Zygotic amplification of secondary piRNAs
during silkworm embryogenesis
SHINPEI KAWAOKA,1YUJI ARAI,1KOJI KADOTA,2YUTAKA SUZUKI,3KAHORI HARA,1SUMIO SUGANO,3
KENTARO SHIMIZU,2YUKIHIDE TOMARI,4,5TORU SHIMADA,1,2and SUSUMU KATSUMA1,6
1Department of Agricultural and Environmental Biology, Graduate School of Agricultural and Life Sciences, The University of Tokyo, Yayoi 1-1-1,
Bunkyo-ku, Tokyo 113-8657, Japan
2Agricultural Bioinformatics Research Unit, Graduate School of Agricultural and Life Sciences, The University of Tokyo, Yayoi 1-1-1, Bunkyo-ku,
Tokyo 113-8657, Japan
3Department of Medical Genome Sciences, Graduate School of Frontier Sciences, The University of Tokyo, 4-6-1 Shirokanedai, Minato-ku, Tokyo
4Institute of Molecular and Cellular Biosciences, The University of Tokyo, Bunkyo-ku, Tokyo 113-0032, Japan
5Department of Medical Genome Sciences, Graduate School of Frontier Sciences, The University of Tokyo, Bunkyo-ku, Tokyo 113-0032, Japan
PIWI-interacting RNAs (piRNAs) are 23–30-nucleotide-long small RNAs that act as sequence-specific silencers of transposable elements
in animal gonads. In flies, genetics and deep sequencing data have led to a hypothesis for piRNA biogenesis called the ping-pong cycle,
where antisense primary piRNAs initiate an amplification loop to generate sense secondary piRNAs. However, to date, the
process of the ping-pong cycle has never been monitored at work. Here, by large-scale profiling of piRNAs from silkworm ovary
and embryos of different developmental stages, we demonstrate that maternally inherited antisense-biased piRNAs trigger acute
amplification of secondary sense piRNA production in zygotes, at a time coinciding with zygotic transcription of sense transposon
mRNAs. These results provide on-site evidence for the ping-pong cycle.
Keywords: silkworm; piRNA; transposon; ping-pong cycle
Controlling transposon activity is essential for host ge-
nomes especially in germ line cells, as an unfavorable
mutation caused by transposition can be transmitted to
the next generation (Malone and Hannon 2009). To pre-
vent this, organisms have evolved an elegant defense system
against transposons (Malone and Hannon 2009). In animal
germ lines, PIWI proteins and associated PIWI-interacting
RNAs (piRNAs) are at the center of this defense system.
piRNAs are 23–30-nucleotide-long small RNAs that act as
guides for PIWI proteins which exhibit slicer activity (Saito
et al. 2006; Gunawardane et al. 2007; Nishida et al. 2007;
Thomson and Lin 2009). Mutations in the piRNA pathway
genes cause transposon derepression and defects in germ
line development (Klattenhoff and Theurkauf 2008).
In contrast tosmall interfering RNAs (siRNAs) and micro-
RNAs (miRNAs), piRNA production does not require Dicer
proteins, suggesting that yet-unidentified piRNA precursors
on slicer activity of PIWI proteins (Brennecke et al. 2007;
proteins, Piwi, Aubergine (Aub), and Argonaute3 (Ago3).
their 59 ends (1U), while Ago3-bound sense piRNAs are
enriched for adenosine at position 10 (10A). Importantly,
Aub-bound antisense and Ago3-bound sense piRNAs often
overlap precisely by 10 nucleotides from their 59 ends.
Moreover, all fly PIWI proteins showed slicer activity (Saito
et al. 2006; Gunawardane et al. 2007; Nishida et al. 2007). As
PIWIs cleave their targets between position 10 and 11 (Saito
et al. 2006; Gunawardane et al. 2007), the 10-nt overlaps
between sense and antisense piRNAs could be explained by
PIWI protein-mediated cleavage. Based on these data, the
current model called the ping-pong cycle describesthat Aub-
bound antisense ‘‘primary’’ piRNAs define 59 ends of Ago3-
of Aub-bound antisense ‘‘secondary’’ antisense piRNAs
Article published online ahead of print. Article and publication date are
RNA (2011), 17:1401–1407. Published by Cold Spring Harbor Laboratory Press. Copyright ? 2011 RNA Society.
this model, loss of ago3 dramatically reduced not only sense
piRNAs but also Aub-bound antisense piRNAs (Li et al.
2009a). The 10-nt overlaps, often called ping-pong signa-
tures, are conserved across phyla, including mice, zebrafish,
and silkworm (Aravin et al. 2008; Houwing et al. 2008;
Kawaoka et al. 2009). However, to date, biochemical evi-
dence for the ping-pong model has been lacking, and the
Silkworm is an emerging model to study piRNAs.
Silkworm ovary and ovary-derived cell line BmN4 express
two PIWI proteins and piRNAs with significant ping-pong
signatures (Kawaoka et al. 2008a,b, 2009). In addition,
silkworm embryos provide an ideal system to monitor
developmental piRNA profiles. After mating, a female moth
continuously lays approximately three hundred eggs within
3–4 hours. Fertilization occurs just before laying. This
allowed us to collect highly developmentally synchronized
egg samples. Here, taking advantage of these points, we
analyzed piRNAs prepared from pharate ovary and de-
veloping embryos. We, for the first time, visualized devel-
opmental, ping-pong-dependent piRNA amplification that
was coupled with zygotic transcription of sense transposon
transcripts. Our study provides on-site evidence for the
Maternal deposition of PIWI/piRNA complexes
in silkworm embryos
After fertilization, an animal embryo at the one-cell stage is
transcriptionally inactive, and early embryogenesis requires
maternal factors (Schier2007). At a distinct point of embryo-
genesis, the embryo initiates zygotic transcription (Schier
2007). This transition is called maternal to zygote transition
(MZT). Transcriptional states of embryos can be monitored
by phosphorylation of RNA polymerase II (Vastenhouw
et al. 2010). To characterize the silkworm embryogenesis,
we performed Western blots with anti-phosphorylated RNA
polymerase II (pol II) antibody (Vastenhouw et al. 2010).
Phosphorylated pol II was detected from 3–4 h post-
fertilization (hpf), suggesting that MZT occurs at this stage
in the silkworm (Fig. 1A).
Next, we asked whether PIWI/piRNA complexes are
maternally inherited in silkworm embryos. We revealed
by Western blotting that Siwi and BmAgo3, two silkworm
PIWI proteins (Kawaoka et al. 2008b) were detected be-
fore zygotic activation (Fig. 1B). Moreover, piRNAs were
abundantly detected in 0–1-hpf embryos (Fig. 1C). These
results proved maternal deposition of PIWI/piRNA com-
plexes in the silkworm. Expressions of PIWI/piRNA com-
plexes lasted as long as we surveyed during embryogenesis
Features of maternally inherited piRNAs
Fly adult ovaries contain gonadal germ line cells and
somatic support cells (Malone et al. 2009). Somatic follicle
epithelium is shed from the laid egg. By comparing ovarian
and early embryonic piRNA libraries, Malone et al. iden-
tified two distinct pools of piRNAs—soma-dominant
piRNAs, and germ line-enriched, maternally inherited
piRNAs (Malone et al. 2009). In the silkworm, ovarian
development occurs exclusively during pupation and pha-
rate adult development before hatching (Yamauchi and
Yoshitake 1984). The structure of pharate adult ovary is
nearly identical to that of flies (Yamauchi and Yoshitake
1984). In contrast, the nature of the silkworm adult ovary is
different from the fly adult ovary, as the silkworm adult ovary
contains only matured eggs, which are mainly composed of
germ line cells.
To compare the characters of both germ line (maternal)
and somatic piRNAs with those in flies, we sequenced and
analyzed piRNAs from pharate ovary and early embryos.
Strong maternal deposition indicates that a piRNA is
enriched in germ line, whereas soma-dominant piRNAs
show weak maternal deposition.
FIGURE 1. Maternal to zygote transition and maternal deposition of
PIWI/piRNA complexes in the silkworm embryos. (A) Proteins from
eggs of 0, 1, 2, 3, 4, and 5 hpf were analyzed with anti-phosphorylated
RNA polymerase II antibody. Vitellin, a major egg protein served as
a loading control. (B) Proteins from eggs of 0, 6, 12, 24, 48, 72, and 96
h post-fertilization (hpf) were analyzed by Western blotting with anti-
Siwi, anti-BmAgo3, and anti-actin antibodies. Actin served as a load-
ing control. (C) 10 mg of total RNAs from 0-, 6-, 12-, 24-, 48-, 72-,
and 96-hpf eggs were analyzed by urea-containing acrylamide gel.
Signals were visualized by SYBRGold staining.
Kawaoka et al.
RNA, Vol. 17, No. 7
We used 60 elements among 121 annotated transposons
that are most heavily targeted by ovarian piRNAs (>500
reads per million [RPM]) (Fig. 2A). We calculated the ma-
ternal deposition and found that 54 of 60 transposon-derived
piRNAs were effectively maternally deposited, indicating
that these piRNAs were enriched in germ line. On the other
hand, six elements exhibited weak maternal deposition,
suggesting that these are expressed predominantly in gonadal
The fly somatic piRNA pool lacks ping-pong signatures,
suggesting that the ping-pong cycle may operate mainly in
gonadal germ line cells (Lau et al. 2009; Li et al. 2009a;
Malone et al. 2009; Saito et al. 2009; Haase et al. 2010;
Olivieri et al. 2010; Saito et al. 2010). To investigate
relationships between maternal deposition and the ping-
pong cycle in the silkworm, we calculated the statistical
significance of a ping-pong signature for individual trans-
posons (see Materials and Methods) (Fig. 2A; Supplemental
Table S1). Within the selected 60 transposons, only R2Bm
lacks a statistically significant ping-pong signature, but
R2Bm-derived piRNAs were efficiently inherited to early
embryos via germ line cells. On the other hand, although
Judo-derived piRNAs harbored significant ping-pong signa-
ture in the ovary, they were soma-dominant. In fact, we
observed no correlation between % ping-pong (Brennecke
et al. 2008; Malone et al. 2009) (see Materials and Methods)
and maternal deposition among 60 transposons (Fig. 2B).
We also confirmed soma-dominant expressions of Judo and
BmRT7 mRNAs by qRT-PCR (Supplemental Fig. S1).
Collectively, as long as we surveyed based on two in-
dependent methods, the ping-pong cycle may operate both
in gonadal soma and germ line cells in the silkworm.
Developmental profiling of the silkworm
In the developmental context, how the ping-pong cycle
initiates is currently unknown. The maternal piRNA pool
can be defined as an input for ping-pong amplification. To
obtain evidence for the initiation of zygotic piRNA bio-
genesis by maternal piRNAs, we analyzed piRNAs from
developing embryos (0, 6, 12, and 24 hpf). We mapped
cloned reads to 121 transposons and investigated strand
bias (Fig. 3). As shown in Figure 3, early embryonic piRNAs
were much more antisense-biased than ovarian piRNAs (six-
fold vs. 3.5-fold), demonstrating that antisense piRNAs are
more enriched in germ lines than in gonadal soma. Strong
antisense bias slightly and gradually declined according
to embryonic development (Fig. 3). Decrease of antisense
piRNAs could account for strand bias changes. Alternatively,
accumulation of sense piRNAs may explain why antisense
strand bias reduced.
To test these possibilities, we constructed a heat map
showing relative abundance of sense and antisense piRNAs
on an element (piRNA expression in 0-hpf embryos = 1)
(Fig. 4). Although MZT occurs at 3–4 hpf in the silkworm
(Fig. 1A), overall piRNA profiles between 0 and 6 hpf were
remarkably similar, indicating that MZT does not affect the
piRNA biogenesis. Throughout embryonic development,
expression levels of antisense piRNAs were comparable to
0-hpf embryos (Fig. 4A). In contrast, after 12 hpf, we found
a significant increase of specific sets of sense piRNAs (Figs.
4B and 5).
Amplification of secondary piRNAs during
The most prominent piRNA change we observed during
embryogenesis was the accumulation of a set of sense
piRNAs at 12–24 hpf (Figs. 4B and 5A). We plotted
expression levels of piRNAs derived from five transposons
whose sense piRNA accumulations were most drastic (Fig.
5A). At 12 hpf, Pao-derived sense piRNAs rapidly increased
(60-fold). The abundance of Pao-derived sense piRNAs
peaked at 24 hpf by 120-fold. The same pattern was obtained
for Yamato, Bmhopper, Benibana, and Bm1modoki.
For detailed analysis, we visualized developmental
piRNA expression as a density plot (Fig. 5B). Pao-derived
FIGURE 2. Maternal piRNAs and ping-pong signatures in silkworm
embryos. (A) Log2 fold ratio between normalized ovarian and 0 hpf-
derived piRNA reads against 60 well annotated transposons. An
element lacking statistically significant ping-pong signature is repre-
sented by white bar. (B) Relationships between maternal deposition
and % ping-pong in pharate ovary are displayed as a scatter plot.
Maternal deposition is defined as normalized early (0 hpf) embryonic
reads/normalized ovarian piRNA reads. % ping-pong is expressed as
the likelihood for the average piRNA mapping to an individual
transposon with ping-pong partners.
On-site evidence for the ping-pong cycle
piRNAs showed a marked antisense bias at 0 hpf. The
piRNA population did not change after MZT (6 hpf). At
12 hpf, we observed an increase in sense piRNAs, followed
by a slight antisense piRNA increase. Strikingly, we noted
that almost all amplification occurred at the ping-pong
site (represented as enlarged pictures in Fig. 5B). Indeed,
>80% of increasing sense piRNAs had their partner.
Consistent with this, we observed an overall increase on
10A bias of sense piRNAs (Supplemental Fig. S2). These
data indicated that slicer activity of PIWI proteins mediate
selective amplifications of sense secondary piRNAs.
We hypothesized that abundance of sense piRNAs de-
pends on zygotic expression of sense transposon mRNA,
a theoretical target of maternal antisense piRNAs. To test
this idea, we investigated the expression profiles of sense
transposon mRNA by using strand- specific RT-PCR (Fig.
5C). In this experiment, we reverse-transcribed total
RNAs with a specific primer for each transposon, rather
than the oligo(dT) primer or random primer; therefore, RT
efficiency is different among transposons. Thus, even though
we used the same amount of total RNAs for input, we
cannot directly compare expression levels among transpo-
sons. This is why we did not see a correlation between the
amount of sense piRNAs produced from a specific trans-
poson (Fig. 5A) and the expression level of corresponding
transposon (Fig. 5C). Sense mRNAs of Pao, Yamato, and
Benibana were not maternally deposited (Fig. 5C). These
mRNAs were immediately transcribed at 12 hpf. Sense Pao
mRNA then rapidly became undetected. Importantly, the
timing of their zygotic expression was coupled with the
timing of sense piRNA accumulation (Fig. 5C).
In contrast to Pao, Yamato, and Benibana, sense mRNAs
of Bmhopper and Bm1modoki were detected as early as
0-hpf embryos and slightly decreased at 6 hpf, suggesting
that these transposon mRNAs were maternally deposited
and being slowly degraded. Although sense mRNAs of
these two elements were present during this period,
corresponding sense piRNAs remained unamplified in
zygotes. At 12 hpf, Bmhopper and Bm1modoki increased
again, most likely reflecting their zygotic transcription. In
response to this, sense piRNAs started to accumulate (Fig.
5A). Sense transcription of Kaede, whose piRNA profile
did not change during embryogenesis (Fig. 4), was not
detected (Fig. 5C). Taken together, we envision that
antisense maternal piRNAs kick-start ping-pong amplifi-
cation of secondary sense piRNAs and that it is coupled
with zygotic transcription of sense transposon mRNAs
(Fig. 5D). Our data do not exclude the possibility that
sense transposon mRNAs could be processed into zygotic
primary sense piRNAs via a ping-pong-independent pri-
mary processing pathway, partially contributing zygotic
sense piRNA accumulation.
Here, by using silkworm developing embryos as a model
system, we monitored the developmental process of piRNA
profiles. We found that many ovarian piRNAs are actually
FIGURE 4. Developmental profiling of embryonic piRNAs. Relative
expressions of antisense (A) and sense (B) piRNAs derived from 60
elements (0-hpf embryo = 1). Blue indicates relative expression (RE)
meets 0.5 # RE # 2.0. Light blue corresponds to RE < 0.5, and red
indicates RE > 2.
FIGURE 3. Strand bias of ovarian and embryonic piRNAs. Strand
bias of 121 transposon-derived piRNAs from pharate ovary, 0-, 6-, 12-,
and 24-hpf embryos.
Kawaoka et al.
RNA, Vol. 17, No. 7
maternally inherited via germ line cells in early embryos
(Figs. 1 and 2). In flies, piRNAs deriving from the flamenco
locus are not effectively maternally deposited (Malone et al.
2009). Importantly, flamenco-derived piRNAs are depleted
of ping-pong signatures (Li et al. 2009a; Malone et al.
2009). Thus there is a clear correlation between maternal
deposition and ping-pong significance in flies. In contrast,
in the silkworm, we observed no correlation between ma-
ternal deposition and ping-pong significance (Fig. 2; Sup-
plemental Table S1). For instance, Judo-derived piRNAs are
FIGURE 5. Zygotic amplification of sense piRNAs by maternal antisense piRNAs. (A) Relative expressions of a set of piRNAs whose sense piRNA
increase was most drastic. (B) Density plot of 0-, 6-, 12-, and 24-hpf piRNAs derived from Pao. Representative ping-pong site was enlarged. Plots
were generated by base calling method. (C) Sense strand specific RT-PCR for Pao, Yamato, Bm1modoki, Benibana, and Kaede. (D) Model
depicting zygotic piRNA amplification by maternal piRNAs in the silkworm.
On-site evidence for the ping-pong cycle
predominant in soma but retain the ping-pong signature,
indicating that the ping-pong cycle may not be specific for
germ line in the silkworm. Although we could not exclude
the possibility that still unannotated transposable elements
show features of flamenco-derived piRNAs, we suggest that
the silkworm genome does not harbor a soma-specific
piRNA (transposon) cluster such as flamenco.
As shown in Figures 3 and 4, maternally deposited
piRNAs are mainly antisense, and their population is not
affected by MZT and subsequent embryonic development
(Figs. 3 and 4). The majority of maternally deposited
out embryogenesis (Fig. 4). We have previously found a
rapid decrease of PIWI mRNAs at 20 hpf (Kawaoka et al.
2008b), whereas protein expressions did not alter until
96 hpf during embryogensis. Collectively, we suggest
that PIWI/piRNA complexes are highly stable during
In contrast to antisense piRNAs, most importantly, our
analyses revealed an acute amplification of a specific set of
secondary piRNAs, which is coupled with zygotic tran-
scription of sense transposon mRNAs. The efficiency of
sense piRNA accumulation is not simply dependent on the
availability of sense transposon mRNAs; for particular
transposons, although we detected maternally transmitted
sense mRNAs at 0 to 6 hpf, corresponding sense piRNAs
were not amplified until their zygotic transcription occurred
at 12 hpf. Therefore, zygotes may detect de novo transcrip-
tion of transposon mRNAs and trigger piRNA-mediated
combat against them. Alternatively, a factor(s) required
for sense piRNA accumulation (i.e., the ping-pong ampli-
fication) in zygotes may be simply missing in 0- to 6-hpf
Accumulating sense piRNAs overlapped by 10 nucleotides
with maternal antisense primary piRNAs, suggesting that
PIWI/antisense piRNA complexes cleaved sense transposon
mRNAs to generate sense piRNAs (Fig. 5). Taken altogether,
our current data provide direct evidence for ping-pong
cycle-dependent ‘‘amplification’’ of piRNAs.
MATERIALS AND METHODS
Wild type (WT) strain p50T was reared on fresh mulberry leaves
in an insect rearing chamber under short-day conditions (12L:
Lysate preparation and Western blotting were performed as de-
scribed previously (Kawaoka et al. 2009). Antibodies used were anti-
Siwi (1:10,000) and anti-BmAgo3 (1:5000) (13), anti-actin (1:1000;
Santa Cruz), and anti-phosphorylated polymerase II (1:250; Santa
piRNA library construction
Total RNA was prepared using Trizol reagent (Invitrogen)
according to the manufacturer’s protocol. The total RNA (10
mg) was loaded onto a 15% denaturing polyacrylamide gel
containing 8 M urea, electrophoresed, and then stained with
SYBRGold (Invitrogen). Signals were visualized using LAS-1000
film (Fujifilm). As silkworm piRNAs are visible as a distinct band
by SYBRGold staining (see Fig.1B), we could easily gel-purify the
piRNA-containing fraction. Small RNA libraries were con-
structed using a small RNA cloning kit (Takara). DNA sequenc-
ing was performed using the Solexa genetic analysis system
(Illumina) (Kawaoka et al. 2009). One nanogram of the prepared
cDNA was used for the sequencing reactions with the Illumina
GA. 10,000–15,000 clusters were generated per ‘‘tile,’’ and 36
cycles of the sequencing reactions were performed. The protocols
of the cluster generation and sequence reactions were according
to the manufacturer’s instructions.
Solexa sequencing generated reads of up to 36 nucleotides in
length. The 39 adaptor sequences were identified and removed,
allowing for up to two mismatches. Reads without adaptor
sequences were discarded. Reads shorter than 23 nucleotides or
longer than 30 nucleotides were excluded, resulting in reads of
23–30 nucleotides. Alignment to the B. mori genome, 121
annotated transposons, and 1668 ReAS clones were performed
with SOAP2 (ver. 2.20) allowing no mismatch (Li et al. 2009b).
In this manuscript, we refer to piRNA as 23–30-nt RNAs
matching transposable elements. The total number of perfect
genome-mapping reads reflects the sequencing depth. To com-
pare the reads among different data sets, reads were expressed in
reads per million (RPM) by normalizing to the total number of
perfect genome-mapping in each library.
Ping-pong pairs were defined as precise 10-nt overlaps between
sense and antisense piRNAs matching each of the 121 trans-
posons. The observed abundance of ping-pong pairs (O) was
defined as the sum of all the piRNA reads that joined in ping-
pong pairs. For estimating the expectation (E), we counted the
number of piRNA reads that form ping-pong pairs when we
randomly map the n number of antisense piRNAs and the m
number of sense piRNAs to each transposon. By computing this
process 100,000 times for each transposon, we calculated the
score X = (O – Ei) (i = 1, 2, ..., 100,000). The P-value represents
k/100,000, where k shows how many times X meets (O – Ei) < 0.
The R-code for this analysis will be provided upon request. In
addition, we evaluated ping-pong significance by calculating the
likelihood for average piRNA mapping to an element to have
a complementary ping-pong partner (% ping-pong) (Brennecke
et al. 2008; Malone et al. 2009).
Strand specific RT-PCR
Total RNAs extracted from embryos were reverse transcribed using
avian myeloblastosis virus (AMV) reverse transcriptase (TaKaRa)
with strand-specific primers listed in Supplemental Table S2, and
PCR was performed using KOD-plus polymerase (TOYOBO).
Kawaoka et al.
RNA, Vol. 17, No. 7
piRNAs sequenced in this study are deposited in DRA000173 and
DRA000317 [DNA Data Bank of Japan (DDBJ)].
Supplemental material is available for this article.
We thank P.B. Kwak for critical comments on the manuscript,
and M. Kawamoto and E. Sekimori for their technical assistance.
Sh.K. is a recipient of fellowships from the Japan Society for the
Promotion of Science. This work was supported in part by Special
Coordination Funds for Promoting Science and Technology from
the Ministry of Education, Culture, Sports, Science, and Tech-
nology of the Japanese Government (MEXT) (No. 22115502 to
Su.K., No. 21710208 to K.K., No. 17018007 to T.S., and the
Professional Program for Agricultural Bioinformatics), by the
National Bio-Resource Project ‘‘Silkworm’’ of MEXT, and Grant-
in-Aid for Scientific Research on Innovative Areas (‘‘Functional
machinery for non-coding RNAs’’) to Su.K. and Y.T.
Received March 6, 2011; accepted April 28, 2011.
Aravin AA, Sachidanandam R, Bourc’his D, Schaefer C, Pezic D, Toth
KF, Bestor T, Hannon GJ. 2008. A piRNA pathway primed by
individual transposons is linked to de novo DNA methylation in
mice. Mol Cell 31: 785–799.
Brennecke J, Aravin AA, Stark A, Dus M, Kellis M, Sachidanandam R,
Hannon GJ. 2007. Discrete small RNA-generating loci as master
regulators of transposon activity in Drosophila. Cell 128: 1089–1103.
Brennecke J, Malone CD, Aravin AA, Sachidanandam R, Stark A,
Hannon GJ. 2008. An epigenetic role for maternally inherited
piRNAs in transposon silencing. Science 322: 1387–1392.
Gunawardane LS, Saito K, Nishida KM, Miyoshi K, Kawamura Y,
Nagami T, Siomi H, Siomi MC. 2007. A slicer-mediated mecha-
nism for repeat-associated siRNA 59 end formation in Drosophila.
Science 315: 1587–1590.
Haase AD, Fenoglio S, Muerdter F, Guzzardo PM, Czech B, Pappin
DJ, Chen C, Gordon A, Hannon GJ. 2010. Probing the initiation
and effector phases of the somatic piRNA pathway in Drosophila.
Genes Dev 24: 2499–2504.
Houwing S, Kamminga LM, Berezikov E, Cronembold D, Girard A,
van den Elst H, Filippov DV, Blaser H, Raz E, Moens CB, et al.
2007. A role for Piwi and piRNAs in germ cell maintenance and
transposon silencing in zebrafish. Cell 129: 69–82.
Houwing S, Berezikov E, Ketting RF. 2008. Zili is required for germ cell
differentiation and meiosis in zebrafish. EMBO J 27: 2702–2711.
Kawaoka S, Hayashi N, Katsuma S, Kishino H, Kohara Y, Mita K,
Shimada T. 2008a. Bombyx small RNAs: genomic defense system
against transposons in the silkworm, Bombyx mori. Insect Biochem
Mol Biol 38: 1058–1065.
Kawaoka S, Minami K, Katsuma S, Mita K, Shimada T. 2008b.
Developmentally synchronized expression of two Bombyx mori
Piwi subfamily genes, SIWI and BmAGO3 in germ-line cells.
Biochem Biophys Res Commun 367: 755–760.
Kawaoka S, Hayashi N, Suzuki Y, Abe H, Sugano S, Tomari Y,
Shimada T, Katsuma S. 2009. The Bombyx ovary-derived cell line
endogenously expresses PIWI/PIWI-interacting RNA complexes.
RNA 15: 1258–1264.
Klattenhoff C, Theurkauf W. 2008. Biogenesis and germline functions
of piRNAs. Development 135: 3–9.
Lau NC, Robine N, Martin R, Chung WJ, Niki Y, Berezikov E, Lai EC.
2009. Abundant primary piRNAs, endo-siRNAs, and microRNAs
in a Drosophila ovary cell line. Genome Res 19: 1776–1785.
Li C, Vagin VV, Lee S, Xu J, Ma S, Xi H, Seitz H, Horwich MD,
Syrzycka M, Honda BM, et al. 2009a. Collapse of germline piRNAs
in the absence of Argonaute3 reveals somatic piRNAs in flies. Cell
Li R, Yu C, Li Y, Lam TW, Yiu SM, Kristiansen K, Wang J. 2009b.
SOAP2: An improved ultrafast tool for short read alignment.
Bioinformatics 25: 1966–1967.
Malone CD, Hannon GJ. 2009. Molecular evolution of piRNA and
transposon control pathways in Drosophila. Cold Spring Harb
Symp Quant Biol 74: 225–234.
Malone CD, Brennecke J, Dus M, Stark A, McCombie WR,
Sachidanandam R, Hannon GJ. 2009. Specialized piRNA pathways
act in germline and somatic tissues of the Drosophila ovary. Cell
Nishida KM, Saito K, Mori T, Kawamura Y, Nagami-Okada T,
Inagaki S, Siomi H, Siomi MC. 2007. Gene silencing mechanisms
mediated by Aubergine piRNA complexes in Drosophila male
gonad. RNA 13: 1911–1922.
Olivieri D, Sykora MM, Sachidanandam R, Mechtler K, Brennecke J.
2010. An in vivo RNAi assay identifies major genetic and cellular
requirements for primary piRNA biogenesis in Drosophila. EMBO
J 29: 3301–3317.
Saito K, Nishida KM, Mori T, Kawamura Y, Miyoshi K, Nagami T,
Siomi H, Siomi MC. 2006. Specific association of Piwi with
rasiRNAs derived from retrotransposon and heterochromatic
regions in the Drosophila genome. Genes Dev 20: 2214–2222.
Saito K, Inagaki S, Mituyama T, Kawamura Y, Ono Y, Sakota E,
Kotani H, Asai K, Siomi H, Siomi MC. 2009. A regulatory circuit
for piwi by the large Maf gene traffic jam in Drosophila. Nature
Saito K, Ishizu H, Komai M, Kotani H, Kawamura Y, Nishida KM,
Siomi H, Siomi MC. 2010. Roles for the Yb body components
Armitage and Yb in primary piRNA biogenesis in Drosophila.
Genes Dev 24: 2493–2498.
Schier AF. 2007. The maternal-zygotic transition: death and birth of
RNAs. Science 316: 406–407.
Thomson T, Lin H. 2009. The biogenesis and function of PIWI
proteins and piRNAs: progress and prospect. Annu Rev Cell Dev
Biol 25: 355–376.
Vagin VV, Sigova A, Li C, Seitz H, Gvozdev V, Zamore PD. 2006. A
distinct small RNA pathway silences selfish genetic elements in the
germline. Science 313: 320–324.
Vastenhouw NL, Zhang Y, Woods IG, Imam F, Regev A, Liu XS, Rinn
J, Schier AF. 2010. Chromatin signature of embryonic pluripo-
tency is established during genome activation. Nature 464: 922–
Yamauchi H, Yoshitake N. 1984. Developmental stages of ovarian
follicles of the silkworm, Bombyx mori L. J Morphol 179: 21–31.
On-site evidence for the ping-pong cycle