Cleavage of Scarecrow-like mRNA Targets Directed by a Class of Arabidopsis miRNA

See allHide authors and affiliations

Science  20 Sep 2002:
Vol. 297, Issue 5589, pp. 2053-2056
DOI: 10.1126/science.1076311


Micro-RNAs (miRNAs) are regulatory molecules that mediate effects by interacting with messenger RNA (mRNA) targets. Here we show that Arabidopsis thaliana miRNA 39 (also known as miR171), a 21-ribonucleotide species that accumulates predominantly in inflorescence tissues, is produced from an intergenic region in chromosome III and functionally interacts with mRNA targets encoding several members of the Scarecrow-like (SCL) family of putative transcription factors. miRNA 39 is complementary to an internal region of three SCL mRNAs. The interaction results in specific cleavage of target mRNA within the region of complementarity, indicating that this class of miRNA functions like small interfering RNA associated with RNA silencing to guide sequence-specific cleavage in a developmentally controlled manner.

Micro-RNAs in eukaryotes are ∼21- to 22-ribonucleotide RNAs that arise from short stem-loop precursors through the activity of the double-stranded ribonuclease Dicer (1–6). The miRNAs from lin-4 and let-7 genes are involved in translational control through interaction with 3′-proximal sequences in target mRNAs in Caenorhabditis elegans(7–11). However, the range of functions for other miRNAs in plants, animals, and microorganisms has yet to be determined.

Arabidopsis contains numerous small RNAs, many of which resemble miRNAs identified in animals (12,13). Several of these, including miRNA 39, have sequence identity or complementarity to mRNAs for protein-coding genes. miRNA 39 accumulates predominantly in inflorescence tissue and was predicted to arise from a precursor gene in an intergenic region (IGR) in chromosome III (12, 13). The miRNA 39 sequence and predicted stem-loop structure of the putative precursor are conserved in the genome of rice (Oryza sativa) (13). We detected miRNA 39–like species in Arabidopsis, rice, andNicotiana benthamiana by blot assay (Fig. 1A). The Arabidopsis, rice, and N. benthamiana miRNA 39 accumulated to relatively high levels in inflorescence and flower tissues (Fig. 1A, lanes 3, 6, and 8) and to lower or nondetectable levels in leaf and stem tissue (Fig. 1A, lanes 1, 2, 4, 5, and 7).

Figure 1

Expression of SCL genes and miRNA 39. (A) RNA blot analysis of miRNA 39 fromArabidopsis and N. benthamiana leaf, stem, and inflorescence/flower tissue and from rice (O. sativa) leaf and flower tissue. (B) RNA blot analysis ofSCL6-III (left) and SCL6-IV (right) mRNAs. Normalized (10 μg per lane) RNA samples from Arabidopsisleaf, stem, and inflorescence tissue were analyzed in duplicate. The proportion of total RNA represented by the large cytoplasmic ribosomal RNA (rRNA) (shown in ethidium bromide–stained gels) differs between tissue types in Arabidopsis, which accounts for the more abundant rRNA in the inflorescence samples. The protein-coding regions (nucleotide numbering begins at the start codon) and the locations of the sequences complementary to miRNA 39 are shown in the expanded diagrams. SCL6-III– andSCL6-IV–specific probes detected both full-length mRNA [SCL6-III(a) and SCL6-IV(a)] and shorter RNA [SCL6-III(b) and SCL6-IV(b)] in inflorescence tissue. The positions corresponding to the 5′ ends of the SCL6-III(b) and SCL6-IV(b) RNAs determined by 5′ RACE, and the number of 5′-RACE clones corresponding to each site, are indicated by arrows.

The miRNA 39 sequence in Arabidopsis is perfectly complementary to an internal sequence in mRNAs of three members [SCL6-II (locus At2g45160), SCL6-III(locus At3g60630), and SCL6-IV (locus At4g00150)] of the Scarecrow-like family of putative transcription factors (12, 13). Four SCL genes in rice have complementarity to miRNA 39 (13). Members of theSCL family control a wide range of developmental processes, including radial patterning in roots and hormone signaling (14–17). No other small RNAs related to sequences from the 3′-proximal regions of SCL6-III orSCL6-IV genes were detected in Arabidopsis leaf, stem, or inflorescence tissues by blot assay [fig. S1A (18)], indicating that miRNA 39 was not part of a larger small interfering RNA (siRNA) population resulting from general RNA silencing triggered against broader segments of the SCLmRNAs. Blot assays indicated that SCL6-III andSCL6-IV genes were expressed in leaf, stem, and inflorescence tissues (Fig. 1B), although SCL6-III andSCL6-IV mRNAs were most abundant in inflorescence tissue (Fig. 1B, lanes 5, 6, 13, and 14). In addition to full-length mRNAs [designated SCL6-III(a) and SCL6-IV(a) RNAs], shorter RNAs [SCL6-III(b) and SCL6-IV(b)] of ∼1.4 and ∼1.3 kilobases (kb), respectively, were detected with a 3′-proximal probe in extracts from inflorescence tissue but not in extracts from stems or leaves (Fig. 1B, lanes 5, 6, 13, and 14).

The 5′ ends of SCL6-III(b) and SCL6-IV(b) RNAs were mapped by 5′ RACE (rapid amplification of cDNA ends) to positions corresponding to the middle of the respective sequences complementary to miRNA 39 (Fig. 1B). The 5′-RACE reactions involved ligation of an adapter to the 5′ end without enzymatic pretreatment, suggesting that both RNAs contained a ligation-competent 5′ monophosphate rather than a conventional 5′ cap. These results are consistent with a model in which the SCL6-III(b) and SCL6-IV(b) RNAs arise by sequence-specific cleavage of the full-length mRNAs at a site directed by miRNA 39. In effect, miRNA 39 may function like a single siRNA (19), produced in trans from a miRNA 39 precursor gene, that guides cleavage of targetSCL mRNAs. As siRNA-guided cleavage of silencing targets occurs at positions centered in the middle of siRNA-target RNA duplexes (20), both siRNA-guided and miRNA 39–guided cleavage might occur by a common mechanism within a sequence-specific nucleolytic complex termed RISC (RNA-induced silencing complex) (21, 22). This model predicts that accumulation of SCL6-III(b) and SCL6-IV(b) RNAs will correlate with accumulation of miRNA 39, which is consistent with the tissue-specific distribution patterns of miRNA 39 and SCL6-III(b) and SCL6-IV(b) RNAs (12) (Fig. 1, A and B).

We tested the hypothesis that miRNA 39 functions to direct cleavage ofSCL mRNAs with an Agrobacterium-mediated delivery (Agro-inoculation) system to coexpress miRNA 39 and SCL6-IV target mRNA in N. benthamiana leaf tissue. We reasoned that a miRNA 39 precursor could be produced through transcription of the miRNA 39–containing IGR, either with an endogenous miRNA 39 promoter or with a 35S promoter, followed by Dicer-like processing (13) of the precursor in N. benthamiana. Three constructs with the complete miRNA 39–containing IGR, or empty vector as a control, were coexpressed with the SCL6-IV construct (35S:SCL6-IV). Two of the constructs (35S:IGR-mi39 and 35S:AS-IGR-mi39) contained a 35S promoter for production of synthetic IGR transcripts of sense (relative to miRNA 39) or antisense orientation, respectively, and one (IGR-mi39) contained the IGR sequence without a 35S promoter (Fig. 2A). A basal level of endogenous miRNA 39 was detected in nontreated leaves (Fig. 2A, lane 2) and in leaves that were co–Agro-inoculated with 35S:SCL6-IV and the empty vector (lanes 3 and 4). High levels of miRNA 39 were produced after co–Agro-inoculation with the sense-oriented 35S:IGR-mi39 construct (Fig. 2A, lanes 7 and 8), indicating that the synthetic RNA transcript was recognized accurately by a Dicer-like enzyme (13). Data showing that miRNA 39 accumulation in these experiments was not the result of generalized RNA silencing of SCL genes or the 35Spromoter–driven constructs are provided in the SOM text (18). Only low levels of miRNA 39 above background were detected in tissues expressing the antisense-oriented35S:AS-IGR-mi39 (Fig. 2A, lanes 9 and 10) or IGR-mi39 constructs (lanes 5 and 6). The low levels of miRNA 39 directed by theIGR-mi39 construct suggest that the IGR sequence contains relatively weak regulatory sequences for transcription of the miRNA 39 precursor gene, or that the miRNA 39 precursor gene is suppressed, inN. benthamiana leaf tissue.

Figure 2

Cleavage of SCL6-IV mRNA is directed by miRNA 39. (A) Constructs containing the miRNA 39–containing IGR and the SCL6-IV coding sequence were co–Agro-inoculated in N. benthamiana leaves, and normalized extracts were tested for miRNA 39 by RNA blot assay 2 and 3 days postinoculation (d p.i.). (B) Analysis of SCL6-IV (a) and (b) RNA forms in normalized extracts from co–Agro-inoculated tissue by RNA blot assay with probes corresponding to 3′ (top) or 5′ (bottom) regions of theSCL6-IV coding sequence. The amount of SCL6-IV(b) RNA, expressed as a percentage of total SCL6-IV RNA [(a) + (b)], is indicated for each sample. The blots were stripped and reanalyzed with a cytoplasmic rRNA probe. The position of sequence complementary to miRNA 39 is indicated in gray in the diagram. The size of the SCL6-IV(b) RNA (∼1.3 kb) corresponds to the 3′-proximal SCL6-IV sequence (∼1.0 kb) plus the 3′-untranslated sequence (∼0.3 kb). At,A. thaliana; Nb, N. benthamiana.

Coexpression of 35S:IGR-mi39 and 35S:SCL6-IVresulted in internal cleavage of the SCL6-IV mRNA (Fig. 2B, top, lanes 8 and 9). About 53 and 79% of theSCL6-IV–related RNA detected 2 and 3 days postinfiltration, respectively, corresponded to the SCL6-IV(b) RNA product. Relatively small proportions of SCL6-IV mRNA were cleaved after coexpression with empty vector (Fig. 2B, top, lanes 4 and 5),IGR-mi39 (lanes 6 and 7), and 35S:AS-IGR-mi39constructs (lanes 10 and 11), which can be explained by the basal or low levels of miRNA 39 in these samples. No cleavage products were detected with the SCL6-IV 5′-end probe (Fig. 2B), suggesting that the 5′ fragment was less stable than the 3′ fragment [SCL6-IV(b)]. Identical results were obtained when the SCL6-III mRNA was used as a target in coexpression assays (23).

To determine whether cleavage of SCL6-IV mRNA depends on perfect complementarity with miRNA 39, we introduced three mismatches into the sequence complementary to miRNA 39 in SCL6-IV RNA (construct 35S:SCL6-IV/mut39) (Fig. 3). Whereas the wild-type SCL6-IV mRNA was cleaved efficiently in the presence of the miRNA 39–producing 35S:IGR-mi39 construct (Fig. 3, lanes 8, 9, 18, and 19), the SCL6-IV/mut39 mRNA was completely resistant to cleavage (lanes 10, 11, 20, and 21). Furthermore, the low level of cleavage of SCL6-IV mRNA in the presence of the empty vector construct was inhibited by the mutations (Fig. 3, lanes 4 to 7), confirming that this activity was due to low levels of endogenous miRNA 39 in N. benthamianatissue.

Figure 3

Cleavage of SCL6-IV mRNA requires sequence complementarity to miRNA 39. Mutations (underlined) were introduced into the 35S:SCL6-IV/mut39 construct, and wild-type and mutant constructs were co–Agro-inoculated with the 35S:IGR-mi39 construct in N. benthamiana. SCL6-IV RNA forms (top) and miRNA 39 (bottom) in extracts at 2 and 3 d p.i. were analyzed by blot assay. The SCL6-IV RNA blot was hybridized with a probe corresponding to the 3′ end of theSCL6-IV coding region and then reanalyzed with a cytoplasmic rRNA probe.

The finding of miRNA-directed cleavage of several SCL mRNA targets in Arabidopsis indicates that there are at least two functional classes of miRNAs. Members of the small temporal RNA (stRNA)–like class, including C. elegans lin-4 andlet-7 miRNAs, down-regulate translation of target mRNAs but do not direct RNA target degradation or site-specific cleavage (7, 8, 10, 11). The stRNA class members do not interact with perfect complementarity to their natural targets, which may explain why they do not exhibit siRNA-like activity (24). In contrast, members of a class represented by Arabidopsis miRNA 39 interact with perfect complementarity and appear to mimic siRNA function to guide cleavage. We propose that miRNA 39 incorporates into a RISC-like complex identical or similar to the RISC complex that mediates target cleavage during RNA silencing (21, 22). Support for this concept also comes from the finding that engineered miRNA-target combinations with perfect complementarity result in target RNA cleavage (25, 26). Finally, miRNA 39–guided cleavage of mRNAs has several possible consequences, including developmentally coordinated inactivation of SCLmRNAs. Internal cleavage might also generate RNA products with novel characteristics or coding potential for truncated SCL proteins. Given the numbers of miRNAs that were recently discovered in eukaryotes (4–6, 12, 13,27, 28), additional members will likely be added to each class.

Supporting Online Material

Materials and Methods

SOM Text

Fig. S1

Table S1


  • * To whom correspondence should be addressed. E-mail: carrington{at}


Stay Connected to Science

Navigate This Article