MIWI N-terminal RG motif promotes efficient pachytene piRNA production and spermatogenesis independent of LINE1 transposon silencing

PIWI proteins and their associated piRNAs act to silence transposons and promote gametogenesis. Murine PIWI proteins MIWI, MILI, and MIWI2 have multiple arginine and glycine (RG)-rich motifs at their N-terminal domains. Despite being known as docking sites for the TDRD family proteins, the in vivo regulatory roles for these RG motifs in directing PIWI in piRNA biogenesis and spermatogenesis remain elusive. To investigate the functional significance of RG motifs in mammalian PIWI proteins in vivo, we genetically engineered an arginine to lysine (RK) point mutation of a conserved N-terminal RG motif in MIWI in mice. We show that this tiny MIWI RG motif is indispensable for piRNA biogenesis and male fertility. The RK mutation in the RG motif disrupts MIWI-TDRKH interaction and impairs enrichment of MIWI to the intermitochondrial cement (IMC) for efficient piRNA production. Despite significant overall piRNA level reduction, piRNA trimming and maturation are not affected by the RK mutation. Consequently, MiwiRK mutant mice show chromatoid body malformation, spermatogenic arrest, and male sterility. Surprisingly, LINE1 transposons are effectively silenced in MiwiRK mutant mice, indicating a LINE1-independent cause of germ cell arrest distinctive from Miwi knockout mice. These findings reveal a crucial function of the RG motif in directing PIWI proteins to engage in efficient piRNA production critical for germ cell progression and highlight the functional importance of the PIWI N-terminal motifs in regulating male fertility.

Pachytene piRNAs are generated by the piRNA biogenesis machinery from precursor transcripts derived from discrete genomic loci at the onset of meiosis [33].While pachytene piR-NAs contain a tiny fraction of transposon-related piRNAs, the overwhelming majority are non-transposon piRNAs derived from intergenic piRNA clusters.The biological function of these pachytene piRNAs still lacks consensus [21][22][23][34][35][36][37][38].Concomitant with pachytene piRNA biogenesis, MILI and MIWI are sequentially expressed and eventually loaded with almost the same sets of pachytene piRNAs [20,39].Unlike MILI, MIWI is not expressed in embryonic germ cells but is exclusively bound to meiotic pachytene piRNAs, starting from mid-pachytene spermatocytes to round spermatids [11,33].MIWI plays critical roles in transposon silencing and germ cell differentiation as Miwi knockout mice show drastically upregulated LINE1 transposons in spermatocytes and round spermatids and display germ cell arrest at the round spermatid step of spermiogenesis [11,27].MIWI slicer activity is required for LINE1 silencing as well as spermatid differentiation [11,27].Despite MIWI's important roles, it remains unclear whether transposon dysregulation causes germ cell arrest due to MIWI deficiency.
Here we use a conserved RG motif of MIWI as a prototype to study the in vivo function of RG motifs in mammalian PIWI proteins.We discovered that the MIWI N-terminal RG motif is crucial for MIWI function and spermatogenesis.Mutations in the N-terminal RG motif disrupt specific MIWI-TDRD interactions and impair pachytene piRNA biogenesis, resulting in male sterility.Importantly, the RG motif mutation separates the two defects of MIWI ablation: spermiogenesis arrest and LINE1 transposon activation.

MIWI N-terminal RG motif selectively interacts with TDRKH in vitro
The N-terminal domain of MIWI contains three RG motifs as TDRD protein binding sites (Fig 1A).We previously reported that TDRKH (TDRD2) specifically interacts with MIWI through the RG motif-1 (termed MIWI N-terminal RG motif thereafter) but not through RG motif-2 or RG motif-3, and this interaction is arginine methylation-independent [28,40].Since MIWI interacts with TDRD proteins and other piRNA biogenesis factors [30], we tested whether MIWI N-terminal RG motif also mediates interaction with other proteins.To abolish arginine-mediated protein interactions (methylation-dependent and methylation-independent), we constructed a Miwi arginine to lysine (Miwi RK ) mutation in which all six arginines (R) in the N-terminal RG motif were mutated to lysines (K) (Fig 1B).This mutation is not expected to change the charge state of MIWI RK mutant protein (MIWI RK ) and therefore minimally impacts the biophysical property of MIWI.Co-expression of FLAG-tagged MIWI or MIWI RK with different GFP-tagged piRNA pathway proteins in HEK293T cells were used to assess MIWI protein interactions by immunoprecipitation and Western blotting.While MIWI interacted with all previously reported MIWI-interacting proteins including TDRD1, TDRKH, RNF17 (TDRD4), TDRD5, TDRD6, STK31 (TDRD8), PNLDC1 and MOV10L1, the RK mutation significantly disrupted the interaction of MIWI RK with TDRKH (Fig 1C).The interaction of MIWI RK with other proteins was minimally affected (Fig 1C).TDRKH is a mitochondrial membrane anchored protein capable of recruiting cytoplasmic MIWI to mitochondria [41,42].Since the Miwi RK mutation strongly disrupted MIWI-TDRKH interaction, we reasoned that MIWI RK would fail to be recruited to mitochondria by TDRKH.We expressed GFP-MIWI or GFP-MIWI RK in HeLa cells, both showing diffused distribution in cytoplasm when singly transfected.When co-expressed with TDRKH-RFP, GFP-MIWI concentrated on mitochondria to significantly overlap with mitochondrial TDRKH-RFP (Fig 1D).However, GFP-MIWI RK still displayed diffused cytoplasmic distribution in the presence of TDRKH-RFP, indicating failure of recruitment by TDRKH to mitochondria (Fig 1E).Together, these data demonstrate that MIWI N-terminal RG motif is required for specific MIWI-TDRKH interaction and regulates MIWI localization to mitochondria to engage in piRNA biogenesis.

RK mutation in MIWI N-terminal RG motif disrupts spermatogenesis in mice
The in vivo function of RG motifs of mammalian PIWI proteins is unknown.We sought to use the MIWI N-terminal RG motif as a prototype to study the functional significance of RG motifs of PIWI proteins during piRNA biogenesis and spermatogenesis in mice (S1 Fig) .Using CRISPR-Cas9 genome editing, we generated a Miwi RK mutant allele in mice in which all six arginines in MIWI N-terminal RG motif were mutated to lysines (S1B Fig) .We confirmed these mutations in Miwi RK mutant mice by Sanger sequencing (S1C Fig) .By breeding the Miwi RK allele into Miwi null background (Miwi RK/-), we next examined whether the Miwi RK mutation affects spermatogenesis.Miwi RK/-mice are viable and grow normally but exhibited reduced testis mass compared to Miwi RK/+ and Miwi +/-control littermates (Fig 2A).Notably, Miwi RK/+ male mice are fertile, indicating that the Miwi RK mutation does not have a dominant-negative effect on germ cell development (Fig 2B).This sharply contrasts to the observed dominant-negative effect of the Miwi catalytically inactivating mutation (slicer mutation) on spermatogenesis, suggesting that the RK mutation did not affect MIWI catalytic activity [27].Histological examination of Miwi RK/-testes revealed that germ cells were arrested at the round spermatid stage and no elongating spermatids were formed (Fig 2B).As a result, numerous round spermatid-like cells were accumulated in epididymides of Miwi RK/-mice (Fig 2B).Together, these data demonstrate that genetic mutation in MIWI N-terminal RG motif blocks spermatogenesis and causes male sterility.
We next explored the nature of the presence of multinucleated giant cells in Miwi RK/-testes.We speculated that multinucleated giant cells in Miwi RK/-testes were being removed by Sertoli cells by phagocytosis, because the giant cells were not observed in the cauda epididymis.To test this, we performed immunostaining of α-tubulin that is highly expressed in Sertoli cell cytoplasm and found that multinucleated giant cells were surrounded by Sertoli cell cytoplasm in Miwi RK/-testes (S2 Fig) .This suggests that the formation of multinucleated giant cells is caused by phagocytic removal of spermatogenic arrested cells by Sertoli cells in Miwi RK/- testes.

MIWI RK fails to interact with TDRKH to localize to IMC despite normal initial expression
We next examined the effect of the RK mutation on MIWI mutant protein (MIWI RK ) expression and localization during spermatogenesis.As expected, MIWI RK in Miwi RK/-testes started to express in early pachytene spermatocytes, resembling the timing of wild-type MIWI (Fig 3A).At P16, MIWI RK expression level and localization pattern in Miwi RK/-testes was similar to that of wild-type MIWI in Miwi +/-testes, indicating that the RK mutation does not affect the initial expression and stability of MIWI (Fig 3A and 3B).However, MIWI RK displayed lower expression levels than wild-type MIWI in P18 and P20 testes as meiosis progresses (Fig 3A).In adult mice, MIWI RK in Miwi RK/RK and Miwi RK/-testes was expressed at a lower level than total MIWI in Miwi RK/+ and Miwi +/-testes, respectively (S3A Fig) .We observed a diffused distribution of MIWI RK in mid-late pachytene and diplotene spermatocytes (Figs 3C and S3A).MILI and TDRKH subcellular localization and expression levels, on the other hand, did not change in Miwi RK/-germ cells (S4 Fig) .The observed MIWI RK mislocalization correlates with the timing of pachytene piRNA production and the initial interaction with TDRKH.To further confirm the MIWI-TDRKH interaction is abolished by the RK mutation in vivo, we performed co-immunoprecipitation and Western blotting and showed that MIWI RK failed to interact with TDRKH in Miwi RK/-testes (Fig 3D).As a control, the interaction of MIWI RK with TDRD6 remained intact (Fig 3D), recapitulating our in vitro observation (Fig 1C) and suggesting the RK mutation selectively disrupts MIWI-TDRKH interaction in vivo.Transmission electron microscopy showed the IMC formation in Miwi RK/-spermatocytes is largely normal, undistinguishable from that in Miwi +/-or Miwi -/-(Fig 3E).We further co-stained MIWI with TDRKH or mitochondrial marker TOMM20 in stage VII-VIII pachytene spermatocytes.While MIWI and TDRKH/TOMM20 completely colocalized at the IMC in Miwi +/-control, MIWI RK only slightly overlapped with TDRKH/TOMM20 at the IMC in Miwi RK/- spermatocytes, with most of its distribution in the cytoplasm (Figs 3F and S3B).This suggests that the efficiency of MIWI RK being recruited by TDRKH to the IMC is severely compromised, which can be explained by the diminished MIWI RK -TDRKH interaction observed in Miwi RK/- testes (Fig 3D).Collectively, the RK mutation did not affect MIWI RK initial expression.Subsequently altered MIWI RK expression and localization are likely due to the loss of MIWI RG motif-mediated protein-protein interactions.

The N-terminal RG motif of MIWI regulates the efficiency of pachytene piRNA biogenesis
We then asked how the RK mutation affects MIWI in participating in piRNA production.We examined the abundance and size of piRNA populations in adult Miwi +/-and Miwi RK/-testes.Radiolabeling of total RNA revealed that total piRNA production was much reduced in Miwi RK/-testes (Fig 4A).We further sequenced these small RNAs from total RNA to show two piRNA populations corresponding to 25-28nt MILI-piRNAs and 29-32nt MIWI-piRNAs in Miwi +/-and Miwi RK/-testes.After normalizing to miRNA counts (21-23nt), the abundance of MILI-piRNAs was unaffected.However, the level of MIWI-piRNAs was significantly reduced in Miwi RK/-testes (Fig 4B).Despite this, the overall piRNA reduction was moderate compared with the severe piRNA loss in Stra8-Cre Tdrkh conditional knockout (Tdrkh cKO ) mice (Fig 4B).To confirm this trend, we analyzed MILI-piRNAs and MIWI-piRNAs by immunoprecipitation of MILI and MIWI followed by RNA labeling.MILI-piRNAs had the same abundance and lengths in Miwi RK/-testes compared to the Miwi +/-control (Fig 4C).Sequencing of MILI-piRNAs further confirmed that MILI-piRNA lengths were normal in Miwi RK/-testes, a clear difference from the extended untrimmed MILI-piRNAs observed in Tdrkh cKO testes (Fig 4D).In contrast, MIWI-piRNAs were significantly decreased in Miwi RK/-testes (Fig 4E).Notably, MIWI level from immunoprecipitation was also reduced in Miwi RK/-testes (Fig 4E).To verify that the piRNA binding ability of MIWI RK was not altered, MIWI immunoprecipitation was re-performed using the same amount of immunoprecipitated MIWI and MIWI RK for RNA labeling.MIWI RK was fully piRNA-loaded in Miwi RK/-testes compared to wild-type MIWI, suggesting that the RK mutation does not change the piRNA binding activity of MIWI RK (Fig 4F ).Sequencing of MIWI-piRNAs in Miwi RK/-testes confirmed that MIWI-piRNA lengths were not affected, indicating that the direct MIWI-TDRKH interaction is not required for piRNA trimming (Fig 4G).
Further analysis of Miwi RK/-piRNAs revealed a strong 5'-end U-bias at the first nucleotides, suggesting that piRNA 5' formation is normal in Miwi RK/-mice (Fig 4H).When mapping piRNA reads to the mouse genome, total piRNA, MILI-piRNAs, and MIWI-piRNAs from Miwi RK/-testes were primarily mapped to piRNA clusters, resembling the wild-type distribution (Fig 4I).The decrease in piRNA amount in Miwi RK/-testes distributed evenly across top pachytene piRNA producing loci (S5 Fig) .This indicates that the RK mutation does not affect the selection of piRNA precursors for processing.Taken together, these data demonstrate for the first time that the MIWI N-terminal RG motif is required for piRNA biogenesis, and the RK mutation results in decreased efficiency of piRNA production without impacting piRNA loading and maturation.

Chromatoid body malformation and round spermatid arrest in Miwi RK/- mice are independent of DNA damage and apoptosis
MIWI is predominately localized to the chromatoid body (CB) in round spermatids.We next examined the fate of MIWI RK in Miwi RK/-round spermatids.Despite the presence of MIWI RK in the CB, transmission electron microscopy revealed CB fragmentation in Miwi RK/-round spermatids (Fig 6A and 6B).The observed CB fragmentation is similar to Miwi -/-and represents a common defect associated with low MIWI expression (Fig 6B).Thus, the RK mutation causes defective CB formation in round spermatids which can be explained by inefficient piRNA biogenesis.
Miwi -/-round spermatids arrest at step 1-3 and undergo DNA damage and apoptosis [11].We then investigated whether DNA damage and/or apoptosis contribute to the late step 8 round spermatid arrest observed in Miwi RK/-testes.While γH2AX positive foci were abundant in Miwi -/-round spermatids indicative of DNA damage, they were not present in Miwi RK/- round spermatids (Fig 6C).TUNEL assay further showed no evident apoptosis in Miwi RK/- round spermatids in contrast to elevated apoptosis observed in Miwi -/-testes (Fig 6D).These data collectively indicate that despite causing piRNA defects, the RK mutation does not induce DNA damage or apoptosis in male germ cells, which is a significant difference from Miwi -/- mice.

Discussion
Our results provide the first evidence that an RG motif of a mammalian PIWI protein is functionally important for piRNA biogenesis, spermatogenesis, and male fertility.Importantly, the Miwi RK mutant mouse model genetically decouples MIWI's function in LINE1 transposon silencing and its role in promoting spermiogenesis (Fig 6E).MIWI N-terminal RG motif is required for efficient piRNA production, suggesting a critical role for this conserved motif in other vertebrate PIWIL1-like proteins to regulate the process of spermiogenesis, the last step of spermatogenesis in diverse species (S1A Fig) .MIWI N-terminal RG motif is required for MIWI-TDRKH interaction.We show here that MIWI-TDRKH interaction is disrupted by the Miwi RK mutation in germ cells.Consequently, MIWI RK fails to be efficiently recruited to the IMC and MIWI RK -bound piRNA production is much reduced.Despite a reduction in piRNA levels, piRNA processing and trimming are complete because there were no obvious length changes for MIWI RK -bound piRNAs, suggesting that MIWI RK can engage in piRNA processing and trimming from start to finish.This is consistent with our observation that the RK mutation in MIWI does not affect TDRKH anchoring on the mitochondrial surface for piRNA trimming.The accessibility of MIWI RK to the piRNA trimming complex can be explained by its residual weak interaction with TDRKH or its direct association with the trimmer PNLDC1 [43,44].Alternatively, other piRNA biogenesis factors may compensate for the recruitment of MIWI RK to the IMC for piRNA processing.Nonetheless, mature piRNAs associated with MIWI RK still follow the abundance pattern derived from piRNA clusters found in wild-type germ cells, indicating it is the piRNA production efficiency but not piRNA precursor selectivity that is affected by the RK mutation.
One surprising finding is that the MIWI N-terminal RG motif is required for primary piRNA biogenesis but not required for LINE1 silencing.This genetically separates MIWI's transposon silencing function from its ability to promote spermiogenesis.The maintenance of LINE1 silencing in Miwi RK mutants is likely due to the slightly reduced but functional LINE1-targeting piRNAs and the normal Ping-Pong signature in mutant germ cells, which differs from Miwi -/-and MIWI slicer mutants that both show significant LINE1 activation [27].Thus, the round spermatid arrest in Miwi RK/-mutants can be dissociated from LINE1 silencing and possibly results from reduction of the bulk of pachytene piRNAs primarily derived from transposon-poor pachytene piRNA clusters.The fragmented CB in Miwi RK/-round spermatids can also be ruled out to be a direct consequence of defective LINE1 silencing.In another genetic model, conditional deletion of piRNA biogenesis factor Mov10l1 by Stra8-Cre causes severe reduction in both MILI-piRNAs and MIWI-piRNAs and spermiogenic arrest without impacting LINE1 silencing, suggesting the critical roles of non-transposon-targeting pachytene piRNAs in promoting spermiogenesis [37].
MIWI/PIWIL1's overall function is carried out by the coordinated actions of its functional domains.The RK mutation abolishes the function of the N-terminal RG motif, while the folding and function of other major domains, PAZ, MID, and PIWI, are not expected to be altered.Indeed, the piRNA-binding ability mediated by the PAZ and MID domains of MIWI RK is unaffected by the RK mutation.The slicer activity conferred by the PIWI domain is also retained by MIWI RK as LINE1 remains silenced in Miwi RK/-mice.A D-box motif in the N-terminal region of MIWI involving ubiquitination has been shown to be important for male fertility [45,46].Mice with D-box mutations display a late spermiogenesis defect differing from that of Miwi RK/-mice, highlighting the designated roles of various functional domains/motifs in directing MIWI function during spermatogenesis.PIWI proteins interact with the TDRD family proteins through arginine-rich motifs and the PIWI-TDRD interaction is conserved among species.However, the binding mechanism varies across species.In mice, TDRKH interacts with the MIWI N-terminal RG motif independent of arginine methylation [28,40].While in Drosophila, Papi, the homolog of TDRKH, interacts with Piwi in an arginine methylation-dependent manner [47].Drosophila PIWI protein Aub interacts with Tudor domain-containing proteins Tud and Krimper through a N-terminal RG motif similar to that found in mouse MIWI2 (PIWIL4) [48,49].By replacing the four arginines in the N-terminal RG motif to lysines, the Aub RK mutant flies show high embryonic lethality and transposon upregulation but intact primary piRNA biogenesis [50,51].By contrast, here we demonstrate that the mouse MIWI N-terminal RG motif is required for spermiogenesis by promoting piRNA biogenesis independent of LINE1 silencing.Our results highlight the functional heterogeneity of RG motifs in different PIWI proteins across diverse species.In mice, the function of the RG motifs in MILI and MIWI2 to regulate piRNA biogenesis, spermatogenesis, and fertility requires further investigation.

Ethics statement
All animal procedures were approved by the Institutional Animal Care and Use Committee of Michigan State University.All experiments with mice were conducted ethically according to the Guide for the Care and Use of Laboratory Animals and institutional guidelines.

Generation of Miwi RK mutant mice
Miwi RK mutant mice were generated by CRISPR-Cas9 targeting of the mouse Piwil1 locus (ENSMUSG00000029423). Wild-type NLS-Cas9 protein, synthetic tracrRNA, crRNA and a single-stranded oligodeoxynucleotide (ssODN) donor template were purchased from Integrated DNA Technologies (Coralville, IA, USA).Protospacer (N) 20 and PAM sequences corresponding to the crRNA used were 5'-TGACTGGCCGAGCCCGAGCT -CGG -3'.Donor ssODNs in the reverse orientation had the following sequence: 5' AGCTATATAAGAATGG TACTCACCGCAGCAGCCCCAACATGCTGCACCGTCTCCTGACCcttTGCCtTGCCcttG GCCttAGCcttGGCcttGCCAGTCATTTTCTGTCAGAGAGGAAAAGCACACGA 3'.Synthetic tracrRNA and crRNA were incubated at 95˚C for 5 min and allowed to cool down in order to form RNA heteroduplexes, which were then incubated with Cas9 protein for 5 min at 37˚C to pre-form ribonucleoprotein (RNP) complexes.RNPs were electroporated into mouse zygotes using a Gene Editor electroporator (BEX CO., LTD, Tokyo, Japan) as previously described [52].Embryos were implanted into pseudo-pregnant recipients according to standard procedures.Editing of founder offspring was assessed using PCR, T7 Endonuclease I assay, and Sanger sequencing of the target region.

Plasmid construction
The full-length mouse Tdrd1, Rnf17, Tdrd5, Tdrd6, Stk31, Pnldc1 and Mov10l1 cDNAs were amplified by PCR and cloned into pEGFP-C1 (GFP tag at N-terminus) expression vector.The full-length mouse Tdrkh cDNA was amplified by PCR and cloned into pEGFP-N1 (GFP tag at C-terminus) expression vector.The full-length mouse Miwi and Miwi RK mutant cDNAs were amplified by PCR and cloned into pcDNA3-FLAG (FLAG tag at N-terminus) and pEGFP-C1 (GFP tag at N-terminus) expression vectors.To obtain RFP-tagged TDRKH, the GFP tag of pEGFP-N1 vector was replaced by TurboRFP tag and the full-length mouse Tdrkh cDNA was cloned into this TurboRFP-tagged (TurboRFP tag at C-terminus) expression vector.

Histology
Mouse testes and epididymides were collected and fixed in Bouin's fixative at 4˚C overnight and embedded in paraffin.Histological sections were cut at 5 μm and stained with hematoxylin and eosin after dewaxing and rehydration.

TUNEL assay
To evaluate apoptosis, TUNEL assay was performed on testis sections using In Situ Cell Death Detection Kit (11684795910, Sigma-aldrich) according to manufacturer's instruction.Briefly, 5 μm PFA-fixed paraffin sections were dewaxed and rehydrated, followed by incubating with Proteinase K for 20 min at room temperature.After washing with PBS, sections were incubated with TUNEL reaction mixture at 37˚C for 1 h and mounted using Vectashield mounting media with DAPI.

Transmission electron microscopy
Mouse testes were fixed with 2.5% glutaraldehyde and 2% PFA in 0.1 M cacodylate buffer for 2 h at room temperature.After washing, the testes were post-fixed with 1% osmium tetroxide in 0.1 M cacodylate buffer for 2 h at room temperature.The testes were dehydrated in increasing concentrations of acetone and then infiltrated and embedded in Spurr's resin.Ultrathin sections were cut at 70 nm and stained with uranyl acetate and lead citrate.Images were taken with JEOL 100CX Transmission Electron Microscope (Japan Electron Optics Laboratory, Japan) at an accelerating voltage of 100kV.

Immunoprecipitation of piRNAs and proteins
Mouse testes were collected and homogenized using lysis buffer (20 mM HEPES pH 7.3, 150 mM NaCl, 2.5 mM MgCl 2 , 0.2% NP-40, and 1 mM DTT) with protease inhibitor and RNase inhibitor.The lysates were pre-cleared using protein-A agarose beads for 2 h at 4˚C.Anti-MILI (PM044, MBL) or anti-MIWI (2079, Cell Signaling Technology) antibodies together with protein-A agarose beads were added to the lysates and incubated for 4 h at 4˚C.The beads were washed in lysis buffer 5 times.Immunoprecipitated RNAs were isolated from the beads using Trizol reagent for piRNA labeling or small RNA library construction.For protein detection, immunoprecipitated beads were boiled in protein loading buffer for 5 min.Western blotting of MILI, MIWI, TDRKH or TDRD6 was performed as described above.

Detection of piRNAs
Total RNA was extracted from mouse testes using Trizol reagent (Thermo Scientific).Total RNA or immunoprecipitated RNA (MILI or MIWI) was de-phosphorylated with Shrimp Alkaline Phosphatase (NEB) and end-labeled using T4 polynucleotide kinase (NEB) and [γ-32P] ATP.The 32P-labeled RNA was separated on a 15% Urea-PAGE gel, and signals were detected by exposing the gel on phosphorimager screen followed by scanning on the Typhoon scanner (GE Healthcare).

Small RNA libraries and bioinformatics
Small RNA libraries from immunoprecipitated RNAs or total RNA were prepared using Small RNA Library Prep Kit (E7300, NEB) following manufacturer's instructions.Multiple libraries with different barcodes were pooled and sequenced with the Illumina HiSeq 4000 platform (MSU Genomic Core Facility).

In situ hybridization
Testes were fixed in 4% PFA overnight at 4˚C.After being immersed in 30% sucrose, testes were embedded in O.C.T compound and frozen before 7 μm sections were cut.Antisense DIG labeled RNA probe was transcribed using DIG RNA Labeling Mix (Roche) from a linearized plasmid containing a full length of LINE1 Orf1 (nucleotides 1741-2814, GenBank: M13002.1).After denaturing the probes for 10 min in hybridization cocktail solution (Amresco), the probes were added to the sections and incubated overnight at 65˚C.After washing and blocking, sections were incubated with alkaline-phosphatase conjugated goat anti-DIG Fab fragments (Roche) overnight.The positive signal was visualized by adding BM Purple (Roche).

Statistical analysis
All data are mean±SEM and all statistical analyses between groups were analyzed by unpaired t-test.

Fig 1 .
Fig 1. MIWI N-terminal RG motif selectively interacts with TDRKH.(A) Schematic illustration of the domain architecture of mouse MIWI.(B) Protein sequences of the N-terminal RG motif (RG motif-1 in (A)) and its RK mutation are shown.(C) The RK mutation (RK) in MIWI N-terminal RG motif selectively diminishes the interaction of MIWI with TDRKH.HEK293T cells were transfected with indicated plasmids.Immunoprecipitation was performed using anti-FLAG resin.GFP-tagged and FLAG-tagged proteins were detected by Western blotting with anti-GFP and anti-FLAG antibodies.(D) TDRKH recruits MIWI to mitochondria.HeLa cells were transfected with GFP-tagged MIWI alone or together with RFP-tagged TDRKH plasmids.After 48 h, the cells were fixed and DNA was stained with DAPI.Scale bar, 20 μm.(E) The RK mutation diminishes the recruitment MIWI to mitochondria by TDRKH.HeLa cells were transfected with GFP-tagged MIWI RK alone, or together with RFP-tagged TDRKH plasmids.After 48 h, cells were fixed and DNA was stained with DAPI.Scale bar, 20 μm.Results shown in (C)-(E) are representative of 3 biological replicates.https://doi.org/10.1371/journal.pgen.1011031.g001

Fig 3 .
Fig 3.The N-terminal RG motif is required for recruitment of MIWI to the IMC in spermatocytes.(A) Western blotting of MIWI at different postnatal (P) ages Miwi +/-and Miwi RK/-testes.β-actin is a loading control.Asterisks indicate non-specific bands.Quantification of intensity of MIWI is shown under the blot (the one in Miwi +/-testis of each group is set as 1.000 after normalization with β-actin).(B) Testes from P16 mice were co-immunostained using MIWI and γH2AX antibodies.DNA was stained by DAPI.Scale bar, 20 μm.(C) The RK mutation causes mislocalization of MIWI in late spermatocytes.Testes from indicated mice were immunostained using MIWI and γH2AX antibodies.DNA was stained by DAPI.Different cell types were distinguished according to γH2AX staining and DAPI staining.Scale bar, 5 μm.(D) The RK mutation selectively disrupts the interaction of MIWI with TDRKH.Immunoprecipitation was performed using anti-MIWI antibody.MIWI, TDRKH and TDRD6 were detected by Western blotting.(E) The RK mutation in MIWI does not affect the formation of the IMC in spermatocytes.Transmission electron microscopy was performed on pachytene spermatocytes (PS) from indicated testes.The IMC region is zoomed and indicated by red arrows.Scale bar, 2 μm.(F) The RK mutation diminishes the recruitment of MIWI to mitochondria in pachytene spermatocytes.Co-immunostaining of MIWI and TDRKH in stage VII-VIII seminiferous tubule from Miwi +/-and Miwi RK/- testes was performed.DNA was stained by DAPI.Scale bar, 10 μm.Results shown in (A)-(F) are representative of 3 biological replicates.https://doi.org/10.1371/journal.pgen.1011031.g003

Fig 4 .
Fig 4. MIWI N-terminal RG motif is required for efficient pachytene piRNA biogenesis.(A) The RK mutation in MIWI N-terminal RG motif causes reduction of total piRNA in adult testes.Total RNA from testes of indicated genotypes were end-labeled with [32P]-ATP and detected by TBE urea gel and autoradiography.18S and 28S ribosomal RNAs served as loading controls.(B) The length distribution of small RNAs from adult testes.Data were normalized by miRNA reads (21-23nt).(C) The RK mutation in MIWI does not affect MILI-piRNA biogenesis.Small RNAs were isolated from immunoprecipitated MILI-RNPs, end-labeled with [32P]-ATP and detected by TBE urea gel and autoradiography.Western blotting was performed using anti-MILI antibody after immunoprecipitation. (D) The length distribution of MILI-piRNAs from adult testes of indicated genotypes.(E) The RK mutation in MIWI causes reduction in MIWI-piRNAs.Small RNAs were isolated from immunoprecipitated MIWI-RNPs, end-labeled with [32P]-ATP and detected by TBE urea gel and autoradiography.Western blotting was performed using anti-MIWI antibody after immunoprecipitation. (F) MIWI RK are properly loaded with piRNAs.Small RNAs were isolated from immunoprecipitated MIWI-RNPs, end-labeled with [32P]-ATP and detected by TBE urea gel and autoradiography.MIWI protein levels after immunoprecipitation served as loading controls.(G) The length distribution of MIWI-piRNAs from adult testes of indicated genotypes.(H) Nucleotide distributions at the first position in total piRNA, MILI-piRNAs and MIWI-piRNAs from adult testes of indicated genotypes.(I) Genomic annotation of total piRNA, MILI-piRNAs and MIWI-piRNAs from adult testes of indicated genotypes are shown.Sequence reads from indicated libraries were aligned to mouse genomic sequence sets in the following order: piRNA clusters, coding RNA, non-coding RNA, repeats, intronic sequences and other.Results shown in (A), (C), (E) and (F) are representative of 3 biological replicates.Small RNA-seq results shown in (B), (D), (G), (H) and (I) are representative of 2 biological replicates.https://doi.org/10.1371/journal.pgen.1011031.g004

Fig 5 .
Fig 5. MIWI RK mutation does not cause LINE1 de-repression.(A) LINE1 ORF1 mRNA was upregulated in Miwi -/-testes but was undetectable in Miwi RK/-testes.In situ hybridization of LINE1 ORF1 mRNA was performed in adult testes.Scale bar, 40 μm.(B) LINE1 ORF1 protein was upregulated in Miwi -/-testes but was undetectable in Miwi RK/-testes.Immunostaining was performed using LINE1 ORF1 antibody on adult testes.DNA was stained with DAPI.Scale bar, 20 μm.(C) The MIWI RK mutation does not affect the Ping-Pong signature of LINE1-derived piRNAs.The 5 0 -5 0 overlaps (Ping-Pong signature) between MIWI-piRNAs from opposite strands of LINE1 elements from Miwi +/-and Miwi RK/-testes were shown.The percentage of pairs of piRNA reads at each position is reported.(D) Graphs showing the distribution of MIWI-piRNAs that were mapped in the sense and antisense orientations to LINE1.Reads were normalized with the total reads from each library.Results shown in (A) and (B) are representative of 3 biological replicates.Small RNA-seq results shown in (D) are representative of 2 biological replicates.https://doi.org/10.1371/journal.pgen.1011031.g005

Fig 6 .
Fig 6.MIWI RK mutation causes chromatoid body malformation but does not induce DNA damage and apoptosis.(A) MIWI RK localization to chromatoid body in round spermatids.Testes from indicated mice were immunostained using MIWI antibody.DNA was stained by DAPI.Scale bar, 5 μm.(B) The RK mutation in MIWI causes chromatoid body fragmentation in round spermatids.Transmission electron microscopy was performed on round spermatids (RS) from Miwi +/-, Miwi RK/-, and Miwi -/-testes.The chromatoid bodies are indicated by dotted line.Scale bar, 1 μm.(C) MIWI RK mutation does not cause DNA damage in arrested round spermatids.Adult testes were immunostained using γH2AX antibody.DNA was stained by DAPI.Scale bar, 20 μm.(D) MIWI RK mutation does not cause apoptosis in testes.TUNEL assays were performed in adult testes.DNA was stained by DAPI.Scale bar, 40 μm.(E) A proposed model illustrating the role of the MIWI N-terminal RG motif in spermiogenesis.Results shown in (A)-(D) are representative of 3 biological replicates.https://doi.org/10.1371/journal.pgen.1011031.g006