In vivo CRISPR screening directly targeting testicular cells

Abstract

CRISPR-Cas9 short guide RNA (sgRNA) library screening is a powerful approach to understand the molecular mechanisms of biological phenomena. However, its in vivo application is currently limited. Here, we developed our previously established in vitro revival screening method into an in vivo one to identify factors involved in spermatogenesis integrity by utilizing sperm capacitation as an indicator. By introducing an sgRNA library into testicular cells, we successfully pinpointed the retinal degeneration 3 (Rd3) gene as a significant factor in spermatogenesis. Single-cell RNA sequencing (scRNA-seq) analysis highlighted the high expression of Rd3 in round spermatids, and proteomics analysis indicated that Rd3 interacts with mitochondria. To search for cell-type-specific signaling pathways based on scRNA-seq and proteomics analyses, we developed a computational tool, Hub-Explorer. Through this, we discovered that Rd3 modulates oxidative stress by regulating mitochondrial distribution upon ciliogenesis induction. Collectively, our screening system provides a valuable in vivo approach to decipher molecular mechanisms in biological processes.

Keywords: CRISPR screening; Hub-Explorer; in vivo genome-wide screening; multi-omics; spermatogenesis.

Graphical abstract

Figure 1

Establishment of in vivo genome-wide screening (A) Experimental scheme of in vivo genome-wide screening toward spermatogenesis. (B) Evaluation of short-term infection efficiency. Lentivirus encoding TagBFP was injected into 11 PND mice testes. One week later, cryo-sectioned testes were stained with anti-TagBFP (cyan) and anti-Gcna1 (magenta) antibodies and DAPI (white), followed by confocal microscopy analysis. Top: virus uninfected; bottom: virus infected. Left: macroscale; right: microscale. Scale bar: 200 μm. (C) Sanger sequencing analysis against Rosa26 region in Basigin (Bsg)⁺ germ cells. Lentivirus encoding sgRosa26 and TagRFP was introduced into 11 PND mice testes. Ten days later, male germ cells were stained with anti-Bsg antibody. TagRFP⁺ and Bsg⁺ germ cells were sorted for gDNA extraction. Indel start position was detected by CRISP-ID, shown as an orchid-colored line. Experiment was repeated twice. (D) Gcna1 knockout (KO) efficiency evaluation. Lentivirus encoding sgGcna1 and TagBFP was introduced into 11 PND mice testes. One week later, dissociated testes were stained with anti-Plzf, anti-Sall4, and anti-Gcna1 antibodies and DAPI. Analyzed cells are shown on the left. Gcna1 signal intensity was evaluated by median fluorescent intensity (MFI) as mean ± SD. Middle: type A spermatogonia (Plzf⁺/Sall4⁺); right: primary spermatocyte (4 chromosomes). Light orchid: uninfected cells; dark orchid: infected cells. Infected samples in type A spermatogonia were analyzed twice (n = 2), and the others were analyzed three times (n = 3). (E) Evaluation of sperm Ca²⁺ influx. Adult mice sperms were capacitated (top; see the STAR Methods) and stained with Fluo-4 AM and PI. Fluo-4 AM signal in PI-negative region was analyzed. Bottom left: non-capacitated sperms; bottom right: capacitated sperms. Experiments were repeated three times (n = 3), and Fluo-4 AM-positive populations were shown as mean ± SD.

Figure 2

Enrichment of essential sgRNAs associated with spermatogenesis (A and B) Library coverage evaluation. Pool B GeCKO-v2 sgRNA library containing 62,804 sgRNAs was introduced into 11 PND mice testes. Three days later, dissociated testes were stained with anti-Plzf and DAPI to sort Plzf⁺ type A spermatogonia for gDNA extraction. The amplified sgRNA regions were sequenced with the next-generation sequencing (NGS). The sgRNA clone coverage is shown in (A) and the sgRNA-targeted gene coverage is shown in (B). sgRNAs from are from input (blue), 9 testes (dark green), and 3 testes (light green). x axis: the sgRNA count (=log₂(CPM + 1)); y axis: the cumulative sgRNA count. (C and D) Sorting diagram of Fluo-4 AM-negative sperms. Left column: pool A; right columns: pool B. (C) First screen. (D) Second screen. (E and F) Significant gene enrichment. Significant genes were extracted above log₂ fold change = 1.5 and enrichment score = 2.0. Candidates in the first library are shown as light blue dots (E) and those in the second library are shown as dark blue dots (F). Enrichment score was calculated as described in the STAR Methods. (G) Enriched candidate genes. Light blue bar: first screening; dark blue bar: second screening. Count shown as log₂(CPM + 1). Fold change (count of second screening/first screening) is shown as colored dots (orchid: enriched genes; light orchid: non-enriched genes).

Figure 3

Rd3 is an essential spermatogenesis factor identified through small-scale functional screening (A) Experimental scheme of shRNA screening. (B) Transcriptome analysis in testes. Each gene expression shown as log₂(TPM + 1) (see the STAR Methods). (C) Testicular weight quantification. Data are shown as violin plot according to the following replicates: n = 24 (shLacZ), 7 (shOvol2), 10 (shRnf215), 22 (shCldn34c4), 15 (shRd3), 8 (shCebpg), 7 (shRbm26), and 10 (shP2ry2). p value: unpaired t test. Significant values (p < 0.05) are colored orange. (D) Sperm number quantification. Data are shown as the following replicates: n = 24 (shLacZ), 8 (shOvol2), 10 (shRnf215), 22 (shCldn34c4), 15 (shRd3), 8 (shCebpg), 7 (shRbm26), and 10 (shP2ry2). Statistical evaluation is as described in (C). (E) Sperm Ca²⁺ influx quantification. Ca²⁺ influx score is as described in Figure S3F. Data are shown as the following replicates: n = 24 (shLacZ), 8 (shOvol2), 10 (shRnf215), 22 (shCldn34c4), 15 (shRd3), 8 (shCebpg), 8 (shRbm26), and 10 (shP2ry2). Statistical evaluation is as described in (C). (F) Histograms for the representative Fluo-4 AM signal intensity.

Figure 4

Round spermatid is the specific cell type with high Rd3 expression in testes (A) Tissue-wide transcriptome analysis of Rd3. Adult mice-derived Rd3 expression is shown as log₂(TPM + 1) values. Testis: magenta; retinal neural layer: midnight blue. (B) t-Distributed stochastic neighbor embedding (tSNE) plot of mouse testicular cells. Single-cell data (total: 6,693 cells) are clustered as the following cell types: spermatogonium (SG), pre-leptotene/zygotene (pL/Z), pachytene (P), diplotene (D), meiotic cell (M), round spermatid (RS), and elongating spermatid (ES). (C) Rd3 expression profiling. x axis: cell types (total: 6,644 cells); y axis: Rd3 expression shown as Z score converted from ln(CPM + 1). (D and E) Rd3 expression on different ages. (D) RT-qPCR. (E) RT-PCR. (F) Hybridization-chain-reaction-based fluorescent in situ hybridization (HCR-FISH). Cryo-sectioned testes were probed with Acrv1 (magenta), Rd3 (green), DAPI (white), and lectin PNA (yellow), detected by confocal microscope. Scale bar: 200 μm. (G) Immunohistochemistry of Rd3. Cryo-sectioned adult testes were stained with DAPI (white), lectin PNA (magenta), and Rd3 (green), detected by confocal microscope. Scale bar: 50 μm.

Figure 5

Identification of ciliogenesis-associated Rd3-mitochondria axis by Hub-Explorer analysis (A) Proteomics for RD3 interactors. Immunoprecipitates obtained by anti-Spot nanobody were analyzed by liquid chromatography-tandem mass spectrometry (LC-MS/MS). Significant proteins (magenta) are shown as p < 0.05 and log₂ fold change (RD3-Spot-expressing cells/RD3^−/− cells) ≧1.5. Experiment was repeated three times (n = 3). (B) Gene expression profile of 269 significant proteins in testicular cells. 269 proteins gene expression levels shown as Z score from testicular gene expression matrix used in Figures 4B, 4C, and S4A. x axis: cell types; y axis: genes. Colored dots indicate the following cell types: SG, pL/Z, P, D, M, RS, and ES. (C) Cell type classification of 269 proteins on gene expression pattern in testis. Highly expressing cell types were determined by the cell type index (CTI ≧ 0.98; see STAR Methods), shown on the y axis, and the count of proteins in each cell type is shown on the x axis. Orange-colored bars: specific group including spermatid. (D) Gene Ontology (GO) analysis. 49 proteins analyzed by DAVID software with mouse (left column) and human (right) Ensembl IDs. (E) Computational process of Hub-Explorer analysis. (F) Similarity profile of 17 Rd3 interactors with hub components (Benjamini-Hochberg false discovery rate [BH-FDR] < 0.05) according to Jaccard similarity index visualized as a heatmap (both axes: 17 interactors). Each cluster is colored by 4 individual colors (C0: wine red, C1: magenta, C2: deep sky blue, and C3: blue). The annotated hierarchical cluster was used for the benchmark of K-means clustering. (G) Profile of hub components and core signaling pathways. x axis: 17 candidates clustered and labeled as indicated in (F); y axis: hub components. Core signaling pathway: scarlet; others: Safrano pink. Numbers on the y axis indicate hub components as follows: 1, CatSper complex; 2, sperm head plasma membrane; 3, ∗centrosome; 4, membrane protein complex; 5, plasma-membrane-bounded cell projection; 6, ∗sperm midpiece; 7, ∗9 + 2 motile cilium; 8, non-membrane-bounded organelle; 9, ∗microtubule cytoskeleton; 10, intracellular non-membrane-bounded organelle; 11, protein-DNA complex; 12, ∗sperm flagellum; 13, ∗motile cilium; 14, nuclear-protein-containing complex; 15, ∗cilium; 16, nucleosome; 17, DNA packaging complex; 18, ∗cytoskeleton; 19, ribonucleoprotein complex; 20, protein-containing complex; 21, phosphatase complex; 22, ∗ciliary basal body; 23, ∗dynein complex; 24, ∗microtubule; 25, catalytic complex; 26, ∗axonemal microtubule; 27, protein serine/threonine phosphatase complex; 28, membrane; 29, ubiquitin ligase complex; 30, cytoplasmic microtubule; 31, cell projection; 32, protein phosphatase type 2A complex; 33, cellular anatomical entity; 34, intracellular-protein-containing complex; and 35, spliceosomal complex. ∗Core signaling pathway (scarlet). (H) Intracellular localization of 17 candidate interactors with hub components in late round spermatids and early elongating spermatids.

Figure 6

RD3 modulates mitochondrial spatial distribution under ciliogenesis-induction-derived oxidative stress (A) RD3 and mitochondria co-localization study. SPOT-tagged RD3 SH-SY5Y cells stained with anti-RD3 (red) and Alexa 488-conjugated Tomm20 (green) antibodies with DAPI (white) and imaged by confocal microscope. Scale bar: 5 μm. Cell: SH-SY5Y. (B and E) Working hypothesis. (C) Quantification of ciliogenesis frequency. Primary cilium frequency quantified using filtered sections (30–50 cells per section). RD3-wild type (WT): 803 cells (navy); KO: 1,269 cells (cobalt blue); and overexpression (OE): 1,141 cells (light green) are shown as a violin plot. Each frequency per section is shown as white dots. p value: an unpaired t test. p < 0.05: orange. Experiments were repeated three times (n = 3). (D) Cilium length quantification. Primary cilium length quantified by Nikon NIS-Element software using filtered sections. RD3-WT: 71 cilia (navy); KO: 106 cilia (cobalt blue); and OE: 72 cilia (light green) are shown as a cumulative plot. Statistical analysis was conducted as described in (C). (F) Mitochondria distribution imaging. Ciliogenesis was induced by serum starvation. Cells were stained with anti-α-tubulin (magenta), Alexa 488-conjugated Tomm20 (green), and Alexa 647-conjugated γ-tubulin (yellow) antibodies and DAPI (white), imaged by confocal microscope. Scale bar: 20 μm. (G) Mitochondrial distribution quantification. The mitochondria distribution index calculated at 5 different spots is defined by the distance from the γ-tubulin signal. Quantification was performed by around 30 cells per replicate. Quantifications were conducted three times (n = 3). p value: an unpaired t test. p < 0.05: orange. (H) Mitochondrial distribution differences. Distribution difference between RD3-WT and KO is shown as barplot. p value: an unpaired t test. (I) Mitochondrial H₂O₂ evaluation. MitoPY1 and Hoechst33342 signals measured by plate reader. Raw MitoPY1 signal was normalized by Hoechst33342, multiplied by 1 × 10⁴, and shown as a bar plot. p value: an unpaired t test. p < 0.05: orange. Experiments were repeated five times (n = 5). (J) Summary of Rd3-mitochondria spatial dynamics under cilium-induction-oriented oxidative stress.