These techniques are being used increasingly to MLN4924 knockout genes in model organisms and cultured cells, but to date there have been few reports comparing the knockout analysis of target genes using these methods with knockdown analysis in cultured cells. Because TALENs are easy to construct compared with ZFNs, and the CRISPR/Cas system may have a problem with specificity, we chose TALENs to knockout claudin-2 in MDCK II cells to investigate its functions in detail. Transcription activator-like effectors are natural bacterial proteins secreted by Xanthomonas sp., which contain tandem repeats of DNA-binding domains that recognize specific nucleotides. TALENs are artificial nucleases generated by fusing a FokI DNA cleavage domain to TALEs. Two TALENs that recognize the left and right arms of the target site form a functional FokI dimer and induce DNA double-strand breaks at the target site. Normally, DNA double-strand breaks are repaired by non-homologous end joining pathways, resulting in the introduction of nucleotide mismatches, insertions, or deletions, and functional gene knockout. In this study, we established claudin-2 knockout clones in MDCK II cells using TALENs in a similar manner as described previously. Surprisingly, claudin-2 knockout increased TER by more than 50-fold in MDCK II cells. Claudin-2 expression restored the TER of claudin-2 knockout cells dependent upon claudin-2 protein levels. Our results indicate that claudin-2 independently determines the ��leaky�� property of TJs in MDCK II cells and suggest the importance of knockout analysis in cultured cells. To compare quantitatively the localization of claudins at TJs between control and their respective knockout cells, we Ponatinib measured the signal intensity of claudins at TJs as described above, and calculated the ratio of signal intensity between control cells to that between control and knockout cells. To compare the relative intensity, we first quantified the relative intensity of ZO-1, -2, and -3 using TALENs constructed in a previous study. The signal intensity of ZO proteins at TJs between control and knockout cells was approximately 50% of that between control cells, and the average of the relative intensity of ZO proteins was 0.45 �� 0.03. The relative intensity of claudin-1, -3, and -4 was similar to that of ZO proteins. In contrast, the relative intensity of claudin-2 and -7 was significantly lower compared with ZO proteins. These results suggest that the efficiency of claudin localization at TJs between control and knockout cells varies among claudins, and claudin-2 and -7 are less efficiently localized at TJs between control and their respective knockout cells. Claudin-2 has been reported to form high conductive cation pores in TJs and be involved in paracellular water permeability. In this study, we succeeded in establishing claudin- 2 knockout clones in MDCK II cells using the TALEN technique. To improve the efficiency of knockout clone selection, we transiently administered G418 and puromycin as described previously, and established five independent claudin-2 knockout clones. Immunoblot analysis with an anti-claudin-2 antibody demonstrated the appearance of faint bands with a molecular mass lower than that of wild-type claudin-2 in claudin-2 knockout clones. Because the open reading frame of canine claudin-2 gene has an in-frame ATG sequence at the 24th codon, these faint bands might reflect an artificial peptide formed by translation using the 24th ATG codon of the claudin-2 gene as its initiating codon, which lacks the intracellular N terminus and most of the first transmembrane helix of claudin-2.