Maternal Piwi regulates primordial germ cell development to ensure the fertility of female progeny in Drosophila

Abstract

In many animals, germline development is initiated by proteins and RNAs that are expressed maternally. PIWI proteins and their associated small noncoding PIWI-interacting RNAs (piRNAs), which guide PIWI to target RNAs by base-pairing, are among the maternal components deposited into the germline of the Drosophila early embryo. Piwi has been extensively studied in the adult ovary and testis, where it is required for transposon suppression, germline stem cell self-renewal, and fertility. Consequently, loss of Piwi in the adult ovary using piwi-null alleles or knockdown from early oogenesis results in complete sterility, limiting investigation into possible embryonic functions of maternal Piwi. In this study, we show that the maternal Piwi protein persists in the embryonic germline through gonad coalescence, suggesting that maternal Piwi can regulate germline development beyond early embryogenesis. Using a maternal knockdown strategy, we find that maternal Piwi is required for the fertility and normal gonad morphology of female, but not male, progeny. Following maternal piwi knockdown, transposons were mildly derepressed in the early embryo but were fully repressed in the ovaries of adult progeny. Furthermore, the maternal piRNA pool was diminished, reducing the capacity of the PIWI/piRNA complex to target zygotic genes during embryogenesis. Examination of embryonic germ cell proliferation and ovarian gene expression showed that the germline of female progeny was partially masculinized by maternal piwi knockdown. Our study reveals a novel role for maternal Piwi in the germline development of female progeny and suggests that the PIWI/piRNA pathway is involved in germline sex determination in Drosophila.

Keywords: Drosophila; Piwi; germline development; maternal effect; oogenesis.

Figure 1

Maternal Piwi persists in germ cells through gonad coalescence. (A) Schematic of crossing strategy to visualize maternal Piwi. (B) Representative images from visualization of Myc-Piwi in late-stage embryos (approximately 16 hours after egg laying) inheriting only maternal Myc-Piwi protein (left panel; n = 56/59 with Myc in germ cells), inheriting both maternal Myc-Piwi protein and the myc-piwi transgene (middle panel; n = 52/56 with Myc in germ cells), and inheriting neither the Myc-Piwi protein nor the myc-piwi transgene (right panel; negative control; n = 0/28 with Myc in germ cells). Coalesced gonads indicated by arrowheads. (C) Representative images of maternal Myc-Piwi protein in UASp-GFP/zfh2-gal4 embryos from the cross depicted in (A), and schematics of germ cell localization at each developmental timepoint. Inset is magnified 2X. PGC, primordial germ cells.

Figure 2

Maternal Piwi is required for fertility of female, but not male, progeny. (A) Schematic of crossing strategy to deplete maternal Piwi. F0 individuals carry both UASp-shRNA and Maternal Alpha Tubulin (MAT)-GAL4, so shRNA is expressed from mid-oogenesis. F1 individuals carry either UASp-shRNA or MAT-GAL4, but not both. (B) Schematic of MAT-GAL4 expression in the MatKD F0 ovary and anti-Piwi immunofluorescence in MatKD F0 ovaries from 2–3-day old flies. The dotted and solid lines in the schematic indicate when MAT-gal4 is expressed and when knockdown effect is observed, respectively. Whole ovary images are at the same scale. (C) Western blot for total Piwi and total GAPDH in 0–2-hour F1 embryos. (D–G) Seven-day fertility tests of individual females mated to two w¹¹¹⁸ males or individual males mated to three w¹¹¹⁸ females (see cross in A). (D) MatKD F0 females (n = 17–25). (E) MatKD F1 females (n = 16–22). (F) MatKD F1 males (n = 17–20). (G) MatKD F2 females (n = 20–24). Upper panels indicate number of eggs laid per day and lower panels indicate the percentage of those eggs that developed to adulthood (only calculated for crosses which produced ≥ 10 eggs). Mean ± SD. One-way ANOVA and Tukey’s multiple comparisons test. n.s., “not significant.”

Figure 3

Adult females depleted of maternal Piwi had arrested ovaries. (A) Representative images of MatKD F1 ovaries from 2 to 3 day old flies from the cross in Figure 2A, and relative frequency of each category of ovary phenotype in each genotype (n = 45–75 per genotype). Chi-square test. Whole ovary images are at the same scale. (B) Anti-Vasa and anti-Piwi immunofluorescence of MatKD arrested ovaries. Piwi-MatKD arrested ovaries arrest around stage 8. An asterisk indicates one example mature egg, and an arrowhead indicates one example arrested egg chamber. (C) Representative images of MatKD F1 testes from 1–2-day old flies, with anti-Vasa and anti-Piwi immunofluorescence, and Differentially Interference Contrast (DIC) to visualize sperm tails.

Figure 4

Transposons were marginally derepressed in the piwi-MatKD F1 generation. (A,B) Differential expression of transposons from RNA-seq on 0–1.5 hours MatKD F1 embryos from the cross in Figure 2A. Red datapoints are significantly changed (P-adjusted < 0.05, fold change > 1.5) in that genotype compared to GFP-MatKD, and large red datapoints are significantly changed (P-adjusted < 0.05, fold change > 1.5) in both piwi-MatKD #1 and #2 compared to GFP-MatKD. (C) Relative proportion of transposon classes in each group. (D) RT-qPCR for transposons in MatKD F1 adult ovaries from 2 to 3 day old flies, zuc^+/- (zucchini^+/HM27), and zuc^-/- [zucchini^{HM27/Df(2L)716}] ovaries. All piwi-MatKD groups were compared to GFP-MatKD, and zuc^-/- was compared to zuc^+/-. The latter comparison serves as a positive control for transposon derepression in the context of piRNA pathway disruption. Two-way ANOVA and Dunnett’s multiple comparisons test, *P < 0.05, **P < 0.01, ***P < 0.0001.

Figure 5

Knockdown of maternal piwi shifts the maternally deposited transcriptome and piRNA pool in the early F1 embryo. (A) Differential expression of nontransposon genes from RNA-seq on 0–1.5-hour MatKD F1 embryos from the cross in Figure 2A. Red datapoints are significantly changed (P-adjusted < 0.05, fold change > 1.5) in that genotype compared to GFP-MatKD, and large red datapoints are significantly changed (P-adjusted < 0.05, fold change > 1.5) in both piwi-MatKD #1 and #2 compared to GFP-MatKD. (B) Small RNAs (20–29 nt) were isolated from total 0 to 1.5 hours MatKD F1 embryo RNA for Small RNA-seq. (C) Size distribution of piRNAs from total Small RNA-seq, after filtering out rRNA, miRNA, and siRNA, normalized to total library size. Two-Way ANOVA, Dunnett’s multiple comparisons test. *P < 0.01, ***P < 0.0001. (D) Sequence distribution from each piRNA library at the first and tenth nucleotide position. (E) Percentage of reads in each piRNA library which have the capacity to target mRNAs, defined as being antisense to a transcribed gene region, allowing up to two mismatches. (F) Differential targetability of Drosophila nontransposon mRNAs in piwi-MatKD vs GFP-MatKD F1 embryos. Red datapoints are significantly changed (P-adjusted < 0.05, fold change > 1.5) in that genotype compared to GFP-MatKD, and large red datapoints are significantly changed (P-adjusted < 0.05, fold change > 1.5) in both piwi-MatKD #1 and #2 compared to GFP-MatKD. (G) Putative antisense piRNA levels (upper panel, from Small RNA-seq) and mRNA levels (lower panel, from RNA-seq) for Jheh2 and Jheh3. Relative levels determined from DESeq2. Significance tested with two-way ANOVA and Dunnett’s multiple comparisons test.

Figure 6

The piwi-MatKD F1 female germline is partially masculinized. (A–C) Germ cell numbers in male and female embryonic MatKD F1 gonads (stages 15–17) from the cross in Figure 2A were visualized (upper panel) and counted (lower panels). Embryos were sexed by paternally inherited Dfd-lacZ on the X chromosome and embryos were staged based on gut morphology. Violin plots indicate germ cell count at each stage and condition; thick dashed line indicates median and thin dotted line indicates 1st and 3rd quartiles. Line indicates regression line of germ cell numbers of each sex within each genotype, β1 indicates the coefficient calculated from Poisson linear regression, and R² indicates fit of the data to regression line. Asterisk indicates significance of the coefficient (**P < 0.01, ***P < 0.001) compared to the null hypothesis (β1 = 0). n = 66–81 per sex of each genotype, with 20–27 per stage of each sex of each genotype. (A) GFP-MatKD F1 embryos. (B) piwi-MatKD #1 F1 embryos. (C) piwi-MatKD #2 F1 embryos. (D) RT-qPCR for key testis-specific mRNAs in MatKD F1 ovaries from 2-3-day old flies, normalized to actin5C RNA levels (left panel). Upper right panel indicates expression level of total phf7 (phf7^TOT) and lower right panel indicates phf7^RC relative to phf7^TOT, after each was normalized to actin5C. Two-way ANOVA and Dunnett’s multiple comparisons test, *P < 0.05, **P < 0.01, ***P < 0.001.