA becomes localized to the posterior of the oocyte where the resulting Grk protein signals to the overlying somatic follicle cells, triggering the disassembly of the posterior MTOC (reviewed in Steinhauer and Kalderon, 2006). The subsequent nucleation of microtubules at the anterior and lateral oocyte cortex leads to a reorganization of the oocyte microtubule cytoskeleton and a bias of microtubule plus ends oriented toward the posterior pole (Cha et al., 2001; Theurkauf et al., 1992; Zimyanin et al., 2008).
One consequence of the reorganization of oocyte microtubules is the relocation of the oocyte nucleus and grk mRNA from the posterior to the dorsal anterior corner of the oocyte (Gonzalez-Reyes et al., 1995). Here, Grk is again synthesized and signals to the overlying follicle cells to specify the dorsoventral (D-V) axis of the embryo (Neuman-Silberberg and Schupbach, 1993). Another consequence is the kinesin-dependent transport of osk mRNA to the posterior of the oocyte, which initiates the assembly of the germ plasm (Brendza et al., 2000). Upon localization, osk is translated and the resulting protein recruits other germ plasm components to the posterior including the RNA helicase Vasa (Vas) (Breitwieser et al., 1996; Hay et al., 1988; Schupbach and Wieschaus, 1986) which together with Osk is required for the localization of nos mRNA later in oogenesis (Wang et al., 1994). osk localization is sufficient to dictate the site of germ plasm assembly, as mistargeting of osk mRNA to the anterior of the oocyte results in ectopic germ plasm assembly, germ cell formation and abdominal development at the anterior of the embryo (Ephrussi and Lehmann, 1992). Interestingly, while translational repression of osk mRNA during its localization to the posterior of the oocyte is required to prevent ectopic Osk function (Kim-Ha et al., 1995), the mechanism by which this repression is alleviated at the posterior pole remains unclear. Additionally, the mechanisms that govern the polarity of the oocyte microtubule cytoskeleton and the restriction of the germ plasm to the posterior pole are not fully understood.
The oocyte cortex itself is polarized by the Par proteins. First characterized in C. elegans, Par proteins are key regulators of cell polarity in many cell types in diverse organisms (reviewed in Goldstein and Macara, 2007). The Par proteins function at the cell cortex where they form two mutually exclusive domains containing either Par-1 or Par-3. In the C. elegans embryo, as well as in mammalian epithelial cells, these complementary cortical domains define an axis of cell polarity and are regulated by antagonistic interactions between the Par proteins (reviewed in Suzuki and Ohno, 2006). In the Drosophila oocyte, Par-1 is localized to the posterior pole where it accumulates by a microfilament based mechanism, prior to osk mRNA localization, and is required for correct polarization of the microtubule cytoskeleton, posterior localization of osk, and consequently the assembly of the germ plasm (Doerflinger et al., 2006; Shulman et al., 2000). In contrast, Bazooka (Baz), the Drosophila homolog of Par-3, is enriched on the anterior and lateral cortex (Benton and St Johnston, 2003b). The function of Baz in the oocyte is poorly understood because all previously characterized baz alleles display defects in oocyte specification (Cox et al., 2001; Huynh et al., 2001). While forcible localization of Baz to the posterior of the oocyte, by expression of an unregulated form of the protein, leads to A-P patterning defects (Benton and St Johnston, 2003b), a bona fide role for Baz in axis determination has yet to be demonstrated.
Here we describe the identification and characterization of a novel allele of baz, bazX-82, which supports oocyte development. We find that bazX-82 causes defects in microtubule organization and germ plasm localization in the oocyte that lead to ectopic foci of germ plasm in the embryo. In addition, we show that mutation of grk, which disrupts microtubule polarity by eliminating the necessary oocyte-to-follicle cell signaling events, leads to ectopic osk translation and germ plasm assembly. Together, these results reveal a novel role for Baz in regulating oocyte microtubule polarity and axis patterning. Furthermore, our results suggest that a focus of microtubule plus ends is sufficient to trigger osk translation and germ plasm assembly even at a location distant from the oocyte cortex.
Oregon-R or y w67c23 were used as wild-type controls. The following mutants and transgenic lines were used: X-82 (Luschnig et al., 2004), bazEH171 (Eberl and Hilliker, 1988), osk-gfp (Snee et al., 2007), vas-gfp (gift of R. Lehmann), khc-lacZ (Clark et al., 1994), nod-lacZ (Clark et al., 1997), pnt-lacZ (Gonzalez-Reyes and St Johnston, 1998), UASp-baz-gfp (Benton and St Johnston, 2003a), baz4 (Muller and Wieschaus, 1996), grk2B and grkHF (Neuman-Silberberg and Schupbach, 1993).
An unlinked lethal mutation present on the original bazX-82 chromosome was removed by recombination of bazX-82 onto the FRT19A chromosome (gift of M. Metzstein). Germline clones were induced by the dominant female sterile technique (Chou et al., 1993). Follicle cell clones were generated using the FRT/UAS-Flp/GAL4 system (Duffy et al., 1998). Clones were induced in females of the following genotype: bazX-82 FRT19A/ubi-GFP FRT19A; e22c-GAL4 UAS-FLP/ + or FRT19A/ubi-GFP FRT19A; e22c-GAL4 UAS-FLP/ + as a control.
UASp-baz-gfp and UASp-bazX-82-gfp were expressed to equivalent levels using the matα-tubulin-GAL-VP16V2H driver at 18° C and the matα-tubulin-GAL-VP16V37 driver at 29° C respectively (Bloomington).
X-82 was mapped proximal to sd by meiotic recombination between the X-82 chromosome and the t v m wy sd os and y sn lz ras v m chromosomes (Bloomington). Recombinants carrying the X-82 mutation were followed by examining embryos under oil for segmentation defects and confirmed in cuticle preparations. X-82 was further mapped using in situ hybridization to nos mRNA in early embryos from females heterozygous for X-82 and each of the following deficiencies: Df(1)ED7355, Df(1)4b18, Df(1)ED7374, Df(1)B25, Df(1)BK10,Df(1)RR79, Df(1)N19, Df(1)Exel6291, Df(1)ED7441, Df(1)JA27 (Bloomington). In this analysis, Df(1)B25 failed to complement the ectopic nos mRNA phenotype of X-82. Of the 14 genes specific to Df(1)B25, we selected baz as a likely candidate gene. Genomic DNA was isolated from individual homozygous X-82 adult males (Mansfield et al., 2002) and exonic regions of the baz gene were amplified by PCR and sequenced. A nonsense mutation that creates a TAG stop codon was identified in exon 7 of baz, and we designate this allele as bazX-82.
Immunoblotting was performed as previously described (Kalifa et al., 2006) except that ovaries were dissected in PBS. Membranes were probed with rabbit anti-Baz (1:2000) (Wodarz et al., 1999) and proteins were detected using Lumi-Light Western Blot Substrate (Roche).
Cuticle preparations and in situ hybridization to embryos for individual mRNAs were performed as previously described (Gavis and Lehmann, 1992). Double fluorescence in situ hybridization (FISH) using tyramide signal amplification was performed as described (Kosman et al., 2004). Immunofluorescence was performed as previously described (Duchow et al., 2005) using embryos heat-fixed in 68 mM NaCl/0.03% Triton X-100, with rabbit anti-Vas (1:10000; gift from R. Lehmann) and Alexa Fluor-568 goat anti-rabbit (1:1000; Molecular Probes). Embryos were mounted in 90% glycerol/100 mM Tris pH 8.0 and imaged with a Zeiss LSM510 confocal microscope.
To visualize microtubules, ovaries from well-fed females were dissected in 10% EM grade formaldehyde (Polysciences Inc.) in PBST (PBS/2% Tween-20). Fixation was stopped 7 minutes after dissection started. Ovaries were washed 3 × 10 minutes in PBST, blocked in PBST/10% BSA for 1 hour and incubated with FITC conjugated anti-α-tubulin (1:250; Sigma) in PBST/10% BSA overnight at 4° C. Subsequently, ovaries were washed 4 × 15 minutes in PBS/0.1% Tween-20, 4 × 5 minutes in methanol and mounted in 40 μl 2:1 benzyl benzoate:benzyl alcohol.
Anti-Grk immunostaining was performed as described (Pane et al., 2007) except that EM grade formaldehyde was used as a fixative. All other immunostaining was performed as described (Shcherbata et al., 2004) except that 4% EM grade formaldehyde was used as a fixative and PBST contained 0.1% Triton X-100. Primary antibodies were used as follows: mouse anti-Grk 1D12 (1:10) (Queenan et al., 1999), rabbit anti-βgalactosidase (1:1000; Cappel), rabbit anti-Osk (1:3000) (Vanzo and Ephrussi, 2002), rabbit anti-Par-1 (1:5000) (Shulman et al., 2000), rabbit anti-Baz (1:1000) (Wodarz et al., 1999). Secondary antibodies: Alexa Fluor 488 goat anti-mouse, Alexa Fluor 488 and Alexa Fluor 633 goat anti-rabbit (Molecular Probes) were used at 1:500. Samples were counterstained with Alexa Fluor 568 phalloidin (1:500) to visualize F-actin and Alexa Fluor 633 wheat germ agglutinin (WGA; 1 μg/ml; Molecular Probes) to label nuclei. Double staining for Par-1 and Osk was performed in sequence interrupted with a 2 hour biotin anti-rabbit (1:1000; Jackson) incubation to prevent crosstalk of secondary antibodies.
FISH to ovaries was performed as described (Vanzo and Ephrussi, 2002) except that samples were blocked in PBST/0.2% BSA/5% normal goat serum for one hour prior to incubation with anti-digoxygenin-POD (1:20; Roche). Samples were mounted in 90% glycerol/100 mM Tris pH 8.0.
Osk-GFP, Vas-GFP, Baz-GFP and BazX-82-GFP were visualized as previously described (Forrest and Gavis, 2003). All fluorescence images were acquired on a Zeiss LSM 510 confocal microscope.
Egg chambers were prepared for time-lapse imaging as described (Weil et al., 2008). Ooplasmic streaming was visualized using 488 nm excitation, which causes autofluorescence of yolk particles. 16 time-lapse images were acquired at 3 second intervals and averaged using a Zeiss LSM 510 confocal microscope.
pattB-UASp was created by excising the pattB multiple cloning site (MCS) with XbaI and BamHI and replacing it with a blunt ended PstI-StuI fragment of pUASp bearing its Gal4 binding sites, promoter, MCS and K10 3'UTR. To create the truncated bazX-82 coding region fused to GFP, an NcoI site was engineered by PCR after nucleotide 3573 of the baz coding region in pCR II-TOPO-Baz (gift of D. St Johnston). This manipulation creates an alanine codon in place of the TAG stop codon found in bazX-82. A fragment spanning from the ApaI site in pCR II-TOPO-Baz to this NcoI site was then joined to a NcoI-NotI fragment from pEGFP-N1(Clontech) in the vector pUC19 to create an in frame fusion to GFP, which was confirmed by sequencing. This bazX-82-GFP fusion was inserted into pattB-UASp and the resulting UASp-bazX-82-GFP transgene was integrated at 68A4, p(CaryP)attP2 (Groth et al., 2004), using the phiC31 integrase system (Bischof et al., 2007).
We examined a set of Drosophila maternal-effect mutants that exhibit defects in abdominal development (Luschnig et al., 2004) for their effect on nos mRNA localization. Embryos derived from females bearing X-82 mutant germline clones develop an average of 5.9 of the wild-type 8 abdominal segments (n=300), and exhibit a previously unreported defect in localization of nos mRNA that can account for this reduction in segment number (see below). We chose this mutant for further study.
We mapped X-82 to 13F-18D by meiotic recombination with visible markers and by complementation with chromosomal deficiencies (see Methods). Sequencing of the mutant chromosome revealed that X-82 is a nonsense mutation in baz that results in a stop codon at amino acid 1192 of the 1464 amino acid protein (Fig. 1A). A truncated Baz protein is observed in bazX-82 ovarian lysates consistent with the observed molecular lesion (Fig. 1B). The BazX-82 protein retains all of the previously identified functional elements of Baz including an oligomerization domain (Benton and St Johnston, 2003a), three PDZ domains and three regulatory phosphorylation sites (Benton and St Johnston, 2003b). A strong allele of baz, bazEH171, that was isolated independently (Eberl and Hilliker, 1988), fails to complement bazX-82 with respect to the observed defect in nos mRNA localization (data not shown). From these results, we conclude that disruption of baz is responsible for the abdominal segmentation phenotype of X-82.
In strong alleles of baz, the oocyte-specific marker Orb accumulates in the nascent oocyte prior to stage 2 of oogenesis but this enrichment is not maintained, leading to a loss of oocyte fate (Cox et al., 2001; Huynh et al., 2001). By contrast, the production of embryos by females bearing bazX-82 germline clones suggests that bazX-82 does not inhibit oocyte fate specification. We therefore visualized localization of Orb in early egg chambers to evaluate oocyte specification in bazX-82 ovaries. At stages 2-5, Orb is enriched in oocytes of wild-type egg chambers (Lantz et al., 1994). Orb also accumulates in similarly staged bazX-82 oocytes (data not shown) suggesting that bazX-82 is a hypomorphic allele that supports oocyte differentiation.
To determine the origin of the abdominal segmentation defect we observed in embryosderived from bazX-82 germline clones (hereafter referred to as bazX-82 embryos), we characterized the effect of bazX-82 on the localization of nos mRNA. In contrast to wild-type embryos in which nos is localized to the posterior pole, about 50% of bazX-82 embryos lack posterior nos localization (Fig. 2 B). Strikingly, approximately 30% of bazX-82 embryos show nos mRNA localized in ectopic cortical patches (Figs. 2A and B) and this phenotype is independent of the posterior localization defect. Localization of nos is absolutely dependent on the prior assembly of the germ plasm and forced mislocalization of germ plasm components to the anterior results in anterior mislocalization of nos (Ephrussi and Lehmann, 1992; Wang et al., 1994). We therefore analyzed the distribution of integral germ plasm components including osk mRNA and Vas protein in bazX-82 embryos. Similarly to nos, osk and Vas are each ectopically localized in patches along the cortex of bazX-82 embryos in addition to exhibiting defects in their localization to the posterior pole (Figs. 2C-F). In contrast, bcd mRNA is unaffected and exhibits wild-type anterior localization indicating that bazX-82 does not disrupt all aspects of A-P polarity in the embryo (Figs. 2G and H). To confirm that the ectopic patches observed for individual germ plasm components correspond to assembled germ plasm, we performed double in situ hybridization experiments for osk and nos. Similarly to their colocalization at the posterior pole of wild-type embryos, nos and osk mRNAs colocalize in the patches found in bazX-82 embryos (Figs. 2I and J).
Targeting of the germ plasm to the anterior of the embryo is sufficient to direct the formation of ectopic germ cell precursors or pole cells (Ephrussi and Lehmann, 1992). We therefore investigated whether the ectopic patches of germ plasm in bazX-82 embryos are competent for pole cell formation. Whereas pole cells form exclusively at the posterior of wild-type embryos (Fig. 2K), 17% of bazX-82 embryos initiate the process of pole cell formation, with the budding of the plasma membrane around germ plasm associated nuclei, at ectopic locations (n=197; Fig. 2L). Unlike wild-type pole buds, these ectopic buds do not progress to form cells, possibly due to a lower level of germ plasm in the ectopic patches than is present at the posterior of the embryo. Taken together these results suggest that the germ plasm assembles properly in bazX-82 oocytes but accumulates at an inappropriate location on the cortex.
To determine the origin of the germ plasm mislocalization in bazX-82 embryos, we characterized the localization of germ plasm components in mutant ovaries. osk mRNA is the first known component of the germ plasm to become localized to the posterior of the oocyte. At stage 8, osk mRNA is observed at both the anterior and posterior of wild-type oocytes. The anterior accumulation is transient, however, and from stage 9 onward, osk is detected exclusively at the posterior pole (Figs. 3A and C) (Kim-Ha et al., 1991). In bazX-82 ovaries, osk mRNA fails to localize to the posterior pole in about 50% of egg chambers and persists at the anterior in 36% of stage 10 oocytes (n=109; Fig. 3B). Moreover, osk mRNA is detected ectopically in the cytoplasm of stage 10 bazX-82 oocytes (23%, n=109; Fig. 3D), with or without posteriorly localized osk. Normally, translation of osk mRNA is repressed until it becomes localized to the posterior cortex. To determine whether ectopically localized osk is translated in bazX-82 oocytes, we examined the expression of Osk protein in mutant oocytes, either by anti-Osk immunostaining or by using Osk-GFP for visualization at later stages of oogenesis. In wild-type oocytes, Osk is detected only at the posterior cortex whereas in bazX-82 oocytes we occasionally detect Osk in an ectopic focus in the cytoplasm at stage 10 (see Fig. 6). Moreover, Vas, a germ plasm component recruited by Osk, is also localized to an ectopic focus in the oocyte cytoplasm during mid-oogenesis as visualized with Vas-GFP (8%, n=118; Figs. 3E and F), indicating that germ plasm assembly is initiated within the cytoplasm. From stage 11 through the end of oogenesis, Osk-GFP and Vas-GFP are found in ectopic cortical patches (Osk 18%, n=145; Vas 12%, n=457; Figs. 3G-J). Thus, the ectopic patches of germ plasm present in bazX-82 embryos appear to result from ectopic localization and translation of osk in the oocyte cytoplasm, leading to inappropriate assembly of germ plasm within the ooplasm and, as oogenesis progresses, ectopic cortical localization.
Reorganization of the microtubule cytoskeleton, triggered by signaling between the oocyte and overlying follicle cells, is required for the polarized transport of osk and bcd mRNAs during mid-oogenesis. When signaling is disrupted, as in mutants for grk or protein kinase A (pka), the posterior MTOC is not disassembled. In these oocytes, the nucleation of microtubules at the anterior and lateral oocyte cortex is thought to result in microtubule minus ends becoming distributed along the entire oocyte cortex, with an enrichment of plus ends directed toward the center of the oocyte. Consequently, osk accumulates in the center of the oocyte in these mutants whereas bcd accumulates at both ends (reviewed in Steinhauer and Kalderon, 2006). We therefore investigated whether the ectopic osk mRNA we observe in bazX-82 ovaries could result from defects in microtubule organization and polarity.
Visualization of oocyte microtubules during mid-oogenesis by α-tubulin immunofluorescence or using the microtubule binding protein Tau-GFP reveals a graded distribution of microtubules within the oocyte with the highest concentration at the oocyte anterior (Fig. 4A and data not shown) (Theurkauf et al., 1992). This distribution is altered in bazX-82 oocytes (Fig. 4B and data not shown). Furthermore, Kin-βgal, a microtubule plus end marker that accumulates at the posterior of wild-type oocytes (Fig. 4C) (Clark et al., 1994) is detected centrally in 21% of bazX-82 egg chambers and remains diffuse in 33% of bazX-82 oocytes (n=148; Fig. 4D). Localization of Nod-βgal, a marker for the minus ends of microtubules, to the anterior of the oocyte is also disrupted in bazX-82 egg chambers, with posterior localization persisting into mid-oogenesis (Figs. 4E and F) (Clark et al., 1997). Another marker for microtubule organization is the migration of the oocyte nucleus to the dorsal anterior corner of the oocyte during mid-oogenesis, which requires a correctly polarized microtubule cytoskeleton (Fig. 4G) (Lei and Warrior, 2000). In bazX-82 egg chambers, nuclear migration is aberrant or fails in about 50% of oocytes (Fig. 4H), and Grk protein is mislocalized together with the oocyte nucleus (Figs. 4G and H). Consistent with this observation, about 40% of bazX-82 embryos display D-V patterning defects reflective of aberrant or absent Grk signaling, including a single dorsal appendage or the absence of appendage material (data not shown). Taken together, these results reveal a role for baz in the organization and polarity of the oocyte microtubule cytoskeleton.
Anterior localization of bcd mRNA occurs by dynein-mediated transport on microtubules (Duncan and Warrior, 2002; Januschke et al., 2002; Pokrywka and Stephenson, 1991; Weil et al., 2006) and thus should also be affected when microtubule organization is altered. While bcd mRNA is properly localized to the anterior of bazX-82 embryos (Figs. 2G and H), bazX-82 oocytes display bcd mislocalized to the lateral oocyte cortex during mid-oogenesis (Figs. 4I and J). bcd mRNA is localized by distinct mechanisms that utilize different populations of microtubules during middle and late stages of oogenesis, with the late-acting mechanism responsible for the majority of bcd transcripts detected at the anterior of the embryo (Weil et al., 2006). The restoration of the wild-type bcd mRNA distribution between mid-oogenesis and embryogenesis indicates that the later bcd localization mechanism is not disrupted in bazX-82 egg chambers.
The production of ectopic germ plasm patches in embryos derived from baz mutant germline clones (Fig. 2) indicates a requirement for baz function in the oocyte itself. However, baz is expressed in follicle cells and therefore could also affect microtubule polarity through a role in posterior follicle cell fate specification or function. Additionally, baz function in the oocyte might be required for the signaling events that establish posterior follicle cell identity. To determine whether the defect in microtubule polarity in bazX-82 oocytes results from a defect in posterior follicle cell specification, we examined the expression of pointed-lacZ (pnt-lacZ), a marker for posterior follicle cell fate (Gonzalez-Reyes and St Johnston, 1998). In wild-type and bazX-82 egg chambers, pnt-lacZ is expressed similarly in a subset of follicle cells posterior to the oocyte (Figs. 4K and L) indicating that the defects in microtubule polarity observed in bazX-82 oocytes are not due to a defect in posterior follicle cell specification. In addition, microtubule polarity, as assayed by the position of the oocyte nucleus, is not affected in egg chambers with large bazX-82 follicle cell clones that encompass the posterior follicle cells (Supplemental Fig. 1). Thus, the observed microtubule defect reflects a role for baz within the oocyte downstream of posterior follicle cell specification.
Around stage 10 of oogenesis, the oocyte microtubules are once again reorganized, with the formation of cortical microtubule arrays (Theurkauf et al., 1992). These cortical microtubules are functionally distinct from the anterior microtubules required for bcd localization in late oocytes (Weil et al., 2006) and drive the churning of the oocyte cytoplasm referred to as ooplasmic streaming that accompanies the rapid emptying of nurse cell contents into the oocyte during the latter period of oogenesis (Gutzeit and Koppa, 1982). In addition to mixing the nurse cell and oocyte cytoplasm, ooplasmic streaming assists in the continued localization of germ plasm components by facilitating diffusion throughout the oocyte (Forrest and Gavis, 2003). Ooplasmic streaming can be visualized by time-lapse imaging of autofluorescence of yolk particles (Fig. 4M). Using this assay, we find that streaming fails to occur in about 75% of late stage bazX-82 oocytes (Fig. 4N). Thus, Baz may play an additional role in the organization of cortical microtubules at late stages of oogenesis. These results, together with the selective effect of bazX-82 on bcd localization during mid-oogenesis (see above), indicate that bazX-82 does not disrupt cytoskeletal polarity globally and suggest that baz may regulate the organization of specific subsets of microtubules within the oocyte.
The mutant BazX-82 protein retains all the identified functional domains of Baz, providing little insight into how the C-terminal truncation compromises Baz function. Since the functions of Par proteins depend on their highly regulated cortical distributions, we investigated the localization of the BazX-82 protein. Baz localizes to the lateral cortex of the oocyte and is excluded from the Par-1 domain at the posterior pole (Benton and St Johnston, 2003b). The role of Baz at the lateral cortex is not known, but inappropriate accumulation of Baz at the posterior of the oocyte disrupts Par-1 function at the posterior and leads to defects in microtubule polarity (Benton and St Johnston, 2003b). To determine if the defects in microtubule polarity and germ plasm localization in bazX-82 oocytes could be due to a defect in Baz localization, we generated bazX-82-gfp transgenic flies and compared the localization of Baz-GFP and BazX-82-GFP in the oocyte. Consistent with previous findings, Baz-GFP localizes to the lateral cortex and is absent from the posterior domain (Fig. 5A) (Benton and St Johnston, 2003b). By contrast, BazX-82-GFP is found diffusely in the cytoplasm and is also targeted weakly and uniformly to the oocyte cortex (Fig. 5B).
In bazX-82 mutants, Par-1 localization mirrors that of Osk, with many oocytes lacking Par-1 at the posterior pole (Figs. 6A and B). Additionally, Par-1 is recruited to ectopic foci of Osk in the oocyte cytoplasm (Figs. 6C-6F). While the phenotypes we observe in bazX-82 oocytes could result from inhibition of Par-1 by BazX-82 at the posterior, the recessive nature of the bazX-82 allele suggests that the defects observed in bazX-82 oocytes are due to the loss of BazX-82 at the lateral cortex rather than its aberrant accumulation at the posterior. Consistent with this hypothesis, expression of BazX-82-GFP in otherwise wild-type ovaries does not produce the bazX-82 phenotype, as assayed by nuclear migration and Grk localization (100% wild-type oocytes; n=225) or abdominal segmentation (93% embryos with eight segments, 7% with partial fusion between two adjacent segments; n=385). The mislocalization of Par-1 in bazX-82 oocytes could thus be due to defects in polarization of the A-P axis of the oocyte rather than via a direct effect of Baz. Indeed, the effects of bazX-82 on microtubule organization are distinct from those observed in par-1 mutant oocytes which display no defects in nuclear migration or ooplasmic streaming (Shulman et al., 2000). Furthermore, the respective apical and basolateral localizations of Baz and Par-1 in the follicle cell epithelium (Benton and St Johnston, 2003a, 2003b) are maintained in bazX-82 egg chambers (Figs. 6G-6J), indicating that the interactions required for their mutually exclusive cortical localizations are still able to occur. Finally, in the strong baz4 allele (Muller and Wieschaus, 1996), we observe ectopic localization of Osk in the small fraction of egg chambers where an oocyte is specified (Fig. 3K and 3L). Together, these results argue that the loss of Baz function on the lateral cortex in bazX-82 oocytes, due to decreased Baz protein localization, decreased Baz protein activity, or both, results in defects in germ plasm localization.
The ectopic synthesis of Osk observed in bazX-82 egg chambers is surprising since Osk translation is tightly coupled to its posterior localization. To determine whether the ectopic translation of osk we observe in bazX-82 egg chambers is a general consequence of defects in oocyte microtubule polarity and the resulting effects on osk mRNA localization, we performed a literature search of previously characterized mutations affecting microtubule organization in the oocyte (Table 1). Ectopic localization of osk mRNA has been reported in numerous mutants documented to affect localization of Kin-βgal and therefore the organization of microtubule plus ends. These mutants include genes of no known function, such as mushroom-body expressed (mub) (Geng and Macdonald, 2007), factors involved in vesicle trafficking including several Rabs, (Coutelis and Ephrussi, 2007; Dollar et al., 2002; Jankovics et al., 2001; Tanaka and Nakamura, 2008), as well as genes predicted to affect the signaling cascade that triggers the reorganization of the microtubule cytoskeleton including lamininA (lanA) and grk (Deng and Ruohola-Baker, 2000; Roth et al., 1995). Although the distribution of Osk protein has not been established in many of these mutants, Osk is ectopically translated in six out of nine mutants where it has been examined, and the ectopic Osk localization frequently mirrors that of ectopic osk mRNA and Kin-βgal (Table 1). This suggests that osk translation is activated at locations where microtubule plus ends are focused. Mutations in three genes affecting vesicle trafficking appear to be the exception, whereby the accumulation of Kin-βgal and osk mRNA in the center of the oocyte does not lead to ectopic Osk expression (Coutelis and Ephrussi, 2007; Dollar et al., 2002; Tanaka and Nakamura, 2008) (see Discussion).
In egg chambers where the follicle cells are mutant for lanA, the signal from the follicle cells to the oocyte required for microtubule repolarization is not transduced. These egg chambers also display ectopic localization of Kin-βgal, osk mRNA and Osk protein to the center of the oocyte (Deng and Ruohola-Baker, 2000), suggesting that altered microtubule polarity alone is sufficient to cause ectopic translation of Osk. To test this hypothesis, we assessed the localization of Osk in grk mutants which similarly disrupt the signaling pathway leading to microtubule reorganization. In grk mutant egg chambers, osk mRNA is localized exclusively to the center of the oocyte (98%, n=55; Figs. 7A and B) (Roth et al., 1995). Endogenous Osk protein, detected by anti-Osk immunostaining, also accumulates ectopically in 37% of mid-stage oocytes (n=127; Figs. 7C and D), supporting the hypothesis that an enrichment of microtubule plus ends is sufficient to cause aberrant translational derepression of osk. Furthermore, Vas-GFP is recruited to the center of grk mutant oocytes (Figs. 7E and F), indicating that the ectopic accumulation of Osk is sufficient to direct germ plasm assembly. Interestingly, ectopic germ plasm assembly occurs at higher frequency and at earlier stages in grk mutant oocytes than in bazX-82 oocytes, which could result if microtubule plus ends are more concentrated centrally during mid-oogenesis in grk mutants than in bazX-82 oocytes. Similarly to bazX-82, grk mutant oocytes exhibit ectopic cortical patches of Vas at late stages (Figs. 7G and H), suggesting that embryos from grk mutant mothers would be likely to display ectopic germ plasm similar to that observed in bazX-82 embryos. We have been unable to test this hypothesis as grk mutant females lay very few eggs.
The assembly of germ plasm at the posterior of the oocyte dictates the future specification of germ cell fate at the posterior of the embryo and, through the localization of nos mRNA, abdominal development. Consequently, mechanisms that limit germ plasm assembly to the posterior of the oocyte are essential for embryonic patterning and fertility. Here, through the analysis of a novel baz allele, bazX-82, that supports oogenesis, we identify a role for Baz in the axis patterning during mid-oogenesis and the organization of the oocyte microtubule cytoskeleton. We show that germ plasm assembly can occur within the oocyte cytoplasm, indicating that cortical anchoring of osk mRNA and germ plasm proteins is not essential, and we propose that an enrichment of microtubule plus ends is sufficient for osk mRNA translation and germ plasm assembly.
Although the precise organization of oocyte microtubules remains unresolved, numerous studies using polarity indicators such as Kin-ßgal and Nod-ßgal as well as osk transport particle components provide evidence that the establishment of axial polarity in the oocyte requires the disassembly of the posterior MTOC by mid-oogenesis and the polymerization of new microtubules from sites along the anterior and lateral oocyte cortex. This reorganization results in an apparent anterior-posterior gradient of microtubules, with a slight enrichment of plus ends directed toward the posterior of the oocyte. This enrichment is in turn required for posterior localization of osk mRNA and ultimately germ plasm assembly (reviewed in Becalska and Gavis, 2009). Our analysis of bazX-82 indicates that baz plays multiple roles in establishing proper oocyte cytoskeletal polarity. The persistence of Nod-βgal at the posterior pole of the oocyte and the failure of nuclear migration indicates that the posterior MTOC is not disassembled in mid-stage bazX-82 oocytes. However, the osk and bcd localization defects produced by bazX-82 differ from those observed in grk mutants, which also fail to disassemble the posterior MTOC (Gonzalez-Reyes et al., 1995). Whereas osk accumulates centrally in grk mutant oocytes as early as stages 8-9, we do not observe ectopic osk mRNA before stage 10 in bazX-82 oocytes. In contrast to the accumulation of bcd at both ends of grk mutant oocytes, bcd is mislocalized along the lateral cortex of bazX-82 oocytes. In addition, the anterior accumulation of osk that occurs transiently during stages 8-9 in wild-type oocytes (Cha et al., 2002) persists to stage 10 in bazX-82 oocytes. We interpret these patterns to reflect a second requirement for baz in the regulation of microtubules nucleated at the anterior and lateral cortex. At later stages of oogenesis, cortical microtubule-dependent ooplasmic streaming is disrupted, whereas anterior microtubule-dependent bcd localization is not, suggesting that bazX-82 may affect the organization of specific subsets of microtubules within the oocyte rather than globally disrupting microtubule organization.
In the C. elegans embryo, Par proteins are segregated into mutually exclusive anterior and posterior cortical domains where they exert a local influence on microtubule dynamics (reviewed in Goldstein and Macara, 2007). In the Drosophila oocyte, Baz-GFP accumulates at the anterior and lateral cortex (Benton and St Johnston, 2003b) (Fig. 5). Unfortunately, we have not been able to visualize the endogenous BazX-82 protein due to the lack of antibodies capable of detecting Baz in the oocyte. However, given that BazX-82-GFP overexpression in the oocyte does not disrupt polarity, the decreased cortical and increased cytoplasmic accumulation of BazX-82-GFP as compared to wild-type Baz-GFP suggests that this altered distribution may be responsible for loss of baz function. In C. elegans Par-3/Baz is thought to promote polarization of the mitotic spindle by stabilizing microtubules at the anterior cortex (Labbe et al., 2003). Similarly, the requirement for Baz in microtubule reorganization could be mediated through the regulation of microtubule stability at the oocyte cortex. For example, a decrease in microtubule stability in bazX-82 oocytes could delay the establishment of the anterior-posterior gradient of microtubules or alter their net orientation, leading to the observed defects in osk localization. Interestingly, Par-1 is thought to promote destabilization of microtubules within the oocyte (Tian and Deng, 2009), suggesting that the distinct cortical domains established by Par1 and Baz could regulate microtubule organization via their differential effects on microtubule stability. Alternatively, bazX-82 could lead to a defect in microtubule modification or interaction with microtubule-associated proteins and thereby affect the organization of specific subsets of microtubules.
In wild-type oocytes, translation of osk mRNA occurs only upon its localization to the posterior. Our results suggest that the ectopic assembly of the germ plasm in bazX-82 oocytes results from the inappropriate translation of osk at an ectopic focus of microtubule plus ends and this hypothesis is further supported by the ectopic Osk observed in grk mutant oocytes. The defects in osk translational control following disruption of microtubule polarity suggest two models for the derepression of osk translation. First, since translation of osk occurs at a location within the oocyte where microtubule plus ends are enriched – either the posterior of a wild-type oocyte, or the center of a grk mutant egg chamber – factors necessary to activate osk translation could be localized toward the plus ends of microtubules, similarly to osk mRNA itself. Transport of these factors in particles distinct from those that transport osk mRNA would prevent premature translation. Consistent with this model, osk mRNA that is localized to the lateral cortex in kinesin heavy chain (khc) mutants, due to unopposed dynein-mediated minus-end directed transport, is translated and recruits the germ plasm component Vas (Cha et al., 2002). Alternatively, the high local concentration of osk mRNA that results from directed transport towards a bias of microtubule plus-ends could be sufficient to trigger translational activation by the local titration of translational repressors. This idea is supported by the ectopic localization and premature translation that occurs when osk is overexpressed (Smith et al., 1992; Zimyanin et al., 2007). In either case, germ plasm assembly may be amplified by a feed-back loop initiated by ectopic Osk, which can recruit additional Par-1, microtubule plus ends, and osk mRNA (Zimyanin et al., 2007). The accumulation of Par-1 together with ectopic Osk protein in the center of bazX-82 oocytes is consistent with this model. Surprisingly, in mutants that affect membrane trafficking in the oocyte and polarization of the oocyte microtubule cytoskeleton, ectopic osk is not translated, suggesting that ongoing vesicular trafficking may be required for the recruitment of factors that activate osk translation.
The ectopic patches we observe in late-stage bazX-82 and grk oocytes likely arise from the subsequent cortical association of germ plasm assembled ectopically in the oocyte cytoplasm. This suggests that while individual components of the germ plasm are normally excluded from the oocyte cortex, a pre-assembled germ plasm complex can become ectopically anchored. Cortical patches are detected only after the onset of nurse cell dumping but their presence does not correlate with defects in ooplasmic streaming, which occur in bazX-82 egg chambers (Fig. 4) but not in grk mutant oocytes (data not shown). Thus, ectopic foci of germ plasm may be displaced to the cortex by the rapid influx of nurse cell cytoplasm or may reach the cortex by diffusion, similarly to nos RNPs (Forrest and Gavis, 2003). Interestingly, localization of Osk protein to the posterior of the oocyte leads to the formation of cortical F-actin projections which may be involved in the actin-dependent anchoring of the germ plasm at the posterior pole (Vanzo et al., 2007). The ectopic germ plasm in bazX-82 and grk oocytes might therefore become cortically anchored by Osk protein-dependent local modification of the actin cytoskeleton. Importantly, our results indicate that polarity of the microtubule cytoskeleton dictates not only the localization of osk mRNA to the posterior of the oocyte, but also spatial restriction of germ plasm assembly. Tight regulation of microtubule organization within the oocyte is thus critical to abdominal segmentation and germ cell development.
Supplemental Fig. 1. bazX-82 acts in the germline to affect axis patterning. (A,B) Confocal images of egg chambers without (A) and with (B) bazX-82 follicle cell clones, stained for F-actin (red) and nuclei (blue). Mutant follicle cell clones are marked by the absence of GFP (green). Arrows point to oocyte nuclei. bazX-82 follicle cell clones do not disrupt oocyte nucleus migration.
We thank C. Nüsslein-Volhard, P. Macdonald, T. Schüpbach, and M. Metzstein for fly stocks, D. St Johnston for fly stocks, baz cDNA, and anti-Par-1 antibody, R. Lehmann for fly stocks and anti-Vas antibody, and A. Wodarz for anti-Baz antibody. We also thank J. Goodhouse for assistance with confocal microscopy and R. Burdine, J. Lee, and K. Sinsimer for comments on the manuscript. This work was supported by a grant from the National Institutes of Health (GM067758) to E.R.G. and by an NSERC postgraduate fellowship (A.N.B.).
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.