Analysis of pri–miR-142 processing in HEK293 cells. (a) Schematic of two pri–miR-142 RNA expression plasmids, one for wild-type, unedited (pri-miR142WT) and the other for pre-edited precursor RNA with A→G substitutions at the +4, +5, +40 and +50 positions (pri-miR142ED). (b) TLC analysis of unedited and pre-edited pri–miR-142 RNAs subjected to in vitro editing by a mixture of rADAR1p110 and rADAR2 proteins (20 ng each). (c) Primers used for differential primer-extension assay are shown, together with primer-extended DNA products of different sizes corresponding to pri–, pre– and mature miR-142 RNAs. A set of three primers complementary to the miR-142-5p sequence, 5pWT (wild-type and unedited), 5pED (pre-edited) and 5pDG (degenerate), were used to monitor the Drosha cleavage step and to quantitate simultaneously the expression levels of pri– and pre–miR-142 and mature miR-142-5p. The 25-nt products represent unprocessed pri–miR-142, whereas the 20-nt primer-extended products represent the pre–miR-142, mature miR-142-5p or both. The primer extension of both pre–miR-142 and mature miR-142-5p RNAs terminates at the same site generated by Drosha cleavage. The primer extension of pri– miR-142 extends five nucleotides upstream of the Drosha cleavage site in the presence of dideoxyATP. A separate set of three primers complementary to the miR-142-3p sequence, 3pWT, 3pED and 3pDG, were used to monitor the Dicer cleavage step and to quantitate simultaneously the expression levels of pri– and pre–miR-142 and mature miR-142-3p. The 26-nt products represent pri– and/or pre–miR-142, whereas the 22-nt primer-extended products represent mature miR-142-3p. The primer extension of both pri– and pre–miR-142 RNAs terminates at the same location in the presence of dideoxyTTP. The primer extension of mature miR-142-3p RNA terminates at the site generated by Dicer cleavage. Degenerate primers (5pDG and 3pDG) were used to monitor the total miR-142 RNAs (unedited and edited). Coloring is as in Figure 1b. The uracil and adenosine residues where the extension reaction is terminated by dideoxyATP and dideoxyTTP, respectively, are highlighted (filled box). Y in the DG primers denotes a random mix of T and C. (d) Analysis of pri– and pre–miR-142 and mature miR-142-5p and miR-142-3p RNAs processed in transfected HEK293 cells. Minor bands of primer-extended products with unexpected sizes (lanes 1 and 2 (22 nt) and lanes 9 and 10 (24 nt)) were also detected. At this time, we do not know how they are generated.

Suppression of pri–miR-142 processing in HEK293 cells transfected with ADAR expression plasmids. (a) Inhibitory effects of ADAR1p110 and ADAR2 on processing of pri–miR-142 in HEK293 cells. Primer-extension assays monitoring for expression levels of miR-142-5p (upper gel) and miR-142-3p (lower gel) as in Figure 2d are shown. (b) Histograms representing data like that in a from three independent experiments. The relative miR-142-5p and miR-142-3p levels in HEK293 cells cotransfected with the ADAR1p150, ADAR1p110 or ADAR2 expression plasmid were compared statistically by individual unpaired Student’s t-tests to those in HEK293 cells transfected only with pri-miR142WT plasmid. Significant differences are indicated as follows: one asterisk, P < 0.01; two asterisks, P < 0.001. Error bars, s.e.m. (c) A quantitative summary (as in Figure 1c) of the editing pattern revealed by sequencing of cDNA clones corresponding to pri–miR-142 RNAs edited in transfected HEK293 cells. White bars represent the results of experiments conducted with HEK 293 cells transfected solely with wild-type pri–miR-142 RNA expression plasmid.

Increased expression levels of miR-142 RNAs in spleen and thymus of ADAR null mice. (a) The wild-type sequence primer 5pWT was used for quantitation of mature miR-142-5p RNA expression levels as in Figure 2d. Identical results were obtained with a separate extension assay done by using the degenerate sequence primer 5pDG (data not shown). (b) The level of the mature miR-142-5p in wild-type and ADAR null mice. Results from three independent experiments are shown. The increase in the relative miR-142-5p levels in comparison to wild-type mouse tissue was examined statistically by individual unpaired Student’s t-tests. Significant differences are indicated as follows: one asterisk, P < 0.01; two asterisks, P < 0.001. Error bars, s.e.m. (c) The editing pattern (quantitated as in Figure 1c) revealed by sequencing of cDNA clones corresponding to the endogenous pri–miR-142 RNAs edited in vivo in spleens of wild-type, ADAR1 null and ADAR2 null mice.

A→I RNA editing of pri-miR RNAs by ADAR. (a) TLC analysis of pri-miRNAs edited in vitro by a mixture of recombinant ADAR1p110 and ADAR2 proteins (rADAR1p110 and rADAR2, 20 ng each). (b) A→I editing sites of four pri-miRNAs. Red, edited adenosine residues (numbered by position with the 5′ end of the mature miRNA sequence counted as +1); green, the region to be processed into the mature miRNA; arrows in pri–miR-142 hairpin structure, cleavage sites for Drosha and Dicer. (c) A quantitative summary of the editing patterns revealed by sequencing of RT-PCR cDNA clones corresponding to pri–miR-142 RNAs edited in vitro by rADAR1p150, rADAR1p110 or rADAR2. Editing frequency is represented as a percentage (number of independent cDNA clones representing the edited pri–miR-142 sequence at that site divided by the total number of cDNA clones examined).

In vitro processing of pri–miR-142 RNAs by miRNA processor complexes. (a) Processing of pre-edited miR-142 precursor RNAs. Unedited pri–miR-142 RNAs as well as three pre-edited pri–miR-142 RNAs containing guanosine residues substituted for adenosine at the +4, +5, +40 and/or +50 sites were subjected to the Drosha cleavage reaction using the Drosha–DGCR complex and, in some experiments, then to the Dicer cleavage reaction using the Dicer–TRBP complex, as described previously28,32. Stability and structural changes introduced in pre-edited pri–miR-142 RNAs by G•U or U•G pairs are equivalent to those of isosteric I•U or U•I pairs. The total length of the pri–miR-142 RNA pre-edited only at the +40 site (203 nt) is shorter than that of the unedited and the two pre-edited RNAs (341 nt). The band positions of the pre–miR-142 RNA correctly cleaved by Drosha (58 nt) and mature miR-142 RNAs generated by Drosha (20 or 22 nt) are indicated. Nonspecific RNase activity of Drosha described previously28 is probably responsible for generation of other RNA products (~35 nt and longer). M_r indicates a lane containing molecular size markers. (b) Processing of pri–miR-142 RNAs that had been edited in vitro to different extents (0%, 2.9% or 7.1% A→I modifications); gel is labeled as in a.

Degradation of highly edited pri–miR-142 RNAs by TSN. (a) In vitro assay for degradation of pri–miR-142 RNAs using purified TSN recombinant proteins. pri–miR-142 RNA edited to different extents (0%, 0.5%, 1.4%, 3.0% or 6.9% A→I modification) or pre-edited (four adenosine residues replaced by guanosine at the +4, +5, +40 and +50 sites) was subjected in vitro to endonucleolytic cleavage by TSN in the presence or absence of the inhibitor pdTp41. (b) Accumulation of highly edited pri–miR-142 RNAs in vivo in transfected HEK293 cells (quantitated as in Figure 1c) in the presence of the pdTp inhibitor.