Ex2 usage results in 70?0 small deletions (2 to 4 nucleotides); both of which are independent of the locus or cell type. Finally, targeting exonuclease activity to the DSB via fusion to the specific endonuclease itself improves efficiency. While it is conceivable that addition of DNA-end processing enzymes could be deleterious for genome integrity by increasing off-target mutagenic rates, we did not observe any toxicity either in survival assays or growth rate analyses for cell populations transfected with meganucleases and/or DNA-end processing enzymes. However, as noted above, target-site accessibility is a crucial parameter in the DSB repair process that could as well affect the potential off-target mutagenesis frequency. Moreover, beyond modulating the nature 
of the mutagenic events, Trex2 over-expression may be beneficial for nuclease-based genome modification. Recently, Bennardo et al. published data suggesting that Trex2 over-expression decreases the frequency of nonhomologous end-joining between distal ends of two DSBs by restricting the persistence of target sites available for repeated cleavage-ligation cycles. In principle this limits the persistence of off-target DNA breaks and thus reduces potential deleterious translocations. Altogether, our study shows that coupling DNA end-processing enzymes with meganucleases enhances targeted mutagenesis and represents an attractive new tool for precise genome engineering purposes.RAG1m (a) DMD21m (b) or CAPNS1m (c) with or without (empty) Trex or scTrex. (EPS)Figure S3 MNs activity and toxicity in the presence of Trex or scTrex. (A) In vivo cleavage activity of MNs alone or fused to scTrex. Activity was monitored using an extrachromasomal SSA assay. CHO-K1 cells were transfected with increasing amounts (0 to 25 ng) of MN or chimeric scTrex-MN and tested against their respective targets. Empty vector and I-SceI are shown as controls. (B) Toxicity of the chimeric scTrex-MN proteins was evaluated in a cell survival assay. CHO-K1 cells were cotransfected with increasing amounts of MN or scTrex-MN in the presence of 10 ng of plasmid JSI-124 cost expressing GFP. Cell survival is expressed as a percentage of cells expressing GFP 6 days 
post transfection 23. (EPS) Figure 1081537 SDeletion pattern induced by MNs and scTrexMNs. Deletion events induced at the RAG1 (A) or CAPNS1 (B) locus by MN alone (blue) or scTrex-MN (green) are shown. Deletion sizes are represented as percentage of total deletion events. Note that 45 of TM induced by RAG1m corresponds to a 9 bp deletion whereas CAPNS1m induces large and diverse deletions that were analyzed by deletion classes (range of deletions size). (EPS)Figure S5 Targeted mutagenesis pattern in DetroitSupporting InformationData S1 Meganucleases sequences. Monomers A and B are linked by the linker GGSDKYNQALSKYNQALSKYNQALSGGGGS. 19S mutation is added in monomer B. (DOC) Data S2 Amino acids sequences of the different hybridcells induced by CAPNS1m in the absence (empty) or presence of scTrex, single-chain Trex2; scTrex-CAPNS1, MN fused to scTrex; Tdt, terminal deoxynucleotidyltransferase. (A) Frequency of small deletion events (2 to 4 nucleotides) among all TM events. (B) Profile of insertion events in the presence (green) or absence (blue) of Tdt. (EPS)Table S1 Examples of sequences with insertion at transgenic locus in presence of Tdt. Sequences marked with BI-78D3 site asterisk (*) correspond to insertion events independent of the TDT activity. They are also found in the sample corresp.Ex2 usage results in 70?0 small deletions (2 to 4 nucleotides); both of which are independent of the locus or cell type. Finally, targeting exonuclease activity to the DSB via fusion to the specific endonuclease itself improves efficiency. While it is conceivable that addition of DNA-end processing enzymes could be deleterious for genome integrity by increasing off-target mutagenic rates, we did not observe any toxicity either in survival assays or growth rate analyses for cell populations transfected with meganucleases and/or DNA-end processing enzymes. However, as noted above, target-site accessibility is a crucial parameter in the DSB repair process that could as well affect the potential off-target mutagenesis frequency. Moreover, beyond modulating the nature of the mutagenic events, Trex2 over-expression may be beneficial for nuclease-based genome modification. Recently, Bennardo et al. published data suggesting that Trex2 over-expression decreases the frequency of nonhomologous end-joining between distal ends of two DSBs by restricting the persistence of target sites available for repeated cleavage-ligation cycles. In principle this limits the persistence of off-target DNA breaks and thus reduces potential deleterious translocations. Altogether, our study shows that coupling DNA end-processing enzymes with meganucleases enhances targeted mutagenesis and represents an attractive new tool for precise genome engineering purposes.RAG1m (a) DMD21m (b) or CAPNS1m (c) with or without (empty) Trex or scTrex. (EPS)Figure S3 MNs activity and toxicity in the presence of Trex or scTrex. (A) In vivo cleavage activity of MNs alone or fused to scTrex. Activity was monitored using an extrachromasomal SSA assay. CHO-K1 cells were transfected with increasing amounts (0 to 25 ng) of MN or chimeric scTrex-MN and tested against their respective targets. Empty vector and I-SceI are shown as controls. (B) Toxicity of the chimeric scTrex-MN proteins was evaluated in a cell survival assay. CHO-K1 cells were cotransfected with increasing amounts of MN or scTrex-MN in the presence of 10 ng of plasmid expressing GFP. Cell survival is expressed as a percentage of cells expressing GFP 6 days post transfection 23. (EPS) Figure 1081537 SDeletion pattern induced by MNs and scTrexMNs. Deletion events induced at the RAG1 (A) or CAPNS1 (B) locus by MN alone (blue) or scTrex-MN (green) are shown. Deletion sizes are represented as percentage of total deletion events. Note that 45 of TM induced by RAG1m corresponds to a 9 bp deletion whereas CAPNS1m induces large and diverse deletions that were analyzed by deletion classes (range of deletions size). (EPS)Figure S5 Targeted mutagenesis pattern in DetroitSupporting InformationData S1 Meganucleases sequences. Monomers A and B are linked by the linker GGSDKYNQALSKYNQALSKYNQALSGGGGS. 19S mutation is added in monomer B. (DOC) Data S2 Amino acids sequences of the different hybridcells induced by CAPNS1m in the absence (empty) or presence of scTrex, single-chain Trex2; scTrex-CAPNS1, MN fused to scTrex; Tdt, terminal deoxynucleotidyltransferase. (A) Frequency of small deletion events (2 to 4 nucleotides) among all TM events. (B) Profile of insertion events in the presence (green) or absence (blue) of Tdt. (EPS)Table S1 Examples of sequences with insertion at transgenic locus in presence of Tdt. Sequences marked with asterisk (*) correspond to insertion events independent of the TDT activity. They are also found in the sample corresp.