Makorin 1 controls embryonic patterning by alleviating Bruno1-mediated repression of oskar translation

Abstract

Makorins are evolutionary conserved proteins that contain C3H-type zinc finger modules and a RING E3 ubiquitin ligase domain. In Drosophila, maternal Makorin 1 (Mkrn1) has been linked to embryonic patterning but the mechanism remained unsolved. Here, we show that Mkrn1 is essential for axis specification and pole plasm assembly by translational activation of oskar (osk). We demonstrate that Mkrn1 interacts with poly(A) binding protein (pAbp) and binds specifically to osk 3' UTR in a region adjacent to A-rich sequences. Using Drosophila S2R+ cultured cells we show that this binding site overlaps with a Bruno1 (Bru1) responsive element (BREs) that regulates osk translation. We observe increased association of the translational repressor Bru1 with osk mRNA upon depletion of Mkrn1, indicating that both proteins compete for osk binding. Consistently, reducing Bru1 dosage partially rescues viability and Osk protein level in ovaries from Mkrn1 females. We conclude that Mkrn1 controls embryonic patterning and germ cell formation by specifically activating osk translation, most likely by competing with Bru1 to bind to osk 3' UTR.

Fig 1. Mkrn1 alteration affects ovarian development.

(A) Relative Mkrn1 mRNA levels (normalized to RpL15 mRNA) at various stages of development, measured by quantitative RT-PCR. Error bars depict Stdev, n = 3. (B) Schematic diagram of the proteins encoded by the Mkrn1 alleles used to analyze its function in vivo. Mkrn1^N is a null allele and produces no protein. (C-F) Bright-field micrographs of entire ovaries from wild-type and Mkrn1 mutant flies. Note the reduced size of Mkrn1^S and Mkrn1^N ovaries. Scale bars, 500 μm. (G-J) Individual egg chambers stained with the DNA marker DAPI. Fewer stage 10 and older egg chambers are present in Mkrn1^S and no late stage egg chambers are present in Mkrn1^N ovaries. Abscission defects resulting from inappropriate follicle cell migration are frequently observed in Mkrn1^N ovaries (J, arrow). Scale bars, 20 μm. (K-N) Individual egg chambers stained with α-Lamin to highlight nuclear membranes. Scale bars, 20 μm. (M) The oocyte nucleus (marked with an arrow in K, L, and M) remains at the posterior of Mkrn1^S oocytes. (N) Some Mkrn1^N egg chambers have 16 germline cells whose nuclei are all of similar size, suggesting a defect in oocyte differentiation. Note also irregularities in the follicle cell monolayer in the Mkrn1^N egg chamber. (O-Q) Dark-field photographs of eggs and embryos produced by wild-type and Mkrn1 mutants Scale bars, 100 μm. (P) Most embryos produced by Mkrn1^W females have a posterior-group phenotype. (Q) Eggs produced by Mkrn1^S females lack dorsal appendages and do not support embryonic development. (R, S) Immunostaining with α-Ftz (red) and α-Vas (green) reveals segmentation defects and the absence of pole cells in Mkrn1^W embryos Scale bars, 50 μm. (T, U) Surface images of embryos immunostained with α-Ftz (red) to better illustrate segmentation defects in Mkrn1^W embryos. Scale bars, 50 μm.

Fig 2. Mkrn1 accumulates in pole plasm.

(A-C) The three panels show the same egg chambers stained for (A) Venus-Mkrn1, (B) Stau, and a (C) merged image. Scale bars, 25 μm. Venus-Mkrn1 expression was driven by nos>Gal4. Colocalization of Venus-Mkrn1 and Stau can be observed in particles that have not yet accumulated at the posterior of the early stage 8 oocyte. (D-F) The three panels show the same stage 10 egg chamber stained for (D) Venus-Mkrn1, (E) Stau and (F) a merged image. Scale bars, 25 μm. There is extensive colocalization of Venus-Mkrn1 and Stau in the posterior pole plasm of the oocyte. (G-I) The three panels show the same egg chambers stained for (G) Venus-Mkrn1, (H) osk mRNA, and (I) a merged image. Scale bars, 25 μm. Colocalization of Venus-Mkrn1 and osk can be observed in an early stage 8 oocyte where osk has not yet fully localized at the posterior of the oocyte. (J-L) The three panels show the same stage 10 egg chamber stained for (J) Venus-Mkrn1, (K) osk mRNA and (L) a merged image. Scale bars, 25 μm. There is extensive colocalization of Venus-Mkrn1 and osk mRNA in the posterior pole plasm of the oocyte. (M-O) The three panels show the same egg chambers stained for (M) Venus-Mkrn1, (N) Aub, and a (O) merged image. Scale bars, 5 μm. Colocalization of Venus-Mkrn1 and Aub can be observed at the nuage surrounding the nurse cell nuclei (P-R). The three panels show the same egg chambers stained for (P) Venus-Mkrn1, (Q) Aub, and a (R) merged image. Scale bars, 20 μm. There is extensive colocalization of Venus-Mkrn1 and Aub in the posterior pole plasm of the oocyte.

Fig 3. Mkrn1 mutations affect accumulation of proteins involved in axis patterning.

(A-D) Posterior accumulation of Osk is greatly reduced in stage 10 Mkrn1^W and Mkrn1^S oocytes as compared with wild-type. Osk is nearly undetectable in Mkrn1^N egg chambers. Scale bars, 25 μm. (E-H) Posterior accumulation of Stau is greatly reduced in stage 10 Mkrn1^W and Mkrn1^S oocytes as compared with wild-type. Stau is nearly undetectable in Mkrn1^N egg chambers. Scale bars, 25 μm. (I-L) Anterodorsal accumulation of Grk is normal in stage 10 Mkrn1^W oocytes. Grk remains associated with the oocyte nucleus and is mislocalized to the posterior in stage 10 Mkrn1^S oocytes. Grk is present at uniformly low levels or undetectable levels in all germ cells in Mkrn1^N egg chambers. Scale bars, (I-K) 20 μm, (L) 25 μm. (M-P) Posterior accumulation of Aub is greatly reduced in stage 10 Mkrn1^W and Mkrn1^S oocytes as compared with wild-type. Aub is present at uniform levels in all germ cells in Mkrn1^N egg chambers. Scale bars, 20 μm. (Q-T) Posterior accumulation of Vas is greatly reduced in stage 10 Mkrn1^W and Mkrn1^S oocytes as compared with wild-type. Vas is present at uniform levels in all germ cells in Mkrn1^N egg chambers. Scale bars, 25 μm. (U-X) Accumulation of Orb is similar in wild-type and Mkrn1^W oocytes, but Orb is more concentrated in the posterior of Mkrn1^S oocytes. In early-stage Mkrn1^N egg chambers there is usually a single Orb-positive cell, indicating that some steps toward oocyte differentiation are able to take place. Scale bars, (U-W) 20 μm, (X) 25 μm.

Fig 4. Translation of osk mRNA is impaired in Mkrn1^W ovaries.

(A) Fluorescent in situ hybridization for osk mRNA (red) with co-immunostaining for Osk protein (green) in wild-type and Mkrn1^W egg chambers. For each genotype and in each column the top and bottom images are of the same egg chamber. In wild-type oocytes posterior accumulation of osk mRNA and Osk protein is robust and stable from stage 9 onward. In Mkrn1^W oocytes accumulation of osk mRNA resembles the wild-type pattern through stage 10A but is not maintained, while Osk protein is rarely detectable at the oocyte posterior at any stage. Scale bars, 25 μm. (B) Western blot analysis from ovary lysates of various genotypes stained for Osk, pAbp and β-tubulin. Osk protein levels are greatly reduced in all Mkrn1 mutant alleles. 1 day-old young females have not yet completed oogenesis and were used as a control for Mkrn1^S and Mkrn1^N ovaries which also lack late-stage egg chambers, where Osk is most abundant. (C) RT-qPCR experiments measuring ovarian osk mRNA levels (normalized to RpL15 mRNA) in the same genotypes as (B). mRNA levels of ovaries from adult females were compared to Mkrn1^W ovaries. For Mkrn1^S and Mkrn1^N ovaries, mRNA levels of 1 day-old young ovaries was used as normalization. Error bars depict Stdev, n = 2.

Fig 5. Mkrn1 interacts strongly with the poly(A) binding protein.

(A) Western blot analysis of co-IP experiments between Venus-Mkrn1 and pAbp. α-Tubulin (lanes 1, 2) and ovaries lacking the Venus-Mkrn1 transgene (lane 4) were used as negative controls. (B) Co-expression of pAbp stabilizes Mkrn1. FLAG-Mkrn1 was co-transfected with increasing levels of HA-pAbp in S2R+ cells. Left: Proteins were examined using immunoblotting. Right: Intensities of FLAG-Mkrn1 levels were quantified and normalized to intensities of β-tubulin. The relative intensity was normalized to Mkrn1 mRNA levels (normalized to RpL15 mRNA) analyzed by RT-qPCR. Error bars depict SEM, n = 9. (C) PAM2 motif alignment in different species. Comparison between Drosophila and human PAM2 motif revealed a Glu to Val substitution (orange) in the consensus sequence. The conserved amino acid sequence to Drosophila (dark purple) is indicated below. The PAM2 motif was mutated using two amino acid substitutions at positions 90 and 92 to alanine (F90A and P92A). (D) Immunoblot analysis of co-IP experiments between FLAG-Mkrn1 and pAbp in S2R+ cells. The interaction of pAbp and Mkrn1 is reduced when the PAM2 motif is mutated. (E) Rescue experiments of wild-type or mutant Mkrn1 in Mkrn1^N mutants. FLAG-Mkrn1 or FLAG-Myc-Mkrn1^PAM2 was overexpressed in ovaries using a nos>Gal4 driver line. Ovarian protein lysates from rescued females were analyzed by immunoblotting together with wild-type and Mkrn1^N as controls. While Mkrn1 overexpression could restore Osk protein level to approximately wild-type, Mkrn1^PAM2 depicted decreased protein levels similar to Mkrn1^N. (F) Egg chambers from (top panel) Mkrn1^W, (middle panel) pAbp/+, and (bottom panel) pAbp/+; Mkrn1^W females stained with DAPI and immunostained for Orb, an oocyte marker. pAbp/+; Mkrn1^W ovaries show diverse developmental defects. Scale bars, 25 μm except for the lower left panel, where the scale bar is 50 μm. (G) Time course of fecundity of Mkrn1^W and pAbp/CyO; Mkrn1^W females.

Fig 6. Mkrn1 associates specifically with the 3’ UTR of osk mRNA.

(A) RIP experiment. Either FLAG-tagged Mkrn1 or Mkrn1^ΔZnF1 was overexpressed in Mkrn1^N ovaries using nos>GAL4 driver. Enrichment of different transcripts was analyzed by RT-qPCR using primers that bind to the respective 3’ UTRs. Fold enrichment is presented relative to the control (nos>GAL4 driver alone, ctrl). Error bars depict SEM, n = 3. Multiple t-test was used to analyze significant changes compared to control RIP. (B) iCLIP results from S2R+ cells showing specific binding of Mkrn1 to osk in a region of the 3’ UTR that partially overlaps with the BRE-C site (yellow). The peaks (purple) indicate crosslinking events of Mkrn1 to osk. Data of two technical replicates for FLAG-Mkrn1 is shown. The same experiment performed with FLAG-GFP (ctrl) did not show specific peaks. (C) RIP experiments of FLAG-Mkrn1^RING in S2R+ cells. Enrichment of luciferase-osk-3’UTR transcript was analyzed by RT-qPCR compared to IP experiments with FLAG-GFP. Mkrn1^RING Binding to osk 3’ UTR is compromised when introducing a deletion of the Mkrn1 binding site (osk^ΔMkrn1, deletion of nucleotides 955–978 of osk 3’ UTR). Error bars depict SEM, n = 4. (D) Binding of Mkrn1^RING to luciferase-osk-3’UTR reporter is reduced in S2R+ cells when using a deletion of the A-rich region of osk 3’ UTR (osk^ΔAR, deletion of nucleotides 987–1019 of osk 3’ UTR). Fold change illustrates the difference of pulled-down osk^ΔAR reporter compared to IP with wild-type osk 3’ UTR. The RIP experiments were normalized to FLAG-tagged GFP. Error bars depict SEM, n = 4. (E) RIP experiments were performed in either control cells or upon depletion of Imp or pAbp. Enrichment was calculated compared to FLAG-GFP. The relative change in Mkrn1^RING binding to luciferase-osk-3’UTR reporter in knockdown cells compared to LacZ-depleted cells is depicted. Depletion of pAbp compromises binding of Mkrn1. Error bars depict SEM, n = 3. (F) RIP experiments showing that FLAG-Mkrn1 binding to luciferase-osk-3’UTR is dependent on the ZnF1 domain as well as on the PAM2 motif. Enrichment was analyzed using RT-qPCR and the relative change in binding compared to RIP of FLAG-Mkrn1 is illustrated. RIP experiments were performed in S2R+ cells using FLAG-GFP as control. Error bars depict SEM, n ≥ 3.

Fig 7. Mkrn1 competes with Bru1 for binding to osk mRNA.

(A) RIP experiments in either control S2R+ cells, or upon knockdown of Mkrn1. Binding of the indicated proteins to the luciferase-osk-3’UTR reporter was monitored by RT-qPCR. The relative fold change in recovered RNA upon Mkrn1 knockdown is illustrated, compared to RIP experiments in LacZ-depleted cells. For every RIP experiment, the enrichment was calculated using GFP. Error bars depict SEM, n ≥ 4. (B) Bru1 binding to luciferase-osk-3’UTR upon pAbp knockdown was analyzed using GFP-RIP with subsequent RT-qPCR. The relative fold change in binding of GFP-Bru1 to luciferase-osk-3’UTR compared to control knockdown is illustrated. The individual enrichments were normalized to IP experiments using GFP alone. Error bars depict SEM, n = 3. (C) RIP experiments in either heterozygous (wild-type) or homozygous Mkrn1^W ovaries using α-Bru1 antibody. The relative fold change in recovered endogenous osk mRNA in wild-type compared to Mkrn1^W ovaries is depicted. As control RIP, normal IgG was used for every condition. Error bars indicate SEM, n = 4. (D-G) Immunostaining experiments showing Bru1 distribution in (D-E) wild-type and (F-G) Mkrn1^W early-stage egg chambers. Note the more prominent accumulation of Bru1 in the oocyte in the Mkrn1 mutant. Scale bars, (D and F) 25 μm; (E and G) 20 μm. (H-M) Stage 10 egg chambers of the indicated genotypes immunostained with α-Osk. Posterior accumulation of Osk is restored to a variable degree (K-M) in Mkrn1^W oocytes when heterozygous for bru1. Scale bars, 25 μm. (N) Quantification of posterior Osk localization in oocytes depicted in (H-M). The thresholds for weak, moderate, and strong are arbitrary. The photographs in the H-M panels illustrate what is meant by the different categories. n = 21 for wild-type, n = 52 for Mkrn1^W, n = 100 for bru1/+; Mkrn1^W. (O) Immunoblot analysis of protein lysates from ovaries depicted in (D-G). Heterozygous mutation of bru1 led to an increase of Osk protein levels in Mkrn1^w ovaries.

Fig 8. Model depicting activation of osk translation via Mkrn1.

Mkrn1 is recruited to the osk 3’ UTR and stabilized by pAbp. The recruitment of Mkrn1 leads to the displacement of Bru1 promoting translational activation at the posterior pole of the oocyte.