Science Wiki

RNA Splicing is the essential post-transcriptional and co-transcriptional molecular process in eukaryotes (and select archaea and bacteriophages) by which non-coding intervening sequences called introns are precisely cleaved from a primary precursor transcript (pre-mRNA, pre-tRNA, or pre-rRNA) and the flanking expressed sequences called exons are covalently joined together to generate a mature, functional RNA molecule. Discovered independently in 1977 by Phillip Sharp and Richard Roberts through electron microscopy of adenovirus hexon transcripts hybridized to genomic DNA (which revealed single-stranded intronic loops or "R-loops"), nuclear pre-mRNA splicing is catalyzed by a dynamic megadalton ribonucleoprotein machine called the spliceosome. Because a single nucleotide frameshift error during splicing destroys the downstream open reading frame prior to translation at the ribosome, the splicing machinery achieves single-bond accuracy across introns ranging from 50 nucleotides to over 500,000 nucleotides in length.

In human protein-coding genes, exons average only ~150 nucleotides in length, whereas introns average ~3,500 to 5,500 nucleotides, meaning that more than 90% of a typical primary transcript synthesized by RNA Polymerase II must be removed. The spliceosome identifies exon–intron junctions through four conserved *cis*-acting consensus sequence motifs located within and adjacent to each intron:

1. The 5' Splice Site (Donor Site): Located at the 5' end of the intron, defined by an almost invariant GU dinucleotide (in >98% of U2-type introns) embedded within a broader consensus sequence MAG|GURAGU (where M = A or C, R = purine A or G, and | denotes the exon–intron cleavage bond).
2. The Branch Point Site (BPS): Situated 18 to 40 nucleotides upstream of the 3' end of the intron, containing a strictly conserved branch point adenosine (A) residue embedded within the consensus motif YNYURAY in mammals (or the strictly conserved UACUAAC box in the budding yeast *Saccharomyces cerevisiae*).
3. The Polypyrimidine Tract (PPT): A stretch of 15 to 20 pyrimidine nucleotides (uracil and cytosine, Y_15–20) positioned immediately between the branch point adenosine and the 3' splice site that recruits early splicing assembly factors.
4. The 3' Splice Site (Acceptor Site): Located at the 3' end of the intron, defined by an invariant AG dinucleotide preceded by a pyrimidine (YAG|G), marking the exact phosphodiester bond cleaved during exon ligation.

Together, the invariant dinucleotides at the intron boundaries establish the universal GU–AG rule (also known as Chambon's rule) for canonical U2-dependent (major) introns, while a rare subset (~0.5%) of U12-dependent (minor) introns utilizes either AT–AC or GU–AG boundaries with a distinct branch site consensus.

Chemically, the excision of an intron and ligation of two exons requires zero net input of adenosine triphosphate (ATP) for the bond-cleavage and bond-formation steps themselves. Instead, RNA splicing proceeds via two sequential S_N2-like nucleophilic transesterification reactions in which two phosphodiester bonds are broken and two new phosphodiester bonds are formed, keeping the total number of high-energy phosphate bonds constant (ΔG° ≈ 0 kJ/mol):

• Step 1 (Branching / Lariat Formation): The 2'-hydroxyl (2'-OH) group on the ribose ring of the conserved branch point adenosine acts as an internal nucleophile. It attacks the phosphorus atom of the phosphodiester bond at the 5' splice site (between the upstream Exon 1 and the first guanine G of the intron). This in-line nucleophilic attack cleaves Exon 1 (leaving a free 3'-OH group on its terminal nucleotide) and covalently links the 5' phosphate of the intronic G to the 2'-OH of the branch point adenosine via an unusual 2',5'-phosphodiester bond. Because the branch point adenosine is already linked to its neighbors via standard 3',5'-phosphodiester bonds, it forms a three-way branched junction, turning the intron into a looped lasso-like structure called an intron lariat intermediate.
• Step 2 (Exon Ligation): The newly liberated 3'-hydroxyl (3'-OH) group at the 3' end of upstream Exon 1 acts as the second nucleophile. Positioned inside the catalytic core, it attacks the phosphorus atom at the 3' splice site (immediately downstream of the conserved intronic AG dinucleotide). This reaction joins Exon 1 and Exon 2 via a standard 3',5'-phosphodiester bond and releases the excised intron lariat with a free 3'-OH tail.

Following release, the excised intron lariat is linearized by the lariat debranching enzyme (DBR1)—which specifically hydrolyzes the 2',5'-phosphodiester bond—and rapidly degraded into ribonucleotide monomers by nuclear exoribonucleases (or occasionally processed into non-coding snoRNAs or miRNAs).

Across living systems, four mechanistically distinct classes of RNA splicing operate, revealing a remarkable evolutionary trajectory from ancient catalytic RNAs (ribozymes) of the RNA world to modern protein-assisted ribonucleoprotein machines:

Spliceosome Assembly and the Two-Metal-Ion Active Site

Although transesterification is isoenergetic, the eukaryotic spliceosome consumes substantial ATP to power DExD/H-box RNA helicases (such as Prp5, Sub2, Prp28, Brr2, Prp2, Prp16, Prp22, and Prp43) that assemble, proofread, and disassemble five small nuclear ribonucleoproteins (snRNPs): U1, U2, U4, U5, and U6:

1. Complex E (Commitment): U1 snRNP base-pairs directly with the 5' splice site via its U1 snRNA, while SF1 (BBP) binds the branch point and U2AF (heterodimer of U2AF65 and U2AF35) binds the polypyrimidine tract and 3' AG dinucleotide.
2. Complex A (Pre-Spliceosome): U2 snRNP displaces SF1 and base-pairs with the branch point consensus, forcing the branch point adenosine to bulge out unpaired from the U2–pre-mRNA helix, priming its 2'-OH for nucleophilic attack.
3. Complex B to B^act (Activation): The pre-assembled U4/U6.U5 tri-snRNP joins the complex. The Brr2 helicase unwinds the extensive base-pairing between U4 and U6 snRNAs, expelling U1 and U4. Freed U6 snRNA base-pairs with both the 5' splice site (replacing U1) and U2 snRNA to form an intramolecular stem-loop (ISL).
4. Complex B* and C (Catalysis): High-resolution cryo-electron microscopy (cryo-EM) structures by Yigong Shi and Reinhard Lührmann proved that the spliceosome is a protein-directed metalloribozyme. The folded U2/U6 snRNA duplex coordinates two catalytic magnesium ions (Mg²⁺_A and Mg²⁺_B) spaced ~3.9 Å apart: Mg²⁺_A activates the attacking 2'-OH (Step 1) or 3'-OH (Step 2) nucleophile, while Mg²⁺_B stabilizes the negatively charged pentavalent phosphorane transition state and leaving group, while U5 snRNA Loop 1 tethers Exon 1 and Exon 2 in alignment.

Group I and Group II Self-Splicing Introns

Discovered by Thomas Cech in 1982 in the 26S rRNA precursor of the ciliated protozoan *Tetrahymena thermophila*, Group I introns fold into a conserved three-dimensional pocket that binds a free guanosine nucleotide (GTP, GDP, or GMP). The 3'-OH of this exogenous guanosine attacks the 5' splice site in Step 1, attaching the G to the 5' end of the intron; Exon 1's 3'-OH then attacks the 3' splice site in Step 2.

In contrast, Group II introns—found in bacteria, archaea, and organellar genomes of mitochondria and chloroplasts—use the exact same bulged adenosine 2'-OH lariat mechanism and two-Mg²⁺ catalytic triad (located in Domain V) as the nuclear spliceosome. Evolutionary biologists universally recognize that nuclear spliceosomal introns and the five snRNAs (U1–U6) evolved during eukaryogenesis when ancestral mobile Group II introns invaded the proto-eukaryotic genome from the α-proteobacterial mitochondrial endosymbiont and fragmented *in trans* into separate catalytic snRNAs and substrate introns.

In living cells, RNA splicing does not wait for transcription to finish. Over 80% of splicing events occur co-transcriptionally while the nascent pre-mRNA is still emerging from the exit channel of RNA Polymerase II. The phosphorylated carboxy-terminal domain (CTD) of the largest Pol II subunit (RPB1)—composed of 52 tandem heptapeptide repeats (YSPTSPS)—directly binds and loads U1 snRNP and U2AF onto nascent splice sites as soon as they are synthesized.

Following exon ligation, the spliceosome deposits a multiprotein Exon Junction Complex (EJC) ~20 to 24 nucleotides upstream of every newly joined exon–exon boundary. The EJC accompanies the mature mRNA through the nuclear pore complex into the cytoplasm and plays a vital role in Nonsense-Mediated mRNA Decay (NMD): if a translating ribosome encounters a premature stop codon (often caused by a splicing mutation or retained intron) located more than 50–55 nucleotides upstream of an EJC, UPF1–UPF3 surveillance factors trigger rapid degradation of the defective transcript before truncated toxic proteins can accumulate. Moreover, through alternative splicing, regulatory proteins can selectively include or skip exons to produce multiple distinct protein isoforms from a single gene.

Why is the spliceosome considered a ribozyme if it contains over 150 proteins?

For decades, researchers debated whether spliceosomal catalysis was performed by protein subunits (such as Prp8) or by the snRNAs. Atomic-resolution cryo-EM structures combined with sulfur-substitution metal-rescue experiments proved that the catalytic active site consists exclusively of U6 and U2 snRNA phosphate oxygens coordinating two catalytic Mg²⁺ ions—architecturally identical to Domain V of Group II self-splicing ribozymes—while Prp8 and other proteins act solely as a structural scaffold.

If transesterification requires zero net energy, why does the spliceosome hydrolyze so much ATP?

Each of the two transesterification steps breaks one phosphodiester bond and forms one phosphodiester bond, making the chemistry isoenergetic (ΔG° ≈ 0). However, the spliceosome consumes 8 or more ATP molecules per intron to power DExD/H-box RNA helicases. These ATPases drive unidirectional conformational rearrangements and perform kinetic proofreading, ejecting suboptimal or mispaired splice sites before irreversible cleavage can occur.

How do mutations in RNA splicing cause human genetic diseases?

Approximately 15% to 35% of all disease-causing single-nucleotide mutations in humans disrupt splicing rather than directly altering amino acid codons. Mutations that destroy the invariant GU or AG dinucleotides (such as in β-thalassemia) or weaken exonic splicing enhancers—such as in Spinal Muscular Atrophy (SMA), where loss of *SMN1* and skipping of exon 7 in *SMN2* causes motor neuron degeneration—lead to exon skipping, intron retention, and loss of functional protein.