Fig. 3. Generation of occludin TM4^- isoform by skipping of exon 4. (A) Aligment of human genomic occludin sequence with TM4⁺ and TM4^- cDNAs obtained by RT-PCR. The region of interest from the genomic sequence of human occludin (AC010355.4) is shown in the top line. Central parts of introns are omitted as indicated by double slashes. Above the genomic sequence, the exon/intron structure is indicated in accordance with the annotation of pertinent genomic contig (accession no. NT_006497.6). Exons are labelled and their extent is marked by square brackets over their first and last nucleotide. Introns are labelled and their extent is marked by indicating conserved dinucleotides of splice donor (GT) and acceptor (AG) sites, respectively. Underneath the genomic sequence, the corresponding sequence of the longer (TM4⁺) RT-PCR product is shown. TM4⁺ product is 100% identical to published occludin cDNA sequence (NM_002538.2), and the numbering shown refers to that sequence. The third line shows the corresponding sequence of smaller RT-PCR product TM4^-, which is 100% identical to TM4⁺ save for a 162 nucleotide deletion in TM4^- compared with TM4⁺. Note that this deletion coincides precisely with the extent of exon 4. Below the alignment, translation of coding sequences is shown based on published cDNA sequence (NM_002538.2). Note that removal of exon 4 is not predicted to affect the reading frame of downstream coding region. (B) Schematic of proposed mechanism. 3' part of exon 3, exon 4, and 5' part of exon 5 of human occludin gene are represented by numbered boxes. Top: proposed alternative splicing event gives rise to TM4^- mRNA isoform by skipping exon 4. Bottom: canonical splicing leads to inclusion of exon 4 in TM4⁺ mRNA isoform.