Single-cell RNA seq reveals injury-induced the cxcl18b-defined MG transitional states

(A) Schematic showing the experimental procedure: five consecutive days of MTZ treatment in Tg(lws2: nfsb-mCherry x mpeg1: GFP) fish to ablate green and red (G/R) cone, starting at 6 dpf and continuing until 11 dpf. The MTZ solution was refreshed every 24 hours, followed by fish fixation for further immunostaining.

(B-D) Quantitative plots showing the dynamic changes in the number of G/R cone (B), recruited microglia (C), and proliferative MG (PCNA+) (D) at different time points after MTZ treatment (uninjured: collected retina number n=14; 24-hpi: n=10; 48-hpi: n=10; 72-hpi: n=11; 96-hpi: n=12; 120-hpi: n=16; mean ± SEM; ****p<0.0001, ***p<0.001, **p<0.01, ns, p>0.05; one-way ANOVA followed by Tukey’s HSD test).

(E) Representative images showing microglia recruitment (mpeg1: GFP, green), G/R cone ablation (lws2: nfsb-mCherry, red), and injury-induced MG proliferation (PCNA, white) in uninjured (E1) and 72-hpi (E2) retinas. The high-magnification images of the boxed area (E3-E4). Scale Bars: 20 μm (E1, E2), 10 μm (E3), 2 μm (E4).

(F) The UMAP plot of 4,172 MG cells was sorted with an increased proportion in response to the G/R cone ablation. Cells were further aggregated into 10 clusters based on previously published scRNA-seq data(Krylov et al., 2023).

(G) Pseudo-time developmental trajectory of MG states identified by Monocle2 analysis shows a main developmental branch originating from Cluster 4 (cx43+), which diverges into two sub-branches: Cluster 0 and Clusters 5/3/6 (pcna+).

(H) Violin plots showing the expression levels of key genes (cx43, glula, gfap, cxcl18b, ascl1α, mki67, pcna, and notch3) in the main developmental branch clusters, progressing from the most original MG states (Cluster 4) to transitional MG states (Clusters 1/2/5), and proliferative MG states (Clusters 3/6).

Clonal analysis reveals the proliferative MG mostly originated from cxcl18b+ MG transitional states

(A) Representative images show dynamic expression of cxcl18b (red, in situ hybridization) and PCNA (white, immunostaining) in Tg(lws2: nfsb-mCherry) retina at different time points following the G/R cone ablation. Scale bars: 20 μm.

(B) Quantitative plots showing the number of cxcl18b+ (red curve, significance shown above the curve) and PCNA+ MG (blue curve, significance shown below the curve) in uninjured (n=11) and injured retina at 24-hpi (n=10), 48-hpi (n=12), 72-hpi (n=11), 96-hpi (n=7), and 120-hpi (n=5). Each injured time point was compared to the uninjured retina (Mean ± SEM; ****p<0.0001, **p<0.01, ns, p>0.05; one-way ANOVA followed by Tukey’s HSD test).

(C) Schematic diagram of the cxcl18b promoter was used to construct the reporter fish line Tg(cxcl18b: GFP) and clonal analysis fish line Tg(cxcl18b: Cre-vmhc-mCherry:: ef1α: loxP-DsRed-loxP-EGFP; lws2: nfsb-mCherry).

(D) Immunostaining of injury-induced cxcl18b+ (green, indicated by Tg(cxcl18b: GFP)) and proliferative (PCNA+, white) MG showing overlapping in the central retina area (yellow arrows, GFP+/PCNA+ MG) at 48 hpi. The high-magnification images of the boxed area (D3’-D3’’’). The area of the retina is labeled with a dashed line, and each layer structure is labeled with dashed lines and marked with the outer nuclear layer (ONL), inner nuclear layer (INL), inner plexiform layer (IPL), ganglion cell layer (GCL) and lens. Scale Bars: 20 μm (D1-D2) and 5 μm (D3’-D3’’’).

(E) Quantitative plots showing the number of cxcl18b+ MG (red) and proliferative MG (PCNA+, blue) in (D1) uninjured (n=4) and (D2) 48-hpi retina (n=7) (Mean ± SEM; ****p<0.0001, ns, p>0.05; two-way ANOVA followed by Tukey’s HSD test).

(F) Clonal analysis of injury-induced cxcl18b+ MG in transgenic fish line Tg(cxcl18b: Cre-vmhc: mCherry:: ef1α: loxP-DsRed-loxP-EGFP:: lws2: nfsb-mCherry) at 72 hpi retina showing overlapping between proliferative (PCNA+, white) MG with cxcl18b+ (green, yellow arrows). The high-magnification images of the boxed area (F3’-F3’’’). Scale Bars: 20 μm (F1-F2) and 5 μm (F3’-F3’’’).

(G) Quantitative analysis at 72 hpi shows no significance in the number of proliferative MG (PCNA+, blue) and double-positive (PCNA+ / cxcl18b+, red) MG (n=14; Mean ± SEM; ns, p>0.05; Unpaired t-test) in (F), with 93±2% of PCNA+ MG also being cxcl18b+.

(H) Representative images show that not all mature MG stained with glutamate synthase (GS+, magenta) are cxcl18b+ (green, labeled by cxcl18b: Cre) in the central retinal area (white dashed lines identified a 45° angular region originating from the optic nerve). The high-magnification images of the boxed area (H2’-H2’’’). Scale Bars: 20 μm (H1) and 5 μm (H2’-H2’’’).

(I) Quantification of GS+/cxcl18b: Cre+ double-positive (blue, yellow arrows in H) and GS+/cxcl18b: Cre- single-positive (red, open yellow arrowheads in H) MG (n=14, Mean ± SEM; *p<0.05; one-way ANOVA followed by Tukey’s HSD test), and the proportion of cxcl18b: Cre+ or - MG within the total population of mature (GS+) MG in the central retina.

(J1-J3) Representative images showing injury-induced cxcl18b+ MG (green) in Tg(lws2: nfsb-mCherry x cxcl18b: GFP) fish retina treated with dexamethasone (Dex) or DMSO at 72 hpi. Scale bars: 20 μm.

(K-M) Quantitative plots showing the number of cxcl18b+ MG (72 hpi: n=14; DMSO: n=6, Dex: n=10) in J1-J3; recruited microglia in Figure 2-figure supplement 1J1-J3 and proliferative MG in Figure 2-figure supplement 1J4-J6(72 hpi: n=9; DMSO: n=7, Dex: n=14) in Tg(mpeg1: GFP; lws2: nfsb-mCherry) retinas after DMSO or Dex treatment at 72 hpi (Mean ± SEM; ****p<0.0001, ***p<0.001, ns, p>0.05; one-way ANOVA followed by Tukey’s HSD test).

The cxcl18b-defined MG transitional states recapitulate the developmental states of RSCs in the CMZ

(A, C) UMAP plots display 5,368 retinal progenitor cells (RPCs) at 24 hpf and 4,625 CMZ progenitor cells at 14 dpf. Clusters are indicated by their cluster-specific marker genes based on previously published scRNA-seq data(Xu et al., 2020).

(B, D) UMAP plots showing expression of fabp11a, cxcl18b, and col15α1b at r24-hpf RPCs (embryonic states) (B) and r14-dpf CMZ-progenitors (post-embryonic states) (D).

(E-H3) In situ hybridization images showing the expression of fabp11a (green), col15α1b (white), two putative makers for post-embryonic retinal stem cells (RSCs), and cxcl18b at 30-hpf (E-F3) and 14-dpf (G-H3) retina. The high-magnification images of the boxed area (F-F3, H-H3). The area of these three in situ signal trouble positive is labeled with a dashed yellow line. Scale Bars: 20 μm (E, G) and 3 μm (F-F3, H-H3).

The nitric oxide metabolic pathway regulates injury-induced MG proliferation

(A) Representative images of microglia recruitment (green, marked by Tg(mpeg1: GFP)), G/R cone ablation (red, marked by Tg(lws2: nfsb-mCherry)) and proliferative MG (white, marked by PCNA+) at 72 hpi with nitric oxide metabolism pathway genes disruption (nos1/nos2a/nos2b/gsnor). Scale Bars: 20 μm.

(B) Quantitative analysis of the number of proliferative MG (PCNA+) at 72 hpi in (A). In total, we collected 11 retinas for WT (n=11), scramble sgRNA-injection (n=7), nos1 sgRNA-injection (n=23), nos2a sgRNA-injection (n=15), nos2b sgRNA-injection (n=22), and gsnor sgRNA-injection (n=13) (Mean ± SEM; ****p<0.0001, **p<0.01, *p<0.05, ns, p>0.05; one-way ANOVA followed by Tukey’s HSD test).

(C) Schematic showing the experimental procedure of nitric oxide synthase (Nos) inhibitors injection or NO scavengers treatment in Tg(lws2: nfsb-mCherry x mpeg1: GFP) retina starting from 5 dpf for five consecutive days to 10 dpf with three consecutive days of MTZ treatment for G/R cone ablation from 7 dpf to 10 dpf. Nos inhibitors, NO scavengers, and MTZ solution were refreshed every 24 hours, and fish fixation was at 10 dpf for further immunostaining.

(D) Representative images of microglial recruitment (green, marked by Tg(mpeg1: GFP)), G/R cone ablation (red, marked by Tg(lws2: nfsb-mCherry)) and proliferative MG (white, marked by PCNA+) at 72 hpi following L-NMMA (10 mM), L-NAME (10 mM), 1400W (200 nM) intravitreal injection, and PBS as control, or C-PTIO (10 mM) treatment. Scale Bars: 20 μm.

(E) Quantitative plots showing the number of proliferative (PCNA+) MG at 72 hpi in (D). Retinas analyzed WT (n=11), PBS-injected (n=16), L_NMMA-injected (n=12), L_NAME-injected (n=12), 1400W-injected (n=10), and C-PTIO treatment (n=27) (Mean ± SEM; ****p<0.0001, ns, p>0.05; one-way ANOVA followed by Tukey’s HSD test).

(F-G) Representative images of proliferative MG (PCNA+, white) and G/R cone ablation (marked by Tg(lws2: nfsb-mCherry), red) at 72 hpi in heterozygous (nos1+/-, nos2a+/-, nos2b+/-, gsnor+/-) (F) and homozygous mutants (G) of nitric oxide metabolism pathway genes (nos1-/-, nos2a-/-, nos2b-/-, gsnor-/-). Scale Bars: 20 μm.

(H-I) Quantitative plots showing the number of proliferative MG (white, PCNA+) at 72 hpi in nos and gsnor mutant fish. In heterozygous (H), analyzed retinas include WT (n=14), nos1+/-(n=19), nos2a+/- (n=13), nos2b+/- (n=20), gsnor+/- (n=12). For homozygous (I), analyzed retinas include nos1-/-(n=18), nos2a-/- (n=20), nos2b-/- (n=20), gsnor-/- (n=27) (Mean ± SEM; ****p<0.0001, ns, p>0.05; one-way ANOVA followed by Tukey’s HSD test).

(J-K) Quantitative plots showing the number of proliferative MG (white, PCNA+) (J) and photoreceptor cells remain (K) at 72 hpi in nos2b hetero- or homozygous mutants. (Mean ± SEM; ****p<0.0001, ***p<0.001, ns, p>0.05; one-way ANOVA followed by Tukey’s HSD test).

RT-qPCR analysis reveals nos2b cell-specific expression in the injury-induced cxcl18b+ MG

(A) Schematic showing the workflow for isolating and enriching for three post-injury MG populations (72-hpi MG, 72-hpi PCNA+ MG, and 72-hpi cxcl18b+ MG) and three control groups (uninjured MG, 72-hpi retinal cells other than MG, and 72-hpi G/R cones) using fluorescence-activated cell sorting (FACS).

(B-C) RT-qPCR analysis of nos1, nos2a, nos2b, and gsnor expression in different cell populations. Expression levels are shown relative to 72-hpi G/R cones (B) and other 72-hpi retinal cell types (C). A total of six independent replicates were performed for cell population enrichment and cDNA template preparation (n=6, Mean ± SEM; ****p<0.0001, ns, p>0.05; two-way ANOVA followed by Tukey’s HSD test).

(D) RT-qPCR analysis comparing nos1, nos2a, nos2b, and gsnor expression in distinct MG states relative to uninjured MG (repeats n=7 in cxcl18b+ MG; n=5 in 72-hpi MG; and repeats n=6 in PCNA+ MG; n=6 in the uninjured retina; Mean ± SEM; ****p<0.0001, ns, p>0.05; two-way ANOVA followed by Tukey’s HSD test).

NO produced by nos2b in cxcl18b+ MG regulates injury-induced proliferation

(A-C) Representative images showing the adenovirus-mediated infection (green, indicated by Y2-GFP) specifically target MG (red, GS staining) in the zebrafish retina. The virus was intravitreally injected into the right eye of Tg(lws2: nfsb-mCherry x ef1α: loxP-DsRed-loxP-EGFP) fish (A2, C), with the left eye as a wild-type (WT) control (A1). The high-magnification images of the boxed area (B-B3). Scale Bars: 20 μm (A1, A2) and 5 μm (B-B3).

(D) Schematic showing the design of the cxcl18b+ MG-specific nos2b knockout system. The viral construct consists of three plasmids: (1) gal4 expression driven by the cxcl18b promoter; (2) UAS-derived Cas9 and Cre elements, and (3) U6 promoters driving two sgRNAs targeting nos2b, with a non-targeting sgRNA as the control.

(E) Schematic showing the procedure of injury process and intravitreal viral injection in Tg(lws2: nfsb-mCherry x ef1α: loxP-DsRed-loxP-EGFP) fish.

(F-G) Representative images showing proliferative MG (PCNA+, white) with cxcl18b+ MG specific knockout nos2b in (G) and control in (F), (GFP+, green, yellow arrows) are defined as virus-infected clones. Upper panels show WT retina (no virus injected), Bottom panels show retinas injected with the virus (two sgRNA targets as nos2b knockout and without sgRNA targets as control). Scale Bars: 20 μm.

(H) Quantification of proliferative (PCNA+/GFP+, red bars) and non-proliferative (PCNA-/GFP+, gray bars) MG clones in (F). For control, ∼75% of virus-infected clones entered the cell cycle (PCNA+; red bars), with 90/120 clones analyzed across 8 independent experiments (n=8). For nos2b knockout clones, ∼23% entered the cell cycle (PCNA+; blue bars), with 22/103 clones analyzed across 6 independent experiments (n=6) (Mean ± SEM; ****p<0.0001; two-way ANOVA followed by Tukey’s HSD test).

NO regulates MG proliferation by suppressing Notch signaling

(A) Dot plot showing the Notch signaling-related gene expression in different MG states. The average expression levels of these genes for all cells in each cluster are coded by the gray level. The percentage of cells expressing each gene within each cluster is coded by dot size.

(B) Representative images showing the dynamic changes of Notch signaling activity (green, indicated by Tg(Tp1: EGFP)) and proliferative MG (white, PCNA+) following injury with nitric oxide (NO) blockade using C-PTIO. Scale Bars: 20 μm.

(C-D) Quantitative plots showing the number of Notch activation (tp1: GFP+ MG) in (C) and proliferative MG (PCNA+ MG) in (D) at different time points post-G/R cone ablation. We collected uninjured retina (24-hpi, n=4; 48-hpi, n=4; 72-hpi, n=3), injured retina (24-hpi, n=9; 48-hpi, n=4; 72-hpi, n=6), and retina treated with C-PTIO (24-hpi, n=6; 48-hpi, n=5; 72-hpi, n=7) (Mean ± SEM; ****p<0.0001, ***p<0.001, **p<0.01, *p<0.05, ns, p>0.05; two-way ANOVA followed by Tukey’s HSD test).

Clusters with increased proportion are identified from the scRNA-seq data

(A) Representative images showing microglia recruitment (green, indicated by Tg(mepg1: GFP)), G/R cone ablation (red, Tg(lws2: nfsb-mCherry)), and injury-induced MG proliferation (white, marked by PCNA) following the injury process from uninjured retina to 120 hpi. Scale Bars: 20 μm.

(B) UMAP plot showing 5,932 cells from uninjured retinas (blue) and 3,999 cells from 72-hpi (red) retinas were obtained, which further aggregated into 13 clusters.

(C) Percentage of MGs from uninjured and 72-hpi retinas in each cluster. Eight clusters (6, 5, 9, 11, 12, 3, 2, 10) were identified and showed an increased proportion for further analysis.

(D) Dot plot showing the expression levels of marker genes for quiescent MG (qMG), reactive MG (rMG), proliferative MG (pMG), and differentiated MG (dMG). The average expression levels of these genes for all cells in each cluster are coded by the gray level. The percentage of cells expressing each gene within each cluster is coded by dot size.

Gene description of cxcl18b does not reduce MG proliferation

(A-A3) In situ hybridization images show injury-induced cxcl18b (red, marked by in situ hybridization) with MG-like morphology, cell-specific expressed in the BLBP+ cells (green, labeled by immunostaining) at 24 hpi in the transgenic fish line Tg(lws2: nfsb-mCherry) retina. In these clones, cxcl18b+ MG are red+/BLBP+ cells (yellow arrowheads). The high-magnification images of the boxed area (A1-A3). The area of the retina structure is labeled with a dashed line, and each layer structure is labeled with dashed lines and marked with the outer nuclear layer (ONL), inner nuclear layer (INL), inner plexiform layer (IPL), ganglion cell layer (GCL) and lens. Scale Bars: 20 μm (A) and 5 μm (A1-A3).

(B-B3) Representative images showing cxcl18b (red, in situ hybridization) co-expressed with PCNA (white, immunostaining) at 72 hpi. In these clones, cxcl18b+ MG was identified as red+/PCNA+ cells (yellow arrowheads). The high-magnification images of the boxed area (B1-B3). Scale Bars: 20 μm (B) and 5 μm (B1-B3).

(C) Schematic showing the design of cxcl18b report fish line and the Cre-loxP transgenic fish line used for clone analysis. The cxcl18b promoter drives GFP or Cre expression, with the heart labeled by mCherry under the vmhc promoter for fish screening.

(D-E3) Representative images showing cxcl18b expression labeled by Tg(cxcl18b: GFP) (green) in (D-D3) at 48 hpi, or three transgenic fish Tg(cxcl18b: Cre-vmhc: mCherry:: ef1α: loxP-DsRed-loxP-EGFP:: lws2: nfsb-mCherry) (green) in (E-E3) at 24 hpi fish retina, merged with cxcl18b in situ hybridization signal (red) (D-E3). The high-magnification images of the boxed area (D1-D3, E1-E3). Scale Bars: 20 μm (D, E) and 5 μm (D1-D3, E1-E3).

(F) Schematic showing the two sgRNA target sites of the cxcl18b genome.

(G) Summary table showing the indel and knockout efficiency of the two sgRNAs targeting cxcl18b.

(H) Representative images of proliferative MG (PCNA+, white) at 72 hpi in Tg(lws2: nfsb-mCherry) fish retina with cxcl18b disruption. Scale Bars: 20 μm.

(I) Quantitative plots showing the number of proliferative MG (PCNA+) at 72 hpi in H. In total, retinas were collected from wild-type (WT, n=11), scramble sgRNA-injected (scrambled, n=7), and cxcl18b sgRNA-injected (cxcl18b KO, n=7) (Mean ± SEM; ns, p>0.05; one-way ANOVA followed by Tukey’s HSD test).

(J1-J6) Representative images showing injury-induced microglia recruitment (green, marked by Tg(mpeg1: GFP)) (J1-J3), and proliferative MG (white, marked by PCNA) (J4-J6) after dexamethasone (Dex) or DMSO treatment in the in Tg(mpeg1: GFP x lws2: nfsb-mCherry) fish retina at 72 hpi. Scale bars: 20 μm.

UMAP plots show the expression of cluster-specific marker genes utilized to identify developmental states

(A, B) UMAP plots showing the expression pattern of embryonic developmental stage cluster-specific marker genes (npm1a, her9, her4.2, fabp7a, dla, and atoh7) in (A) and post-embryonic developmental stage marker genes (her4.2, atoh7, vsx1, mafba, otx5, and rem1) in (B).

(C) UMAP plots showing the expression pattern of fabp11a, cxcl18b, and col15α1b at MG clusters with increased population after G/R cone ablation at 72 hpi in Figure 1F.

(D) Representative images showing the cxcl18b expression (green, marked by Tg(cxcl18b: GFP)) are located in the most peripheral region of the CMZ (white, marked by PCNA) at 5 dpf. The high-magnification images of the boxed area (D1-D3). Scale Bars: 20 μm (D) and 5 μm (D1-D3).

Genetic disruption of the nitric oxide pathway modulates G/R cone ablation and microglial recruitment

(A) Dot plot showing the redox gene expression profiles of regulating glutathione (GSH), hydrogen peroxide (H2O2), nicotinamide adenine dinucleotide phosphate (NADPH), nitric oxide (NO), metal ion, and sulfhydryl group (-SH) in the uninjured and cxcl18b+ MG in 72-hpi scRNA-seq data (Krylov et al., 2023). The average expression levels of these genes for all cells in each cluster were coded by the gray level. The percentage of cells expressing each gene within each cluster was coded by dot size.

(B) Schematic showing the two sgRNA target sites of the nitric oxide metabolism pathway genome (nos1/nos2a/nos2b/gsnor).

(C) Summary table showing the indel and knockout efficiency of the two sgRNAs targeting nos1/nos2a/nos2b/gsnor.

(D) Representative images of microglial recruitment (marked by Tg(mpeg1: GFP), green) and G/R cone ablation (marked by Tg(lws2: nfsb-mCherry), red) at 72 hpi with nos1/nos2a/nos2b/gsnor disruption, with scrambled sgRNA-injected as the control.

Scale Bars: 20 μm.

(E-F). Quantitative plots showing the number of recruited microglia (GFP+) (D) and G/R cone ablation (mCherry+) (E) at 72 hpi with nos1/nos2a/nos2b/gsnor disruption. In total, we collected 11 retinas for WT (n=11), scrambled sgRNA-injected (n=7), nos1 sgRNA-injected (n=10), nos2a sgRNA-injected (n=6), nos2b sgRNA-injected (n=12), and gsnor sgRNA-injected (n=13) (Mean ± SEM; *p<0.05; ns, p>0.05; one-way ANOVA followed by Tukey’s HSD test).

The genotype of nitric oxide metabolism pathway mutants

(A) Sequence alignment of WT nitric oxide metabolism pathway genes (nos1/nos2a/nos2b/gsnor) and their CRISPR-cas9-induced mutant alleles.

(B) Summary table showing the premature stop codon and unknown polypeptides sides of each mutant.

In situ hybridization reveals nos2b cell-specific expression in the cxcl18b-defined transitional MG states

(A-A3) Representative in situ hybridization images showing nos2b expression (red, in situ hybridization) cell-specific in the injury-induced cxcl18b+ MG (GFP+, green) in Tg(lws2: nfsb-mCherry x cxcl18b: GFP) fish retina at 72 hpi. Clones with green/red/white trouble-positive are cxcl18b+/nos2b+/PCNA+ MG cells (yellow arrowheads). The high-magnification images of the boxed area (C-C3). Scale Bars: 20 μm (A, B) and 5 μm (C1-C3).

(D-E) RT-qPCR analysis of glula (pink bar) and glulb (green bar) expression in different cell populations relative to 72-hpi G/R cone (D), or other retinal cell types (E). The significance of 72-hpi G/R cone as control is noted by * for comparisons with 72-hpi G/R cones and # for comparisons with uninjured MG. In total, we do 6 independent replicates to enrich the different cell populations and get the cDNA templates (Mean ± SEM; **** and ####p<0.0001, **p<0.001, **p<0.01, ns, p>0.05; two-way ANOVA followed by Tukey’s HSD test).

(F) RT-qPCR analysis of cxcl18b (blue bar) and pcna (orange bar) expression in different MG states. (Mean ± SEM; ****p<0.0001, **p<0.01, *p< 0.05, ns, p>0.05; two-way ANOVA followed by Tukey’s HSD test).

Analysis of cell-specific knockout efficiency for nos2b in virus-infected clones

(A) Schematic showing the cell-specific gene knockout system design. Injury-induced cxcl18b promoter directs the expression of gal4, which in combination with the UAS element, drives the cell-specific expression of Cas9 and Cre. The U6 promoter enables the broad expression of sgRNAs targeting nos2b, and Cas9 mediates gene knockout within the targeted cells through these sgRNAs. The ef1α promoter controls the loxP-DsRed-loxP-EGFP reporter system, allowing for the switch from DsRed to EGFP upon Cas9 and Cre activity, and GFP+ MG clones are identified with successful nos2b knockout. This system operates with high specificity via double adenovirus infection.

(B) Schematic showing the procedure of virus-infected clones are sorted and enriched by FACS for further knockout efficiency analysis.

(C) Representative gel electrophoresis image showing nos2b deletions in knockout cells compared to controls. Single-virus injections were used to assess Cre system leakage. While minimal Cre system leakage was observed, it did not result in detectable nos2b knockout.

(D) Summary table showing the knockout efficiency of each virus-injected clone.

sgRNA sequences and genotyping primers, related to STAR Methods.

Plasmid construction primer sequences, related to STAR Methods.

qPCR primer sequences, related to STAR Methods.