Post-transcriptional repression of CFP-1 expands the regulatory repertoire of LIN-41/TRIM71

Abstract The Caenorhabditis elegans LIN-41/TRIM71 is a well-studied example of a versatile regulator of mRNA fate, which plays different biological functions involving distinct post-transcriptional mechanisms. In the soma, LIN-41 determines the timing of developmental transitions between larval stages. The somatic LIN-41 recognizes specific mRNAs via LREs (LIN-41 Recognition Elements) and elicits either mRNA decay or translational repression. In the germline, LIN-41 controls the oocyte-to-embryo transition (OET), although the relevant targets and regulatory mechanisms are poorly understood. The germline LIN-41 was suggested to regulate mRNAs indirectly by associating with another RNA-binding protein. We show here that LIN-41 can also regulate germline mRNAs via the LREs. Through a computational-experimental analysis, we identified the germline mRNAs potentially controlled via LREs and validated one target, the cfp-1 mRNA, encoding a conserved chromatin modifier. Our analysis suggests that cfp-1 may be a long-sought target whose LIN-41-mediated regulation during OET facilitates the transcriptional reprogramming underlying the switch from germ- to somatic cell identity.


INTRODUCTION
RNA-binding proteins (RBPs) are versatile regulators of mRNA fate ( 1 ).They employ di v erse RNA binding domains (RBDs) to recognize specific sequences and / or structural features of RNA ( 2 ).The NHL domain, named after the Caenorhabditis elegans proteins N CL-1, H T2A and L IN-41, is an RBD found in the conserved TRIM-NHL family of RBPs ( 3 ).The NHL repeats form a beta-propeller structure whose top surface can bind RNA in various ways ( 4 ).The TRIM-NHL proteins control cell proliferation versus dif ferentia tion decisions in many species.Underscoring the importance of RN A regulation, m utations in their NHL domains are associated with numerous diseases, including cancer and neurological disorders (5)(6)(7)(8).
The C. elegans LIN-41 is a well-characterized TRIM-NHL protein that associates with specific RNAs via triloop RNA hairpins called the LIN-41 Response Elements (LREs) ( 4 ).LIN-41 is best known as a player in the socalled heter ochr onic pa thway tha t regula tes de v elopmental transitions in the soma ( 9 , 10 ).The somatic LIN-41 regulates target mRNAs via either degradation or translational r epr ession, depending on whether it binds within the 3 or 5 untranslated regions (UTRs) of those mRNAs ( 11 ).Howe v er, the e xact molecular mechanisms remain unknown.Additionally, LIN-41 functions in the germline, where it controls various aspects of the OET, including the meiotic progression and reprograming into pluripotency (12)(13)(14).In the absence of LIN-41, germ cells abort meiosis, proliferate and abnormally differentiate into somatic cells forming the inv ertebrate equi valent of a human teratoma ( 12 ).The teratomatous differentiation reflects the pr ematur e onset of transcriptional reprogramming underlying the so-called embryonic (or zygotic) genome activation, which, during wild-type de v elopment, occurs in an early embryo ( 12 ).While the function of LIN-41 in the heter ochr onic pathway is linked to specific mRNAs, the identity of germline targets relevant for the various aspects of OET is less clear.Involvement of LIN-41 in the progression through meiosis I entails the regulation of cdc-25.3mRNA, which encodes an activator of the cell cycle kinase CDK-1 ( 13 ).Howe v er, which target(s) are relevant for the cell fate r eprogramming r emains unknown.
Intriguingl y, the cdc-25.3mRN A does not harbor LREs in its UTRs, arguing against the direct regulation by LIN-41.Instead, it has been proposed that LIN-41 is recruited to this and additional germline mRNAs via the association with distinct RBPs called OMA-1 and OMA-2 (collecti v ely r eferr ed to as OMA), which bind the cdc-25.3mRNA via OMA binding sites (OBSs, UAA / U) ( 15 ).The OMA-dependent association of LIN-41 with mRNAs in the germline could reflect fundamental differences in the RNA-binding mechanisms used by LIN-41 in the soma versus germline.Alternati v ely, both OMA-mediated and LREmediated RNA binding could co-exist in the germline to regulate distinct targets with different biological functions.Indeed, we show here that the LRE-mediated regulation is also present in the germline and characterize one target, the cfp-1 mRNA, which is translationally r epr essed via LREs.Unlike in the soma, we show that the translational r epr ession (or r epr ession for simplicity) of cfp-1 involves poly(A) tail shortening via the CCR4-NOT deadenylase.CFP-1 is a conserved CXXC zinc finger protein functioning in different chromatin-modifying complexes (16)(17)(18).Our observations implicate it in the tr anscriptional reprogr amming underlying the germline-to-soma transition and suggest that its LIN-41-mediated r epr ession facilitates an orderly OET.

Experimental model: C. elegans
C. elegans strains were maintained by incubation at 20 • C on 2% Nematode Growth Medium (NGM) plates seeded with Esc heric hia coli strain OP50 ( 19 ).N2 Bristol strain was used as a wild-type r efer ence of C. elegans .All other strains used are listed in Supplementary Table S1, together with their genotype and the strain ID used in the Ciosk lab.PHX3876, PHX5469 and PHX5817 were generated by SunyBiotech.To obtain synchronous worm populations, embryos were extr acted from gr avid adults with a bleaching solution (30% (v / v) sodium hypochlorite (5% chlorine) reagent (Thermo Fisher Scientific; 419550010), 750 mM KOH) and incubated overnight in the absence of food, at room temperature in M9 buffer (42 mM Na 2 HPO 4 , 22 mM KH 2 PO 4 , 86 mM NaCl, 1 mM MgSO 4 ).Arrested L1 larvae were plated on food and incuba ted a t 25 • C for the desired number of hours.For RNAi experiments, arrested L1 larvae (unless otherwise specified) were plated on RNAi-inducing NGM agar plates containing E. coli HT115 bacteria with plasmids targeting the gene of interest ( 20 ).

Analysis of RIP-seq data and calculation of predicted binding score
The LIN-41 RIP-Seq data and the LRE model used here are described in a pre vious pub lication ( 4 ).Briefly, LIN-41 binds a tri-loop RN A structure, w here the type of nucleotide bases at specific positions in the RNA stem-loop determines the binding strength.Additionally, the binding strength depends on the LRE numbers and whether the LREs are located within 5 or 3 UTRs of mRNAs.Thus, to determine predicted binding scores for all mRNAs, the C. elegans transcriptome was scanned for putati v e LREs within 5 and 3 UTRs; we used the C. elegans transcript annotations from the WormBase version WS259, as described earlier ( 4 ).Putati v e LREs were grouped into four categories (minimal, weak, medium, strong), with twofold threshold steps (0.225, 0.45, 0.9, 1.8) and the predicted binding scores were calculated per mRNA.Finally, the binding scores were plotted against RIP-Seq enrichment values (log 2 ) of mR-NAs that were enriched 4 folds or more in the RIP-Seq experiment.

Construction of reporter strains
The germline GFP reporter with the LREs in the 3 UTR was constructed as described earlier ( 4 ).Briefly, using the MultiSite Gateway Technology (Thermo Fisher Scientific) three entry plasmids with the promoter ( mex-5 ), gene body (PEST::GFP::H2B) and 3 UTR ( mab-10 3 UTR LRE sequences within unc-54 3 UTR) were combined with the destination vector pCFJ150 ( 21 ) resulting in a plasmid containing a promoter, 5 UTR, coding sequence and a 3 UTR.Transgenic animals were obtained by single-copy integration into the ttTi5605 locus on chromosome II, using the protocol for injection with low DNA concentration ( 22 ).

Microscopy
DIC (Differ ential Interfer ence Contrast) and fluor escence imaging was carried out with the Zeiss Axio Imager Z1 microscope equipped with a Zeiss AxioCam MRm camera.The images were processed in an identical manner using Fiji and Adobe Illustrator.For whole animal microscopy, animals were mounted on thin agarose pads in 20 mM levamisole.C. elegans gonads were dissected from adult animals on glass microscope slides with reaction wells using syringe needles (BD Microlance ™ 3).Animal were placed in a droplet of dissecting buffer (1 l of 10% Tween 20, 12 l of 100 mM levamisole, 10 l of 10 × M9 and 77 l of ddH 2 O) ( 23 ).Using the two syringe needles, the head of the animal was cut off in a single rapid motion just below the pharynx, resulting in the gonads and gut popping out of the animal.Thereafter, the gut and the body were carefully removed to isolate the gonads for imaging.

RN A e xtraction and cDN A synthesis
For w hole animals, RN A was purified by the PureLink ™ RN A Mini Kit (Invitro gen) using the protocol provided by the supplier.For dissected gonads, RNA was extracted using the protocol described for single worm RN A anal ysis ( 24 ).Briefly, ten dissected gonads per sample per biological r eplicate wer e lysed by hea ting a t 65 • C in the lysis buffer (5 mM Tris pH 8.0, 0.5% Triton X-100, 0.5% Tween 20, 0.25 mM EDTA and 1 mg / ml proteinase K).Extracted RNA was then treated with dsDNase (Thermo Scientific) to get rid of genomic DNA.cDNA was synthesized using Maxima H Minus First Strand cDNA Synthesis Kit (for dissected gonads) and SuperScript ™ IV First-Strand Synthesis System (for whole worms) using pr otocol fr om the suppliers.Total and pol yA RN A was converted to cDNA by using random hexamers and oligo d(T) 20 primers, respectively.

Reverse transcription quantitative PCR (RT-qPCR)
RT-qPCR was performed with the cDNA as template and gene-specific primers (Supplementary Table S2) for amplification using HOT FIREPol ® EvaGreen ® qPCR Mix Plus (Solis Biodyne) in a LightCycler96 qPCR machine.For qPCR from whole worm RNA, the housekeeping gene act-1 was used as an internal control to normalize the PCRs for the amount of RNA used in the re v erse transcription reaction.Analysis f or f old change was perf ormed with the 2 − C q method ( 25 ).For qPCR from gonads, relati v e abundance for different mRNAs was measured using a standard curve ( 26 ).The statistical significance of difference between the two conditions was calculated using the unpaired t-test.All primers are listed in Supplementary Table S2.

Rapid amplification of cDNA ends (RACE)
Total RNA used for the 3 RACE experiment was extracted from embryos pr epar ed from wild-type animals as described above.The 3 RACE System from Invitrogen ™ (Catalog number: 18373019) was used for cDNA synthesis according to the kit protocol with 2.5 g total RNA per reaction.To control RNA and primer quality, cDNA was also synthesized using the FIREScript ® RT cDNA synthesis MIX with Oligo (dT) primers from Solis BioDyne.PCR r eactions wer e set up using the 5X FIREPol ® Master Mix (Solis BioDyne).Each reaction was set up identically using 2 l cDNA, 1 l gene specific primer (10 M) and 1 l uni v ersal adapter primer pr ovided fr om the Invitr ogen ™ 3 RACE System kit in a total reaction volume of 50 l.The gene specific cfp-1 primers used were: cfp-1 F1 and cfp-1 F2 (Supplementary Table S2).For the Oligo (dT) primed cDNA controls cfp-1 R1 or cfp-1 R2 (Supplementary Table S2) were used instead of the uni v ersal adapter primer from the RACE kit.PCR products were separated on a 1.5% agarose T AE (Tris-acetate-EDT A) gel and individual bands were cut out, purified and sequenced.

Quantification of the cfp-1 3 UTR reporter GFP and the CFP-1::mCherry-myc fusion protein
To estima te dif fer ences in the expr ession of cfp-1 3 UTR GFP reporters (with and without LRE-disrupting muta-tions), we measured GFP abundance based on mean fluor escence pix el intensity.The intensities wer e measur ed within oocyte nuclei, except in the -1 oocyte where LIN-41 is being degraded ( 27 ).To normalize the data, the values were subsequently divided by the mean pixel intensity from arbitrarily selected areas in the distal gonad where LIN-41 is absent.The measurements were taken using Fiji.The means of the normalized oocyte intensities per animal were calculated for 7-10 animals for each reporter strain.The levels of CFP-1::mCherry-Myc protein, in wild-type and lin-41(tn1487ts) backgrounds, were quantified based on the mCherry fluorescence, as explained above for the GFP.The data was plotted as a boxplot using R, and the statistical significance of difference between the two strains was calculated using the unpaired two-samples Wilco x on test.

Quantification of the somatic LRE reporters
To estima te dif fer ences in the expr ession of the somatic LRE GFP reporters between mock and different RNAi-treated animals, we measured GFP abundance based on mean fluor escence pix el intensity, which was measured for 6-12 hypodermal cells per animal.To normalize the data, background mean fluorescence pixel intensity was measured in areas next to the hypodermal cells and this signal was subsequently subtracted from the reporter signal.The measurements were taken using Fiji from at least fiv e animals for each condition, resulting in the measurement of at least 30 cells per condition.The data was plotted as a boxplot using R, and the statistical significance of difference between mock treated and RNAi treated animals was calculated using the unpaired two-samples Wilco x on test.

LIN-41 can regulate mRNAs in the germline via LREs
We have previously shown that the NHL domain of LIN-41 binds to target mRNAs in the soma via LREs (Figure 1 A and ( 4)).Howe v er, the association of LIN-41 with se veral germline mRNAs is reported to be indirect, depending on its interaction with the OMA RBPs that bind a distinct RNA motif (Figure 1 A and ( 13 , 28 ).To examine w hether LRE-mediated mRN A regulation is possible in the germline, we constructed a reporter strain expressing GFP (fused to PEST for rapid turnover and histone H2B to concentrate the signal in the nuclei, facilitating imaging) from a germline promoter (P mex-5 ) and under the control of a 3 UTR containing LREs from a somatic LIN-41 target ( mab-10 ) (Figure 1 B).In the wild type, this reporter was expressed in the distal germline but not in the oocytes expressing LIN-41.Consistent with LIN-41-mediated regulation, the reporter was de-repressed upon lin-41 RNAi but not oma RNAi (Figure 1 C).Change in the reporter GFP expression could reflect either the degradation or translational r epr ession of the r eporter mRNA in wild-type / mock RNAi-ed animals.To distinguish between these two scenarios, we compared the le v els of total reporter mRNA in mock versus lin-41 RNAi-treated animals by RT-qPCR (using random hexamers to synthesize cDNA).We observed no significant difference, arguing against mRNA degradation (Figure 1 D).Regulation of the poly(A) tail length is a common mechanism that controls the efficiency of mRNA translation during oogenesis and early embryonic de v elopment ( 29 , 30 ).Ther efor e, we also compar ed the le v els of pol yadenylated reporter mRN A by RT-qPCR (using oligo dT primers to synthesize cDNA).We observed enrichment in the poly(A) fraction of reporter mRNA upon lin-41 RNAi (Figure 1 D).This suggests either the retention or extension of the poly(A) tail in the absence of LIN-41.Concluding, LIN-41 appears to be capable of translationally r epr essing mRNA via LREs also in the germline, through a mechanism that may involve cytoplasmic deadenylation.
Ther efor e, we set out to identify the physiological germline targets of LIN-41 regulated by this mechanism.

The LRE-containing germline mRNAs include the cfp-1 mRNA
Our starting point were over 600 mRNAs that co-purify with LIN-41; they were identified by RN A imm unoprecipitation, followed by RNA sequencing (RIP-seq) ( 4 ).These included the functionally relevant LIN-41 targets in the soma ( mab-10, mab-3, lin-29A and dmd-3) and in the germline ( cdc-25.or der to e xclude OMA-dependent binding, we remov ed mRN As significantl y enriched in OMA RIP ( 28 ).This analysis resulted in 11 candidate mRN As, w hich contained LREs in the 3 UTR but none in the 5 UTRs (green dots, Figure 2 A).In order to check if these candidate mRNAs display LIN-41-dependent changes in polyadenylation, as observed for the LRE reporter (Figure 1 D), we compared the le v els of total and poly-adenylated candidate mRNAs (by RT-qPCR) between wild-type and lin-41(rrr3) null mutant animals (Figure 2 C).Three candidates were enriched in the poly(A) but not total RNA fraction in lin-41(rrr3) mutants (Figure 2 C).Amongst these candidates, o rc-1 mRNA has been previously reported to be regulated by LIN-41 ( 28 ).ORC-1 is an essential component of the origin recognition complex r equir ed for DNA replication and may have a function in cell-cycle and meiotic ma tura tion.However, its relationship to lin-41 or oma phenotypes is not reported.F14H3.4 codes for a y et unchar acterized protein with a coiled-coiled domain and two disordered regions.By contrast, CFP-1 is the homolog of mammalian CXXC1 (CFP1) and yeast SPP1, which are chromatin-targeting subunits of the COMPASS (Complex Proteins Associated with Set1) histone methyltr ansfer ase complex ( 32 ).Thus, we decided to pursue it further.

LIN-41 regulates cfp-1 via LREs
To verify LIN-41-mediated regulation of cfp-1 mRNA, we cr eated a r eporter strain (similar to the LRE reporter in Figur e 1 B) wher e the expr ession of GFP is under the control of cfp-1 3 UTR.Two 3 UTR isoforms were annotated for cfp-1 , of which one lacks the region containing thr ee pr edicted LREs (Supplementary Figur e S1A, Worm-Base WBGene00009924#0-9f-10).To experimentally validate these isoform(s), we performed a 3 RACE experiment using total RNA extracted from embryos of wild-type animals (Supplementary Figure S1B).We identified only one 3 RACE product corresponding to the cfp-1 3 UTR that matched the longer 3 UTR isoform containing the three predicted LREs (Supplementary Figure S1C-E and Figure 3 A).This does not exclude the existence of other isoforms expressed at lower levels and / or other developmen-tal stages.Nonetheless, to evalua te LIN-41-media ted regulation through the identified 3 UTR isoform, we created the corresponding cfp-1 3 UTR GFP reporter strain and e xamined the GFP e xpression in animals subjected to either mock or lin-41 RNAi.In control animals, we observed a stronger GFP expression in the distal gonad but significantly reduced in the de v eloping oocytes, which increased again in the ovulating (-1) oocyte (Figure 3 B).This expression pa ttern anti-correla tes with LIN-41, as LIN-41 is expressed in the de v eloping oocytes except in the ovulating (-1) oocyte, where it is acti v ely degraded via the proteasome ( 27 ).In agreement with LIN-41 mediated regulation, the r eporter GFP expr ession was no longer r educed in lin-41 RNAi-treated gonads (Figure 3 B).Finally, to confirm that the cfp-1 LREs mediate the r epr ession, we cr eated a variant of the cfp-1 3 UTR GFP r eporter, wher e single point mutations disrupted the LRE structures necessary for the association with LIN-41 (Supplementary Figure S2).Consistent with the de-r epr ession of this reporter variant, the GFP signal in the oocytes was significantly increased compared to the wild-type reporter (Figure 3 C).Taken together, our observations support the LRE-mediated translational r epr ession of cfp-1 by LIN-41.

The cfp-1 regulation involves the CCR4 NOT deadenylase complex
Our RT-qPCR experiments suggest that LRE-mediated regulation in the germline involves target mRNA deadenylation (GFP mRNA in Figure 1 D and cfp-1 mRNA in Figure 2 C).Purification and identification of LIN-41 partner proteins by immunoprecipitation followed by massspectrometry has shown that LIN-41 interacts with the components of CCR4-NOT deadenylase complex ( 28 ).Additionall y, by systematicall y anal yzing the function of different deadenylases in C. elegans germline de v elopment, it has been determined that CCR4-NOT complex is the major deadenylase complex in C. elegans germ cells and CCF-1 is the main deadenylase ( 33 ).To test if the CCR4-NOT deadenylase complex regulates cfp-1 mRNA, we RNAi-depleted mRNAs encoding the scaffolding protein NOT1 ( ntl-1 in C. elegans ) and the deadenylase CCF-1 and monitored the ex-pression of the cfp-1 3 UTR GFP reporter.Knocking down these factors resulted in se v ere de v elopmental defects including in the germline, as observed previously ( 33 ).Therefore, we optimized the RNAi-mediated depletion, treating animals from mid-L4 to young adult stage (for 12-14 h), which had no overt impact on oocyte appearance.Such short treatment was sufficient to observe an increase in the expression of the cfp-1 3 UTR GFP reporter upon RNAi against either not-1 or ccf-1 (Figure 4 ).These observations suggest a model, where LRE-associated LIN-41 regulates the associated mRNA by recruiting the CCR4-NOT deadenylase complex.Interestingly, this mechanism appears to be specific to the germline, as the CCR4-NOT complex seems dispensable for the regulation of somatic LIN-41 targets (Supplementary Figure S3).Thus, LIN-41 utilizes different mechanisms to regulate germline versus somatic targets.

CFP-1 protein is o v er-expressed in lin-41 mutant gonads
The experiments so far showed that LIN-41 elicits LREmedia ted transla tional r epr ession of a r eporter GFP.To confirm that this mechanism also impacts the le v els of CFP-1 protein, we generated a strain expressing endogenous CFP-1 tagged (by CRISPR-Cas9 editing) with mCherry-Myc ( cfp-1(syb3876) , henceforth cfp-1::mCherry-m y c ).We observed expression of the CFP-1 fusion protein through larval de v elopment both in the soma and the germline (Supplementary Figure S4).Howe v er, its expression was strongly reduced in the proximal gonad of wild-type animals containing LIN-41 (Figure 5 A).To confirm that the le v els of CFP-1 depend on LIN-41, we subjected the cfp-1::mCherry-m y c animals to lin-41 RNAi (from the L1 to young adult stage).Consistent with the regulation by LIN-41, the expression of CFP-1::mCherry-My c e xtended to the proximal gonad in the RNAi-treated animals (Supplementary Figure S5B).To rule out secondary effects due to oocyte malformation, we also examined a lin-41 temperature sensitive mutant, lin-41(tn1487ts) , whose gonads at restricti v e temperature still contain oocyte-like cells rather than a teratoma ( 13 ).Using this mutant, we also observed the ectopic expression of CFP-1::mCherry-Myc in the oocyte-like cells (Figure 5 B, C).Additionally, we examined the CFP-1::mCherry-My c e xpression in wild-type embryos.We detected CFP-1 expression starting from the 4-8 cell-stage, with subsequent increase in older embryos (Figure 5 D).We noticed a small difference in the expression patterns between the CFP-1 fusion protein and the cfp-1 3 UTR GFP r eporter (Figur e 3 B and 5A ).While the reporter GFP was visible already in 2 cell-stage embryos, the fusion protein accumulated with some delay (Figure 5 D and Supplementary Figure S5A), possibly suggesting regulation by additional mechanisms (see Discussion).Taken together, our results support LIN-41-media ted transla tional r epr ession of CFP-1 in the proximal gonad.

CFP-1 promotes the expression of early embryonic genes
CFP-1 is a conserved protein studied from yeast to humans ( 32 ).The mammalian CFP1 / CXXC1 recruits the SET1 / COMPASS to induce the tri-methylation of histone H3 at Lys4 (H3K4me3) at acti v ely transcribed genes.Additionally, recent studies using C. elegans show that CFP-1 interacts also with additional chromatin factors, including components of the SIN3 / HDAC complex ( 16 , 17 ).Since CFP-1 is absent in the de v eloping oocytes but accumulates in early embryos, we wondered if its abnormal accumulation in lin-41 mutants could facilitate the transcriptional reprogramming resulting in a teratoma.We first asked whether there is a relationship between CFP-1 and the genes transcribed during embryonic genome activation ( 26 ).Specifically, we examined if the early embryonic genes (EEGs) are down or upregulated in cfp-1 mutants using published transcriptomics data ( 16 ).We found a striking enrichment of EEGs among the downregulated but not upregulated genes (Figure 6 A).Gene Ontology enrichment analysis showed that EEGs down-regulated in cfp-1 mutants are functionally related to tr anscription (dsDNA binding, tr anscription r egulatory r egion nucleic acid binding, transcription factor activity etc.; Supplementary Figure S6).
We then asked if the ectopic expression of CFP-1 may promote the induction of embryonic genes in lin-41 gonads.To do that, we dissected gonads from lin-41(rrr3) animals subjected to either mock or cfp-1 RNAi and perf ormed RT-qPCR f or specific genes.We selected genes rep-r esenting differ ent aspects of embryonic differ entiation that wer e pr eviousl y shown to be abnormall y induced in LIN-41-deficient gonads ( 12 ).Expectedly, genes r epr esenting EEGs ( v et-4 , v et-6 ), somatic lineage specific genes ( hlh-1, unc-120, end-1 ) and hox genes ( mab-5, ceh-13 ), were all upregulated in lin-41(rrr3) mutants compared with wild-type (Figure 6 B).Howe v er, RNAi-mediated depletion of CFP-1 strongly reduced the le v els of most tested transcripts (Figure 6 B).Taken together, our experiments suggest that LIN-41-media ted transla tional r epr ession of cfp-1 mRNA helps maintain the germ cell fate in the de v eloping oocytes by pre v enting an untimely onset of embryonic transcription.Importantl y, the underl ying mechanism of translational repression is distinct from what is reported for other known LIN-41 mRNA targets.Unlike cdc-25.3, the r epr ession of cfp-1 does not r equir e the OMA RBP.Although the r epr ession of cfp-1 is mediated by LREs, like the r epr ession of somatic targets, the CCR4-NOT deadenylase complex is essential only for the former.Thus, LIN-41 appears to have e volv ed unique solutions to the regulation of different targets performing distinct biological roles.

DISCUSSION
The cfp-1 mRNA adds to a growing number of LIN-41 targets regulated by target-specific mechanisms (Figure 7 ).Pr evious r eports postulated an indir ect r ecruitment of LIN-41 to some germline transcripts, via its association with the OMA RBP.By contrast, our data suggest a direct recruitment of LIN-41 to cfp-1 through the LREs.Our results suggest that this leads to translational r epr ession involving the CCR4-NOT deadenylase complex.Controlling gene expression through polyaden ylation / deaden ylation is widespread in germ cells and early embryos ( 29,30 ).It was suggested that transcripts associating with LIN-41 are substrates of the GLD-2 pol yA pol ymerase, and both GLD-2 with its co-factors and CCR4-NOT components were identified in the LIN-41 pull-downs ( 28 ).Thus, a tug-ofwar between cytoplasmic deadenylation and polyadenylation could decide the fate of cfp-1 mRNA, as was previously suggested for other germline transcripts ( 34 ).Interestingly, like LIN-41, also the fly TRIM-NHL protein BRAT utilizes the CCR4-NOT complex to control mRNA translation ( 35 , 36 ).The CCR4-NOT recruitment is mediated by BRAT's interactions with another RBP, Nanos, which directly associates with the Not1 and Not3 components of the CCR4-NOT complex ( 37 ).Whether LIN-41 recruits the CCR4-NO T complex directl y or thr ough interacting pr oteins remains to be tested.Although the transient translational r epr ession of cfp-1 via deadenylation could largely explain the CFP-1 expression pattern, other mechanisms may contribute.We noticed some differences between the expression patterns of the cfp-1 3 UTR GFP reporter and the endogenous CFP-1 protein.The reporter expression anticorrelates with the abundance of LIN-41.While it is repressed in most oocytes, it is de-r epr essed in the ovulating (-1) oocytes and earl y embryos, w here LIN-41 is degraded ( 27 ).By contrast, the CFP-1 protein is detectable in neither the -1 oocytes nor early embryos.While other explanations remain possible, these observations suggest an   additional layer of CFP-1 regulation, possibly involving its proteolysis.
Regardless of the actual mechanism(s), why is CFP-1 s expression controlled?CFP-1 is r equir ed for fertility and cfp-1(tm6369) mutants have reduced brood size when grown at 25 • C, e v entually leading to complete sterility within two generations ( 16 , 17 ).Our observations implicate CFP-1 in the expression of early embryonic genes.Whether this function is related to the gonadal or embryonic expression of CFP-1, and whether it is connected to the sterility phenotype, remains to be tested.Howe v er, loss of cfp-1 in the germline results in a phenotype stronger than COM-PASS inactivation, most likely reflecting additional functions such as SIN3 recruitment to H3K3me3 enriched promoters ( 17 ).It is possible that CFP-1 acts as a hub to coordinate H3K4 tri methylation and HDAC activity.An important question for the future is the relati v e importance of the CFP-1-containing complexes for the expression of embryonic genes.Similar to CFP-1, the murine CFP1 is important for various aspects of oocyte de v elopment and zygotic genome activation ( 38 , 39 ).Consistent with a role in embryonic transcription, Cfp1 −/ -murine embryonic stem (ES) cells fail to dif ferentia te (40)(41)(42).Intriguingly, like the nematode protein, CFP1 is temporally r epr essed in maturing murine oocytes and early embryos ( 38 , 39 ).Thus, the transient r epr ession of CFP-1 / CFP1 could be a conserved phenomenon important for the epigenetic reprogr amming under lying the switch from germ-to embryonic transcription.
In other models, so-called pioneering transcription factors (Zelda in flies, Pou5f3, Sox19b and Nanog in fish, DUX and NFY in mice, and OCT4 in humans) play critical roles in activating the embryonic genome by increasing chromatin accessibility to other transcription factors ( 43 ).Among them, the fly Zelda and human OCT4 were shown to be post-transcriptionally regulated.Zelda is the major activator of the zygotic genome and its expression in fly embryos is translationally r epr essed by the LIN-41 ortholog BRAT ( 44 , 45 ).As for OCT4 , it activates transcription in the mouse embryos at the 2-cell stage when the zygotic transcription begins ( 46 ).In human embryonic stem cells and germinal vesicle-stage of pig oocytes, the OCT4 mRNA associates with another RBP, DND1 ( 47 , 48 ).This association could be functionally relevant for the germline, as OCT4 expression is post-transcriptionally downregulated in the male germ cells w here, analo gous to LIN-41, DND1 maintains germline identity, pre v enting the onset of testicular teratomas ( 49 , 50 ).Also, similar to LIN-41 and BRAT ( 37 ), DND1-mediated translational r epr ession involves the r ecruitment of CCR4-NOT Binding to the 3 UTRs results in mRN A degradation, w hereas binding to the 5 UTR in translational r epr ession ( 11 ).The underlying mechanisms and effectors remain unknown as indicated by the red question mark.The known somatic mRNA targets of LIN-41 encode transcription factors.Lower panel: in the germline, LIN-41 was suggested to bind se v eral mRNAs (including the cdc-25.3encoding a cell cycle regulator) indirectl y, possibl y via the association with the OMA RBP, which binds this and other mRNAs via OMA binding sites (OBSs; consisting of a repetiti v e UAA / U motif).Whether cdc-25.3 and / or other indirect LIN-41 targets are regulated via deadenylation is not clear, as indicated by the red question mark ( 28 ).Additionally, using the example of cfp-1 mRNA encoding a chromatin modifier, we showed that LIN-41 regulates germline mRNAs by directly associating with them via LREs.In the case of translational r epr ession of cfp-1 mRNA, this involves deadenylation via the CCR4-NOT complex.deadenylase ( 36 ).Thus, while the gener al str ategy of controlling the transcriptional reprograming during OET by posttranscriptional 'roadblocks' appears to be conserved ( 51 ), the mRNA targets may differ.Zelda is not conserved outside insects, and the nematode homologs of OCT4 and SOX2 are required for a transdif ferentia tion e v ent in the soma, where the rectal epithelial cell Y transdif ferentia tes into PDA neurons ( 52 ).Perple xingly, howe v er, these factors appear to have no role in the transcriptional reprogramming during OET, nor were their functional equivalents identified so far.In nematode embryos, histone acetyltr ansfer ases (HATs) promote most somatic differentiation programs likely by antagonizing histone deacetylase activities ( 53 ).In murine embryos, ov ere xpression of a dominant-negati v e form of HDAC1 / 2 leads to a de v elopmental arrest at the two-cell stage ( 54 ).In these embryos, 64% of the downregulated genes are EEGs and the authors proposed that HDAC activity is critical for the activation of the zygotic genome by cr eating corr ect transcription-acti v e and -r epr essi v e states at the chromatin le v el.These e xamples stress the importance of epigenetic reprogramming for embryonic differentia tion.W hile it remains possible that nematodes use a yet-to-be-found pioneering factor, our findings suggest that chroma tin modifica tions media ted by CFP-1 and possibly other chromatin interacting proteins could play the main role in the switch from the germline-specific to embryonic transcription.In this scenario, transient r epr ession and subsequent r e-expr ession of CFP-1 could contribute to the erasure of germline-specific chromatin states and the subsequent establishment of a chr omatin envir onment compatible with embryonic dif ferentia tion.

DA T A A V AILABILITY
The data underlying this article are available in the article and in its online supplementary material.

Figure 1 .
Figure 1.LIN-41 can regulate mRNAs in the germline via LREs.( A ) Tissue-specific mRNA regulation modes by LIN-41.In the somatic tissue,

Fold. 1 Figure 2 .
Figure 2. Identification of germline mRNAs potentially regulated via LREs.( A ) Scatterplot showing (lack of) correlation between mRNAs enriched in LIN-41 RIP (log 2 scale) and mRNAs predicted to bind LIN-41 via LREs.The mab-10 mRNA, a somatic LIN-41 target containing LREs is shown in cyan, and cdc-25.3, a germline mRNA lacking LREs but regulated by LIN-41 is shown in r ed.Gr een dots r epr esent putati v e LIN-41 germline targets containing LREs identified as described in (B).( B ) A discovery pipeline used to identify potential LIN-41 mRNA targets in the germline regulated via LREs.( C ) RT-qPCR analysis comparing relati v e le v els of total and polyadenylated candidate mRNAs between wild-type and lin-41(rrr3) mutant animals.The bars r epr esent the ratio of expression levels in lin-41(rrr3) mutants to wild-type animals and are therefore compared to 1 as indicated by the red line.Error bars r epr esent standard deviation from three biological replicates.*** denotes a P -value of < 0.05 and ** a P -value of < 0.01, by unpaired t -test.Blue colored *** indicate significant change in the le v els of polyadenylated mRNA whereas black colored *** indicate significant change in the le v els of total mRNA.

Figure 3 .
Figure 3.The cfp-1 mRNA is a bona fide LIN-41 target.( A ) Schematic of the cfp-1 3 UTR determined by 3 RACE.The 3 UTR is 458 nt-long and contains thr ee pr edicted LREs at the indicated positions.The red asterisks indicate the nucleotides that are mutated in panel C. ( B ) Top: Diagram r epr esenting a germline r eporter, wher ein GFP fused to PEST and H2B fragments is expressed under the control of mex-5 promoter and cfp-1 3 UTR.The LREs are sho wn as stem-loops.Lo wer panels: Fluorescent micro gra phs of gonads (outlined) dissected from animals expressing the cfp-1 3 UTR reporter, subjected to mock or lin-41 RNAi.Oocyte nuclei expressing strong GFP fluorescence are marked with yello w arro wheads.The ovulating (-1) oocyte in the mocktreated gonad is marked with a red arrowhead.The numbers indicate how many animals out of total display the presented phenotype.Scale bar = 25 m. ( C ) Left: Schematics r epr esenting the 3 UTR variations used in the r eporters.The r ed asterisks indicate the mutant nucleotides marked in panel A. Middle: fluorescent micro gra phs of gonads (outlined) dissected from animals expr essing the GFP r eporter under the control of wild-type or mutated cfp-1 3 UTR.Oocyte nuclei expressing strong GFP are marked with yellow arrowheads.The ovulating (-1) oocytes are marked with red arrowheads.Scale bar = 25 m.Right: quantification of pixel intensity illustrating the change in reporter GFP fluorescence, calculated as ratio between fluorescence in the oocyte nuclei versus in the distal gonad (oocyte GFP expression / distal gonad GFP expression).The ovulating (-1) oocytes were not included in this analysis.Mean values are marked by white crosses.The P -value was calculated using the unpaired two-sample Wilco x on test.

Figure 4 .
Figure 4.The CCR4-NOT deadenylase complex is r equir ed for the translational r epr ession of cfp-1.Left: on top is shown a diagr am illustr ating the cfp-1 3 UTR reporter as in Figure 3 B. Below are fluorescent micro gra phs of gonads (outlined) in li v e animals expressing the cfp-1 3 UTR reporter and subjected to mock, ccf-1 , or not-1 RNAi.Yellow arrowheads point to examples of nuclei with strong GFP fluorescence.The -1 ovulating oocytes are indicated by red arrowheads.Scale bar = 50 m.Right: quantification of pixel intensity illustrating the change in reporter GFP fluorescence, calculated as ratio between fluorescence in the oocyte nuclei versus in the distal gonad (oocyte GFP expression / distal gonad GFP expression).The ovulating (-1) oocytes were not included in this analysis.Mean values are marked by white crosses.The P -value was calculated using the unpaired two-sample Wilco x on test.

Figure 5 .
Figure 5. LIN-41 r epr esses CFP-1 expr ession in the de v eloping oocytes.( A -C ) DIC and fluorescence micro gra phs from the cfp-1::mCherr y-m y c strain, expressing endogenous CFP-1 fused to mCherry and myc.(A) Gonad dissected from the animal expressing CFP-1::mCherry-Myc.Red arrowheads point to the oocyte nuclei with weak or no mCherry fluorescence.Scale bar = 25 m.(B) Gonad dissected from the lin-41(tn1487ts); cfp-1::mCherry-m y c animal.Note the abnormal oocytes in the proximal gonad.Yellow arrowheads point to the nuclei with stronger mCherry fluorescence.Scale bar = 25 m.(C) Quantification of pixel intensity illustrating the difference in mCherry fluorescence between cfp-1::mCherry-m y c and lin-41(tn1487ts); cfp-1::mCherrym y c animals, calcula ted as ra tio between fluorescence in the oocyte nuclei versus in the distal gonad (oocyte mCherry expression / distal gonad mCherry expression).Mean values are marked by white crosses.The P -value was calculated using the unpaired two-sample Wilco x on test.( D ) Embryos at the indicated de v elopmental stages.Note that the C. elegans embryonic transcription begins around the 4-cell stage.Red arrowheads point to the nuclei with weak or no mCherry fluorescence.Yellow arrowheads point to the nuclei with stronger mCherry fluorescence.Scale bar = 25 m.

Figure 6 .
Figure 6.CFP-1 is r equir ed for the expression of early embryonic genes in wild-type embryos and lin-41(-) mutant gonads.( A ) Venn diagram showing an overlap between early embryonic genes and differentially expressed genes in the cfp-1(tm6369) animals.*** indicates the P -value of 1.87e-33, by hypergeometric test.The cfp-1 ( tm6369 ) mutant gene expression data set is taken from Beurton et al. and the EEG list is taken from Fassnacht et al .Dn = downregulated and Up = upregulated.( B ) RT-qPCR analysis comparing the abundance of selected mRNAs in the gonads dissected from animals of the indicated genotypes.act-1 served as a negati v e control, the remaining mRNAs are normally expressed in embryos.Error bars represent standard deviation from three biological replicates and *** denotes a P -value of < 0.05, by unpaired t -test.Note that, compared with wild type, the le v els of embryonic mRNAs were higher in the lin-41(rrr3) gonads, but that increase was suppressed upon cfp-1 RNAi.

Figure 7 .
Figure 7.The summary of LIN-41-mediated mRNA regulation Upper panel: in the soma, LIN-41 binds to mRNAs via LREs located in the 3 or 5 UTRs.Binding to the 3 UTRs results in mRN A degradation, w hereas binding to the 5 UTR in translational r epr ession( 11 ).The underlying mechanisms and effectors remain unknown as indicated by the red question mark.The known somatic mRNA targets of LIN-41 encode transcription factors.Lower panel: in the germline, LIN-41 was suggested to bind se v eral mRNAs (including the cdc-25.3encoding a cell cycle regulator) indirectl y, possibl y via the association with the OMA RBP, which binds this and other mRNAs via OMA binding sites (OBSs; consisting of a repetiti v e UAA / U motif).Whether cdc-25.3 and / or other indirect LIN-41 targets are regulated via deadenylation is not clear, as indicated by the red question mark( 28 ).Additionally, using the example of cfp-1 mRNA encoding a chromatin modifier, we showed that LIN-41 regulates germline mRNAs by directly associating with them via LREs.In the case of translational r epr ession of cfp-1 mRNA, this involves deadenylation via the CCR4-NOT complex.