Rhythmic testicular transcriptome in zebrafish and mice. (a and b) Heatmaps of the 1082 zebrafish rhythmic transcripts (a) and 1882 mouse rhythmic transcripts (b), revealed by the MetaCycle analysis of the 12-timepoint transcriptome data (P < 0.05, n = 2 × 12). (c–n) Heatmaps and phases of the rhythmically expressed zebrafish and mouse genes involved in circadian regulation (zebrafish: c, d; mouse: e and f), testicular functions (zebrafish:g, h; mouse: i, j), and retinoic acid signaling (zebrafish: k, l; mouse: m, n). The gene names (zebrafish: c, g, k; mouse: e, i, m) and their phases (zebrafish: d, h, l; mouse: f, j, n) are color-coded and annotated. (o and p) PCA of all transcripts from the 12 timepoints in the zebrafish (o) and mouse (p) testes shows a single clustering of the rhythmic transcriptome (red dots in the enlarged panel), respectively. (q and r) Phase distribution over the 24-hour cycle of the enriched KEGG pathways in the zebrafish (q) or mouse (r) testes (P < 0.05) (Fig. S1).

The zebrafish testis clock ticks in a cell-specific manner. (a and b) Bioluminescence recordings (a) and their periods (b) of the testis cultured ex vivo from Tg(per3: luc) (black), Tg(per3: luc); clock1a^−/− (red) and Tg(bmal1b: luc) (cyan), detrended by 24-hour moving average (n = 10). (c) Locomotor rhythm analysis of 5- to 9-dpf (days postfertilization) WT (black) and clock1a^−/− (red) zebrafish larvae under DD condition (n = 24). (d) Detrended bioluminescence of Tg(per3: luc) testis recordings under 10-hour advanced conditions after the culture ex vivo for 3 days under normal LD (n = 4). (e) Confocal images of the testis from clock1a-KI-tdTomato zebrafish (n = 3). (f) Confocal images of IHC staining in clock1a^−/− and WT control testes with a mouse CLOCK antibody (n = 6). (g and h) Expression of per1b and per2 in the Sertoli cells (g) and spermatogonia (h), and their downregulation in the clock1a^−/− Sertoli cells or spermatogonia, as shown by confocal images of the testes from the Tg(per1b: EGFP; gsdf: mCherry), Tg(per1b: EGFP; gsdf: mCherry); clock1a^−/−, Tg(per2: EGFP; gsdf: mCherry), and Tg(per2: EGFP; gsdf: mCherry); clock1a^−/− zebrafish lines (n = 5–14). In (e–h), all nuclei were counterstained with Hoechst 33342. Arrowheads indicate Sertoli cells, and arrows spermatogonia. Scale bars =100 μm (Fig. S2 and Movie S1).

Circadian regulation of RA signaling in zebrafish. (a) scRNA-seq analysis identifies seven zebrafish testicular clusters (n = ∼32 000). (b) Expression of circadian clock genes and testicular genes in the testicular cell clusters. (c–e) Single-cell co-expression (blue dots) of aldh1a2 (red dots) with per1b (green dots) in Sertoli cells (c), ztbt16a (red dots) with rarga (green dots) in SPG undiff. (d), and izumol (red dots) with rxrba (green dots) in spermatids (e). (f) Plot of the rhythmic expression of aldh1a2 in Sertoli cells and its arrhythmic expression in non-Sertoli cells at four timepoints of a day, revealed by scRNA-seq. (g and h) qRT-PCR analyses of clock1a (red), per1b (blue), and aldh1a2 (green) with RNAs from FACS-selected Sertoli cells, non-Sertoli cells, mixed cells from testicular cells of Tg(gsdf : mCherry) (g) and Tg(gsdf : mCherry); clock1a^−/− (h) (n = 3 × 3). (i) Co-localization of clock1a (green) and aldh1a2 (red) in Sertoli cells, as shown by double FISH (n = 3 × 4). All nuclei were counterstained with Hoechst 33342. Arrowheads indicate Sertoli cells. Scale bar, 50 μm. (j) Confocal images of the testes from Tg(RARE-gata2a : NLS-EYFP; gsdf : mCherry) and Tg(RARE-gata2a : NLS-EYFP; gsdf : mCherry); clock1a^−/− zebrafish lines (n = 5–7). Sg, spermatogonia; Sc, spermatocyte; Sz, spermatozoon. Scale bar, 50 μm. (k) Schematic diagrams of the aldh1a2 and rarga loci harboring E-boxes. (l, m), Luciferase report assays of the aldh1a2 (l) and rarga regulatory fragments (m) (n = 3). (n) Quantitative ChIP assays (n = 2). *P < 0.05; **P < 0.01 (Fig. S3).

Spermatogenesis and fertilization defects in clock1a^−/− and Sertoli cell clock1a mutant zebrafish. (a–f) Images of the H&E staining (a), fluorescent Tg(nanos3 : EGFP) (b) and Tg(piwil1 : EGFP) (c), IHC with a Vasa antibody (d), FISH with kita probe (e) and sycp3 probe (f) of WT control, clock1a^−/−, and Sertoli cell clock1a mutant testes (n = 9–20). (g) Quantification of the images in (b–f) (n = 6–16). (h) Detrended bioluminescence recordings of the testes ex vivo from Tg(per3 : luc) (black), Tg(per3 : luc); clock1a^−/− (red), Tg(per3 : luc; gsdf : Cas9; CG; u6a : clock1a gRNA; LC) (green), and Tg(per3 : luc; piwil1 : Cas9-RFP; CG; u6 : clock1a 3gRNAs; LC) (cyan) zebrafish lines (n = 3). (i and j) Sperm density (i) and motility (j) of WT control, clock1a^−/−, and Sertoli cell clock1a mutant zebrafish (n = 5). (k–m) GSI (k), morphology (l), and images of fertilization at 3/4 hours postfertilization (HPF) (m) of WT control, clock1a^−/−, Sertoli cell clock1a mutant, and spermatogonia clock1a mutant testes (n = 6–10). Scale bar, 500 μm. (n and o) Fertilization rates by pairwise crosses of WT control, clock1a^−/−, Sertoli cell clock1a mutant, and spermatogonia clock1a mutant males with WT control females by natural crossings (n) (n = 10–24) or IVF (o) (n = 8). (a) Scale bar, 33 μm. SC, spermatocyte; SZ, spermatozoon. (b–f) Scale bar, 100 μm. (b, c) All nuclei were counterstained with Hoechst 33342. Arrowheads indicate Sertoli cells and arrows spermatogonia. *P < 0.05; **P < 0.01; ***P < 0.001 (Fig. S4 and Movie S2).

atRA treatment rescues the clock1a^−/− spermatogenesis and fertilization defects via zbtb16a and izumo1 in zebrafish. (a) Schedules of the experiments. (b) FISH images of the WT and clock1a^−/− testis sections with the kita probe at days 1 to 4 following the atRA or vehicle treatment i.p. at ZT12 (n = 3 × 5). The signals were superimposed with the Hoechst33423 nuclear counterstain. Scale bar, 100 μm. (c) Quantification of kita-positive cells in b (n = 4). (d) Fertilization rates by pairwise crosses of WT males or clock1a^−/− males with WT females 1, 3, and 5 days after atRA or vehicle treatment at ZT0 or ZT12 (n = 9–18). (e and f) Sperm density (e) and motility (f) of the clock1a^−/− mutant males after atRA or vehicle treatment (n = 6). (g and h) Fertilization rates by pairwise crosses of WT females with clock1a^−/− males or WT males treated with BMS753 (g), and WT females with WT males treated with BMS493 (h) at ZT0 or ZT12 (n = 10–18). (i) The ratios of gene expression levels in the clock1a^−/− testis 2 hours after atRA treatment to those after vehicle treatment (n = 3). (j) FISH images of WT and clock1a^−/− testes using zbtb16a or izumo1 probe (n = 4). Scale bar, 100 μm. (k) FISH images of clock1a^−/− testes treated with RA or vehicle using zbtb16a or izumo1 probe (n = 4). (l and m) Quantification of zbtb16a-positive and izumo1-positive cells in (j and k) (n = 4). (n) Schematic diagrams of the RARE motifs-containing izumo1 and zbtb16a promoters. (o and p) Luciferase reporter assays of izumo1 (o) and zbtb16a (p) promoters (n = 3). (b, j and k) Arrows indicate spermatogonia. Sc, spermatocyte; Sz, spermatozoon. *P < 0.05; **P < 0.01;***P < 0.001 (Fig. S5).

The conserved role of the Sertoli cell clock in mice. (a) scRNA-seq analysis identifies nine mouse testicular clusters (n = ∼35 000). (b) Expression of Clock and Bmal1 in the mouse testicular cell clusters. (c–e) Single-cell co-expression (blue) of Aldh1a2 (red) with Clock (green) in Sertoli cells (c), Zbtb16 (red) with Rarg (green) in SPG undiff. (d), and Izumo1 (red) with Rxra (green) in spermatids (e). (f) Confocal IHC image shows co-localization of ALDH1A2 (green) and BMAL1 (red) in Sertoli cells (n = 3 × 3). (g) Fluorescent image of EGFP-labeled Sertoli cells of Amh-ROSA26-EGFP^fl/+ mice (n = 3 × 3). (h) Bioluminescence analysis of FACS-selected Sertoli cells (black), non-Sertoli cells (red) and mixed testicular cells (blue) from Amh-Rosa26-EGFP^fl/+; PER2: : LUC mouse testis (n = 3). (i) Bioluminescence recordings of the testes ex vivo from PER2: : LUC (black), PER2: : LUC; Bmal1^−/− (red), PER2: : LUC; Amh-Bmal1^−/− (cyan), and PER2: : LUC; Amh-Bmal1-RE ^fl/fl(green) mice (n = 3–6). (j–l) Fluorescent images of WT control, Bmal1^−/− KO, Amh-Bmal1^−/−, Amh-Bmal1-RE ^fl/fl, and Nms-Bmal1^−/− testes with ALDH1A2 (j) and KIT (k) antibodies, and their elongating spermatids (l) (n = 5–10). (m) Quantification of the images in (j–l) (n = 3–8). (n and o) Epididymal spermatozoa (n) of WT control, Bmal1^−/− KO, Amh-Bmal1^−/−, and Amh-Bmal1-RE ^fl/fl(n = 3), and fertilization rates of their sperms with WT oocytes by IVF (n = 3) (o). (p) IHC images of KIT in Bmal1^−/− KO testes 1 day after atRA treatment at ZT12 (n = 4). (q) Quantification of the KIT cells in p (n = 4). (f to g, j to l, and p) ALL nuclei were counterstained with Hoechst 33342. Arrowheads indicate Sertoli cells and arrows spermatogonia. Scale bar, 100 μm. *P < 0.05; **P < 0.01; ***P < 0.001 (Figs S6 and S7).

Temporally desynchronizing the circadian clock results in arrested spermatogonial differentiation and reduced fertilization in zebrafish and mice. (a and b) Confocal images of the testes from Tg(RARE-gata2a : NLS-EYFP; hsp70l : clock1a; CG2) and sibling males after heat shock at ZT0 (a) or ZT12 (b) for 7 consecutive days (n = 9–15). (c) FISH images of Tg(hsp70l : clock1a; CG2) and sibling male testes with kita or sycp3 probe, and confocal images of Tg(hsp70l : clock1a; CG2; piwil1 : mCherry) and sibling male testes, after heat shock at ZT0 or ZT12 for 7 consecutive days (n = 3 × 5). (d) Quantification of the images in (c) (n = 6 −18). (e–g) Fertilization rates by pairwise crosses of WT females and Tg(hsp70l : clock1a; CG2) males or sibling males following heat shock just once (e) or for 7 consecutive days (f) starting at ZT0 or ZT12 or 7 days after removing the treatment from heat shock starting at ZT0 or ZT12 for 7 consecutive days (g) (n = 3 × 10). (h) Fertilization rates of desynchronization and control groups’ sperms with WT oocytes by IVF (n = 3). (i) Heatmaps of loss-of-rhythmicity genes in the desynchronization and control groups. (j) Top-20 enriched GO terms BP of the loss-of-rhythmicity genes in the desynchronization group. (k) Expression of Cry1, Smcp, Zbtb16 and Cyp26b1 in the control (black) and desynchronization (red) groups. (a–c), Arrows indicate spermatogonia. Sg, spermatogonia; Sc, spermatocyte; Sz, spermatozoon. Scale bar, 100 μm. **P < 0.01; ***P < 0.001 (Figs S8 and S9).