= Solution
\b[During <amino acid starvation>, cells reduce bulk <biological translation> while selectively promoting an adaptive response.] It has both a direct substrate effect on elongation and regulated effects on initiation. Their balance depends on which <amino acid> is limiting, how severely it is depleted, the organism and the duration of deprivation.
The immediate substrate problem is reduced charging of the relevant <transfer RNAs> by <aminoacyl-tRNA synthetases>. A <ribosome> encountering a cognate <codon> waits longer for an appropriate charged adaptor, potentially creating queues or <ribosome collisions>. The response is not necessarily identical at every synonymous <codon>, because <transfer RNA> isoacceptors, their charging and their competition differ. In mammalian <biological cells>, comparative <ribosome profiling> under particular <amino acid> limitations demonstrated selective <codon>-associated pausing rather than an invariant slowdown at every sense <codon>.<Translational control through differential ribosome pausing during amino acid limitation in mammalian cells by Alicia M. Darnell and others>[Summary and Results: amino-acid-specific pausing, tRNA charging and translational responses]{ref} Persistent stalls can engage <ribosome-associated quality control> to rescue stalled machinery and dispose of incomplete products. Increased <ribosome> occupancy at a site therefore need not mean increased productive synthesis.
In <eukaryotes>, <GCN2> links <amino acid> limitation to the <integrated stress response>. Deacylated <transfer RNA> and signals from starved <ribosomes>, supported by <GCN1>, promote <GCN2> activation.<Activation of GCN2 by the ribosomal P-stalk by Alison J. Inglis and others>[Abstract and Results: ribosome-dependent activation and relationship to translational stress]{ref} Its <protein kinase> activity phosphorylates the alpha subunit of <eIF2> at serine 51. Normal <eIF2> brings initiating <methionyl-tRNA> to the small <ribosomal subunit> as part of a ternary complex:
$$
\mathrm{eIF2\!\cdot\!GTP\!\cdot\!Met\!\! -\!tRNA_i}.
$$
After initiation, <eIF2> must exchange bound <GDP> for <GTP>, a reaction promoted by <eIF2B>. Phosphorylated <eIF2> inhibits this recycling activity, reducing ternary-complex supply and initiation on many <messenger RNAs>. This conserves scarce substrates and limits addition of further <ribosomes> to already stressed templates. Genetic experiments established that <GCN2>-dependent regulation can repress bulk synthesis while increasing production of <ATF4>.<Regulated translation initiation controls stress-induced gene expression in mammalian cells by Heather P. Harding and others>[Abstract and Results: eIF2 kinase dependence and selective ATF4 expression]{ref}
The apparent paradox is resolved by <upstream open reading frames>. In the canonical mammalian <ATF4> leader, a short permissive <upstream open reading frame> is followed by an inhibitory one that overlaps the <ATF4> coding region in a different <reading frame>. When ternary complex is plentiful, scanning small <ribosomal subunits> that have completed the first short frame reacquire it rapidly and initiate at the inhibitory frame. After <eIF2> <phosphorylation>, reacquisition is delayed: some scanning subunits pass that inhibitory start and become competent before the downstream <ATF4> start. \b[Reduced initiation capacity can therefore increase initiation at a particular downstream coding region.] Experiments using <upstream open reading frame> <mutations> and reporter assays directly tested this reinitiation mechanism.<Reinitiation involving upstream ORFs regulates ATF4 mRNA translation in mammalian cells by Krishna M. Vattem and Ronald C. Wek>[Abstract and Results: uORF mutations, spacing and delayed reinitiation]{ref} The increase is not unlimited: if ternary complex becomes too scarce, initiation at <ATF4> also fails.
<ATF4> then changes <transcription> of adaptive <genes>, including those supporting <amino acid> supply, transport and stress management. Yeast uses the related <GCN2>–<eIF2> logic to increase <GCN4> production through a leader with several <upstream open reading frames>; it should not be drawn as the identical two-frame <ATF4> arrangement. The response can recover after replenishment through restoration of charging and dephosphorylation of <eIF2>. Prolonged, unresolved deprivation can instead favor damage and <apoptosis>. These later <gene expression> changes complement the initial translational response rather than explaining its earliest onset.
A parallel mammalian pathway regulates <mTORC1>. In nutrient-replete conditions, <mTORC1> promotes cap-dependent initiation partly by phosphorylating <4E-BP proteins> and activating <S6 kinase>. Phosphorylated <4E-BP proteins> release <eIF4E>, allowing it to associate with <eIF4G> and assemble an active <eIF4F> initiation complex. <amino acid starvation> reduces <mTORC1> activity, allowing hypophosphorylated <4E-BP proteins> to bind <eIF4E> and restrain initiation. A concrete upstream sensor is <Sestrin2>: loss of <leucine> promotes its inhibitory interaction with <GATOR2>, influencing <Rag GTPases> and nutrient-dependent recruitment of <mTORC1> to the <lysosome>.<Sestrin2 is a leucine sensor for the mTORC1 pathway by Rachel L. Wolfson and others>[Abstract and Results: leucine binding, sensor mutants and GATOR2 interactions]{ref} This route measures nutrient sufficiency through a different mechanism from the <GCN2> pathway; neither should be reduced to the other.
Reduced <mTORC1> signaling also helps limit investment in <ribosome> production and permits <autophagy>, which can recycle cellular material to supply <amino acids>. These adaptations operate on different timescales from waiting for a missing charged <transfer RNA>. The outcome is selective redistribution of translational capacity, not simply uniform cessation of every <protein> synthesis event. Some cap-dependent transcripts are especially sensitive to initiation-factor availability, while stress-regulated leaders can favor others.
To distinguish the mechanisms, compare a defined single-<amino acid> withdrawal with matched complete medium and replenishment. Control residual <amino acids> in supplements and measure viability. Measure global nascent <protein> synthesis, <transfer RNA> charging, <eIF2> <phosphorylation>, <ATF4> synthesis, and <mTORC1> substrate <phosphorylation> over time. Combine <ribosome profiling> with <RNA sequencing> and external normalization; otherwise globally reduced synthesis can disappear through library normalization. Compare normal <biological cells> with <GCN2> loss or a nonphosphorylatable <eIF2> variant, and use <ATF4> leader reporters with separately mutated <upstream open reading frames>. Independently perturb <4E-BP proteins> or <mTORC1> signaling. Recovery of initiation signaling without recovery of elongation would show that substrate shortage still limits output. A loss of large <polysomes> suggests reduced initiation, whereas maintained or increased occupancy can accompany stalled elongation; neither pattern alone measures productive <protein> synthesis.
In <bacteria>, a related sensing principle feeds a different control system: in <Escherichia coli>, <RelA> recognizes starved <ribosomes> containing deacylated <transfer RNA> and synthesizes <(p)ppGpp>, initiating the <stringent response>. Structural work explains how the starved <ribosome> activates <RelA>.<Ribosome-dependent activation of stringent control by Alan Brown and others>[Abstract and structural Results: deacylated tRNA, RelA activation and alarmone synthesis]{ref} The <stringent response> reallocates <gene expression> and restrains investment in growth machinery; it is not an <eIF2>-based mechanism. \b[Across these systems, the common strategy is to lower demand while preserving or increasing the capacity to restore <amino acid> supply.]
Back to article page
