1

10

100

1000

gypsy

fold

cha

nge

in s

tead

y st

ate

RN

A le

vels

GFP arm

i

mag

o

tsu

Acn

RnpS

1

btz

siRNA mediated depletion against indicated genes in OSCs

Hayashi_FigS1

Supplementary Figure S1. The EJC is required for transposon silencing in OSCs.Fold changes in steady-state RNA levels of gypsy in OSCs treated with siRNAs against indicated genes, showing that the depletion of EJC factors except for btz leads to an enhanced expression of gypsy. RNA levels are normalized to rp49 levels, and values indicate averages of three biologi-cal replicates relative to control knockdowns (error bars: stdev.).

A

C

BAc

n cl

onal

KD

fs(1

)Yb

clon

al K

Dts

u cl

onal

KD

DNA GFP Armi

0

0.5

1.0

1.5

A

A B C D E F

amplicons

BCDEF

fold

cha

nge

in s

tead

y st

ate

flam

enco

RN

A le

vels

GFP tsu Acn

siRNA mediated depletion against indicated genes in OSCs

50 kB

400

-100

GFP knockdown

400

-100

tsu knockdown

400

-100

Acn knockdown

flamenco qRT-PCR

RN

A-se

q re

ads

per m

illion

Armi

Hayashi_FigS2

Supplementary Figure S2. Early steps in piRNA biogenesis are not affected in EJC-depleted somatic cells.(A) Confocal sections (scale bars = 10μm) of egg-chambers with clones of somatic follicle cells (marked by GFP expression) expressing shRNAs against EJC factors stained for Armi (red) and DNA (blue). Yb bodies (piRNA processing sites marked by cytoplasmic Armi foci) are abolished in Fs(1)Yb-depleted cells (control), but remain intact upon knockdowns of Tsu or Acn.(B and C) Changes in steady-state levels of flamenco in OSCs treated with siRNAs against indicat-ed genes, showing that the depletion of Tsu or Acn lead only to mild reductions of flamenco levels. RNA levels were measured by qRT-PCR (B; normalized to rp49 levels; values indicate averages of five biological replicates relative to control knockdowns; error bars: stdev.) or by RNA-seq (C; normalized to one million genome-mappable reads). RNA-seq profiles at the flamenco locus were obtained from rRNA-depleted total RNA from OSCs treated with siRNAs against indicated genes. LINE transposons (red) and LTR transposons (black) are indicated. The RNA-seq tracks of Tsu- and Acn-depleted samples (red) are overlayed to the tracks of the GFP-depleted control sample (gray).

1/10

1/2

1

2

10

fold

cha

nge

in s

tead

y st

ate

piw

i RN

A le

vels

tsu

Acn

white

germline specific knockdown againstindicated gene

exon 3 (amplicon D)intron 4 spliced (amplicon B)intron 4 unspliced (amplicon C)

Hayashi_FigS3

Supplementary Figure S3. The EJC is required for splicing of piwi int4 in ovaries.The fold changes in steady-state RNA levels of all piwi transcripts (exon3, black bars) and those piwi transcripts with spliced (grey bars) or unspliced (red bars) int4, showing the decrease of spliced int4 and the increase of unspliced int4 in Tsu- or Acn-depleted ovaries. Total RNA isolated from ovaries depleted for indicated genes in the germline was used for qRT-PCR. Values were normalized to rp49 levels and averages of three biological replicates relative to control knock-downs are shown with error bars indicating Stdev.

GFP-piwi [wildtype]

whi

te S

KDar

mi S

KDts

u SK

DAc

n SK

D

GFP-piwi [Δ int4]100.0 ± 18.1 % 100.0 ± 12.1 %

N.D.

N.D.

N.D.

103.4 ± 30.0 %

333 ± 23 % 106.2 ± 14.2 %

mdg1

1

0.1

10

100

white armi tsu

*

*

fold

cha

nge

in s

tead

y sy

ate

RN

A le

vel

+ none+ GFP-piwi [wt]+ GFP-piwi [Δint4]

soma specific knockdown against indicated gene

A

E

int4

GLKD GFP GLKD tsu

GLKD btzGLKD Acn GLKD RnpS1

GLKD mago

B

C

anti-Ago3 staining

ago3 RT-PCR D ago3 qRT-PCR (exon 5)

white tsu

Acn

1.0 kb

2.0 kb

3.0 kb

0

0.5

1.0

1.5

fold

cha

nge

in s

tead

y st

ate

ago

3 R

NA

leve

ls

tsu

Acn

white

germline specific knockdownagainst indicated gene

germline specific knockdownagainst indicated gene

Hayashi_FigS4

Supplementary Figure S4. A piwi transgene lacking int4 rescues TE de-repression in Tsu-de-pleted somatic follicle cells.(A) The depletion of tsu or Acn affects the expression of wildtype GFP-piwi, but not of GFP-piwi Δint4 in follicle cells. Shown are confocal sections (scale bars = 10μm) through egg chambers expressing wildtype GFP-piwi or GFP-piwi that lacks int4 (Δint4) depleted for indicated genes in somatic follicle cells. All images report relative GFP levels to white SKD from each GFP-piwi trans-gene. Indicated fluorescence levels (± Stdev.) were calculated from GFP intensities in nurse cell nuclei normalized to those in follicle cell nuclei from three egg chambers. N.D.: not determined.(B) Confocal sections (scale bars = 10μm) through stage 3-5 egg-chambers expressing shRNAs against EJC factors stained for AGO3 proteins, showing that AGO3 protein levels are decreased upon knockdown of mago or tsu.(C) Agarose/EtBr gel showing AGO3 RT-PCR products amplified from poly-A selected RNA isolated from ovaries depleted for indicated genes. Primers mapping to the first and the last exon were used. No mis-spliced species were detected in any knockdown.(D) Shown are the fold changes in steady-state RNA levels of AGO3 transcripts (exon5) amplified from poly-A selected RNA isolated from ovaries depleted for indicated genes. Values were normal-ized to rp49 levels and averages of four biological replicates relative to control knockdowns are shown with error bars indicating Stdev. AGO3 transcript levels are decreased in Tsu-, but not in Acn-depleted cells (*: p < 0.01)(E) Shown are fold changes in steady state RNA levels of mdg1 in ovaries expressing indicated GFP-piwi transgenes and depleted for white, armi or tsu in the soma. The GFP-piwi construct lacking int4 rescues the de-repression of mdg1 upon depletion of tsu (*: p < 0.01). RNA levels were normalized to rp49 levels and averages of three biological replicates relative to control knockdowns are shown (error bars indicate Stdev.).

light

siRNA mediated depletion against indicated gene in OSCs

GFP arm

i

mag

o

tsu

Acn

RnpS

1

btz

} products with skipped exons

full length (2143 bp)

1.0 kb

2.0 kb

Hayashi_FigS5

Supplementary Figure S5. Depletion of core EJC factors but not Acn results in exon-skip-ping of light transcripts.Agarose/EtBr gel image showing RT-PCR products of light transcripts amplified from total RNA from OSCs depleted for indicated genes using the primers mapping to the first and the last exons. The depletion of mago, tsu and to a lesser extent RnpS1 leads to exon-skipping of light transcripts whereas the depletion of Acn does not affect the splicing.

Supplementary Table S2. (Illumina sequencing datasets used in this study)

Origin Method Genotype GEO

OSC RNA-seq (Poly-A+) si-GFP KD (replicate #1) GSE58830

RNA-seq (Poly-A+) si-GFP KD (replicate #2) GSE58830

RNA-seq (Poly-A+) si-tsunagi KD (replicate #1) GSE58830

RNA-seq (Poly-A+) si-tsunagi KD (replicate #2) GSE58830

RNA-seq (Poly-A+) si-Acinus KD (replicate #1) GSE58830

RNA-seq (Poly-A+) si-Acinus KD (replicate #2) GSE58830

RNA-seq (ΔrRNA) si-GFP KD GSE58830

RNA-seq (ΔrRNA) si-tsunagi KD GSE58830

RNA-seq (ΔrRNA) si-Acinus KD GSE58830

Supplementary Table S3. (Oligos used for cloning of shRNA constructs)

Gene Sequence

tsunagi fw:  ctagcagtCTGCGATTACGGAGAAATCAAtagttatattcaagcataTTGATTTCTCCGTAATCGCAGgcg  

rv:  aattcgcCTGCGATTACGGAGAAATCAAtatgcttgaatataactaTTGATTTCTCCGTAATCGCAGactg  

eIF4AIII fw:  ctagcagtAAGGGTTTCAAGGAACAGATAtagttatattcaagcataTATCTGTTCCTTGAAACCCTTgcg  

rv:  aattcgcAAGGGTTTCAAGGAACAGATAtatgcttgaatataactaTATCTGTTCCTTGAAACCCTTActg  

barentsz fw:  ctagcagtTGGCATGGAATTTAAGAAGAAtagttatattcaagcataTTCTTCTTAAATTCCATGCCAgcg  

rv:  aattcgcTGGCATGGAATTTAAGAAGAAtatgcttgaatataactaTTCTTCTTAAATTCCATGCCAactg  

Others  

GFP (Handler et al. 2013)  

white (Handler et al. 2013)  

armitage (Handler et al. 2013)  

mago nashi (Handler et al. 2013)  

Acinus (Handler et al. 2013)  

RnpS1 (Handler et al. 2013)  

Supplementary Table S4. (siRNAs used for RNAi in OSCs)

Gene sequence

mago nashi Guide: UCCUCCUGCAUGAUCUCCGUU

Passenger: CGGAGAUCAUGCAGGAGGAUU

tsunagi Guide: UUGAUUUCUCCGUAAUCGCUU

Passenger: GCGAUUACGGAGAAAUCAAUU

Acinus Guide: UUGUUAUCGCGGUCUUUGCUU

Passenger: GCAAAGACCGCGAUAACAAUU

RnpS1 Guide: UUAACAUUAUUGCUAUCGGUU

Passenger: CCGAUAGCAAUAAUGUUAAUU

barentsz Guide: UUCUUCUUAAAUUCCAUGCUU

Passenger: GCAUGGAAUUUAAGAAGAAUU

Others

GFP (Handler et al. 2013)

armitage (Handler et al. 2013)

Supplementary Table S5. (Primers for QPCR analysis)

Gene Sequence

flamenco (amplicon A) fw: GAAGTCTTGGGACACTCATAGGT

rv: CCAGAAAATTAAGCGGAAGC

flamenco (amplicon B) fw: TGCAATTCCCCAAATCTCGATCC

rv: GGACACATGGAAGCTTCGAAGAA

flamenco (amplicon C) fw: TGTAGTCTTTGCAGTGTCAGTGT

rv: CCTGCATAAACGGATCGGTGATA

flamenco (amplicon D) fw: TTGCCTCAGTGAAACGCCTAAAA

rv: TTCCCTTATTGAACATCACCGCC

flamenco (amplicon E) fw: GTCAAGTGTCTTTTGCTGTGCTG

rv: CCCACCCTTGCATAGTCTCTCTA

flamenco (amplicon F) fw: AATGTTTCGGGTTCTAGGGTAGC

rv: TCAGCAACGAACATGATTACCTAGT

piwi intron4_spliced fw: AATTCCTGAGCTCTGCCGAGTG

rv: TAACTGCTCATGGCACGCATAA

piwi intron4_unspliced fw: AAACGAAGAAAAATCTGAGCACT

rv: TAACTGCTCATGGCACGCATAA

piwi exon3 fw: TTGGGCACCGAAATAACTCA

rv: TTACCCGTACTTCGTCCTGATG

ago3 exon5 fw: CCAACAAGATAAGTGGGCGTCA

rv: TTGGTCACAGCATGTTTGGAGA

cg8671 intron1_spliced fw: CAAACGAAAGCAAAGGCAGTTG

rv: ATATAGCGGAGACACATCCGAA

cg8671 intron1_unspliced fw: TGTTCCGTGCTGTTGTGTTTGT

rv: ATATAGCGGAGACACATCCGAA

garnet intron7_spliced fw: CCTTAGAACCCCGGCTAGGAAA

rv: CAGAGCTGAATGGAGGCACTGT

garnet intron7_unspliced fw: CGAACCAAAAACCAAACCCAAC

rv: CAGAGCTGAATGGAGGCACTGT

Others

rp49 (Handler et al. 2013)

HeT-A (Handler et al. 2013)

blood (Handler et al. 2013)

mdg1 (Handler et al. 2013)

gypsy (Muerdter et al. 2013)

Supplementary Table S6. (Primers for RT-PCR analysis)

Gene Sequence

piwi fw: CAGGCGTCCACTTAACGAAG

rv: GTGTGCTCTTCGCAATGTCA

gasz fw: AGCAACCTTTGCAAATACGG

rv: CTTTTTGCGTTTTCGCTCTG

shutdown fw: TGGAAGAAAACTTCGAACCGTA

rv: CCAAGTTGTACTCGCCCAGA

ago3 fw: TTAGTTTTCTGTTTTTGTGGTA

rv: CATCTCGTTGAATACTTTGTCC

rolled (Ashton-Beaucage et al. 2010)

light (Ashton-Beaucage et al. 2010)

REFERENCES

Ashton-Beaucage D, Udell CM, Lavoie H, Baril C, Lefrancois M, Chagnon P, Gendron P, Caron-Lizotte O, Bonneil E, Thibault P et al. 2010. The exon junction complex controls the splicing of MAPK and other long intron-containing transcripts in Drosophila. Cell 143: 251-262.

Handler D, Meixner K, Pizka M, Lauss K, Schmied C, Gruber FS, Brennecke J. 2013. The Genetic Makeup of the Drosophila piRNA Pathway. Mol Cell 50: 762-777.

Muerdter F, Guzzardo PM, Gillis J, Luo Y, Yu Y, Chen C, Fekete R, Hannon GJ. 2013. A Genome-wide RNAi Screen Draws a Genetic Framework for Transposon Control and Primary piRNA Biogenesis in Drosophila. Mol Cell 50: 736-748.