Frameshift Mutation

A frameshift mutation is a class of genetic mutation caused by the insertion or deletion (indel) of a number of nucleotide base pairs that is not evenly divisible by three (Δn ≢ 0 mod 3, such as ±1 bp, ±2 bp, ±4 bp, or ±5 bp) within a protein-coding open reading frame (ORF) or across an exon-intron splice site. Because ribosomes translate messenger RNA (mRNA) in a sequential, unpunctuated, non-overlapping series of three-nucleotide codons anchored by the 5'-AUG-3' translation initiation start codon, any indel whose length is not a multiple of three shifts the translational reading frame of every subsequent triplet downstream of the mutation site. Unlike a single-nucleotide substitution (point mutation) or an in-frame 3n indel—which alters or inserts/deletes a single amino acid while leaving the downstream polypeptide sequence intact—a frameshift mutation replaces the downstream amino acid sequence with an unrelated frame-shifted peptide and commonly encounters an out-of-frame premature termination codon (PTC: UAA, UAG, or UGA), triggering transcript degradation via nonsense-mediated mRNA decay (NMD) or synthesis of a truncated polypeptide. Exploited experimentally in 1961 by Francis Crick, Sydney Brenner, Leslie Barnett, and R. J. Watts-Tobin to prove the triplet nature of the genetic code, frameshift mutations arise from polymerase slippage during DNA replication and defective DNA repair.

Phase Shift Arithmetic (+1 and −1 Reading Frames)

Within any mature protein-coding mRNA, a nucleotide sequence contains three potential reading frames (Frame 0, Frame +1, and Frame +2 / −1). The 43S pre-initiation complex scans from the 5' m7G cap until it recognizes the start codon AUG (encoding methionine) within a Kozak sequence (5'-gccRccAUGG-3'), locking the ribosome into the canonical 0 reading frame (3 nucleotides = 1 codon per translocation cycle). When an indel of length k nucleotides occurs within the coding region, it shifts the downstream reading phase by k mod 3:

• −1 Frameshift (or +2 Insertion, e.g., −1 bp, +2 bp, −4 bp): Shifts one nucleotide backward (or two forward), grouping the 3rd base of codon i with the 1st and 2nd bases of codon i + 1.
• +1 Frameshift (or −2 Deletion, e.g., +1 bp, −2 bp, +4 bp): Shifts one nucleotide forward (or two backward), grouping the 2nd and 3rd bases of codon i with the 1st base of codon i + 1.
• In-Frame Indel (k ≡ 0 mod 3, e.g., ±3 bp, ±6 bp): Adds or removes an integral number of amino acids (as in c.1521_1523delCTT removing Phe508 in CFTR) while preserving the downstream reading frame.

Out-of-Frame Premature Termination Codons and Nonsense-Mediated Decay (NMD)

Because 3 of the 64 triplets in the standard genetic code are translation termination (stop) codons (UAA, UAG, UGA), random out-of-frame nucleotide sequences contain a stop codon at a statistical probability of roughly 3 / 64 ≈ 1 in 21.3 codons. Consequently, a frameshift mutation typically causes the translating ribosome to stall at an out-of-frame Premature Termination Codon (PTC) within 15 to 30 codons downstream of the indel.

In mammalian cells, the fate of an mRNA harboring a frameshift-induced PTC is governed by the 50–55 nucleotide rule of Nonsense-Mediated mRNA Decay (NMD):

1. PTC Upstream of the Last Exon Junction (>50–55 nt 5' to the final boundary): During splicing, the spliceosome deposits an Exon Junction Complex (EJC: eIF4A3, MAGOH, Y14, MLN51) 20–24 nt upstream of each exon-exon junction. When a frameshift introduces a PTC >50–55 nt upstream of the final junction, terminating release factors (eRF1–eRF3) interact with the retained downstream EJC via UPF1, UPF2, and UPF3B, activating SMG1 and endoribonuclease SMG6 to degrade the mRNA, producing a loss-of-function null allele.
2. PTC in the Final Exon (or within 50 nt of the final junction): No downstream EJC remains bound when the ribosome reaches the PTC; the transcript escapes NMD and translates into a truncated polypeptide with a frameshifted tail that can exert dominant-negative or gain-of-function effects.

Acridine Mutagenesis in Bacteriophage T4 rII and Intragenic Suppression

In December 1961, Francis Crick, Sydney Brenner, Leslie Barnett, and R. J. Watts-Tobin published "General nature of the genetic code for proteins" in Nature, using frameshift mutants in the rII (rIIB cistron) locus of bacteriophage T4 to prove deductively that the genetic code is read in non-overlapping, comma-free triplets from a fixed starting point. Wild-type T4 phage grows and forms plaques on Escherichia coli strain K12(λ), whereas rII mutants cannot grow on E. coli K12(λ), providing positive selection for functional revertants. Crick and Brenner mutagenized phage T4 with proflavin (an acridine dye that intercalates into duplex DNA and induces single-base +1 insertions or −1 deletions), starting with an rIIB mutant designated FC0 (assigned the phase sign +):

• Double Mutants (+ paired with −): Plating FC0 (+) on E. coli K12(λ) selected pseudo-wild-type revertants that had acquired a nearby intragenic suppressor mutation of opposite sign (FC1 (−), FC7 (−)). Combining a + insertion with a nearby − deletion (+ −) restored wild-type function because only the short interval between the two mutations was read out of frame.
• Pairs (+ + or − −) vs. Triples (+ + + and − − −): Combining any two mutations of the same sign (+ + or − −) remained nonfunctional. However, constructing triple mutants combining three insertions (+ + +) or three deletions (− − −) restored wild-type phage growth on E. coli K12(λ), proving (+1 + 1 + 1 = +3 ≡ 0 mod 3) that the genetic code is a degenerate, non-overlapping triplet code.

Streisinger Slipped-Strand Mispairing and Microsatellite Instability (MSI)

In 1966, George Streisinger proposed the slipped-strand mispairing (polymerase slippage) model to explain why spontaneous frameshift mutations cluster within homopolymeric nucleotide runs (e.g., AAAAAAAA) and microsatellites. During DNA replication, transient dissociation and re-annealing of the nascent strand against the template strand allows repeat units to loop out: a loop on the nascent primer strand causes a +1 insertion, whereas a loop on the template strand causes a −1 deletion. Normally, insertion/deletion loops are corrected by DNA mismatch repair (MMR) (MSH2–MSH6, MLH1–PMS2). In hereditary Lynch syndrome and sporadic MMR-deficient tumors, loss of MMR produces Microsatellite Instability (MSI-High), causing coding frameshifts in mononucleotide tracts of tumor suppressor genes such as TGFBR2 (A10 tract) and BAX (G8 tract).

Muscular Dystrophy Reading-Frame Rule and Programmed Ribosomal Frameshifting

In 1988, Anthony P. Monaco and Louis M. Kunkel formulated the reading-frame rule in the 79-exon DMD gene encoding dystrophin: multi-exon deletions that join exons with incompatible boundary phases (such as deleting exon 50, 109 bp) shift the reading frame, trigger NMD, eliminate dystrophin, and cause Duchenne muscular dystrophy (DMD), whereas even large multi-exon deletions that preserve matching exon phases (in-frame deletions) produce internally shortened dystrophin and milder Becker muscular dystrophy (BMD). Antisense oligonucleotide (ASO) exon-skipping drugs (eteplirsen skipping exon 51, golodirsen skipping exon 53, casimersen skipping exon 45) restore the reading frame (109 bp + 233 bp = 342 bp ≡ 0 mod 3) to convert Duchenne into Becker transcripts.

Conversely, retroviruses (HIV-1 gag-pol at 5'-U UUU UUA-3') and coronaviruses (SARS-CoV-2 ORF1a/1b at 5'-U UUA AAC-3') utilize Programmed Ribosomal Frameshifting (PRF), where a heptameric slippery sequence paired with a downstream RNA pseudoknot induces controlled −1 ribosomal slippage to synthesize viral polymerases in fixed stoichiometric ratios.

Why is a 3-base-pair insertion or deletion not considered a frameshift mutation?

Because the genetic code is translated in consecutive triplets (3 nucleotides = 1 codon), inserting or deleting any multiple of three nucleotides (±3 bp, ±6 bp, ±9 bp) adds or removes one, two, or three amino acids at the mutation site (and potentially changes a single boundary amino acid if the indel straddles two codons), but leaves the reading frame of every subsequent downstream codon unchanged. For example, the cystic fibrosis mutation c.1521_1523delCTT (F508del) deletes 3 base pairs across codons 507 and 508, removing a single phenylalanine residue while preserving the remaining 972 downstream amino acids of the 1,480-residue CFTR channel in their normal reading frame.

Why are frameshift mutations located near the 3' end of a gene sometimes more clinically harmful than those near the 5' end?

When a frameshift mutation occurs in an early (5') or middle exon, the resulting out-of-frame premature termination codon (PTC) lies more than 50–55 nucleotides upstream of the final exon-exon junction. Consequently, the Nonsense-Mediated mRNA Decay (NMD) pathway recognizes the retained downstream Exon Junction Complex (EJC) and degrades the mutant mRNA, producing a simple loss-of-function null allele (which is often benign in heterozygous carriers of haplosufficient genes). Conversely, a frameshift occurring in the final (3') exon places its PTC downstream of all EJCs, allowing the mutant mRNA to escape NMD and be translated into a truncated protein with an aberrant C-terminal peptide tail that can poison wild-type complexes (dominant-negative) or resist degradation (gain-of-function).

How does a splice-site point mutation cause a frameshift if no bases were inserted or deleted in the coding exons themselves?

A single-nucleotide point substitution in an invariant intronic splice donor (5'-GU) or splice acceptor (3'-AG) dinucleotide typically causes the spliceosome to skip the adjacent exon (or activate a cryptic intronic splice site). If the skipped exon contains a number of nucleotides not divisible by three (for example, skipping a 109-bp or 142-bp exon), joining the flanking upstream and downstream exons shifts the translational reading frame of all subsequent exons by +1 or +2 nucleotides, producing the same downstream frameshift and premature termination codon as an intragenic indel.

1. Crick, F. H. C., Barnett, L., Brenner, S., & Watts-Tobin, R. J. (1961). "General nature of the genetic code for proteins." Nature, 192(4809), 1227–1232. https://doi.org/10.1038/1921227a0
2. Streisinger, G., Okada, Y., Emrich, J., Newton, J., Tsugita, A., Terzaghi, E., & Inouye, M. (1966). "Frameshift mutations and the genetic code." Cold Spring Harbor Symposia on Quantitative Biology, 31, 77–84. https://doi.org/10.1101/SQB.1966.031.01.014
3. Monaco, A. P., Bertelson, C. J., Liechti-Gallati, S., Moser, H., & Kunkel, L. M. (1988). "An explanation for the phenotypic differences between patients bearing partial deletions of the DMD locus." Genomics, 2(1), 90–95. https://doi.org/10.1016/0888-7543(88)90113-990113-9)
4. Kurosaki, T., Popp, M. W., & Maquat, L. E. (2019). "Quality and quantity control of gene expression by nonsense-mediated mRNA decay." Nature Reviews Molecular Cell Biology, 20(7), 406–420. https://doi.org/10.1038/s41580-019-0126-2