Animals and tissue collection
This study was granted by the Animal Care and Use Committee of Shanghai General Hospital (Shanghai, China). We used wide-type (WT) C57BL/6 mice (both male and female, the Experimental Animal Center of the Second Military Medical University, Shanghai, China). We obtained Crybb2-/- mice from the Ingenious Targeting Laboratory (Ronkonkoma, NY, USA). We maintained the mice in a pathogen-free room with 12-h light/dark cycles and free access to food and water ad libitum. The study was performed in accordance with the ethical standards of the 1964 Declaration of Helsinki and later amendments.This study was conducted in accordance with institutional guidelines and approved by the Animal Care and Use Committee, Wusong Center Hospital (permission number 2020-Y-10).
We anesthetized six months old WT and Crybb2-/- mice (n=3 per group) with pentobarbitone (150 mg/kg), dissected their lens, immediately frozen them in liquid nitrogen and kept at -80°C.
RNA extraction and RNA-Seq
We extracted total RNAs from the samples and digested ribosomal RNA using the ribo-zero kit (Illumia，Cat.No.RS-122-2301). After determined the RNA concentrations using the NanoDrop ND-2000 (ThermoFisher Scientific, Waltham, MA, USA), we generated a RNA library and broke the circRNA into short fragments. RNA-seq library preparation was carried out using the NEBNext® Ultra™ Directional RNA Library Prep Kit for Illumina.Subsequently, we reversely transcribed the RNA fragments into cDNA and evaluated its quality in the Agilent 2100 Bioanalyzer. Finally, we sequenced the cDNA library in an Illumina HiSeqTM 2500 system following the vendor’s recommended protocol (TruSeqR RNA Sample Preparation v2 Guide, Illumina Company Ltd). HTSeq software was used to count the reads numbers mapped to each circRNA.
CircRNA sequencing analysis
We used the NGSQC Tool kit (version number v2.3.3) to evaluate the quality control and removed the low-quality reads to obtain high-quality clean reads. Simultaneously, we used the FASTQC2 (version number v0.10.1) software to evaluate the data quality. Next, we employed the BWA MEM algorithm (version 0.7.5a) to align clean data against the reference gene sequence and detected circRNA using the CIRI software. According the circRNA position and the protein-coding gene annotation in the database, we obtained the sequences and counted the number of circRNA in each sample. We statistically analyzed the circRNAs for their length distribution, the chromosome distribution, and the number of circRNA exons. Data are expressed as RPM4 (reads per million mapping) in box plots and Venn diagrams using online tools (www.omicshare.com/tools/Home/Soft/box and www.omicshare.com/tools/Home/Soft/venn).
Unsupervised hierarchical clustering and PCA
We employed the Deseq25 package to perform principal component analysis (PCA), and investigate the distribution and relationship between the samples. In addition, we clustered the samples to investigate their similarity. We also performed unsupervised hierarchical clustering analyses using the gmodels in R (https://cran.r-project.org/web/packages/gmodels/index.html) and PCA using the ggplot2 in R (https://cran.r-project.org/web/packages/ggplot2/index.html)  .
GO and KEGG pathway analyses
GO annotations are useful for predicting the functions of gene products across numerous species. The GO categories were derived from the GO database. KEGG is a main public pathway-related database. The KEGG database is used to perform Pathway analysis of genes from different circRNA sources. We estimated the molecular function (MF), biological process (BP) and cellular component (CC) of differentially expressed circRNAs by GO analysis and predicted the functional enrichment of differentially expressed circRNAs by the KEGG pathway analysis. We determined the differentially expressed circRNAs using the edgeR package, based on FC≥2 and p<0.05, and generated volcano plots using the ggplot2 in R.
Prediction of circRNA-miRNA interactions
As circRNAs were shown to have an impact on miRNA-mediated regulation of gene expression through miRNA sequestration，we constructed a network of differentially expressed circRNAs using the TargetScan and Miranda.We selected the high confidence miRNAs in humans. Hypergeometric distribution test was used to identify the miRNAs that have greater influence in the differential circRNA. The calculated result returned an enrichment significance p value. The small p value indicated that the differential circRNA was enriched in the miRNA. For the enrichment results of total differential circRNA, we sorted by p value, top 300 miRNA-circRNA interaction pairs with smaller p value were extracted, and predicted a circRNA-miRNA regulation network using the Cytoscape (V3.2.0).
Quantitative real-time polymerase chain reaction (qRT-PCR)
We extracted total RNA from the lens using TRIzol reagent (Invitrogen) and reversely transcribed each sample into cDNA using the RevertAid First Stand cDNA Synthesis Kit (GeneCopoeia, Cat AORT-0020). The relative levels of each differentially expressed circRNA transcript to the control GAPDH were determined by qRT-PCR in the LightCycler480 Ⅱ Real-time PCR Instrument (Roche, Swiss) using the QuantiFast SYBR Green PCR Master Mix (Qiagen, Germany) and specific primers. We performed the PCR reactions in triplicate as the following protocol: 95℃ for 5 min, and 40 cycles of 95℃ for 10 s, 60℃ for 30 s. The primer sequences were circRNA_03118: Forward 5’-GTCTTCTTTGGTGGAACCTACA-3’, Reverse 5’-TTGGAGAAATGGTGCTTTGC-3’; circRNA_05302: Forward 5’-CCACATGGAGACAGCAGA-3’ Reverse 5’-GATAAGCCACTTCGTCCC-3’;
circRNA_09966: Forward 5’- CATAGCAGGAGTGGAGAG-3’ Reverse 5’-AAACAGCTACAGGAAACCTT-3’; circRNA_10276: Forward 5’-GTGAGCCCTGGCTTTGA-3’, Reverse 5’-TACTTTGTCCTGTTTGGCTG-3’;
circRNA_14121: Forward 5’-CATCCGGAGACCCATACTG-3’, Reverse5’-GGATCAAACACTTCTGAAGCCA-3’; circRNA_16505: Forward5’-CCACAACAACAGGAAACG-3’, Reverse 5’-GCTGGTGTAACACTGAAGAT-3’; circRNA_34578: Forward 5’-CTGGCACTTCAAGTCTTCAT-3’, Reverse 5’-AAGCCAAGAGCTGCTTTAG-3’; circRNA_36601: Forward 5’-GCTACGAGCAGGCTAATTG-3’, Reverse 5’-CTGCGTCCCACTTTGATG-3’; GAPDH: Forward 5’-TCATCCCAGAGCTGAACG-3’, Reverse 5’-TCATACTTGGCAGGTTTCTCC-3’. The data were analyzed by the 2-ΔΔCt method (Livak and Schmittgen, 2001).
The data were analyzed by Student’s t-test using R software version 3.2.1(http://www.r-project.org/). A two-side P-value of < 0.05 was considered statistically significant.
RNA‐seq data were submitted to Sequence Read Archive (SRA) under accession number PRJNA728889 (https://www.ncbi.nlm.nih.gov/sra/PRJNA728889).