eIF5A: A Key Protein in Cellular Function and Disease

eIF5A is a small protein found in every eukaryotic cell, from yeast to humans, and it carries a distinction no other protein can claim: it is the sole known carrier of the amino acid hypusine, a molecular modification so critical that cells cannot survive without it. Although originally classified as a translation initiation factor (hence the “eIF” in its name), decades of research have revealed that eIF5A does far more than help start protein production. It keeps ribosomes from stalling mid-sentence, supports cellular recycling and energy production, and shows up in disease after disease, from cancer and diabetes to brain aging and viral infection.

What eIF5A Actually Does Inside a Cell

For years after its discovery, researchers assumed eIF5A mainly helped kick-start protein synthesis. That turned out to be only part of the story. Ribosome profiling of cells depleted of eIF5A revealed a widespread slowdown in protein production, with ribosomes stalling at many different sequences throughout the genome, not just at the handful of trouble spots scientists originally expected.¹ The same study found that eIF5A also speeds up the final step of protein production, boosting the rate at which ribosomes release their finished product by more than 17-fold.

One of eIF5A’s best-characterized jobs is rescuing ribosomes that get stuck while building stretches of the amino acid proline. Three prolines in a row are enough to bring a ribosome to a halt, and eIF5A relieves that stall both in the test tube and in living cells.² But the problem extends well beyond proline. Ribosomes also pause at sequences rich in glycine and charged amino acids, and eIF5A helps at more than 200 different tripeptide motifs where stalling occurs.³ In practical terms, eIF5A functions as a general-purpose anti-stalling factor for the protein-building machinery.

The Hypusine Modification

What makes eIF5A biologically unusual is hypusine, an amino acid found nowhere else in the cell’s protein repertoire. Hypusine is not encoded in DNA. Instead, it is built onto eIF5A after the protein is already made, through a two-step chemical process. First, an enzyme called deoxyhypusine synthase (DHPS) takes part of the polyamine spermidine and attaches it to a specific lysine residue on eIF5A. Then a second enzyme, deoxyhypusine hydroxylase (DOHH), adds a hydroxyl group to complete the modification.⁴ Without this two-step activation, eIF5A cannot do its job. The modification is essential for eIF5A’s activity in regulating protein synthesis across all eukaryotes.⁵

Because spermidine serves as the raw material for hypusination, anything that changes spermidine levels in a cell changes how much active eIF5A is available. This makes eIF5A a metabolic sensor of sorts: the protein links a cell’s polyamine supply to its capacity for protein production and downstream processes. Blocking DHPS with a compound called GC7, or inhibiting DOHH with the drug ciclopirox, shuts down eIF5A’s activity and with it a cascade of cellular functions.⁶

How eIF5A Sits on the Ribosome

Structural biologists have captured detailed snapshots of eIF5A bound to the ribosome, and these images explain a lot about how the protein works. A cryo-electron microscopy reconstruction at 3.9 ångströms resolution showed eIF5A wedged between the P-site and E-site of the ribosome, where it makes contact with the transfer RNA carrying the growing protein chain.⁷ A crystal structure at 3.25 ångströms resolution confirmed this positioning and showed eIF5A interacting with two flexible parts of the large ribosomal subunit, the L1-stalk and a helix of the ribosomal RNA.⁸

Mutagenesis experiments pinpointed which surface features of eIF5A matter most for holding onto the ribosome. A cluster of charged residues proved critical: when these were mutated to alanine, eIF5A’s grip on the ribosome weakened substantially. The data suggest that eIF5A’s interaction with a ribosomal protein called L1 contributes more to stable binding than its contacts with the transfer RNA itself.⁹

Autophagy and Mitochondrial Health

eIF5A’s influence reaches well beyond keeping ribosomes moving. One of its most consequential downstream roles is in autophagy, the process cells use to clean up damaged components and recycle them. A high-throughput screen identified eIF5A as necessary for the formation of autophagosomes, the membrane-bound compartments that engulf cellular debris. The reason traces back to translation: eIF5A is needed to efficiently produce ATG3, a protein required for a key step in autophagosome assembly.¹⁰ eIF5A also drives production of TFEB, a master regulator of autophagy-related genes. When spermidine levels fall, as they do in aging, TFEB production drops and autophagy falters.¹¹

Mitochondrial function also depends on eIF5A. Cells lacking eIF5A show reduced oxygen consumption, lower ATP production, and decreased levels of several enzymes involved in energy metabolism.¹² This appears to happen because hypusinated eIF5A supports the production of a subset of mitochondrial proteins involved in the TCA cycle and the electron transport chain. When eIF5A activation is impaired, cells show reduced flux through these core energy-generating pathways.⁶

Cancer and eIF5A’s Two Isoforms

Humans carry two versions of eIF5A, called eIF5A1 and eIF5A2. They share about 84% of their sequence, but behave very differently. eIF5A1 is produced in virtually every cell type and is considered the housekeeping isoform. eIF5A2 is a different story: its protein is normally detectable in very few tissues, but it shows up at high levels in many cancers and has been flagged as a candidate oncogene.¹³ Elevated expression of either isoform carries unfavorable prognostic implications for several cancer types, and both have been proposed as biomarkers.¹⁴

Research into prostate cancer illustrates how targeting this pathway might work therapeutically. Knocking down DHPS or treating cells with the inhibitor GC7 cut cell proliferation, migration, and invasion. In patient-derived tumor models, GC7 reduced the ability of cancer cells to invade surrounding tissue by about 60%. In mice, both genetic and chemical suppression of eIF5A hypusination significantly shrank tumors and dramatically reduced metastasis. In an orthotopic model where cancer cells were implanted directly into the prostate, almost all control mice developed distant metastases, while the majority of mice with suppressed eIF5A hypusination did not.¹⁵ Similar antiproliferative effects have been seen in neuroblastoma, where GC7 inhibited tumor cell growth in a dose-dependent fashion by triggering the cell-cycle brake p21.¹⁶

The eIF5A axis is also regulated by non-coding RNA networks. In non-small cell lung cancer, a long non-coding RNA called OGFRP1 acts as a molecular sponge, soaking up a microRNA that would otherwise keep eIF5A levels in check. When OGFRP1 is silenced, eIF5A expression drops and tumor-promoting behavior is blunted.¹⁷ Analogous sponge circuits involving eIF5A2 have been documented in colorectal cancer.¹⁸ These findings suggest that eIF5A levels can be hijacked by the cancer cell’s own gene-regulatory machinery, not just by changes to eIF5A itself.

Diabetes and Pancreatic Inflammation

Outside of cancer, one of the most developed disease connections is between eIF5A and diabetes. In mouse models of type 1 diabetes, hypusinated eIF5A contributes to the inflammatory cascade that destroys insulin-producing beta cells. The mechanism runs through inducible nitric oxide synthase (iNOS), an enzyme that generates toxic levels of nitric oxide in inflamed tissue. eIF5A’s hypusine modification is needed for efficient transport and translation of iNOS-encoding messenger RNA. When eIF5A was depleted or its hypusination was blocked, iNOS levels fell, beta cells survived longer, and mice were protected from developing high blood sugar.¹⁹ Pharmacological blockade of DHPS improved glucose tolerance and preserved beta cell mass in both type 1 and type 2 diabetic mouse models, and hypusinated eIF5A has been detected in the pancreas of humans with both forms of diabetes.²⁰

Brain Aging and Spermidine

Some of the most attention-grabbing eIF5A research concerns aging, particularly in the brain. In fruit flies, hypusinated eIF5A levels in the brain decline with age, and that decline tracks with worsening mitochondrial function, reduced locomotion, and impaired memory. Genetically reducing eIF5A hypusination in neurons produced a pattern of mitochondrial decay resembling premature aging.²¹ When the researchers knocked down DHS (the fly equivalent of DHPS) specifically in neurons, proteins related to mitochondrial respiration and oxidative phosphorylation were significantly less abundant in the brain.²²

The encouraging finding is that dietary spermidine supplementation can push back against this decline. In aged mice, spermidine delivered in drinking water or food increased eIF5A hypusination in the hippocampus, the brain region central to memory formation.²³ Across multiple model organisms, dietary spermidine extended lifespan, activated autophagy, and improved mitochondrial function, with the spermidine-to-eIF5A-hypusination axis identified as a key mediator.²⁴ Whether these effects translate meaningfully to human brain aging is still an open question, but the consistency of the finding across yeast, flies, and mice has made this one of the more watched areas in aging research.

Viral Infections and eIF5A

Several viruses exploit eIF5A for their own replication. HIV-1 is the best-studied example. The virus depends on eIF5A hypusination for efficient gene expression, and partially knocking down eIF5A with small interfering RNA suppresses HIV-1 gene output.²⁵ The mechanism appears to involve an internal ribosome entry site in the HIV-1 messenger RNA, a structure that lets the virus initiate protein production without the normal cap-dependent machinery. Blocking DOHH with deferiprone or ciclopirox reduced expression of HIV-1 proteins and impaired the virus’s ability to use this alternative translation route.²⁶ The same study found that other retroviruses, including HTLV-1 and MMTV, also rely on DOHH activity for their cap-independent translation. Because these drugs target the host cell’s own machinery rather than viral proteins, they sidestep the problem of viral mutation-driven drug resistance, though they also carry the risk of disrupting normal eIF5A function.

When eIF5A Itself Is Broken

Rare genetic mutations in the eIF5A pathway cause severe developmental disorders in children. Mutations in the gene for DHPS, which catalyzes the first step of hypusination, are associated with a neurodevelopmental disorder in individuals who carry two defective copies.²⁷ Mutations in the eIF5A gene itself, arising spontaneously rather than being inherited from both parents, cause a condition marked by developmental delay, abnormally small head size, a small jaw, and distinctive facial features. In experimental models, spermidine supplementation partially rescued some of the defects caused by impaired eIF5A function, hinting that boosting the polyamine supply might help compensate for a weakened protein.²⁸

An Ancient and Deeply Conserved Protein

eIF5A is not just a eukaryotic invention. Its bacterial counterpart, EF-P, performs a strikingly similar ribosome-rescue function, and archaea carry their own version, called aIF5A. A crystal structure of the archaeal aIF5A-DHS complex from the hot-spring organism Saccharolobus islandicus revealed that the interaction between the protein and its modifying enzyme is remarkably similar to the human version, with the critical hypusine loop residues identical across the two domains of life.²⁹ This degree of conservation across billions of years of evolution underscores how fundamental the hypusination mechanism is to life.

Plants have their own eIF5A genes, often in expanded families with specialized roles. In the desert shrub Tamarix androssowii, an eIF5A homolog was shown to boost stress tolerance: transgenic poplar trees expressing it had higher antioxidant enzyme activity, less membrane damage, and more chlorophyll under salt stress conditions.³⁰ Across plant species, different eIF5A family members have been linked to meristem activity, vascular development, organ growth, senescence, programmed cell death, and immune defense.³¹ The breadth of these roles echoes what researchers see in animals: eIF5A sits at a translational chokepoint that affects whatever biological process relies on the proteins it helps produce.

A Life Beyond Translation

Most research has focused on eIF5A’s cytoplasmic role in translation, but evidence is accumulating that the protein also works inside the nucleus. A preprint from yeast experiments reported that eIF5A binds to chromatin, the DNA-protein complex that makes up chromosomes, and represses transcription by preventing RNA polymerase II from accessing certain genes. The genes most affected are the same ones that depend on eIF5A for their translation, creating a feedback loop: eIF5A both regulates how fast a gene’s messenger RNA is made and how efficiently that messenger RNA is turned into protein.³² If this finding holds up in mammalian systems, it would substantially broaden the picture of what eIF5A does, repositioning it as a gene-expression coordinator rather than purely a translation helper.

Meanwhile, the connections between eIF5A and non-coding RNA continue to expand beyond cancer. In vascular smooth muscle cells, the long non-coding RNA SNHG12 promotes cell proliferation and migration partly by shielding eIF5A messenger RNA from a microRNA that would otherwise degrade it.³³ This sort of regulatory wiring means eIF5A levels in a given tissue are not just set by how much spermidine is around; they are also tuned by an entire layer of RNA-based gene regulation that researchers are only beginning to map. For a protein once dismissed as a minor initiation factor, eIF5A has turned out to be remarkably central to cell biology, sitting at the crossroads of metabolism, protein production, and disease in ways that are still being uncovered.