Introduction and Historical Background

Biological revolutions are, sometimes, born out of experimental artifacts or from completely unexpected experimental results. Such is the case for the RNA interference (RNAi) and microRNA (miRNA) revolution.

I will provide a brief historical account of the birth of this fascinating field; I offer my sincere apologies for any unintended omissions. The reviews that follow present in more detail our current understanding of the field and how it impacts neurobiology and neuropathology, especially brain tumors and neurodegenerative diseases.

WHITE PETUNIAS, WHITE FUNGI AND TWITCHING NEMATODES

To enhance the purple color of petunia flowers, the Jorgensen and Mol labs introduced in petunia plants a transgene encoding the pigment producing gene chalcone synthase (21, 28). Contrary to expectations, in many plants the flowers turned out white or variegated, implying that the overexpressed transgene suppressed the expression of itself and of the endogenous gene. A similar phenomenon (termed quelling) was observed by the Macino lab when orange (wild‐type) strains of the fungus Neurospora crassa, were transformed with albino transgenes. Albino genes are normally present in the genome of Neurospora and are required for carotenoid biosynthesis. Many of the transformed fungi showed a white phenotype suggesting that the overexpressed transgene inactivated (silenced) itself and the endogenous albino genes (3, 25). Studies in plants showed that viruses can also elicit gene silencing, and work by many laboratories indicated that repression of gene expression in plants and fungi may occur both at the transcriptional (termed transcriptional gene silencing) or post‐transcriptional (termed post‐transcriptional gene silencing; PTGS) levels (reviewed in (29, 30). The trigger for target gene silencing was postulated to involve the synthesis of “aberrant” RNAs homologous to the silenced gene. At the same time, studies of embryonic polarity in the nematode Caenorhabditis elegans by the Kemphues lab using antisense RNA approaches, led to the unexpected discovery that RNAs with sense polarity were as potent in knocking down target gene expression as antisense RNAs (9). The significance of these observations became clear when the Fire and Mello labs demonstrated that the critical trigger for target gene silencing is actually double‐stranded RNA (dsRNA), which almost always contaminates RNA prepared by in vitro transcription (7). Injection in worms of minute amounts of dsRNA corresponding to the endogenous C. elegans unc‐22 mRNA (encoding a myofilament protein), produced a strong twitching phenotype diagnostic for unc‐22 loss of function. dsRNA injection experiments against various mRNAs showed that dsRNA‐mediated gene silencing was a general cellular response and provided a versatile tool to “knockdown”C. elegans genes and they termed this phenomenon “RNA interference”—RNAi (7). The levels of targeted mRNAs in these studies were dramatically reduced suggesting that the dsRNA had destabilized its target mRNA (7). These experiments gave the valuable clue that dsRNA lies at the heart of RNA‐mediated silencing phenomena; for their seminal discovery Drs. Fire and Mello received the 2006 Nobel Prize in Physiology or Medicine. Another central piece of the RNA‐mediated gene silencing puzzle was put in place by the Baulcombe lab with the discovery of a species of small ∼22–25 nt RNAs later named small interfering RNAs (siRNAs) by the Tuschl lab. siRNAs were found in plants undergoing PTGS, and were of both sense and antisense polarity (10). Collectively, these observations suggested that the numerous manifestations of RNA‐mediated gene silencing observed in diverse organisms might have a common mechanism: dsRNA (endogenous or experimentally introduced) gave rise to ∼22 nt RNAs that targeted and silenced homologous mRNA sequences.

DEFECTIVE NEMATODES, RNAi AND RNP ASSEMBLYOSOMES

Genetic characterization of the heterochronic gene pathway (ie, the temporal progression of developmental events in C. elegans), by the Ambros lab uncovered a ∼22 nt non‐coding RNA as the product of the lin‐4 gene (19). lin‐4 was proposed to be derived from a ∼65 nt hairpin‐forming RNA precursor. Mature lin‐4 RNA repressed the protein levels of lin‐14, another gene of the heterochronic pathway. Furthermore, it was shown that lin‐14 mRNA carried imperfect sites of complementarity in its 3′‐untranslated region (3′‐UTR) for the lin‐4 RNA and in the presence of lin‐4 the levels of lin‐14 protein were reduced significantly (19, 31). For several years, lin‐4 was the only example of an endogenous ∼22 nt regulatory RNA. But this changed when the Ruvkun lab discovered let‐7, a second ∼22 nt RNA that also functioned in the heterochronic gene pathway (24). Like lin‐4, let‐7 RNA derived from a larger (∼75 nt) precursor that had the capacity to form a single stem‐loop structure. let‐7 RNA recognized sequences present in the 3′‐UTR of the lin‐41 mRNA and repressed its protein levels; lin‐4 and let‐7 RNAs were named small temporal RNAs (stRNAs) (24). An important observation was that let‐7 was evolutionary conserved (22), suggesting that other ∼22 nt regulatory RNAs might exist in many organisms and might be derived, like stRNAs from dsRNA‐hairpin precursors.

Meanwhile, biochemical dissection of RNAi by the Hannon, Tuschl, Zamore, Bartel and Sharp labs provided concrete evidence that ∼22 nt siRNAs were indeed the critical determinants of RNAi (4, 11, 27, 32). Furthermore, the isolation of Dicer, an RNAse III‐type nuclease as the enzyme that processed dsRNA into siRNAs (1) and of an siRNA‐containing multisubunit ribonuclease termed RNA‐induced Silencing Complex (RISC) (11) by the Hannon lab, solidified the mechanistic model of target mRNA destruction in RNAi. The connection between siRNAs and stRNAs was strengthened with the demonstration by the Zamore lab that Dicer also processed the let‐7 stRNA from its longer stem‐loop precursor (12). The central role of Dicer, in RNAi and in stRNA processing was almost simultaneously reported by the Plasterk, Hannon, Mello and Bass labs and was supported both genetically and biochemically (8, 13, 14). In addition to Dicer, the Argonaute (Ago) family of proteins were emerging as important factors in RNAi and in the function of stRNAs. Three Ago proteins, RDE‐1 in C. elegans, QDE‐2 in Neurospora and AGO1 in Arabidopsis were identified in genetic screens as essential for RNAi‐type silencing by the Mello, Cogoni and Vaucheret labs, respectively (2, 6, 26). And alg‐1 and alg‐2, two C. elegans Ago proteins, were found by the Mello lab to be required for the stability and function of stRNAs (8). These studies showed conclusively that siRNAs and stRNAs were intimately related. The two stRNAs were the first glimpse of a very large family whose members awaited their day under the sun, unnoticed at the bottom of RNA gels.

The Tuschl, Bartel and Ambros labs and soon thereafter the Dreyfuss lab, reported the existence in different organisms of >100 endogenous ∼22 nt RNAs. These RNAs were derived from ∼65 nt hairpin‐like precursors and were named miRNAs (16, 17, 18, 20). A clever directional cloning method, developed by the Tuschl lab to study the sequences of siRNAs from in vitro RNAi reactions prepared from Drosophila embryo extracts (4), led to three seminal discoveries. The first was the discovery that siRNAs acted as guides to direct endonucleolytic cleavage of their target RNAs across from the center of the guide siRNA (4). The second was the precise characterization of siRNA duplexes. This finding led to their utilization to knockdown genes experimentally (5). The third was the fortuitous cloning of miRNAs from Drosophila (and later their purposeful cloning from human cells) (5, 16). The Bartel lab similarly cloned C. elegans miRNAs from total RNA (17), while the Ambros lab identified C. elegans miRNAs by combining a bioinformatics and cloning approach from cDNA libraries prepared from small RNAs (18). The Dreyfuss lab discovered human miRNAs by studying the biochemistry of the Survival of Motor Neurons (SMN) complex, an assemblyosome of diverse Ribonucleoproteins (RNPs) most notably the RNPs that are required for pre‐mRNA splicing, termed small nuclear RNPs (snRNPs) (15, 23). Loss of function of SMN leads to Spinal Muscular Atrophy (23). The SMN complex is composed of eight proteins: the SMN protein and seven Gemins (Gemins 2–8) (15). One of the Gemins (Gemin 3) has protein domains typical of RNA helicases. As the SMN complex plays a major role in the assembly of spliceosomal snRNPs and as Gemin 3, a putative RNA helicase, is found in the SMN complex, it was thought that U snRNAs (the RNAs that are found in snRNPs) would associate with Gemin 3. To look for these RNAs, the Gemin 3 protein was immunoprecipitated and its associated RNAs were isolated, 3′‐end labeled, to visualize them by autoradiography, and fractionated on a gel to separate them by size. However, instead of U snRNAs (whose sizes are larger than 100 nt), a prominent band at ∼22 nt was evident. These small RNAs from the Gemin 3 immunoprecipitates were cloned using the directional cloning method described above and were found to be miRNAs (20). Further biochemical experiments showed that Gemin 3 and Gemin 4 were also found together outside of the SMN complex. Immunopurification of this separate complex and biochemical characterization of its protein and RNA composition showed that Ago2 (previously known as eIF2C2—eukaryotic translation Initiation factor 2C2, a human Argonaute––Ago‐protein) was also part of the Gemin3/Gemin4 complex and brought with it miRNAs (20). Furthermore, it was shown that miRNAs associated with Ago protein. The Ago/miRNA complex and its associated proteins was termed the microRNP (20). We now know that Ago proteins are the key mediators of miRNA function and isolation and cloning of small RNAs that associate with Argonaute proteins is a very important discovery tool for new small regulatory RNAs.

There has been tremendous progress, and in a short period of time, since these early reports. The authors of this symposium review the biology of miRNAs and their impact in neurodegenerative diseases and in brain tumors.