A late phase of germ plasm accumulation during Drosophila oogenesis requires lost and rumpelstiltskin

Abstract

Asymmetric mRNA localization is an effective mechanism for establishing cellular and developmental polarity. Posterior localization of oskar in the Drosophila oocyte targets the synthesis of Oskar to the posterior, where Oskar initiates the assembly of the germ plasm. In addition to harboring germline determinants, the germ plasm is required for localization and translation of the abdominal determinant nanos. Consequently, failure of oskar localization during oogenesis results in embryos lacking germ cells and abdominal segments. oskar accumulates at the oocyte posterior during mid-oogenesis through a well-studied process involving kinesin-mediated transport. Through live imaging of oskar mRNA, we have uncovered a second, mechanistically distinct phase of oskar localization that occurs during late oogenesis and results in amplification of the germ plasm. Analysis of two newly identified oskar localization factors, Rumpelstiltskin and Lost, that are required specifically for this late phase of oskar localization shows that germ plasm amplification ensures robust abdomen and germ cell formation during embryogenesis. In addition, our results indicate the importance of mechanisms for adapting mRNAs to utilize multiple localization pathways as necessitated by the dramatic changes in ovarian physiology that occur during oogenesis.

Fig. 1.

Lost interacts with the nos localization factor Rump. (A,B) Co-immunoprecipitation analysis of Rump (A) and Lost (B) from wild-type (WT), rump¹ (r¹), or lost¹ (l¹) Drosophila ovary extracts with (+) or without (–) RNase treatment. Lost and Rump are detected by immunoblotting with the respective antibodies as indicated. The input samples contain 1-2% of the amount of extract used for immunoprecipitation (IP). The enhancement of co-immunoprecipitation in the presence of RNase could be due to protein conformational changes that result in enhanced antibody or increased protein accessibility.

Fig. 2.

Identification and characterization of Lost. (A) Genomic region surrounding the CG14648 (lost) locus, including upstream genes CG14647 and CG9853. The organization of the lost transcript, including the predicted transcriptional start site, is shown with coding sequence indicated in gray and 5′ and 3′ UTRs in white. FlyBase predicts an additional transcriptional start site but the coding potential of the transcript is unclear and we have no evidence of an additional isoform. The arrowhead marks the site of the GFP gene-trap P element insertion CG14648^ZCL3169 in the first intron; the extent of the lost¹ lesion created by imprecise excision of this P element is indicated by the dashed line and parentheses. The deletion removes both of the predicted CG14648 transcriptional start sites. The region contained in the genomic lost rescue transgene (glost) is indicated. (B) Northern blot analysis of lost mRNA in wild-type (WT), CG14648^ZCL3169 (GFP-lost) and lost¹ ovaries. rp49 serves as a loading control. (C-E) Confocal images of GFP-Lost (green) in fixed stage 7-8 (C), stage 10 (D) and stage 13 (E) egg chambers. Arrowheads indicate anterior enrichment, arrows indicate posterior enrichment. nc, nurse cells; oo, oocyte. (F-K) Confocal images showing the posterior poles of live pre-blastoderm (F-H) and blastoderm stage (I-K) embryos expressing GFP-Lost (green) and nos^*RFP (red). pc, pole cells. (L) In situ hybridization to nos mRNA in early wild-type embryos (WT), lost¹ embryos or lost¹ embryos with the genomic lost rescue transgene (glost; lost¹). Anterior is to the left. Right-hand panels show enlargements of the posterior for each genotype. Data from in situ hybridization were used to classify posterior localization as wild type, diffuse or reduced; quantification of these patterns is shown on the right (WT, n=116; lost¹, n=292; glost; lost¹, n=103).

Fig. 3.

Lost and Rump are required for osk localization. (A,B) In situ hybridization to nos (A) or osk (B) in early wild-type (WT) or lost^– rump^– embryos. Right-hand panels show enlargements of the posterior for each genotype. Quantification of posterior localization patterns observed by in situ hybridization is shown on the right for each mRNA (nos: WT, n=116; lost^– rump^–, n=85. osk: WT, n=166; lost^– rump^–, n=60). (C) Immunoblot analysis of Osk levels in wild-type (WT) or lost^– rump^– 0- to 2-hour-old embryo extracts. Both protein isoforms, Osk^L and Osk^S, can be detected in all wild-type and lost^– rump^– samples. Densitometry analysis shows that Osk is reduced by ∼75% in lost^– rump^– embryos compared with wild type. Snf serves as loading control. (D) Confocal z-series projections showing the posterior of wild-type (WT) or lost^– rump^– (l^– r^–) blastoderm embryos immunostained for Vas (Blue). Nuclei are labeled with DAPI (green). Graph shows the total number of pole cells per embryo (solid circles for WT, open circles for lost^– rump^–), determined by counting cells that were both Vas- and DAPI-positive. The average number of pole cells in lost^– rump^– embryos is significantly less than the wild-type average (P<0.0001). (E) Fluorescence in situ hybridization to osk (red) in wild-type (WT) or lost^– rump^– egg chambers at the indicated stages. Nuclei are stained with DAPI (blue). Anterior is to the left. All of stage 9 lost^– rump^– oocytes (n=33) and 98% of stage 10 lost^– rump^– oocytes (n=52) exhibit normal posterior accumulation of osk mRNA. (F) Immunoblot analysis of Osk in extracts from wild-type (WT) or lost^– rump^– (l^– r^–) oocytes at mid-oogenesis (stage 9/10a) and late oogenesis (stage 12/13). Kinesin heavy chain (Khc) provides a loading control. Densitometry analysis shows no difference in the total amount of Osk protein between mid-stage wild-type and lost^– rump^– oocytes (<5%, darker exposure). By contrast, total Osk is reduced by 30% in late-stage lost^– rump^– oocytes compared with wild type (lighter exposure).

Fig. 4.

Continued accumulation of germ plasm during late stages of oogenesis. (A-D) Time-lapse confocal projections of the posterior of a live Drosophila oocyte expressing osk^*GFP. The time course captures the localization of osk^*GFP starting prior to the onset of nurse cell dumping, stage 10B (A), and ending after nurse cell dumping is complete, stage 12 (D). Elapsed time (t) is indicated; posterior is up. (E-H) Fluorescence recovery after photobleaching (FRAP) experiment performed on osk^*GFP at the posterior of a wild-type stage 12 oocyte. The photobleached region of interest (ROI) is indicated by the rectangle. The zero time point corresponds to the completion of photobleaching in the ROI. Time elapsed in the FRAP recovery period is indicated; posterior is up. (I-L) Time-lapse confocal projections of osk^*GFP at the posterior of a live lost^– rump^– oocyte from stage 10B (I) to stage 12 (L). Elapsed time is indicated; posterior is up. (M-P) FRAP experiment performed on osk^*GFP at the posterior of a stage 12 lost^– rump^– oocyte. The photobleached ROI is indicated by the rectangle. The zero time point corresponds to the completion of photobleaching in the ROI. Time elapsed in the FRAP recovery period is indicated; posterior is up. (Q,Q′) The total amount of posterior osk^*GFP fluorescence in time-lapse confocal z-stack images was quantified for six wild-type and five lost^– rump^– oocytes (see Materials and methods) and the net change in fluorescence intensity between the initial time point (equivalent to panels A or I, labeled as t₀) and final time point (equivalent to panels D or L, labeled as t₃) was plotted for each oocyte (Q). Plot of the average net change in fluorescence intensity ± s.e.m. for each genotype is shown in Q′. (R,S) Quantification of the FRAP experiments from the wild-type (R) or lost^– rump^– (S) oocytes shown in E-H and M-P. Similar results were observed for two additional wild-type and three additional lost^– rump^– stage 12 egg chambers. (T,U) Quantification of the volume (μm³) occupied by osk^*GFP at the posterior of individual wild-type (solid circles) and lost^– rump^– (open circles) oocytes at stage 10A (T) and stage 12/13 (U). The mean value for the volume of osk^*GFP fluorescence (horizontal bars) in lost^– rump^– oocytes is not significantly different from wild type at stage 10A but is significantly smaller by stage 12/13 (P<0.0086).

Fig. 5

. Continued posterior accumulation of GFP-Vasa during late oogenesis is dependent on lost and rump. (A-H) Time-lapse confocal projections of the posterior of live wild-type (A-D) and lost^– rump^– (E-H) oocytes expressing GFP-Vas, from stage 10B (A,E) to stage 12 (D,H). Elapsed time (t) is indicated; posterior is up. (I) The net change in the total amount of GFP-Vas fluorescence at the posterior of individual wild-type (WT, solid circles; n=6) and lost^– rump^– (open circles; n=9) oocytes between t₀ and t₃ was quantified as described in Fig. 4.

Fig. 6.

Rump associates with osk both in vitro and in vivo. (A) UV-crosslinking of MBP-Rump (^*Rump) or MBP-Lost (^*Lost) to ³²P-labeled nos +2′, osk 5′UTR, or osk 3′UTR RNA probes. The protein-RNA complex is indicated by the arrow. NP, no protein added. (B) UV-crosslinking of MBP-Rump to ³²P-labeled probes comprising osk 3′UTR nucleotides 1-738, nucleotides 731-1004, or the entire osk 3′UTR (nucleotides 1-1004). Binding reactions contained increasing concentrations of MBP-Rump as indicated. The Rump-osk complex is indicated by the arrow. (C) RT-PCR to detect osk co-immunoprecipitated with Rump from wild-type (WT) and rump^– ovary extracts. Reactions were performed without (–RT) or with (+RT) reverse transcriptase. rump^– extract serves as a negative control to demonstrate antibody specificity.