Argonaute catalytic activity is required for maternal mRNA clearance in embryos

: Argonaute proteins and their interacting small RNAs play a key role in regulating complementary mRNA targets during animal development. Here, we investigate a novel and essential function of the catalytically active Argonaute protein CSR-1 in maternal mRNA degradation during early embryogenesis in Caenorhabditis elegans . We show that CSR-1 interacts with endogenous small RNAs antisense to hundreds of cleared maternal mRNAs in embryos, and preferentially cleaves mRNAs no longer engaged in translation. The depletion of CSR-1 during maternal to zygotic transition leads to embryonic lethality in a catalytic-dependent manner and impairs the degradations of its embryonic mRNA targets. Given the conservation of Argonaute catalytic activity, we propose that a similar mechanism operate to clear maternal mRNAs during maternal to zygotic transition across species.

4 pathway are also maternally inherited in the embryos (14,20). However, a specific function of CSR-1 during early embryogenesis remains unknown.
To precisely study the localization of CSR-1 during embryogenesis and in animal development, we have generated different CRISPR-Cas9 tagged versions of CSR-1 (22). As previously observed (23), CSR-1 is expressed mainly in the germline of adult worms and localizes in the cytoplasm and germ granules (Fig. S1 A, B). However, in early embryos, CSR-1 localizes both in germline and somatic blastomeres (Fig. 1A, and Fig. S1B). The localization of CSR-1 in the somatic blastomeres is mainly cytoplasmic and persists for several cell divisions until it becomes restricted to the perinuclear region of the germline blastomeres (Fig. 1A, and Fig. S1B).
This is in contrast with other germline Argonaute proteins (24,25), such as PIWI, which exclusively segregate with the germline blastomere from the first embryonic cleavage (Fig. 1A, and Fig. S1B). To study the specific role of CSR-1 during embryogenesis, we decided to deplete CSR-1 protein specifically during oocyte and embryo production. For this purpose, we adopted the Auxin Inducible Degradation (AID) system (26) to trigger CSR-1 protein depletion by exposure to auxin at chosen developmental stages. We confirmed that depletion of CSR-1 starting from the first larval stage (L1) recapitulates the loss of fertility and embryonic lethality phenotypes previously observed in csr-1 (tm892) mutant worms (14) (Fig. 1B). Next, we depleted CSR-1 in adult worms starting from oogenesis (Fig. S2A). This treatment did not significantly decrease the fertility of the animals (number of eggs laid, Fig 1B), yet all the embryos failed to develop in larvae ( Fig. 1C). This result suggests that CSR-1 plays an essential role during embryogenesis. To test whether the catalytic activity of CSR-1 is essential for embryonic viability, we attempted to rescue the depletion of endogenous CSR-1 with a transgenic germline expression of CSR-1. For this purpose we introduced by mosSCI (27), a single copy version of CSR-1 with or without an alanine substitution of the first aspartate residues within the catalytic DDH motif (Fig. 1D). While the (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint this version posted February 3, 2020. . https://doi.org/10.1101/2020.02.03.919597 doi: bioRxiv preprint transgenic expression of CSR-1 catalytic dead mutant (ADH) failed to rescue the embryonic lethality caused by the oogenic specific depletion of CSR-1 (Fig. 1D), expression of a catalytic functional transgenic CSR-1 (DDH) in the same conditions rescued the embryonic lethality (Fig.   1D). These results demonstrate that CSR-1 catalytic activity is essential during embryogenesis. The peculiar localization of maternal-inherited CSR-1 in the somatic blastomeres during early embryogenesis suggests that CSR-1 and its interacting 22G-RNAs might regulate specific embryonic targets in the soma. To better understand the potential and specific regulatory function of CSR-1 in the embryo, we identified and compared CSR-1 -interacting 22G-RNAs in the embryo and adult worms. We immunoprecipitated CSR-1 in populations of embryos ranging from 1-to 20cells stage (Fig. S3A) or young adult worms followed by RNA-seq to detect interacting 22G-RNAs complementary to mRNA targets. Our analysis revealed that even though the majority of CSR-1-22G-RNA targets in the adult germlines and early embryos largely overlap (66 %) ( Fig. 2A), the relative abundance of 22G-RNAs on these targets is different between adult germlines and early embryos ( Fig. 2A, B). Indeed, only 31 % of CSR-1 targets with highest levels of small RNAs (top 10 %) in adult and embryos are shared ( Fig. 2A, B). Given that the impact of CSR-1 cleavage activity on target mRNAs depends on the abundance of 22G-RNAs (20), our results suggest that CSR-1 might regulate a different subset of mRNAs in the embryos. To evaluate gene expression changes upon CSR-1 depletion in embryos, we depleted CSR-1 in adult worms starting from the beginning of oogenesis and collected embryo populations fully depleted of CSR-1 in early stages of embryonic development (from 1-cell to 100-cell stages) ( Fig. S2A and Fig. S3B). The analysis of nascent transcription by Global Run-On sequencing (GRO-seq) revealed that zygotic transcription in CSR-1 depleted embryos was largely unaffected compared to control untreated embryos (Fig. 2C), indicating that embryos were able to normally initiate zygotic transcription in the absence of CSR-1. However, the analysis of the steady-state mRNA levels by total RNA-seq (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint this version posted February 3, 2020. . https://doi.org/10.1101/2020.02.03.919597 doi: bioRxiv preprint 6 revealed increased CSR-1 embryonic mRNA targets (Fig. 2D). Moreover, the level of upregulation of these target mRNAs correlated with the levels of complementary antisense 22G-RNAs loaded into CSR-1 (Fig. 2D). These results suggest that CSR-1 directly regulates these target mRNAs at the post-transcriptional level in a dose-dependent manner. To gain insight into the type of mRNA targets regulated by CSR-1, we decided to investigate whether CSR-1 embryonic targets corresponded to the maternal or the newly transcribed zygotic mRNA transcripts. For this reason, we adapted a previously developed sorting strategy (28) to precisely sort populations of embryos at 1-cell stage (99 % pure), or enriched from 4-to 20-cells (early embryos), or having more than 20-cells (late embryos) (Fig. S3C). We performed stranded-specific RNA-seq on these three sorted embryo populations to profile the dynamics of gene expression during early embryonic development. Following this approach, we were able to detect 7088 maternal mRNAs inherited in 1-cell embryos (TPM > 1) (Table S1). We also identified three different classes of genes. One class of genes, which we called maternal cleared, corresponded to 1633 genes whose maternal mRNAs are inherited in 1-cell embryos and their levels gradually diminished in early and late stages (  (Table S1). These genes belong to the previously characterized Maternal class II genes, which are maternally inherited in the embryos and degraded in somatic blastomeres (12). The second class of genes, which we called maternal stable, correspond to 1014 genes whose maternal mRNAs are inherited in 1-cell and their levels are stably maintained in early and late embryos (Fig. 3A) (Table   S1). Lastly, we identified 1201 genes whose mRNAs accumulate in early and late embryos but were undetectable in 1-cell embryos (TPM < 1), suggesting that they are zygotically transcribed ( Fig. 3A) (Table S1). Next, we analyzed whether the levels of embryonic CSR-1 targets followed one of these three categories of genes. Indeed, we could identify that CSR-1 embryonic mRNA targets are inherited in 1-cell embryos and gradually depleted in early and late embryos, indicating they are maternal mRNAs cleared in somatic blastomeres during embryogenesis (Fig. 3B). These (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint this version posted February 3, 2020. . https://doi.org/10.1101/2020.02.03.919597 doi: bioRxiv preprint results suggest that CSR-1 can contribute to the clearance of maternal mRNAs in somatic blastomeres. To prove this, we analyzed the levels of maternal cleared mRNAs in CSR-1 depleted embryos and found reduced degradation of maternal mRNAs that are also targets of CSR-1 in the embryo (Fig. 3C). Furthermore, the accumulation of cleared maternal mRNA targets, correlates with the abundance of antisense small RNAs loaded onto CSR-1 (Fig. 3C). To demonstrate that CSR-1 is directly regulating these maternal mRNA targets in embryos, we performed RNA-seq on CSR-1 depleted embryos enriched in 1-to 4-cells stage or more advanced embryos enriched in 4to 20 cells stage ( Fig. S3E and Materials and Methods). CSR-1 depletion in 1-to 4-cell embryos did not cause global changes in the levels of maternal cleared mRNA targets. However, CSR-1 depleted embryos enriched in 4-to 20-cells showed increased levels of maternal cleared mRNA targets compared to 1-to 4-cell embryos (Fig. 3D). These results demonstrated that CSR-1 and its interacting small RNAs actively degrade their maternal mRNA targets during early embryogenesis.
To confirm these results with an independent method we performed RNA Fluorescence In Situ Hybridization (RNA FISH) on a maternal cleared CSR-1 target in CSR-1 depleted and control embryos. As expected, we observed reduced maternal mRNA degradation of the CSR-1 target in advanced CSR-1 depleted embryos compared to control (Fig. S4). Collectively, these results suggest that the maternal inherited CSR-1 loaded with 22G-RNAs contribute to initiating the degradation of complementary maternal mRNA targets in somatic blastomeres. To better understand the dynamics of CSR-1-dependent mRNA clearance during embryogenesis, we further divided the maternal mRNA cleared genes in earlier and later degraded mRNAs. We selected 653 genes that show at least a 2-fold reduction in mRNA levels in a population of sorted embryos enriched in 4-to 20-cells compared to 1-cell embryos (earlier degraded) (Fig. 4A). Moreover, we selected 1399 genes that show stable levels of mRNAs in 4-to 20-cells compared to 1-cell embryos and that are decreased at least 2-fold in populations of embryos with more than 20 cells (later (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint this version posted February 3, 2020. . https://doi.org/10.1101/2020.02.03.919597 doi: bioRxiv preprint 8 degraded) (Fig. 4A). Because the profile of 22G-RNAs interacting with CSR-1 in the embryo has been performed in populations of embryos enriched from 1-to 20-cells, we hypothesized that CSR-1 targets would be enriched for earlier degraded maternal mRNAs. Indeed, the levels of 22G-RNAs targeting the earlier degraded mRNAs were significantly higher compared to the later degraded mRNAs in total RNAs and CSR-1 IPs (Fig. 4B, C). These results suggest that the abundance of 22G-RNAs correlates with the temporal decay of maternal cleared mRNAs during embryogenesis.
We also generated metaprofiles of CSR-1-interacting 22G-RNAs levels along the gene body of earlier and later degraded target mRNAs (Fig. 4D). We observed that the 22G-RNAs loaded into CSR-1 were enriched along the whole coding sequence of the early degraded mRNAs (Fig. 4D).
Instead, the later degraded mRNAs showed an enrichment of 22G-RNAs loaded into CSR-1 mainly at the 3' end of the genes in a region corresponding to the 3'UTR (Fig. 4D). Previous reports have shown that the rate of mRNA clearance during maternal to zygotic transition is influenced by translational efficiency (7,8). Thus, the different translational status of earlier and later degraded mRNAs may influence the accessibility and production of CSR-1-loaded 22G-RNAs antisense to the coding sequences of cleared mRNAs. To test this hypothesis, we measured ribosomal occupancy on maternal mRNAs by Ribo-seq in a population of embryos enriched in 4-to 20-cells ( Fig. S3D). Our analysis showed that earlier degraded maternal mRNA targets have very low ribosomal occupancy and translational efficiency compared to later degraded mRNA targets (Fig.   4E, F). These results suggest that the loading of ribosomes on maternal mRNAs can influence their coding sequence accessibility for the 22G-RNA-mediated cleavage by CSR-1 in somatic blastomeres. Previous studies have shown that the codon usage of mRNAs can influence their translational efficiency and maternal mRNA decay during maternal to zygotic transition (7,8).
However, the earlier and later degraded mRNAs in C. elegans embryos did not show any differences in their codon usage (Fig. S5). Because the translation of germline mRNAs is primarily (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint this version posted February 3, 2020. . https://doi.org/10.1101/2020.02.03.919597 doi: bioRxiv preprint regulated by their 3'UTR in C. elegans (29), we decided to investigate whether different 3'UTR from earlier and later degraded mRNAs can impact the translation and the rate of mRNA degradation in early embryos. For this reason, we generated a single-copy transgene expressing a germline mCherry::h2b mRNA reporter with a 3'UTR from an earlier (egg-6 3'UTR) or later (tbb-2 3'UTR) degraded mRNA targets (Fig. 4G). The mCherry::h2b reporter fused with egg-6 3'UTR was indeed poorly translated in early embryos compared to mCherry::h2b reporter fused with tbb-2 3'UTR. Moreover, the level of mCherry::h2b mRNA reporter fused with egg-6 3'UTR was significantly lower in the embryos even though its level was similar in adult worms, indicating a faster rate of degradation in early embryos (Fig. 4H). To prove that CSR-1 22G-RNAs contribute to the faster decay of mCherry::h2b reporter fused with egg-6 3'UTR, we sequenced small RNAs from the two transgenic strains in early embryo preparations. Our analysis showed increased levels of small RNAs on the coding sequences of the mCherry transgenic reporter with egg-6 3'UTR compared to the one containing tbb-2 3'UTR ( Fig. 4I, J). These results suggest that the translation level of maternal mRNAs can influence the rate of CSR-1-mediated mRNA cleavage in somatic blastomeres.
In conclusion, we have shown that the maternal inherited CSR-1 protein and its interacting 22G-RNAs trigger the cleavage of hundreds of complementary maternal mRNAs targets in somatic blastomeres. CSR-1 is the only catalytically active Argonaute protein with cleavage activity detected in C. elegans extract (30). We propose that CSR-1 catalytic activity is required during embryogenesis to cleave untranslated maternal mRNAs accumulated in somatic blastomeres after the first embryonic divisions. Catalytically active Argonaute proteins are conserved in animals, including humans. In Drosophila embryos, the Argonaute protein Aubergine and its interacting piRNAs have been shown to participate in maternal mRNAs clearance (3,4). However, it is still not known whether the catalytic activity of the protein is required for this process. The inheritance (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint this version posted February 3, 2020. . https://doi.org/10.1101/2020.02.03.919597 doi: bioRxiv preprint of maternal Ago2, the only catalytically active Argonaute in mouse, is essential for early embryonic mouse development, and it has been suggested to play a role in maternal to zygotic transition (31).
Moreover, endogenous small RNAs targeting protein-coding genes have been detected in the mouse oocyte and embryonic stem cells (32-34). Therefore, we propose that an Argonautemediated mRNA cleavage might be a conserved mechanism to clear maternal mRNAs during maternal to zygotic transition.   (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

C. elegans strains and maintenance
Strains were grown at 20 °C using standard methods (35). The wild-type reference strain used was Bristol N2. A complete list of strains used in this study is provided in Table S1.
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

3'UTR replacement experiment
A sequence of 800bp downstream of the stop codon of egg-6 was amplified from genomic DNA by PCR and used as 3' UTR. We used the single-copy transgene mCherry::his-11::tbb-2 3'UTR as entry strain to insert egg-6 3'UTR by CRISPR-Cas9 as described above. egg-6 3'UTR was inserted right after the STOP codon of his-11. The specific expression of the inserted UTR was validated by RT-qPCR.

Auxin-inducible depletion of CSR-1
Auxin-inducible depletion has been performed as described in (26). 250 mM Auxin stock solution was prepared in Ethanol and stored at 4°C. Auxin plates or Ethanol plates were prepared by the addition of Auxin or only Ethanol to NGM plates (final concentration: 500 µM auxin, 0.5 % Ethanol for Auxin plates and 0.5 % Ethanol for Ethanol plates). Plates were seeded with OP50 E.
Coli, stored at 4°C and warmed at room temperature before the experiment. Worms were placed on Auxin or Ethanol plates from L1 or at the beginning of oogenesis as explained for each experiment.
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

WT and CSR-1 mutants
Single L1 larvae were manually picked and placed onto NGM plates seeded with OP50 E. coli and grown at 20°C until adulthood and then transferred on a new plate every 24 hours for a total of 2 transfers. The brood size of each worm was calculated by counting the number of embryos and larvae laid on the 3 plates. For the auxin-depleted experiments NGM plates were supplemented with Ethanol (control) or Auxin.

Brood size of CSR-1 auxin-depleted worms during oogenesis
For CSR-1 depletion during oogenesis: brood size assay has been performed as described above with the exception that worms were grown from L1 on regular NGM plates and transferred onto Auxin or Ethanol plates 44h after hatching and transferred on new corresponding Auxin and Ethanol plates every 24 hours for a total of 2 transfers.
Embryonic lethality assay of CSR-1 rescue experiments Single L1 larvae of strains carrying Degron::CSR-1 complemented with the single-copy insertion Mex-5P::GFP::CSR-1::tbb-2 3'UTR or Mex-5P::GFP::CSR-1(ADH)::tbb-2 3'UTR were manually picked and placed onto NGM plates seeded with OP50 E. coli and grown at 20°C until adulthood. Since transgene was prone to silencing, adult worms were allowed to lay at least 65 embryos and then adults were removed from the plates and imaged to check for GFP::CSR-1 expression. Only plates from GFP expressing adults were used for the assay. The percentage of embryonic lethality is calculated by dividing the number of dead embryos for the total number of laid embryos.
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Collection of early embryos populations
Synchronous populations of worms were grown on NGM plates until adulthood and were carefully monitored using a stereomicroscope and bleached shortly after worms started to produce the first embryos. After bleaching, Early embryos were washed with cold M9 buffer to slow down embryonic development and immediately frozen in dry ice. A small aliquot of embryo pellet (2 µL) was taken right before freezing and mixed with 10 µL VECTASHIELD® Antifade Mounting Medium with DAPI (Vector laboratories) and immediately frozen in dry ice. DAPI stained embryos were defrosted on ice and used for counting cell nuclei and score the embryonic cell stage of the population.

Collection of CSR-1 depleted early embryos populations
Early embryos were collected as described above with the exception that worms were transferred on Auxin or Ethanol plates at the beginning of oogenesis.

Collection of early and late CSR-1 depleted embryo populations
Early populations of CSR-1 depleted embryos were collected as described above with the exception that harvesting was performed at the very beginning of embryo production to further enrich the population in early staged embryos. After bleaching, an aliquot of the same population was allowed to develop further for 1 to 2h in M9 containing 500 µM Auxin, 0.5 % Ethanol or 0.5 % Ethanol.
Embryonic cell stage was scored, and harvesting was performed when embryo populations reached the desired developmental stage.

sRNA-seq library preparation
Total RNA from staged embryo preparations with RIN>9 was used to generate sRNA libraries as previously described (22).
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

RNA IP
A synchronous population of 40,000 worms (48h after hatching) or a preparation of at least 150,000 early embryos was collected and suspended in extraction buffer (50 mM HEPES pH 7.5, 300 mM NaCl, 5 mM MgCl2, 10 % glycerol, 0.25 % NP-40, protease inhibitor cocktails (Thermo Scientific), 40 U/mL RiboLock RNase inhibitors (Thermo Scientific)). Samples were crushed in a metal dounce on ice performing at least 40 strokes. Crude protein extracts were centrifuged at 12,000 rpm at 4°C for 10 minutes. Protein was quantified using Pierce™ 660 nm Protein Assay Reagent (Thermo Scientific) and 1 mg (for adults) or 700 µg (for embryos) of protein extract was used for RNA immunoprecipitation as described in (22) and used for sRNA-seq library preparation.

GRO-seq on CSR-1 depleted embryos:
Populations containing at least 40,000 CSR-1 depleted early embryos were collected as described above. Early embryos were resuspended in 1,5mL Nuclei extraction buffer (3 mM CaCl2, 2 mM

Nuclear Run-On reaction and RNA extraction
Nuclear Run-On reaction was performed by addition of 100 µL NRO 2x buffer (10 mM Tris HCl, 5 mM MgCl2, 1 mM DTT, 300 mM KCl, 1 % Sarkosyl, 0.5 mM ATP, CTP and GTP and 0.8 U/µL RNase inhibitor) and using 1 mM Bio-11-UTP final concentration and incubation for 5 minutes at 30°C. NRO reaction was stopped by the addition of TRIzol LS reagent (Ambion) and RNA (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint this version posted February 3, 2020. . https://doi.org/10.1101/2020.02.03.919597 doi: bioRxiv preprint extraction was performed following the manufacturer's instructions. Purified RNA was fragmented by the addition of reverse transcriptase buffer and incubation for 7 minutes at 95°C.

Ligation of 3' and 5' Adapter Oligos and 2 nd and 3 rd Biotin enrichments
RNA was ligated to 3' end adapter using T4 RNA ligase 2 Truncated KQ (home-made) for 16h at 15˚C. After ligation RNA was purified using solid-phase reversible immobilization beads (SPRI beads) and biotinylated RNA was enriched as described above. After purification RNA was ligated at 5' end using T4 RNA ligase 1 for 2 h at 25˚C. RNA was purified using SPRI beads and biotinylated RNA was enriched for a third time as described above.

Reverse transcription and amplification of cDNA libraries
Purified RNA was reverse transcribed using SuperScript IV Reverse Transcriptase (Thermo Fisher Scientific) following manufacturer conditions except that reaction was incubated for 1h at 50˚C.
cDNA was PCR amplified with specific primers using Phusion High fidelity PCR master mix 2x (New England Biolab) for 18-20 cycles and sequenced on Illumina Next 500 system.
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Sorting of C. elegans embryos
Sorting of C. elegans embryos was performed as described in (28) with the following modifications: a strain expressing both mCherry::OMA-1 and PIE-1::GFP was used to collect embryos at three different developmental stages (see Fig. S3C); embryos after bleaching were fixed with 2% formaldehyde in M9 to block cell division. After sorting, embryos were reverse crosslinked in 250µL RIPA buffer with RNase inhibitors for 30 minutes at 70°C and RNA was extracted with TRIzol LS (Ambion) following manufacturer instructions.

Strand-specific RNA-seq library preparation
Strand-specific RNA-seq libraries were prepared as described previously (22).

RT-qPCR
1 µg DNase treated total RNA was used as a template for cDNA synthesis using random hexamers and M-MLV reverse transcriptase. qPCR reaction was performed using Applied Biosystems Power up SYBR Green PCR Master mix following the manufacturer's instructions and using an Applied Biosystems QuantStudio 3 Real-Time PCR System. Primers used for qPCR are listed in Table S1.

Single-molecule inexpensive FISH
Single-molecule inexpensive FISH (smiFISH) was performed as described in (39). Briefly, embryos were harvested by bleaching, immediately resuspended in Methanol at -20˚C, freeze cracked in liquid nitrogen for 1 minute and incubated at -20˚C overnight. Embryos were washed once in wash buffer (10 % formamide, 2x SSC buffer (Ambion)) and hybridized with the corresponding FLAP containing probes in 100µL hybridization buffer (10 % dextran sulfate, 2 mM vanadyl-ribonucleoside complex, 0.02 % RNAse-free BSA, 50 µg E. coli tRNA, 2xSSC, 10 % formamide) at 30˚C overnight. Hybridized embryos were washed twice with wash buffer and once (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Sequencing data analyses
Analysis for RNA-seq, sRNA-seq, and GRO-seq have been performed as previously described (22). Metaprofiles were generated using RPM from sRNA-seq analysis by summarizing normalized coverage information (taken from bigwig files and averaged across replicates) along early degraded targets or late degraded targets using deeptools (40).

Ribo-seq
Ribo-seq have been performed as described in (41) with some modifications. Briefly, embryos harvested by bleaching were lysate in Polysome buffer and fractionated on sucrose gradient by ultracentrifugation. RNA from monosome fraction was Dnase treated and fragments of 28-30 nucleotides were size selected after running on a TBE-Urea gel 15%. 28-30 Ribosome protected fragments were cloned with the previously described sRNA-seq library preparation approach with the following modifications: 3'phosphate was removed and 5' end was phosphorylated by treating RNA with Polynucleotide kinase.
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Gene lists
Gene lists are provided in Table S1.
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint this version posted February 3, 2020. . https://doi.org/10.1101/2020.02.03.919597 doi: bioRxiv preprint   (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.  (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.    (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.
The copyright holder for this preprint this version posted February 3, 2020. . https://doi.org/10.1101/2020.02.03.919597 doi: bioRxiv preprint  (which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Fig. S1
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Fig. S2
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Fig. S3
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.

Fig. S5
(which was not certified by peer review) is the author/funder. All rights reserved. No reuse allowed without permission.