© 2017. Published by The Company of Biologists Ltd.

Tex19 paralogs are new members of the piRNA pathway controlling retrotransposon

suppression

Yara Tarabay1,5,6, Mayada Achour1,6, Marius Teletin1,2, Tao Ye1, Aurélie Teissandier3,

Manuel Mark1,2, Déborah Bourc’his3, Stéphane Viville1,4,*°

1Institut de Génétique et de Biologie Moléculaire et Cellulaire (IGBMC), Institut National de

Santé et de Recherche Médicale (INSERM) U964/ Centre National de Recherche Scientifique

(CNRS) UMR 1704/Université de Strasbourg, 67404 Illkirch, France 2Service de Biologie de la Reproduction, Centre Hospitalier Universitaire, 67000 Strasbourg,

France 3Institut Curie, department of Genetics and Developmental Biology, CNRS UMR3215, INSERM

U934, Paris, France 4Centre Hospitalier Universitaire, 67000 Strasbourg, France 5present address: Institut de génétique humaine (IGH), 141 rue de la Cardonille, 34396

Montpellier, France 6Co-first authors *To whom correspondence should be addressed

° present addresses: Laboratoire de diagnostic génetique, UF3472 – Infertilité, Nouvel Hôpital

Civil, 1 place de l’Hôpital, 67091 Strasbourg cedex and IPPTS, 3, rue Koeberlé, 67000

Strasbourg

Contact information: stephane.viville@unistra.fr

Tel : +33 3 68 85 37 00

Fax : + 33 3 68 85 38 09

Jo

urna

l of C

ell S

cien

ce •

Adv

ance

art

icle

JCS Advance Online Article. Posted on 2 March 2017

Summary statement

Through in vitro and in vivo experiments, the present study demonstrates that the two TEX19

paralogs belong to the family of piRNA-interacting proteins.

Abstract

Tex19 genes are mammalian specific and duplicated in Tex19.1 and Tex19.2 in some species,

such as the mouse and rat. It has been demonstrated that mutant Tex19.1 males display a

variable degree of infertility whereas they all upregulate MMERVK10C transposons in their

germ line. In order to study the function of both paralogs in the mouse, we generated and

studied double knockout (Tex19DKO) mutant mice. Adult Tex19DKO males exhibited a fully

penetrant phenotype, similar to the most severe phenotype observed in single Tex19.1KO

mice, with small testes and impaired spermatogenesis, defects in meiotic chromosome

synapsis, persistence of DNA double-strand breaks during meiosis, lack of post-meiotic germ

cells and upregulation of MMERVK10C expression. The phenotypic similarities with Piwi

KO mice prompted us to check and then demonstrate, by immunoprecipitation and GST

pulldown followed by mass spectrometry analyses, that TEX19 paralogs interact with PIWI

proteins and their VPTEL domain directly binds piRNAs in adult testes. We therefore

identified two new members of the postnatal piRNA pathway.

Key words:

Meiosis, piRNA pathway, Infertility, Transposable elements

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

Introduction

In all sexually reproducing species, germ cells are the cell lineage that is specialized in

transferring genetic information from one generation to the next (Hajkova, 2011). The

genome of germ cells needs to be maintained stably and protected in all individuals

throughout development. In particular, the genome must be protected against mobile

transposable elements (TEs). These TEs can move, amplify themselves and transpose into

new genomic loci prompting a serious threat to the genome. Although most TEs are inert,

some elements have retained activity and become typically expressed in the germ line where

they can transpose into new copies and pass onto the next generation. TEs account for almost

half of the human and mouse genomes. They are classified into two categories, class I, which

represents the majority of the TE families and class II, which accounts for approximately 3%

of the mammalian genome (Zamudio and Bourc'his, 2010, Goodier and Kazazian, 2008).

Class I TEs are believed to be ancient traces of viral infections; they transpose themselves

using a “copy and paste” mode involving an RNA intermediate, which is converted into

cDNA before its integration. Class II TEs are DNA transposons, which transpose using a “cut

and paste” mode without any transcription needed (Goodier and Kazazian, 2008, Lander et

al., 2001, Zamudio and Bourc'his, 2010).

Despite an undeniable role in evolution (Furano et al., 2004, Ostertag and Kazazian, 2001,

Zamudio and Bourc'his, 2010), TEs can be responsible for drastic mutagenic effects

illustrated by the growing number of human diseases caused by their transposition and

expression (Cowley and Oakey, 2013). Germ cells have developed sophisticated strategies to

control TE activities at the transcriptional level by DNA methylation and the post-

transcriptional level by RNA degradation or by blocking their integration for example (Gasior

et al., 2006, Obbard et al., 2009, Rollins et al., 2006, Yoder et al., 1997, Zamudio and

Bourc'his, 2010). For reasons that remain unclear, the control of TEs in the mammalian germ

line is gender-specific. Indeed, it has been shown that the protection of male germ line

genomes relies on the PIWI family of ARGONAUTE proteins, which produce and bind small

RNAs around 24-30 nucleotides long (named piRNAs) (Ghildiyal and Zamore, 2009). In

mice, the PIWI clade contains three members, namely MIWI (PIWIL1), MILI (PIWIL2) and

MIWI2 (PIWIL4) (Kuramochi-Miyagawa et al., 2004), which interact with germ line-specific

members of the TUDOR family proteins such as TDRD1, TDRD5, TDRD6, TDRD7 and

TDRD9 (Chuma et al., 2003, Hosokawa et al., 2007, Vasileva et al., 2009, Zheng and Wang,

2012). They also interact with other factors such as MOV10L1, MVH (DDX4), MAEL or

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

GASZ (Costa et al., 2006, Ma et al., 2009, Soper et al., 2008, Zheng et al., 2010, Kotaja et al.,

2006, Kuramochi-Miyagawa et al., 2004). Mutation of these piRNA pathway members give

rise to a similar phenotype consisting of male infertility due to meiosis defects that can take

place at different stages of meiosis and lead to an up-regulation of most TEs (Zamudio and

Bourc'his, 2010). Despite a large number of studies, particularly involving loss-of-function

mice, the exact mechanism whereby piRNAs control TE expression and the link between the

loss of TE control and meiosis defects are far from fully understood. Currently, only the male

process is beginning to be deciphered while the female one remains to be discovered.

In 2001, Wang et al described a set of testis expressed (Tex) genes, including Tex19 (Wang et

al., 2001). In a previous study, we showed that Tex19 is a mammal-specific gene, which is

duplicated in the mouse genome into two paralogs, Tex19.1 and Tex19.2. By multiple

sequence alignment of mammalian TEX19 proteins, two highly conserved domains, named

MCP and VPTEL respectively, were identified. However, none of them shares homology

with known proteins, preventing functional prediction (Kuntz et al., 2008). Both genes are co-

expressed in the ectoderm, then in primordial germ cells (PGCs). Later on, they are co-

expressed from embryonic day 13.5 until adulthood in testes. However, only Tex19.1

transcripts are present in developing and adult ovaries as well as in the placenta (Celebi et al.,

2012). Tex19.1 Knockout (KO) mice show a variable degree of male infertility due to

impaired meiosis associated with an up-regulation of MMERVK10C retrotransposon

expression (Ollinger et al., 2008, Tarabay et al., 2013). More recently, it was shown that

Tex19.1 plays an important role in placenta function, as Tex19.1KO mouse embryos exhibit

intra-uterine growth retardation and have small placentas due to reduced number of

spongiotrophoblast, glycogen trophoblast and sinusoidal trophoblast giant cells (Reichmann et

al., 2013, Tarabay et al., 2013). Tex19.2KO mice are fertile presenting only a subtle

phenotype with discrete seminiferous tubule degeneration in adult male testes (unpublished

observations Tarabay et al). The co-expression of Tex19.1 and Tex19.2 in PGCs, gonocytes

and spermatocytes suggests that these two genes could play redundant functions during

spermatogenesis.

By generating a double knockout of both Tex19.1 and Tex19.2 genes (Tex19DKO), we

demonstrate here that TEX19 paralogs exhibit redundant functions, which are essential for

male fertility. We show that the Tex19DKO phenotype is fully penetrant, mimicking the most

severe phenotype observed in single Tex19.1KO mice. All adult DKO males are infertile and

display impaired spermatogenesis : meiosis is blocked at the pachytene stage- during which

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

chromosome synapsis is not correctly formed or incomplete- and this leads to testis

degeneration. Furthermore, MMERVK10C retrotransposon expression is up-regulated during

meiosis in Tex19DKO testes. Through GST pull-down and immunoprecipitation experiments,

we found that TEX19 paralogs interact with PIWI proteins and their accessory partners; they

have moreover the ability to bind RNAs of around 30 nucleotides, a diagnostic size for

piRNAs, but are not required for their biogenesis. Detailed analysis further showed that

TEX19 directly interacts with piRNAs via its VPTEL domain. Altogether, our study provides

novel evidence that TEX19 proteins are new members of the PIWI/piRNA pathway acting

during post-natal spermatogenesis.

Results

Tex19 paralogs play an essential role in male fertility

Tex19.1KO mice show a variable level of male infertility. : one third of the KO male mice are

almost indistinguishable from wild type (WT) mice, one third display a mild phenotype and

the last third a severe phenotype with a complete meiosis arrest at pachytene (Ollinger et al.,

2008, Tarabay et al., 2013). Independently of the severity of the phenotype, a two-fold up-

regulation of MMERVK10C elements was consistently observed among Tex19.1KO animals

(Ollinger et al., 2008, Tarabay et al., 2013). In contrast, Tex19.2KO mice are fertile and their

testes are undistinguishable from controls. However, some seminiferous tubules show

epithelium vacuolization (unpublished observations Tarabay et al). Considering the

overlapping profile of Tex19.1 and Tex19.2 expression (Celebi et al., 2012), we wondered

whether there was any functional redundancy between the two genes in testes. We therefore

generated conditional double KO mice for Tex19.1 and Tex19.2 genes, which exist as tandem

duplication (Fig 1A).

The KO of both alleles (so called Tex19DKO) was confirmed by PCR (Fig 1B) and the

absence of the encoded proteins by western blot using testes extracts from double wild-type

(DWT), Tex19.1KO, Tex19.2KO or Tex19DKO samples (Fig 1C and data not shown).

Interbreeding of Tex19 double heterozygote (DHZ) animals resulted in a statistically

significant deviation of the frequency of double homozygous animals from the expected

Mendelian 1:2:1 ratio (207 (29.3%) Tex19DWT mice, 440 (62.3%) Tex19 double

heterozygote (DHZ) mice and 59 (8.3%) Tex19DKO mice, p<0.05, Student t-test). Yang et

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

al., 2010 previously described a more important female lethality among Tex19.1KO animals,

an observation which was not shared by other independent studies including ours (Ollinger et

al., 2008, Tarabay et al., 2013, Yang et al., 2010). Nonetheless, we found that the number of

Tex19DKO females was significantly reduced when compared to Tex19DKO males (32.2% of

females versus 67.8% males, P<0.05), suggesting a gender-specific lethality. Surviving

Tex19DKO mice did not present gross somatic abnormalities. Females displayed normal

fertility, but males were sterile.

Tex19DKO males displayed a constant and severe reduction in testis size (Fig 2A). The

weight of Tex19DKO adult testes (ranking from 19mg up to 35mg, mean 24,4mg ±10.5mg

per testis) was three-fold less than that of WT testes (ranking from 109mg up to 135mg, mean

113.5mg ±5.52mg per testis) (Fig 2B). In contrast to WT seminiferous tubules (Fig 2C),

Tex19DKO testes histology showed degenerate seminiferous epithelium with tubules

containing only early spermatogenic cells, but a complete lack of post-meiotic germ cells,

suggesting a meiotic arrest (Fig 2D); consequently, the epididymis was free of spermatozoa

(Fig S1A, B). Adult Tex19DKO seminiferous tubules exhibited a drastically increased level of

apoptosis when compared to WT (Fig S1C). This phenotype is comparable to the most severe

phenotype observed in the single Tex19.1KO (Tarabay et al., 2013). Thus, Tex19 paralogs are

required for male meiosis and are essential for male fertility.

To further investigate the meiotic defect of Tex19DKO mutants, we tested the assembly of the

synaptonemal complex by co-immunostaining of the central and lateral elements with

antibodies against SYCP1 and SYCP3, respectively. In contrast to WT spermatocytes where

all autosomes were fully synapsed at pachytene (Fig 2E), synapsis failed to occur properly in

Tex19DKO, as evidenced by the complete or partial absence of the SYCP1 signal (Fig 2F).

Furthermore, as progression of homologous recombination and chromosome synapsis are

interdependent, we checked whether homologous recombination was also affected by

studying DNA double strand break (DSB) distribution, by immunostaining against

phosphorylated histone γH2AX. γH2AX staining is normally present throughout chromatin

during the zygotene stage. As synapsis proceeds, the DSBs are resolved, resulting in γH2AX

disappearance from autosomes, but not from the sex chromosomes. In WT pachytene cells,

synapsis was complete and only the sex chromosomes were stained for γH2AX (Fig 2G).

However, the incompletely synapsed pachytene chromosomes in Tex19DKO nuclear spreads

exhibited strong diffuse γH2AX staining, suggesting that the disruption of Tex19 paralogs

caused meiotic arrest at the pachytene stage because of the persistence of DSBs (Fig 2H).

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

Alternatively, γH2AX retention in spermatocytes from Tex19DKO may be a consequence of

asynapsis.

In order to date the developmental onset of the Tex19DKO phenotype, we analyzed testis

sections at post-natal (p) day 10, 16, 20 and at 16 weeks of age (Fig 3). At p10, Tex19DKO

were histologically indistinguishable from WT and Tex19DHZ testes (Fig 3A-C). Tex19DKO

showed vacuolization and seminiferous epithelium degeneration as soon as p16 (Fig 3D-F).

Thus, the Tex19DKO testes phenotype became histologically visible between p10 and p16,

corresponding to the timing when the first spermatocytes reach the pachytene stage during the

first post-natal wave of spermatogenesis. This confirmed a spermatogenesis arrest occurring

between the zygotene and the early pachytene stages of meiosis prophase I.

Finally, increased TE expression has been proposed to be responsible for impaired

chromosome synapsis and meiotic defects during spermatogenesis in Dnmt3L, Miwi2 and Mili

mutant mice (Aravin et al., 2007, Bestor and Bourc'his, 2004, Carmell et al., 2007,

Kuramochi-Miyagawa et al., 2004, Zamudio et al., 2015). Overexpression of MMERVK10C

retrotransposons could similarly be responsible for the meiotic defects seen in Tex19.1KO

mice (Ollinger et al., 2008, Tarabay et al., 2013). In Tex19.DKO mice, we observed a two-

fold and four-fold up-regulation of MMERVK10C expression at p10 and p16, respectively;

no upregulation was observed at other TEs (LINEs and IAP) (Fig S2), confirming the specific

impact of Tex19 deficiency in the control of MMERVK10C.

TEX19 associate with PIWI proteins and pachytene piRNAs

To gain insight into TEX19 function, we went on seeking its protein partners through GST

pull-down and immunoprecipitation (IP) experiments. For the GST-pull-down experiment,

GST alone, GST-TEX19.1 or GST-Tex19.2 was incubated with total testicular protein extract

of adult mice (Fig S3); complexes were then analyzed by mass spectrometry (MS). In GST-

TEX19.1 and GST-TEX19.2 samples, we detected bands corresponding to MAEL, MIWI,

MILI, TDRD6, RANBP9 and MVH while these bands were absent in GST alone (Fig 4B).

These interactions were further confirmed by western blot (WB) in co-immunoprecipitation

experiments. Using a specific anti-TEX19.1 monoclonal antibody, we further immuno-

purified TEX19.1 complexes in adult testes and identified individual polypeptides by MS.

Among the identified proteins, we found MAEL, MIWI, MILI, and MVH which were already

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

identified by GST pull-down, but also TDRD8, TDRD6, EDC4, PABPC2, PSMC3,

RANBP9, ANGEL1, SRPK1, DDX20, DDX46. We confirmed the presence of MAEL,

MIWI, MILI, MVH, EDC4, TDRD6 and RANBP9 in TEX19.1 immuno-purified complexes

isolated from cytoplasmic adult testes extracts by WB (Fig 4C, D). These interactions were

further validated by the reciprocal detection of TEX19.1 in immuno-purified MAEL, MIWI,

MILI, MVH, EDC4 and RANBP9 complexes (Fig 4D). Altogether these results suggest that

TEX19.1 and TEX19.2 interact with proteins of the post-natal PIWI/piRNA pathway, namely

core components such as MIWI, MILI, and accessory proteins such as EDC4, TDRD6,

RANBP9 and MVH.

The identification of TEX19 paralogs as partners of the PIWI proteins along with the

phenotypic similarities observed between Tex19KO mice and piRNA mutant mice prompted

us to check if TEX19s could bind small RNAs, and in particular piRNAs. For this purpose,

we immuno-precipitated TEX19.1, TEX19.2 and MILI (as a positive control) in adult mouse

(aged 12 weeks) testes extracts and assessed for the presence of small RNAs by 32P 5’ end

labeling and gel electrophoresis (Fig 5). As expected, MILI associated with ~26-30 nucleotide

(nt) small RNAs. TEX19.1 purification also revealed an interaction with small RNAs ranging

in size between 26 and 30 nt, migrating similarly to MILI associated piRNAs. Using the

6Tex:4D2 antibody that recognized both TEX19 paralogs on an adult Tex19.1KO testicular

extract from a normal size testis from a 12 weeks old male, we showed that TEX19.2 also

interacted with small RNAs of the same size range (Fig 5). All bands disappeared upon

RNase treatment, confirming the RNA nature of these entities. Furthermore, no such band

was detected in the ascitic fluid control sample or in the sample precipitated with a specific

anti-TEX19.1 monoclonal antibody in Tex19.1KO testicular extract, confirming the

specificity of the interaction between TEX19.1 and small RNAs. Thus both TEX19 paralogs

interact with small RNAs in a size range compatible with piRNAs.

To further identify the small RNA species associated with TEX19 proteins, we generated

small RNA-seq libraries by Illumina sequencing after TEX19 and TEX19.1

immunoprecipitation from adult WT testes. To infer the TEX19.2-bound fraction, we

analyzed TEX19-complexes in Tex19.1KO testes. Size distribution and genomic content were

then compared to available datasets from MILI and MIWI-bound piRNAs from adult testes

(Robine et al., 2009). MIWI was formerly shown to preferentially associate with 29-30nt

piRNA species, while the main piRNA size in MILI complexes is shorter, around 26-27nt.

Interestingly, TEX19-, TEX19.1- and TEX19.2-associated small RNAs were found to

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

segregate in size with MIWI-bound piRNAs, with a diagnostic 30nt size (Fig 6A). TEX19.2

showed a stronger association to 22-23nt long RNAs, which is a diagnostic size for

microRNAs: as small RNA libraries are expressed as relative rather than absolute values, this

may reflect a weaker affinity of TEX19.2 towards piRNAs but a genuine preferential

association with miRNAs cannot be excluded.

Because the three libraries were very similar in size and content, but the TEX19.1 sample had

a better quality, we focused the rest of our analysis on the latter. Out of 1.6 million reads, we

noticed a strong preference (94.42%) for a uridine at position one (U1 bias), in agreement

with the biogenesis mode of piRNAs (Aravin et al., 2006, Frank et al., 2010, Kawaoka et al.,

2011) (Table S2, Fig S4). The genomic distribution appeared identical to both MILI and

MIWI IPs (Fig 6B): the majority (70%) of the TEX19.1-derived reads originated from

intergenic regions, while less than 20% of the reads mapped to repeated elements. This is in

line with the known genomic distribution of pachytene piRNA species, which are produced in

adult testes from a discrete number of unique and intergenic piRNA clusters . Accordingly,

90% TEX19.1-associated reads uniquely mapped onto the genome (Fig 6C), and we moreover

found the large majority of TEX19.1-associated small RNAs to be derived from the same

piRNA loci clusters previously defined as regulated by MILI and MIWI (Fig 6D) (Beyret et

al., 2012). We nonetheless examined the content of the repeated fraction of the TEX19.1

library and found that around 50% were derived from LTR retrotransposons, among which

around 1% belonged to the MMERVK10C class (Fig 6B). This means that along all 25-30nt-

bound RNAs, only 0.1% derived from MMERVK10C elements. Finally, to address the role of

TEX19.1 in the biogenesis of piRNAs, we sequenced whole testes small RNA libraries in

three adult (12 weeks of age) Tex19.1KO animals showing normal size testis: the size

distribution appeared similar to WT testes, and in particular there was no relative

redistribution towards 22-23nt microRNA species, as an indication of impaired piRNA

production (Fig 6E). Enrichment in 30nt-centered small RNAs was even higher in the

Tex19.1KO background compared to WT: this likely reflects different cellular composition,

and in particular an overrepresentation of pachytene cells in Tex19.1KO testes, due to the

meiotic arrest. Altogether, these results suggest that TEX19 proteins sample the same set of

primary, unique piRNA species as MILI and MIWI, at the time of their production at the

pachytene stage in the adult testis. However, they are not required for their synthesis, and

likely act as simple cargo.

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

TEX19 paralogs interact directly with piRNAs

In order to assess a direct interaction between TEX19.1, TEX19.2 and piRNAs, we performed

electrophoretic mobility shift assays (EMSA) using GST-TEX19.1 and GST-TEX19.2

proteins as well as different domains of TEX19.1 fused to GST (Fig S5 and S6) against the

piR-117061 piRNA. We choose this piRNA because it has a very significant absolute count

(2361 and 2028) in two biological replicates prepared using anti-TEX19.1 antibody (Table

S3) and it has a U at position one “U1”. We found that the piR-117061 was ‘shifted' up upon

electrophoresis in the presence of GST-TEX19.1, GST-TEX19.2, GST-D2Cter (125-351aa)

or GST-VPTEL of TEX19.1, but not in the presence of GST-alone or GST-MCP (1-124aa) or

GST-Cter (163-351aa) of TEX19.1. This result stresses the specificity of the interaction

between piR-117061 and TEX19.2 and TEX19.1 via the VPTEL domain (Fig 7A). The

specificity of this interaction was further confirmed by showing that a modified form of piR-

117061, in which we replaced U1 and A10 by a C was unable to bind TEX19.1, its VPTEL

domain or TEX19.2 (Fig 7B) and was not able to replace piR-117061 in a competition

experiment (Fig 7C). By adding an increased concentration of the non-radiolabeled piR-

117061 to a fixed concentration of GST-VPTEL (5µM pmole) and radiolabeled piR-117061,

we showed that the non-radiolabeled piR-117061 completely displaced the radiolabeled piR-

117061 at one hundred fold concentration (Fig 7D). Finally, we showed that the quantity of

piR-117061 ‘shift' up was proportional to the quantity of the TEX19.1 VPTEL domain (Fig

7D).

Discussion

Tex19 genes are specific to mammals and can be found as tandem duplication in the mouse

and rat genomes. In the mouse, Tex19.1 and Tex19.2 expression is restricted to the germ line

and the placenta (Celebi et al., 2012). Analyses of Tex19.1KO mice have previously

highlighted a variable phenotype. In the severest form, Tex19.1 deficient males were infertile

and presented an arrest at the pachytene stage of meiosis, in association with a failure to

silence MMERVK10C elements (Ollinger et al., 2008, Tarabay et al., 2013). Tex19.2KO

males are fertile; however a subtle phenotype of testes degeneration could be observed in

adult mice (unpublished observations Tarabay et al, 2013). The present study describes the

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

functional consequences of deleting both TEX19 paralogs and the involvement of these

proteins in the piRNA pathway. Surviving Tex19DKO male and female mice were perfectly

healthy, but males exhibited a fully penetrant phenotype similar to the severest Tex19.1KO

phenotype (Ollinger et al., 2008, Tarabay et al., 2013). Indeed, Tex19DKO males have highly

degenerated testes, with a lack of germ cells beyond the pachytene stage, linked to defects in

chromosome synapsis during meiosis. Compared to Tex19.1KO testes (Ollinger et al., 2008,

Tarabay et al., 2013), we showed here a stronger up-regulation of MMERVK10C

retrotransposon in double Tex19DKO testes, which was two-fold at p10 and four-fold at p16

compared to WT counterparts. This fully penetrant phenotype comforts the hypothesis of a

functional redundancy between Tex19.1 and Tex19.2. Moreover, this phenotype bears striking

resemblance with the Dnmt3L, Miwi2 and Mili mutant phenotypes (Aravin et al., 2007, Bestor

and Bourc'his, 2004, Carmell et al., 2007, Kuramochi-Miyagawa et al., 2004). One

explanation was proposed by Zamudio et al, who suggested that improper TE silencing may

alter the meiotic process through local remodeling of the chromatin landscape and (Zamudio

et al., 2015).

The essential role of TEX19 may extend beyond the context of spermatogenesis as we report

here a severe lethality of Tex19DKO animals, with only 8.3%, instead of the expected 25% of

mutant animals. Moreover, this lethality was sexually dysmorphic as among the survivors,

only one third were females. We did not observe such lethality in our single Tex19.1KO

mouse models (Tarabay et al., 2013), but Yang et al., 2010 also reported such female-specific

lethality in their Tex19.1KO mice and suggested that the variable severity among different

studies could be linked to strain specificities (Yang et al., 2010, Kwon et al., 2003). However,

our studies on Tex19.1KO or Tex19DKO mice were conducted onto the same genetic

background (C57Bl/6). A strong bias in sexual lethality was also observed in C57Bl/6 mice

by Reichmann et al, but only when females were lactating and nursing a pre-existing litter

during pregnancy (Reichmann et al., 2013). Yet, none of our studies were conducted under

such lactating conditions. Consequently, this difference between Tex19.1KO and Tex19DKO

phenotypes is quite surprising since Tex19.2 is not expressed in the placenta, neither in WT

nor in Tex19.1KO animals. This early postnatal lethality could be indicative of defects in

placenta (Reichmann et al., 2013, Tarabay et al., 2013), but also suggests a functional

redundancy between Tex19.1 and Tex19.2 during embryo development. The exact reason for

the preferential female lethality of the Tex19DKO mutation remains unclear, but may be

linked to X chromosome-specific effects.

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

The main findings provided by our work relate to the identification of the protein partners of

TEX19 in adult testes. Through GST pulldown and IP experiments, we indeed showed for the

first time that TEX19.1 exists in a complex with members of the PIWI/piRNA pathway:

MAEL, MIWI, MILI and MVH. We also confirmed a significant level of interaction between

TEX19.1 and UBR2 by IP, as it had been previously reported (Yang et al., 2010). Most

importantly, we found that both TEX19 paralogs bind small RNAs that have all the

characteristics of piRNAs: around 30nt in size and with a strong preference for a uridine as

the first nucleotide (Frank et al., 2010, Kawaoka et al., 2011, Brennecke et al., 2007). Finally,

this interaction seems to be direct and to occur through the well-conserved VPTEL domain of

TEX19 proteins. We can therefore conclude that TEX19 paralogs are new members of the

family of proteins that bind piRNAs and that the VPTEL domain could be a new small RNA-

binding motif. While this study may reveal the piRNA-binding function of the VPTEL

domain, we cannot speculate about the function of the MCP domain, which lacks homology

with any known proteins. PIWI proteins have, in addition to their ability to bind piRNAs, an

RNA-guided nuclease function linked to a catalytic triad DDH localized in the PIWI domain

and which is required for piRNA biogenesis (Martinez and Tuschl, 2004). No such triad can

be found in TEX19 proteins. Accordingly, the production of 25-30 nt small RNAs is not

altered in Tex19-deficient testes. This suggests that while TEX19 proteins bind to piRNAs,

they are not involved in the catalysis of piRNA synthesis.

Finally, it is now well established that TEX19 paralogs are required for the control of

MMERVK10C transposons in testes (Ollinger et al., 2008, Tarabay et al., 2013). Despite a

consistent effect of Tex19 deficiency on these elements, we found that MMERVK10C-

derived reads only represent 0.1% of all 25-30nt TEX19-bound small RNAs. In contrast,

TEX19 mostly binds unique, non-transposon derived small RNAs in adult testes, which are

similar to the MILI and MIWI-associated piRNAs. This is not surprising, as the post-natal

PIWI machinery is responsible for producing piRNAs derived from discrete intergenic

regions as male germ cells enter the pachytene stage (Aravin et al., 2006, Girard et al., 2006)

although the function of these mammalian post-natal piRNAs remains enigmatic (Goh et al.,

2015). This poor enrichment of TEX19-bound small RNAs in MMERVK10C fragments

raises the question as to whether the reactivation of MMERVK10C in Tex19.1KO and

Tex19DKO mice reflects a direct role of TEX19 proteins in silencing these elements-likely

via post-transcriptional mechanisms- or an indirect effect of the Tex19 mutation. Further work

should be dedicated to resolve this question.

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

In conclusion, our analysis of the double-mutant mice for the two TEX19 paralogs

undoubtedly reveals the crucial role of these proteins for male fertility in mammals, likely via

their association with the PIWI/piRNA pathway.

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

Material and methods

Generation of Tex19 double knockout mice

The Tex19 double mutant mouse line was established at the MCI/ICS (Mouse Clinical

Institute- Institut Clinique de la Souris-, Illkirch, France; http://www-mci.u-strasbg.fr). For

Tex19.1, the targeting vector was constructed as follows (Fig 1A): three PCR fragments (from

129S2/SvPas ES cell genomic DNA) one 2.2 Kb fragment corresponding to inter-loxP

fragment encompassing exons 2 and 3, a 4.5 kb and a 3.1kb fragment corresponding to the 5’

and 3’ arms respectively were sequentially subcloned in a MCI proprietary vector. This MCI

vector contains a LoxP site as well as a floxed and flipped Neomycin resistance cassette. A

unique PacI restriction site was inserted downstream to the 3’ loxP site. The linearized

construct was electroporated in 129S2/SvPas mouse embryonic stem (ES) cells. After G418

selection, targeted clones were identified by PCR using external primers and further

confirmed by Southern blot with 5’ and 3’ external probes.

For Tex19.2, the targeting vector was constructed as follows: three PCR fragments (from

129S2/SvPas ES cell genomic DNA) a 2.0 Kb fragment corresponding to inter-lox P fragment

encompassing exons 2 and 3, a 4.4 kb and a 3.5kb fragment corresponding to the 5’ and the 3’

arms respectively were sequentially subcloned in a MCI proprietary vector. This MCI vector

contains a Lox5171 site as well as a F3 surrounded Hygromycin resistance cassette. A unique

PacI restriction site was inserted upstream to the 5’ loxP site. The linearized construct was

electroporated in one fully valid Tex19.1 KO clone. After hygromycin selection, targeted

clones were identified by PCR using external primers and further confirmed by Southern blot

with 5’ and 3’ external probes. Restriction digestion by PacI followed by Southern blot

allowed selection of a clone showing targeting of both constructs on the same allele. This

clone was injected into C57BL/6J blastocysts, and male chimeras tested for germline

transmission. The NeoR and HygroR cassettes were deleted by breeding the F1 mice with Flp

mice (Rodriguez et al., 2000). Excision of both Tex19.1 and Tex19.2 inter-Lox fragments was

done by breeding the mice with a Cre deleter line (Dupe et al., 1997). The absence of Tex19.1

and Tex19.2 mRNA and protein was tested by RT-PCR using oligonucleotides described

previously (Tarabay et al., 2013) and by western blotting respectively.

All animal experimental procedures were performed according to the European authority

guidelines.

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

Antibody production, Western blotting, preparation of cytoplasmic and nuclear

fractions

Injecting the entire protein or the peptide DLGPEDAEWTQALPWRC into mice allowed the

generation of anti-TEX19 monoclonal antibodies. The specificity of both antibodies was

tested as described elsewhere (Tarabay et al., 2013). The 7Tex:1F11 antibody is TEX19.1

specific, whereas 6Tex:4D2 antibody detects both TEX19.1 and TEX19.2.

Cytoplasmic and nuclear fraction were prepared as described

http://openwetware.org/wiki/Cytoplasm_and_nuclear_protein_extraction

RT-PCR and quantitative RT-PCR

RNA was prepared using the RNeasy mini or micro kit (Qiagen) following the manufacturer’s

instructions. After DNase I digestion (Roche, Mississauga, ON, Canada), 1µg RNA was

reverse-transcribed by random priming using Superscript II (Invitrogen/Life Technologies,

Burlington, ON, Canada). QPCR reaction was performed using SYBR® green JumpStartTM

Taq ReadyMixTM (Sigma-Aldrich, Oakville, ON, Canada) and LightCycler 480 (Roche,

Mississauga, ON, Canada). The efficiency and specificity of each primer pair was checked

using a cDNA standard curve. All samples were normalized to β-actin and Rrm2 expression

(Tarabay et al., 2013).

Histology

Testes were collected and fixed in Bouin’s fluid for 48 hours and then embedded in paraffin.

Adult male testis were dissected at 16 weeks of age. For histological analyses, 5 µm thick

sections were stained with hematoxylin/eosin. All slides were examined using a DMLA

microscope (Leica) with 10X, 20X, 40X and 100X objectives with apertures of 0.3, 0.5, 0.7

and 1.3, respectively. Images were taken with digital camera (CoolSnap; Photometrix) using

CoolSnap v.1.2 software and then processed with Photoshop CS2 v.9.0.2 (Adobe).

Immunostaining of meiotic chromosome spreads

Meiotic chromosome spreads were prepared as described previously (Mark et al., 2008).

TEX19 protein production and purification

Tex19.1 cDNA or its domains were cloned, by RT- PCR starting from a testis cDNA library,

into pCR®4-TOPO (Invitogen/Life Technologies, Burlington, ON, Canada) then into pENTR

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

1A (Life Technologies) then by LR recombinase into pHGGWA or pGGWA. The

recombinant proteins GST-TEX19, MCP, VPTEL-Cter, VPTEL, Cter, and protein GST-alone

were individually expressed in E. coli BL21 cells. The induction by 0.1mM of IPTG was

performed at a cell density of A600 = 0.8 at 20°C overnight and a portion of the proteins

examined by 10% SDS-PAGE. Then, sonication was performed 3s ON and 3s OFF for 5 min

at amplitude 33%. Cleared lysates were prepared in lysis buffer (Tris HCl 50mM pH9, NaCl

400mM, DTT 2mM, triton 1% and protease inhibiting cocktail) and centrifuged for 1 hour at

20000 rpm. The soluble fusion proteins were purified by batch procedure incubating the

lysate with 500µL of Glutathione Sepharose (Amersham Biosciences) for 1 hour at 4°C with

gentle shaking. Then the Glutathione Sepharose was collected by centrifugation and washed

five times with lysis buffer. After boiling, the bound proteins were analyzed by 10% SDS‐

PAGE followed by Western blotting using anti-TEX19 monoclonal antibodies 7Tex:1F11 or

6Tex:4D2.

For GST pulldown assays, total testicular extracts of adult mice (aged 12 weeks) were added

to recombinant protein fixed to Gluthatione Sepharose beads for 2h00 at 4°C. After fives

washes, proteins were separated on 10% SDS–PAGE and subsequently stained with Bio-Safe

Coomassie (Bio-Rad, Hercules, CA, USA). The gel was then sent to Taplin Biological Mass

Spectrometry Facility to be analyzed by LC/MS/MS (Taplin, Boston, MA, USA).

Purified recombinant proteins were used in EMSA assays as described hereafter.

Co-Immunoprecipitation

Antibodies were chemically cross-linked to protein G–Sepharose beads using dimethyl

pimelimidate (DMP; Sigma-Aldrich D-8388) and used to purify TEX19.1, TEX19.2, MAEL,

MIWI, MILI, MVH complexes from adult mouse testes (aged 12 weeks) cytoplasmic

fractions. 30mg of clarified DWT testes cytoplasmic extracts were incubated with 300µl

protein G-Sepharose beads coupled to antibodies; anti-TEX19.1 (7Tex:1F11,100µl), anti-

TEX19.1/TEX19.2 (6Tex:4D2, 10ml of hybridoma supernatant), anti-MAEL (18µg, Abcam,

Toronto, ON, Canada), anti-MIWI (20µg, Aviva System Biology, San Diego, CA, USA),

anti-MILI (40µg, Abcam), anti-MVH (12µg, Abcam), anti-TDRD66 (50µg,

Millipore/Upstate, Etobicode, ON, Canada), anti-EDC4 (12µg, Abcam), anti-RANBP9 (8µg,

Proteintech, Rosemonte, IL, USA) or an ascitic fluid control (negative control, 100µl) in Co-

IP buffer (50 mM Tris, pH 8, 150 mM NaCl, 1Mm EDTA, complete protease inhibitor)

overnight at 4°C. After five washes with Co-IP buffer, the retained proteins were denatured

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

by Laemmli buffer with 5% -Mercaptoethanol and boiled to 100°C, then separated by SDS-

PAGE as described hereafter. If required, gels were stained with Coomassie blue and sent to

Taplin Biological Mass Spectrometry Facility to be analyzed by LC/MS/MS (Taplin, Boston,

MA, USA).

Small RNA immunoprecipitation and purification

TEX19 libraries were prepared as described elsewhere (Reuter et al., 2009).

Electrophoretic Mobility Shift Assay (EMSA)

For 32P-labeling, piR-117061 (5’UAGGACCUGAGAACUUAACCUUGUUAUGGG 3’)

and piR-nonspecific (5’CAGGACCUGCGAACUUAACCUUGUUAUGGG 3’) were

incubated for 60min at 37°C in 20µl kinase buffer containing 5μl RNA (0.5 nmol), 2μl buffer

B, 4µl PEG, 10U T4 PNK kinase (10U/µl, Fermentas), 40U RNAsine (40U/µl), 2 μl [γ-

32P]ATP (6000 Ci/mmole) (Perkin Elmer), 5μl ddH2O.

EMSA assays were performed with 5 µM GST-alone, GST-TEX19.1, GST-MCP, GST-

VPTEL/Cter, GST-VPTEL, GST-Cter and 125 femtomoles (8000cpm) denatured

radiolabeled piRNA in the binding buffer (250µg Herpain, 10Mm MOPS pH 7, 50mM

MgCl2, 1 mM DTT,10% glycerol, 40U RNAsine, 10ng/µl tRNA of E.coli) as described

(Bendak et al., 2012).

Small RNA sequencing and analysis

Small RNA-seq of whole adult testes and TEX19-IP were done using Illumina HiSeq 2000 on

libraries (18-45nt RNAs size-fractionated on acrylamide gel) prepared with the Illumina

TruSeq small RNA protocol. For the analysis, adapters were removed using Cutadapt (Martin,

2011). The trimmed sequencing reads of each library were then mapped and annotated with

ncPRO-seq (Chen et al., 2012) onto the mouse reference genome (mm9). Briefly, read

alignment was done using Bowtie v0.12.8, allowing a sum of qualities of mismatching bases

lower than 50 (-e 50). A number of 5,000 hits per aligned read were allowed. Aligned reads

were then annotated using RepeatMasker and refGene databases, and the positional read

coverage was weighted by mapping site numbers.

The piRNA cluster analysis was performed as described (Beyret et al., 2012). A cluster was

defined as a group of piRNAs where the reads were less than 1,500bp away from each other.

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

Clusters with at least 1,000 normalized read count were reported. The UCSC genome tracks

were generated using the HOMER software (v4.3) using unique mapped reads.

Funding

The study was funded by Agence Nationale de la Recherche (ANR-11-BSV2-002

« TranspoFertil ») and l’Agence de BioMédecine (« AMP, diagnostic prénatal et diagnostic

génétique »).

Acknowledgment

Y.T., M.A., M.T., T.Y., A.T., M.M., D.B., S.V.: Data analysis and interpretation, manuscript

writing, final approval of manuscript. We would also like to thank Robert Drillien for his

critical reading of the manuscript. We are grateful to the Institute of Genetics and Molecular

and Cellular Biology (IGBMC) platforms. The study was funded by Agence Nationale de la

Recherche (ANR-11-BSV2-002 « TranspoFertil ») and l’Agence de BioMédecine (« AMP,

diagnostic prénatal et diagnostic génétique »). This work was supported by the French Centre

National de la Recherche Scientifique (CNRS), Institut National de la Santé et de la

Recherche Médicale (INSERM), the Ministère de l’Education Nationale et de l’Enseignement

Supérieur et de la Recherche, the University of Strasbourg, the University Hospital of

Strasbourg.

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

References

ARAVIN, A., GAIDATZIS, D., PFEFFER, S., LAGOS-QUINTANA, M., LANDGRAF, P.,

IOVINO, N., MORRIS, P., BROWNSTEIN, M. J., KURAMOCHI-MIYAGAWA, S.,

NAKANO, T., CHIEN, M., RUSSO, J. J., JU, J., SHERIDAN, R., SANDER, C.,

ZAVOLAN, M. & TUSCHL, T. 2006. A novel class of small RNAs bind to MILI

protein in mouse testes. Nature, 442, 203-7.

ARAVIN, A. A., HANNON, G. J. & BRENNECKE, J. 2007. The Piwi-piRNA pathway

provides an adaptive defense in the transposon arms race. Science, 318, 761-4.

BENDAK, K., LOUGHLIN, F. E., CHEUNG, V., O'CONNELL, M. R., CROSSLEY, M. &

MACKAY, J. P. 2012. A rapid method for assessing the RNA-binding potential of a

protein. Nucleic Acids Res, 40, e105.

BESTOR, T. H. & BOURC'HIS, D. 2004. Transposon silencing and imprint establishment in

mammalian germ cells. Cold Spring Harb Symp Quant Biol, 69, 381-7.

BEYRET, E., LIU, N. & LIN, H. 2012. piRNA biogenesis during adult spermatogenesis in

mice is independent of the ping-pong mechanism. Cell Res, 22, 1429-39.

BRENNECKE, J., ARAVIN, A. A., STARK, A., DUS, M., KELLIS, M.,

SACHIDANANDAM, R. & HANNON, G. J. 2007. Discrete small RNA-generating

loci as master regulators of transposon activity in Drosophila. Cell, 128, 1089-103.

CARMELL, M. A., GIRARD, A., VAN DE KANT, H. J., BOURC'HIS, D., BESTOR, T. H.,

DE ROOIJ, D. G. & HANNON, G. J. 2007. MIWI2 is essential for spermatogenesis

and repression of transposons in the mouse male germline. Dev Cell, 12, 503-14.

CELEBI, C., VAN MONTFOORT, A., SKORY, V., KIEFFER, E., KUNTZ, S., MARK, M.

& VIVILLE, S. 2012. Tex 19 paralogs exhibit a gonad and placenta-specific

expression in the mouse. J Reprod Dev, 58, 360-5.

CHEN, C. J., SERVANT, N., TOEDLING, J., SARAZIN, A., MARCHAIS, A.,

DUVERNOIS-BERTHET, E., COGNAT, V., COLOT, V., VOINNET, O., HEARD,

E., CIAUDO, C. & BARILLOT, E. 2012. ncPRO-seq: a tool for annotation and

profiling of ncRNAs in sRNA-seq data. Bioinformatics, 28, 3147-9.

CHUMA, S., HIYOSHI, M., YAMAMOTO, A., HOSOKAWA, M., TAKAMUNE, K. &

NAKATSUJI, N. 2003. Mouse Tudor Repeat-1 (MTR-1) is a novel component of

chromatoid bodies/nuages in male germ cells and forms a complex with snRNPs.

Mech Dev, 120, 979-90.

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

COSTA, Y., SPEED, R. M., GAUTIER, P., SEMPLE, C. A., MARATOU, K., TURNER, J.

M. & COOKE, H. J. 2006. Mouse MAELSTROM: the link between meiotic silencing

of unsynapsed chromatin and microRNA pathway? Hum Mol Genet, 15, 2324-34.

COWLEY, M. & OAKEY, R. J. 2013. Transposable elements re-wire and fine-tune the

transcriptome. PLoS Genet, 9, e1003234.

DUPE, V., DAVENNE, M., BROCARD, J., DOLLE, P., MARK, M., DIERICH, A.,

CHAMBON, P. & RIJLI, F. M. 1997. In vivo functional analysis of the Hoxa-1 3'

retinoic acid response element (3'RARE). Development, 124, 399-410.

FRANK, F., SONENBERG, N. & NAGAR, B. 2010. Structural basis for 5'-nucleotide base-

specific recognition of guide RNA by human AGO2. Nature, 465, 818-22.

FURANO, A. V., DUVERNELL, D. D. & BOISSINOT, S. 2004. L1 (LINE-1)

retrotransposon diversity differs dramatically between mammals and fish. Trends

Genet, 20, 9-14.

GASIOR, S. L., WAKEMAN, T. P., XU, B. & DEININGER, P. L. 2006. The human LINE-1

retrotransposon creates DNA double-strand breaks. J Mol Biol, 357, 1383-93.

GHILDIYAL, M. & ZAMORE, P. D. 2009. Small silencing RNAs: an expanding universe.

Nat Rev Genet, 10, 94-108.

GIRARD, A., SACHIDANANDAM, R., HANNON, G. J. & CARMELL, M. A. 2006. A

germline-specific class of small RNAs binds mammalian Piwi proteins. Nature, 442,

199-202.

GOH, W. S., FALCIATORI, I., TAM, O. H., BURGESS, R., MEIKAR, O., KOTAJA, N.,

HAMMELL, M. & HANNON, G. J. 2015. piRNA-directed cleavage of meiotic

transcripts regulates spermatogenesis. Genes Dev, 29, 1032-44.

GOODIER, J. L. & KAZAZIAN, H. H., JR. 2008. Retrotransposons revisited: the restraint

and rehabilitation of parasites. Cell, 135, 23-35.

HAJKOVA, P. 2011. Epigenetic reprogramming in the germline: towards the ground state of

the epigenome. Philos Trans R Soc Lond B Biol Sci, 366, 2266-73.

HOSOKAWA, M., SHOJI, M., KITAMURA, K., TANAKA, T., NOCE, T., CHUMA, S. &

NAKATSUJI, N. 2007. Tudor-related proteins TDRD1/MTR-1, TDRD6 and

TDRD7/TRAP: domain composition, intracellular localization, and function in male

germ cells in mice. Dev Biol, 301, 38-52.

KAWAOKA, S., IZUMI, N., KATSUMA, S. & TOMARI, Y. 2011. 3' end formation of

PIWI-interacting RNAs in vitro. Mol Cell, 43, 1015-22.

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

KOTAJA, N., BHATTACHARYYA, S. N., JASKIEWICZ, L., KIMMINS, S., PARVINEN,

M., FILIPOWICZ, W. & SASSONE-CORSI, P. 2006. The chromatoid body of male

germ cells: similarity with processing bodies and presence of Dicer and microRNA

pathway components. Proc Natl Acad Sci U S A, 103, 2647-52.

KUNTZ, S., KIEFFER, E., BIANCHETTI, L., LAMOUREUX, N., FUHRMANN, G. &

VIVILLE, S. 2008. Tex19, a mammalian-specific protein with a restricted expression

in pluripotent stem cells and germ line. Stem Cells, 26, 734-44.

KURAMOCHI-MIYAGAWA, S., KIMURA, T., IJIRI, T. W., ISOBE, T., ASADA, N.,

FUJITA, Y., IKAWA, M., IWAI, N., OKABE, M., DENG, W., LIN, H., MATSUDA,

Y. & NAKANO, T. 2004. Mili, a mammalian member of piwi family gene, is

essential for spermatogenesis. Development, 131, 839-49.

KWON, Y. T., XIA, Z., AN, J. Y., TASAKI, T., DAVYDOV, I. V., SEO, J. W., SHENG, J.,

XIE, Y. & VARSHAVSKY, A. 2003. Female lethality and apoptosis of spermatocytes

in mice lacking the UBR2 ubiquitin ligase of the N-end rule pathway. Mol Cell Biol,

23, 8255-71.

LANDER, E. S., LINTON, L. M., BIRREN, B., NUSBAUM, C., ZODY, M. C., BALDWIN,

J., DEVON, K., DEWAR, K., DOYLE, M., FITZHUGH, W., FUNKE, R., GAGE, D.,

HARRIS, K., HEAFORD, A., HOWLAND, J., KANN, L., LEHOCZKY, J., LEVINE,

R., MCEWAN, P., MCKERNAN, K., MELDRIM, J., MESIROV, J. P., MIRANDA,

C., MORRIS, W., NAYLOR, J., RAYMOND, C., ROSETTI, M., SANTOS, R.,

SHERIDAN, A., SOUGNEZ, C., STANGE-THOMANN, N., STOJANOVIC, N.,

SUBRAMANIAN, A., WYMAN, D., ROGERS, J., SULSTON, J., AINSCOUGH, R.,

BECK, S., BENTLEY, D., BURTON, J., CLEE, C., CARTER, N., COULSON, A.,

DEADMAN, R., DELOUKAS, P., DUNHAM, A., DUNHAM, I., DURBIN, R.,

FRENCH, L., GRAFHAM, D., GREGORY, S., HUBBARD, T., HUMPHRAY, S.,

HUNT, A., JONES, M., LLOYD, C., MCMURRAY, A., MATTHEWS, L.,

MERCER, S., MILNE, S., MULLIKIN, J. C., MUNGALL, A., PLUMB, R., ROSS,

M., SHOWNKEEN, R., SIMS, S., WATERSTON, R. H., WILSON, R. K., HILLIER,

L. W., MCPHERSON, J. D., MARRA, M. A., MARDIS, E. R., FULTON, L. A.,

CHINWALLA, A. T., PEPIN, K. H., GISH, W. R., CHISSOE, S. L., WENDL, M. C.,

DELEHAUNTY, K. D., MINER, T. L., DELEHAUNTY, A., KRAMER, J. B.,

COOK, L. L., FULTON, R. S., JOHNSON, D. L., MINX, P. J., CLIFTON, S. W.,

HAWKINS, T., BRANSCOMB, E., PREDKI, P., RICHARDSON, P., WENNING, S.,

SLEZAK, T., DOGGETT, N., CHENG, J. F., OLSEN, A., LUCAS, S., ELKIN, C.,

UBERBACHER, E., FRAZIER, M., et al. 2001. Initial sequencing and analysis of the

human genome. Nature, 409, 860-921.

MA, L., BUCHOLD, G. M., GREENBAUM, M. P., ROY, A., BURNS, K. H., ZHU, H.,

HAN, D. Y., HARRIS, R. A., COARFA, C., GUNARATNE, P. H., YAN, W. &

MATZUK, M. M. 2009. GASZ is essential for male meiosis and suppression of

retrotransposon expression in the male germline. PLoS Genet, 5, e1000635.

MARK, M., JACOBS, H., OULAD-ABDELGHANI, M., DENNEFELD, C., FERET, B.,

VERNET, N., CODREANU, C. A., CHAMBON, P. & GHYSELINCK, N. B. 2008.

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

STRA8-deficient spermatocytes initiate, but fail to complete, meiosis and undergo

premature chromosome condensation. J Cell Sci, 121, 3233-42.

MARTIN, M. 2011. Cutadapt removes adapter sequences from high-throughput sequencing

reads. EMBnet.journal, 17, 10.

MARTINEZ, J. & TUSCHL, T. 2004. RISC is a 5' phosphomonoester-producing RNA

endonuclease. Genes Dev, 18, 975-80.

OBBARD, D. J., GORDON, K. H., BUCK, A. H. & JIGGINS, F. M. 2009. The evolution of

RNAi as a defence against viruses and transposable elements. Philos Trans R Soc

Lond B Biol Sci, 364, 99-115.

OLLINGER, R., CHILDS, A. J., BURGESS, H. M., SPEED, R. M., LUNDEGAARD, P. R.,

REYNOLDS, N., GRAY, N. K., COOKE, H. J. & ADAMS, I. R. 2008. Deletion of

the pluripotency-associated Tex19.1 gene causes activation of endogenous retroviruses

and defective spermatogenesis in mice. PLoS Genet, 4, e1000199.

OSTERTAG, E. M. & KAZAZIAN, H. H., JR. 2001. Biology of mammalian L1

retrotransposons. Annu Rev Genet, 35, 501-38.

REICHMANN, J., REDDINGTON, J. P., BEST, D., READ, D., OLLINGER, R., MEEHAN,

R. R. & ADAMS, I. R. 2013. The genome-defence gene Tex19.1 suppresses LINE-1

retrotransposons in the placenta and prevents intra-uterine growth retardation in mice.

Hum Mol Genet, 22, 1791-806.

REUTER, M., CHUMA, S., TANAKA, T., FRANZ, T., STARK, A. & PILLAI, R. S. 2009.

Loss of the Mili-interacting Tudor domain-containing protein-1 activates transposons

and alters the Mili-associated small RNA profile. Nat Struct Mol Biol, 16, 639-46.

ROBINE, N., LAU, N. C., BALLA, S., JIN, Z., OKAMURA, K., KURAMOCHI-

MIYAGAWA, S., BLOWER, M. D. & LAI, E. C. 2009. A broadly conserved

pathway generates 3'UTR-directed primary piRNAs. Curr Biol, 19, 2066-76.

RODRIGUEZ, C. I., BUCHHOLZ, F., GALLOWAY, J., SEQUERRA, R., KASPER, J.,

AYALA, R., STEWART, A. F. & DYMECKI, S. M. 2000. High-efficiency deleter

mice show that FLPe is an alternative to Cre-loxP. Nat Genet, 25, 139-40.

ROLLINS, R. A., HAGHIGHI, F., EDWARDS, J. R., DAS, R., ZHANG, M. Q., JU, J. &

BESTOR, T. H. 2006. Large-scale structure of genomic methylation patterns. Genome

Res, 16, 157-63.

SOPER, S. F., VAN DER HEIJDEN, G. W., HARDIMAN, T. C., GOODHEART, M.,

MARTIN, S. L., DE BOER, P. & BORTVIN, A. 2008. Mouse maelstrom, a

component of nuage, is essential for spermatogenesis and transposon repression in

meiosis. Dev Cell, 15, 285-97.

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

TARABAY, Y., KIEFFER, E., TELETIN, M., CELEBI, C., VAN MONTFOORT, A.,

ZAMUDIO, N., ACHOUR, M., EL RAMY, R., GAZDAG, E., TROPEL, P., MARK,

M., BOURC'HIS, D. & VIVILLE, S. 2013. The mammalian-specific Tex19.1 gene

plays an essential role in spermatogenesis and placenta-supported development. Hum

Reprod, 28, 2201-14.

VASILEVA, A., TIEDAU, D., FIROOZNIA, A., MULLER-REICHERT, T. &

JESSBERGER, R. 2009. Tdrd6 is required for spermiogenesis, chromatoid body

architecture, and regulation of miRNA expression. Curr Biol, 19, 630-9.

WANG, P. J., MCCARREY, J. R., YANG, F. & PAGE, D. C. 2001. An abundance of X-

linked genes expressed in spermatogonia. Nat Genet, 27, 422-6.

YANG, F., CHENG, Y., AN, J. Y., KWON, Y. T., ECKARDT, S., LEU, N. A.,

MCLAUGHLIN, K. J. & WANG, P. J. 2010. The ubiquitin ligase Ubr2, a recognition

E3 component of the N-end rule pathway, stabilizes Tex19.1 during spermatogenesis.

PLoS One, 5, e14017.

YODER, J. A., WALSH, C. P. & BESTOR, T. H. 1997. Cytosine methylation and the

ecology of intragenomic parasites. Trends Genet, 13, 335-40.

ZAMUDIO, N., BARAU, J., TEISSANDIER, A., WALTER, M., BORSOS, M., SERVANT,

N. & BOURC'HIS, D. 2015. DNA methylation restrains transposons from adopting a

chromatin signature permissive for meiotic recombination. Genes Dev, 29, 1256-70.

ZAMUDIO, N. & BOURC'HIS, D. 2010. Transposable elements in the mammalian germline:

a comfortable niche or a deadly trap? Heredity (Edinb), 105, 92-104.

ZHENG, K. & WANG, P. J. 2012. Blockade of pachytene piRNA biogenesis reveals a novel

requirement for maintaining post-meiotic germline genome integrity. PLoS Genet, 8,

e1003038.

ZHENG, K., XIOL, J., REUTER, M., ECKARDT, S., LEU, N. A., MCLAUGHLIN, K. J.,

STARK, A., SACHIDANANDAM, R., PILLAI, R. S. & WANG, P. J. 2010. Mouse

MOV10L1 associates with Piwi proteins and is an essential component of the Piwi-

interacting RNA (piRNA) pathway. Proc Natl Acad Sci U S A, 107, 11841-6.

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

Figures

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

Fig 1: Genetic targeting of Tex19 paralogs in mice. (A) The loci of the Tex19 paralogs are

shown before and after homologous recombination upon removal of the neomycin (Neo) gene

by crossing the mice with the Flp-recombinase strain. LoxP sites surrounding Tex19.1,

Lox5171 sites surrounding Tex19.2 and PacI restriction sites (PacI RS) are depicted; (B)

Examples of PCR genotyping; one example of each genotype, Tex19.1KO (+/+ and -/-),

Tex19.2KO (+/+ and -/-) and Tex19DKO is shown. Wild type alleles are indicated by +/+

whereas mutant alleles are indicated by -/-. (C) Western blot using the 6Tex:4D2 antibody

recognizing both paralogs on adult testes extracts from WT, Tex19.1KO and Tex19DKO

showing the absence of TEX19 proteins in the Tex19DKO. β-Tubulin is used as a loading

control. M: molecular size marker.

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

Fig 2: Tex19DKO testicular defects. (A) Testes from Tex19DKO (DKO) mice are smaller

than testes from double wild type (DWT) littermates (8-week old); (B) Adult testes weight in

mg. Mean testes weight is significantly reduced in Tex19DKO mice (Student t-test *p<0.01)

(C, D) Histological sections stained with hematoxylin and eosin through the testes of adult

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

mice. (E-H) Double immunostaining of spread nuclei using antibodies against SYCP1 and

either SYCP3 or γH2AX, as indicated. In E, F, the synaptonemal complexes (SC) appear

yellow from overlapping of SYCP3 (green) and SYCP1 (red) signals. In both DWT and

Tex19DKO spermatocytes at the zygotene stage, short fragments of SC begin to form. In

DWT spermatocytes at the pachytene stage, all nineteen bivalents are fully synapsed. In about

20% of the Tex19DKO spermatocytes at the pachytene stage, only few chromosomes are fully

(arrowheads) or partially (arrow) synapsed, while the majority is unsynapsed (thin green

strands). (I, K) DWT and Tex19DKO spermatocytes at the zygotene stage express γH2AX

(green signal) throughout the nucleus. (J) In DWT spermatocytes at the pachytene stage,

γH2AX expression becomes restricted to the (almost) unsynapsed XY body. (L) In about 80%

of the Tex19DKO spermatocytes at the pachytene stage, γH2AX remains widely distributed

thought the nucleus, even in regions that are apparently synapsed as assessed from the

presence of thick SYCP3-positive strands (in red). Bar (in D): 100 m (in L): 10 m.

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

Fig 3: Testes histology in double wild type (DWT), double heterozygous (DHZ) and

double knockout (DKO) animals. In DWT (A, D, G, J) and DHZ (B, E, H, K) testis

histology is normal. The seminiferous tubules contain Sertoli cells and germ line cells are

found at all studied ages. (L) Tex19DKO testes exhibit a severe phenotype with a diminished

tubule diameter, vacuolization of the seminiferous epithelium and a reduction of the germ cell

number at 16 weeks. (C) An abnormal testicular phenotype starts to be visible at 16 days of

age (p16) (F) and increase at 20 days of age (p20) days (I, L) but is absent at 10 days of age

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

(p10) (C). Thus, the histological phenotype starts to be visible between p10 and p16. Bar (in

L): 100 μm.

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

Fig 4: GST pull-down and immunoprecipitation of TEX19 paralogs. (A) Western blot

(WB) showing GST-TEX19.1 and GST-TEX19.2. (B) Western blot analysis of the GST

pulldown revealing that MAEL, MIWI, MILI, MVH, TDRD6 and RANBP9 are present in the

analyzed pull down samples. (C) Reciprocal immunoprecipitation experiment confirming the

association of MAEL, MILI, MIWI, MVH, RANBP9, EDC4 with TEX19.1 by WB. (D) Co-

immunoprecipitation of TEX19.1 by anti-TEX19.1 (7Tex: 1F11) and the detection by WB of

either MAEL, MILI, MIWI, MVH, TDRD6, RANBP9 or EDC4. IBIAB (Immuno blot with

IP antibody) is indicated with the corresponding IP antibody. Ascitic fluid from an

unimmunized animal is used as a negative control.

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

Fig 5: Interaction of TEX19.1 and TEX19.2 with ~ 30 nt RNAs. MILI, TEX19.1 or

TEX19.2 were immunoprecipited (IP) from two month-old mouse testis protein extract by

specific antibodies and bound RNA were then 32P-end-labeled and separated on a 15%

denaturing urea-polyacrylamide gel. Immunoprecipited RNAs are predominantly ~ 30 nt

long. The size and mobility of the oligoribonucleotide markers is indicated on the right. The

asterisk indicates RNAase treatment. Ctrl- indicates negative control; Ascitic fluid from an

unimmunized animal is used as a negative control.

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

Fig 6: PiRNA populations isolated from TEX19 paralogs and MILI complexes were

cloned, sequenced and annotated. A) Size distribution of Small RNA-seq libraries

immunoprecipitated with TEX19 and TEX19.1 from WT adult testes, and with TEX19 from

Tex19.1-ko. These libraries exhibit a peak around 29-30nt, which is reminiscent of MIWI-

associated piRNAs ; in contrast, MILI-associated piRNAs range in size between 26-27nt. (B)

Genomic annotation of TEX19.1-associated reads compared to MILI and MIWI-associated

reads. In agreement with piRNA size, the analysis was restricted to 25-32nt long reads. (C)

The vast majority of TEX19.1 reads have unique hits on the genome, similarly to MILI and

MIWI-associated Small RNAs. (D) Venn diagram distribution of piRNA cluster defined by

the piRNA population associated with TEX19.1 compared to MILI and MIWI. (E) Size

distribution of whole adult testes Small RNA-seq libraries obtained from a WT and three

Tex19.1-KO animals. The 22-23nt peak corresponds to microRNAs, the piRNAs are clustered

in two peaks, at 26-27nt and at 29-30nt.

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

Fig 7: Tex19 paralogs interact directly with ~26-30 nt piRNA: (A) and (B) Electrophoretic

Mobility Shift Assay “EMSA” was performed in the presence of 5µM of GST-alone, GST-

TEX19.1, GST-D1 (1-124aa), GST-D2Cter (125-351aa), GST-D2 (125-163aa “VPTEL”) or

GST-Cter (164-351aa) of TEX19.1 or GST-TEX19.2. All proteins were incubated with 125

femtomoles (8000cpm) of radiolabeled piR-117061 or modified piR-117061. (C) An

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

increased concentration of the modified non-radiolabeled piR-117061(2C) was used with a

fixed concentration (5µM) of TEX19.1 VPTEL domain and fixed quantity of radiolabeled

piR-117061 (8000 cpm). (D) An increased concentration of the non-radiolabeled piR-117061

was used with a fixed concentration of the VPTEL domain of TEX19.1 (5 µM) and a fixed

quantity of radiolabeled piR-117061 (8000 cpm).

Jour

nal o

f Cel

l Sci

ence

• A

dvan

ce a

rtic

le

DWT DKO

Adu

ltp1

2

A B

DWT DKO

C

Fig S1: (A) and (B) Histological sections stained with hematoxylin andeosin through the caudal epididymis of adult mice, note absence ofsperm in the mutant epididymis (C) TUNEL assays: the positive signalwas converted to a red false color and superimposed with the DAPInuclear stain (blue false color); Note a normal, low, proportion ofapoptotic germ cells in DWT testes but an increase in the proportion ofTUNEL- positive germ cells in Tex19DKO testes (black arrows).

J. Cell Sci. 130: doi:10.1242/jcs.188763: Supplementary information

Jour

nal o

f Cel

l Sci

ence

• S

uppl

emen

tary

info

rmat

ion

0

5

10

15

20

25

L1ORF2 expression in testis at p16

0

1

2

3

4

5

6

IAP∆1 expression in testis at p16

0

5

10

15

20

25

MuERV1 expression in testis at p16

DWT DKO(n=3) (n=4)

0

50

100

150

200

250

300

350

400

450

MMERVK10C expression in Tex19-/-

testes at P16

0

50

100

150

200

250

300

MMERVK10C expression in Tex19-/-

testes at P10

* *

DWT DKO(n=2) (n=3)

DWT DKO(n=2) (n=3)

DWT DKO(n=2) (n=3)

DWT DKO(n=2) (n=3)

Fig S2: RT-qPCR of MMERVK10C, L1ORF2, IAP∆1 and MuERV1retrotransposon expression in Tex19DKO (DKO) p10 and p16 micetestes compared to double wildtype (DWT). Expression levels werenormalized to the β-actin or Rrm2 housekeeping gene, n valuesindicate the number of biological replicates, DWT: double wild type;DKO: double knockout,*p<0.03 (Student t-test).

J. Cell Sci. 130: doi:10.1242/jcs.188763: Supplementary information

Jour

nal o

f Cel

l Sci

ence

• S

uppl

emen

tary

info

rmat

ion

GST

-TE

X19

.1

G-s

epha

rose

be

ads

GST

28

36

557295

130150

kDa

GSTTEX19.1 full length

GST-D1 1-124 aa GST-D2 125 -351 aa

BI AI 1 2 3

BI AI 1 2 3 BI AI 1 2 3

BI AI 1 2 3

GST

GSTVPTEL 125-163 aa

28

36

557295

130150

GSTTEX19.2 full lengh

Tex19.2

BI 1 2 3

VPTELD1

D2

GST

Tex19.1

kDa

28

36

557295

130150

kDa

B

A

Fig S3 : (A) Silver staining of GST-Tex19.1 pull down using G-sepharose beads. G-sepharose beads and GST combined to G-sepharose beads are deposited as control. (B) Production andpurification of GST, GST-TEX19.1, GST-TEX19.2, GST-MCP (1-124aa), GST-VPTEL Cter (125-351aa) or GST-VPTEL (125-163aa) ofTEX19. Gel 10% and bleu Coomassie coloration were used.

J. Cell Sci. 130: doi:10.1242/jcs.188763: Supplementary information

Jour

nal o

f Cel

l Sci

ence

• S

uppl

emen

tary

info

rmat

ion

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

U1 A10

VPTEL (125-163aa)

D2(125-351aa)

MCPP VPTELTEX19.1

MCPPD1(1-124aa)

VPTEL

VPTEL

Cter(164-351aa)

M C P P V S V R H G A R G M S C L Y E A W L Y H L V H G E Q T K I C F A C F K A A F

L L N K L Y L E M G D W Q E E E E E E E E E D A D L L E Y L S E S E S E S E Q E P G P

E Q D A W R G L G S L Y V P Q S V S E G S G V L L P T P V W T Q G I L F S I F V P T E L

F P Q E A V P L D L G P E D A E W T Q A L P W R L D G L F P C S H Q L I P P L T W W D

I F D V M P S P G Q P V L L E L R C H W P L D Q T V A Q S W L Q D Q K F V L L L D S

V Q S R C H L L S M R V R W V V R T Q V Q H W Q V L L D P G E M W V A H F R K E V

G Q H G L Y H Q S L N P W R L S I L T A S E L G M E L L P A T C Y L W N K G F W V G

S F L P W H I N M P E T W S W E P G E R L F I T D A T I C G T D Y H L A Q S F L D S H

P T P H P L L T L T P

TEX19.1

MC

PV

PT

EL

94.42%

30%

TEX19.1 piRNA library

Fig S4: A: Percentage of the distribution of TEX19.1 piRNA. TEX19.1

J. Cell Sci. 130: doi:10.1242/jcs.188763: Supplementary information

Jour

nal o

f Cel

l Sci

ence

• S

uppl

emen

tary

info

rmat

ion

Adapter triming 15-45bp Alignment, 2 mismatches

Sample Out percent

Input Too-short Adapter Non-clipped

Output Aligned reads

Failed Suppressed-m

Output reads

Tex19.1 in WT (1)

27.68% 6985974 4980436 3241 68727 1933570 1647203 (85.19%)

283557 (14.66%)

2810 (0.15%)

15875788

Tex19.1 in WT (2)

22.64% 6368438 4826897 2150 97851 1441540 1214051 (84.22%)

225055 (15.61%)

2434 (0.17%)

12610662

Tex19.2 in Tex19.1-/-

78.52% 15596146 3035682 176061 139045 12245358 9700101 (79.21%)

2524130 (20.61%)

21127 (0.17%)

117286602

Tex19.1 & 2 in WT

72.35% 13615661 3590834 23129 150175 9851523 8202474 (83.26%)

1637355 (16.62%)

11694 (0.12%)

83224246

Table S1: Adapter trimming and alignment

piRNA name Absolute count Tex19.1 in WT (1)

Absolute count Tex19.1 in WT (2)

piRNA sequence

gi|109016873|gb|DQ727168.1|piR-142490 5367 4008 ACUCGAGACUGUGUAGAACCUUCAUAGGCgi|108994920|gb|DQ710717.1|piR-126039 3828 2505 UUUGAAAUCAGGACUGUUCUGGAAAUUCCgi|108994167|gb|DQ710038.1|piR-125360 3009 2480 CUCCAGAUUAAAAAAUGUCCTAAGUUAGAGCUgi|108955533|gb|DQ687738.1|piR-103060 2742 2323 UGAAGUGAACCUGAGCAAUUCCAUGUGGA

gi|109001017|gb|DQ715342.1|piR-130664 2684 1962 UAUGAGUAGAACAAAUGGGGGCAAACCCUAGCgi|108984009|gb|DQ702024.1|piR-117346 2363 1839 UGGAUAAUAUUCCGGAUAAUAAACUGAGC

gi|108983531|gb|DQ701739.1|piR-117061 2361 2028 UAGGACCUGAGAACUUAACCUUGUUAUGGGgi|109013632|gb|DQ724840.1|piR-140162 2180 1319 UCUAGGUCUCUGAAAUUUACUAUGCACAGGGUgi|108036260|gb|DQ564985.1|piR-5073 2015 1516 UACCUCGGUCUCUGUAGAAUCCAGUUUUgi|108995087|gb|DQ710862.1|piR -126184 1989 1397 UACAGACUGAUAAACGAUUUAAAGUAAAAAGU

gi|108958333|gb|DQ688999.1|piR-104321 1635 1122 UACUGGAACACUGGAUUAACAGAACGUUCAgi|108954985|gb|DQ687488.1|piR-102810 1594 1421 UAAAUGGACUUUGGCCCCAGGUCAAUUCUGGC

gi|109012695|gb|DQ724202.1|piR-139524 1479 1030 UGAUAGAAUGUAGCCUCUGACCGGGAAUAgi|108951359|gb|DQ684704.1|piR-100026 1464 1009 UAGACUGUGAGCUCCUUGAGGACAGGGACAGUgi|108969977|gb|DQ694365.1|piR-109687 1399 1366 UACGAGAUCGUCCCUUAAAGCUCAAAGGACA

Table S2: Tex19.1 reads display a strong preference for a uridine at position one (U1-bias)

J. Cell Sci. 130: doi:10.1242/jcs.188763: Supplementary information

Jour

nal o

f Cel

l Sci

ence

• S

uppl

emen

tary

info

rmat

ion