Specific miRNA-GPCR networks regulate Sox9a/Sox9b activities to promote gonadal renewal in zebrafish

Fertility and endocrine function rely on a tightly regulated synchronicity within the hypothalamic-pituitary gonadal (HPG) axis. FSH/cAMP/MAPK/ Sox9 axis signaling and its regulated specific miRNAs are thought to regulate vertebrate gonadal development and sex differentiation, and yet the regulatory networks are largely unknown. Here we construct small RNA and mRNA libraries from sexually matured ovary and testis of zebrafish to identify specific miRNA-target pairs. Integration of Targetscan prediction and in vivo induced gene expression highlight four specific miRNAs that conditionally target three G protein–coupled receptor (GPCR) x-Sox9 signaling genes, and implicate two regulatory circuits of miR430a-Sox9a in the testis and miR218a-Sox9b in the ovary. Co-injected Sox9a-miR430a mixture increases the proportion of spermatogonia but degenerates primary oocyte, while Sox9b-miR218a mixture induces renewal of ovarian follicles. Co-immunoprecipitation and mass-spectrometry analyses further reveal that miR430a and Sox9a synergistically activate testicular PKC/Rock1 signals while miR218a and Sox9b constrict ovary PKC/PI3K/Rock1 signaling. These results clarify specific miRNAs-GPCR regulatory networks of Sox9a/Sox9b switch, and also provide mechanistic insight into gonadal rejuvenation and plasticity.


Introduction 29
Although cell phenotypic changes often occur in stem cell-based differentiation, dedifferentiation or 30 transdifferentiation, both stem cells and differentiated cells are able to reversibly dedifferentiate and transform 31 into cells of different lineages under certain conditions. This phenomenon, called environmentally guided cell 32 plasticity is the basis of tissue homeostasis, tissue regeneration, cancer cell recover (Galliot & Ghila, 2010), 33 and sex reversal (Sun, Zhang et al., 2013). One excellent system to study cell plasticity is the vertebrate gonad, 34 where distinct populations of somatic and germ stem cells give rise to different sex-dependent morphotypes: 35 male testiculogenesis and female folliculogenesis (Chen, Zheng et al., 2013, Gassei & Schlatt, 2007. 36 The early gonad is an undifferentiated primordium composed of bipotential somatic stem cells: precursors for 37 supporting cells and steroid-secreting cells. When the primordial germ cells (PGCs) reside in the gonadal 38 ridge, the supporting cell precursors develop into either testis-specific Sertoli cells or ovary-specific follicle 39 Histologically, testiculogenesis with cord formation and Leyding cell differentiation supports spermatogenesis 45 while folliculogenesis accompanies ootidogenesis, where the ovarian follicle surrounding the oocyte develops 46 from a primordial follicle to a preovulatory one. There is a close anatomical relationship between the 47 5 hormone (LH), are glycoprotein peptide hormones. Gonadotropin receptors are coupled to 98 the G-protein system. We searched the two transcriptome datasets for GPCR-sox9 axis genes. These included 99 92 GPCR, 35 cAMP, 34 MAPKs, and 21 p53 orthologs as well as 30 sex determinant genes. The vast 100 majority of the GPCR-sox9 axis genes (201/212) did not show significant differential expression between 101 testis and ovary (Table S5). This result conformed to the phenomena that miRNAs only modestly 102 downregulate the mRNA level of their target genes (Selbach, Schwanhausser et al., 2008). 103

Identification of specific miRNAs and GPCR-sox9 axis target genes for gonadal differentiation 104
To identify specific miRNAs-target pairs in reproductive cycle, we analyzed two small RNA libraries, which 105 were simultaneously constructed from the same ovary or testis total RNA samples as the mRNA libraries were 106 made. After aligning the small RNA sequences with the miRBase (zebrafish miRNA database), bioinformatic 107 analysis identified 350 unique mature miRNAs, and 346 miRNA precursors. 314 miRNAs were found in 108 either of the two libraries, of which 51 miRNAs showed 2-fold higher transcription in testis (refer to testis 109 miRNAs) and 106 miRNAs were preferentially transcribed in ovary (ovary miRNAs). 110 Using TargetScan to search for the target pairs between 157 sex-biased miRNAs and 209 GPCR-sox9 axis 111 genes, we found that 97 miRNAs targeted 183 GPCR-Sox9 axis genes, including those multiple and cross 112 targets (Table S7). For example sox9a was targeted by 25 miRNAs and sox9b was targeted by 20 miRNAs. 113

Interaction modes between specific miRNAs and GPCR-sox9 targets in testis and ovary 129
To investigate the physiological interactions between four specific miRNAs and GPCR-sox9 axis targets, we 130 devised an in vitro gonadal microinjection strategy to enforce exogenous expression of Sox9a/Sox9b and four 131 specific miRNAs in the gonads, and evaluated their effects on the expression of GPCR-sox9 axis targets 132 (Table S8). When the Targetscan prediction and RT-PCR examinations were integrated together, four 133 interaction modes emerged (Table 2) : 1) a negative specific miRNA-target pair: a predicted target was 134 downregulated in both gonads by a specific miRNA; 2) a positive specific miRNA-target pair: a predicted 135 target was upregulated in both gonads by a specific miRNA; 3) a conditionally specific miRNA-target pair: a 136 predicted target was upregulated by a specific miRNA in one sex gonad but downregulated in the other gonad; 137 and 4) an indirect specific miRNA-target: a gene without specific miRNA binding sites was downregulated or 138 upregulated. Totally 92.3% (12 of 13) of predicted targets of GPCR-sox9 genes were correctly scored as direct 139 targets in testis and/or ovary by at least one of four specific miRNAs. Fig.1   Previous study has identified several sox9 downstream genes (cyp19a, cyp19b, amh) as potential targets of 152 miR430 family during sexual transformation of the rice field eel . Here we found that 153 miR734 and miR141 formed target-pairs or conditional target pairs with sox9a, amh and/or sox9b, while 154 miR430a and miR218a only showed indirect interactions with these genes and cyp19a1a (Fig.1). Referred to 155 sox9a and sox9b transcription, increased miR141 and miR734 showed similar regulatory functions as 156 miR430a (Table S8). Additionally three miR430a direct target genes (lgr4, grk5l, grk4) were also pair-targets 157 or conditional pair-targets for miR141, miR734 and miR218a (Fig.1). Thus, the four miRNAs may coordinate 158 to regulate Sox9a/Sox9b activity through GPCR-signaling networks. 159 We then investigated the regulatory modes between the four specific miRNAs and two Sox9 isoforms. 160 Sustained Sox9b expression increased the transcription of four pre-miRNAs in both ovary and testis, whereas 161 enforced Sox9a only induced pre-miR430a transcription in testis (Fig.S2). Similarly, overexpressed Sox9a and 162 Sox9b activated all tested GPCR-sox9 genes in the ovary whereas a few genes were suppressed in testis.
cyp19a1a. In contrast to miR430a-Sox9a reciprocal activation in the testis, miR218a downregulated 166 transcripts of sox9a and cyp19a1a but modestly upregulated sox9b transcript in both testis and ovary. 167 These results implicated two transcriptional regulatory circuits: miR430a-Sox9a, and miR218a-Sox9b. 168 Relative to miR218a-Sox9b's extensive and ovary-biased regulation, miR430a-Sox9a synergic regulation 169 was specific in the testis. When we constructed a GFP reporter for 3'UTR of sox9a and sox9b, and carried out 170 GFP reporter assays in zebrafish embryonic fibroblast cells (Pac2), we found that miR430a and miR218a not 171 only regulates the two Sox9 genes' transcription, but also confines the two proteins' subcellular distribution 172 ( Fig.S4, S5). 173

Gonadal microinjection of specific miRNAs-Sox9 mixture cause renewal of gonocytes 174
To evaluate two miRNA-Sox9 regulatory circuits and its role in the gonadal differentiation, we conducted a 175 combinatory microinjection into mature gonads (5mpf), and then compared cellular events of somatic cells 176 and germline cells (Fig.2A，2B). Enforced Sox9a-miR430a mixture expression increased the proportion of 177 mitotic condensed chromatin nuclear cell (Cn) and undifferentiated gonocytes (Gc), and increased 178 degeneration of primary oocytes (Fig.2Aa, d). Co-injected Sox9a-miR141-miR734 cocktails increased 179 proportion of spermatocytes (Sc), and also induced ovary-testis transition (indicated by oocyte degeneration 180 and multinucleate cell generation, Fig.2Ab, 2Ae). Conversely, enforced Sox9b-miR218a expression increased 181 each stage of follicles (oocytogenesis), but reduced testis Cn and Sc with increase of deforming Gc (Fig. 2Ac,  182 2Af). We didn't found consistent roles of other Sox9a/ Sox9b-miRNAs mixtures in testiculogenesis and 183 folliclulogensis (Fig.2C, Fig.S3). 184 Since Sox9a, Sox9b and pre-miRNAs transcripts were differentially transcribed in the testis and ovary 185 (Fig.S1), and naturally reduced with age, we posited that exogenous specific miRNAs-coupled Sox9a or 186 Sox9b may induce certain "rejuvenation" effects on the old gonads. To explore this possibility, we tested 187 1.5-2 years old of zebrafish. As expected, Sox9a-miR430a mixture significantly increased the proportion of 188 Gc and Cn whereas Sox9b-miR218a mixture increased all stages of follicles. Sox9b+miR430a increased the 189 testis Cn and Sc (Fig.2Bc, 2Cc) as well as the primordial follicles (Fig.2Be). Compared to the injection into 190 the young fish (5 months olds) ( Fig.2A), Sox9b-miR218a mixture ( Fig.2A-2B)  Immunoblotting analyses showed that specific Sox9a and Sox9b antibodies recognized several bands, which 222 molecular-weight sizes were bigger than the predicted; and the band patterns were changed after 223 cotransfections of miR430a or miR218a (Fig.3A). In the transfected Pac2 cells, the presumed protein 224 modifications (Fig.3.B) were not significant as those occurred in the microinjected gonadal tissues. 225

Fig.3. Immunoblotting analyses of endogenous and exogenous Sox9a/Sox9b protein modifications. (A) 227
The testes were microinjected with pSox9a-myc (Sox9a) or pSox9a-myc and miR430a mix (9a+miR430a); 228 the ovaries were injected with pSox9b-myc (Sox9b) or pSox9b-myc and miR218a mix (9b+miR218a). (B). In 229 the in vitro cultured Pac2 cells were transfected with pSox9a-myc, pSox9a-myc/miR430a mix, pSox9b-myc, 230 pSox9b-myc/miR218a mix, or empty-control. (C). Sex gonads were microinjected as described in (A). The 231 expression of each signaling component was examined. Relative signal intensity was calculated compared to 232 total Akt. 233 234 Co-immunoprecipitation (Co-IP) coupled-mass spectrometer analysis indicated several reverse changes in 235 covalent modifications (Table 3) We then verified whether the identified kinases linked to our proposed miR430a/miR218a-GPCR-Sox9 247 networks. Enforced Sox9b increased Rock1, p-PI3K and p-PKC in the ovary while co-injected miR218a 248 suppressed Rock1 and p-PKC, partially suppressed p-PI3K, and increased p-Akt, suggesting that miR218 249 could modulate Sox9b activations (Fig.3C). In the testis, Sox9a did not activate Rock1, p-PI3K and p-PKC; 250 However, Sox9a and miR430a mixture activated p-PKC and Rock1signaling. 251 252  together with the results shown in Fig.3, Fig.S2, Table2 and Table 3, we can make following conclusions: 1) 265  Second, our analysis on the global mRNA-miRNA expression profiles that concur in the same sex gonads 306 represents the first effort to systematically identify the gonadotropins-GPCR-Sox9 regulatory networks in 307 testiculogenesis and folliculogenesis. We extend the previous finding that miRNAs could target multiple 308 genes and form a complex miRNA regulatory network (Tao et al., 2016), and summarize four interaction 309 modes between specific miRNAs and the target genes, i.e., downreguatory target-pair, upregulatory target-pair, 310 conditionally regulatory target-pair, and indirectly regulatory target-pair. 311 Previous studies have revealed dual regulatory module of miRNAs from repression to activation at 312 transcriptional and translational levels (Vasudevan, Tong et al., 2007). The miR-430 allows the 313 post-transcriptional activation of nanos1 in germline cells, but induces deadenylation and translational 314 repression in somatic cells (Mishima, Giraldez et al., 2006). Now we present the other mechanism that 315 miR430a could modulate expression and intercellular localization of Sox9a and Sox9b proteins through 316 chemical modifications, such as phosphorylation, acetylation, oxidation and ubiquitination (Table 3) Finally, we initiated gonadal injections of the specific miRNA-target cocktail to rejuvenate gonadal reserve. 320 Replacement therapy with hormones, and stem cells, blood from aged male, and blood mononuclear cells conditions. In the present study, we found that gonadal injection of miR430a-Sox9a mixture could increase 327 testicular gonocytes, and miR218a-Sox9b mixture could increase primordial follicles in old ovaries. Although 328 it is unknown about the source origins from proliferation of reserved stem cell or dedifferentiation of adult 329 gonadal cells, miR430a-Sox9a and miR430a-Sox9b at least could expand the testis spermatogonia reserve and 330 ovarian follicle reserve respectively. If equivalent efficacy can be found in human gonads, miRNA-based 331 testis rejuvenation and prevention of premature ovarian failure may one day be realized. The miRNA-target 332 pair data reported here will help identify more potential therapeutic targets and promote to develop new 333 therapeutic interventions for aging-related ovarian failure and age-related testicular regression. 334 335

Fish husbandry 337
Zebrafish and vasa-GFP transgenic medaka were bred and maintained as previously described (Li, Guan et al.,  relative size and location of the sex gonads, injection site was about 1 mm upward from the intersect between 346 pectoral fin upper horizontal and the front end vertical line of spelvic fin. By this way overexpression 347 plasmids (100ng), and/or specific miRNAs (20pmol each oligos) (Table S10) were microinjected into sex 348 gonads with lipohigh transfection reagent (Sangon Biotech, Shanghai). Microinjection efficiency was 349 confirmed by western blotting analyses of enforced Sox9a-Myc and Sox9b-Myc (Fig.S6).