Codex @codex  0

Production of different mature RNA sequences from the same precursor by changing which splice sites or exons are used. A cassette exon can be included in one transcript isoform and skipped in another.

A segment retained in the mature RNA after RNA splicing, or the corresponding segment of a gene. An exon need not encode a protein: the ribosomal RNA exons in the Tetrahymena ribozyme experiment are an example.

Coordinated recognition of the boundaries of an exon, helping specify splice-site pairing. Disrupting communication can repress exon use without directly occupying one essential site.

Sequencing designed to span long portions of individual RNA molecules or derived Complementary DNA. Linking distant exons on one molecule helps resolve complete transcript isoforms, while per-base errors and sampling depth still require control.

A very short exon. Specialized alignment and junction validation can be needed to detect its inclusion; the exact length cutoff should be stated for a given analysis.

A member of a neural RNA-binding protein family with position-dependent effects on alternative splicing. Binding-site position helps predict activation or repression of exon inclusion.^[ref]

A guide specifies the target; the attached Cas9 machinery determines whether expression changes through DNA repair or direct transcriptional regulation. In CRISPR genome editing, a guide RNA supplies sequence complementarity to a genomic target adjacent to a suitable protospacer adjacent motif. The commonly used Cas9 from Streptococcus pyogenes recognizes an NGG motif, where N is any base. Recognition involves both that motif and pairing between the guide and target DNA; an arbitrary complementary sequence is therefore insufficient. In mammalian experiments, a nuclear-localized Cas9 and guide can be delivered by expression constructs or as a preassembled ribonucleoprotein. Multiplexing means supplying several guides, rather than engineering a new DNA-binding protein for every target. The original mammalian demonstration combined programmable targeting with endogenous repair to change genomic sequences.^[ref]

For a loss-of-function experiment, a cut in an early constitutive exon is repaired by non-homologous end joining and related end-joining pathways. Insertions or deletions can change the reading frame, introduce a premature stop, and trigger nonsense-mediated decay. Neither cleavage nor an insertion/deletion automatically proves a gene knockout: repair can preserve the reading frame, alternative splicing can bypass the exon, and a mammalian cell may retain an intact allele. Targeting a promoter or enhancer can instead alter the amount or cell specificity of expression. Two guides can delete a defined regulatory segment, but the resulting junction and possible larger rearrangements need verification. homology-directed repair with a donor can introduce an expression tag, repair a mutation, or replace a regulatory sequence; its efficiency depends on cell state and competing repair pathways. Changing the genomic coding sequence can change protein activity without changing transcript abundance, so these outcomes should be distinguished.

Expression can also be controlled without cleavage. Catalytically inactive dCas9 retains guide-directed binding. Positioning it at an appropriate regulatory region can obstruct transcription; recruiting a repressive domain such as KRAB strengthens CRISPR interference in mammalian cells. Fusing an activation domain, or recruiting several activators, gives CRISPR activation at an endogenous promoter. These approaches preserve the coding sequence and permit graded or inducible perturbation, although chromatin context and guide position strongly affect performance. Mammalian experiments comparing regulatory fusions established that the same targeting platform can repress or activate transcription.^[ref]

The readout should match the proposed mechanism: sequence the target for editing, measure messenger RNA by RT-qPCR, and measure the protein and its biological activity. Use multiple independent guides, a nontargeting guide, delivery-matched controls, and several independently derived clones or a controlled pooled experiment. Rescue with a guide-resistant functional gene supports causality; overexpression rescue alone can have dosage artefacts. Check relevant predicted off-target effects, assess cell viability, and distinguish direct expression changes from selection of surviving cell states. A nuclease perturbation can itself activate a DNA-damage response. Thus the strongest conclusion connects a verified molecular perturbation to RNA, protein and phenotype, rather than treating a guide-dependent phenotype as sufficient evidence.

Shared preference for pyrimidine-rich RNA does not prescribe one splicing outcome: position, multivalent binding and interactions with the spliceosome determine the regulated step. Sex-lethal acts in the Drosophila sex-determination programme, whereas PTBP1 regulates many mammalian precursor RNAs in a cell-type-dependent manner. Their effects must be interpreted against competing splice sites and the proteins normally recognizing them.

During early spliceosome assembly, U1 snRNP recognizes a 5′ splice site, while factors including U2AF help recognize a polypyrimidine tract and 3′ splice-site region and promote assembly around the branch point. Alternative sites compete within this context. Binding of another RNA-binding protein can occlude an essential interaction, organize a repressive complex or favour use of a competing site; recognition need not be determined solely by an isolated site's consensus strength.

A particularly clear SXL mechanism operates on precursor RNA encoding Transformer protein. In the absence of SXL, U2AF recognizes a strong default 3′ site, producing an unproductive transcript. In female biological cells, SXL binds the pyrimidine-rich region at that site and competes with U2AF. The weaker alternative female site can then be used, producing functional Transformer protein. Reconstituted splicing experiments showed the antagonism, and adding a U2AF effector domain to SXL converted its activity from repression toward activation.^[ref] The domain-fusion result is important: related RNA recognition can direct different outcomes because the proteins recruit different downstream activities. SXL need not directly activate the female acceptor to shift usage toward it; repression of the stronger competitor is sufficient in this substrate.

SXL also maintains its own productive expression through exclusion of a male-specific exon containing a premature stop. This provides positive feedback in female biological cells. It is not adequately described as another instance of simple U2AF displacement. Work on the SXL autoregulatory substrate identified a specialized arrangement of AG dinucleotides and a pyrimidine tract: U2AF and SPF45 recognize different features, and interaction with SXL inhibits SPF45-dependent progression at the second catalytic step.^[ref] SXL can consequently alter site selection after early assembly has occurred. Accumulation of a first-step intermediate in an appropriate biochemical assay distinguishes such a block from failure to recruit the spliceosome initially. The same regulator can use different mechanisms on different RNAs.

PTBP1 has four RNA recognition motifs and commonly binds CU-rich sequences at more than one position. Structural work showed distinct RNA-recognition arrangements and an organization of its third and fourth domains compatible with bridging separated RNA segments.^[ref] Förster resonance energy transfer, NMR and splicing assays subsequently supported RNA looping by these domains: contacts on either side of a segment can organize an RNA geometry that disfavors exon use.^[ref] These observations support a physical repression mechanism, but a short model RNA loop does not prove that every endogenous PTBP1-regulated exon is looped out.

Other PTBP1 arrangements can interfere with early assembly or communication between sites without directly replacing U2AF at the same tract. Binding near or within an alternative exon can impair its definition, whereas in another arrangement PTBP1 can favour exon inclusion, including by suppressing a competing event. Genome-wide perturbation and positional analyses revealed both activation and repression associated with different binding contexts.^[ref] Thus neither “all pyrimidine binders are repressors” nor “PTBP1 always causes skipping” is a valid general rule. protein abundance, RNA architecture, neighbouring factors and related PTBP family members change the effective regulatory complex.

To discriminate mechanisms, map direct contacts by crosslinking and immunoprecipitation, measure exon-junction usage after acute depletion, and rescue with wild-type or selected RNA-binding/interface mutants. Use a minigene splicing reporter to mutate individual tracts, change their spacing or move them relative to a regulated exon, then confirm important effects at the endogenous locus. Assays of U2AF or U1 snRNP association and native spliceosomal assembly can reveal an early block; time-resolved splicing-intermediate analysis can reveal a late one. Control reporter transcription and precursor abundance, and account for nonsense-mediated decay, because an unstable mature isoform can conceal the splice choice that generated it. Mutants must retain expression and the intended remaining contacts. SXL supplies examples of both competition and late catalytic control; PTBP1 illustrates multivalent, context-dependent organization of splice-site recognition. RNA affinity is only one part of specificity.

A convincing splicing code must predict alternative splicing from both the RNA sequence and the cellular context, then survive causal perturbation. A list of exons specific to neurons or enriched sequence motifs is only a starting point. In the central nervous system, distinguish neurons from glial cells, individual neuron classes, anatomical regions, developmental stages and activity states. Otherwise a bulk difference can arise because two samples contain different proportions of biological cells, even when no individual cell type changes its RNA splicing.

Start with replicated, well-matched samples from several central nervous system regions and non-neural comparison biological tissues, recording donor, age, sex, preparation batch and RNA integrity. Obtain deep, stranded RNA sequencing with reads spanning splice junctions; combine this with long-read RNA sequencing to establish complete transcript isoforms. Use purified cell types or sufficiently deep, full-length single-cell RNA sequencing to separate cellular composition from regulation. Routine assays that sequence only the $3^{\prime }$ end cannot measure most internal exons; single-nucleus RNA sequencing additionally enriches incompletely processed RNA, so retained introns require careful interpretation. A matched bulk or purified-cell dataset supplies coverage that sparse individual-cell observations lack.

Align reads with a splice-aware sequence alignment method and quantify cassette exons, alternative splice sites, mutually exclusive exons and retained introns. Search explicitly for microexons, whose short length can defeat standard alignment filters. For a cassette exon, summarize inclusion by percent spliced in, written as a fractionwhere $I$ and $S$ estimate abundances of included and skipped molecules after correcting the different informative read opportunities. Adding the two inclusion-junction counts and comparing their sum directly with a single skipping-junction count would bias this estimate. Report uncertainty and coverage, and model between-donor variation; select changes by both effect size and false discovery rate. Validate selected splice junctions using RT-PCR and Sanger sequencing. Distinguish genuine altered processing from differential destruction of transcript isoforms through nonsense-mediated decay, using newly synthesized or nuclear RNA and a short, controlled perturbation of that pathway where needed.

Build a feature set around each regulated exon: strengths of the two splice sites, RNA splicing branch point and polypyrimidine tract; exon and intron lengths; position-specific cis-regulatory elements; predicted or measured RNA secondary structure; and evolutionary conservation. Include the abundance, protein level, nuclear location and activity of candidate RNA-binding proteins, rather than treating their messenger RNA abundance as a perfect proxy. A splicing code can combine many weak sequence features: the original tissue-prediction study demonstrated the value of combinations rather than one universal neural motif.^[ref]

Fit an interpretable statistical model first, allowing interactions between motif position and RNA-binding protein abundance. Compare it with more flexible machine learning, using cross-validation that holds out whole genes, donors and, for a strong transfer test, a cell type. Randomly separating junctions from the same gene between training and testing risks information leakage. Compare against sequence-only, context-only and expression-matched baseline predictors. Successful prediction of held-out percent spliced in is more informative than classifying the training set. In particular, an enriched motif can reflect nucleotide composition, intron length or another factor with similar specificity; enrichment alone does not establish binding or function.

Map candidate RNA-binding protein occupancy with crosslinking and immunoprecipitation in the relevant preparations of neurons. Include input material, a no-crosslink control where appropriate, a validated antibody, and biological replication; account for transcript abundance and crosslinking preferences. Integrate binding positions with RNA sequencing after acute depletion. NOVA proteins give a concrete precedent: binding to YCAY clusters has position-dependent effects, with exonic binding able to inhibit inclusion and downstream intronic binding able to promote it. Here Y denotes a pyrimidine nucleobase. Thus the same motif can support opposite predictions at different positions.^[ref] Test other neural regulators, such as RBFOX proteins and PTBP proteins, as potentially overlapping or antagonistic inputs rather than assuming every neural exon follows the NOVA protein rule.

Finally, make the predictions experimentally vulnerable. For representative exons, mutate or move individual cis-regulatory elements in a minigene splicing reporter, preserving splice sites and controlling total reporter abundance. Test gain and loss of inclusion in both neural and non-neural biological cells. Edit the endogenous element because a short minigene splicing reporter can omit distant regulation within introns or chromatin context. Acutely deplete the proposed RNA-binding protein and rescue with physiological amounts of normal protein or an RNA-binding-defective mutant. Combined depletion tests redundancy; motif mutation plus factor depletion tests whether both act through the same element. A particularly strong result is the predicted reversal after moving a binding site from a repressive to an activating position. Measure cell differentiation and viability so that a change in cellular identity is not mistaken for direct RNA splicing regulation. The resulting code is a tested, context-dependent set of regulatory rules, with an explicit domain of validity.

The inclusion fraction of an alternative exon or event. With effective included and skipped molecule abundances $I$ and $S$, $\psi =I/(I+S)$; expressing this fraction as a percentage gives $100\psi$. Raw junction counts can require opportunity or length correction before substitution.

A ribozyme is an RNA molecule whose folded structure provides catalysis of a chemical reaction. A protein enzyme builds its active site from an amino acid chain; a ribozyme builds a working active site from RNA, sometimes with metal ions and associated proteins. Base pairing can help recognize a substrate, but recognition alone does not establish catalysis: the RNA must promote chemical conversion. A passive messenger RNA template being read by a ribosome is not thereby a ribozyme. See Cech and Herschlag, Group I Ribozymes: Substrate Recognition, Catalytic Strategies, and Comparative Mechanistic Analysis, sections 1–3.

The surprise is that an information-bearing polymer can also perform chemistry. Examples include the self-splicing group I intron of Tetrahymena thermophila and the catalytic RNA component of bacterial RNase P, discussed below. The ribosome supplies an everyday example: its peptide bond-forming active site is made of ribosomal RNA, although the complete machine also contains proteins. Nissen and colleagues examined structures of a large ribosome subunit with substrate analogues and found RNA contacting those analogues at the catalytic centre. This supports an RNA active site, not a claim that isolated ribosomal RNA can carry out all of biological translation. See Nissen, Hansen, Ban, Moore and Steitz, The structural basis of ribosome activity in peptide bond synthesis, Science 289 (2000), 920–930.

For the Tetrahymena group I intron, RNA splicing removes the intron and joins its flanking exons. The first chemical step uses the free $3^{\prime }$ hydroxyl group of an external guanosine to attack the phosphodiester bond at the $5^{\prime }$ splice junction. This releases the upstream exon with a free $3^{\prime }$ hydroxyl group and attaches the guanosine to the intron. That newly freed hydroxyl group then attacks the phosphodiester bond at the other splice junction, joining the exons and releasing the linear intron. Each step exchanges one backbone connection for another; it is not simply separation of base pairs. See Cech, Zaug and Grabowski, 1981: the phosphoester-transfer mechanism.

The motivating experiment began with an unexpected control. As Cech recounts, Zaug purified a ribosomal RNA precursor from Tetrahymena using a detergent and phenol extraction, then compared test tubes with and without added cell nucleus extract. Incubation in salts and nucleotides released the same amount of small intron RNA in both cases, detected by gel electrophoresis. Repetition made a missing ingredient or accidental omission less plausible. Activity also survived treatments intended to destroy proteins. Nevertheless, absence of added extract did not by itself exclude a tightly attached, resistant protein, or prove that the small RNA had been newly excised rather than released from an already processed complex. See Thomas R. Cech, 1989 Nobel lecture, pp. 655–659: the unexpected control and alternative explanations.

An independent chemical readout addressed the latter problem. Cech mixed precursor RNA with phosphorus-32-labelled guanosine bearing three phosphate groups (GTP), and used gel electrophoresis followed by exposure to X-ray film to locate the label. Label appeared in the excised intron with magnesium present; it was absent when magnesium was omitted or labelled ATP replaced the labelled guanosine reagent. This experiment is described in Cech, 1989 Nobel lecture, pp. 657–658: labelled-guanosine controls. The 1981 study established attachment through the guanosine $3^{\prime }$ hydroxyl group and showed that guanosine without a triphosphate also supported excision: breaking down the added nucleotide to supply energy was unnecessary. See Cech, Zaug and Grabowski, In vitro splicing of the ribosomal RNA precursor of Tetrahymena, Cell 27 (1981), 487–496: primary abstract. Separately, sequencing and terminal analysis found an extra $5^{\prime }$ guanosine absent from the corresponding DNA sequence. See Zaug and Cech, The intervening sequence excised from the ribosomal RNA precursor of Tetrahymena contains a 5-terminal guanosine residue not encoded by the DNA, Nucleic Acids Research 10 (1982), 2823–2838. Together, incorporation and sequence evidence support a new covalent bond, beyond a change in how RNA associates.

Kruger, Grabowski, Zaug, Sands, Gottschling and Cech's 1982 test removed dependence on a precursor purified from Tetrahymena. They cloned the intron and adjacent exons into a DNA template, made precursor RNA with purified bacterial RNA polymerase under conditions suppressing RNA splicing, and removed proteins by detergent–phenol extraction. Denaturing gel electrophoresis separated precursor species. Purified precursor was then incubated with the guanosine reagent, magnesium and salt; matched samples without this incubation supplied a precursor baseline. Their standard assay used $10\mathrm{mM}$ magnesium chloride, $0.1\mathrm{mM}$ of that guanosine reagent, $100\mathrm{mM}$ ammonium sulfate and a buffer solution at pH $7.5$, for $30$ minutes at temperature $37{}^{\circ }\mathrm{C}$, with specified variations in individual experiments. Gel electrophoresis and radioactive-film detection showed excised intron RNA; sequencing verified its $5^{\prime }$ junction. See Kruger et al., Self-Splicing RNA, Cell 31 (1982), 147–157, figures 3–4 and Experimental Procedures.

Excision alone could leave disconnected exons. To test joining, the authors paired labelled RNA with DNA spanning the joined-exon sequence, removed exposed RNA with enzymes, and separated the protected pieces by denaturing gel electrophoresis. Spliced samples yielded protected RNA long enough to span the junction. See Kruger et al., 1982, figure 6: independent evidence for exon ligation. The experimental logic matters: an intron-sized band identifies a candidate product, while sequence and junction assays test its chemical identity.

These controls support sufficiency of the prepared RNA under the assay conditions. They do not show that every RNA self-splices, or that proteins are dispensable inside a biological cell. Many group I introns use proteins to help attain an active fold. See Cech and Herschlag, Group I Ribozymes, section 6: protein facilitation. Nor does a failed assay necessarily identify an inactive nucleotide sequence: unsuitable ions, pH, or temperature can prevent a working active site from forming.

The distinction between self-processing and repeated use became especially clear with bacterial RNase P. In 1983 Guerrier-Takada, Gardiner, Marsh, Pace and Altman separated the RNA and protein components of RNase P from two bacteria. Purified RNA cleaved precursor transfer RNA under conditions containing $60\mathrm{mM}$ magnesium ions, or $10\mathrm{mM}$ magnesium ions plus $1\mathrm{mM}$ spermidine. The protein components alone had no catalytic activity. At $5$–$10\mathrm{mM}$ magnesium without spermidine, both components were needed. Thus the comparison demonstrated an RNA catalyst while also explaining why protein assistance could matter under other conditions. RNase P produces the mature $5^{\prime }$ end of transfer RNA by removing its leader; it acts on a separate substrate and can be used again. See Guerrier-Takada et al., The RNA moiety of ribonuclease P is the catalytic subunit of the enzyme, Cell 35 (1983), 849–857: primary abstract.

Self-splicing demonstrates RNA chemistry, but one precursor completing one rearrangement is not yet evidence for catalytic turnover: repeated conversion of separate substrates by a regenerated catalyst. The self-splicing precursor changes its own connectivity and incorporates external guanosine. In 1986 Zaug and Cech showed that a shortened Tetrahymena intron RNA could cleave and rejoin separate short RNA substrates in a sequence-dependent reaction. This demonstrated an RNA enzyme acting beyond its own self-processing. See Zaug and Cech, The intervening sequence RNA of Tetrahymena is an enzyme, Science 231 (1986), 470–475: primary abstract.

Read the second figure as a counting argument. In a stipulated model, $A$ denotes unprocessed precursor and $P_{\mathrm{self}}$ counts completed self-processing events. Set $A(0)=R_0$, $P_{\mathrm{self}}(0)=0$ and $dA/d\tau =-A$. The exponential function solution gives $P_{\mathrm{self}}/R_0=1-e^{-\tau }$, never exceeding one. Here guanosine and other necessary reagents are taken to be available in excess; $P_{\mathrm{self}}$ counts events, not every separate reaction product molecule.

For a second stipulated model, let an unchanged ribozyme $R$ perform $R+S⟶R+P$ on a separate substrate $S$. Assume the effective law of mass action $dS/dt=-kRS$, no reverse reaction, no ribozyme loss and no inhibited or long-lived bound states. Set $R=R_0$, $S(0)=5R_0$, $P(0)=0$ and $\tau =k{R}_{0}t$. Then $S/R_0=5e^{-\tau }$ and $P/R_0=5(1-e^{-\tau })$, with $S+P=5R_0$. At $\tau =1$, self-processing gives about $0.632$ completed events per initial RNA, whereas the regenerated ribozyme gives about $3.161$ reaction product molecules per initial ribozyme. More than one substrate conversion per initial ribozyme, with appropriate background and identity controls, is evidence for turnover. Product accumulation below that threshold alone cannot distinguish a reusable catalyst from one-shot chemistry. These models illustrate that distinction; actual enzyme kinetics can include folding delays, binding, saturation and product release.

Processing of a precursor RNA that removes an intron and joins the flanking exons. The ribozyme experiments show one self-splicing mechanism; RNA splicing need not always be autonomous.

The junction between segments joined during RNA splicing. A read crossing an exon–exon junction can distinguish a processed transcript from the corresponding unspliced precursor.

A predictive set of rules connecting RNA sequence features and cellular regulators to alternative splicing outcomes. A useful code predicts the extent of exon inclusion in a specified cell type, rather than merely listing enriched motifs.