Fly strains

All fly stocks were maintained at 25°C or 18°C. Either the yw strain or the respective heterozygote was used as a control. A TRiP line, HMS04468 (BL#57024, Bloomington Drosophila Stock Center, IN, USA), was used to knockdown dMarf1 expression. The following drivers were used for transgene or RNAi expression: NGT40 and nosGal4-VP16, Matα Gal4-VP16 (BL#7063) (for germline cells) and traffic jam-Gal4 (for somatic follicle cells). A loss-of-function allele of dMarf1, dMarf1^KO321, was generated using the CRISPR-Cas9 system (details are described in the S1 Protocol). The resultant mutant strain was back-crossed three times with the wild-type fly, and maintained over CyO balancer chromosome. To generate flies expressing dMarf1 transgene variants, the coding sequences of full-length dMarf1 or its truncated mutants (ΔN, ΔOST, ΔRRM, ΔC, and ΔC47) were PCR amplified using the cDNA clone RE61489 as a template. Triple repeats of the Myc-epitope tag were added to the N-terminal region. Purified PCR products were assembled using the In-Fusion system (Takara Bio, Japan) with an XbaI-digested pUASp-attB vector (Drosophila Genomics Resource Center). The resultant plasmids were injected into embryos for insertion at the phiC31 integration site—attP-3B—at 65B2 (BL#32543) using the standard protocol [9,10]. The primers used in this study are shown in S5 Table.

Isolation of egg chambers

Female flies were fed with wet-yeast paste for 2 to 3 days, and their ovaries were dissected in oocyte isolation buffer (100 mM HEPES-Na, pH 7.4, 55 mM NaOAc, 40 mM KOAc, 110 mM sucrose, 1.2 mM MgCl₂, 1 mM CaCl₂) to prevent oocyte activation [11] Egg chambers were dissected from ovarioles by using tungsten needles under the dissection microscope. Stages of egg chambers were assigned based on the standard King’s criteria [12]. Each group of egg chambers were collected into a 1.5 ml tubes, homogenized by the plastic pestle, and examined for protein expression by Western blotting analysis.

Immunostaining

Germaria and egg chambers (stage 1–12) were immunostained as described previously [13]. For staining stage 13–14 egg chambers, 10–20 ovaries were dissected from the adult females and immediately transferred to 1.5 ml tube containing ~1 ml of 100% methanol. To disrupt the eggshell and facilitate antibody penetration, ovaries were sonicated 20 times for 0.5 second each by using the sonicator (TOMY UD-201) with the following settings: Output 2 and Duty 50. The sonicated ovaries were washed with 100% methanol and sequentially incubated at 25°C in 90% methanol for 60 min, in 75% methanol/PBS for 30 min, in 50% methanol/PBS for 30 min, in PBS for 30 min, and in PBS containing 1.0% Triton X-100 for 60 min, followed by the conventional immunostaining as described previously [13]. For immunostaining of unfertilized eggs, females were mated with can¹² homozygous mutant males [14], which cannot produce mature sperms. Thus, females were induced to lay unfertilized eggs. Immunostaining was performed using the following primary antibodies: mouse monoclonal anti-Myc (1:400 dilution) (9E10, Wako, Japan), mouse anti-α-tubulin (1:400) (DM1A, Santa Cruz Biotechnology, TX, USA), and rabbit anti-phospho-histone H3 (1:400) (#06–570, Merck Millipore, MA, USA), and the following secondary antibodies: Alexa Fluor 488-, or 555-conjugated goat anti-rabbit, and anti-mouse IgG (1:400) (Thermo Fisher Scientific, MA, USA). The chromosomes and actin filament were stained with 4′,6-Diamidine-2′-phenylindole (DAPI, Sigma-Aldrich, MO, USA) and rhodamine-phalloidin (Cytoskeleton, CO, USA), respectively. Images were acquired using an Olympus FV1000 or Zeiss LSM780 confocal microscope.

Immunoprecipitation

Myc-tagged full-length dMarf1 (Myc-dMarf1-FL) was expressed in germline cells by crossing with the Matα driver. Approximately 300 ovaries were collected and homogenized using a hand pestle in the oocyte isolation buffer containing proteinase inhibitor cocktail (Roche, Switzerland) and 0.1% Triton X-100 (Wako). The lysate was cleared by two rounds of centrifugation at 20,000 Relative Centrifugal Force (RCF). Myc-dMarf1-FL protein was immunoprecipitated from the cleared ovary lysate using anti-c-Myc antibody-conjugated magnetic beads (Thermo Fisher Scientific). Anti-HA antibody magnetic beads (Thermo Fisher Scientific) were used as a negative control. After washing the beads with phosphate-buffered saline (PBS) containing 0.1% Triton X-100, TRIzol^TM (Invitrogen, OR, USA) was added to extract the bound RNA.

Generation of anti-dMarf1 antibody

The C-terminal nucleotide sequence of dMarf1 (901–1305 residues) was amplified and ligated into pENTR^TM/D-TOPO (Invitrogen), and recombined into pDEST17 to generate a His-tag fusion protein, His-dMarf1-C, using the Gateway system (Invitrogen). The His-dMarf1-C protein expressed in E. coli was purified using Ni Sepharose^TM 6 beads (GE Healthcare) according to the manufacturer’s protocol. The purified protein combined with complete adjuvant was injected into guinea pigs. Serum was obtained by centrifugation and heat inactivated at 56°C for 30 min.

Quantitative RT-PCR

RNA was extracted from the ovaries or immunoprecipitates using TRIzol^TM reagent (Invitrogen), according to the manufacturer's protocol. The purified RNA (~1 μg) was treated with DNase I (New England Biolabs, MA, USA) to remove any contaminating genomic DNA. DNase I was inactivated at 70°C for 10 min in the presence of EDTA, and the complementary DNA was synthesized using superscript III (Invitrogen). Real-time PCR was performed using Fast SYBR Green system (Thermo Fisher Scientific) with StepOnePlus (Applied Biosystems, CA, USA). Act5C was used for normalization.

mRNA sequencing and analyses

Total RNAs were extracted from hand-dissected ovaries from dMarf1^KO321/CyO heterozygous and dMarf1^KO321 homozygous females using TRizol (ThermoFisher) in accordance with the manufacturer’s instruction. Quality of the purified RNA was examined by Agilent 2100 Bioanalyzer (Agilent) prior to the library construction. RNAs with poly(A) tails were enriched by using oligo-dT beads. RNAs containing poly(A) tails were purified with oligo-dT beads, fragmented and reverse transcribed. Adapters were ligated and the libraries were amplified by PCR. DNA fragments with small size (less than 200 bp) were removed by AMPure XP beads (Beckman Coulter). Libraries were sequenced with Illumina HiSeq sequencer (Illumina) on a 100 cycle Single End Read sequencing run. The cDNA library preparation and next-generation sequencing were performed at Hokkaido System Science (Hokkaido, Japan). Total reads obtained by the Illumina HiSeq sequencing were 23,867,733 and 23,608,608 reads for dMarf1^KO321 heterozygous control and homozygous mutant, respectively. Reads were mapped to the Drosophila genome (Release 6) using the TopHat2 program [15]. The mapping rate was 88.8% and 88.6% for dMarf1^KO321 heterozygous control and homozygous mutant, respectively. The differentially-expressed genes were detected by Cuffdiff program from the Cufflinks package [16].

RNA-IP-seq and analyses

Myc-dMarf1 was immunoprecipitated from the ovarian lysate using anti-c-Myc antibody-conjugated magnetic beads as described in Materials and Methods. Immunoprecipitate with anti-HA antibody-conjugated magnetic beads was used as a negative control. RNAs bound to the magnetic beads were isolated by TRizol (ThermoFisher) in accordance with the manufacturer’s instruction. Quality of the purified RNA was checked by Agilent 2100 Bioanalyzer prior to the library construction. The following RNA purification and the library construction were conducted as described above. Libraries were sequenced with Illumina HiSeq sequencer on a 150 cycle Pair-End Read sequencing. The cDNA library preparation and next-generation sequencing were performed at Gene-Nex (Japan). Total reads obtained by the Illumina HiSeq sequencing were 14,273,308 pairs and 10,686,509 pairs for Myc-dMarf1 bound and control fraction, respectively. Among reads, low-quality reads and adaptor sequences (PE1; TACACTCTTTCCCTACACGACGCTCTTCCGATCT, PE1_rc; AGATCGGAAGAGCGTCGTGTAGGGAAAGAGTGTA, PE2; GTGACTGGAGTTCAGACGTGTGCTCTTCCGATCT, PE2_rc; AGATCGGAAGAGCACACGTCTGAACTCCAGTCAC) were trimmed by Trimmomatic software [17]. Around 2.6% reads were trimmed in both libraries. The resulting reads were mapped to the Drosophila genome (Release 6) using the TopHat2 program [15]. The mapping rate was 89.3% and 89.8% for Myc-dMarf1 bound and control fraction, respectively. The differentially expressed genes were detected by Cuffdiff program from the Cufflinks package [16].

Quantitative proteome analysis

The heterozygous and homozygous ovaries of dMarf1^KO321 were dissected in oocyte isolation buffer. 50 of stage 14 egg chambers were picked up into 1.5 ml tubes, and the buffer was replaced with 50 μl of 50 mM Tris-Cl (pH 8.0) buffer containing 2% deoxycholate and proteinase inhibitor cocktail. The egg chambers were homogenized by the plastic pestle, and the lysates were cleared by centrifugation at 20,000 RCF for 1 min at 4°C. Homogenization and centrifugation were repeated twice more. After third homogenization, lysates were further cleared by centrifugation at 20,000 RCF for 15 min at 4°C. The supernatants (~30 μl) were collected in low-protein binding tubes (Eppendorf, Germany). The protein concentrations of the lysates were between 0.95 to 1.37 μg/ml. Proteins extracted from egg chambers were reduced, alkylated, and digested with trypsin. The peptides were labeled using the Tandem Mass Tag^TM system (Thermo Fisher Scientific). The pooled peptides were separated using the UltiMate 3000 RSLCnano system and analyzed by Q-Exactive mass spectrometry (Thermo Fisher Scientific). The differential protein expression levels between control and mutant were determined based on the signal strengths of the corresponding peptides detected for each protein. We had two biological replicates for each control and mutant samples. P-value was calculated for proteins with more than two peptides detected by using a standard software, Scaffold (Proteome Software, Oregon) and shown in S3 Table.

In vitro oocyte activation

Mature but inactive egg chambers were isolated from dissected ovaries in the oocyte isolation buffer at 25°C. Oocyte activation was initiated by exchanging the isolation buffer with a hypotonic buffer (20 mM PIPES, 10 mM NaCl, 50 mM KCl, 5% PEG6000, 2 mM CaCl₂, pH 6.4) [11,18]. After incubation for 30 min, the egg chambers were treated with 50% bleach for 2 min to lyse the inactive egg chambers. The resistant egg chambers were considered activated oocytes.

Western blotting

Protein lysates from ovaries (one-half ovary equivalent per lane) were electrophoresed on a 7.5% or 10% sodium dodecyl sulfate-polyacrylamide gel. The following primary antibodies were used: guinea pig anti-dMarf1 (1:1,000, this study), mouse anti-Myc 9E10 (1:5000, Wako), mouse anti-CycA (1:1000, DSHB, IN, USA), mouse anti-CycB (1:1000, DSHB), rabbit anti-Nanos (1:1000, from Dr. Nakamura), rabbit anti-Dhd (1:1000, from Dr. Loppin), mouse anti-Polo (1:400, from Dr. Glover), mouse anti-glorund (1:1000, DSHB), and mouse anti-αTubulin DM1A (1:1000, Santa Cruz Biotechnology). Immunoreactive bands were visualized using HRP-conjugated goat anti-guinea pig (1:1000, Dako, Denmark), anti-rabbit, or anti-mouse secondary antibodies (1:3000, Bio-Rad, CA, USA), and immunoblots were detected using LAS-1000 (GE Healthcare) with an ECL chemiluminescent substrate (Thermo Fisher Scientific).

Data deposition

The next generation sequencing data shown in this study are deposited to the DDBJ sequence Read Archive (DRA). The submission IDs for RNA-seq and RIP-seq data are SSUB012974 and SSUB012977, respectively.

MARF1 is an RRM and OST motif-containing protein conserved in animals

The genome of Drosophila melanogaster was screened for the mouse homolog of MARF1 using the method of polypeptide similarity search. The protein encoded by the CG17018 gene showed the highest score. The full-length CG17018 polypeptide consists of 1,305 amino acid residues and shares 31% sequence identity with mouse MARF1 in their middle regions that contain the RNA Recognition Motif (RRM) and OST/Lotus domains (Fig 1A). The OST/Lotus domain is highly conserved from bacteria to humans, and proteins harboring this domain are exclusively expressed in germline cells of animals [19,20]. The presence of multiple OST/Lotus domains is a characteristic of the MARF1 family. MARF1 homologs in most animals, except Drosophila, contain a nuclease domain (NYN) (Fig 1A and S1 Fig). Hereafter, CG17018 is referred to as dMarf1, a Drosophila homolog of MARF1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 1. dMarf1 is a conserved protein and functions in the late stages of oogenesis.

(A) Schematic model of mouse MARF1 and Drosophila dMarf1 proteins. NYN, RRM, and OST/Lotus correspond to Nedd4BP1 and YacP nuclease domain, RNA Recognition Motif, and Oskar-TDRD5/7 domain, respectively. The polypeptide sequence similarity in the conserved domain region is shown. The loss-of-function allele, KO321, contains a two-nucleotide deletion, which results in premature termination of translation. (B) Western blotting of ovarian lysates from the control and dMarf1^KO321 homozygous mutant females. Each lane contains equivalents of 0.5 ovaries. (C) Western blotting showing temporal regulation of dMarf1 protein expression at different stages of egg chambers and embryo development. α-Tubulin was used as a loading control. dMarf1 expression was upregulated in stage 12 egg chambers, and immediately reduced after the onset of embryogenesis. (D) Embryos were collected from the heterozygous control and dMarf1^KO321 homozygous mutant mothers were collected for 6-hrs after egg laying and stained with DAPI (blue), anti-α-Tubulin (green), and anti-phospho-Histone H3 (red). Young control embryos (0-2-h old) exhibit the characteristic rosette structure containing three pronuclei, while 0–6 h-old embryos from the mutant mothers do not. Scale bar, 10 μm.

https://doi.org/10.1371/journal.pone.0231114.g001

Secondary structure prediction was performed to examine the structural features of dMarf1. While the RRM and OST domains are highly structured, the N-terminal region of approximately 200 residues and C-terminal region of approximately 300 residues were predicted to be unstructured (S2 Fig). The unstructured C-terminal region is not evolutionary conserved among all animal groups. However, it is well conserved among higher animal groups such as humans, mouse, Xenopus, and Zebrafish (S3 Fig). Moreover, the C-terminal region of human MARF1 is proposed to be a binding site for Ge-1, a key component of the decapping complex [21]. The C-terminal sequence of dMarf1 is different from those of higher animals. However, in contrast to the highly diverse N-terminal region, the C-terminal region consisting of 50 amino acids is well conserved among insects (S4 Fig). This finding suggests that the C-terminal regions is functionally important for dMarf1.

dMarf1 functions during late oogenesis

We generated a loss-of-function allele dMarf1^KO321 using the CRISPR-Cas9 system to study the in vivo function of dMarf1. Genome editing resulted in a two-nucleotide deletion in the N-terminal region of dMarf1 (Fig 1A), causing a frameshift mutation and premature termination of translation. dMarf1^KO321 males did not show any defect; normal egg hatching was observed when the control wild-type females were mated with dMarf1^KO321 males (S5A Fig). By contrast, dMarf1^KO321 females were sterile. Although the mutant females produced a normal number of eggs upon mating with control males, none of them hatched (S5B Fig). We further confirmed this defect by downregulating dMarf1 expression using small RNA-based strategy [22]. The TRiP line flies carrying the HMS04468 transgene to express shRNA complementary to dMarf1 were crossed with two different germline driver lines—NGT40 and nosGal4-VP16—to drive expression from the germarium to the late stages of egg chambers, and Matα-Gal4 in germline cells after stage 2 of egg chambers [23,24]. We also used a somatic driver, traffic jam-Gal4 (tj-Gal4), as a control. The mutant females in which dMarf1 expression was downregulated using either germline driver line laid a similar number of eggs compared to the control females (S5C Fig). However, the eggs from the mutant females did not hatch (S5D Fig). By contrast, the downregulation of dMarf1 with somatic tj-Gal4 did not cause any defects. These results suggest that dMarf1 plays an important role in germline cells after stage 2 of oogenesis.

Next, we generated an antibody against dMarf1 to examine the protein expression of dMarf1. Western blotting revealed the presence of a unique band in the control ovary lysate that was absent in the dMarf1^KO321 homozygous mutant ovary lysate (Fig 1B). This result indicated that the antibody specifically targets dMarf1 in the control ovaries and dMarf1 expression is lost in the mutant ovaries. Consistent with its predicted size, the apparent molecular weight of dMarf1 was approximately 130 kDa. dMarf1 was first detected in stage 10 of oogenesis, and its expression increased until stage 14 (Fig 1C). dMarf1 expression was detected during the initial 2 h after egg laying, and immediately decreased thereafter. These results suggest that dMarf1 functions in the late stages of oogenesis.

Because the eggs laid by dMarf1^KO321 females did not hatch (S5B Fig), we next examined meiotic progression in the absence of dMarf1^KO321 by immunostaining α-tubulin and phospho-histone H3 (pH3), an M-phase marker. The fertilized eggs of the control females exhibited the characteristic rosette structure containing three pronuclei, indicating the completion of meiosis (arrow, Fig 1D). In contrast, the eggs of the dMarf1^KO321 females contained fragmented or over-replicated chromosomes that were often surrounded with microtubules, but not the rosette structure (arrowhead, Fig 1D). These results suggest that meiosis was incomplete in the dMarf1^KO321 oocytes.

dMarf1 is required for meiosis I to II transition

Next, we examined late-stage oogenesis defects in dMarf1^KO321 mutant ovaries. The egg chambers of heterozygous control and dMarf1^KO321 ovaries were categorized based on the King’s criteria [12] and counted at each stage. The distribution of the egg chamber stages was not significantly different between the control and dMarf1^KO321 ovaries (Fig 2A). In addition, dMarf1^KO321 mutant females laid similar number of eggs compared to the control (S5B Fig). These results suggested that oogenesis progresses normally in the absence of dMarf1 until stage 14. This result also suggests that dMarf1 is not required for the release of prophase I arrest of stage 13 egg chamber, unlike mouse MARF1 that is required for the release of prophase I arrest of meiotic division (GV stage)[6].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 2. dMarf1 functions during the late stages of oogenesis.

(A) The ratio of each stage is plotted for the egg chambers from heterozygous control ovaries (upper bar) and dMarf1^KO321 ovaries (lower bar), respectively. Stages of egg chambers were assigned based on the standard King’s criteria [12]. (B) In vitro activation of stage 14 egg chambers. The average ratio ± standard deviation is shown for the control (n = 94) and dMarf1^KO321(n = 158) ovaries (p < 0.05; Student’s t-test). (C-D) Meiotic spindles in stage 14 egg chambers. Most spindles in the control were round, long, or separated. By contrast, 20% of stage 14 egg chambers in the dMarf1^KO321 ovaries contained abnormal (broken or mis-aligned) spindles. The ratio of each spindle shape type is plotted as a bar graph in (C) (n = 58 for control, and n = 91 for dMarf1^KO321), and immunostaining of the corresponding spindle types are shown in (D). DAPI in blue, α-tubulin in green, phospho-histone H3 (an M-phase marker) in red. Scale bar, 5 μm.

https://doi.org/10.1371/journal.pone.0231114.g002

To further examine whether dMarf1 plays a role in the maturation process after the release of the first arrest point, we performed the egg activation assay in vitro [11]. Dehydrated and inactive stage 14 egg chambers were dissected from ovaries in a hypertonic buffer to retain dehydration, and subsequently activated by hydration in a hypotonic buffer [11]. Of all the control eggs, 60% were activated and showed resistance to bleach by crosslinking of vitelline membrane proteins in meiosis II (Fig 2B). By contrast, only 25% of the dMarf1^KO321 mutant eggs were resistant to bleach, indicating the failure of vitelline membrane maturation in the absence of dMarf1. This result prompted us to examine meiotic spindles of the metaphase I-arrested oocytes. Symmetric spindles (round, long, and separated) were detected in 97% of the control oocytes at stage 14, and only 3% showed abnormally shaped spindles. However, 22% of the stage 14 dMarf1^KO321 oocytes exhibited broken spindles or misaligned chromosomes (Fig 2C and 2D). These results suggest that dMarf1 is required to stabilize meiosis I spindle structure.

Furthermore, the dMarf1 mutant eggs exhibited severe defects in meiosis I to II progression (Fig 3). Oocytes in a second arrest at metaphase I can be released from the arrest by ovulation to resume meiosis I to II progression without requiring fertilization. Of all the heterozygous control eggs, 71% showed the characteristic rosette structure with pronuclei and microtubules, indicating the completion of meiosis (Fig 3A, top panel, and B). In another 18%, two to four additional microtubules with the rosette structure were observed. However, no rosette structure was observed in the dMarf1^KO321 mutant eggs (Fig 3A: lower panels). Among those, 36% of the mutant eggs showed a single metaphase I-like spindle, indicative of prolonged arrest state (Fig 3B), 15% showed two spindle-like structures that were not connected to each other as transiently observed during meiosis II [11], and 19% exhibited abnormal and disrupted spindle structures. In the dMarf1^KO321 oocytes, fragmented or over-duplicated chromosomes were often surrounded by microtubules (Fig 3A: bottom panels). These results indicate that meiosis is incomplete and meiosis I to II transition is compromised in the absence of dMarf1.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 3. Meiosis I to II transition is compromised in dMarf1^KO321 mutant females.

(A-B) Meiotic products in the unfertilized egg from the heterozygous control or homozygous dMarf1^KO321 females. The microtubule structures were classified into the following categories: control: rosette structure (a fused structure of pronuclei), rosette with two to four extra spindles, and an uncategorized structure; dMarf1^KO321: one spindle, two spindles, disrupted spindles/microtubules, and an uncategorized structure. The ratio of each spindle type was plotted in a bar graph (B) (n = 181 for control, and n = 140 for dMarf1^KO321), and immunostaining of the corresponding spindle type is shown in (A). DAPI in blue, α-tubulin in green, phospho-histone H3 (an M-phase marker) in red. Scale bar, 10 μm.

https://doi.org/10.1371/journal.pone.0231114.g003

dMarf1 is expressed in nurse cells and deposited into ooplasm

Western blotting revealed that dMarf1 expression increased in the late stages of oogenesis (Fig 1C). Hence, we analyzed the subcellular localization of dMarf1 in the egg chambers. Because the anti-dMarf1 antibody did not work for immunostaining, we used a transgenic fly expressing Myc-tagged full-length dMarf1 (Myc-dMarf1-FL). Myc-dMarf1-FL driven by a Matα-Gal4 germline driver was detected both in oocytes and nurse cells using the anti-Myc antibody (Fig 4). In nurse cells, the nuclear expression of Myc was slightly higher than its cytoplasmic expression, suggesting that dMarf1 is localized into the nucleus. The expression of Myc-dMarf1-FL in a dMarf1^KO321 mutant background significantly increased the hatching rate of eggs (up to 17%); however, none of the dMarf1^KO321 mutant eggs hatched (S6A Fig). Consistently, embryogenesis of the dMarf1^KO321 mutant eggs was recovered, normal mitotic spindles were observed in the rescued eggs (S6B Fig), and 20% of the mutant eggs passed through the blastoderm within 6 h after egg laying. However, the dMarf1^KO321 mutant eggs lacking Myc-dMarf1-FL expression did not show any sign of development (S6C Fig).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 4. Localization of full-length or truncated variants of Myc-dMarf1 in germline cells.

Full-length (FL) Myc-dMarf1 and its truncated variants, lacking 252 N-terminal amino acid residues (ΔN), RRM domain (ΔRRM), OST domain (ΔOST), C-terminal region (ΔC), and 47 C-terminal amino acid residues (ΔC47), were expressed in female germline cells and immunostained with an anti-Myc antibody. DAPI in magenta, α-Myc in green. Scale bar, 50 μm.

https://doi.org/10.1371/journal.pone.0231114.g004

Next, truncated variants of dMarf1 transgenes lacking either the N-terminal region, RRM domain, OST domain, C-terminal region, or 47 C-terminal amino acid residues, were generated and expressed in female germline cells to examine the role of the conserved domains (S7 Fig; Fig 4). Similar to the full-length Myc-dMarf1, all truncated dMarf1 proteins except Myc-dMarf1-ΔN lacking 252 N-terminal residues were localized to the nurse cell nucleus. Myc-dMarf1-ΔN showed faint staining in the nurse cell nuclei, suggesting that the N-terminal residues are required for nuclear localization. Although other truncated proteins predominantly accumulated in nurse cell nucleus and ooplasm, none including Myc-dMarf1-ΔN rescued the hatching rate of eggs (S1 Table). Even Myc-dMarf1-ΔC47 lacking 47 C-terminal residues was unable to rescue the hatching rate, indicating that this small but conserved region is important for dMarf1 function in insects.

Dhd thioredoxin required for pronuclei fusion is downregulated in dMarf1 mutant ovaries

Consistent with the finding that dMarf1 is expressed in the late stages of oogenesis (Fig 1C), meiotic progression beyond metaphase I—the second arrest point—was not observed in the absence of dMarf1 (Fig 3), suggesting that dMarf1 is required for proper oocyte maturation. In an attempt to obtain candidate genes those expressions were affected by the loss of dMarf1, we first conducted the next-generation sequencing (mRNA-seq) of each one RNA library from dMarf1^KO321 mutant and the heterozygous control ovaries. The mRNA expression levels of 10 genes were significantly decreased or increased in dMarf1^KO321 ovaries compared to those in the control (p-value < 0.0001, Fig 5A, S2 Table). We further validated these changes with three biological replicated qRT-PCR, and identified four genes (CR44701, Cyp4e2, pdk, and CG18088) whose expression was downregulated (Fig 5B). pdk encodes pyruvate dehydrogenase kinase, a negative regulator of the pyruvate dehydrogenase complex. Pyruvate metabolism regulates the balance between glucose uptake and glycogen storage in oocytes during maturation [25]. Thus, a decrease in Pdk levels may affect the metabolome of dMarf1 mutant oocytes and perturb embryogenesis.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 5. Global analysis of gene expression in dMarf1^KO321 ovaries and dMarf1-bound mRNA in Myc-dMarf1 expressing ovaries.

(A) MA plot for mRNA-seq analysis. The average RPKM (Reads Per kilobase per Million reads) values for each gene in dMarf1^KO321 and the control ovaries are plotted on the X-axis. The log2-fold change in the expression of each gene in dMarf1^KO321 compared to that in the control ovaries are plotted on the Y-axis. Red dots indicate genes whose expressions are significantly different between dMarf1^KO321 mutant and the control ovaries (p-value < 0.0001). (B) qRT-PCR validation of differentially expressed genes between the control and dMarf1^KO321 ovaries. Values represent the mean and standard deviation of three replicates. The significant differences of student’s t-test are indicated by ** (p-value < 0.01) or *** (p-value < 0.001). (C) MA plot for quantitative proteome analysis of stage 14 egg chambers. The average log2 values in dMarf1^KO321 mutant and the control ovaries for each detected protein are plotted on the X-axis. The log2-fold changes of each detected protein in dMarf1^KO321 mutant compared to that in the control ovaries are plotted on the Y-axis. Red dots indicate proteins that show significantly different expression levels between the mutant and the control ovaries (p-value < 0.05). (D) Western blot showing Dhd expression in stage 14 dMarf1^KO321 egg chambers and control (upper panel). The band intensity represents the mean of three biological replicates. The mean and standard deviation are shown in the lower panel (p < 0.0001; Student’s t-test). α-Tubulin was used as the loading control. (E) MA plot for dMarf1-bound RNA. RNAs bound to Myc-dMarf1 were immunoprecipitated and analyzed by next-generation sequencing. The average RPKM values of each gene whose transcript bound to dMarf1 and the control beads are plotted on the X-axis. The log2-fold change of each gene whose transcript bound to dMarf1 compared to control are plotted on the Y-axis. Red dots indicate genes whose transcripts are significantly enriched with dMarf1 (p-value < 0.005, and log2-fold change > 1.8). (F) qRT-PCR validation of enriched mRNAs in Myc-dMarf1 between the control and dMarf1^KO321 ovaries. Values represent the mean and standard deviation of three replicates. The significant differences of student’s t-test are indicated by * (p-value < 0.05) or ** (p-value < 0.01).

https://doi.org/10.1371/journal.pone.0231114.g005

Other genes whose mRNA expression was altered in dMarf1 mutant ovaries were not related to meiosis. We also examined defects in protein expression of stage 14 dMarf1 mutant egg chambers. Of the 700 proteins that were detected, Dhd expression was significantly reduced to 20% in dMarf1^KO321 compared to that in the control (Fig 5C; S3 Table). Western blotting confirmed that only 10% of Dhd was expressed in the dMarf1 mutant relative to that in the control (Fig 5D). In dMarf1^KO321 ovaries, however, the mRNA expression of Dhd was not affected despite a significant reduction in its protein expression (S8A Fig). This result suggests that Dhd expression is post-transcriptionally regulated. Dhd is a thioredoxin whose expression dramatically increases during oocyte maturation [26]. Unlike the dMarf1^KO321 mutant, dhd mutant females exhibit defects in unpacking paternal chromosomes before the fusion of pronuclei [27], suggesting that the reduction in Dhd expression is not the causative factor for meiotic defect in dMarf1^KO321 oocytes. Our results suggest that dMarf1 regulates not only meiosis but also subsequent steps including pro-nuclear fusion at the onset of embryogenesis.

dMarf1 binds to nos mRNA to regulate its expression

It is known that dMarf1 harbors a conserved RRM domain; hence, we screened for RNA molecules bound by dMarf1 to identify its direct effectors. Myc-dMarf1-FL protein expressed in the germline cells was immunoprecipitated. To obtain the potential genes whose mRNAs bound by dMarf1 protein, the co-precipitated RNAs were analyzed by next-generation sequencing. Six different RNAs were enriched in the immunoprecipitated fraction (Fig 5E and S4 Table; log2 fold change > 1.5 and p-value < 0.01). Of these, the enrichment of nanos (nos), transformer 2 (tra2), and abnormal oocyte (abo) mRNAs in the Myc-dMarf1 bound fraction was further confirmed by qRT-PCR (Fig 5F). tra2 encodes an RNA-binding protein that functions together with transformer (tra) to promote sex-specific splicing of doublesex (dsx) mRNA [28]. abo is a negative regulator of histone gene transcription [29]. Although both tra2 and abo are not directly involved in meiosis or oocyte maturation, dMarf1 is suggested to function by binding with tra2 and abo mRNAs, which will be discussed later.

Another dMarf1-bound mRNA, nos, encodes a translational repressor of cycB, the protein of which complexes with Cdk1 to form maturation promoting factor (MPF) [30]. To examine whether dMarf1 regulates nos expression by binding to its RNA, we analyzed Nos protein expression in the late stages of oogenesis. Consistent with the results of the previous study, Nos was highly expressed in stage 10 egg chambers and its expression rapidly decreased in the later stages [31] (Fig 6A, middle panel). Nos protein levels in stage 10 dMarf1^KO321 egg chambers were comparable to that in stage 10 control egg chambers (Fig 6A). However, Nos expression decreased gradually, but remained detectable even in the stage 14 egg chambers; Nos expression in stage 14 dMarf1^KO321 egg chambers was approximately 7-fold higher than that in the corresponding control egg chambers (Fig 6B). By contrast, the mRNA expression of nos in stage 14 dMarf1^KO321 egg chambers remained unchanged (S8A Fig). These results suggest that dMarf1 can function as a translational repressor of nos mRNA, although the possibility of delayed or aberrant degradation of Nos in dMarf1^KO321 ovaries cannot be ruled out.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Fig 6. dMarf1 translationally represses Nos in the late stages of oogenesis.

(A) Western blot showing protein expression at each stage of oogenesis. Upper panel, dMarf1; middle panel, Nos; bottom panel, Cyclin B, and the loading control, α-tubulin. (B) Quantitative analysis of Nos protein expression in stage 14 egg chambers. The mean and standard deviation of Nos expression are shown in terms of arbitrary unit (p < 0.01; Student’s t-test). (C) Quantitative analysis of CycB protein expression in stage 14 egg chambers. The mean and standard deviation of CycB expression are shown in terms of arbitrary unit (p < 0.001; Student’s t-test). (D) Schematic model of dMarf1 function. In wild type females, dMarf1 coordinates the late stages of oogenesis. dMarf1 binds to nos mRNA and inhibits its protein translation. Consequently, CycB expression is upregulated at the onset of oocyte maturation, which results in the formation of the MPF complex to promote meiosis. In the absence of dMarf1, Nos is expressed until stage14 and CycB is downregulated, which in turn reduces MPF activity.

https://doi.org/10.1371/journal.pone.0231114.g006

CycB protein expression is reduced in dMarf1 mutant ovaries

Because Nos, together with Pumilio, is known to translationally repress CycB in embryos [32], we examined CycB expression in stage 14 dMarf1^KO321 egg chambers where Nos protein expression is remarkably upregulated. In the control ovaries, CycB accumulation was observed in stage 14 egg chambers along with the downregulation of Nos. However, CycB protein expression in stage 14 egg chambers of dMarf1^KO321 was five-fold lower than that in the control (Fig 6A, bottom panel, and 6C). Moreover, cycB mRNA expression was not significantly affected (S8 Fig). Our results suggest that Nos functions as a translational repressor of cycB not only during embryonic development but also oocyte maturation. The downregulation of CycB expression in the late stages of oogenesis in dMarf1^KO321 females reduces MPF activity, and can potentially inhibit meiotic progression.

CycB3 is another cyclin that is required for anaphase progression in meiosis I and II. [4]. CycB3 expression was only slightly decreased in dMarf1^KO321 (S9B Fig), confirming that CycB, rather than CycB3, is a major effector of dMarf1 regulation. MPF activity is also required to induce the expression of PNG kinase complex components (Png/Plu/Gnu) that are important for the activation of stage 14 oocytes [33]. Consistent with the reduction of MPF activity in dMarf1^KO321, only the protein expression but not the mRNA expression of Gnu was also reduced (S9B Fig, S8A Fig). Taken together, our results suggest that dMarf1 functions to coordinate oocyte maturation, at least in part, through regulation of the MPF component, CycB, via Nos-induced translational repression.

dMarf1 may regulate Nos during late oogenesis

Nos is an evolutionary conserved protein that play important roles in early embryogenesis, formation of primordial germ cells and maintenance of germline stem cells (GSCs) [31,34,35]. However, the function of Nos during late oogenesis has not been described yet. Nos protein expression reaches the maximum around stage 10 of oogenesis and immediately reduces thereafter [31] (Fig 6A). This suggests that Nos expression is tightly regulated during late oogenesis. Our results indicate that dMarf1 regulates Nos expression by inhibiting its translation in late oogenesis. In the absence of dMarf1 expression during early-to-mid oogenesis, Nos might repress cycB to prevent the premature release of meiotic arrest until stage 10 of oogenesis. Our results suggest that dMarf1 may coordinate oocyte maturation during late oogenesis in Drosophila (Fig 6D); moreover, dMarf1 is dominantly expressed after stage 10, and binds to nos mRNA to repress its translation and reduce its expression. Consequently, Nos expression almost disappears at stages 12–14. Reduced expression of Nos induces the release of cycB mRNA from repression and promotes CycB translation. Subsequently, CycB binds to Cdk1 to form the active MPF complex to promote meiosis. Therefore, dMarf1 plays an important role in the release of the second meiotic arrest to drive embryogenesis.

The conserved domain and region of MARF1 protein have distinct functions

The MARF1 gene is evolutionarily conserved in animals; most proteins of the MARF1 family contain three major domains: NYN, OST (also known as LOTUS), and RRM (S1 Fig). Although the ribonuclease activity of NYN is essential for MARF1 function in mouse [7], it is not required in Drosophila dMarf1. The OST domain is present in the proteins of several species ranging from bacteria to humans [19,20]. Drosophila melanogaster has four members of OST domain-containing proteins: dMarf1, Oskar (Osk), Tejas (Tej), and Tapas (Tap) [13,36,37]. All of these proteins except dMarf1 contain a single OST domain, and are predominantly expressed in germline cells. Structural and biochemical studies of Osk, Tej, and Tap OST domains have revealed the ability of this domain to bind to Vasa, an RNA helicase expressed exclusively in germline cells [38]. Interestingly, the OST domain(s) of MARF1 family members is smaller than those of other proteins and does not bind to Vasa. Instead, a recent study reported that dMarf1 could bind to CCR4-Not deadenylase complex via the OST domain(s); however, the importance and cooperation between multiple OST domains remains elusive [8]. In addition to these two domains, the MARF1 family members contain one or two RRM domains. In this study, we showed that dMarf1 translationally repressed nos mRNA. nos may not be the only target of dMarf1; other mRNAs such as tra2 and abo can also be targeted by dMarf1, although the biological significance of these interactions remains unclear. Zhu et al. have recently reported that dMarf1 can bind to cyclin A mRNA via its RRM domain [8]. However, our RNA-IP analysis did not detect cyclin A in the dMarf1-bound mRNA fraction (S4 Table). Further studies of the molecular mechanism underlying the specificity of dMarf1 RNA binding will reveal the range of mRNA regulation during oocyte maturation.

The C-terminal region of the MARF1 family members is highly conserved among higher animals, except insects (S3 and S4 Figs), despite not forming any secondary structure. The C-terminal region of human MARF1 has been shown to directly interact with the decapping complex component Ge-1 [21]; however, this interaction was not observed in Drosophila [8]. Moreover, the C-terminal region of the MARF1 family members is conserved among different insect species, but different from that of higher animals (S4 Fig), indicating that it may bind to a unique partner in insects. This hypothesis was supported by our results showing that transgenic dMarf1 mutant lacking 47 C-terminal residues (ΔC47) could not rescue dMarf1^KO321 mutant phenotype (S1 Table).

Both Drosophila and mouse Marf1 are involved in RNA regulation, but their modes of function are different

In addition to binding to the RNA decapping complex subunit [21,39], human MARF1 can localize to processing body (P-body), which is often related to translational repression and mRNA decay [40]. Mouse MARF1 has also been shown to degrade target mRNA via its NYN domain, which is absent in Drosophila dMarf1 [7]. Transcriptome analysis of mouse MARF1 mutant oocytes revealed that 1,470 transcripts were upregulated in the steady state, whereas 103 transcripts were downregulated [7], indicating a global impact on RNA homeostasis in the mutant oocytes. By contrast, the expression of a few RNAs was downregulated and dp1 mRNA expression was upregulated in Drosophila dMarf1 mutant ovaries (Fig 5A, S2 Table). These results suggest that mammalian MARF1 may regulate the global transcriptome predominantly by degradation, while dMarf1 represses translation of target proteins such as CycB and CycA by modulating Nos/Pumilio and the CCR4-NOT deadenylase complex, respectively [37].

In addition to nos mRNA, dMarf1 can bind to other mRNAs, including tra2 and abo. The mRNA expression of tra2 was not significantly affected in stage 14 dMarf1^KO312 egg chambers (S8A Fig), suggesting that dMarf1 post-transcriptionally regulates tra2 expression. abo is a negative regulator of histone gene expression and its expression is downregulated in mature oocytes to produce more histones [41]. The expression of abo mRNA was upregulated by approximately three-fold in stage 14 dMarf1^KO321 egg chambers (S8A Fig). This may result in the overexpression of Abo protein in dMarf1 mutant oocytes, which in turn causes the downregulation of histone proteins that are required for embryogenesis.

The Ppp2cb gene encodes a protein phosphatase that is involved in cell cycle regulation. Ppp2cb has been previously reported as a major downstream effector of mouse MARF1 [42]. The high expression of Ppp2CB phosphatase in the MARF1 mutant ovaries of mouse can disrupt meiosis. However, the expression of mts, a Drosophila homolog of Ppp2cb, was not affected in dMarf1 mutant ovaries (S8A Fig), suggesting that the signaling pathway for the activation of the M phase promoting factor, CycB/Cdk1, is not conserved between mouse and Drosophila. Although the direct activation of MPF in mouse MARF1 mutant oocytes by the inhibition of Ppp2cb rescued meiotic defect, embryogenesis of mutant oocytes was affected [42], suggesting that Ppp2cb may have additional functions in addition to MPF activation. Similarly, dMarf1^KO321 ovaries exhibited not only meiotic defects, but also translationally downregulated some proteins required for embryogenesis, such as Dhd and Gnu (Fig 5D, S8A Fig, and S9B Fig). In conclusion, MARF1 may trigger oocyte maturation and coordinate multiple events during late oogenesis and fertilization.