Supporting

The cell lysate for soluble Top1 fraction was prepared. Western blot was performed using 2.5, 5 or 10 ul of the cell lysate. The intensity of the bands visualized by anti-Top1 antibody was measured by ImageJ. As a result, the measured Top1 protein amounts in 2.5, 5

The sequence of constructed plasmids was verified using Sanger sequencing (BigDye Terminator v3.1 and 3130xl Genetic Analyzer, Thermo Fisher Scientific).CH12-dCas9-KRAB#40 cells were transiently transfected with the gRNA plasmids, stimulated using CIT stimulation 48 hours after transfection and recovered at 24 hours after stimulation for evaluation of IgA expression by flowcytometry and knockdown efficiency by RT-PCR.RNA was purified by Sepazol RNA Super II (Nacalaitesque) or Maxwell® RSC simplyRNA Cells Kit (Promega).cDNA was synthesized using SuperScript® IV Reverse Transcriptase (ThermoFisher Scientific).Sequences of RT-PCR primers are provided in the Dataset S5.

Assay for IgA switching
Staining of the cells and evaluation by FACS Calibur (BD Biosciences) were performed as previously described (9).In general, the majority of non-stimulated cells of cultured CH12 cells remain IgM positive, with less than 1.5% IgA positive cells.Surface IgA expression in AIDactivated CH12 cells is due to genetic recombination between Sµ and Sa regions as shown in the literature (10).By these reasons, AID-dependent increase of surface IgA expression is described by using the term of class switch recombination (CSR) in this manuscript.

PAR-CLIP
The PAR-CLIP (Photoactivatable-Ribonucleoside-Enhanced Crosslinking and Immunoprecipitation) method was performed according to previous work (13) with some modifications.293T-AID-ER and 293T-KSS-ER were established through stable-transfection with AID-ER inserted in pEF1alpha-IRES-ZsGreen1 vector (Clontech) or KSS-ER inserted in pEF1alpha-IRES-DsRed-Express2 vector (Clontech).These plasmids were constructed by amplifying AID-ER and KSS-ER using the primers listed in Dataset S5 and inserted to the vectors with the restriction enzyme sites NheI and EcoRI.The cells were incubated with 100 µm 4-SU for 13 hours, exposed to UV 365 nm, and harvested.After IP and bead washing, protein and RNA were extracted from 5% and 95% of the recovered beads, respectively.RNA extraction and cDNA preparation methods were as described in the section of FA-RNA-IP.
Top1 mRNA Amplicons were prepared using the human Top1 mRNA-specific primers.
The PCR primers used in this study are listed in Dataset S5.The small RNA libraries from 293T-AID-ER and 293T-KSS-ER obtained by anti-Ago2 antibody were initially constructed using Ion Total RNA-seq Kit (ThermoFisher Scientific) following the protocol for "the amplified small RNA library" which includes the size-selection steps.These libraries were converted for sequencing by SOLiD 5500xl, using the specific primers including the P2 sequence, the barcodes (003 for AID-ER and 004 for KSS-ER), IA sequence and the Primer A for aligning the Ion library (total 79 bp).
The used primer's sequences are described in Dataset S5.The distal 110 nt-region of 3'UTR was omitted due to PCR difficulty and the Top1 5'UTR was not analyzed because of a poor conservation within the vertebral species.

Sequencing and analysis of the IPed RNAs
The prepared Top1 mRNA amplicon libraries were sequenced using ION Torrent (ThermoFisher Scientific) and the automatically processed data by TMAP tool was further manipulated by SAMtools to count nucleotide variation in Top1 mRNA sequences in the Center for Genomic Medicine, Graduate School of Medicine, Kyoto University (Fig. 1D-E, Fig. S3 and Dataset S1-2).
The sequencing of the RNA without immunoprecipitation with anti-Ago2 antibody was performed with Sanger sequencing (3500Genetic analyzer, Thermo Fisher Scientific) (Fig. 1F-G,

Dataset S3).
SOLiD sequencing (ABI) for Ago2-binding small RNAs was also performed in the Center for Genomic Medicine.The data obtained by SOLiD 5500xl was trimmed for removing adaptor sequences using BLASTN and mapped to hg19 genome using BWA software.The mapped reads were grouped into the "clusters" consisted of more than 50 reads and used for the subsequent analysis to pick-up the candidate miRNAs binding to Top1 mRNA (Fig. S13, Dataset S4).

Top1 3'UTR KO by CRISPR/Cas9
Four plasmids were constructed using pX335-U6-Chimeric BB-CBh-hSpCas9n (D10A) (Addgene).Since CH12 cells were revealed to harbor three copies of the Top1 gene, three different targeting vectors were prepared using the selection markers of zeocin-and blasticidin-resistant genes, and the ZsGreen gene.Each allele of 3'UTR of Top1 gene was serially targeted (Fig. S4A-B).The three complete knockout clones of Top1 3'UTR were obtained and named 3'UTRKO-A, -B, and -C (Fig. S4B).The recombination of each clone was confirmed using conventional PCR (Fig. S4C).The 3'UTRKO-A line lost zeocin resistant gene cassette during the final recombination, which was replaced with the BSD resistant gene or the ZsGreen cassette.To observe the effect of 3'UTR knockout without the influence of endogenous AID, exogenous and inducible AID-ER was introduced.

Western blot analysis
Western blot was performed as described in previous work (9)  Chemiluminescence was detected using Chemi-Lumi One L or Chemi-Lumi One Super (Nacalaitesque) with an ImageQuantLAS4000 system (cytiva).

SHM analysis
Wild-type and Top1 3'UTR KO CH12 cells with AID-ER and with or without 4-OHT stimulation were used for SHM analysis.DNA was harvested 48 hours following the start of 4-OHT.399 bp of the 5' of Sµ core region were amplified using Q5 DNA polymerase (NEB).The primer sequences are displayed in Dataset S5.The PCR products were treated with ExoSAP-IT (Thermo Fisher Scientific) and purified with Agencourt AMPure XP beads (Beckman).Purified PCR products were cloned into pCRII Blunt-TOPO using a TOPO cloning kit.Subsequent steps follow a previously described protocol (15).

DNA break assay with biotinylated d-UTP by terminal deoxy-transferase (TdT)
This assay is performed following the published method with minor modifications (16,17).Briefly, the fixed 5 × 10

Polysome fractionation
CH12 with AID-ER and CH12 Top1 3'UTR KO-C cells were stimulated by CD40L and IL-4 for 24 hours.0.1 µg/ml CHX (Sigma C1988) was added to the cell culture medium 5 min prior to the end of cell culture.Cells were washed by PBS containing 0.1 µg/ml CHX twice and lysed in a lysis buffer (20 mM HEPES-KOH, pH 7.5, 10 mM MgCl2, 0.1 M KCl, 1 mM DTT, 0.25% NP-40, 1 mg/ml CHX, RNaseOUT (LifeTechnologies)) and cOmplete EDTA-free (Roche).After vigorous syringe and needle homogenization, cell lysates were cleared by centrifugation at 15,000 rpm for 5 min at 4˚C.Cleared lysates were loaded onto a 10 -45% sucrose gradient made by Gradient Master (Biocomp) and centrifuged at 23,400 rpm for 2 hours at 4˚C in a P40ST rotor and himac CP70MX (Hitachi).Fractions were collected with Piston Gradient Fractionator (Biocomp).Optical density was measured at a wavelength of 254 nm using a SpectraMax (Molecular Device).Eight serial fractions obtained by the fractionator were pooled for RNA extraction using the kit of Maxwell RSC miRNA from plasma and serum (Promega).cDNA was synthesized from these RNA samples by SuperScript IV (LifeTechnologies) and oligo dT20 (LifeTechnologies).Top1 mRNA and b2microglobulin (b2M) mRNA were measured by qPCR with the primers, as described in Dataset S5.
The 80S ribosome peaks are unclear due to the lower speed of the centrifuge (81,000 g for 2 hours in P40ST rotor, Hitachi) compared to the general polysome fractionation (100,000 g for 4 hr or at 230,000 g for 2 h in SW40Ti rotor, Beckman-Coulter) (18,19).The lower centrifuge speed than the typical centrifuge speed was necessary to analyze the Top1 mRNA (3,856 nt in wild-type and 3.7~4.0k nt in knockout alleles), because the peak of Top1 mRNA distribution had shifted to the heavy side and gone out from the analysis range if the general ultracentrifuge condition is applied to.
In order to verify the separation of polysomes, centrifugation was performed on the CH12 cells lysate on the same sucrose gradient at the higher speed (143,435 g) for 2 h.As the result, the typical OD profile was yielded (Fig. S18A).Accordingly, the peak of the b2M mRNA (860 nt in mouse and shorter than Top1 mRNA) was positioned in right half (Fig. S18B), whereas the same b2M mRNA peak was positioned in left half in the lower speed centrifuge (Figs.4F, S9D and
Briefly, 360 ul protein solution was mixed with 830 and 208 µl of ice-cold methanol and chloroform, respectively, and vortexed.554 µl of distilled water was added and centrifuged at 14,000 g for 5 minutes at 4 ˚C.After discarding aqueous phase, 623 µl methanol was added, vortexed and centrifuged at 14,000 g for 5 minutes at 4˚C.Supernatant was removed and the pellet was air-dried.
The protein pellet was resuspended with 2x SDS sample buffer with 2M urea and used for western blot.The antibodies used are listed in Dataset S5.Particularly, anti-AID #488-5 antibody was used to detect endogenous AID and AID-ER.This antibody was developed by Dr. Reiko Shinkura (Institute for Quantitative Biosciences, The University of Tokyo), Dr. Kyoichi Matsumoto (Mikuri Immunology Laboratory) and Dr. Tasuku Honjo.

Overexpression and inhibition of the candidate miRNAs
For the candidate miRNAs, miRIDIAN microRNA Mimics (Horizon Discovery) for the candidate miRNAs were purchased and electroporated to the CH12 cells.As the negative control, miRIDIAN microRNA Mimic Negative Control #2 (CN-002000-01-05) was used.IgA% was evaluated under 4-OHT stimulation.
The miR-92a inhibitor of LNA oligo was originally designed and purchased from AJINOMOTO BIOPHARMA.The sequence of LNA oligo is listed in Dataset S5.CH12 cells with AID-ER was used to evaluate its effect on IgA% under 4-OHT stimulation.

Knockdown of miRNA by Synthetic Tough Decoy (S-TuD) oligo
S-TuD oligos were purchased from Ajinomoto Bio-Pharma Services.The negative control oligo uses the sequence NC5 Negative Control (Mouse) for IDT miRNA inhibitors.25 pmol of the oligos was electroporated to 8 ´ 10 5 cells using the amaxa 96-well Shuttle system (LONZA) with Amaxa SF Cell Line 96-well Nucleofector Kit (LONZA).

Quantification of miR-92a-3p
RNA was recovered using the Maxwell RSC miRNA Tissue Kit (Promega) according to the manufacturer's protocol.Reverse transcription and qPCR were performed using the Taqman MicroRNA Reverse Transcription Kit, TaqMan MicroRNA Assays, and QuantStudio5 Real-Time PCR System (ThermoFisher Scientific).

Fig. S16 A possible RNA editing site in miR-92a-3p for the stronger interaction between them
If C-to-U editing by AID occurs in g19 cytosine in miR-92a-3p, base-pairing between this edited miRNA and Top1 mRNA will be further strengthened.Dataset S3 (separate file)."Dataset_S3_PARCLIP_noIP", the excel table summarizing the sequencing results presented in Fig. 1F-G.

C to U?
Dataset S4 (separate file)."Dataset_S4_hek293t_mirna_solid20130212-re", data supporting Fig. S13A-C.the excel table based on fastq files obtained by SOLiD 5500xl, processed by blast

Fig.±
Fig. S1 Difference of DNA 3' ends between DNA base excision repair (BER)-induced single-strand breaks (SSBs) and DNA topoisomerase1-induced SSBs A. In BER-induced SSBs, damaged or modified bases are removed by DNA glycosylases.By the function of apurinic/apyrimidinic endo/exonuclease1 (APE1), abasic ribose results in 5'deoxyribose-phosphate (5'dRP).DNA 3'end takes the form of a hydroxyl group.B. DNA topoisomerase 1 (Top1)-induced SSBs.A catalytic tyrosine residue in Top1 is covalently bound to 3' phosphate of DNA in Top1-cleavage complex (Top1-cc) which forms DNA cleavage.The proteasome degrades Top1 and the binding tyrosine is removed by TDP1.As a result, 3' phosphate remains.This 3'phosphate will be further processed by PNKP, aprataxin and so on for complete repair, or hypothetically, for class switch recombination.Deficiencies of these processing enzymes result in genomic instability, cell and tissue damages and the neurodegenerative or developmental disorders.Quoted from "DNA single-strand break repair and human genetic disease", Caldecott, 2022, Trends in Cell Biology and "Abasic sites in DNA: repair and biological consequences in Saccharomyces cerevisiae",Boiteux and Guillet, 2004, DNA Repair (22, 23).

Fig. S4 Generation of Top1 3 'Fig. S5 Analysis of Top1 3 'Fig. S6 Germline transcripts in the cloned Top1 3 'Fig.
Fig. S4 Generation of Top1 3'UTR knockout (3'UTRKO) cells from CH12 cells A. The targeting scheme of 3'UTR of Top1 mRNA in CH12 cells using CRISPR/Cas9.Because there are three alleles corresponding to the Top1 gene in CH12 wild-type cells, three different targeting vectors harboring the zeocin resistant gene, blasticidin resistant gene, and ZsGreen gene were used serially in the knockout procedure.Details of the establishment of 3'UTRKO cells are described in Supplementary Information.B. The family tree of generated Top1 3'UTR KO cells.

Fig. S11
Fig. S11 Localization of Ago2, AID-ER and AID in polysome fractions shown in Fig.S9 (A) and S10 (B) Figure Fig. S17 Possible localization of Top1 mRNA in P-body separate file)."Dataset_S1_antiAgo2_top1-coding", data supporting Fig. 1D.The excel table based on fastq files obtained by Ion PGM, processed by SAMtools, showing U-to-C conversion frequencies in the coding region of Top1 mRNA, caused by 4-SU uptake and binding to Ago2.Dataset S2 (separate file)."Dataset_S2_antiAgo2_Top1-3UTR_Exp1-2", data supporting Fig. 1E.The excel table similarly obtained to Dataset S1, based on the fastq files obtained by Ion PGM, processed by SAMtools, showing U-to-C conversion frequencies of the 3'UTR of Top1 mRNA to Ago2.

Table S1 .
Details of SHM of the mass stable wild-type and Top1 3'UTR KO cells transfected with AID-ER presented in Fig.S5C.

Table S2 .
Details of SHM of the cloned wild-type and Top1 3'UTRKO cells with AID-ER, presented in Fig.