A single female-specific piRNA is the primary determiner of sex in the silkworm

Abstract

The silkworm Bombyx mori uses a WZ sex determination system that is analogous to the one found in birds and some reptiles. In this system, males have two Z sex chromosomes, whereas females have Z and W sex chromosomes. The silkworm W chromosome has a dominant role in female determination, suggesting the existence of a dominant feminizing gene in this chromosome. However, the W chromosome is almost fully occupied by transposable element sequences, and no functional protein-coding gene has been identified so far. Female-enriched PIWI-interacting RNAs (piRNAs) are the only known transcripts that are produced from the sex-determining region of the W chromosome, but the function(s) of these piRNAs are unknown. Here we show that a W-chromosome-derived, female-specific piRNA is the feminizing factor of B. mori. This piRNA is produced from a piRNA precursor which we named Fem. Fem sequences were arranged in tandem in the sex-determining region of the W chromosome. Inhibition of Fem-derived piRNA-mediated signalling in female embryos led to the production of the male-specific splice variants of B. mori doublesex (Bmdsx), a gene which acts at the downstream end of the sex differentiation cascade. A target gene of Fem-derived piRNA was identified on the Z chromosome of B. mori. This gene, which we named Masc, encoded a CCCH-type zinc finger protein. We show that the silencing of Masc messenger RNA by Fem piRNA is required for the production of female-specific isoforms of Bmdsx in female embryos, and that Masc protein controls both dosage compensation and masculinization in male embryos. Our study characterizes a single small RNA that is responsible for primary sex determination in the WZ sex determination system.

Full text

LETTER doi:10.1038/nature13315
A single female-specific piRNA is the primary
determiner of sex in the silkworm
Takashi Kiuchi
1
, Hikaru Koga
1
*, Munetaka Kawamoto
1
*, Keisuke Shoji
1
*, Hiroki Sakai
2
, Yuji Arai
1
, Genki Ishihara
1
,
Shinpei Kawaoka
1
, Sumio Sugano
3
, Toru Shimada
1
, Yutaka Suzuki
3
, Masataka G. Suzuki
2
& Susumu Katsuma
1
The silkworm
Bombyx mori
uses a WZ sex determination system
that is analogous to the one found in birds and some reptiles. In this
system, males have two Z sex chromosomes, whereas females have Z
and W sex chromosomes. The silkworm W chromosome has a dom-
inant role in female determination
1,2
, suggesting the existence of a
dominant feminizing gene in this chromosome. However, the W chro-
mosome is almost fully occupied by transposable element sequences
3–5
,
and no functional protein-coding gene has been identified so far.
Female-enriched PIWI-interacting RNAs (piRNAs) are the only known
transcripts that are produced from the sex-determining region of
the W chromosome
6
, but the function(s) of these piRNAs are unknown.
Here we show that a W-chromosome-derived,female-specific piRNA
is the feminizing factor of
B. mori
. This piRNA is produced from a
piRNA precursor which we named
Fem
.
Fem
sequences were arranged
in tandem in the sex-determining region of the W chromosome. Inhi-
bition of
Fem
-derived piRNA-mediated signalling in female embryos
led to the production of the male-specific splice variants of
B. mori
doublesex
(
Bmdsx
), a gene which acts at the downstream end of the
sex differentiation cascade
7,8
. A target gene of
Fem
-derived piRNA
was identified on the Z chromosome of
B. mori
. This gene, which we
named
Masc
, encoded a CCCH-type zinc finger protein. We show
that the silencing of
Masc
messengerRNA by
Fem
piRNA is required
for the production of female-specific isoforms of
Bmdsx
in female
embryos, and that Masc protein controls both dosage compensation
and masculinization in male embryos. Our study characterizes a
single small RNA that is responsible for primary sex determination
in the WZ sex determination system.
In Bombyx mori,sex determination is probably established at an early
stage of embryogenesis. We prepared the sexed RNA from individual
silkworm embryos genotyped by three W chromosome-specific ran-
domly amplified polymorphicDNA (RAPD) markers
4
(Extended Data
Fig. 1a), and examined the splicing pattern of a doublesex orthologue of
B. mori (Bmdsx). Bmdsx produces female- and male-specific RNAs by
sex-specific alternative splicing
9
that have essential roles in silkworm
sexual development
7,8
. Female-specific splicevariants of Bmdsx were the
default transcripts during an early stage of development. Whereas the
male-specificsplice variants clearly appeared in male embryos, only faint
bands were observed in females, from 21–24 h post-oviposition (hpo)
(Fig. 1a). This indicated that the feminizing signal is transmitted from
the W chromosome before 21 hpo. Thus, we performed deep sequenc-
ing of RNAs (RNA-seq) isolated from male and female embryos at 15,
18, 21 and 24 hpo, and identified differentially expressed transcripts
between male and female embryos.
One contig, comp73859_c0, was consistently identified in female
embryos at all of the developmental times tested (Extended Data Fig. 1b, c).
This sequence was amplified by PCR only when female genomic DNA
or complementary DNA was used as a template (Fig. 1b). The sequence
of comp73859_c0 did not show significant identity with any sequence
in the draft male silkworm genome sequence
10
, suggesting that this contig
is localized on and transcribed from the W chromosome. In addition,
this sequence was amplified from genomic DNA isolated from female
wild silkmoth Bombyxmandarina, but notfrom males (Fig. 1b). Reverse
transcription followed by quantitative PCR (RT–qPCR) showed that
the expression level peaked at 18–21 hpo, and then gradually declined
during embryogenesis (Fig. 1c). This transcript was also detected in
the ovary and other somatic tissues (Extended Data Fig. 2a, b, d). Long
PCR demonstrated that there are multiple copies of this sequence on
the W chromosome (Extended Data Fig. 2c). The sequences were occa-
sionally arranged in tandem and some were probably expressedas long
transcriptional units. Northern blot analysis revealed that transcripts
of approximately 0.8 and 1.4 kilobase are major units, and antisense
transcripts were not detected (Extended Data Fig. 2d). The copy number
of this contig per haploid genome was estimated at more than 30 in the
genome of B. mori.
This contig did not show homology with any known sequence, nor
did it seem to encode a functional protein. Instead, this transcript seemed
to be a piRNA precursor. Mapping of embryonic or ovarian piRNAs
6,11,12
onto this transcript and northern blotting revealed a 29-nucleotide-long
piRNA-producing region (Fig. 1d and Extended Data Fig. 3a). This piRNA
was poorly transmitted from the mother moth, accumulated from 15 hpo,
*These authors contributed equally to this work.
1
Department of Agricultural and Environmental Biology, Graduate School of Agricultural and Life Sciences, The University of Tokyo, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113-8657, Japan.
2
Department of
Integrated Biosciences, Graduate School of Frontier Sciences, The University of Tokyo, 5-1-5 Kashiwanoha, Kashiwa, Chiba 277-8562, Japan.
3
Department of Medical Genome Sciences, Graduate School of
Frontier Sciences, The University of Tokyo, 4-6-1 Shirokanedai, Minato-ku, Tokyo 108-8639, Japan.
]
]
F
M
a
gDNA cDNA
b
cd
comp73859_c0
1.0
0.5
kb
1F2R
B. mori B. mandarina
(hpo)
1.06
0.50
kb
767 bases
gDNA
0
0.5
1.0
1.5
2.0
2.5
3.0
15 18 21 24 27 15 18 21 24 27
mRNA level
(contig/rp49)
hpo
0
Sense
piRNA
Reads per million
200
100
0
668 bp
UACCUCUUUUUGUCAAUUCAUAAAGUCAU
Antisense
15 18 21 24 27
Figure 1
|
Characterization of a female-specific piRNA precursor in early
silkworm embryos. a, Splicing patterns of Bmdsx in early embryos. The F and
M indicate female- and male-type splicing of Bmdsx, respectively. Similar
results were obtained in three independent experiments. b, Detection of a
W chromosome-derived transcript. Genomic DNA (gDNA) and cDNA were
prepared from female and male embryos of B.mori at 24 hpo (left panel) or
adult B. mandarina (right panel). c, Expression profile of the female-specific
contig in early embryos. Data shown are means6s.d. of three embryos.
d, Mapping of embryonic piRNAs (24 hpo) onto the comp73859_c0. The
relative location, abundance and sequence of the 29 base-long piRNA (shown
in green) are indicated.
00 MONTH 2014 | VOL 000 | NATURE | 1
Macmillan Publishers Limited. All rights reserved
©2014

and increasedrapidly between 18 and21 hpo (Extended Data Fig.3b, c).
By screening piRNA libraries that were generated from three B. mori
strains that each possess a unique W chromosome structure
6
, we found
that this piRNA was produced from the sex-determining region of W
chromosome (Extended Data Fig. 3d).
The silkworm KG strain, which possesses a mutation(s) in the W
chromosome, shows various degrees of female masculinization features
13
.
Expression of the contig-derived piRNA in the masculinized ovary was
markedly lower than in wild type (Extended Data Fig. 4a), indicating
that a certain amount of this piRNA might be required for complete
feminization of adult moths. To investigate the role of this piRNA, we
used a unique RNA-based inhibitor that seemed to function in BmN4
cells (Fig. 2a and Extended Data Fig. 4b) and investigated the effect of
the inhibitor on Bmdsx splicing in early embryos. The Bmdsx splicing
was markedly altered to produce the male-type isoform when the inhibitor
was injected into female embryos, whereas the pattern was not affected
in male embryos (Fig. 2b–c). The splicing pattern, however, was not
altered in newly hatched larvae, presumably because thisinhibitor does
not possess the long-term inhibitory activity (Extended Data Fig. 4c).
These results demonstrated that the targeted piRNA is required for the
female-type splicing of Bmdsx. Thus, we named the precursor of this
piRNA, Feminizer (Fem).
We performed RNA interference (RNAi) experiments that targeted
two core components of thesilkworm piRNA biogenesis pathway called
Siwi and BmAgo3 (Extended Data Fig. 4d)
14,15
. Small interfering RNA
(siRNA)-mediated knockdown of Siwi in female embryos commonly
led to the production of the male-type Bmdsx transcripts, whereas little
effect was observed in male embryos (Fig. 2d and Extended Data Fig. 4e).
BmAgo3 RNAi did not affect Bmdsx splicing in either female or male
embryos (Fig. 2d and Extended Data Fig. 4e). These results indicated
that Siwi expression is crucial for the female-type splicing of Bmdsx in
female embryos.Our hypothesis as to why BmAgo3 knockdown did not
affect Bmdsx splicing in early embryos is discussed later.
We identified only one genomic locus where the Fem piRNA sequence
was extensively complementary (Fig. 3a). This locus was present within
the ninth exon of an uncharacterized gene located on the Z chromosome
(Fig. 3a and Extended DataFig. 5a). We named this gene Masculinizer
(Masc). Masc potentially encoded a novel CCCH-tandem zinc finger
protein (Extended Data Fig. 5b). Phylogenetic analysis suggested that
this protein forms a novel lepidopteran-specific protein family (Extended
Data Fig. 5c). PIWI–piRNA complexes are known to cleave their com-
plementary target sequences across from positions 10 and 11 of the guide
piRNA
16,17
. By a modified 59rapid amplification of cDNA ends (mod-
ified RACE) method
18
, we found that all of the cloned 59ends of the
a piRNA 5′
UAACAauggagaaaaacaguuaaguauuucaguaAAUGC
uaccucuuuuugucaauucauaaagucau
3′
3′
5′
(2′-O-Me)
(2′-O-Me)
piRNA inhibitor
d
0
20
40
60
80
100
(%)
GFP siRNA: Siwi Ago3 GFP Siwi Ago3
InhibitorControl
InhibitorControl InhibitorControl
]
]
F
M0
20
40
60
80
100
F-type M-type F+M-type
F-type M-type F+M-type
(%)
c
13
16 11 11 8 898714 14
11 6 11
b
1212 121 2
Figure 2
|
Female-specific piRNA is a primary sex determinant of
B. mori
.
a, Structure of the female-specific piRNA and its inhibitor. b,c, Effect of the
RNA inhibitor on the Bmdsx splicing. The splicing patterns were examined at
72 h post-injection. Representative splicing patterns are shown in band the
data are summarized in c. The number indicates the sample size. F-type,
female-type; M-type, male-type; F1M-type, both variants are mixed.
d, Splicing of Bmdsx in embryos that were injected with Siwi or BmAgo3
siRNAs. Two types of siRNAs for each target were used. The splicing patterns
were examined at 72h post-injection. The numbers in the columns indicate
sample size.
a
b
-aaauggcuuugugaaucgacaaaaagagguaacaauugaagcuaaucagaagaaaa-
aaaagagguaacaauugaagcuaaucaga
uacugaaauacuuaacuguuuuucuccau
-gugacuuacugaaauacuuaacuguuuuucuccauuguuacuuu-
aaaagagguaacaauugaagcuaaucaga
-aaauggcuuugugaaucgacaaaaagagguaac-
uacugaaauacuuaacuguuuuucuccau
Fem piRNA 5′
Fem piRNA 5′
Fem 5′
Masc piRNA 5′
Masc piRNA 5′
Masc mRNA 5′
Fem 3′
Masc mRNA 5′
-acuguuuuucuccauuguuacuuucuuuuagucguguuu-
Cleavage site
Cleava
g
e site
Z chromosome (Bm_scaf26)
ATG TA G
0
1
2
3
4
mRNA level (Masc/rp49)
(control female = 1)
Control Inhibitor Control
12 9 7
c
**
Figure 3
|
Masc
mRNA is the target of
Fem
piRNA. a, Genomic structure of
the Masc gene on the Z chromosome of B.mori. The putative cleavage site
by the Fem piRNA–Siwi complex is shown by the red line. b, Identification
of Masc piRNA. The ping-pong signature (within the red box) and putative
cleavage site (red line) of Fem by the Masc piRNA–BmAgo3 complex are
shown. c,Masc expression in the piRNA inhibitor-injected embryos at 18 hpo.
The numbers in the columns indicate sample size. Data shown are means1s.d.
Data were subjected to Kruskal–Wallis analysis with post hoc Dunn’s test.
*P,0.05.
RESEARCH LETTER
2 | NATURE | VOL 000 | 00 MONTH 2014
Macmillan Publishers Limited. All rights reserved
©2014

Masc-derived RNA fragments from early embryos mapped precisely
to the predicted Fem piRNA cleavage site (Extended Data Fig. 6a), indi-
cating that Masc mRNA is the target of Fem piRNA.
The piRNA biogenesis occurs through a ping-pong mechanism that
involves two different PIWI proteins. A 10-nucleotide overlap between
sense and antisense piRNAs, called a ping-pong signature, is the hall-
mark of the cleavage reaction catalysed by PIWI proteins
16,17
.Wefound
piRNAs that havea perfect 10-nucleotideoverlap with Fem piRNA. The
most abundant of these piRNAs were those that perfectly matched to
the Masc coding region (Fig. 3b and Extended Data Fig. 6b, c), indicat-
ing that the Masc mRNA-derived piRNA (Masc piRNA) is a ping-pong
partner of Fem piRNA. The Masc piRNA was extensively complemen-
tary to Fem (Fig. 3b), indicating that the PIWI–Masc piRNA complex
will reliably slice Fem RNA. Fem piRNA preferentially bound to Siwi,
whereas Masc piRNA preferentially bound to BmAgo3(Extended Data
Fig. 6d). Thus, our findings indicate a ping-pong amplification model
for Fem and Masc piRNAs (Extended DataFig. 6e). This modelis exper-
imentally supported by the introduction of the inhibitor for Fem piRNA
(Fig. 3c)or siRNAs for Siwi (Extended DataFig. 7a) into female embryos
showing enhanced Masc levels. Unlike Fem piRNA, a moderate amount
of Masc piRNA was maternally transmitted (Extended Data Fig. 6f, g).
Together with the fact that BmAgo3 is also maternally transmitted
12
,
this suggests that a moderate amount of the Masc piRNA–BmAgo3 com-
plex exists even in newly laid eggs. The presence of this complex helps
to explain why BmAgo3 RNAi in female embryos had little effect on
Bmdsx splicing (Fig. 2d). Embryonic RNAi for BmAgo3 did not alter
the Bmdsx splicing, but enhanced Masc expression in newly hatched
female larvae (Extended Data Fig. 7b–d), supporting the roleof the Masc
piRNA–BmAgo3 complex in sex determination. Higher levels of Masc
piRNA were detected from 21–27hpo (Extended Data Fig. 6f, g); this
increase correlated with a massive accumulation of Fem piRNA (Extended
Data Fig. 3b, c).
In male embryos, Masc expression rapidly increased, then rapidly
decreased between 15 and 18, and 18 and 21 hpo,respectively (Fig. 4a).
In contrast,Ma sc expressionin female embryos graduallydeclined from
15 hpo, and remained at a low level compared with that found in males
(Fig. 4a). These data indicate that Fem piRNA-mediated cleavage of
Masc mRNA results in low-level accumulation of Masc mR NA in female
embryos. Injection of Masc siRNA into male embryos reduced Masc
expression to levels that were found in control female embryos at 18 hpo
(Fig. 4b), and resulted in the production of female-type variants of Bmdsx
throughout the embryonic stage (Fig. 4c and Extended Data Fig. 8a, b).
Female embryos injected with Masc siRNA hatched normally, whereas
male embryos did not (Fig. 4d), indicating that inhibition of the Masc
pathway at the embryonic stage results in male-specific lethality. This
probably mimics the way that an arthropod pathogen Wolbachia induces
a male-killing phenotype in lepidopteran insects
19
. The Fem piRNA-
resistant Masc (Masc-R) mRNA was more accumulated than the wild-
type Masc mRNA in Masc cDNA-transfected BmN4 cells, whereas Masc
piRNA was poorly detected in Masc-R cDNA-transfected cells (Extended
Data Fig. 9a–c). The Masc-R mRNA was not cleaved, which changed the
Bmdsx splicing pattern in BmN4 cells to the male-type completely, and
induced a growth inhibition (Extended Data Fig. 9d–f), indicating that
the Fem piRNA-mediated cleavage of Masc mRNA is essential for silk-
worm feminization.
RNA-seq analyses of Masc siRNA-injected embryosrevealed that the
transcripts differentially expressed in males were mapped predominantly
onto the Z chromosome (chromosome 1, 51%), whereas such a bias was
not observed in females (Fig. 4e). Most of the Z-chromosome-derived
transcripts expressed differentially in males (97%) were expressed higher
in Masc RNAi embryos (Fig. 4e) and randomly dispersed throughout
this chromosome (Extended Data Fig. 10). These results demonstrate
that Masc protein globally represses gene expression from the male Z
chromosome at the embryonic stage. Taken together, Mascprotein con-
trols both dosage compensation and masculinization (Fig. 4f). In Dro-
sophila, Sex-lethal, a master switch for sex determination, controls dosage
compensation by inhibiting translation of male-specific lethal 2 (msl-2)
20
.
Loss of msl-2 causes male lethality, owing to the failure of hypertran-
scription from the male X chromosome. A failure of dosage compensation
is probably involved in male-specific lethality of Masc mRNA-depleted
male embryos.
We unravelled a question that has perplexed insect geneticists for
more than eight decades. Our study answersthe question of how the W
chromosome determines the femaleness of the silkworm B. mori. The
silkworm feminizer Fem is the precursor of a 29-nucleotide-long small
RNA. To our knowledge, this is the first example of the identification
of a primary sex-determining factor in Lepidoptera, and the first exper-
imental evidenceshowing a piRNA-mediated sex determination mech-
anism. Our findings also suggestthat Masc levels may be involved in sex
determination in lepidopteran species that are monosomic (Z0) in females
and ZZ in males
21
. We are now experimentally surveying this hypo-
thesis using moth species with a Z0/ZZ sex chromosome constitution.
0
10
20
30
40
50
60
0
10
20
30
40
50
60 siGFP > siMasc
siGFP < siMasc
1 5 10 15 20 25 NA
siGFP > siMasc
siGFP < siMasc
Chromosome number
e
0
0.5
1.0
1.5
2.0
hpo
15 18 21 24 27 0
50
100
150
Relative level
(%, siGFP male = 100)
MascMasc
MascMascMascMasc
16 11 9 5 8 9
0
20
40
60
80
100
(%)
siRNA: GFP 12
siRNA: GFP 12GFP 12siRNA: GFP 12GFP 12
GFP 12
a b
f
0
20
40
60
80
100
F-type M-type F+M-typeF-type M-type F+M-type (%)
d
c
Not hatched
66610455888
Mapped contigs (%)
WZ
Z
Female-specic
splicing
Male-specic
splicing
Masc mRNA
Masc mRNA
Fem piRNA
Fem
Bmdsx
Bmdsx
Dosage
compensation
Protein
mRNA level
(Masc/rp49)
**
**
**
Figure 4
|
Masc protein controls both masculinization and dosage
compensation in male embryos. a, Expression profile of Masc in early
embryos. Data shown are means6s.d. of three embryos. b, Knockdown of
Masc mRNA in B.mori embryos. The embryos were injected with two types
of siRNAs for Masc, and Masc expression was examined by RT–qPCR at
18 hpo. Data shown are means 1s.d. The number indicates the sample size.
One-way ANOVA was performed with post hoc Tukey’s test. *P,0.05.
c,d, Splicing of Bmdsx in Masc siRNA-injected embryos. The splicing pattern
was determined at 72 h (c) and about 240 h (d, immediately after hatching)
post-injection.The number indicates the sample size. e, Differentiallyexpressed
transcripts in Masc RNAi embryos. siGFP, GFP siRNA-injected embryos;
siMasc, Masc siRNA-injected embryos; NA, not assigned. f, A proposed model
for the sex determination pathway in B. mori.
LETTER RESEARCH
00 MONTH 2014 | VOL 000 | NATURE | 3
Macmillan Publishers Limited. All rights reserved
©2014

METHODS SUMMARY
Sex-specific splicingof Bmdsx,piRNA mapping, qRT–PCRof piRNA, and transfec-
tion experiments using BmN4 cells were performed as described previously
6,12,13,14
.
Embryonic RNAi was performed by injecting embryos with two different siRNAs
for each gene investigated.
Online Content Any additional Methods, ExtendedData display items and Source
Data are available in the online version of the paper; references unique to these
sections appear only in the online paper.
Received 2 October 2013; accepted 8 April 2014.
Published online 14 May 2014.
1. Hasimoto, H. The role of the W-chromosome in the sex determination of Bombyx
mori [in Japanese]. Jpn. J. Genet. 8, 245–247 (1933).
2. Tajima, Y. Studies on chromosome aberrations in the silkworm. II. Translocation
involving second and W-chromosomes [in Japanese]. Bull. Seric. Exp. Stn. 12,
109–181 (1944).
3. Abe, H. et al. Identification of novel random amplified polymorphic DNAs (RAPDs)
on the W chromosome of the domesticated silkworm, Bombyx mori, and the wild
silkworm, B. mandarina, and their retrotransposable element-related nucleotide
sequences. Genes Genet. Syst. 73, 243–254 (1998).
4. Abe, H. et al. Partial deletions of the W chromosomedue to reciprocal translocation
in the silkworm Bombyx mori. Insect Mol. Biol. 14, 339–352 (2005).
5. Abe, H., Mita, K., Yasukochi, Y., Ohshiki, T. & Shimada, T. Retrotransposable
elementson the W chromosome of the silkworm, Bombyxmori. Cytogenet. Genome
Res. 110, 144–151 (2005).
6. Kawaoka, S. et al. The silkworm W chromosome is a source of female-enriched
piRNAs. RNA 17, 2144–2151 (2011).
7. Suzuki, M. G., Funaguma, S., Kanda, T., Tamura, T. & Shimada, T. Analysis of the
biological functions of a doublesex homologue in Bombyx mori. Dev. Genes Evol.
213, 345–354 (2003).
8. Suzuki, M. G., Funaguma, S., Kanda, T., Tamura, T. & Shimada, T. Role of the male
BmDSX protein in the sexual differentiation of Bombyx mori. Evol. Dev. 7, 58–68
(2005).
9. Ohbayashi, F., Suzuki, M. G., Mita, K., Okano, K. & Shimada, T. A homologue of the
Drosophila doublesex gene is transcribed into sex-specific mRNA isoforms in the
silkworm, Bombyx mori. Comp. Biochem. Physiol. B Biochem. Mol. Biol. 128,
145–158 (2001).
10. The International Silkworm Genome Consortium. The genome of a lepidopteran
model insect,the silkworm Bombyx mori.Insect Biochem. Mol.Biol. 38, 1036–1045
(2008).
11. Kawaoka, S. et al. Bombyx small RNAs: genomic defense system against
transposons in the silkworm, Bombyx mori. Insect Biochem. Mol. Biol. 38,
1058–1065 (2008).
12. Kawaoka, S. et al. Zygotic amplification of secondary piRNAs during silkworm
embryogenesis. RNA 17, 1401–1407 (2011).
13. Hara, K. et al. Alteredexpression of testis-specific genes, piRNAs, and transposons
in the silkwormovary masculinizedby a W chromosome mutation.BMC Genomics
13, 119 (2012).
14. Kawaoka, S. et al. The Bombyx ovary-derived cell line endogenously
expresses PIWI/PIWI-interacting RNA complexes. RNA 15, 1258–1264
(2009).
15. Kawaoka, S., Minami, K., Katsuma, S., Mita, K. & Shimada, T. Developmentally
synchronized expression of two Bombyx mori Piwi subfamily genes, SIWI and
BmAGO3 in germ-line cells. Biochem. Biophys. Res. Commun. 367, 755–760
(2008).
16. Brennecke, J. et al. Discrete small RNA-generating loci as master regulators of
transposon activity in Drosophila. Cell 128, 1089–1103 (2007).
17. Gunawardane, L. S. et al. A slicer-mediated mechanism for repeat-associated
siRNA 59end formation in Drosophila. Science 315, 1587–1590 (2007).
18. Watanabe,T. et al. Role of piRNAs and noncodingRNA in de novo DNA methylation
of the imprinted mouse Rasgrf1 locus. Science 332, 848–852 (2011).
19. Sugimoto, T. N. & Ishikawa, Y. A male-killing Wolbachia carries a feminizing factor
and is associated with degradation of the sex-determining system of its host. Biol.
Lett. 8, 412–415 (2012).
20. Penalva, L. O. & Sa
´nchez, L. RNA binding protein sex-lethal (Sxl) and control of
Drosophila sex determination and d osage compensation. Microbiol. Mol. Biol. Rev.
67, 343–359 (2003).
21. Traut, W., Sahara, K. & Marec, F. Sex chromosomes and sex determination in
Lepidoptera. Sex Dev. 1, 332–346 (2007).
Supplementary Information is available in the online version of the paper.
Acknowledgements We thank S. G. Kamita for critical reading of the manuscript;
Y. Tomari for critical reading of the manuscript and technical suggestions. This work
was supported by the Program for Promotion of Basic and Applied Researches for
Innovations in Bio-oriented Industry to Su.K. and Grants-in-Aid for Scientific Research
on Innovative Areas (Nos. 22115502 and 22128004) to Su.K. and T.S.
Author Contributions Su.K., T.K. and M.G.S. conceived and designed the experiments.
T.K., H.K., K.S., H.S., G.I., Y.A., Sh.K., M.G.S. and Su.K. performed molecular biological
experiments.M.K. and K.S. performed most of the bioinformatic analyses. S.S.and Y.S.
performed deep sequencing and data analysis. T.S. provided essential reagents and
expertise. All of the authors discussed the data and helped manuscript preparation.
Su.K. wrote the manuscript with intellectual input from all authors. Su.K. supervised
the project.
Author Information The nucleotide sequences of Fem and Masc have been deposited
in the DDBJ/EMBL/GenBank databank under the accession numbers AB840787 and
AB840788. Deep sequencing data obtained in this study are available under the
accession numbers DRA001104 and DRA001338 (DDBJ), respectively. Reprints and
permissions information is available at www.nature.com/reprints. The authors declare
no competing financial interests. Readers are welcome to comment on the online
version of the paper. Correspondenceand requests for materials shouldbe addressed
to Su.K. (katsuma@ss.ab.a.u-tokyo.ac.j p).
RESEARCH LETTER
4 | NATURE | VOL 000 | 00 MONTH 2014
Macmillan Publishers Limited. All rights reserved
©2014

METHODS
Insects and cell lines. Larval B. mori (p50T, N4, and F1 hybrid Kinshu 3Showa)
and B. mandarina were reared as described previously
6
. The silkworm ovary-derived
BmN4 cells were grown at 27 uC in TC-100 or IPL-41 medium supplemented with
10% fetal bovine serum
14
.
Molecular sexing. Total RNA and genomic DNA were prepared simultaneously
from a single embryo using TRIzol reagent (Invitrogen) according to the manu-
facturer’s protocol. We previously reported that the polar-body-derived W chromo-
some fragment can be detected at the early stage of embryogenesis
22
. To perform
accurate molecular sexing of each embryo, we used three sets of W chromosome
primers for PCR (Supplementary Table 1) or performed RT–qPCR for Fem.
RNA-seq. Libraries for RNA sequencingwere generated from 15, 18, 21, 24 hpo of
molecularly sexed embryos using the TruSeq RNA Sample Preparation kit (Illumina)
and were analysed using the IlluminaHiSeq 2000 platform with 101-bp paired-end
reads (normal embryo samples, 8 data set) or HiSeq 2500 platform with 100-bp
paired-end reads (RNAi embryo samples, 4 data set) according to the manufac-
turer’s protocol
23
.
Quantifications of
Fem
copy number. We estimated Fem copy number per hap-
loid genome by quantitative PCR as reported previously
24
. Genomic DNA was
extracted from larval tissues using standard procedures. Siwi was used as a single
copy control gene on the autosome. qPCR analyses were performed using a KAPA
SYBR FAST qPCR kit (KapaBiosystems) and specific primers listedin Supplemen-
tary Table 1.
RT–PCR. Total RNA was prepared using TRIzol reagent (Invitrogen) according
to the manufacturer’s protocol and subjected to reverse transcription using avian
myeloblastosis virus (AMV) reverse transcriptase with an oligo-dT primer (TaKaRa).
PCR was carried out with KOD FX-neo DNA polymerase (TOYOBO). Sex-specific
splicing of Bmdsx was examined by PCR with primers listed in Supplementary
Table 1
25
. RT–qPCR analyses were performed using a KAPA SYBR FAST qPCR
kit (Kapa Biosystems) and specific primers listed in Supplementary Table 1. RT–
qPCR of piRNAs was performed as described previously
6
. In brief, small RNA
fractions were enriched with the aid of a mirVana miRNA isolation kit (Ambion)
and reverse transcribed using a miScript Reverse Transcription Kit (QIAGEN).
qPCR was performed using a miScript PCR System (QIAGEN). The qPCR pro-
ducts were verified by cloning and DNA sequencing. let-7, one of the well-known
silkworm microRNAs, was used as a control. The primers used in this experiment
are described in Supplementary Table 1.
Embryonic RNAi. The short interfering RNA (siRNA) sequences listed in Sup-
plementary Table 1 were designed based on the ORF sequences of thetarget genes
and enhancedgreen fluorescent protein (GFP,control). Two different siRNAswere
designed for each gene (that is, Siwi-1 and Siwi-2). Double-stranded siRNAs were
purchased from FASMAC Corp (Japan), dissolved in annealing buffer (100 mM
potassium acetate, 2 mM magnesium acetate, 30 mM HEPES-KOH; pH7.4), and
stored at 280 uC for later use. The B. mori N4 eggs used for siRNA injection were
prepared as described previously
26
. Injection was performed accordin g to the method
described previously
27
using a microinjector (IM 300 Microinjector, Narishige Japan).
One to 5 nl of each siRNA solution (50 mM for Siwi, 100 mM for BmAgo3 and
Masc (18 hpo and 72 h post-injection), and 500 mM for Masc (144, 216 and about
240 h post-injection)) was injected into each egg within 4–8 h after oviposition.
The injected embryos were incubated at 25 uC in a humidified Petri dish. At 72 h
post-injection, the expression level of the target gene was quantified by RT–qPCR,
and samples whose target mRNA level (Siwi and BmAgo3) was knocked down
by at least 80% was used for further analysis. Masc expression levels in embryos
that were injected with siRNA were analysed at 18 hpo. Randomization and blind-
ing were not applied to determine how embryo samples were allocated to experi-
mental groups, because it is not possible to visually distinguish female and male
embryos of silkworm N4 strain. The expression levels of rp49 were used to nor-
malize transcript levels. Primers used for RT–qPCR are listed in Supplementary
Table 1.
Injection of the piRNA inhibitor into embryos. We designed a unique RNA-
based inhibitor by modification of a previously described strategy
28
(Fig. 2a). We first
tested the efficacyof our inhibitor using BmN4 cell line, a silkworm-ovary-derived,
W-chromosome-harbouring cell line. BmN4 cells express the corresponding piRNA
precursors (Extended Data Fig. 2b, d) as well as female-type Bmdsx transcripts
(Extended Data Fig. 4b), and possess a complete piRNA pathway
14
. The male-type
splice variant of Bmdsx was enhanced in BmN4 cells when transfected with the
inhibitor (Extended Data Fig. 4b), indicating that our RNAinhibitor functioned to
inhibit the piRNA-mediated signalling cascade.
One to five nl of a 1 mM RNA solution (anti-Fem piRNA or anti-GFP piRNA, Sup-
plementary Table 1) was injected into the B.mori N4 strain eggs within 4–8 h after
oviposition as described above. Masc expression levels in embryos that wereinjected
with the inhibitor of Fem pi RNA were analysed at 18 hpo.
RNA transfection in BmN4 cells. BmN4 cells (2.5 310
5
cells per 60-mm diameter
dish) were transfected with single-stranded RNAs (250 pmol per dish, Supplementary
Table 1) using X-tremeGENE HP (Roche)
29
. Following incubation for 12 h, the
culture medium was removedand fresh medium was added. Cells were collected at
48 h after transfection, and total RNA was isolated. For transfection experiments
using BmN4 cells, at least three independent experiments were performed.
Transient expression of
Masc
mRNA in BmN4 cells. The Fem piRNA-resistant
Masc (Masc-R) cDNA was constructed by using PrimeSTAR Mutagenesis Basal
Kit (TaKaRa). Five nucleotide mutations that do not result in amino acid sub-
stitutions for the Masc protein were introduced (Extended Data Fig. 9a). Masc or
Masc-R cDNA was cloned into the pIZ/V5-His vector (Invitrogen). BmN4 cells
(2.5 310
5
cells per 35-mm diameter dish) were transfected with plasmid DNAs
(0.5 mg) using FuGENE HD (Promega)
29
. Cells were collected at 72 h after trans-
fection. mRNA was prepared using Micro-FastTrack 2.0 Kit (Invitrogen) and sub-
jected to RT–qPCR. Masc mRNA level wasnormalized to that of rp49.Masc piRNA
was also quantified by RT–qPCR as described above.
Generation of BmN4 cells stably expressing Masc proteins. BmN4 cells stably
expressing empty vector (pIZ/V5-His), Masc or Ma sc-R were generated as described
previously
14
. Three daysafter transfection,zeocin (finalconcentration, 500 mgml
21
)
was added to the medium. Six days after drug selection, the splicing patterns of
Bmdsx were examined by RT–PCR.
Northern blot analysis. Total RNA was separated by electrophoresis, transferred
to a nylon membrane, and probed with strand-specific oligonucleotide probes as
described previously
30
with some modifications. Small RNA fractions for piRNA
detection were prepared from early embryos whose diapause was artificially ter-
minated. The probe sequences are listed in Supplementary Table 1.
Modified RACE. The Masc mRNA-derived RNA fragments weredetermined by a
modified RACE procedure as described previously
18
. To detect the cleaved frag-
ments from exogenously introduced Masc, we used the primers designed on the
pIZ/V5-His vector (Extended Data Fig. 9d, Supplementary Table 1).
RNA-seq analysis. De novo assembly of RNA-seqdata from 8 data sets (15, 18, 21,
24 hpo of each sex, 303,483,056 reads in total) was performed using Trinity
31
, and
221,677 contigs (170,255 kinds of transcripts) were produced. Transcript abun-
dance in each contig was quantifiedby RSEM
32
. Differentiallyexpressed transcripts
(adjusted Pvalue ,0.05) between female and male embryos were identified by
the R/Bioconductor package, DESeq
33
. Contigs with more than 10 transcripts per
million at any data set were selected and 157 contigs were used for further analysis.
Fem contig was the only transcript showing significantly statistical scoresbetween
female and male at all time points examined (adjusted Pvalues were 1.89 310
26
at 15 hpo, 1.42 310
228
at 18 hpo, 1.92 310
27
at 21 hpo, and 3.24 310
283
at 24 hpo).
The R-code for this analysis is available as Supplementary Information.
Analysis of RNA-seq data from Masc RNAi experiments (GFP and Masc RNAi
embryos of each sex, 72h post-injection, 4 data sets) was performed as described
above. We selected585 and 608 differentially expressed transcripts(Pvalue ,0.05,
GFP siRNA versus Masc siRNA-1) inmale andfemale, respectively. The chromo-
some on which each transcriptis localized was identified by mappingthe contigs to
the silkworm genome scaffolds.
Raw RNA-seq data from control and Masc RNAi embryos were also mapped to
the silkworm genome scaffolds by Bowtie
34
without mismatches. The coverage at
each nucleotide position was esti mated by coverageBed (included in BEDtools). The
total mapped reads in each RNA-seq library to the scaffolds were used for normal-
ization. The average coverage across each 1-kb window was determined and com-
pared between the two RNA-seq libraries. The genome regions where the average
coverage was more than 10 in either library were selected, grouped into three cate-
gories (siGFP/siMasc .2, 0.5 #siGFP/siMasc #2, and siGFP/siMasc,0.5) and
visualized as Extended Data Fig. 10.
piRNA mapping. piRNA mapping was performed allowing two mismatches by
Bowtie as desc ribed previously
6
. The total mappedreads in each piRNA library
6,11–14
to B. mori repetitive sequences (121 annotated transposons and 1,690 ReAS clones)
were used for normalization.
To determine the genomic locus from which Fem piRNA is produced, we used
piRNA libraries prepared from three B. mori strains that each possess a unique W
chromosome structure
6
(Extended Data Fig. 3d). The sex-limited yellow (LY) strain
35
has a W chromosome that is approximately 90% shorter than the W chromosome
of wild-type B.mori. This extensively truncated W chromosome, however, retains
the ability to determine femaleness
35
, indicating that this W fragment contains the
putative sex-determining region. Of 12 RAPD markers identifiedin the normal W
chromosomes, the LY W chromosome contained only one (W-Rikishi). The sex-
determining region can be defined as the region where the W-Rikishi marker
exists
35
. The DfZ-DfW strain (‘without Fem’, WF) on the other hand has a trun-
cated W chromosome (approximately 75% shorter than the wild-type chromo-
some) that is attached to a Z chromosome
36
. This W chromosome fragment is not
sufficient for determining femaleness, and indicates that it does not contain the
LETTER RESEARCH
Macmillan Publishers Limited. All rights reserved
©2014

sex-determiningregion
36
. The Mandarina W (MW) strain of B.morihas a W chro-
mosome that originates from B. mandarina
6
. When B. mandarina is crossed with
B. mori, fertile hybrids are produced, indicating that the W chromosome of B.
mandarinacan determine the femaleness ofB. mori, and implying that both species
use the same sex-determination system. Examining abundance of Fem piRNA in
each piRNA library
6,11,12,14
revealed that Fem piRNA wasexpressed in the ovaries of
wild-type B.mori but notin the testes (Extended Data Fig. 3d). In ovariesfrom the
LY and MW strains, this piRNA was expressed at 9% and 26%, respectively, of the
level found in wild-type B.mori.Testes from the WF strain expressed an extremely
low level (0.4% of the wild-type) of this piRNA even though the Z chromosome
of this strain retains one-fourth of the W chromosome (Extended Data Fig. 3d).
These results indicated that the sex-determining region of W chromosome pro-
duces Fem piRNA.
Target search for
Fem
piRNA. Base pairing of 11 or 12 nucleotides (nucleotides
2–12 or 2–13) at the 59end of a target sequence of the piRNA is required for effi-
cient target cleavage by the mouse Piwi protein homologue Miwi
37
. To identify a
potential target of Fem piRNA, we searched for genomic sequences of B. mori that
were completely identicalto nucleotides 2–12 of the 59-end Fem piRNA. From this
search we identified three candidate loci, among which Masc showed the lowest
Evalue of 0.008, whereas the other two candidate loci showed Evalues that were
.0.1. Bioinformatic analysis using the silkwormtranscriptome and genome data-
bases revealed that thesetwo loci were notlocated within a predicted protein-coding
gene or transcriptional unit. The Masc locus wasthus predictedas the primarytarget
of Fem piRNAs.
Phylogenetic analysis. The amino acid sequences of proteins in the NCBI data-
base that showed significant homology (Evalue of ,1310
29
) to residues 51–122
of Masc were identified using the BLAST program. A neighbour-joining tree was
constructed using 39 sequences and the reliability of the tree was tested by boot-
strap analysis with 1,000 replications.
Statistical analysis. The sample size in each experiment was adjusted depending
on the initial experimental results. Data distribution and normality were assessed
by Prism 5 software (Graphpad). The data for Fem piRNA inhibitor (Fig. 3c) and
Siwi RNAi (Extended Data Fig. 7a) experiments were subjected to Kruskal–Wallis
analysis with post hoc Dunn’s test. For Masc RNAi (Fig. 4b) experiment, one-way
analyses of variance (ANOVA) was performed with post hoc Tukey’s test. The
data for BmAgo3 RNAi (Extended Data Fig. 7b, d) experiments were subjected to
Mann–Whitney test.
22. Sakai, H., Yokoyama, T., Abe, H., Fujii, T. & Suzuki, M. G. Appearance of
differentiated cells derived from polar body nuclei in the silkworm, Bombyx mori.
Front. Physiol. 4, 235 (2013).
23. Sato, Y. et al. Integrated molecular analysis of clear-cell renal cell carcinoma.
Nature Genet. 45, 860–867 (2013).
24. Sakudoh, T. et al. Diversity in copy number and structure of a silkworm
morphogeneticgene as a result of domestication.Genetics 187, 965–976 ( 2011).
25. Suzuki, M. G. et al. Establishment of a novel in vivo sex-specific splicing assay
system toidentify a trans-actingfactor that negativelyregulates splicingof Bombyx
mori dsx female exons. Mol. Cell. Biol. 28, 333–343 (2008).
26. Wang, L. et al. Mutation of a novel ABC transporter gene is responsible for the
failure to incorporate uric acid in the epidermis of ok mutants of the silkworm,
Bombyx mori. Insect Biochem. Mol. Biol. 43, 562–571 (2013).
27. Yamaguchi, J., Mizoguchi, T. & Fujiwara, H. siRNAs induce efficient RNAi response
in Bombyx mori embryos. PLoS ONE 6, e25469 (2011).
28. Hutva
´gner, G., Simard, M. J., Mello, C. C. & Zamore, P. D. Sequence-specific
inhibition of small RNA function. PLoS Biol. 2, e98 (2004).
29. Shoji, K. et al. Characterizationof a novel chromodomain-containing genefrom the
silkworm, Bombyx mori. Gene 527, 649–654 (2013).
30. Katsuma, S. et al. Novel macula-like virus identified in Bombyx mori cultured cells.
J. Virol. 79, 5577–5584 (2005).
31. Grabherr, M. G. et al. Full-length transcriptome assembly from RNA-Seq data
without a reference genome. Nature Biotechnol. 29, 644–652 (2011).
32. Li, B. & Dewey, C. N. RSEM: accurate transcript quantification from RNA-Seq data
with or without a reference genome. BMC Bioinformatics 12, 323 (2011).
33. Anders, S. & Huber, W. Differential expression analysis for sequence count data.
Genome Biol. 11, R106 (2010).
34. Langmead,B., Trapnell, C., Pop, M.& Salzberg, S. L. Ultrafastand memory-efficient
alignment of short DNA sequences to the human genome. Genome Biol. 10, R25
(2009).
35. Abe, H. et al. Identificationof the female-determining regionof the W chromosome
in Bombyx mori. Genetica 133, 269–282 (2008).
36. Fujii, T. et al. The female killing chromosome of the silkworm, Bombyx mori,was
generated by translocation between the Z and W chromosomes. Genetica 127,
253–265 (2006).
37. Reuter, M. et al. Miwi catalysis is required for piRNA amplification-independent
LINE1 transposon silencing. Nature 480, 264–267 (2011).
RESEARCH LETTER
Macmillan Publishers Limited. All rights reserved
©2014

Extended Data Figure 1
|
Molecular sexing and comparative transcriptome
analysis of embryonic
B. mori
.a, Molecular sexing of individual embryos
at 21 hpo. Musashi,Sasuke and Bonsai are W chromosome RAPD markers.
‘Chr2’ control bands are generated from a primer set that amplifies a sequence
within the 2nd chromosome of B.mori.b, MA plots of RNA-seq data. The
comp73859_c0 contig is indicated by red dots and highlighted by arrows. The
axes show: A(x-axis) 5(log
2
(transcripts per million in male) 1log
2
(transcripts
per million in female))/2. M(y-axis) 5log
2
(transcripts per million
in male) 2log
2
(transcripts per million in female). c, Number of the
comp73859_c0-derived transcripts in each RNA-seq library. Note that the
comp73859_c0-derived transcripts detected in male libraries may be derived
from incorrectly sexed embryos or RNA produced by polar bodies. Combined
with RT–qPCR results of Fig. 1c, the expression level of this contig peaks
around 18–21 hpo in the B.mori embryo.
LETTER RESEARCH
Macmillan Publishers Limited. All rights reserved
©2014

Extended Data Figure 2
|
Expression profile of the female-specific
comp73859_c0 contig. a, Developmental expression profile of the
female-specific contig in ovary during the larval (4th and 5th instars) and pupal
stages. RT–qPCR was performed using total RNA that was isolated from ovary
of 4th and 5th instar larvae, and pupae (p50T). This contig was detected in
the ovary of 4th and 5th instar larvae, and pupae of B.mori with a strong peak
expression at an early pupal stage. rp49 was used as an internal control. Data
shown are mean 1s.d. of three individuals, except for day 0 of 5th instar
(n52). b, The mRNA expression in 17 different tissues from day 3, 5th instar
larvae (p50T). RT–qPCR was performed using total RNA from brain (BR),
prothoracic gland (PG), salivary gland (SG), fat body (FB), trachea (TR),
haemocyte (HC), testis (TES), ovary (OV), anterior silkgland (ASG), middle
silkgland (MSG), posterior silkgland (PSG), foregut (FG), midgut (MG),
hindgut (HG), Malpighian tubules (MT), integument (IG) of male and female
larvae (except for testis and ovary) or BmN4 cells (BmN). rp49 was used as an
internal control. c, Amplification of the female-specific transcript. Long PCR
using female gDNA and cDNA as templates was performed with primers 1F
and 1R. Black arrows show bands corresponding to single or multiple units
of this transcript. The predicted structure of each unit was also indicated.
d, Northern blot analysis of total RNA that was prepared from embryos
(24 hpo) and tissues from day 3 5th instar F1 hybrid Kinshu 3Showa larvae
(ovary, testis, fat body, and silk gland), and BmN cells. The asterisks show
major transcripts.
RESEARCH LETTER
Macmillan Publishers Limited. All rights reserved
©2014

Extended Data Figure 3
|
Characterization of the female-specific piRNA.
a, Detection of contig-derived piRNA and piRNA-1 (control). Northern blot
analysis was performed using total RNA prepared from early embryos.
The asterisks show the location of each piRNA. b, Normalized reads of the
female-specific piRNA in embryonic piRNA libraries of B.mori
12
generated at
0, 6, 12, and 24hpo. Reads of 26–29 nucleotides that showed 2 or fewer
mismatches to the corresponding piRNA sequence were scored as a positive
match. c, RT–qPCR estimation of the female-specific piRNA levels in early
embryos. The piRNA level was normalized to that of let-7.d, Normalized reads
of the female-specific piRNA in piRNA libraries
6
from ovary and testis of
wild-type B.mori or W chromosome mutants. Schematic representation of
sex chromosomes of each strain is shown below the panel. The putative
sex-determining region is represented by the green box. The orange bar
represents the W chromosome derived from B.mandarina. OV, ovary from
wild-type; TES, testis from wild-type; MW, ovary from MW strain; LY, ovary
from LY strain; WF, testis from WF strain.
LETTER RESEARCH
Macmillan Publishers Limited. All rights reserved
©2014

Extended Data Figure 4
|
Effect of inhibition of the piRNA pathway on the
splicing of
Bmdsx
transcripts. a, Abundance of the female-specific piRNA in
three piRNA libraries constructed from three KG individual ovaries (KG12,
KG41 and KG42)
13
. Of these, two (KG41 and KG42) showed a severe
masculinized phenotype, and the rest (KG12) showed a weak phenotype.KG12
expressed a slightly lower amountof this piRNA than that of LY (82.4% of LY),
whereas its expression in the ovary of severe masculinized individuals
(KG41 and KG42) were markedly lower than LY’s (12.1 and 29.7% of LY,
respectively). The abbreviations are the same as in Extended Data Fig. 3d.
b, Effect of the inhibitor RNA on the Bmdsx splicing. BmN4 cells were
transfected with the inhibitor RNA or control RNA (that is, inhibitor for
GFP piRNA), and the splicing patterns of Bmdsx were examined by RT–PCR.
The F and M indicate female- and male-type splicing of Bmdsx, respectively.
Similar results were obtained in three independent experiments. c, Effect of the
RNA inhibitor on the Bmdsx splicing. The Bmdsx splicing patterns were
examined at about 240h post-injection (immediately after hatching). The
abbreviations are the same as in Fig. 2c. The number indicates the sample size.
d, Knockdown of Siwi or BmAgo3 mRNAs in female and male embryos.
The embryos were injected with two types of siRNAs that target Siwi (Siwi-1
and Siwi-2) or BmAgo3 (Ago3-1 and Ago3-2) or a control siRNA that targets
GFP. Total RNA was isolated from female or male siRNA-injected embryos
at 72 h post-injection and RT–qPCR was performed. The data shown are
mean 1s.d. The number above each bar indicates the sample size of each
group. e, Representative patterns of the Bmdsx splicing in siRNA-injected
embryos. The F and M indicate female- and male-type splicing of Bmdsx,
respectively.
RESEARCH LETTER
Macmillan Publishers Limited. All rights reserved
©2014

Extended Data Figure 5
|
Characterization of
Masc
.a, Structure of Masc
mRNA. Five Masc transcripts (A–E) that encode full-length Masc proteins but
show unique splicing patterns in the 39-untranslated region as well as one
transcript (F) that encodes a truncated Masc protein are found. b, Domain
structure of the Masc protein. The hexagons show the location of two
CCCH-type zinc finger domains. The amino acid sequences of these domains
are shown below. The conserved CCCH residues are shown in red.
c, Phylogenetic analysis of Masc proteins. The neighbour-joining tree was
generated using the amino acid sequencesof zinc finger domains from proteins
showing homology to Masc. The numbers on the internal branches represent
the support value in the bootstraps of 1,000 replicates.
LETTER RESEARCH
Macmillan Publishers Limited. All rights reserved
©2014

Extended Data Figure 6
|
Cleavage of
Masc
mRNA. a, Identification of the
cleavage site of Masc mRNA. The Masc mRNA-derived RNA fragments were
amplified by a modified RACE method, cloned, and sequenced. The RACE
adaptor and the cloned 59-end are indicated. Thirteen 59-ends were
determined and allshowed identical sequences. Nucleotidesidentical to the top
sequence are represented by asterisks. b, Detection of Masc piRNA. Northern
blot analysis was performed using total RNA prepared from early embryos.
The asterisk shows the location of Masc piRNA. c, Mapping of embryonic
piRNAs (24 hpo) onto Masc mRNA. The relative location of ORF of Masc is
shown below. d, Normalized reads of Fem piRNA and Masc piRNA in Siwi- or
BmAgo3-immunoprecipitated libraries from BmN4 cells
14
.e, A ping-pong
amplification model of Fem piRNA/Masc piRNA. f, Normalized reads of Masc
piRNA in embryonic piRNA libraries. Reads of 26–29 nucleotides that showed
2 or fewer mismatches to the Masc piRNA sequence were scored as positive.
g, RT–qPCR estimation of Masc piRNA in early embryos. The Masc piRNA
level was normalized to that of let-7.
RESEARCH LETTER
Macmillan Publishers Limited. All rights reserved
©2014

Extended Data Figure 7
|
Effects of
Siwi
or
BmAgo3
knockdown on the
Bmdsx
splicing and
Masc
expression. a,Masc expression in female embryos
injected with two types of siRNAs that target Siwi (Siwi-1 and Siwi-2) or a
control siRNA that targets GFP. Total RNA was isolated from female
siRNA-injected embryos at 18 hpo and RT–qPCR was performed. The data
shown are mean 1s.d. The number at the baseof each bar indicates the sample
size of each group. Data were subjected to Kruskal–Wallis analysis with post
hoc Dunn’s test. *P,0.05. The expression levels of Siwi mRNA decreased to
23 and 44% after injecting Siwi-1 and Siwi-2 siRNAs, respectively, compared
with that in GFP-siRNA-injected embryos. b, Knockdown of BmAgo3 mRNA
in newly hatched larvae. The embryos were injected with BmAgo3 or GFP
(control) siRNA. Total RNA was isolated from newly hatched larvae (at about
240 h post-injection) and RT–qPCR was performed. The data shown are mean
1s.d. The number indicates the sample size of each group. *P,0.05,
one-sided Mann–Whitney test. c, Splicing of Bmdsx in newly hatched larvae
that were injected with BmAgo3 siRNA. The Bmdsx splicing patterns were
examined at about 240h post-injection. The number indicates the sample size.
The abbreviations are the same as in Fig. 2c. d,Masc expression in newly
hatched larvae that were injected with BmAgo3 siRNA. Total RNA was isolated
from siRNA-injected newly hatched larvae (at about 240h post-injection)
and RT–qPCR was performed. The data shown are mean 1s.d. The
number indicates the sample size of each group. *P,0.05, one-sided
Mann–Whitney test.
LETTER RESEARCH
Macmillan Publishers Limited. All rights reserved
©2014

Extended Data Figure 8
|
Splicing of
Bmdsx
in
Masc
siRNA-injected
embryos. a,b, The Bmdsx splicing pattern was determined at 144 h (a) and
216 h (b) post-injection. The abbreviations are the same as in Fig. 2c. The
number indicates the sample size.
RESEARCH LETTER
Macmillan Publishers Limited. All rights reserved
©2014

Extended Data Figure 9
|
Functional analysis of the
Fem
piRNA-resistant
Masc
transcript. a, Sequence of the Fem piRNA-resistant Masc (Masc-R)
mRNA. Five nucleotide mutations that do not result in amino acid
substitutions for the Masc protein are shown by red letters. The putative
cleavage site by the Fem piRNA–Siwi complex is shown by the red line.
b, RT–qPCR of Masc mRNA in cDNA-transfected BmN4 cells. BmN4 cells
were transfected with Masc expression vectors or control vector. The Masc
mRNA level was normalized to that of rp49. Data shown are means of
duplicates. c, RT–qPCR of Masc piRNA in BmN4 cells transfected with Masc
expression vectors or control vector. The Masc piRNA level was normalized to
that of let-7. Data shown are means of duplicates. d, Identification of the
cleavage site of exogenously introduced Masc. BmN4 cells were transfected
with Masc expression vectors or control vector. Three days after transfection,
zeocin (final concentration, 500 mgml
21
) was added to the medium. Six days
after drug selection, the Masc mRNA-derived RNA fragment (shown by the
red asterisk) expressed from the transfected plasmids was amplified by a
modified RACE method. The fragment was cloned, sequenced, and identified
as the Masc mRNA-derived one. The locations of the primers are shown by
arrows. e, Effect of Masc transfection on the Bmdsx splicing in BmN4 cells.
The splicing patterns of Bmdsx in stably transfected BmN4 cells (six days after
drug selection) were examined by RT–PCR. The F and M indicate female- and
male-type splicing of Bmdsx, respectively. Similar results were obtained in
two independent experiments. f, Light microscopic observations of BmN4 cells
stably transfected with Masc expression vectors or control vector (2 weeks after
drug selection).
LETTER RESEARCH
Macmillan Publishers Limited. All rights reserved
©2014

Extended Data Figure 10
|
Distribution of
Masc
-regulated genes
throughout the silkworm genome. a,b, The genome loci where
Masc-regulated genes are located were identified using RNA-seq data from
male (a) and female (b) embryos injected with control (siGFP) and Masc
(siMasc) siRNAs (72 h post-injection).
RESEARCH LETTER
Macmillan Publishers Limited. All rights reserved
©2014