We present rna22, a method for identifying microRNA binding sites and their corresponding heteroduplexes. Rna22 does not rely upon cross-species conservation, is resilient to noise, and, unlike previous methods, it first finds putative microRNA binding sites in the sequence of interest, then identifies the targeting microRNA. Computationally, we show that rna22 identifies most of the currently known heteroduplexes. Experimentally, with luciferase assays, we demonstrate average repressions of 30% or more for 168 of 226 tested targets. The analysis suggests that some microRNAs may have as many as a few thousand targets, and that between 74% and 92% of the gene transcripts in four model genomes are likely under microRNA control through their untranslated and amino acid coding regions. We also extended the method's key idea to a low-error microRNA-precursor-discovery scheme; our studies suggest that the number of microRNA precursors in mammalian genomes likely ranges in the tens of thousands.

Urinary exosomes or microvesicles are being studied intensively to identify potential new biomarkers for renal disease. We sought to identify whether these microvesicles contain nucleic acids. We isolated microvesicles from human urine in the same density range as that previously described for urinary exosomes and found them to have an RNA integrity profile similar to that of kidney tissue, including 18S and 28S rRNA. This profile was better preserved in urinary microvesicles compared with whole cells isolated from urine, suggesting that microvesicles may protect RNA during urine passage. We were able to detect mRNA in the human urinary microvesicles encoding proteins from all regions of the nephron and the collecting duct. Further, to provide a proof of principle, we found that microvesicles isolated from the urine of the V-ATPase B1 subunit knockout mice lacked mRNA of this subunit while containing a normal amount of the B2 subunit and aquaporin 2. The microvesicles were found to be contaminated with extraneous DNA potentially on their surface; therefore, we developed a rapid and reliable means to isolate nucleic acids from within urine microvesicles devoid of this extraneous contamination. Our study provides an experimental strategy for the routine isolation and use of urinary microvesicles as a novel and non-invasive source of nucleic acids to further renal disease biomarker discovery.

E-cadherin is a major adherens junction protein of epithelial cells, with a central role in cell-cell adhesion and cell polarity. Newly synthesized E-cadherin is targeted to the basolateral cell surface. We analyzed targeting information in the cytoplasmic tail of E-cadherin by utilizing chimeras of E-cadherin fused to the ectodomain of the interleukin-2␣ (IL-2␣) receptor expressed in Madin-Darby canine kidney and LLC-PK 1 epithelial cells. Chimeras containing the full-length or membraneproximal half of the E-cadherin cytoplasmic tail were correctly targeted to the basolateral domain. Sequence analysis of the membrane-proximal tail region revealed the presence of a highly conserved dileucine motif, which was analyzed as a putative targeting signal by mutagenesis. Elimination of this motif resulted in the loss of Tac/E-cadherin basolateral localization, pinpointing this dileucine signal as being both necessary and sufficient for basolateral targeting of E-cadherin. Truncation mutants unable to bind ␤-catenin were correctly targeted, showing, contrary to current understanding, that ␤-catenin is not required for basolateral trafficking. Our results also provide evidence that dileucine-mediated targeting is maintained in LLC-PK 1 cells despite the altered polarity of basolateral proteins with tyrosine-based signals in this cell line. These results provide the first direct insights into how E-cadherin is targeted to the basolateral membrane.