Main source: Jones-Rhoades, M. W., Bartel, D. P., and Bartel, B. (2006) MicroRNAs and their regulatory roles in plants. Annual Reviews in Plant Biology 57, 19-53

1. Introduction

Recently discovered (which means about 20 years ago), the small RNAs are one of the most actively researched field in modern molecular biology. Those small RNAs are found playing very important roles in gene expression’s regulation in all eukaryotic species, the phenomenon called RNA interference (RNAi). Two most prominent of them are: micro RNAs (miRNAs) and small interfering RNAs (siRNAs). The structure, biogenesis and mechanism of action of miRNA, in comparison with those of siRNA,  will be described later in this review.

2. miRNAs-Overview

miRNAs were first discovered in C. elegans, a nematode (roundworm) chosen as model organism. They are short (20-24 nucleotide long), endogenously expressed and non-translated RNAs, generated from the stem-loop region of a longer precursor RNA. miRNAs are very conversed in both plants and animals, though there are some significant differences between those two. They play an important role in post-transcriptional and translational gene silencing, in which the sequences of target mRNA is found strikingly complementary to those of miRNAs. This extensive complementarity provides useful information for miRNAs’ prediction and discovery.

3. miRNAs’ discovery

Three basic approaches for miRNA’s discovery are: direct cloning, forward genetics screening and bioinformatics, followed by experimental verification.

a) Cloning is the most straightforward method. miRNAs were isolated from total RNA, amplified by RT-PCR and then sequenced. The disadvantage of this method is: it is very difficult to clone low-expressed miRNAs.

b) Forward genetics screening: the main idea is to search for the phenotype of interest among mutated population. It is the very method with which miRNAs were first discovered in C. elegans. However, this method is not very succesful in plants, since only a very limited number of miRNAs were discovered using it. The reasons could be: 1) the target region is very small, and 2) nearly all evolutionarily conversed plant miRNAs are encoded by gene families. Familiy members are likely to have overlapping functions, so the loss of function of one member doesn’t affect much. This redundance can be circumvented either by overexpression,  mutation in the targeted mRNA sequence.

c) Bioinformatics both important characteristics of miRNAs: they’re highly conversed and generate from stem-loop-precursors, are used to discover miRNAs. The whole genomes of species of interest and their related ones are screened for potential miRNA genes. The candidates are now experimentally verified.

d) And results: miRNAs in plants

118 potential miRNA genes are found in Arabidopsis. These 118 can be grouped into 42 families, 21 of thems (92 genes) are also conversed in other species. Although the sequences of mature miRNAs are highly conversed, the same cannot be said for the sequence, secondary structure and length of the intervening loop-regions.  Interestingly, the mismacht schemes are also found conversed, which may play an important role guiding the DLC1 (Dicer-like 1) to cut at the correct positions.

e) Nonconversed miRNAs and their challenges: most of miRNAs are highly conversed among flowering plants, but not all of them. Some miRNAs are found only in a single sequenced genome. Here a new problem arises: lacking conversion both of sequence and of secondary structure, it is difficult to say if this RNA is originated from a stem-loop precusor (i.e miRNAs) or from a double-stranded RNA (i.e miRNA).  More criteria are therefore needed for a confident miRNA annotation.

4. miRNA biogenesis

From miRNA gene, a primary transcript (pri-miRNA) is transcribed, which folds into a stem-loop structure. It is subsequently cleaved by ribonuclease complexes to a miRNA-miRNA* duplex. miRNA* is degraded, releasing the mature miRNA. Mature RNA is incorporated in a RISC (RNA-induced silecing complex), where it detects the target mRNA and inactivates it either by mRNA degradation or by translational inhibition. Methylation and cytoplasmic export are also involved in miRNA-processing.

The biogenesis of miRNA differs significantly between animals and plants.

a) Transcription of pri-miRNA

Most, if not all, plant miRNAs are produced from their own transcriptional units, which contrasts to animals where some miRNAs are found encoded in intron regions of protein-coding genes. At least some plant miRNAs are confirmed to be spliced, capped and polyadenylated. miRNA genes are usually preceded by typical TATA-box motifs. Therefore, it is suggested that plant miRNAs are transcribed by RNA-Polymerase II, like in animals. Little is known about transcriptional regulation, but it is expected to be similar to that of protein-producing transcription.

b) Processing and export

miRNA processing in animals (From: Novina, C. D. and Sharp, P. A. (2004) The RNAi revolution. Nature 430, 161-164)

These steps involve RNase III-type (double stranded RNA cutting) endonucleases.

In animals, those are Drosha, a nuclear localized enzyme which cut the flanking sequences releasing “pre-miRNA”, and Dicer, a cytoplasmic enzyme which cuts the loop region releasing miRNA-miRNA*-duplex. This pair of enzymes works stepwise: pri-miRNA is cut in nuclear by Drosha, exported to cytoplasma and cut again by Dicer.  miRNA-miRNA*-duplex often has a two-nucleotide overhang, similar to those of siRNA duplexes (also cut by Dicer).

In plants, DLC1 (Dicer-like 1) is responsible for both cuts. How DLC1 recognizes the cleavage positions is still unknown, but some researches suggest that not only the secondary structure but also the primary sequence ( the mismatch scheme) that may play a role in determining cleavage sites. DLC1 is also found forming a complex with HYL1, a dsRNA-binding protein.

In addition, another enzyme complex, HEN1 is also involved, which methylates the duplex in both 3′-ends on their 2′-OH groups. The methylation is believed to stabilize the duplex. Subsequently, miRNA-miRNA*-duplex is exported to cytoplasm.

c) Incorporation to RISC

The incorporation mechanism is not fully understood, but some similarities to siRNA-siRNA*-duplex are expected. In case of siRNA-siRNA*-duplex, the 5′-end of siRNA* is more stable than that of siRNA and bound to a Dicer-R2D2 complex. The other 5′-end (of siRNA) is bound to RISC. Upon loading the duplex to the Argonaute protein (APO), siRNA* is cleaved.

APO is the central component of all RISC. It contains 2 conversed domains, a RNA-binding PAZ-domain and a RNase-H domain Piwi.

5. miRNA’s mechanisms of action

Three basic suggested mechanisms for miRNA-induced gene silencing are: mRNA cleavage, translational repression and transcriptional silencing.

a) Small RNA-directed mRNA-cleavage is the best understood one. miRNA binds to complementary sequence in mRNA, guiding the Argonaute component of RISC to cut a phosphodiester bond within the sequence.  The fragments are released, allowing RISC to recognize and cut another mRNA molecule. The cleavage activity is believed to be related to Piwi domain of Argonaute proteins.

b) Additional mechanisms

– Translational repression: miRNA-mRNa-interaction is not always enough to trigger the cleavage of target mRNA. Instead, a different mechanism of silencing  is suggested: miRNA-mRNA binding directs the mRNA to the Processing bodies (P-bodies), the main sites of RNA storage and degradation, hence separates it from translational machinery and destabilizes it.

– Small RNA-directed transcriptional silencing: RNA-induced transcriptional silencing (RITS) complex, like RISC also containing Dicer-produced small RNA and Argonaute protein, in yeast is found to be able to intervene in heterochromatic modifications and thus disrupt the transcription. There is evidence that it might also be the case in plants.