The jellyfish that ages backward has been sequenced

Researchers at the University of Oviedo sequenced the genome of Turritopsis dohrnii, the jellyfish that reverts to its juvenile stage, and found doubled copies of genes for DNA repair and protection.

A medusa that turns back into a polyp

Most jellyfish follow one direction. A fertilized egg becomes a larva, the larva settles and becomes a polyp, the polyp buds off medusae, and the medusae reproduce and die. In 1996 Piraino, Boero and colleagues reported in The Biological Bulletin that the hydrozoan then known as Turritopsis nutricula did not. All stages of its medusa, from newly released to fully mature, could transform back into colonial hydroids, either directly or after a resting period. It was, they wrote, the first animal known to revert to a colonial juvenile form after reaching sexual maturity. Excision experiments showed that the transformation required differentiated cells from the outer epidermis of the bell and part of the gut, revealing what they called a transformation potential "unparalleled in the animal kingdom."

In 2007 Schmich, Piraino and colleagues showed the process could be triggered on demand. Exposure to cesium chloride, a chemical known to induce larval metamorphosis in cnidarians, also induced reverse development in Turritopsis dohrnii, the species now recognised as the one that rejuvenates without limit. The reversal proceeded through a cascade of events that resembled metamorphosis run in the opposite direction.

The cyst stage

The reversal passes through an intermediate. A damaged, starved or ageing medusa shrinks until every structure of the adult disappears into a homogeneous opaque mass called the cyst. From the cyst a stolon grows and polyp buds appear, becoming full polyps. In 2019 Matsumoto, Piraino and Miglietta sequenced the RNA of the polyp, the newborn medusa and the cyst of T. dohrnii, assembling 92,569, 74,639 and 86,373 contigs respectively. Compared with the two stable stages, the cyst showed changed activity in telomerase, in the regulation of transposable elements and in DNA repair systems, alongside suppression of cell signalling, cell division and differentiation. Cells in the cyst were transdifferentiating, changing from one specialised type to another without passing through an embryo.

Two genomes, side by side

A transcriptome shows which genes are active; it does not show what genes an animal has. In 2022 Maria Pascual-Torner, Carlos López-Otín and colleagues at the University of Oviedo published in PNAS the first genome of T. dohrnii, and alongside it the genome of Turritopsis rubra, a related species with no reported ability to rejuvenate after reproduction. The comparison was the point. Both animals are Turritopsis; only one is immortal.

Both genomes were sequenced on the Illumina platform. The T. dohrnii assembly came to an estimated 390 megabases in 74,829 scaffolds, consistent with an earlier estimate of 383 megabases; T. rubra came to 210 megabases in 9,508 scaffolds. Automatic annotation predicted 17,468 genes in T. dohrnii and 9,324 in T. rubra. Repetitive elements made up about half of each genome. Completeness, measured against a benchmark set of single-copy orthologs, was 78.88 percent for T. dohrnii and 88.78 percent for T. rubra. The authors then manually annotated almost 1,000 genes involved in ageing and DNA repair in both species and in three other cnidarians, Hydra vulgaris, Clytia hemisphaerica and Aurelia aurita, and validated gene copies by PCR with primers designed to distinguish them.

Extra copies of the repair machinery

The immortal species carried more copies of several genes. POLD1, encoding DNA polymerase delta, was present in four copies in T. dohrnii against one in T. rubra; POLA2, encoding a subunit of DNA polymerase alpha, in two against one; TOP3B, a topoisomerase, in four against three. Among repair genes, XRCC5, GEN1, RAD51C and MSH2 were each duplicated in T. dohrnii and single in T. rubra. In the response to oxidative stress, thioredoxin, TXN, had five copies in T. dohrnii against two in T. rubra, and glutathione reductase, GSR, was duplicated in T. dohrnii alone among the five cnidarians examined. Both were switched on during life cycle reversal. The rate of amplification among the manually annotated ageing genes was 3.77 times higher than in a random set of 1,000 genes from the same genome, which argued against a general tendency of the T. dohrnii genome to duplicate.

Telomeres and stem cells

At the telomeres, T. dohrnii had two copies of GAR1, a component of the telomerase complex, where the other cnidarians had one. In the shelterin protein POT1, which caps telomeres and restrains telomerase, both Turritopsis species carried rare variants at the same highly conserved position, G272N in T. dohrnii and G272R in T. rubra. When the researchers engineered these substitutions into the human protein and measured binding to telomere-like DNA, the T. dohrnii variant reduced binding dramatically, suggesting a weaker brake on telomere maintenance. In stem cell pathways, GLI3, a transcription factor in Hedgehog signalling, was duplicated in T. dohrnii, single in T. rubra and absent from the other three cnidarians, and every copy was active during reversal. BMP7, which modulates apoptosis, had eight copies in T. dohrnii, five of them active during reversal, against five in T. rubra and three in Hydra.

The reprogramming signature

The transcriptome of reversal supplied the last piece. Gene set enrichment analysis found that the targets of polycomb repressive complex 2, a chromatin silencer whose activity is required to establish and keep pluripotency in embryonic stem cells, were switched off, most strongly in the cyst stage. At the same time the targets of the pluripotency factors NANOG, OCT4, SOX2 and MYC were switched on. These are two of the main mechanisms by which differentiated cells are reprogrammed in the laboratory. The authors proposed that in T. dohrnii they operate naturally, in the whole animal, every time it rewinds.

What the story does and does not show

The genomes are drafts. The T. dohrnii assembly is fragmented, with a completeness score of 78.88 percent, and the authors acknowledged that a higher number of gene copies in a more fragmented assembly could be an artefact; they addressed this with conservative annotation rules and PCR validation, and discarded copies that could not be confirmed. The two species compared are not close relatives, so some of the differences between them have nothing to do with immortality, and the authors wrote that a closer mortal relative such as T. nutricula will be needed to confirm the interpretation. The paper was corrected twice in 2023. The word immortal is defined in the paper by a specific observation: sexually mature medusae of T. dohrnii rejuvenate with a probability of up to 100 percent and no apparent limit on the number of cycles. It is a statement about the laboratory record, not a measurement of a maximum lifespan in the sea.

What the paper does show is a list of candidates. In an animal that rewinds its development, the genes for copying DNA, repairing it, protecting it from oxidation, maintaining telomeres and holding cells in a stem-like state are present in more copies, or in altered forms, than in a relative that cannot. Which of them are necessary, and which are passengers, is a question the genome poses and does not answer. The instruction set has been read; what each line does remains to be tested.

It can live forever. Now we can read how.