8.2: Nucleic Acids - RNA Structure and Function

Search Fundamentals of Biochemistry

Learning Goals 

(Learning goals written by Claude, Sonnet 4.6, Anthropic)

RNA Structure and Coding RNA

• Explain how the presence of the 2'-OH group in RNA distinguishes it chemically and structurally from DNA — describing how this single modification (1) makes RNA more susceptible to hydrolysis, (2) prevents RNA-DNA hybrids and dsRNA from adopting the B-form helix (forcing the A-form instead), (3) enables the formation of diverse tertiary structures through 2'-OH hydrogen bonding, and (4) makes dsRNA a pathogen-associated molecular pattern (PAMP) recognized by TLR3 — and describe the overall pathway from heteronuclear RNA (containing both exons and introns) through spliceosome-catalyzed intron removal to mature mRNA, and its translation on ribosomes.
• Explain the molecular basis of mRNA vaccine technology — describing how Karikó and Weissman's discovery that substituting uridine with pseudouridine (Ψ) or N1-methylpseudouridine prevents TLR-mediated PAMP recognition while preserving base-pairing with adenine and translation capacity, how mRNA is encapsulated in lipid nanoparticles (containing ionizable/cationic lipids, phospholipids, cholesterol, and PEGylated lipids) for cellular delivery, and why this platform allows rapid vaccine development against emerging pathogens like SARS-CoV-2 and avian influenza H5N1 — connecting this to the 2023 Nobel Prize in Physiology or Medicine.

Long and Short Noncoding RNAs

• Classify noncoding RNAs by size (lncRNAs >200 nt; sncRNAs <200 nt) and describe the roles of major lncRNA and sncRNA classes — ribosomal RNA (rRNA, the catalytic core of ribosomes), transfer RNA (tRNA, 76–90 nt cloverleaf adapter molecule linking codons to amino acids through aminoacyl-tRNA synthetase-mediated acylation and anticodon-codon Watson-Crick/wobble base pairing), small nuclear RNA (snRNA, U1/U2/U4/U5/U6 RNPs assembling the spliceosome with catalysis by U6), and small nucleolar RNA (snoRNA, C/D box guide RNAs directing 2'-O-methylation of rRNA residues by base-pairing with the target sequence within the C/D ribonucleoprotein complex).
• Describe how microRNAs (miRNAs) and small interfering RNAs (siRNAs) silence gene expression through the RNA interference (RNAi) pathway — explaining the canonical biogenesis pathway (RNA Pol II transcription → Drosha cleavage of pri-miRNA → nuclear export by Exportin-5 → Dicer cleavage to 19–25 nt guide-passenger duplex → loading of guide strand into Argonaute-2-containing RISC → base-pairing of guide RNA's conserved 7-mer seed sequence with the 3'-UTR of target mRNA → translational repression or endonuclease cleavage) — distinguishing miRNA (imperfect complementarity to target, can silence multiple mRNAs) from siRNA (perfect complementarity, cleaves a single target mRNA), and explaining why let-7/lin-4 discovery in C. elegans earned the 2024 Nobel Prize in Physiology or Medicine (Ambros and Ruvkun).
• Explain how lncRNAs regulate cellular processes through RNA-protein interactions — using mamRNA (which binds Mmi1 and Mei2 in fission yeast to regulate the mitosis-to-meiosis transition by promoting ubiquitination and proteolysis of Mei2) and ToxI (a lncRNA pseudoknot-repeat antitoxin that inhibits the ToxN mRNA endonuclease in a bacterial toxin-antitoxin system) as examples — and connect the discovery of diverse novel lncRNA assemblies (OLE, ROOL, GOLLD RNA cage structures; EIciRNAs; circRNAs) to the emerging view that lncRNAs significantly expand regulatory complexity beyond the ~20,000 human protein-coding genes.

RNA: Structure and Function

Ribonucleic acids are very similar in chemical structure to DNA, except they contain ribose instead of deoxyribose. They also have the pyrimidine base uracil instead of thymine, as shown in Figures 1 and 2 above. These two small changes (but mostly the first) confer on it a very different set of biological functions than DNA. This should not surprise us, and the basis of all chemistry and biochemistry is that chemical structure determines chemical and biochemical functions and activities. In the previous section, we discussed how RNA can adopt complex tertiary structures that require additional noncanonical base pairs and chemical modifications of bases. In this section, we will explore the various types of RNA structures and their functions.

The sequence of RNA is made from DNA through a process called transcription (converting the information of DNA, a nucleic acid, into RNA, another nucleic acid). RNA can form double-stranded helices, but typically, these are viral in origin. dsRNA is a pathogen-associated molecular pattern (PAMP) that binds Toll-like receptor 3 (TLR3), as seen in Chapter 5.5. If both strands of DNA are transcribed, the resulting strands can anneal to form dsRNA. In addition, a single strand of RNA can fold on itself if the 5' and 3' ends are complementary to form a stem-hairpin loop. Figure \(\PageIndex{1}\) shows a stem-loop from a messenger RNA (4QOZ) when it is bound to a specific RNA-binding protein (not shown).

Figure \(\PageIndex{1}\) A stem-loop from a messenger RNA (4QOZ) when it is bound to a specific RNA-binding protein (not shown).

Larger ssRNA can form tertiary structures with many regions of intrastrand hydrogen bonds forming secondary structures, as shown in Figure \(\PageIndex{2}\) for one type of RNA called a transfer RNA

Figure \(\PageIndex{2}\): Secondary and tertiary structure of tRNA showing unpaired (gray) and hydrogen-bonded secondary structure regions (color). Kyle Schneider https://commons.wikimedia.org/w/inde...curid=12309266

Figure \(\PageIndex{3}\) shows a computed model for secondary structure within a much larger single RNA molecule, S11, that is part of the ribosome. The figure shows color-coded differences in accessibility when the S11 RNA is free (blue) and bound to the protein NSP2 (red), which induces structural rearrangements.

Figure \(\PageIndex{3}\): Model showing secondary structure on ribosomal S11 RNA and accessibility changes on binding the protein NSP2.

You can imagine that a different set of intrachain H-bonded double-stranded regions could easily form, with the most likely determined by sequence, local environment, and protein binding partners. Each RNA molecule would have a thermodynamic folding landscape similar to that of a protein. Programs are available to determine secondary structures from RNA sequences. Each RNA molecule would have a thermodynamic folding landscape similar to that of a protein. Also, structures are dynamic, as seen with proteins.

Another feature that complicates RNA is that many different types of RNA are made from DNA using RNA polymerases. They are loosely divided into two types of RNA. One is coding RNA, which contains the sequence information that will be translated into a protein sequence. The other type is called noncoding RNA. These RNAs regulate many cellular processes, including transcription, which produces coding RNA.

Coding RNA

The DNA template from which the coding sequence of a translatable RNA is produced is called a gene. The coding RNA, which encodes the protein, is called messenger RNA (mRNA). The exact sequence of RNA in mRNA that encodes a protein is derived from a longer contiguous DNA sequence in the nucleus, from which sections called intervening sequences or introns have been removed. The coding sequences of DNA, which are separated by introns, are called exons. When DNA is transcribed, a long, contiguous sequence containing both exons and introns is transcribed into a single primary transcript, called heteronuclear RNA. The introns in the heteronuclear RNA are removed in a splicing reaction catalyzed by a large complex called the spliceosome to form mRNA. The process is illustrated in Figure \(\PageIndex{4}\). The first RNA sequence made is the heteronuclear RNA.

Figure \(\PageIndex{4}\): https://commons.wikimedia.org/wiki/F...ns_introns.gif. This work is in the public domain in the United States

The long, single-stranded mRNA molecule binds to ribosomes and nanomachines that orchestrate the translation of its sequence into a protein sequence. Around 20,000 human genes produce an even larger number of mRNA that arise from differential splicing of the primary transcript.

RNA Vaccines

Amazing progress has been made in creating mRNA vaccines, as demonstrated by the Moderna and Pfizer RNA vaccines targeting the spike protein of the SARS-CoV-2 virus. Models show that just in the US  through 2022, COVID-19 vaccines saved the US $1.15 trillion and 3 million lives.  Just in the first year, the vaccines are estimated to have saved upwards of 20 million lives worldwide, an accomplishment worthy of every prize in the world!  Many more could have been saved in the developing world if the vaccine had been more widely available.

As we learned previously, RNA is much more labile to hydrolysis than DNA since RNA has a 2' OH group. Methods to stabilize RNA were required before an RNA vaccine could become a reality. More importantly, dsRNA is a danger signal that a viral infection may be present. dsRNA, a pathogen-associated molecular pattern (PAMP), binds to Toll-like Receptor 3 (TLR3) and initiates an inflammatory response that eliminates the RNA before it can be decoded into a protein sequence that could elicit an antibody response, a requirement for a vaccine.

Katalin Karikó and Drew Weissman found that modifying uracil bases to pseudouracil (Ψ) prevents the PAMP response and allows RNA to persist long enough to encode the protein sequence (i.e., increasing RNA stability against hydrolysis). Pseudouracil (Ψ) is found in structural RNAs (transfer, ribosomal, small nuclear, and small nucleolar), and is the most common modification found in RNA. It is even metabolized by a naturally occurring pathway back to uracil.   

Figure (\PageIndex{22}\) below shows the structures of uracil and the N1-methyl derivative of pseudouracil attached to ribose in an RNA and their base pairing to adenine. The N1-methylpseudouracil base still bases pairs with an adenine base.  Hence, the modification does not affect the RNA structure and its functional ability to be decoded into a protein sequence.  In short, mRNAs modified to contain pseudouracil have a much greater translational capacity and stability.  In addition, methylation of uracil decreases the immunogenicity of RNAs that contain it.

Figure (\PageIndex{22}\): Structural and base pairing features of uracil and N1-methylpseudouracil

Figure (\PageIndex{23}\) below shows a cartoon version of the modified mRNA (a) Covid-19 and its encapsulation into a lipid nanoparticle (b) for an mRNA vaccine.

Figure (\PageIndex{23}\): Modified mRNA (a) Covid-19 and its encapsulation into a lipid nanoparticle (b) for an mRNA vaccine. Heinz, F.X., Stiasny, K. Distinguishing features of current COVID-19 vaccines: knowns and unknowns of antigen presentation and modes of action. npj Vaccines 6, 104 (2021). https://doi.org/10.1038/s41541-021-00369-6. Creative Commons Attribution 4.0 International License

Panel (a) shows the generalized structure of the mRNA vaccine for the S gene, which encodes the virus's surface spike protein. Panel b shows the lipid nanoparticle that encapsulates and protects the mRNA vaccine. The lipids include proprietary mixtures of phospholipids, cholesterol, cationic (ionizable) lipids, and polyethylene glycol (PEG)- modified lipids. Without the basic research of Karikó and Weissman (and many others), the world would not have had the Covid-19 vaccine in time to save so many lives. They were awarded the Nobel Prize in Medicine in 2023 for their work.  

Biosynthesis of pseudouridine

The biosynthesis of pseudouridine, the most common modification of cellular RNA, is done after DNA transcription (i.e., post-transcriptionally) using the enzyme pseudouridine synthases (PUS), which is found in all kingdoms in life. The reaction involves

• cleavage of the C-N-glycosidic bond of uridine in RNA
• rotation of the cleaved uracil to align C5 of uracil and C1′ of the ribose
• formation of the C1′-C5 carbon–carbon bond. 

These processes are illustrated in Figure (\PageIndex{24}\) below.

Figure (\PageIndex{24}\): Post-transcriptional modification of uridine to pseudouridine.  Czudnochowski N. et al. The mechanism of pseudouridine synthases from a covalent complex with RNA, and alternate specificity for U2605 versus U2604 between close homologs. Nucleic Acids Res. 2014 Feb;42(3):2037-48. doi: 10.1093/nar/gkt1050. Epub 2013 Nov 7. PMID: 24214967; PMCID: PMC3919597. Creative Commons Attribution License (http://creativecommons.org/licenses/by/3.0/)

https://www.nature.com/articles/s41467-025-64829-6

The influenza virus has been a great killer and a cause of recurring pandemics.  Vaccines against it are only 10-60% effective each year, since they are made in chicken eggs and the influenza strain used is chosen well in advance of when it's given, so the prevailing strain at the time of immunization may differ significantly from the strain used to produce the vaccine.  Most influenza vaccines are made using live-attenuated or inactivated virus.  They are not suited for emerging pandemic strains, such as those that could arise from virulent avian influenza viruses.  It would be ideal if an mRNA-based influenza vaccine could be developed, as it could be made quickly against an emerging pandemic virus.  It would be better if the vaccine could be administered nasally to stimulate the mucosal membrane, since the virus gains entry to the body through the nose and lungs.  In addition, they are much easier to produce in large quantities compared to egg-based vaccines.  A nasal mis vaccine would also address the question of vaccine hesitancy since not intramuscular needle injection would b necessary.  

Intranasal injection simulates the secretory IgA immunoglobulins and also memory B and T cells in the lungs.  It could help in the early stages of infection, before widespread viral replication, and prevent transmission. A thermostable vaccine, a replicon (i.e., mRNA)-nanostructured lipid carrier, has been developed targeting the H5 and H7 strains of the influenza virus (high risk).  The vaccine protects mice and confers full protection against H5N1 and H7N9 viruses in ferrets.  Ferrets' respiratory tracts are anatomically and physiologically similar to ours.  They also express similar sialic acid receptors in their upper airways as humans.  They can also transmit the virus to other ferrets through aerosol drops from sneezing.  

---

mRNA molecules have a polyA sequence connected to the 3' end (see Chapter 25.2: RNA Processing for more details). PolyA binding proteins (PABP) interact with the polyA tail and are likely to stabilize the RNA molecule.  A PABP domain was fused to another protein called a major vault protein (MVP).  The MVP (100K molecular mass) forms a large multimeric "vault" with a large internal volume that can encapsulate large molecules (A Cellular ‘vaults’ deployed to spy on gene activity), such as RNAs.  The assembled vault may act as a scaffold for proteins involved in signal transduction and in transport between the nucleus and cytoplasm.  During assembly, the vault captures mRNA molecules produced by the cell for up to a week, enabling real-time analysis of transcription. Table \(\PageIndex{1}\) below shows two iCn3D models of oligomeric vaults created by the MVP

Human Vault Cage (9BW5) Click the image for a popup or use this external link: https://www.ncbi.nlm.nih.gov/Structu...0508a84051ead8
Rat Half vault structure (7PKY) Click the image for a popup or use this external link:https://www.ncbi.nlm.nih.gov/Structu...23ab413f7914d9 (long load time)

Table \(\PageIndex{1}\):  iCn3D models of oligomeric vaults created by the MVP. The top is the full vault and the bottom the half vault. One monomer in the half vault is shown in spacefill (brown).  

Pablo Guerra et al. Symmetry disruption commits vault particles to disassembly. Sci. Adv.8,eabj7795(2022). DOI:10.1126/sciadv.abj7795

Half-vaults (39mers) from oligomers of MVP then assemble into the full vault (78mer). In mammals, the 193-kDa vault poly(adenosine 5′-diphosphate) ribose polymerase (VPARP) protein, the 290-kDa telomerase-associated protein 1 (TEP1), or small untranslated RNAs can be found inside.  

---

Noncoding RNA (ncRNA)

Not long ago, few thought about possible RNA transcripts from non-coding regions of the genome, except for two types of RNA required to translate mRNA. These two are ribosomal RNAs (rRNAs) found in ribosomes, and transfer RNAs, to which amino acids are esterified and transferred to a growing protein chain on the ribosome. Many more classes have been discovered and given names that confuse those more familiar with protein structures. One way to classify noncoding RNAs (ncRNAs) is based on size.

• short noncoding RNAs (sncRNAs) are <200 nucleotides
• long noncoding RNAs (lncRNAs)are >200 nucleotides

These function to regulate gene expression at both the transcription and post-transcriptional levels. Some have catalytic functions. Some affect chromosome structure and chemical modification.

Long Noncoding RNAs (lncRNAs)

There may be between 16,000 to over 100,000 human lncRNAs encoded into the genome, which adds much complexity to our understanding of the function of RNA transcripts. An online lncipedia is a database of searchable lncRNA sequences. There are many types of lncRNAs. The first we will consider is ribosomal RNA.

a. Ribosomal RNA (rRNA):

These RNAs fit the simple definition of lncRNAs (>200 nucleotides and are not protein-coding), but most would not think of them as lncRNAs since they have always been in their own category of a nonprotein-coding gene. rRNAs vary in length from between 1500 and 3000 nucleotides in bacteria and about 1800 and 5000 nucleotides in humans, and are the core structure of ribosomes. These nanomachines translate bound mRNA into a protein sequence.

Figure \(\PageIndex{5}\) shows an interactive iCn3D model of the structure of 23S rRNA of the large ribosomal subunit from Deinococcus radiodurans (2O44) (long load time).

Figure \(\PageIndex{5}\): Structure of 23S rRNA of the large ribosomal subunit from Deinococcus radiodurans (2O44). (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/icn3d/share.html?Tb2BVwSTe6LviPou5

The red (highlighted yellow) spacefill is the 5' start of the rRNA. The chain has a complex tertiary structure, like a protein, and ends at the cyan space-filling 3' end. It has 2880 nucleotides.

b.  Other types

Let's focus on more classic examples of long noncoding RNAs (i.e., not rRNA). One way to categorize them is by the position in the genome where they are encoded. The different types include long intergenic noncoding RNAs (lincRNAs), intronic lncRNAs, antisense RNAs (as lncRNAs), and other variants. These are illustrated in Figure \(\PageIndex{6}\), where the lncRNA is pink.

Figure \(\PageIndex{6\): https://en.Wikipedia.org/wiki/Long_non-coding_RNA. Creative Commons Attribution-Share Alike 3.0 Unported

Another variant is exon and intron-containing circRNAs (EIciRNAs), as illustrated in panel B in Figure \(\PageIndex{13}\). These are presumably produced from pre-mRNA for a given mRNA and appear to regulate gene expression through RNA-RNA interactions with U1 snRNA, which starts the assembly of the spliceosome on pre-mRNA when it binds to the 5′ pre-mRNA splice site.

circRNAs are found throughout the biological world and have been associated with diseases such as cancer, cardiovascular disease, and brain disorders like Alzheimer's.  Their possible functions include binding to miRNAs and RNA-binding proteins, altering their activities, and binding to DNA to regulate transcription.  CircAtlas is a database of circRNAs that two techniques have experimentally validated

Now let's consider specific examples of long non-coding RNAs (lncRNAs), which are often bound to target proteins.

• mamRNA (a lnc RNA)

The lncRNA mamRNA (Mmi1 and Mei2-associated RNA) binds the proteins Mmi1 and Mei2 in Schizosaccharomyces pombe, which control the balance between meiosis and mitosis in yeast. (Schizosaccharomyces pombe is a "fission" yeast that divides by fission, not budding.) The MamRNA has two variants, 550 and 700 nucleotides in length. The binding of mamRNA leads to the ubiquitinylation of Mei2 in the complex. Mmi1 R is an RNA-binding protein that binds to a modified version of adenosine methylated at N6 and is found internally in mRNA. Mei2 (meiosis protein 2) is necessary of meiosis. The binding of mamRNA leads to the ubiquitinylation of Mei2 in the complex, targeting it for proteolysis. Mei2 concentrations increase relatively, shifting yeast from mitosis to meiosis. Figure \(\PageIndex{7}\) shows a cartoon depicting these interactions.

Figure \(\PageIndex{7}\): Binding of the lncRNA mam2 to yeast proteins Mei2 and Mmi1. Andric, V. et al. Yeast. Non-coding RNA 2021, 7, 34. https://doi.org/10.3390/ncrna7020034. Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/).

Figure \(\PageIndex{8}\) shows an interactive iCn3D model of the S. pombe Mei2 RRM3 protein domain bound to the Mei2 binding "domain" of mamRNA (6YYM), which in this structure is only eight nucleotides long (not the full length of this lncRNA, which is 550 and 700 nucleotides long).

Figure \(\PageIndex{8}\): S. pombe Mei2 RRM3 protein domain bound to an eight-nucleotide section of the yeast lncRNA mam (6YYM). (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...5xmQjwAmQL9on7

• ToxI - a lncRNA inhibitor of the endonuclease ToxN

Those with a more chemistry-centric background might be surprised that viruses also "infect" bacteria. These viruses are called bacteriophages. It is estimated that there are over 10³⁰ in nature. Some covalently incorporate into genomes, where they reside permanently. They are a major driver of bacterial genome evolution, shaping bacteria's immune responses and adaptations.

One very interesting example is the type III toxin-antitoxin (TA) system in E. Coli. It consists of a toxin, ToxN, which is a nuclease that cleaves internally after the second A in a AAA sequence. It acts on mRNA, but especially pre-mRNA sequences. It is inhibited by the binding of a lncRNA called ToxI (toxin inhibitor). The RNA sequence of the ToxI inhibitor contains 36 "domain" repeats of a pseudoknot, each of which is sufficient to inhibit ToxN. The ToxN endonuclease cleaves the ToxI lncRNA as it assembles the complex. It also cleaves its mRNA. Figure \(\PageIndex{9}\) shows an interactive iCn3D model of the protein toxin (ToxN):lncRNA (ToxI), which is a shortened version of 29 29-nucleotide section from Pectobacterium atrosepticum (2xdb).

Figure \(\PageIndex{9}\): Protein toxin (ToxN) - 29 NT fragment of lncRNA antitoxin ToxI complex (2xdb). (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...KjNPGghiUyAXr9

As we discussed in Chapter 3.2: The Structure of Proteins - An Overview, some long noncoding RNAs have noncanonical open reading frames (ncORFs) that are translated into proteins, adding to the complexity of RNA.

Recent Updates:  January 18, 2026

New types of long noncoding RNAs are routinely discovered.  Some are termed Natural RNA-Only Assemblies. These are often found in bacteria and viruses that infect them (bacteriophages). Some examples of iCn3D structures are shown in Table \(\PageIndex{2}\) below.

Ornate Large Extremophilic (OLE) RNAs (pdb_00009lcr) Click the image for a popup or use this external link: https://www.ncbi.nlm.nih.gov/Structu...d35637ad01dfd1
ROOL (Rumen-Originating, Ornate, Large) RNA hexamer from bacteria found in cow stomachs (pdb_00009j6y) Click the image for a popup or use this external link: https://www.ncbi.nlm.nih.gov/Structu...0f5f943d312af6
GOLLD (Giant, Ornate, Lake- and Lactobacillales-Derived) RNAs from bacteria 10mer (pdb_00009l0r) Click the image for a popup or use this external link: https://www.ncbi.nlm.nih.gov/Structu...9c2b7ccc40d727

Table \(\PageIndex{2}\):  iCn3D structures of Natural RNA-Only Assemblies

ROOL and GOLLD structures form cages with internal spaces.  For more information, visit this PDP 101 page on Natural RNA-Only Assemblies by Janet Iwasa.

Short Noncoding RNA

Short noncoding RNAs (sncRNAs) are less than 200 nucleotides in length. By definition, this would include transfer RNAs (tRNAs), which bring to the ribosome amino acids covalently attached to their 3' end of the tRNA for incorporation into a growing protein chain during the translation of mRNA. As with rRNA for lncRNAs, these are really in a class of their own. Others include small nuclear RNAs (snRNAs) involved in splicing, small nucleolar RNAs (snoRNAs) involved in the modification of rRNAs, and microRNAs (miRNAs), involved in the inhibition of translation and transcription, PIWI-interacting RNAs (piRNAs), and endogenous small interfering RNAs (siRNAs). It is difficult to remember the subtle differences among these, which makes them hard to understand. We will tell their stories with a few targeted examples.

a. Transfer RNA:

Transfer RNAs act as adapter molecules between transcription and translation. They are 76-90 nucleotides long and have a cloverleaf shape. An enzyme, aminoacyl-tRNA synthase, covalently attaches a select amino acid at its 3' end. Another end of the tRNA hydrogen bonds through 3 nucleotides (the anticodon) to a triplet nucleotide (the codon) on the mRNA that encodes a specific amino acid at that triplet position. Figure \(\PageIndex{10}\) shows an interactive iCn3D model of the structure of yeast phenylalanine tRNA (1EHZ)

Figure \(\PageIndex{10}\): Yeast phenylalanine tRNA (1EHZ) (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/icn3d/share.html?iAWpKknDWbQm5hHL6

b. Small nuclear RNA (snRNA):

The spliceosome is a nanoparticle that catalyzes the removal of introns from pre-mRNA in eukaryotes (prokaryotes appear devoid of introns). The yeast spliceosome has a molecular weight of 1.3 million and contains five small ribonucleoproteins (RNPs) with many other associated proteins. Each of the 5 RNPs has a small nuclear RNA (U1, U2, U4, U5, and U6) enriched in uracils. U6 is highly conserved and is directly involved in catalysis. Figure \(\PageIndex{11}\) shows an interactive iCn3D model of the core structure of the U6 small nuclear ribonucleoprotein complex with most of the U6 RNA bound.

Figure \(\PageIndex{11}\): Core structure of the U6 small nuclear RNA - protein (ribonucleoprotein) complex (4N0T). (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...P3SLVRAKhP7W57

c. MicroRNAs (miRNAs) and small inhibitory RNAs (siRNAs)

MicroRNAs (miRNAs) control the expression of thousands of genes in plants and animals. They are single-stranded but fold on themselves to form a stem-hairpin. The miRBase is a microRNA database containing almost 40,000 miRNA sequences. miRNAs are highly conserved and are found in animals, plants, and some unicellular eukaryotes. They interact with the 3′ untranslated regions of mRNAs and inhibit or prevent their translation. Several key proteins, RNA polymerase II, Drosha, and Dicer are involved in the canonical pathway.  The others appear to be independent of Drosha, which is a ribonuclease III double-stranded (ds) RNA endoribonuclease.

Dicer is a dsRNA) endoribonuclease, which cleaves long dsRNAs and short hairpin pre-microRNAs (miRNA) into fragments of either 21-23 nucleotides (short interfering RNA) or 19-25 nucleotides (microRNAs). Each has two nucleotides that are unpaired at the 3' end. These bind to the enzyme complex RISC ( RNA-induced silencing complex), which then targets them to mRNA complementary to the siRNA/miRNA (RISC), causing cleavage of the mRNA and hence inhibiting translation.

Small inhibitory RNAs (siRNAs) are very similar to miRNAs (to the point that differentiating between them is somewhat arbitrary). They both engage in RNA interference (RNAi) of mRNA translation. Here are some reported differences:

• The substrate for dicer cleavage is dsRNA (that could be added exogenously) of length 30-100+ for siRNA, but the actual pre-miRNA of length 7-100 nucleotides that may contain hairpins with some mismatches (i.e., not a perfect stem and hairpin)
• The final RNA after dicer processing is double-stranded for both and 21-23 nucleotides long for siRNA and 19-25 for miRNA
• siRNAs are perfectly complementary to the target mRNA, while miRNAs, which are not necessarily perfectly complementary, typically bind to the 3' untranslated end of the mRNA
• Because of the perfect complementarity to target mRNA, siRNA interacts with only one mRN,A while miRNAs, given that they are not perfectly complementary to their target sequences, can bind different mRNAs
• Given their higher affinity for binding, siRNAs lead to Dicer endonuclease cleavage of the target mRNA. In contrast, inhibition of mRNA translation by miRNAs arises from binding of the miRNA to the mRNA or, if the match between the miRNA and mRNA is high enough, endonuclease cleavage of the mRNA.

Figure \(\PageIndex{12}\) shows canonical and several alternative pathways for their transcription and processing from the noncoding miRNA genes.

Figure \(\PageIndex{12}\):

Figure 1. Canonical and non-canonical pathways of microRNA biogenesis. (A) Canonical pathway—microRNA gene is transcribed by RNA polymerase II into primary microRNA (pri-miRNA), cleaved by microprocessor complex Drosha/DGCR8, and precursor microRNA (pre-miRNA) is exported from the nucleus to the cytoplasm by Exportin 5 (XPO5) and further processed by Dicer and its partners into 18–25 nucleotide long microRNA duplex with 2-nucleotide 30 overhangs. The guide strand is subsequently bound by the Argonaute proteins 1-4 (AGO1-4) and retained in the microRNA-induced silencing complex to target mRNAs for post-transcriptional silencing. (B) Mirtrons—generated through mRNA splicing independently of the Drosha-mediated processing step. (C) Small nucleolar RNA-derived microRNAs—Drosha-independent pathway. (D) Exportin 5-independent transport of pre-miRNAs from the nucleus to the cytoplasm has been described in the case of miR-320 family. (E) Dicer-independent processing of miR-451—pre-miR-451 is directly loaded into AGO2, cleaved, and trimmed by poly(A)-specific ribonuclease PARN to produce mature miR-451. Gregorova et al. Cancers 2021, 13, 1333. https://doi.org/10.3390/cancers13061333. Creative Commons Attribution (CC BY) license (https://creativecommons.org/licenses/by/4.0/)

siRNA that are perfect matches to specific mRNA can be easily designed and purchased for translation inhibition and through gene silencing studies. Both miRNAs and siRNAs are potentially therapeutic as it is much simpler to design a drug that targets an mRNA sequence (a 1D sequence target) than a protein active site (a 3D target). Additionally, they can be used to inhibit the synthesis of target proteins that lack a "druggable" active site.

The protein Argonaute is involved in miRNA- and siRNA-mediated silencing of genes through their mRNAs in the RISC (RNA-induced silencing complex). RISC contains the protein argonaute 2 (AGO2) bound to a "guide" RNA, which is either microRNA (miRNA) or short interfering RNA (siRNA). The miRNA or siRNA directly interacts with the "target" - the mRNA.

• Example miRNA:

Figure \(\PageIndex{13}\) shows an interactive iCn3D model of human Argonaute2 Bound to a Guide (miRNA) and Target RNA (4W5O). The two RNA sequences are 5' UUCACAUUGCCCAAGUCUUU 3' and 5' CAAUGUGAAA 3'.

Figure \(\PageIndex{13}\): human Argonaute2 Bound to a Guide and Target RNA (miRNA) (4W5O). (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/icn3d/share.html?9fvyQmPkk4Kctei59

Recent Updates:  October 9, 2024

2024 Nobel Prize in Physiology and Medicine for discovery of miRNAs

Victor Ambros and Gary Ruvkun won the Nobel Prize (October 7, 2024) for discovering miRNA and its function in the roundworm C. elegans, which has fewer than 1000 cells.  Their basic research on miRNAs' role in protein synthesis regulation was extended to other organisms.  Humans express over 1000 miRNAs that regulate translation.  

Ruvkun found that two small RNAs, lin-4 (22 nucleotides) and let-7 (21 nucleotides), are required to progress from larval to adult stages in the worm.   They were nonhomologous but complementary to parts of the 3'-untranslated regions (UTRs) of mRNAs encoding proteins that are downregulated during development.  Let-4 is found in many animal species, including humans, and binds to the 3'-UTR of the mRNA for the protein lin-41. Ambrose discovered that a small RNA lin-4 from a non-coding gene was complementary to regions in the 3'-UTR of the mRNA for a protein lin-14.

The seed sequences of miRNAs are a conserved 7-mer at positions 2-7 from the miRNA 5'-end. This part of the miRNA must be exactly complementary to the 3'-UTR of the target mRNA, but the other bases don't have to match exactly.

The sequence of let-7 miRNA from C. elegans is 5'-UGAGGUAGUAGGUUGUAUAGU-3'.  Figure \(\PageIndex{i}\) below shows the aligned sequences of a let-7 miRNA and lin-41 mRNA (A) and analogs made for NMR structure determination (B,C).  

Figure \(\PageIndex{i}\): Sequences of a let-7 miRNA and lin-41 mRNA (A) and analogs.

( A ) Schematic representation of the complex of let-7 miRNA with 3′-UTR of the lin-41 mRNA (LCS 2). The residues in green and blue match the sequences used in the monomolecular and dimeric constructs. ( B ) The 33-nt monomolecular RNA construct mimicking the complex. ( C ) The dimeric RNA construct. The numbering in (B) has been preserved to facilitate comparison of the two constructs. The residues originating from the lin-41 mRNA sequence are labeled by asterisk.  Cevec, M, Thibaudeau, C and Plavec, J.  Nucleic Acids Research, Volume 36, Issue 7, 1 April 2008, Pages 2330–2337, https://doi.org/10.1093/nar/gkn088.  Creative Commons CC-BY-NC license. 

Two analog structures preserved most of the matching sequences from the miRNA and mRNA (green for the let-7 miRNA and blue for the lin-41 mRNA.  The seed sequence of the miRNA was completely preserved. One was a single strand that formed a stem-loop structure. Figure \(\PageIndex{j}\) shows an interactive iCn3D model of the solution structure of a let-7 miRNA:lin-41 mRNA complex from C. elegans (2JXV) using the single-stranded analog.

Figure \(\PageIndex{j}\): solution structure of a let-7 miRNA:lin-41 mRNA complex from C. elegans (2JXV). (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...xJh4paMJRiTQ47

G1 (red spacefill) is the 5'-end and C33 (yellow) is the 3'-end of the single-stranded analog. The green bases match those from the figure above. For clarity, the matching base partners of the lin-41 mRNA are shown in cyan rather than blue.

The model shows two stems with an asymmetric loop containing 3 Us and 2 A.  A GU wobble base is seen in the first stem between U6 and G28.

 

• Example: siRNA (Small interfering RNA)

Virus genomes are ultimately decoded into new viruses by the host's replication, transcription, and translation machinery. Host cells have evolved mechanisms to silence viral mRNAs. Unfortunately, viruses evolve ways to suppress host RNA silencing. Many viral proteins are used by the host to suppress silencing. The viral p19 protein preferentially binds to host short interfering RNAs (siRNAs) rather than to microRNAs (miRNAs). A single mutation in the viral p19 protein changes its selectivity, allowing it to bind a specific human miRNA, miR-122. This shows the subtle complexities of protein:RNA interactions. Figure \(\PageIndex{14}\) shows an interactive iCn3D model of the viral suppressor of RNA silencing protein and a 21-residue small interfering RNA (6BJV)

Figure \(\PageIndex{14}\): Viral suppressor of RNA silencing protein p19 mutants and a double-stranded 21 nucleotide small interfering RNAs (6BJV). (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/i...VXzSgezTEPcPt9

• Example: piRNA (a specific miRNA)

Piwi proteins are RNA-binding proteins in plants and animals and are structurally similar to argonaute. They bind a guide RNA called piwi-interacting RNA (piRNA) and lead to the silencing of transposable elements that can move around the genome. piWi has endonuclease activity and can cleave mRNA. The piRNAs are just one type of miRNA. Figure \(\PageIndex{15}\) shows an interactive iCn3D model of Ephydatia fluviatilis (a sponge) PiwiA with a guide (piRNA) and-target RNA(7KX9)

Figure \(\PageIndex{15}\): Ephydatia fluviatilis PiwiA-piRNA-target complex (7KX9). (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/icn3d/share.html?j9ikApUgsVvqsiM5A

lncRNAs and miRNAs in the brain

After the decoding of the human genome, many have been struggling to understand how the complexity of the human brain (large size, greater connectivity among neurons) arises, given that we appear to have only around 20,000 protein genes encoded by the genome (not counting small proteins of less than 100 amino acids). Long noncoding RNAs (lncRNAs) and miRNAs appear to be significant pieces of this puzzle. Their ability to regulate transcription during development may hold the key. Additional roles for these RNAs beyond transcriptional regulation are being discovered. Some are transported away from the nucleus to serve other functions in axons, dendrites, etc. For example, the lncRNA Gm38257 binds to proteins that structure the synapse (components of the spectrin/ankyrin complex) instead of simply regulating gene expression.

The miRNA repertoire appears to be significantly expanded in "intelligent" organisms such as humans and octopuses. For example, a large increase (179) in miRNAs occurs in the evolutionary scale from mice (which have about 24,000 protein-encoding genes) to humans (around 20,000). miRNAs and lncRNAs may be involved (causative or correlative?) with brain disease. An example is miR-124, which is significantly elevated (3.5X) in hippocampal cells from mouse models of Alzheimer's compared to normal mice. Altered expression of the lncRNA named Gomafu, RNCR2 or MIAT) appears to affect certain psychiatric diseases.

c. small nucleolar RNA

The nucleolus is a small nuclear structure that helps assemble the ribosomal RNAs synthesized in the nucleus. They are then transported through the nuclear membrane into the cytoplasm, where they combine with cytoplasmic proteins to form complete ribosomes. As described below, rRNA is chemically modified by enzymes (much like the post-translational modification of proteins). One such modification is 2'-O-methylation in archaea and eukaryotes. A class of small nucleolar RNAs (snoRNAs) that vary from 10-21 base pairs is called C/D RNAs, and they "guide" the modification. Hence, they are also called "guide" RNAs. These snoRNAs bind to 3-4 proteins into ribonucleoproteins. Figure \(\PageIndex{16}\) shows an interactive iCn3D model of the box C/D ribonucleoprotein 40 nt snoRNA "guide" and a 10-nucleotide RNA target substrate. It appears that the maximal duplex RNA formed (from the guide and target) is 10 base pairs long.

Figure \(\PageIndex{16}\): Box C/D ribonucleoprotein with a small nucleolar RNA (snoRNA) 12 nucleotide "guide" and 13 nucleotide target RNA for 2'-O-methylation (5GIO). (Copyright; author via source). Click the image for a popup or use this external link: https://structure.ncbi.nlm.nih.gov/icn3d/share.html?q45tA8LEko87mYTw6

Summary

(Summary written by Claude, Sonnet 4.6, Anthropic)

This chapter surveys the structural and functional diversity of RNA, extending from its chemical distinction from DNA through the multiple classes of noncoding RNA that regulate virtually every aspect of gene expression — establishing that RNA is far more than a passive messenger between DNA and protein.

RNA structure and its chemical basis differ from DNA in two fundamental ways: the presence of a 2'-OH group on ribose and the substitution of uracil for thymine. The 2'-OH makes RNA more susceptible to hydrolytic cleavage (the 2'-O⁻ can attack the adjacent phosphodiester bond in an intramolecular transesterification), explains why RNA-DNA hybrids adopt the A-form helix rather than B-form (the 2'-OH sterically clashes with the minor groove in B-form geometry), enables additional hydrogen bonding within RNA that stabilizes complex tertiary structures, and makes double-stranded RNA (dsRNA) a danger signal recognized by TLR3 as a PAMP. While DNA primarily serves as a stable repository of genetic information in double-helical form, single-stranded RNA folds into complex three-dimensional tertiary structures — containing stems, hairpin loops, bulges, pseudoknots, and G-quadruplexes stabilized by noncanonical base pairing — that functionally resemble proteins. Like proteins, each RNA has a thermodynamic folding landscape, and RNA structure is dynamic and modulated by interaction with proteins and small molecules.

Coding RNA (mRNA) carries the sequence information that is translated into protein. The primary nuclear transcript (heteronuclear RNA) contains both exons (protein-coding sequences) and introns (intervening sequences), which are removed by the spliceosome (a ~1.3 MDa ribonucleoprotein containing U1, U2, U4, U5, and U6 snRNPs) through two transesterification reactions analogous to Group II intron self-splicing, yielding mature mRNA. About 20,000 human protein-coding genes produce a much larger number of mature mRNAs through alternative splicing. The mRNA vaccine platform — enabled by the landmark discovery of Karikó and Weissman (2023 Nobel Prize in Physiology or Medicine) that substituting uridine with pseudouridine (Ψ) or N1-methylpseudouridine eliminates TLR3/7/8-mediated PAMP recognition while preserving adenine base-pairing and ribosomal decoding — allows rapid vaccine production against emerging pathogens. Modified mRNA is encapsulated in lipid nanoparticles (containing ionizable/cationic lipids, phospholipids, cholesterol, and PEGylated lipids for stability and cellular uptake) for delivery. The COVID-19 mRNA vaccines saved an estimated 20+ million lives in the first year and $1.15 trillion in US healthcare costs through 2022. This platform is now being extended to nasal mRNA influenza vaccines (targeting H5 and H7 strains) that stimulate mucosal IgA and lung memory B and T cells.

Long noncoding RNAs (lncRNAs, >200 nt) comprise a vast and diverse class of regulatory RNAs. Ribosomal RNAs (rRNAs, 1500–5000 nt) are the structural and catalytic core of ribosomes, with the 23S/28S rRNAs providing the peptidyl transferase activity at the peptidyl transferase center. Beyond rRNA, true lncRNAs include long intergenic noncoding RNAs (lincRNAs), intronic lncRNAs, antisense lncRNAs, and circular RNAs (circRNAs). The lncRNA mamRNA in fission yeast simultaneously binds the proteins Mmi1 and Mei2, promoting the ubiquitination and proteolysis of Mei2 and thereby regulating the mitosis-to-meiosis transition; structural studies reveal that even an 8-nucleotide segment of mamRNA is sufficient to bind Mei2's RNA-recognition motif. In bacteria, ToxI — a lncRNA with 36 pseudoknot repeats — functions as an antitoxin by binding and inhibiting the mRNA endonuclease ToxN in a type III toxin-antitoxin system, protecting the bacterium from phage infection. Novel "Natural RNA-Only Assemblies" (OLE, ROOL, GOLLD RNAs) form cage-like RNA quaternary structures analogous to protein complexes. CircRNAs — circular RNAs without 5' caps or 3' poly-A tails — have been associated with cancer, cardiovascular disease, and neurological disorders. The biological complexity encoded in the estimated 16,000–100,000 human lncRNAs may explain aspects of human brain complexity and disease that the ~20,000 protein-coding genes alone cannot account for.

Short noncoding RNAs (sncRNAs, <200 nt) include tRNAs, snRNAs, snoRNAs, miRNAs, siRNAs, and piRNAs. Transfer RNAs (76–90 nt) adopt the iconic cloverleaf/L-shaped tertiary structure, with specific amino acids esterified to the 3'-CCA end by aminoacyl-tRNA synthetases; the anticodon loop reads the mRNA codon through Watson-Crick and wobble base pairing. Small nuclear RNAs (snRNAs U1–U6) assemble with proteins into snRNPs that form the spliceosome; U6 is the most conserved and directly catalyzes splicing. Small nucleolar RNAs (snoRNAs), particularly C/D box snoRNAs, base-pair with specific rRNA sequences to direct site-specific 2'-O-methylation by the associated methyltransferase in the C/D ribonucleoprotein complex. MicroRNAs (miRNAs, 19–25 nt) are perhaps the most numerous and broadly important sncRNAs: expressed from ~1000+ human genes, they are processed from pri-miRNA transcripts by Drosha (nuclear) and Dicer (cytoplasmic) and loaded into Argonaute-2 (AGO2) within the RISC complex. The conserved 7-mer "seed sequence" at positions 2–7 from the 5' end must base-pair with the target mRNA 3'-UTR; imperfect complementarity causes translational repression, while perfect complementarity (as in siRNA) triggers endonuclease cleavage. The discovery of lin-4 (22 nt) and let-7 (21 nt) miRNAs by Victor Ambros and Gary Ruvkun in C. elegans, showing these small RNAs regulate developmental timing by base-pairing with 3'-UTRs of developmental mRNAs, earned the 2024 Nobel Prize in Physiology or Medicine. Let-7 is conserved from worms to humans and targets mRNAs including lin-41; its loss of regulation is implicated in cancer. siRNAs differ from miRNAs in that they are perfectly complementary to their single target mRNA, causing obligatory endonuclease cleavage; they can be rationally synthesized against any mRNA sequence and used as research tools or therapeutics. piRNAs guide Piwi-clade Argonaute proteins to silence transposable elements in germline cells. The expanding repertoire of miRNAs and lncRNAs in evolutionarily complex organisms — with 179 new miRNAs arising in the human lineage relative to mice — suggests that noncoding RNA expansion may underlie the evolution of cognitive and neural complexity.

References

Börner, R., Kowerko, D., Miserachs, H.G., Shaffer, M., and Sigel, R.K.O. (2016) Metal ion induced heterogeneity in RNA folding studied by smFRET. Coordination Chemistry Reviews 327 DOI: 10.1016/j.ccr.2016.06.002 Available at: https://www.researchgate.net/publication/303846502_Metal_ion_induced_heterogeneity_in_RNA_folding_studied_by_smFRET

Hardison, R. (2019) B-Form, A-Form, and Z-Form of DNA. Chapter in: R. Hardison’s Working with Molecular Genetics. Published by LibreTexts. Available at: https://bio.libretexts.org/Bookshelves/Genetics/Book%3A_Working_with_Molecular_Genetics_(Hardison)/Unit_I%3A_Genes%2C_Nucleic_Acids%2C_Genomes_and_Chromosomes/2%3A_Structures_of_Nucleic_Acids/2.5%3A_B-Form%2C_A-Form%2C_and_Z-Form_of_DNA

Lenglet, G., David-Cordonnier, M-H., (2010) DNA-destabilizing agents as an alternative approach for targeting DNA: Mechanisms of action and cellular consequences. Journal of Nucleic Acids 2010, Article ID: 290935, DOI: 10.4061/2010/290935 Available at: https://www.hindawi.com/journals/jna/2010/290935/

Mechanobiology Institute (2018) What are chromosomes and chromosome territories? Produced by the National University of Singapore. Available at: https://www.mechanobio.info/genome-regulation/what-are-chromosomes-and-chromosome-territories/

National Human Genome Research Institute (2019) The Human Genome Project. National Institutes of Health. Available at: https://www.genome.gov/human-genome-project

Wikipedia contributors. (2019, July 8). DNA. In Wikipedia, The Free Encyclopedia. Retrieved 02:41, July 22, 2019, from https://en.Wikipedia.org/w/index.php?title=DNA&oldid=905364161

Wikipedia contributors. (2019, July 22). Chromosome. In Wikipedia, The Free Encyclopedia. Retrieved 15:18, July 23, 2019, from en.Wikipedia.org/w/index.php?title=Chromosome&oldid=907355235

Wikilectures. Prokaryotic Chromosomes (2017) In MediaWiki, Available at: https://www.wikilectures.eu/w/Prokaryotic_Chromosomes

Wikipedia contributors. (2019, May 15). DNA supercoil. In Wikipedia, The Free Encyclopedia. Retrieved 19:40, July 25, 2019, from en.Wikipedia.org/w/index.php?title=DNA_supercoil&oldid=897160342

Wikipedia contributors. (2019, July 23). Histone. In Wikipedia, The Free Encyclopedia. Retrieved 16:19, July 26, 2019, from en.Wikipedia.org/w/index.php?title=Histone&oldid=907472227

Wikipedia contributors. (2019, July 17). Nucleosome. In Wikipedia, The Free Encyclopedia. Retrieved 17:17, July 26, 2019, from en.Wikipedia.org/w/index.php?title=Nucleosome&oldid=906654745

Wikipedia contributors. (2019, July 26). Human genome. In Wikipedia, The Free Encyclopedia. Retrieved 06:12, July 27, 2019, from en.Wikipedia.org/w/index.php?title=Human_genome&oldid=908031878

Wikipedia contributors. (2019, July 19). Gene structure. In Wikipedia, The Free Encyclopedia. Retrieved 06:16, July 27, 2019, from en.Wikipedia.org/w/index.php?title=Gene_structure&oldid=906938498