Ligase I and ligase III mediate the DNA double-strand break ligation in alternative end-joining

Significance

DNA repair is not only extremely important for the genome stability in normal cells but also extensively involved in carcinogenesis and evolution. DNA double-strand breaks (DSBs), one of the most harmful types of DNA damage, are repaired mainly by homologous repair (HR) and nonhomologous end-joining (NHEJ). Although the components of NHEJ have been studied extensively in last two decades, in cells deficient for core classic NHEJ factors, significant DSB end-joining activities have been observed in various situations, suggesting the existence of the alternative end-joining (A-EJ) activities. Here, we dissected the roles of DNA ligases in mediating the last step of A-EJ. The results suggested the existence of multiple DNA ligase-containing complexes in A-EJ.

Abstract

In eukaryotes, DNA double-strand breaks (DSBs), one of the most harmful types of DNA damage, are repaired by homologous repair (HR) and nonhomologous end-joining (NHEJ). Surprisingly, in cells deficient for core classic NHEJ factors such as DNA ligase IV (Lig4), substantial end-joining activities have been observed in various situations, suggesting the existence of alternative end-joining (A-EJ) activities. Several putative A-EJ factors have been proposed, although results are mostly controversial. By using a clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (Cas9) system, we generated mouse CH12F3 cell lines in which, in addition to Lig4, either Lig1 or nuclear Lig3, representing the cells containing a single DNA ligase (Lig3 or Lig1, respectively) in their nucleus, was completely ablated. Surprisingly, we found that both Lig1- and Lig3-containing complexes could efficiently catalyze A-EJ for class switching recombination (CSR) in the IgH locus and chromosomal deletions between DSBs generated by CRISPR/Cas9 in cis-chromosomes. However, only deletion of nuclear Lig3, but not Lig1, could significantly reduce the interchromosomal translocations in Lig4^−/− cells, suggesting the unique role of Lig3 in catalyzing chromosome translocation. Additional sequence analysis of chromosome translocation junction microhomology revealed the specificity of different ligase-containing complexes. The data suggested the existence of multiple DNA ligase-containing complexes in A-EJ.

Mammalian genomes are subjected substantial DNA damage from both endogenous processes [e.g., DNA replication, class switching recombination (CSR), etc.] and exogenous resources (e.g., ionic radiation, DNA-damaging chemicals, etc.). Evolutionarily conserved DNA repair pathways are essential to maintain both the structure integrity and the information accuracy of the genome (1). In eukaryotes, DNA double-strand breaks (DSBs), one of the most dangerous and severe types of DNA damage, are repaired mainly by two evolutionarily conserved repair pathways: homologous repair (HR) and nonhomologous end-joining (NHEJ) (2). NHEJ directly ligates two broken DSB ends, whereas HR uses a homologous template (in most of cases, the sister chromosome) for repair. Although NHEJ is simpler and faster than HR, repair by NHEJ often leads to change of sequences in the repair junctions via deletion, insertion, and mutations. Most importantly, NHEJ can also directly ligate two distant DSBs (both in cis- and trans-chromosomes), therefore leading to chromosomal deletions and translocations that are tightly linked to the genome evolution and carcinogenesis (3).

The NHEJ in eukaryotes has been extensively studied in the last two decades (4–6). Mechanistically, NHEJ could be separated into several steps. First, in the DSB binding and tethering step, the DSB ends are recognized and bound by Ku70/Ku80 heterodimers that function as docking sites for other NHEJ factors. In the next end-processing step, nuclease and DNA polymerase activities are recruited to remove damaged or mismatched nucleotides and prepare the broken ends for ligation. Finally, DNA ligase IV (Lig4) complex, consisting of DNA Lig4 and its cofactor X-ray repair cross-complementing protein 4 (XRCC4), as well as the newly identified XRCC4-interacting factor (XLF), reseals the DSBs. Both V(D)J recombination and class switching recombination (CSR) have been used as important in vivo models to study the NHEJ (7). During the V(D)J recombination, DSBs generated by recombination-activating gene 1 (RAG1)/RAG2 endonuclease are joined exclusively by NHEJ. Deficiency of core NHEJ factors in mice, such as Lig4 and XRCC4, leads to complete abolishment of V(D)J recombination and lack of mature B and T cells (8, 9). When combined with a checkpoint-deficient genetic background (e.g., p53^−/−), unresolved RAG-generated DSB ends in B and T cells in core NHEJ factor-deficient mice could result in dramatic genomic instability. In particular, oncogenic chromosome translocations lead to the development of lymphoma and leukemia in those mice (10, 11). The fact that those chromosome translocations are formed by end-joining in the absence of core NHEJ factors suggested the existence of alternative end-joining pathway(s) (A-EJ). CSR, by which mammalian mature B cells change their production of Ig constant region type from one to the other, represents an excellent model to study classic NHEJ (c-NHEJ) and A-EJ (12, 13). DSBs in the IgH class switching (S) regions induced by activation-induced cytidine deaminase (AID) are repaired by NHEJ via a loop-out and deletion mechanism (14). In the absence of c-NHEJ core factors (such as Lig4, XRCC4, and Ku70/80), significant CSR activities, mediated by A-EJ, have been observed in both animals and cell lines (12, 13). There is no doubt that A-EJ contributes to all end-joining activities in the absence of c-NHEJ. However, the contribution of A-EJ in the presence of c-NHEJ is still debatable. For example, it has been suggested that A-EJ is the main end-joining activity to catalyze chromosomal translocations in murine (15) but not in human cells (16).

Although A-EJ activities have been observed in many cell types and biological processes (12, 17–19), A-EJ’s exact components and mechanisms have been still not clearly revealed and sometimes are controversial (5, 20, 21). For example, whether A-EJ is a completely independent new pathway or an alternative c-NHEJ pathway in which alternative components could substitute the missing c-NHEJ factors is still debatable. Comparing with large numbers of factors and pathways involved in the early DSB repair steps, there are only three known DNA ligases (DNA Lig1, DNA Lig3, and DNA Lig4) in mammalian cells to finish the last ligation step (22). It has been proposed that those three DNA ligases function differently in various DNA metabolism processes. Although all three mammalian DNA ligases have highly homologous catalytic cores (including DBD, AdD, and OB-Fold domains), through their distinct N- and C-terminal regions, the DNA ligases may interact with different partners, which could confer functional specificity. In DSB repair, the role of Lig4 has been mostly restricted to c-NHEJ, whereas both Lig1 and Lig3 have been suggested to mediate the A-EJ in vitro and in vivo (23–28). Here, we used clustered regularly interspaced short palindromic repeats (CRISPR)/CRISPR-associated protein 9 (Cas9) to generate cell lines in which Lig1 or Lig3 were completely depleted, and we tried to unequivocally reveal the ligases’ roles in A-EJ.

Results

Establishment of Mammalian Cell Lines Containing only a Single DNA Ligase in Nucleus.

To unequivocally study the repair of endogenous chromosomal DSBs by the A-EJ pathway(s) in vivo, we used CH12F3 (CH12) cells, which are proficient or deficient of the core c-NHEJ factor Lig4 (13, 29). To completely inactivate Lig1 in those cells, CRISPR/Cas9 and two single-guide RNAs (sgRNAs) were designed to delete exons 18 and 19 of the mouse Lig1 gene (Fig. 1A). PCR strategy was used to identify the cell lines in which both alleles of Lig1 were deleted (Fig. 1C). Western blot results showed that there was no detectable Lig1 protein in Lig1^−/− cells (Fig. 1E). For Lig3 inactivation, exons 8–13 of the Lig3 gene, which encode catalytic core of Lig3 protein, were deleted by Cas9 and two sgRNAs (Fig. 1B). Because the mitochondria isoform of Lig3 protein is essential for the maintenance of normal mitochondrial function and cell viability, to deplete the nuclear form of Lig3, we had to first stably express a MtLig3-ΔBRCT-GFP-NES protein that locates exclusively in mitochondria and cytoplasm but not in nucleus (30) (Fig. S1 and Fig. 1D). Western blot and PCR genotyping were used to identify the cell lines specifically deficient for nuclear Lig3 (Fig. 1 D and F). In the rest of this report, we simplified such nuclear Lig3-deficient cells as Lig3^−/− cells.

Fig. 1.

Generation of mouse CH12 cells deficient for Lig1 and nuclear Lig3. (A and B) Schematic representation for the genomic organization of WT and CRISPR/Cas9/sgRNA-deleted Lig1 (A) and Lig3 (B) loci. Target sites of CRISPR/Cas9/sgRNA as well as primers used for genotyping are indicated. (C) PCR genotyping of Lig1^+/+ (lane 1), Lig1^+/− (lane 2), and Lig1^−/− (lane 3) cells. (D) PCR genotyping of Lig3^+/+ (lane 1), Lig3^+/− (lane 2), and Lig3^−/− (lane 3) cells. (E) Western blot analysis of Lig1 protein. In Lig1^−/− cells (lane 1), the Lig1 protein is completely lost. Lane 2 is the WT control cell line. (F) Western blot analysis of Lig3 protein. In nuclear Lig3^−/− cells (lane 1), the WT Lig3 protein (∼105 kDa) is completely lost. In lane 1, the rescued Mtlig3-ΔBRCT-GFP-NES protein (∼130 kDa) is marked by an asterisk. Lane 2 is the WT control cell line.

Fig. S1.

Subcellular localization of GFP-tagged Lig3 constructs in CH12 cells. Mitochondria and the nucleus were labeled with MitoTracker Red CMXRos (Invitrogen) and Hoechst 33342 (Invitrogen), respectively. (Upper) Wild-type Lig3-GFP is in both the nucleus and mitochondria. (Lower) MtLig3-ΔBRCT-GFP-NES is no longer in the nucleus.

From the Lig4-deficient cell line (13), we generated CH12 cells deficient for either both Lig1 and Lig4 (Lig1^−/−Lig4^−/−) or both Lig3 and Lig4 (Lig3^−/−Lig4^−/−) (Fig. S2). In other words, in Lig1^−/−Lig4^−/− cells, the only remaining nuclear DNA ligase is Lig3, whereas in Lig3^−/−Lig4^−/− cells, the only nuclear DNA ligase is Lig1. We also tried to obtain cell lines in which both Lig1 and nuclear Lig3 are deleted (i.e., a Lig4-only cell line). However, such cell lines could not be obtained (Table S1).

Fig. S2.

Western blot results for the cell lines used in current study. Analysis Lig1 (A) and Lig3 (B) protein expression level in WT (lane 1), Lig4^−/− (lane 2), Lig1,4^−/− clone 1 (lane 3), Lig1,4^−/− clone 2 (lane 4), Lig1,4^−/− clone 3 (lane 5), Lig3,4^−/− clone 1 (lane 6), Lig3,4^−/− clone 2 (lane 7), and Lig3,4^−/− clone 3 (lane 8).

Table S1.

Lig1 and nuclear Lig3 could not be deleted at the same time in CH12 cells

CH12	Rescue	Cas9+sgRNA	Total clones	Knockout clones
WT	MtLig3-ΔBRCT-GFP-NES	Lig1 sg1/2	558	7
Lig3−/−		Lig1 sg1/2	471	0
WT		Lig3 sg3/4	300	6
Lig1−/−		Lig3 sg3/4	257	0

Comparing with Lig4^−/− cells, additional deletion of either Lig1 (Lig1^−/−Lig4^−/−) or nuclear Lig3 (Lig3^−/−Lig4^−/−) did not significantly affect the cell proliferation without or with cytokine stimulation (Fig. S3 A and B, respectively) and did not affect the cells’ sensitivity to hydroxyurea (HU) (Fig. S3C), suggesting that Lig1 and Lig3 can completely replace each other’s function in DNA replication. However, both double-deficient cells showed slightly but significantly more sensitivity to methyl methanesulfonate (MMS) and PARP inhibitor (Olaparib) (Fig. S3 D and E, respectively). In addition, comparing with Lig4^−/− and Lig1^−/−Lig4^−/−, Lig3^−/−Lig4^−/− cells also showed slightly more sensitivity to Zeocin (Fig. S3F). These similarities and differences in response to various DNA-damaging reagents suggest that Lig1 and Lig3 play both complementary and noncomplementary functions in DNA repair.

Fig. S3.

Proliferation and drug sensitivity of Lig1- or nuclear Lig3-deficient CH12 cells. (A and B) Live cell counts of WT, Lig4^−/−, Lig1,4^−/−, and Lig3,4^−/− CH12 cells in cultures without (Unstimulated) (A) or with (Stimulated) (B) cytokines for inducing CSR. (C–F) Sensitivity of each cell line to HU (C), MMS (D), PARP inhibitor (Olaparib) (E), and Zeocin (F). The experiments were performed independently for three times. Error bars represent the SDs of three replicates.

Deletion of Either Lig1 or Nuclear Lig3 Could Not Affect A-EJ–Mediated IgH CSR.

We examined the cytokine-induced IgH CSR in various CH12 cell lines (Fig. 2 and Fig. S4). Surprisingly, deletion of either Lig1 or nuclear Lig3 in Lig4-deficient cells could not further reduce the percentage of IgA-positive (IgA⁺) cells after CSR; these IgA⁺ cells represent the successful IgH CSR products catalyzed by A-EJ (Fig. 2A, Upper). At the same time, the percentages of IgA and IgM double-negative (IgA⁻IgM⁻) cells, representing cells in which CSR failed after DSBs were introduced in switching regions, are similar in Lig4^−/−, Lig1^−/−Lig4^−/−, and Lig3^−/−Lig4^−/− cells (Fig. 2A, Lower).

Fig. 2.

Deletion of either Lig1 or nuclear Lig3 could not affect A-EJ–mediated CSR in Lig4-deficient CH12 cells. (A) Summary of FACS analysis results from cytokine-induced CSR in various CH12 cells after 48 h (Left) and 72 h (Right). Results of surface staining of IgA positive (IgA⁺) cells (Upper) and IgA/IgM double-negative (IgA⁻IgM⁻) cells (Lower) are shown. (B) Summary of FACS analysis results from Cas9/sgRNA induced CSR in various CH12 cells after 48 h (Lower Left) and 72 h (Lower Right). (B, Upper) Schematic representation of the assay and the localization of sgRNAs used. Mean percentages of IgA⁺ or IgA⁻IgM⁻ for each cell line from three independent experiments were calculated. For Lig1,4^−/− and Lig3,4^−/− cell lines, the results from three independent single clones were combined. Background from normal culture without cytokine was subtracted in all experiments. Error bars indicate the SD.

Fig. S4.

Representative FACS analysis for cytokine-induced CSR. Representative flow cytometry analysis by surface staining of IgA and IgM of WT, Lig4^−/−, Lig1,4^−/− clone 1, Lig1,4^−/− clone 2, Lig1,4^−/− clone 3, Lig3,4^−/− clone 1, Lig3,4^−/− clone 2, and Lig3,4^−/− clone 3 cells at 0, 48, and 72 h after cytokine stimulation.

To further refine the direct roles of Lig1 and Lig3 in the DSB repair step during CSR, we developed an AID-independent CSR assay, in which Cas9/sgRNAs targeting Sμ and Sα regions were cotransfected into normal cultured CH12 cells without cytokine stimulation (Fig. 2B and Fig. S5). Such Cas9/sgRNA-induced DSBs in class switching regions could be efficiently joined by c-NHEJ and A-EJ activities in both wild-type (WT) and Lig4-deficient CH12 cells (Fig. S5). However, the end-joining efficiencies between such AID-independent DSBs within the IgH locus in Lig4^−/−, Lig1^−/−Lig4^−/−, and Lig3^−/−Lig4^−/− cells are not significantly different, suggesting such A-EJ activities are not solely dependent on either Lig1 or Lig3.

Fig. S5.

Representative FACS analysis for Cas9/sgRNA-induced CSR. Representative flow cytometry analysis by surface staining of IgA and IgM of WT, Lig4^−/−, Lig1,4^−/− clone 1, Lig1,4^−/− clone 2, Lig1,4^−/− clone 3, Lig3,4^−/− clone 1, Lig3,4^−/− clone 2, and Lig3,4^−/− clone 3 cells at 0, 48, and 72 h after transfection of Cas9/sgRNAs.

Intrachromosomal DSB Deletional A-EJ in Lig1- or Nuclear Lig3-Deficient CH12 Cells.

CSR in the IgH locus may represent a special form of A-EJ that is complicated by IgH locus-specific synapse formation, AID-induced DSBs, and end-processing (14, 31). Therefore, we further investigated the A-EJ activities between two non-IgH chromosomal DSBs. We designed four sgRNAs targeting different genomic loci in chromosome 8 (Fig. 3A). By cotransfecting distinct Cas9/sgRNA pairs, the deletional end-joining between two sgRNA-induced DSBs separated by various distances (4, 40, and 370 kb) could be analyzed by quantitative real-time PCR. The results showed that the efficiencies for those deletional A-EJ activities in Lig4^−/−, Lig1^−/−Lig4^−/−, and Lig3^−/−Lig4^−/− cells are not significantly different (Fig. 3B). The above results further confirmed that Lig1 and Lig3 could completely replace each other’s role in deletional A-EJ.

Fig. 3.

Intrachromosomal DSB repair by A-EJ in Lig1- or nuclear Lig3-deficient CH12 cells. (A) Schematic representation of the intrachromosomal deletion assay. The CRISPR/Cas9 and sgRNAs were designed to introduce individual DSB spanning a 370-kb region on chromosome 8. (B) Summary of the quantitative real-time PCR analysis of intrachromosomal deletion efficiencies between two sgRNA-induced DSBs separating by various distances (4, 40, and 370 kb). Relative deletion frequencies in different cell lines were normalized to WT cells and also adjusted by transfection efficiency. Error bars indicate the SD of four independent experiments. For Lig1,4^−/− and Lig3,4^−/− cell lines, the results from three independent single clones were combined.

Interchromosomal Translocations by A-EJ in the Absence of Either Lig1 or Nuclear Lig3.

We set up a nested-PCR assay for chromosomal translocations between Cas9/sgRNA-induced DSBs in nonhomologous chromosomes (Fig. S6 A and B). Translocation frequencies between the Tet2 locus in chromosome 3 and the HDAC6 locus in chromosome X (Tet2/HDAC6; Fig. 4A and Fig. S6C) and between the Tet2 locus and the ADA2 locus in chromosome 2 (Tet2/ADA2; Fig. 4B and Fig. S6D) increased significantly in Lig4-deficient CH12 cells in comparison with WT. Further deletion of Lig1 (Lig1^−/−Lig4^−/−) could not affect the frequencies of both chromosome translocations. Surprisingly, deletion of nuclear Lig3 (Lig3^−/−Lig4^−/−) resulted in a significant reduction of translocations catalyzed by A-EJ in Lig4-deficient cells. In other words, in Lig3^−/−Lig4^−/− cells, Lig1 could not fully replace the translocation ligation activities of nuclear Lig3. However, in Lig1^−/−Lig4^−/− cells, Lig3 could largely compensate the loss of Lig1 in chromosome translocations. These results suggested that for A-EJ activities mediating chromosome translocations, Lig1 and Lig3 have both complementary and noncomplementary functions.

Fig. S6.

Interchromosomal translocations induced by Cas9/sgRNA-cleaved DSB pairs. (A and B) Schematic presentation of the translocation assays. Target sequences and their chromosomal positions are indicated for all sgRNAs. Nested PCR using indicated primer sets (black arrows) flanking the cleavage sites was used to detect and quantify the translocations. (C and D) Summary of the relative chromosomal translocation efficiencies of t(Tet2/HDAC6) (C) and t(Tet2/ADA2) (D). This is a biological repeat for the experiment in Fig. 4 A and B. (E) Deletion lengths of t(Tet2/HDAC6) junctions in indicated cell lines. (F) Percentage of junctions with the indicated length of microhomology in t(Tet2/HDAC6) junctions from WT, Lig4^−/−, Lig1,4^−/−, and Lig3,4^−/− CH12 cells.

Fig. 4.

Interchromosomal translocation mediated by A-EJ in Lig1- or nuclear Lig3-deficient CH12 cells. (A and B) Summary of the relative chromosomal translocation efficiencies of t(Tet2/HDAC6) (A) and t(Tet2/ADA2) (B). Error bars indicate the SD (of four independent experiments). For Lig1,4^−/− and Lig3,4^−/− cell lines, the results from three independent single clones were combined. (C) Distribution of t(Tet2/HDAC6) translocation junctions. Percentages of direct joints (black), deletional junctions (red), and insertions (blue) were calculated for WT (n = 62), Lig4^−/− (n = 64), Lig1,4^−/− (n = 71), and Lig3,4^−/− (n = 97) cell lines. (D) Distribution of t(Tet2/HDAC6) junction microhomologies in indicated cell lines. Junctions with insertions were excluded. Mann–Whitney U test was performed.

The individual Tet2/HDAC6 translocation junctions were isolated for sequence analysis. As shown in Fig. 4C, Table S2, and Dataset S1, the direct joints, which represent the direct ligation of DSB ends generated by precise cleavage of Cas9/sgRNA complexes, decreased significantly (10/62 in WT vs. 1/64 in Lig4^−/−) when Lig4 was deleted. Therefore, DSBs without end-processing might the preferential substrates for Lig4 and c-NHEJ during chromosome translocation. On the other hand, it is also possible that c-NHEJ is kinetically faster than A-EJ and therefore will be used first before the significant DSB end-processing happens. We also compared the distribution of microhomologies in junctions isolated from various cell lines (Fig. 4D and Fig. S6F). As expected, the junctions in Lig4^−/− cells showed significantly longer microhomologies than in WT cells. In particular, the percentage of blunt junctions [without microhomology (MH = 0)] was reduced from 50% in WT to 10% in Lig4^−/− cells. Interestingly, in Lig3^−/−Lig4^−/− but not in Lig1^−/−Lig4^−/− cells, such blunt junctions are further reduced. At the same time, in Lig1-only (Lig3^−/−Lig4^−/−) cells, translocation junctions have significantly longer microhomology. The results suggest that during A-EJ the blunt junctions are likely preferential substrates of Lig3, whereas Lig1 ligates preferentially junctions with longer microhomologies.

Table S2.

Information about the t(Tet2/HDAC6) chromosomal translocation junction

CH12	Total reads	Direct joint	Deletion	Insertion
WT	62	10	45	7
Lig4−/−	64	1	50	13
Lig1,4−/−	71	1	60	10
Lig3,4−/−	97	1	65	31

Discussion

DNA ligases catalyze the formation of a phosphodiester bond to join broken DNA strands together and therefore play essential roles in nuclear DNA metabolism, including DNA replication, base excision repair (BER), nucleotide excision repair (NER), single-strand break repair, mismatch repair, homology-mediated DSB repair, and NHEJ of DSB (22). Whereas the function of Lig4 is mostly restricted to c-NHEJ, how Lig1 and Lig3 contribute to the other DNA repair processes has not been studied in details in vivo. We have generated various cell lines genetically deficient for one or two DNA ligases; these cell lines provide useful tools to clarify the in vivo functions and specificity of DNA ligases in nuclear DNA metabolism.

In particular, both Lig1 and Lig3 have been proposed to mediate A-EJ (23–28). By analyzing the A-EJ activities in cells containing only Lig1 (Lig3^−/−Lig4^−/−) or Lig3 (Lig1^−/−Lig4^−/−), we found that both Lig1- and Lig3-containing complexes could efficiently catalyze A-EJ for chromosomal deletions. However, for A-EJ–mediated interchromosomal translocations, Lig1- and Lig3-containing complexes showed different activities. More specifically, deletion of nuclear Lig3 could significantly reduce the translocation frequency in Lig4^−/− cells, whereas deletion of Lig1 could not. In addition, analysis of translocation junction microhomology revealed the specificity of different ligase complexes. During chromosome translocation, chromosomal DSB ends without end-processing are the preferential substrates for Lig4 complex(s); during A-EJ, the blunt junctions without microhomology are preferential substrates of Lig3 complex(s), whereas Lig1-complex(s) ligates preferentially junctions with longer microhomologies. We propose that those three kinds of ligase complexes may compete for the final ligation step during chromosome translocation. The components of such ligase complexes are still waiting to be systematically characterized.

Materials and Methods

Plasmids.

The pcDNA3.3-hCas9 and pX330-U6-Chimeric_BB-CBh-hSpCas9 plasmids were obtained from Addgene (no. 41815 and no. 42230). The pCR-Blunt II-TOPO-sgRNA empty cloning vector was obtained from Addgene (no. 41824), and the sgRNAs for Lig1 and Lig3 deletions and for chromosome translocation assays were constructed following the Church Laboratory sgRNA Synthesis Protocol (www.addgene.org/static/data/93/40/adf4a4fe-5e77-11e2-9c30-003048dd6500.pdf). Other sgRNA plasmids were constructed following the Zhang Laboratory gRNA Synthesis Protocol (www.genome-engineering.org/crispr/wp-content/uploads/2014/05/CRISPR-Reagent-Description-Rev20140509.pdf). Lig3-GFP and MtLig3-ΔBRCT-GFP-NES cDNAs were assembled as described by Simsek et al. (26) and cloned into the pPB-LR51-EF1α-puro2ACas9 (a gift from KosukeYusa, Wellcome Trust Sanger Institute, London). All PCR-generated fragments were verified by Sanger sequencing. See oligo information in Table S3.

Table S3.

Oligo information

Oligo name	Sequence (5′→3′)	Note
Lig1-sgRNA1-F	TTCTTGGCTTTATATATCTTGTGGAAAGGACGAAACACCCAGAGCAAGGCAGGCAGCTG	Lig1 sgRNA construction; pCR-Blunt as backbone
Lig1-sgRNA1-R	ACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTAAAACCAGCTGCCTGCCTTGCTCTG
Lig1-sgRNA2-F	TTCTTGGCTTTATATATCTTGTGGAAAGGACGAAACACCGGATGGTAGACTAGCAGCTG
Lig1-sgRNA2-R	ACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTAAAACCAGCTGCTAGTCTACCATCC
Lig3-sgRNA3-F	TTCTTGGCTTTATATATCTTGTGGAAAGGACGAAACACCAAAGACAAAAGAGAGAATTC	Lig3 sgRNA construction; pCR-Blunt as backbone
Lig3-sgRNA3-R	ACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTAAAACGAATTCTCTCTTTTGTCTTT
Lig3-sgRNA4-F	TTCTTGGCTTTATATATCTTGTGGAAAGGACGAAACACCGCTGGTCCTAGATCTGAAAA
Lig3-sgRNA4-R	ACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTAAAACTTTTCAGATCTAGGACCAGC
Lig1-sgRNA1-RFLP-F	GTGAACAACTATACCCAGACAG	Primer a
Lig1-sgRNA1-RFLP-R	GTGTCTTTGGACTACTACTCAC
Lig1-sgRNA2-RFLP-F	CTTAGAGACACAGACCACCTG
Lig1-sgRNA2-RFLP-R	AGGACGAGGTCAAAAGACAGG	Primer b
Lig3-sgRNA3-RFLP-F	ACCTTAAGTGCATTATTCGGC	Primer e
Lig3-sgRNA3-RFLP-R	TATAGAAGTGGGCACATGGAG
Lig3-sgRNA4-RFLP-F	TGCAGTCTGCTTCTCAAAGG
Lig3-sgRNA4-RFLP-R	ATCAGCTGTATCAGCCATGG	Primer h
Lig1-het-F	GATCTTCAGCAGGAACCAGG	Primer c
Lig1-het-R	CTGCTCTGCCTCTGAGCTAC	Primer d
Lig3-het-F	TACTCTGAAGACAGCAGGGG	Primer f
Lig3-het-R	TAGTCCTTGAAGTGGGCCAC	Primer g
Sμ-sg-f	CACCG TGGGGTGAGCTGAGCTGAGC	sgRNAs construction for Cas9/sgRNA-induced CSR assay; pX330 as backbone
Sμ-sg-r	AAAC GCTCAGCTCAGCTCACCCCA C
Sα-sg1-f	CACCG TGAGCTGAGCTGGAATGAGC
Sα-sg1-r	AAAC GCTCATTCCAGCTCAGCTCA C
Sα-sg2-f	CACCG TGGGCTAGGCTGAGCTGAGC
Sα-sg2-r	AAAC GCTCAGCTCAGCCTAGCCCA C
DNAJB1-SG1-F	CACCGCTGGTCGCACCGAGATCTAG	sgRNAs construction for deletion; pX330 as backbone
DNAJB1-SG1-R	AAACCTAGATCTCGGTGCGACCAGC
DNAJB1-SG2-F	CACCGGTGATGTGCATAATGCACTC
DNAJB1-SG2-R	AAACGAGTGCATTATGCACATCACC
DNAJB1-SG3-F	CACCGCCGCTGCAATAGCCGGGAAC
DNAJB1-SG3-R	AAACGTTCCCGGCTATTGCAGCGGC
DNAJB1-SG4-F	CACCGGCTGTTCCTCTGGGCCCATC
DNAJB1-SG4-R	AAACGATGGGCCCAGAGGAACAGCC
DNAJB1-sg1-qPCR-F	CGTAGGAGCTTCTCTGTAAGATG	Quantitative real-time PCR for deletion frequency
DNAJB1-sg2-qPCR-R	GGCACAGTCTTGGTAAAAGTTTC
DNAJB1-sg3-qPCR-R	AAACTTACCACATACCATGGCAC
DNAJB1-sg4-qPCR-R	GCTTTAGTCCCCAGAAAGTGCAG
ADA2-sgRNA-F	TTCTTGGCTTTATATATCTTGTGGAAAGGACGAAACACCTGATGCATATACATACAAGC	sgRNAs construction for chromosomal translocation; pX330 as backbone
ADA2-sgRNA-R	ACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTAAAACGCTTGTATGTATATGCATCA
HDAC6-sgRNA-F	TTCTTGGCTTTATATATCTTGTGGAAAGGACGAAACACCGTGGATTGGGATGTTCATCA
HDAC6-sgRNA-R	ACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTAAAACTGATGAACATCCCAATCCAC
Tet2-sgRNA-F	TTCTTGGCTTTATATATCTTGTGGAAAGGACGAAACACCGAAAGTGCCAACAGATATCC
Tet2-sgRNA-R	ACTAGCCTTATTTTAACTTGCTATTTCTAGCTCTAAAACGGATATCTGTTGGCACTTTC
Tet2-F	CAGATGCTTAGGCCAATCAAG	Chromosomal translocation
Nest-Tet2-F	CCGCCACAAGAAAATATGTCCAG
Hdac6-R	CATGCAGCCAGGTAATCAGCATCA	Tet2/HDAC6 translocation
Nest-HDAC6-R	AGGGCACATTGACAGTGAAGC
ADA2-R	TCATGTGGCATCTCATGAGTCC	Tet2/ADA translocation

F, forward; qPCR, quantitative real-time PCR; R, reverse.

Antibodies.

Antibodies to Lig1 (18051-1-AP; Proteintech Group), Lig3 (611876; BD Biosciences), and α-tubulin (T6199; Sigma-Aldrich) were used in Western blot. FITC-conjugated anti-mouse IgA (559354; BD Biosciences) and allophycocyanin (APC)-conjugated anti-mouse IgM (17-5790-82; Affymetrix/eBioscience) were used in FACS analysis.

Cell Culture and Transfection.

CH12 cells were cultured in RPMI 1640 medium supplemented with 10% (vol/vol) FBS, 1× nonessential amino acids, 1 mM sodium pyruvate, 20 mM Hepes, 2 mM l-glutamine, 50 mM β-mercaptoethanol, and 5% (vol/vol) NCTC-109; 1 μg of indicated plasmid was transfected into 1 × 10⁶ CH12 cells following the NeonTransfection System protocol with 1,100 voltages, 20 microseconds, and 2 pulses.

Generation of Lig1 and Nuclear Lig3-Deficient Cell Lines.

For Lig1 and Lig3 deletion, pcDNA3.3-hCas9, indicated sgRNA expression plasmids, and a venus expression vector were cotransfected into CH12 cells; 72 h after transfection, venus-positive cells were sorted as single clones into 96-well plates by BD FACSAria II. Individual clones were genotyped by PCR, and positive clones were verified by Western blot.

For Lig3 deletion, before Cas9/sgRNA deletion, cell lines stably expressing Lig3-GFP and MtLig3-ΔBRCT-GFP-NES proteins were established by the piggyBac transposon system, as described previously (32). See oligo information in Table S3.

Cell-Proliferation and Drug-Sensitivity Assays.

To monitor cell proliferation and drug sensitivity, indicated cells were seeded at 5 × 10⁴ cells/mL with various chemicals added at different concentrations in a 96-well plate and cultured for 24 or 48 h. Then, 20 μL of reconstituted CellTiter-Glo reagent (Promega G7570) with 80 μL of cell culture medium was added to each well. The contents were mixed on an orbital shaker for 5 min to induce cell lysis and were subsequently incubated at room temperature for 5 min to stabilize the luminescent signal, which was recorded using a Multimode Plate Readers (2300-001M; PerkinElmer).

Cytokine-Stimulated and Cas9/sgRNA-Induced CSR Assays.

For cytokine-induced CSR assay, CH12 cells were seeded at 5 × 10⁴ cells/mL in the presence of 1 ng/mL anti-CD40 (16-0402-86; eBioscience), 5 ng/mL of IL-4 (404-ML; R&D Systems), and 0.5 ng/mL TGF-β1 (240-B; R&D Systems) and cultured for 48 and 72 h.

For Cas9/sgRNA-induced CSR assay, 1 × 10⁶ CH12 cells were transfected with 0.5 μg of pX330-Sμ-sg and 0.5 μg of pX330-Sα-sg and cultured for 48 and 72 h. Cells were stained with FITC-conjugated anti-mouse IgA and APC-conjugated anti-mouse IgM antibodies and then analyzed with an Accuri C6 Flow Cytometer (BD Biosciences). CSR efficiencies were determined as the percentage of IgA-positive cells.

Cas9/sgRNA-Induced Chromosomal Deletion Assay.

A total of 1 × 10⁶ CH12 cells were electroporated with indicated sgRNA and Cas9 vectors. Chromosomal deletion frequency was determined by quantitative real-time PCR using GAPDH as an internal control using genomic DNA isolated after 72 h posttransfection. Relative deletion frequencies were also normalized to transfection efficiency. See oligo information in Table S3.

Cas9/sgRNA-Mediated Chromosomal Translocation.

For translocations, 2 × 10⁶ of CH12 cells were electroporated with indicated sgRNA and Cas9 vectors. Genomic DNAs were collected at 72 h posttransfection. Chromosomal translocations were detected and quantified by a nested-PCR approach. The results were also normalized to transfection efficiency. Individual chromosomal translocation junctions were sequenced to verify translocations and determine the junction sequences. See oligo information in Table S3.