by Lucy Chen
Introduction
Humans are highly complex, multicellular organisms with millions of cells specialised for particular functions via distinct morphologies and chemistry. The process by which various cell types are generated from one universal type during development is known as cellular differentiation, and is driven by changes in gene expression in response to environmental changes. Different patterns of genes generate different protein products, which then undertake different functions; a cell generally expresses only a fraction of the genes it contains.[1] However, if a neuron and a liver cell have the same underlying genome (or total genetic information) stored in their nuclei, how does the cell ensure these changes in gene expression occur at the correct time and location during development? The answer lies in the central dogma; DNA is converted to RNA which is converted to protein in a multi-step process and involves many points at which the steps can be manipulated to alter the final expression of the gene (Figure 1).[2]
There are many methods of doing this: controlling the point in time and frequency at which a particular gene is transcribed, regulating the splicing of the resulting RNA transcripts, selectively exporting chosen mRNAs out of the nucleus to predetermined locations in the rest of the cell, or altering the stability of both mRNA and proteins. In this article, we will explore the underlying molecular mechanisms of the latter, and discuss the therapeutic potential of these regulatory RNAs towards cancer and drug delivery. In this article, one specific mechanism of translational control will be discussed; namely, that of RNA interference (RNAi).[2] RNAi evolved as a mechanism of protection against RNA viruses by cleaving the genome into fragments to use as a honing mechanism against translation of viral proteins.
Figure 1: Diagram showing various points of control during the eukaryotic transcription and translation process.[2]
What is RNAi?
RNA interference involves changes in the secondary structure of protein-RNA interactions and is used for large-scale screening of random genes.[3] As RNA is a single-stranded nucleic acid, it is not subject to the same rotational restrictions as DNA, and thus can adopt many different conformations, much like a shift in protein structure results in modulation of its function. Thus, RNAi exploits the steric nature of translation to control which genes are expressed in a cell. This unique specificity is ensured through base pairing the RNAi molecule to the product of the transcribed gene, allowing the RNAi molecule to identify its complementary target with accuracy.[4] This also leads to different fates for the targeted mRNA, disrupting its stability and affecting the ability of ribosomes to bind correctly.
In eukaryotes, RNA interference is conducted by specific classes of non-coding RNA (ncRNA). Non-coding RNA molecules have varied functions, and are characterised by the fact that they are transcribed, but not translated.[5] There are many different types of ncRNA specialised for various cellular processes, including transfer RNAs (tRNAs) and ribosomal RNAs (rRNAs) to drive translation. Small ncRNAs such as microRNAs (miRNA) and small interfering RNAs (siRNAs) are characterised by their comparatively shorter nucleotide length and are responsible for gene regulation.[5] Piwi-interacting RNAs (piRNA) constitute the largest class of small ncRNA and are characterised by their function in the epigenetic silencing of transcripts that contain a high frequency of repeat sequences, most notably transposable elements, and their regulatory role in haploid germline cells.[6]
Figure 2: Schematic of general differences in pathways of biogenesis and action for small non-coding RNA classes.[7]
What are miRNAs and siRNAs?
MicroRNAs are ncRNAs of approximately 22 nucleotides in length and form a secondary stem-loop structure.[2,7] They function extensively throughout the cell; the human genome contains approximately 1500 genes to encode regulatory miRNAs, and are involved in nearly every cellular process.[1] These regulatory molecules are crucial in ensuring normal cell differentiation and homeostasis; mutations in their RNA genes that result in loss of function have been linked to neurodegenerative diseases, autoimmune conditions, and certain types of cancer.[8,9] miRNAs are derived from the introns of coding genes, large ncRNAs, or pseudogenes, which are genes that have such heavily mutated sequences that they are essentially no longer functional.[1] The transcription of these RNA genes is a temporal and tissue type-dependent process.
Small interfering RNAs (siRNAs) are distinguished from miRNAs in that they are double-stranded molecules, roughly 20-25 nucleotides in length. Unlike miRNAs, siRNAs are produced from long dsRNA, or small hairpin RNAs, generating a duplex structure that has a 2nt overhang on the 3’ end of both strands.[10] They may also occur during viral infection, having been introduced into the cell via transfection; like piRNAs, they regulate transposons, which could be the basis of their variable evolution.[1,2] Whereas miRNAs have generally conserved mechanisms of action, siRNAs originate from different locations and exert effects in various ways.
Biogenesis of miRNAs
miRNA is produced through many different pathways; this article will discuss the canonical, or most prevalent, mechanism of biogenesis.[11] Primary microRNA (pri-miRNA) is transcribed from its corresponding gene; its sequence is highly complementary, with 95% pairing homology, which facilitates the formation of a stable stem-loop structure.[1,11] Although there is not complete complementarity, there are a few common patterns between miRNA sequences: the nucleotide at position 1 usually base-pairs with an mRNA molecule with adenosine in the corresponding position, as will be further discussed later. The seed region close to the 5’ end from positions 2-8 is characterised with perfect base pairing, and is critical to correct mRNA binding.[1,4] Nucleotides from positions 13-16 usually have high complementarity as well, but not to the same extent as the seed sequence.
Figure 3: Conserved regions of the miRNA:mRNA binding interaction.[4]
This is then processed into complete miRNA via two reactions, the first using the microprocessor complex, which is made up of two proteins: DGCR8, an RNA-binding protein, and Drosha, a ribonuclease present in the nucleus.[12] DGCR8 recognises and binds a consensus sequence GGAC, then orients Drosha to cleave at the correct position at the base of the hairpin. The result is a hairpin-shaped 70nt fragment with a phosphorylated 5’ end and a 3’ two nucleotide-length overhang; this is known as pre-miRNA.[2,10] The pre-miRNA is transported out of the nucleus via exportin 5, which recognises the overhang.[13] In the cytoplasm, the hairpin is cleaved by Dicer, another RNase, which cuts the loop connecting the 3’/5’ arms, forming an RNA duplex with a two nucleotide-length 3’ overhang, where one end of the strand extends beyond the other.[14]
Figure 4: Processing of pre-miRNA to mature miRNA from the nucleus to the cytoplasm.
To ensure the duplex is stable and not targeted for degradation by cytoplasmic enzymes, the 3’ hydroxyls undergo O-methylation where a CH3 group is added to the ends of the molecule.[1] Further modifications include ADAR adenosine deaminase-mediated base editing, which involves the conversion of adenosine to inosine, leading to a A-I base pairing instead of the correct A-U. This destabilises the duplex and may change its target specificity.[15]
Figure 5: Regulation of RNA processing via recognition of the stem-loop structure. Proteins directly bind to the loop and induce negative regulation, leading to degradation. MCPIP1 and Lin28 are both negative regulators; MCPIP1 cleaves the loop whereas Lin28 recruits a uridylyl transferase to tag the loop for degradation via attachment of a poly(U) tail. KSRP is a positive regulator.[16]
Both strands of the miRNA fragment are able to act functionally, but only one strand is loaded onto the RNA-induced silencing complex (RISC). This is facilitated by the N-terminal helicase domain of Dicer, which unwinds the double-stranded RNA molecule into single strands. The strand loaded into the RISC complex is known as the guide strand, and is preferentially determined by its lower thermodynamic stability, whereas the other passenger strand is degraded.[17] This process is performed by the RISC-loading complex (RLC), which is the core structure of RISC and has three components: Dicer, as previously mentioned; transactivating response RNA-binding protein TRBP; and Argonaute 2 (Ago2), which has catalytic RNase activity for removal of the passenger strand.[18]
Figure 6: Schematic representation of the components of the RISC-loading complex.[20]
Argonaute 2 belongs to the Argonaute protein family, which all bind miRNA, siRNA and piRNA. The role of the small ncRNAs is to guide Argonaute proteins to their targets; the protein will then cleave the mRNA or inhibit its translation. All members of the Argonaute protein family are characterised by four domains: N-terminal, PAZ, Mid, and PIWI. The PAZ domain is conserved and recognises the 3’ end of siRNA and miRNA, and is responsible for the method of regulation chosen after the ncRNA has bound to its mRNA target.[19,20] Mid contains a MC motif, which is involved in binding 5’ caps and mediates translational control via a mechanism that will be discussed in more detail later. PIWI has RNase activity and functions to degrade the mRNA target.
Figure 7: Ribbon representation of secondary structure of domains of RLC complex.[20]
Once the RISC is guided to the target mRNA via its bound miRNA strand, the miRNA interacts with complementary regions of the RNA strands, known as miRNA response elements (MREs) which are primarily located in the 3’ untranslated region (UTR).[21] The degree of complementarity determines the downstream mechanism of regulation; there are two main mechanisms for control of expression: first, direct mRNA degradation; and second, steric inhibition of translation.[1,2] In animals, miRNA is primarily used for translation inhibition, although both pathways are possible, whereas siRNA is used for degradation specifically. The reason for this pathway distinction is due to the degree of complementarity in the mRNA:miRNA pairing; the higher the level of base pairing, the more likely the mRNA is to be degraded.[22] In general, the miRNA-mRNA pairing is not fully complementary, such that one miRNA may have multiple targets, and one mRNA can be bound by many different miRNA molecules.[2] This flexibility generates a network where the expression of a gene can be altered to a highly sensitive degree.
The RISC complex translocates down the target mRNA in a sliding manner from 5’ to 3’, reading the sequence as it moves. Recognition of an initial 2-4 nucleotide region of homology upstream of the characteristic seed region allows for full pairing of the rest of the miRNA molecule to the mRNA.[11]
If imperfect base pairing occurs, RISC then interferes with translation by inhibiting initiation, disrupting the function of the eukaryotic initiation factor 4A (eIF4A), which is a member of the translational machinery. The mechanism by which this is done is still not well understood, but the result is eIF4A dissociation from the mRNA, which thus prevents ribosome scanning and initiation complex assembly.[23] In actuality, there are approximately nine proposed mechanisms of action through which miRNAs may exert their effects, but the supporting evidence for some is limited, with contradictory conclusions. Many models have been postulated towards unifying all these mechanisms together; one particular study derived a mathematical model through which all possible modes of action were able to operate simultaneously in the cell.[24] The dominant mechanism chosen when regulating a specific pathway thus depends on a set of intrinsic parameters (concentration of translated protein, mRNA, and number of ribosomes bound to one mRNA after production of corresponding miRNA) that affect the relative kinetic rates of translation and degradation.[24] A few of these mechanisms are described below, as shown in figure 8. In all mechanisms, the Argonaute component of the RISC recruits a glycine-tryptophan protein (GW182) family member to enforce translational repression.[25]
Inhibition of elongation: Occurs after initiation machinery has been assembled; Ago proteins stimulate ribosomal dropoff, although the mechanism of this is unknown.
Co-translational protein degradation: the nascent, newly-made polypeptide is degraded and translated simultaneously.
Competition for the 5’ cap: elongation initiation factor 4E (eIF4E) is outcompeted by the Ago protein, which binds to the methylated cap and prevents recruitment of further initiation complexes.
Inhibition of ribosomal assembly: Ago proteins recruit elongation initiation factor 6 (eIF6) to prevent the association of the small and large ribosomal subunits.
Inhibition of mRNA circularisation and deadenylation: The mechanism is not well understood,but may involve deadenylation via GW182 interaction with polyadenylate-binding protein (PABPC), which then recruits the 3’-poly(A) tail-nuclease deadenylation complex subunit 2 (PAN2)-PAN3 and other associated complexes. Deadenylation facilitates the enzymatic decapping of the 5’ end via the DCP1-DCP2 complex. Removal of the cap leads to instability, increasing the susceptibility to degradation.[26]
Figure 8: Methods of translational repression by the miRNA:RISC complex.[26]
Biogenesis of siRNAs
Unlike miRNAs, siRNAs target the genes they were derived from, thus participating in autoregulation. They are processed using the same proteins as miRNAs, but result in different molecules. For example, Dicer cleaves the double-stranded RNA hairpin in two places to generate fragments with lengths of approximately 20 nucleotides; in contrast to the biogenesis of miRNAs, this step occurs in the cytosol instead of the nucleus. Overhangs of two nucleotides are generated on both ends of the 20-nt fragment.[27]
Figure 9: Diagram of siRNA biogenesis.[28]
The process of unwinding, strand selection and loading onto the RISC complex is the same for siRNA as it is for miRNA. However, the major difference in regulation is based on the complementarity of mRNA binding; siRNA has complete complementarity to its target, which then induces degradation.[1,2] Initially, the mRNA is cleaved into fragments between the nucleotides at position 10 and 11, catalysed by the PIWI domain of RISC. The remaining fragments are then degraded further by exonucleases such as XRN1.[29] Thus, the mRNA is no longer available to be translated into protein, and the gene is silenced.
Therapeutic applications
As these RNA molecules have specific roles in gene regulation, they can act as highly sensitive representatives of changes in internal cellular conditions as the environment is heavily influenced by expression levels. Thus, many RNAi molecules are associated with disease pathologies, and as many siRNAs can be specifically engineered to target any gene, they may also be used as therapeutic agents as well as biomarkers.[30] Studies into miR-506 have proven its effectiveness as a tumour antagonist.[31] There may also be a possible role for miR-129 in synaptic plasticity towards the development of Alzheimer’s disease.[9] RNAi is a diverse and constantly-changing field of research, and has many promising implications for the future of disease therapy and the healthcare industry.
Glossary:
DGCR8: DiGeorge syndrome critical region 8, a protein within the microprocessor complex.
Domain: A conserved region of a protein that folds independently of the rest of the polypeptide with specific function and tertiary structure.
MC motif: a structure homologous to the cap-binding motif seen in eukaryotic initiation factor 4E.
N-terminal helicase: a domain at the amino terminus of the polypeptide that has helicase catalytic activity, enabling the unwinding of nucleic acid strands.
Ribonuclease: an enzyme that catalyses the degradation of RNA into smaller fragments
References:
[1] Krebs, J. E., Elliott S. Goldstein, Stephen T. Kilpatrick. Lewin’s Genes XII. 12th ed. Massachusetts: Jones & Bartlett Learning, 2018.
[2] Alberts, B., Alexander Johnson, Julian Lewis, David Morgan, Martin Raff, Keith Roberts, and Peter Walter. Molecular Biology of the Cell. 6th ed. New York: Garland Science, 2015.
[3] Wahid, F. Adeeb Shehzad, Taous Khan,You Young Kim. “MicroRNAs: synthesis, mechanism, and recent clinical trials.” Biochimica et Biophysica Acta – Molecular Cell Research. 1803, 11 (Nov 2010): 1231-1243. doi:10.1016/j.bbamcr.2010.06.013
[4] Bartel, D. P. “MicroRNA Target Recognition and Regulatory Functions.” Cell. 136, 2 (Jan 2009): 215-233. doi:10.1016/j.cell.2009.01.002
[5] Palazzo, A. F. and Eliza S. Lee. “Non-coding RNA: what is functional and what is junk?” Frontiers in Genetics. 6, 2 (Jan 2015). doi:10.3389/fgene.2015.00002
[6] Seto, A. G. Robert E. Kingston, Nelson C. Lau. “The coming of age for Piwi proteins.” Molecular Cell. 26, 5 (Jun 2007). doi:10.1016/j.molcel.2007.05.021
[7] Jeang, K. “RNAi in the regulation of mammalian viral infections.” Biomedical Central Biology. 10, 58 (Jun 2012). doi:10.1186/1741-7007-10-58
[8] Akcakaya, P., Susanne Ekelund, Iryna Kolosenko, Stefano Caramuta, Deniz M. Ozata, Hong Xie, Ulrik Lindforss, Hans Olivecrona, and Weng-Onn Lui. “miR-185 and miR-133b deregulation is associated with overall survival and metastasis in colorectal cancer.” International Journal of Oncology. 39, 2 (Aug 2011). doi:10.3892/ijo.2011.1043
[9] Hosseinian, S. Ehsan Arefian, Sayed-Amir Marashi. “A meta-analysis of gene expression data highlights synaptic dysfunction in the hippocampus of brains with Alzheimer’s disease.” Nature Scientific Reports. 10, 8384 (May 2020). doi:10.1038/s41598-020-64452-z
[10] Lagana, A., Dario Veneziano, Francesco Russo, Alfredo Pulvirenti, Rosalba Giugno, Carlo Maria Croce, Alfredo Ferro. “Computational design of artificial RNA molecules for gene regulation.” RNA Bioinformatics Methods in Molecular Biology. 1269 (Oct 2014): 393-412. doi:10.1007/978-1-4939-2291-8_25
[11] Gebert, L. F. R. and Ian J. MacRae. “Regulation of microRNA function in animals.” Nature Reviews Molecular Cell Biology. 20 (Jan 2019): 21-37. doi:10.1038/s41580-018-0045-7
[12] Nguyen, T. A, Myung Hyun Jo, Yeon-Gil Choi, Joha Park, S. C. Kwon, Sungchul Hohng, V. N. Kim, and Jae-Sung Woo. “Functional Anatomy of the Human Microprocessor.” Cell. 161, 6 (Jun 2015): 1374-1387. doi:10.1016/j.cell.2015.05.010
[13] Okada, C. Eiki Yamashita, Soo J. Lee, Satoshi Shibata, Jun Katahira, Atsushi Nakagawa, Yoshihiro Yoneda, Tomitake Tsukihara. “A high-resolution structure of the pre-microRNA nuclear export machinery.” Science. 326, 5957 (Nov 2009): 1275-1279. doi:10.1126/science.1178705
[14] Tsutsumi A., Tomoko Kawamata, and Yukihide Tomari. “Recognition of the pre-miRNA structure by Drosophila Dicer-1.” Nature Structural and Molecular Biology. 18 (Sep 2011): 1153-1158. doi:10.1038/nsmb.2125
[15] Nishikura K. “A-to-I editing of coding and non-coding RNAs by ADARs.” Nature Reviews Molecular Cell Biology. 17 (Feb 2016): 83-96. doi:10.1038/nrm.2015.4
[16] Bajan, S. and Gyorgy Hutvagner. “Another ‘loophole’ in miRNA processing.” Molecular Cell. 44, 3 (Nov 2011): 345-347. doi:10.1016/j.molcel.2011.10.010
[17] MacRae, I. J. Enbo Ma, Min Zhou, Carol V. Robinson, and Jennifer A. Doudna. “In vitro reconstitution of the human RISC-loading complex.” Proceedings of the National Academy of Sciences of the United States of America. 105, 2 (Jan 2015): 512-517. doi:10.1073/pnas.0710869105
[18] Pratt, A. J. and Ian J. MacRae. “The RNA-induced silencing complex: a versatile gene-silencing machine.” Journal of Biological Chemistry. 284, 27 (Jul 2009): 17897-17901. doi:10.1074/jbc.R900012200
[19] Agarwal V., George W. Bell, Jin-Wu Nam, and David P. Bartel. “Predicting effective microRNA target sites in mammalian RNAs.” eLife. (Aug 2015). doi:10.7554/eLife.05005
[20] Sheu-Gruttadauria, J. and Ian J. MacRae. “Structural foundations of RNA silencing by Argonaute.” Journal of Molecular Biology. 429, 17 (Aug 2017). doi:10.1016/j.jmb.2017.07.018
[21] O’Brien, J., Heyam Hayder, Yara Zayed and Chun Peng. “Overview of microRNA biogenesis, mechanisms of actions, and circulation.” Frontiers in Endocrinology. 9, 402 (Aug 2018). doi:10.3389/fendo.2018.00402
[22] Lim, L. P., Nelson C. Lau, Philip Garrett-Engele, Andrew Grimson, Janell M. Schelter, John Castle, David P. Bartel, Peter S. Linsley, Jason M. Johnson. “Microarray analysis shows that some microRNAs downregulate large numbers of target mRNAs.” Nature. 433, 7027 (Feb 2005). doi:10.1038/nature03315
[23] Mathonnet, G., Marc R. Fabian, Yuri V. Svitkin, Armen Parsyan, Laurent Huck, Takayuki Murata, Stefano Biffo, William C. Merrick, Edward Darzynkiewicz, Ramesh S. Pillai, Witold Filipowicz, Thomas F. Duchaine, Nahum Sonenberg. “MicroRNA inhibition of translation initiation in vitro by targeting the cap-binding complex eIF4F.” Science. 317, 5845 (Sep 2007). doi:10.1126/science.1146067
[24] Morozova, N., Andrei Zinovyev, Nora Nonne, Linda-Louise Pritchard, Alexander N. Gorban, Annick Harel-Bellan. “Kinetic signatures of microRNA modes of action.” RNA. 18, 9 (Sep 2012). doi:10.1261/rna.032284.112
[25] Braun, J. E., Eric Huntzinger, Elisa Izaurralde. “The role of GW182 proteins in miRNA-mediated gene silencing.” Advances in experimental medicine and biology. 768 (Nov 2012): 147-163. doi:10.1007/978-1-4614-5107-5_9
[26] Eulalio, A., Eric Huntzinger, Elisa Izaurralde. “Getting to the root of miRNA-mediated gene silencing.” Cell. 132, 1 (Jan 2008): 14. doi:10.1016/j.cell.2007.12.024
[27] Lee, Y. S., Kenji Nakahara, John W. Pham, Kevin Kim, Zhenying He, Erik J. Sontheimer, Richard W. Carthew. “Distinct roles for Drosophila Dicer-1 and Dicer-2 in the siRNA/miRNA silencing pathways.” Cell. 117, 1 (Apr 2004): 69-81. doi:10.1016/S0092-8674(04)00261-2
[28] Mendonca, L., Maria Pedroso de Lima, S. Simoes. “Targeted lipid-based systems for siRNA delivery.” Journal of Drug Delivery Science and Technology. 22, 1 (Dec 2012): 65-73. doi:10.1016/S1773-2247(12)50006-7
[29] Orban, T. I. and Elisa Izaurralde. “Decay of mRNAs targeted by RISC requires XRN1, the Ski complex, and the exosome.” RNA. 11,4 (Apr 2005): 459-469. doi:10.1261/rna.7231505
[30] Bobbin, M. L., and John J. Rossi. “RNA interference (RNAi)-based therapeutics: delivering on the promise?” Annual Review of Pharmacology and Toxicology. 56 (Jan 2016): 103-122. doi:10.1146/annurev-pharmtox-010715-103633
[31] Liu, G., Yan Sun, Ping Ji, Xia Li, David Cogdell, Da Yang, Britany C. P. Kerrigan, Ilya Shmulevich, Kexin Chen, Anil K. Sood, Fengxia Xue, Wei Zhang. “MiR-506 suppresses proliferation and induces senescence by directly targeting the CDK4/6-FOXM1 axis in ovarian cancer.” Journal of Pathology. 233, 3 (Mar 2014). doi:10.1002/path.4348
RNA interference: the molecules that govern genetic control
Introduction
Humans are highly complex, multicellular organisms with millions of cells specialised for particular functions via distinct morphologies and chemistry. The process by which various cell types are generated from one universal type during development is known as cellular differentiation, and is driven by changes in gene expression in response to environmental changes. Different patterns of genes generate different protein products, which then undertake different functions; a cell generally expresses only a fraction of the genes it contains.[1] However, if a neuron and a liver cell have the same underlying genome (or total genetic information) stored in their nuclei, how does the cell ensure these changes in gene expression occur at the correct time and location during development? The answer lies in the central dogma; DNA is converted to RNA which is converted to protein in a multi-step process and involves many points at which the steps can be manipulated to alter the final expression of the gene (Figure 1).[2]
There are many methods of doing this: controlling the point in time and frequency at which a particular gene is transcribed, regulating the splicing of the resulting RNA transcripts, selectively exporting chosen mRNAs out of the nucleus to predetermined locations in the rest of the cell, or altering the stability of both mRNA and proteins. In this article, we will explore the underlying molecular mechanisms of the latter, and discuss the therapeutic potential of these regulatory RNAs towards cancer and drug delivery. In this article, one specific mechanism of translational control will be discussed; namely, that of RNA interference (RNAi).[2] RNAi evolved as a mechanism of protection against RNA viruses by cleaving the genome into fragments to use as a honing mechanism against translation of viral proteins.
Figure 1: Diagram showing various points of control during the eukaryotic transcription and translation process.[2]
What is RNAi?
RNA interference involves changes in the secondary structure of protein-RNA interactions and is used for large-scale screening of random genes.[3] As RNA is a single-stranded nucleic acid, it is not subject to the same rotational restrictions as DNA, and thus can adopt many different conformations, much like a shift in protein structure results in modulation of its function. Thus, RNAi exploits the steric nature of translation to control which genes are expressed in a cell. This unique specificity is ensured through base pairing the RNAi molecule to the product of the transcribed gene, allowing the RNAi molecule to identify its complementary target with accuracy.[4] This also leads to different fates for the targeted mRNA, disrupting its stability and affecting the ability of ribosomes to bind correctly.
In eukaryotes, RNA interference is conducted by specific classes of non-coding RNA (ncRNA). Non-coding RNA molecules have varied functions, and are characterised by the fact that they are transcribed, but not translated.[5] There are many different types of ncRNA specialised for various cellular processes, including transfer RNAs (tRNAs) and ribosomal RNAs (rRNAs) to drive translation. Small ncRNAs such as microRNAs (miRNA) and small interfering RNAs (siRNAs) are characterised by their comparatively shorter nucleotide length and are responsible for gene regulation.[5] Piwi-interacting RNAs (piRNA) constitute the largest class of small ncRNA and are characterised by their function in the epigenetic silencing of transcripts that contain a high frequency of repeat sequences, most notably transposable elements, and their regulatory role in haploid germline cells.[6]
Figure 2: Schematic of general differences in pathways of biogenesis and action for small non-coding RNA classes.[7]
What are miRNAs and siRNAs?
MicroRNAs are ncRNAs of approximately 22 nucleotides in length and form a secondary stem-loop structure.[2,7] They function extensively throughout the cell; the human genome contains approximately 1500 genes to encode regulatory miRNAs, and are involved in nearly every cellular process.[1] These regulatory molecules are crucial in ensuring normal cell differentiation and homeostasis; mutations in their RNA genes that result in loss of function have been linked to neurodegenerative diseases, autoimmune conditions, and certain types of cancer.[8,9] miRNAs are derived from the introns of coding genes, large ncRNAs, or pseudogenes, which are genes that have such heavily mutated sequences that they are essentially no longer functional.[1] The transcription of these RNA genes is a temporal and tissue type-dependent process.
Small interfering RNAs (siRNAs) are distinguished from miRNAs in that they are double-stranded molecules, roughly 20-25 nucleotides in length. Unlike miRNAs, siRNAs are produced from long dsRNA, or small hairpin RNAs, generating a duplex structure that has a 2nt overhang on the 3’ end of both strands.[10] They may also occur during viral infection, having been introduced into the cell via transfection; like piRNAs, they regulate transposons, which could be the basis of their variable evolution.[1,2] Whereas miRNAs have generally conserved mechanisms of action, siRNAs originate from different locations and exert effects in various ways.
Biogenesis of miRNAs
miRNA is produced through many different pathways; this article will discuss the canonical, or most prevalent, mechanism of biogenesis.[11] Primary microRNA (pri-miRNA) is transcribed from its corresponding gene; its sequence is highly complementary, with 95% pairing homology, which facilitates the formation of a stable stem-loop structure.[1,11] Although there is not complete complementarity, there are a few common patterns between miRNA sequences: the nucleotide at position 1 usually base-pairs with an mRNA molecule with adenosine in the corresponding position, as will be further discussed later. The seed region close to the 5’ end from positions 2-8 is characterised with perfect base pairing, and is critical to correct mRNA binding.[1,4] Nucleotides from positions 13-16 usually have high complementarity as well, but not to the same extent as the seed sequence.
Figure 3: Conserved regions of the miRNA:mRNA binding interaction.[4]
This is then processed into complete miRNA via two reactions, the first using the microprocessor complex, which is made up of two proteins: DGCR8, an RNA-binding protein, and Drosha, a ribonuclease present in the nucleus.[12] DGCR8 recognises and binds a consensus sequence GGAC, then orients Drosha to cleave at the correct position at the base of the hairpin. The result is a hairpin-shaped 70nt fragment with a phosphorylated 5’ end and a 3’ two nucleotide-length overhang; this is known as pre-miRNA.[2,10] The pre-miRNA is transported out of the nucleus via exportin 5, which recognises the overhang.[13] In the cytoplasm, the hairpin is cleaved by Dicer, another RNase, which cuts the loop connecting the 3’/5’ arms, forming an RNA duplex with a two nucleotide-length 3’ overhang, where one end of the strand extends beyond the other.[14]
Figure 4: Processing of pre-miRNA to mature miRNA from the nucleus to the cytoplasm.
To ensure the duplex is stable and not targeted for degradation by cytoplasmic enzymes, the 3’ hydroxyls undergo O-methylation where a CH3 group is added to the ends of the molecule.[1] Further modifications include ADAR adenosine deaminase-mediated base editing, which involves the conversion of adenosine to inosine, leading to a A-I base pairing instead of the correct A-U. This destabilises the duplex and may change its target specificity.[15]
Figure 5: Regulation of RNA processing via recognition of the stem-loop structure. Proteins directly bind to the loop and induce negative regulation, leading to degradation. MCPIP1 and Lin28 are both negative regulators; MCPIP1 cleaves the loop whereas Lin28 recruits a uridylyl transferase to tag the loop for degradation via attachment of a poly(U) tail. KSRP is a positive regulator.[16]
Both strands of the miRNA fragment are able to act functionally, but only one strand is loaded onto the RNA-induced silencing complex (RISC). This is facilitated by the N-terminal helicase domain of Dicer, which unwinds the double-stranded RNA molecule into single strands. The strand loaded into the RISC complex is known as the guide strand, and is preferentially determined by its lower thermodynamic stability, whereas the other passenger strand is degraded.[17] This process is performed by the RISC-loading complex (RLC), which is the core structure of RISC and has three components: Dicer, as previously mentioned; transactivating response RNA-binding protein TRBP; and Argonaute 2 (Ago2), which has catalytic RNase activity for removal of the passenger strand.[18]
Figure 6: Schematic representation of the components of the RISC-loading complex.[20]
Argonaute 2 belongs to the Argonaute protein family, which all bind miRNA, siRNA and piRNA. The role of the small ncRNAs is to guide Argonaute proteins to their targets; the protein will then cleave the mRNA or inhibit its translation. All members of the Argonaute protein family are characterised by four domains: N-terminal, PAZ, Mid, and PIWI. The PAZ domain is conserved and recognises the 3’ end of siRNA and miRNA, and is responsible for the method of regulation chosen after the ncRNA has bound to its mRNA target.[19,20] Mid contains a MC motif, which is involved in binding 5’ caps and mediates translational control via a mechanism that will be discussed in more detail later. PIWI has RNase activity and functions to degrade the mRNA target.
Figure 7: Ribbon representation of secondary structure of domains of RLC complex.[20]
Once the RISC is guided to the target mRNA via its bound miRNA strand, the miRNA interacts with complementary regions of the RNA strands, known as miRNA response elements (MREs) which are primarily located in the 3’ untranslated region (UTR).[21] The degree of complementarity determines the downstream mechanism of regulation; there are two main mechanisms for control of expression: first, direct mRNA degradation; and second, steric inhibition of translation.[1,2] In animals, miRNA is primarily used for translation inhibition, although both pathways are possible, whereas siRNA is used for degradation specifically. The reason for this pathway distinction is due to the degree of complementarity in the mRNA:miRNA pairing; the higher the level of base pairing, the more likely the mRNA is to be degraded.[22] In general, the miRNA-mRNA pairing is not fully complementary, such that one miRNA may have multiple targets, and one mRNA can be bound by many different miRNA molecules.[2] This flexibility generates a network where the expression of a gene can be altered to a highly sensitive degree.
The RISC complex translocates down the target mRNA in a sliding manner from 5’ to 3’, reading the sequence as it moves. Recognition of an initial 2-4 nucleotide region of homology upstream of the characteristic seed region allows for full pairing of the rest of the miRNA molecule to the mRNA.[11]
If imperfect base pairing occurs, RISC then interferes with translation by inhibiting initiation, disrupting the function of the eukaryotic initiation factor 4A (eIF4A), which is a member of the translational machinery. The mechanism by which this is done is still not well understood, but the result is eIF4A dissociation from the mRNA, which thus prevents ribosome scanning and initiation complex assembly.[23] In actuality, there are approximately nine proposed mechanisms of action through which miRNAs may exert their effects, but the supporting evidence for some is limited, with contradictory conclusions. Many models have been postulated towards unifying all these mechanisms together; one particular study derived a mathematical model through which all possible modes of action were able to operate simultaneously in the cell.[24] The dominant mechanism chosen when regulating a specific pathway thus depends on a set of intrinsic parameters (concentration of translated protein, mRNA, and number of ribosomes bound to one mRNA after production of corresponding miRNA) that affect the relative kinetic rates of translation and degradation.[24] A few of these mechanisms are described below, as shown in figure 8. In all mechanisms, the Argonaute component of the RISC recruits a glycine-tryptophan protein (GW182) family member to enforce translational repression.[25]
Inhibition of elongation: Occurs after initiation machinery has been assembled; Ago proteins stimulate ribosomal dropoff, although the mechanism of this is unknown.
Co-translational protein degradation: the nascent, newly-made polypeptide is degraded and translated simultaneously.
Competition for the 5’ cap: elongation initiation factor 4E (eIF4E) is outcompeted by the Ago protein, which binds to the methylated cap and prevents recruitment of further initiation complexes.
Inhibition of ribosomal assembly: Ago proteins recruit elongation initiation factor 6 (eIF6) to prevent the association of the small and large ribosomal subunits.
Inhibition of mRNA circularisation and deadenylation: The mechanism is not well understood,but may involve deadenylation via GW182 interaction with polyadenylate-binding protein (PABPC), which then recruits the 3’-poly(A) tail-nuclease deadenylation complex subunit 2 (PAN2)-PAN3 and other associated complexes. Deadenylation facilitates the enzymatic decapping of the 5’ end via the DCP1-DCP2 complex. Removal of the cap leads to instability, increasing the susceptibility to degradation.[26]
Figure 8: Methods of translational repression by the miRNA:RISC complex.[26]
Biogenesis of siRNAs
Unlike miRNAs, siRNAs target the genes they were derived from, thus participating in autoregulation. They are processed using the same proteins as miRNAs, but result in different molecules. For example, Dicer cleaves the double-stranded RNA hairpin in two places to generate fragments with lengths of approximately 20 nucleotides; in contrast to the biogenesis of miRNAs, this step occurs in the cytosol instead of the nucleus. Overhangs of two nucleotides are generated on both ends of the 20-nt fragment.[27]
Figure 9: Diagram of siRNA biogenesis.[28]
The process of unwinding, strand selection and loading onto the RISC complex is the same for siRNA as it is for miRNA. However, the major difference in regulation is based on the complementarity of mRNA binding; siRNA has complete complementarity to its target, which then induces degradation.[1,2] Initially, the mRNA is cleaved into fragments between the nucleotides at position 10 and 11, catalysed by the PIWI domain of RISC. The remaining fragments are then degraded further by exonucleases such as XRN1.[29] Thus, the mRNA is no longer available to be translated into protein, and the gene is silenced.
Therapeutic applications
As these RNA molecules have specific roles in gene regulation, they can act as highly sensitive representatives of changes in internal cellular conditions as the environment is heavily influenced by expression levels. Thus, many RNAi molecules are associated with disease pathologies, and as many siRNAs can be specifically engineered to target any gene, they may also be used as therapeutic agents as well as biomarkers.[30] Studies into miR-506 have proven its effectiveness as a tumour antagonist.[31] There may also be a possible role for miR-129 in synaptic plasticity towards the development of Alzheimer’s disease.[9] RNAi is a diverse and constantly-changing field of research, and has many promising implications for the future of disease therapy and the healthcare industry.
Glossary:
DGCR8: DiGeorge syndrome critical region 8, a protein within the microprocessor complex.
Domain: A conserved region of a protein that folds independently of the rest of the polypeptide with specific function and tertiary structure.
MC motif: a structure homologous to the cap-binding motif seen in eukaryotic initiation factor 4E.
N-terminal helicase: a domain at the amino terminus of the polypeptide that has helicase catalytic activity, enabling the unwinding of nucleic acid strands.
Ribonuclease: an enzyme that catalyses the degradation of RNA into smaller fragments
References:
[1] Krebs, J. E., Elliott S. Goldstein, Stephen T. Kilpatrick. Lewin’s Genes XII. 12th ed. Massachusetts: Jones & Bartlett Learning, 2018.
[2] Alberts, B., Alexander Johnson, Julian Lewis, David Morgan, Martin Raff, Keith Roberts, and Peter Walter. Molecular Biology of the Cell. 6th ed. New York: Garland Science, 2015.
[3] Wahid, F. Adeeb Shehzad, Taous Khan,You Young Kim. “MicroRNAs: synthesis, mechanism, and recent clinical trials.” Biochimica et Biophysica Acta – Molecular Cell Research. 1803, 11 (Nov 2010): 1231-1243. doi:10.1016/j.bbamcr.2010.06.013
[4] Bartel, D. P. “MicroRNA Target Recognition and Regulatory Functions.” Cell. 136, 2 (Jan 2009): 215-233. doi:10.1016/j.cell.2009.01.002
[5] Palazzo, A. F. and Eliza S. Lee. “Non-coding RNA: what is functional and what is junk?” Frontiers in Genetics. 6, 2 (Jan 2015). doi:10.3389/fgene.2015.00002
[6] Seto, A. G. Robert E. Kingston, Nelson C. Lau. “The coming of age for Piwi proteins.” Molecular Cell. 26, 5 (Jun 2007). doi:10.1016/j.molcel.2007.05.021
[7] Jeang, K. “RNAi in the regulation of mammalian viral infections.” Biomedical Central Biology. 10, 58 (Jun 2012). doi:10.1186/1741-7007-10-58
[8] Akcakaya, P., Susanne Ekelund, Iryna Kolosenko, Stefano Caramuta, Deniz M. Ozata, Hong Xie, Ulrik Lindforss, Hans Olivecrona, and Weng-Onn Lui. “miR-185 and miR-133b deregulation is associated with overall survival and metastasis in colorectal cancer.” International Journal of Oncology. 39, 2 (Aug 2011). doi:10.3892/ijo.2011.1043
[9] Hosseinian, S. Ehsan Arefian, Sayed-Amir Marashi. “A meta-analysis of gene expression data highlights synaptic dysfunction in the hippocampus of brains with Alzheimer’s disease.” Nature Scientific Reports. 10, 8384 (May 2020). doi:10.1038/s41598-020-64452-z
[10] Lagana, A., Dario Veneziano, Francesco Russo, Alfredo Pulvirenti, Rosalba Giugno, Carlo Maria Croce, Alfredo Ferro. “Computational design of artificial RNA molecules for gene regulation.” RNA Bioinformatics Methods in Molecular Biology. 1269 (Oct 2014): 393-412. doi:10.1007/978-1-4939-2291-8_25
[11] Gebert, L. F. R. and Ian J. MacRae. “Regulation of microRNA function in animals.” Nature Reviews Molecular Cell Biology. 20 (Jan 2019): 21-37. doi:10.1038/s41580-018-0045-7
[12] Nguyen, T. A, Myung Hyun Jo, Yeon-Gil Choi, Joha Park, S. C. Kwon, Sungchul Hohng, V. N. Kim, and Jae-Sung Woo. “Functional Anatomy of the Human Microprocessor.” Cell. 161, 6 (Jun 2015): 1374-1387. doi:10.1016/j.cell.2015.05.010
[13] Okada, C. Eiki Yamashita, Soo J. Lee, Satoshi Shibata, Jun Katahira, Atsushi Nakagawa, Yoshihiro Yoneda, Tomitake Tsukihara. “A high-resolution structure of the pre-microRNA nuclear export machinery.” Science. 326, 5957 (Nov 2009): 1275-1279. doi:10.1126/science.1178705
[14] Tsutsumi A., Tomoko Kawamata, and Yukihide Tomari. “Recognition of the pre-miRNA structure by Drosophila Dicer-1.” Nature Structural and Molecular Biology. 18 (Sep 2011): 1153-1158. doi:10.1038/nsmb.2125
[15] Nishikura K. “A-to-I editing of coding and non-coding RNAs by ADARs.” Nature Reviews Molecular Cell Biology. 17 (Feb 2016): 83-96. doi:10.1038/nrm.2015.4
[16] Bajan, S. and Gyorgy Hutvagner. “Another ‘loophole’ in miRNA processing.” Molecular Cell. 44, 3 (Nov 2011): 345-347. doi:10.1016/j.molcel.2011.10.010
[17] MacRae, I. J. Enbo Ma, Min Zhou, Carol V. Robinson, and Jennifer A. Doudna. “In vitro reconstitution of the human RISC-loading complex.” Proceedings of the National Academy of Sciences of the United States of America. 105, 2 (Jan 2015): 512-517. doi:10.1073/pnas.0710869105
[18] Pratt, A. J. and Ian J. MacRae. “The RNA-induced silencing complex: a versatile gene-silencing machine.” Journal of Biological Chemistry. 284, 27 (Jul 2009): 17897-17901. doi:10.1074/jbc.R900012200
[19] Agarwal V., George W. Bell, Jin-Wu Nam, and David P. Bartel. “Predicting effective microRNA target sites in mammalian RNAs.” eLife. (Aug 2015). doi:10.7554/eLife.05005
[20] Sheu-Gruttadauria, J. and Ian J. MacRae. “Structural foundations of RNA silencing by Argonaute.” Journal of Molecular Biology. 429, 17 (Aug 2017). doi:10.1016/j.jmb.2017.07.018
[21] O’Brien, J., Heyam Hayder, Yara Zayed and Chun Peng. “Overview of microRNA biogenesis, mechanisms of actions, and circulation.” Frontiers in Endocrinology. 9, 402 (Aug 2018). doi:10.3389/fendo.2018.00402
[22] Lim, L. P., Nelson C. Lau, Philip Garrett-Engele, Andrew Grimson, Janell M. Schelter, John Castle, David P. Bartel, Peter S. Linsley, Jason M. Johnson. “Microarray analysis shows that some microRNAs downregulate large numbers of target mRNAs.” Nature. 433, 7027 (Feb 2005). doi:10.1038/nature03315
[23] Mathonnet, G., Marc R. Fabian, Yuri V. Svitkin, Armen Parsyan, Laurent Huck, Takayuki Murata, Stefano Biffo, William C. Merrick, Edward Darzynkiewicz, Ramesh S. Pillai, Witold Filipowicz, Thomas F. Duchaine, Nahum Sonenberg. “MicroRNA inhibition of translation initiation in vitro by targeting the cap-binding complex eIF4F.” Science. 317, 5845 (Sep 2007). doi:10.1126/science.1146067
[24] Morozova, N., Andrei Zinovyev, Nora Nonne, Linda-Louise Pritchard, Alexander N. Gorban, Annick Harel-Bellan. “Kinetic signatures of microRNA modes of action.” RNA. 18, 9 (Sep 2012). doi:10.1261/rna.032284.112
[25] Braun, J. E., Eric Huntzinger, Elisa Izaurralde. “The role of GW182 proteins in miRNA-mediated gene silencing.” Advances in experimental medicine and biology. 768 (Nov 2012): 147-163. doi:10.1007/978-1-4614-5107-5_9
[26] Eulalio, A., Eric Huntzinger, Elisa Izaurralde. “Getting to the root of miRNA-mediated gene silencing.” Cell. 132, 1 (Jan 2008): 14. doi:10.1016/j.cell.2007.12.024
[27] Lee, Y. S., Kenji Nakahara, John W. Pham, Kevin Kim, Zhenying He, Erik J. Sontheimer, Richard W. Carthew. “Distinct roles for Drosophila Dicer-1 and Dicer-2 in the siRNA/miRNA silencing pathways.” Cell. 117, 1 (Apr 2004): 69-81. doi:10.1016/S0092-8674(04)00261-2
[28] Mendonca, L., Maria Pedroso de Lima, S. Simoes. “Targeted lipid-based systems for siRNA delivery.” Journal of Drug Delivery Science and Technology. 22, 1 (Dec 2012): 65-73. doi:10.1016/S1773-2247(12)50006-7
[29] Orban, T. I. and Elisa Izaurralde. “Decay of mRNAs targeted by RISC requires XRN1, the Ski complex, and the exosome.” RNA. 11,4 (Apr 2005): 459-469. doi:10.1261/rna.7231505
[30] Bobbin, M. L., and John J. Rossi. “RNA interference (RNAi)-based therapeutics: delivering on the promise?” Annual Review of Pharmacology and Toxicology. 56 (Jan 2016): 103-122. doi:10.1146/annurev-pharmtox-010715-103633
[31] Liu, G., Yan Sun, Ping Ji, Xia Li, David Cogdell, Da Yang, Britany C. P. Kerrigan, Ilya Shmulevich, Kexin Chen, Anil K. Sood, Fengxia Xue, Wei Zhang. “MiR-506 suppresses proliferation and induces senescence by directly targeting the CDK4/6-FOXM1 axis in ovarian cancer.” Journal of Pathology. 233, 3 (Mar 2014). doi:10.1002/path.4348