MicroRNAs and their regulatory roles in animals and plants

Abstract

microRNAs (miRNAs) are an abundant class of newly identified endogenous non-protein-coding small RNAs. They exist in animals, plants, and viruses, and play an important role in gene silencing. Translational repression, mRNA cleavage, and mRNA decay initiated by miRNA-directed deadenylation of targeted mRNAs are three mechanisms of miRNA-guided gene regulation at the post-transcriptional levels. Many miRNAs are highly conserved in animals and plants, suggesting that they play an essential function in plants and animals. Lots of investigations indicate that miRNAs are involved in multiple biological processes, including stem cell differentiation, organ development, phase change, signaling, disease, cancer, and response to biotic and abiotic environmental stresses. This review provides a general background and current advance on the discovery, history, biogenesis, genomics, mechanisms, and functions of miRNAs. J. Cell. Physiol. 210: 279–289, 2007. © 2006 Wiley-Liss, Inc.

MicroRNAs (miRNAs) are an abundant class of newly identified endogenous non-protein-coding small RNAs with 20–25 nucleotide length (Ambros, 2001, 2004; Carrington and Ambros, 2003; Bartel, 2004). A majority of identified miRNAs are highly evolutionarily conserved among many distantly related species, some from worms to human in animals (Pasquinelli et al., 2000), and mosses to high flowering eudicots in plants (Axtell and Bartel, 2005; Zhang et al., 2006c), suggesting that miRNAs play a very important role in essential biological processes, including developmental timing (Lee et al., 1993), stem cell differentiation (Houbaviy et al., 2003; Hatfield et al., 2005; Zhang et al., 2006b), signaling transduction (Guo et al., 2005; Karp and Ambros, 2005; Kwon et al., 2005), disease (Labourier et al., 2004; Alvarez-Garcia and Miska, 2005), and cancer (Hayashita et al., 2005; Lu et al., 2005b). Currently, miRNAs have been considered one of the most important regulatory molecules, which regulate gene expression at the posttranscriptional levels by targeting mRNAs for direct cleavage of mRNAs or repression of mRNA translation.

HISTORY

One decade ago, two research groups surprisingly discovered that a small 21-nucleotide RNA molecule called lin-4 controls developmental timing in Caenorhabditis elegans by the posttranscriptional regulation of the heterochronic gene lin-14 (Lee et al., 1993; Wightman et al., 1993) which plays an important role in controlling the temporal pattern formation (Ambros and Horvitz, 1987). They also observed that lin-4 didnot code protein, and contained antisense sequences complementary to a repeated sequence element in the 3′ untranslated region (UTR) of the lin-14 mRNA (Lee et al., 1993; Wightman et al., 1993). Although they hypothesized that lin-4 downregulated lin-14 expression via an antisense RNA–RNA interaction (Lee et al., 1993; Wightman et al., 1993), almost all scientists did not pay any attention on this new class of small RNAs, and a majority of them considered it as an oddity in C. elegans genome. Seven years later, let-7 was discovered as another small regulatory RNA in C. elegans with an exactly same regulatory mechanism as lin-4, and let-7 regulates gene expression which controls developmental timing and cellular differentiation in C. elegans (Reinhart et al., 2000). More interestingly, both sequence and developmental expression pattern of let-7 were highly conserved in a wide range of animal species, including vertebrate, ascidian, hemichordate, mollusc, annelid, and arthropod (Pasquinelli et al., 2000), suggesting that let-7 plays more important roles in biological processes than our previous thought. This finding attracted lots of attention from scientists, evidence by three papers related to this class of small RNA were published in a same issue of the most famous journal Science in the following year. In the three papers, an extensive number of small RNAs similar to lin-4 and let-7 were identified in invertebrates and vertebrates as well as in C. elegans (Lagos-Quintana et al., 2001; Lau et al., 2001; Lee and Ambros, 2001). It was also the first time to recognize this abundant class of newly identified small RNAs as miRNAs. At that time, miRNAs lin-4 and let-7 were recognized as the founding members of miRNAs.

Since miRNAs were recognized in 2001, the broad significance of these new identified small RNAs is becoming clear and is being fully appreciated, because more and more evidences suggest that miRNAs play an essential role in multiple biological processes. In the past 5 years, a huge amount of papers related to miRNA research have been published, and miRNA-related research has become one of the hottest research fields in biology. Currently, about 4,000 of miRNAs have been identified in a variety of animals, plants, and viruses, and have been deposited in publicly available databases, such as miRBase (Griffiths-Jones et al., 2006). Computational approaches have estimated that organisms probably contain about 1–5% miRNA genes of the total protein-coding genes (Lai et al., 2003; Lim et al., 2003b; Lewis et al., 2005), and about 30% of protein-coding genes may be regulated by miRNAs (Lewis et al., 2005). Right now, the mechanisms of miRNA-guided gene regulation and the functions of miRNAs in animals and plants are becoming clear. All of these achievements benefit from the following technological progress.

First, cloning technology makes it possible to identify new miRNAs or to confirm miRNAs predicted by computational approaches. Although the founding members of miRNAs, lin-4 and let-7, were identified by a genetic screening technology (Lee et al., 1993; Wightman et al., 1993; Pasquinelli et al., 2000), the application of this method is limited because it is expensive, time consuming, and dominated by chance. Recently, scientists have improved the cloning technology to better clone small RNAs including miRNAs from different organisms. Using the modified method, lots of new miRNAs have been identified in animals (Lagos-Quintana et al., 2001; Fu et al., 2005) and plants (Lu et al., 2005a; Sunkar et al., 2005). One of the most important advantages of direct cloning is that this is a unique method to clone new miRNAs, especially for tissue-specific or species-specific miRNAs, no matter whose genomes are encoded or not. Currently, several companies have developed commercial kits for miRNA isolation and cloning. The modified cloning method usually includes the following five steps: isolating small RNAs from biological samples, ligating the isolated small RNAs to an adaptor of oligonucleotides, reverse transcription of the ligated small RNAs, amplification by PCR, and sequencing.

Second, computational approaches provide a useful complement method to direct cloning, and the application of computational approaches has dramatically increased the numbers of identified miRNAs. At present, several computational programs, such as miRScan (Lim et al., 2003b), miRseeker (Lai et al., 2003), findMiRNA (Adai et al., 2005), and miRAlign (Wang et al., 2005b), have been developed and successfully applied to predict miRNAs in a various animal and plant species.

Third, expressed sequence tag (EST) and genomic sequence survey (GSS) analysis enhances the traditional computational approaches and genetic approaches to find miRNA homologs. Although computational approaches are good approaches to predict miRNAs, the application of these methods are limited because they require genome sequences which are only available for several model organisms, such as human and C. elegens in animals, and Arabidopsis thaliana and rice in plants. For many identified miRNAs, they are evolutionarily conserved from species to species, suggesting a powerful tool to identify miRNA homologs using the huge EST and GSS databases which are publicly available. EST analysis has been employed to identify 100s of miRNAs in animals (Weber, 2005) and plants (Zhang et al., 2005, 2006c; Dezulian et al., 2006). EST analysis was also used to find some evidences that miRNAs are conserved among different species (Floyd and Bowman, 2004; Jones-Rhoades and Bartel, 2004). One important advantage of EST analysis is that mining EST databases in a systematic way could provide a deeper insight in the distribution and conservation of miRNAs, and EST analysis can be used to identify miRNAs in any species no matter their genome is encoded or not. Using EST and GSS analysis, Zhang and colleagues (2006) identified a total of 481 miRNAs, belonging to 37 miRNA families in 71 different plant species, and found that many miRNA families were evolutionarily conserved across all major lineages of plants, including mosses, gymnosperms, monocots, and eudicots (Zhang et al., 2006a, c). This finding suggests that regulation of gene expression by miRNAs appears to have existed at the earliest stages of plant evolution and has been tightly constrained (functionally) for more than 425 million years (Zhang et al., 2006c).

Fourth, quantitive PCR and microarray technology make it possible to better study the expression patterns and functions of miRNAs. Hundreds of miRNAs have been identified in a single cell, it is impossible to detect the expression patterns of all miRNAs in a single time using traditional Northern blotting, which is important to study the functions of miRNAs in multiple biological processes, especially in disease and cancer. miRNA-specific quantitive PCR and microarray technology make this study become easier. In the past 1 year, several quantitive PCR (Chen et al., 2005; Raymond et al., 2005; Shi and Chiang, 2005) and microarray technology (Babak et al., 2004; Barad et al., 2004; Nelson et al., 2004b; Thomson et al., 2004; Liang et al., 2005) have been developed, and successfully used to study the functions of miRNAs in organism development (Miska et al., 2004), disease and cancer (Lu et al., 2005b). At present, several companies have developed miRNA chips for studying the expression profiles at a specific situation.

MicroRNA biogenesis

Mature miRNAs only contain ∼20–22 nucleotides. However, the genes coding miRNAs are much longer than mature miRNAs; they usually contain several 10s of nucleotides even 100s of nucleotides. The length of miRNA genes varies from miRNAs to miRNAs and from species to species. For example, miRNA genes in plant species are usually longer than in animals.

It is a multiple-step biological process to generate a mature miRNA from a miRNA gene, and several enzymes play critical roles in the process. First, miRNA genes are transcribed to primary miRNAs (pri-miRNAs). This process is facilitated by polymerase II enzyme (pol II), and a polymerase II-depend promoter was identified in several miRNA genes (Lee et al., 2004). Lots of miRNA genes exist in plants and animals. Computational approaches establish that the number of miRNA genes may be more than 1–5% of the total protein-coding genes (Lai et al., 2003; Lim et al., 2003a, b; Lewis et al., 2005). Generally, miRNA genes exist in any location of a genome, including introns and exons. However, a majority of the characterized miRNA genes are located at the internal space of two protein-coding genes and they have a different transcribed direction to neighboring protein-coding genes. This indicates that miRNA genes may be transcribed as independent units. However, there is still some evidence indicating that miRNAs may be transcribed together with protein-coding genes. Smalheiser (2003) found that pri-miRNA-mRNA transcripts existed in human and mouse EST databases. Like protein-coding mRNAs, pri-miRNAs can be spliced, may be caped by 7-methyl guanosine at its 5′ site, and added a polyadenylated tail at its 3′ site (Cai et al., 2004) although they may not contain an open reading frame (ORF). After pol II transcribes miRNA genes, the 5′ capped and 3′ polyadenylated pri-miRNA forms a specific hairpin-shaped stem-loop secondary structure and enters a large complex called microprocessor complex (500–650 kDa) which constitutes of a Drosha (a RNase III endonuclease) and an essential cofactor DGCR8/Pasha (a protein contains two double-stranded RNA binding domains) (Denli et al., 2004; Gregory et al., 2004; Han et al., 2004a; Landthaler et al., 2004). In this new identified microprocessor complex, DGCR8 first recognizes the distinct stem-loop structures and binds to the pri-miRNAs (Denli et al., 2004; Gregory et al., 2004; Han et al., 2004a; Landthaler et al., 2004), then Drosha asymmetrically and specifically cut the both strands of the hairpin-shaped stem at the sites near the base of the stem loop; and finally release a 60- to 70-nt pre-miRNAs that have a 5′ phosphate and a 3′ 2 nt overhang. This mechanism has been deeply understood in animals (Denli et al., 2004; Gregory et al., 2004; Han et al., 2004a; Landthaler et al., 2004). The pre-miRNAs are then transported to the cytoplasm by Exportin-5 (Exp5) (a member of the Ran transport receptor family) (Bohnsack et al., 2004; Yi et al., 2003; Lund et al., 2004). This transport process requires energy, a specific hairpin secondary structure, and another important factor Ran (Yi et al., 2003; Bohnsack et al., 2004; Lund et al., 2004). Once in the cytoplasm, pre-miRNAs are further processed by Dicer, a second RNase III endonuclease (Grishok et al., 2001; Hutvagner et al., 2001; Ketting et al., 2001). In this step, the PAZ domain of Dicer was thought to first recognize the 2-nucleotide 3′ overhang, then Dicer cuts off the about 20 ∼ 22 nucleotide length double strand miRNA:miRNA* duplex with 5′ phosphate and a 3′ 2 nt overhang from the end of the hairpin structure stem (Zhang et al., 2004). Finally, miRNA:miRNA* duplex is unwound by helicase into two single strands, mature miRNA and miRNA*, and miRNA* is degraded by an unknown enzyme nuclease while mature miRNA is incorporated into an ribonucleoprotein effector complex known as RNA-induced silencing complex (RISC) which induces gene silence at a posttranscriptional level (Schwarz et al., 2003; Hammond, 2005; Khvorova et al., 2003).

Although miRNA biogenesis in plants is similar within animals, both are transcribed from miRNA genes by pol II, and sequentially produce pri-miRNAs and pre-miRNAs, and finally form mature miRNAs. However, there are several differences between animal miRNA biogenesis and plant miRNA biogenesis. First, no homolog of Drosha and DGCR8/Pasha have been identified in plants, so the similar mechanism processing pri-miRNAs to pre-miRNAs in plants is still unclear. However, evidences have demonstrated that plant pre-miRNA formation is mediated by the Dicer-like protein 1 (DCL1) (Park et al., 2002; Reinhart et al., 2002; Papp et al., 2003; Kurihara and Watanabe, 2004). With the helps of other factors, such as HYL1 (a two dsRBD-containing nuclear protein) (Han et al., 2004b; Vazquez et al., 2004) and HEN1 (a protein with a dsRBD and a methyltransferase domain) (Boutet et al., 2003; Park et al., 2005; Yu et al., 2005), DCL1 further cleaves a pre-miRNA to a miRNA:miRNA* duplex in nucleus instead of in cytoplasm which usually happen in animal miRNA biogenesis (Park et al., 2002; Reinhart et al., 2002; Papp et al., 2003; Kurihara and Watanabe, 2004). Recent study shows that it is a crucial step in plant miRNA biogenesis to methylate miRNA:miRNA* duplex at the 2′ hydroxyl groups on the 3′ most nucleotides (Yu et al., 2005). However, the biochemical function of the 2′ hydroxyl groups on miRNA biogenesis and similar mechanism existing in other organisms are still unclear. Following the formation of miRNA:miRNA* duplex, the duplex is transported to cytoplasm by HASTY, the plant ortholog of exp5 (Park et al., 2005), and then unwound by helicase and release mature miRNAs for mediating gene expression.

In RISC complex, miRNAs bind to targeted messenger RNA (mRNA) and inhibit gene expression by direct cleavage of targeted mRNAs or repression of translation through perfect or near-perfect complementarity between the miRNAs and the targeted mRNAs.

Mechanisms of microRNA-mediated gene regulation

Three mechanisms have been described for miRNA-mediated gene regulation: mRNA degradation, translational repression, and miRNA-mediated mRNA decay. No matter what kind of mechanisms, all miRNAs regulate gene expression at the posttranscriptional level.

Translational repression and mRNA degradation are two common mechanisms for miRNA-mediated gene regulation. In most cases, it is governed by the complementarity between miRNAs and targeted mRNAs. When an miRNA perfectly or near-perfectly pairs to the targeted mRNAs, it was thought that mRNA cleavage is the primary mechanism for miRNA-mediated gene regulation (Rhoades et al., 2002; Bartel, 2004). Otherwise, if a miRNA imperfectly pairs to its targeted mRNAs, translational repression is thought to be occurred. The degree of repression is associated with the number of miRNA-binding sites in a targeted mRNA (Cuellar and McManus, 2005). In translational repression, a majority of miRNAs bind to their targeted mRNAs at the 3′ UTRs; however, some miRNAs can also bind to the 5′ UTR and/or the ORF (Zeng et al., 2002; Doench and Sharp, 2004). Although the precise mechanism of miRNA-mediated translational repression has not been elucidated, studies suggest that miRNA may hamper ribosome movement along the mRNAs, and repress protein translation (Carrington and Ambros, 2003). However, not all miRNAs follow this role to regulation gene expression. For example, miR 172 regulates gene expression by repressing translation although it can perfectly complement to the targeted APETALA2 (AP2) mRNA (Aukerman and Sakai, 2003; Chen, 2004).

A majority of miRNAs downregulate gene expression by translational repression in animals while by mRNA degradation in plants. However, some miRNAs downregulates gene expression by translational repression in plants. For example, miR 172 regulate AP2 through translational repression despite miR 172 can perfectly complement with AP2 mRNA (Aukerman and Sakai, 2003; Chen, 2004). In animals, there are also miRNAs which directly degrade their targeted mRNAs. For example, miR-196 directly cleaves the mRNA of HOXB8 (Mansfield et al., 2004; Yekta et al., 2004; Hornstein et al., 2005), which play important role in animal development (van den Akker et al., 1999; Greer and Capecchi, 2002; Reilly, 2002; El-Mounayri et al., 2005).

Recently studies suggest a third mechanism of miRNA-mediated gene expression. After mRNA transcription, a ploy(A) tail always is added to the 3′ of the mRNA to keep mRNA more stable and avoid the occurrence of mRNA decay (Jacobson and Peltz, 1996; Coller and Parker, 2004). Recently, Wu et al. (2006) found that miRNAs are involved in mRNA decay by directing rapid deadenylation of mRNAs. Their finding indicates that miRNAs destabilize mRNAs by accelerating poly(A) tail removal as an initial step in mRNA degradation. Giraldez et al. (2006) also observed the same phenomena. In their study, they found that miR-430 accelerated the deadenylation of target mRNAs in zebrafish, and facilitated the deadenylation and clearance of maternal mRNAs during early embryogenesis and affected embryo development. This conclusion has been confirmed by several in vitro and in vivo studies (Bagga et al., 2005; Jing et al., 2005; Wu and Belasco, 2005), in which the cellular concentrations of target mRNAs were reduced by miRNAs that do not perfectly or near-perfectly complement to their targeted mRNAs.

Targets of microRNAs

miRNAs regulate gene expression through targeting mRNA for cleavage, translational repression or mRNA decay. Thus, identifying miRNA targets is a very important step to study miRNA functions in animal and plant development. In the past years, several approaches have been employed to identify miRNA targets. A genetic approach is the first approach to identify miRNA targets, which is based on the abnormal expression of targeted mRNAs in the miRNA loss-of-function mutants. This approach has been used to identify several miRNA targets that play an important role in worm development (Lee et al., 1993; Wightman et al., 1993; Reinhart et al., 2000; Johnston and Hobert, 2003). For all known miRNA targets, they have conserved perfect or near-perfect complementary sites of miRNAs (Llave et al., 2002; Pasquinelli and Ruvkun, 2002; Saxena et al., 2003; Bartel, 2004; Mallory et al., 2004a; Meister et al., 2004b; Ota et al., 2004; Vella et al., 2004; Bagga et al., 2005; Brown and Sanseau, 2005; Millar and Waterhouse, 2005), especially for plant miRNAs (Aukerman and Sakai, 2003; Wang et al., 2005a; Williams et al., 2005b). This suggests a powerful strategy to predict miRNA targets by computational approaches. Based on this characteristic, several laboratories have developed different computational strategies to predict miRNA targets in available genome database, and successfully identified 100s of miRNA targets for given miRNAs (Rhoades et al., 2002; Enright et al., 2003; Lewis et al., 2003; Stark et al., 2003; Axton, 2004; Bonnet et al., 2004; John et al., 2004; Jones-Rhoades and Bartel, 2004; Kiriakidou et al., 2004; Lai, 2004; Rajewsky and Socci, 2004; Rehmsmeier et al., 2004; Wang et al., 2004; Axtell and Bartel, 2005; Bentwich, 2005; Brennecke et al., 2005; Brown and Sanseau, 2005; Burgler and Macdonald, 2005; Grun et al., 2005; Hariharan et al., 2005; Kawasaki and Taira, 2005; Krek et al., 2005; Legendre et al., 2005; Lewis et al., 2005; Li and Zhang, 2005; Nakahara et al., 2005; Robins et al., 2005; Saetrom et al., 2005; Williams et al., 2005a; Xie et al., 2005; Yoon and De Micheli, 2005; Zhang, 2005). These computer software programs include TargetScan (Lewis et al., 2003), TargetScanS (Lewis et al., 2005), miRanda (Enright et al., 2003; John et al., 2004), MovingTargets (Burgler and Macdonald, 2005), PicTar (Grun et al., 2005; Krek et al., 2005), RNAhybrid (Rehmsmeier et al., 2004), DIAN-AmicroT (Lim et al., 2005) for animals; and MIRcheck (Jones-Rhoades and Bartel, 2004), findMiRNA (Adai et al., 2005), miRU (Zhang, 2005), and PatScan* (Dsouza et al., 1997; Rhoades et al., 2002) for plants. Microarray technology was also recently employed to identify miRNA targets, and successfully identified 100s of miRNA targets (Lim et al., 2005).

Predicting and identifying miRNA targets by computational approaches is much easier in plants than in animals. This is due to the fact that complementarity between miRNAs and targeted mRNAs is much higher in plants than in animals for a majority of targets (Carrington and Ambros, 2003; Ambros, 2004; Bartel, 2004). Thus, for a majority of plant miRNAs have predicted targets although some of them have not been validated by experimental approaches. However, there are no targets found for a majority of animal miRNAs although more miRNAs are identified in animals than in plants. There are only one or a few targets for a majority of plant miRNAs (Rhoades et al., 2002), while there are lots, even 100s of targets for each animal miRNA (Enright et al., 2003; Lewis et al., 2003, 2005; Stark et al., 2003; John et al., 2004; Lai, 2004; Rajewsky and Socci, 2004; Bentwich, 2005; Krek et al., 2005).

Currently, a majority of plant miRNAs have identified targets. For example, miR 172 targets AP2 in Arabidopsis thaliana and gossy15 in maize for regulating flower development and developmental timing switch (Aukerman and Sakai, 2003; Chen, 2004; Lauter et al., 2005). For animal miRNAs, a majority of miRNAs have not been experimentally identified their targets although 100s of targets are predicted for each miRNA. However, about 30% of protein-coding genes are predicted to be negatively regulated by miRNAs (Lewis et al., 2005). This suggests miRNAs are the biggest regulator in gene regulation.

Functions of microRNAs in plants

Plant miRNAs play an important role in many aspects, including organ development, phage change, signal transduction, and response to environmental stress.

miRNAs regulate organ development in plants

Dicer-like enzyme 1 (DCL1) is an important enzyme that processes pri-miRNAs to pre-miRNAs then to miRNA:miRNA* duplexes (Park et al., 2002; Reinhart et al., 2002; Papp et al., 2003; Kurihara and Watanabe, 2004; Liu et al., 2005; Kurihara et al., 2006). Loss-of-function of the dcl1 gene reduced the expression level of mature miRNAs and consequently caused many developmental abnormalities (Liu et al., 2005; Kurihara et al., 2006). These developmental abnormalities include arrested embryos at early stages, altered leaf shape and morphology, delayed floral transition, and female sterility (Park et al., 2002; Reinhart et al., 2002; Liu et al., 2005). HASTY, an ortholog of exportin 5, is also an important protein in mature miRNA biogenesis, which transports the miRNA:miRNA* duplex from the nucleus to the cytoplasm in Arabidopsis (Park et al., 2005). Loss-of-function of hasty gene also caused pleiotropic developmental abnormalities, such as disrupting leaf shape and flower morphology, accelerating phase change, and reducing fertility (Bollman et al., 2003). All these finding indicate that miRNAs play an important role in a variety of developmental process in plants, and miRNAs regulate plant development at different organ level, including roots, stems, shoots, and flowers.

miRNAs control leaf development by regulating the expression of class-III homeodomain leucine zipper (HD-ZIP) transcription factor genes, which control leaf asymmetry pattern along adaxial/abaxial (upper/lower) axis (Juarez and Timmermans, 2004; Juarez et al., 2004). PHABULOSA (PHB), PHAVOLUTA (PHV), and REVOLUTA (REV) are three closely related Arabidopsis HD-ZIP transcription factors. Dominant mutations in any of these three transcription factor genes (phb, phv, and rev) results in radialization and adaxialization of leaf and vascular bundles in the stem (McConnell et al., 2001; Emery et al., 2003). Recently, several laboratories demonstrated that all of the three transcription factors are the targets of miR165 and miR166, and are regulated by these two miRNAs (Emery et al., 2003; Bao et al., 2004; Bowman, 2004; Juarez et al., 2004; Mallory et al., 2004b; Zhong and Ye, 2004; Kim et al., 2005; Williams et al., 2005b; Ko et al., 2006). miR165 and miR166 are evolutionarily conserved in all lineages of land plants, including mosses, ferns, gymnosperms, and angiosperms (Zhang et al., 2006c). Misexpression of miR165 and miR166 resulted in leaf developmental abnormalities in many plant species, including Arabidopsis and corn (Juarez et al., 2004). Except miR165 and miR166, miR159/Jaw also controls leaf development by regulating a subset of TCP transcription factor genes (Palatnik et al., 2003). Loss-of-function of miR159/Jaw caused uneven leaf shape and curvature (Palatnik et al., 2003).

miRNAs control apical meristem development by targeting several members of the NAM/ATAF/CUC(NAC)-domain transcription factors that play an important role in both embryogenic, floral, shoot and root development (Aida et al., 1997; Takada et al., 2001; Hibara et al., 2003). Inappropriate expression of miR164 resulted in abnormal expression of NAC-domian transcription factors, and caused many developmental abnormality, including shoot and root development (Laufs et al., 2004; Mallory et al., 2004a; Guo et al., 2005).

miRNAs regulate plant development most likely by regulating signal transduction. miR164 regulates auxin signals for lateral root development by targeting NAC mRNA cleavage (Guo et al., 2005). miR167 regulates plant development by negatively regulating the expression of several AUXIN RESPONSE FACTORS (ARF), including ARF 2, ARF 3, ARF 4, ARF 10, ARF 16, and ARF 17 (Mallory et al., 2005; Sorin et al., 2005; Williams et al., 2005a; Yang et al., 2006).

miRNAs regulate phase change in plants

It is an important transition in plants from vegetable growth to reproductive growth. This is always evidenced by the appearance and formation of appropriate floral organs. More and more evidences show that plant phase change is controlled by miRNAs. APETALA 2 (AP2) is one of the class A genes that control plant flowering time and floral morphology (Lohmann and Weigel, 2002). Overexpression of miR172 inhibited translation of AP2 and AP2-like mRNAs, and resulted in early flowering and disrupting the specification of floral organ identity in Arabidopsis and maize (Aukerman and Sakai, 2003; Chen, 2004; Lauter et al., 2005).

miR159, miR156, and miR171 are also involved in phase change and floral development (Llave et al., 2002; Achard et al., 2004; Schwab et al., 2005). miR171 is predominantly expressed in inflorescence and floral tissues instead of in leaf and other vegetative tissues (Llave et al., 2002). Overexpression of miR159 results in a low level expression of LEAFY and further perturbed anther development and delayed flowering in a short-day photoperiod (Achard et al., 2004). miR156-affected plant phase transition, including quickly initiating the formation of rosette leaves, by negatively regulating the expression of a Squamosa promoter binding protien like (SPL) plant-specific transcription factor (Schwab et al., 2005).

miRNAs regulate plant responses to environmental stresses

Abiotic and biotic stresses are a big issue for plant growth and development, lots of field studies show that environmental stress caused about 20–30% yield loss, and some may destroy crop yield (Zhang et al., 2000; Mahajan and Tuteja, 2005; Vinocur and Altman, 2005; Yamaguchi and Blumwald, 2005). In the evolution, crops have evolved different mechanisms to resist different environmental stress, including salinity, cold, drought, and pests. Although several genes have been identified and isolated from plants (Zhu, 2001; Nakashima and Yamaguchi-Shinozaki, 2006; Valliyodan and Nguyen, 2006), the principle mechanism of plant resistance still remains unknown. Recently, increasing evidences suggest that miRNAs may play an important role in plant response to biotic and abiotic stresses (Zhang et al., 2006d).

miRNAs are involved in plant diseases. Some of these miRNAs may get involved in virus-induced gene silencing. Helper component-proteinase (HC-Pro), p19, p21, and p69, are unrelated viral suppressors of gene silencing, and they play important roles in the virus response to plant antiviral silencing response (Plisson et al., 2003; Chapman et al., 2004). Several investigations demonstrated that several miRNAs are related the activity of these viral suppressors (Kasschau et al., 2003; Chapman et al., 2004; Chen et al., 2004; Llave, 2004). HC-Pro inhibited the expression level and activity of miR171, and caused miR171-related developmental deficiency (Kasschau et al., 2003). P69 enhanced the expression and activity of miRNAs, and caused rapid degradation of miRNA-targeted mRNAs, and consequently enhance plant resistance to pathogens (Chen et al., 2004). A recent study demonstrated that bacterial flagellin-derived peptide induced an overexpression of miR393 in Arabidopsis (Navarro et al., 2006). miR393 negatively regulated F-box auxin receptors (TIR1, AFB1, AFB2, and AFB3), and resulted in inhibit bacteria Pseudomonas syringae growth and increased plant resistance to pathogens (Navarro et al., 2006).

miRNAs are also involved in plant response to other stresses. Our study indicated that 25.8% of ESTs containing miRNAs were identified in stress-induced plant tissues, including pathogen-, salt-, drought-induced tissues (Zhang et al., 2005). miR395 was overexpressed under sulfate starvation conditions (Jones-Rhoades and Bartel, 2004). miR319 was induced by either cold or other stress (Sunkar and Zhu, 2004). miR402 was strongly induced by drought, cold, and/or salinity (Sunkar and Zhu, 2004). Recently, Lu et al. (2005c) identified 48 miRNA sequences from the Populus genome, and some of them were induced by mechanical stress and may function in a critical defense system for structural and mechanical fitness.

Functions of microRNAs in animals

Increasing evidences suggest that miRNAs have versatile multiple biological functions in animals, although only few targets of animal miRNAs have been identified and the function of very few miRNAs have been worked out in details. The function of animal miRNAs has been studied by several approaches. As we know, Dicer and Argonaute2 are two important enzymes in biogenesis and functions of miRNAs (Hutvagner et al., 2001; Tijsterman and Plasterk, 2004; Chendrimada et al., 2005). Loss-of-function of these two enzymes decreases the expression of global miRNAs, and can be useful to study the global functions of miRNAs (Meister et al., 2004b; Karube et al., 2005). Knockdown and/or knockout or overexpression of specific miRNAs is a good approach to investigate the specific function of an unique miRNA (Hutvagner et al., 2004; Meister et al., 2004a; Lee et al., 2005). Recently developed miRNA microarray technology and miRNA-specific real-time PCR also provide useful information on miRNA functions (Miska et al., 2004; Baskerville and Bartel, 2005; Liang et al., 2005).

miRNAs regulate developmental timing

Two founding members of miRNAs, also the two best-studied miRNAs, lin-4 and let-7, regulate developmental timing in C. elegans (Lee et al., 1993; Wightman et al., 1993; Reinhart et al., 2000). Loss-of-function of lin-4 and let-7 result in retarded development. However, lin-4 controls worm development at an early stage at the first larval stage (L1) and controls worm developmental transition from the L1 stage to the L2 stage (Lee et al., 1993); let-7 controls worm development at a late stage and controls worm developmental transition from the L2 stages to the L3 stage (Reinhart et al., 2000). Lin-4 control worm development by negatively regulating the expression of two genes lin-14 and lin-28, in which both have antisense complementary sites to lin-4 (Lee et al., 1993; Wightman et al., 1993; Moss et al., 1997).

miRNAs regulate animal development

miRNAs regulate animal development at multiple tissues and at multiple developmental stages. Emerging evidences suggest that miRNAs are essential for the normal development of almost all animal tissues (Lagos-Quintana et al., 2002), including stem cell (Houbaviy et al., 2003; Suh et al., 2004; Forstemann et al., 2005; Hatfield et al., 2005; Kanellopoulou et al., 2005; Lechman et al., 2005; Lee et al., 2005; Murchison et al., 2005), embryo (Kloosterman et al., 2004; Aboobaker et al., 2005; Alvarez-Garcia and Miska, 2005; Biemar et al., 2005; Kanellopoulou et al., 2005; Schulman et al., 2005; Wienholds et al., 2005; Wienholds and Plasterk, 2005; Yang et al., 2005), brain (Krichevsky et al., 2003; Miska et al., 2004; Rogelj and Giese, 2004; Sempere et al., 2004; Giraldez et al., 2005; Lugli et al., 2005; Nelson and Mourelatos, 2005; Rogaev, 2005; Rowan, 2005; Wu and Belasco, 2005; Nelson et al., 2006; Schratt et al., 2006), heart (Lagos-Quintana et al., 2002; Schubert, 2005; Zhao et al., 2005b), and limb (van den Akker et al., 1999; Harfe et al., 2005; Hornstein et al., 2005; Lancman et al., 2005; Maatouk et al., 2005), liver and other tissues (Lagos-Quintana et al., 2002). Specific miRNAs control specific tissue development at a specific developmental stage (Lagos-Quintana et al., 2002). Dicer mutant in zebrafish embryos developed normal at the beginning, but embryo development arrested ∼8 days after fertilization (Wienholds et al., 2003). This suggests that embryo development need appropriate expression of certain miRNAs. HOX is an important gene in animal development, and it is negatively regulate by miR-196 and miR-181 (Yekta et al., 2004; Naguibneva et al., 2006). Misexpression of these miRNAs caused abnormal expression of HOX, and results in animal developmental abnormality (Mansfield et al., 2004; Yekta et al., 2004; Guenther et al., 2005; Naguibneva et al., 2006). One of the muscle-specific miRNAs, miR-1, controls the balance between differentiation and proliferation during cardiogenesis by negatively regulating the critical cardiac regulatory proteins (Zhao et al., 2005a).

miRNAs and cancer

One of the biggest progresses on miRNAs is to find that miRNAs play an important role in cancer pathogenesis. Cancer is the most difficult cured human disease. Although several genes, including oncogenes and tumor suppressor genes, have been identified in human and/or other model animal genomes, to solve the mechanism of cancer formation is still far away from what we thought. However, more and more evidences suggest that miRNAs play an important role in cancer, and better understanding the function of miRNAs in cancers will provide a new insight for cancer research.

The function of miRNAs in cancers is initially recognized from two finding: one is that a majority of miRNAs were located in the cancer-associated genomic regions or in fragile sites (Calin et al., 2004); another is that miRNAs were differential expression between cancer cells and the related normal tissue cells (Lu et al., 2005b). The first evidence for miRNAs involved in cancer came from a molecular study characterizing the 13q14 deletion in human chronic lymphocytic leukemia (CLL) (Calin et al., 2002), the most common form of adult leukemia in the Western world (Dohner et al., 2000). With the development of mature miRNA microarray technology (Babak et al., 2004; Barad et al., 2004; Liu et al., 2004; Nelson et al., 2004a, b; Thomson et al., 2004; Liang et al., 2005; Lu et al., 2005b), more and more evidences become available to connect miRNAs with cancers.

Almost all cancers have been found to be related with the abnormal expression of miRNAs, and some miRNAs may play a critical role in many common and important cancers, including cancers in lung, brain, breast, blood, liver, colon, lymphomas, thyroid, and testicular germ cell (see review Zhang et al., 2006f). One good example is lung cancer. Lung cancer is one of the common cancers and the leading cause of cancer-related deaths, especially in many economically developed countries. Several evidences indicates that miRNA let-7 may controls lung cancer development, or at least plays an essential role in the pathogenesis of lung cancer (Takamizawa et al., 2004; Eder and Scherr, 2005; Johnson et al., 2005; Morris and McManus, 2005). The expression of let-7 was significantly reduced in lung cancer, and reduced let-7 expression was associated with shortened postoperative survival (Takamizawa et al., 2004). In contrast, overexpression of let-7 inhibited lung cancer cell growth (Takamizawa et al., 2004). Following study indicates that let-7 regulated lung cancer pathogenesis by negatively regulating the expression of oncogene RAS (Johnson et al., 2005). RAS is a well-studied oncogene which interferes with p53 pathway in lung cancer, and it has multiple complementary sites to let-7 in its 3′UTR (Johnson et al., 2005). Thus, miRNA let-7 inhibited RAS mRNA translation, and inhibited lung tumor cell differentiation and growth (Johnson et al., 2005). Recent study demonstrated that the expression of let-7 was significantly reduced while the expression of RAS was dramatically increased in lung tumor tissues (Johnson et al., 2005). This suggests that miRNA let-7 may functions as a tumor suppressor gene in human and other mammalians.

miRNAs and disease

Except cancers, miRNAs are also involved in a broad spectrum of human disease. A link between miRNAs and human disease came from the identification of an essential multisubunit protein complex termed microprocessor that is necessary and sufficient for processing miRNA precursor RNAs (Denli et al., 2004; Gregory et al., 2004; Han et al., 2004a; Landthaler et al., 2004; Gregory and Shiekhattar, 2005). In the microprocessor complex, an essential cofactor is DiGeorge syndrome critical region gene 8 (DGCR8) (Denli et al., 2004; Gregory et al., 2004; Han et al., 2004a; Landthaler et al., 2004), which is located at chromosomal region 22q11.2 and is commonly deleted in DiGeorge syndrome (Denli et al., 2004; Gregory et al., 2004; Han et al., 2004a; Landthaler et al., 2004). DiGeorge syndrome is a rare congenital disease. Although the symptoms vary greatly between individual patients, the common symptoms include a history of recurrent infection, heart defects, and characteristic facial features (Baldini, 2004).

miRNAs maybe also involved in virus-related or -induced disease and immune defense. Lecellier et al. (2005) found that a cellular miRNA, miR-32, mediated antiviral defense in human cells, and regulated primate foamy virus type I (PFV-1) proliferation. Several mammalian viruses, including the herpesvirus family, such as Epstein-Barr virus (EBV), simian virus 40 (SV40), and Kaposi sarcoma-associated virus, have been found to code miRNAs (Pfeffer et al., 2004, 2005; Cai et al., 2005; Grey et al., 2005; Omoto and Fujii, 2005; Sullivan et al., 2005; Sullivan and Ganem, 2005a, 2005b; Cai and Cullen, 2006; Jiang et al., 2006; Schuetz and Sarnow, 2006; Simon-Mateo and Garcia, 2006). Although the function of these coding miRNAs is still unknown, it may play a role in human disease development.

Except these functions mentioned, miRNAs also regulate programmed cell death (Brennecke et al., 2003; Xu et al., 2003; Cheng et al., 2005; Cimmino et al., 2005), miRNA and siRNA biogenesis (Bartel, 2005), insulin secretion (Poy et al., 2004), and metabolic processes (Xu et al., 2003; Esau et al., 2006).

CONCLUSIONS AND FUTURE PERSPECTIVES

miRNAs have attracted lots of interests from scientists due to their versatile functions in development, signaling, disease and cancers, and become one of the most important gene regulators. However, miRNA-related research and their application are still in infants, and several mysteries are still in secret, such as what is the origin of miRNAs, what regulates miRNA expression. Although thousands of miRNAs have been identified, their targets are still unclear, especially for a majority of animal miRNAs. The future application of miRNA-related researches will become brighter if miRNA targets are identified and the regulatory mechanisms are clear.

Recently, Zhang and colleague proposed that designed artificial miRNAs may be used to suppress gene expression for knockdown of targeted genes and studying gene function (Zhang et al., 2006d). This hypothesis was quickly confirmed in multiple plant species by two recent studies at two individual laboratories (Alvarez et al., 2006; Schwab et al., 2006). Schwab et al. (2006) designed a range of artificial miRNAs to target various endogenous protein-coding genes in Arabidopsis thaliana. Their results indicated that artificial miRNAs were highly expressed while their proposed targets were downexpressed in Arabidopsis thaliana. This finding was also confirmed in other plant species, tobacco and tomato (Alvarez et al., 2006). Both studies indicated that artificial miRNAs only knocked down their predicted targets. This suggests that artificial miRNAs could be widely used to knock down gene expression in a range of organisms. One advantage is that artificial miRNAs not only knock down one single gene, but also they can knock down several related, but not identical, targeted genes simultaneously. This provides a powerful tool to study gene function with multiple genes in a gene family, which usually exist in animals and plants. Both groups also found that artificial miRNAs were expressed at an inducible and/or tissue-specific manner with limited nonautonomous effects by using inducible or tissue-specific promoters which is an advantage over siRNAs (Alvarez et al., 2006; Schwab et al., 2006). These studies suggest that artificial miRNAs may become a new genetic approach to study and improve gene function in organisms. Due to the fact that miRNAs are highly conserved from species to species in animals (Pasquinelli et al., 2000) and plants (Zhang et al., 2006c), the similar strategy may also be used to cancer clinic and prevention study, such as preventing cancer development by knockout of oncogene.