concept · intermediate · partly written
Translation initiation: scanning and start-codon choice
How the small ribosomal subunit loads at the cap, scans and picks a start codon, how that choice is measured, and what it does to the protein.
Also called: translation initiation, scanning, Kozak sequence, uORF
In short
A small ribosomal subunit loaded with the first transfer RNA binds the cap, slides 5' to 3', and starts at the first AUG in decent context. A weak context lets some subunits skip it (leaky scanning) and start further on, which can give a second, N-terminally shorter protein. This is real for named genes, and the second form is the minor one in every case that has any number. 12
Reported, not yet replicated
The machine, in order more
The eIF2 complex carries the initiator methionine transfer RNA with GTP. With eIF1, eIF1A, eIF3 and eIF5 it forms the 43S complex, which attaches to the cap through eIF4F (eIF4E binds the cap, eIF4G is the scaffold, eIF4A is an RNA helicase). 'eIF' is the plain expansion eukaryotic initiation factor; the number-and-letter convention was not checked against HGNC. The complex scans, spending ATP, testing each triplet. On AUG pairing eIF1 leaves, eIF2 hydrolyses GTP, and eIF5B brings in the large subunit. A review states that a precise kinetic analysis of scanning is 'acutely lacking'. 123
Reported, not yet replicated
What sets the choice: context, near-cognate codons, upstream ORFs more
Purine at position -3 and G at +4 mark a strong AUG. A fluorescence-sorting screen of all 65,536 sequences from -6 to +5 in a mouse pre-B cell line, with an artificial reporter, gave a 12-fold range across AUG contexts; by the simple rule 40% of human genes (8,629) sit in a highly efficient context and 9.6% (2,065) have neither feature. That range is a model prediction from one mouse line, not measured per human gene. Near-cognate starts, in a reporter library in a mouse cell line, as percent of an AUG in best context: CUG about 50, ACG about 40, GUG about 20, AUU/AUA/AUC/UUG 5 to 10; absolute efficiency differed about 4-fold between two reporter genes. Another table, cited from a review whose primary was not opened, gives CUG 9.9%, ACG 4.1% and GUG 2.5% of AUG. The two tables disagree and neither is a constant. Reinitiation after a short upstream ORF is under 10% when the spacing is under 11 nt and about 70% at 79 nt or more (review, read through a summarising tool). Stress phosphorylates eIF2 alpha at Ser51, lowering most initiation but raising skipping of inhibitory upstream ORFs; ATF4 (activating transcription factor 4) uses delayed reinitiation to switch on. 4536
Reported, not yet replicated
How starts are seen, and how many more
Ribosome profiling with start-freezing drugs gave thousands of candidate starts: 13,454 in about 5,000 transcripts in mouse embryonic stem cells with harringtonine; 7,251 putative starts in 4,364 transcripts in human THP-1 cells, about 85% at NUG codons. Per-protocol counts for other human protocols could not be re-opened and are withheld. At protein level, N-terminal proteomics in Jurkat cells found 2,578 N-terminal peptides, 2,135 from annotated starts and 443 (about 17%) from alternative starts; 335 proteoforms from 133 genes with two or more N-termini; 101 of the 443 (23%) explained by splicing. Ribosome profiling agreed with 89% of annotated starts but only 29% (130 of 443) of the alternative ones. In HEK293T cytosol 31.6% of 3,306 N-terminal peptides pointed to proteoforms, with an extra western band for 34 genes. The footprint numbers are site counts, not protein. 78910
Reported, not yet replicated
Named genes with two N-terminal forms more
CEBPA (CCAAT/enhancer-binding protein alpha): a long form from the first AUG and a short form from an internal AUG by reinitiation after a six-codon upstream ORF; the short form keeps the DNA-binding domain and lacks most of the activation domain, and the two act oppositely in myeloid development. In a reporter, about 67% long and 33% short; lengthening the upstream ORF to 30 nt cut short-form output about 40%. PTEN (phosphatase and tensin homologue): PTENalpha from a CUG and PTENbeta from an AUU, with the N-terminal peptide confirmed by mass spectrometry and mutation of the AUU abolishing PTENbeta. In ten human genes with a weak annotated start, SNAP-tag reporters showed more initiation at the downstream AUG, but peptides unique to the long form were missing or about 100-fold rarer in all but one gene; ribosome profiling agreed with the reporters for CMPK1 and not for AASDHPPT. Gene names here are the expansions the sources print; they were not checked against HGNC. 11121314
Reported, not yet replicated
Function or noise: the field disagrees more
Reading A: alternative initiation 'predominantly results from molecular error' (abstract; the phrase was relayed by a fetch tool and not checked against the page). Reading B: ribosome-profiling non-AUG extensions (390 genes) far outnumber those with evolutionary support (60 genes), only 8 in both; the authors say lack of evidence of selection does not show an extension is non-functional. Neither paper reconciles the two. 1516
Contested
Where the measurements break more
Cycloheximide, used to freeze ribosomes, distorts yeast profiles; human 293T profiles were largely stable (R2 at least 0.92, fewer than 24 of 12,330 transcripts changed) but density still rose over the first 15 or so codons. In HeLa reporters cycloheximide did not stop near-cognate translation. These are caveats on every start-site list built from the drug. 1718
Reported, not yet replicated
Rates and errors, as far as measured more
Single human U2OS reporter mRNAs were translated at about 3.1 to 3.5 codons per second in one imaging study (the source check could not re-open it, so treat as unverified). A bacterial ribosome in vitro gives about 1e-5 to 1e-3 errors per codon depending on method. Mass-spectrometry detection of mistranslation exists for E. coli and yeast. No human per-codon error rate was found. 192021
Unverified: from our own report, source not opened
His idea, in his words
"There's also a reverse truncation where, I just realized, you may get two start codons. Now you have two versions of the protein: one that's transcribing from the front to the end, and one that's transcribing from the middle to the end." 22
Amit's hypothesis
The sourced account beside his statement more
The process is translation, not transcription: a second start codon is used by a ribosome on an mRNA already made. No source used the phrase 'reverse truncation'. The core is right: a downstream in-frame start gives a protein with the same end and a shorter beginning (CEBPA and PTEN above). It is incomplete: many starts extend the protein (546 putative N-terminal extensions in THP-1 cells); a second start can be in another frame; an upstream ORF usually lowers the main product; and the same N-terminal difference also arises from alternative promoters or splicing (23% of Jurkat alternatives). The second form is the minor one wherever a number exists. The report found no source that gives a share of human proteins with two confirmed forms. 1112891
Reported, not yet replicated
Plain answer
Translation starts when a small ribosomal subunit scans from the cap and recognises a start codon. Weak context or a near-match to AUG lets some subunits start elsewhere, giving a second N-terminal form for some genes; ribosome profiling finds thousands of such starts, and N-terminal proteomics confirms hundreds. 19
Reported, not yet replicated
Why it matters
The start codon sets the reading frame and the protein's first residues, and sometimes its address in the cell; start-site lists from footprints and from peptides disagree by about 70%. 9
Reported, not yet replicated
Still open here
- No human scanning rate was found; the only direct one is yeast, reconstituted, about 100 nt/s (abstract).
- No published share of human proteins with a confirmed second N-terminus was found; the nearest numbers are 133 genes in Jurkat cells and 34 genes by western blot in HEK293T.
- No endogenous human ratio of the two forms from a primary paper: the MYC 10 to 15% traces through a review to a 1992 paper in avian and mouse cells.
- Human translation error rate: not measured in anything opened (bacterial in vitro rates only).
- Termination, ribosome recycling and ribosome-stall quality control: not sourced.
- Non-AUG share of ribosome-profiling starts differs by protocol; per-protocol counts for Lee 2012, Gao 2015 and Ingolia 2011 could not be re-opened and are not given.
Not yet written: Equation, Real numbers.
Linked from
- N-terminal proteomics (COFRADIC)
- Ribosome profiling (Ribo-seq)
- Alternative start codons and two protein forms
- Conservation of the gene-expression machinery across bacteria, archaea and eukaryotes
- RNA capping and the poly(A) tail
- Nicholas T. Ingolia
- For how many human genes does a second start codon make a second protein with its own job?
- How many translated noncanonical ORFs make functional products?
Sources
- Hinnebusch AG, Ivanov IP, Ivanov IP, et al.. Translational control by 5'-untranslated regions of eukaryotic mRNAs.. Science (2016). paper; opened 2026-11-17 all records citing this
- Kearse MG, Wilusz JE. Non-AUG translation: a new start for protein synthesis in eukaryotes. Genes Dev (2017). review; opened 2026-11-17 all records citing this
- Dever TE, Ivanov IP, Hinnebusch AG. Translational regulation by uORFs and start codon selection stringency.. Genes Dev (2023). paper; opened 2026-11-17 all records citing this
- Noderer WL, Flockhart RJ, Bhaduri A, et al.. Quantitative analysis of mammalian translation initiation sites by FACS-seq.. Mol Syst Biol (2014). paper; opened 2026-11-17 all records citing this
- Diaz de Arce AJ, Noderer WL, Wang CL. Complete motif analysis of sequence requirements for translation initiation at non-AUG start codons.. Nucleic Acids Res (2018). paper; opened 2026-11-17 all records citing this
- Vattem KM, Wek RC. Reinitiation involving upstream ORFs regulates ATF4 mRNA translation in mammalian cells.. Proc Natl Acad Sci U S A (2004). paper; opened 2026-11-17 all records citing this
- Ingolia NT, Lareau LF, Weissman JS. Ribosome profiling of mouse embryonic stem cells reveals the complexity and dynamics of mammalian proteomes.. Cell (2011). paper; opened 2026-11-17 all records citing this
- Fritsch C, Herrmann A, Nothnagel M, et al.. Genome-wide search for novel human uORFs and N-terminal protein extensions using ribosomal footprinting.. Genome Res (2012). paper; opened 2026-11-17 all records citing this
- Gawron D, Ndah E, Gevaert K, et al.. Positional proteomics reveals differences in N-terminal proteoform stability.. Mol Syst Biol (2016). paper; opened 2026-11-17 all records citing this
- Bogaert A, Fijalkowska D, Staes A, et al.. N-terminal proteoforms may engage in different protein complexes.. Life Sci Alliance (2023). paper; opened 2026-11-17 all records citing this
- Fernandez SG, Ferguson L, Ingolia NT. Ribosome rescue factor PELOTA modulates translation start site choice for C/EBPα protein isoforms.. Life Sci Alliance (2024). paper; opened 2026-11-17 all records citing this
- Liang H, Chen X, Yin Q, et al.. PTENβ is an alternatively translated isoform of PTEN that regulates rDNA transcription.. Nat Commun (2017). paper; opened 2026-11-17 all records citing this
- Liang H, He S, Yang J, et al.. PTENα, a PTEN isoform translated through alternative initiation, regulates mitochondrial function and energy metabolism.. Cell Metab (2014). paper; opened 2026-11-17 all records citing this
- Benitez-Cantos MS, Yordanova MM, O'Connor PBF, et al.. Translation initiation downstream from annotated start codons in human mRNAs coevolves with the Kozak context.. Genome Res (2020). paper; opened 2026-11-17 all records citing this
- Xu C, Zhang J. Mammalian Alternative Translation Initiation Is Mostly Nonadaptive.. Mol Biol Evol (2020). paper; opened 2026-11-17 all records citing this
- Fedorova AD, Kiniry SJ, Andreev DE, et al.. Thousands of human non-AUG extended proteoforms lack evidence of evolutionary selection among mammals.. Nat Commun (2022). paper; opened 2026-11-17 all records citing this
- Sharma P, Wu J, Nilges BS, et al.. Humans and other commonly used model organisms are resistant to cycloheximide-mediated biases in ribosome profiling experiments.. Nat Commun (2021). paper; opened 2026-11-17 all records citing this
- Kearse MG, Goldman DH, Choi J, et al.. Ribosome queuing enables non-AUG translation to be resistant to multiple protein synthesis inhibitors.. Genes Dev (2019). paper; opened 2026-11-17 all records citing this
- Yan X, Hoek TA, Vale RD, et al.. Dynamics of Translation of Single mRNA Molecules In Vivo.. Cell (2016). paper; opened 2026-11-17 all records citing this
- Wohlgemuth I, Pohl C, Mittelstaet J, et al.. Evolutionary optimization of speed and accuracy of decoding on the ribosome.. Philos Trans R Soc Lond B Biol Sci (2011). paper; opened 2026-11-17 all records citing this
- Mordret E, Dahan O, Asraf O, et al.. Systematic Detection of Amino Acid Substitutions in Proteomes Reveals Mechanistic Basis of Ribosome Errors and Selection for Translation Fidelity.. Mol Cell (2019). paper; opened 2026-11-17 all records citing this
- Amit's dictation of 2026-11-17: what the atlas must contain. Kept in the atlas repository; not published.
- Lee S, Liu B, Lee S, et al.. Global mapping of translation initiation sites in mammalian cells at single-nucleotide resolution.. Proc Natl Acad Sci U S A (2012). paper; opened 2026-11-17 all records citing this
Where it sits on the map
History of this record (6)
This record has been changed 6 times, on the dates below. The note kept against each change is written for the people who maintain the atlas, in their own shorthand, so it is not shown here; it is in the record itself.
- Status
- partly written
- Last updated
- 2026-11-19
- Record
- content/concepts/translation-initiation-and-start-codon-choice.json