The Slug That Eats Its Own Solar Panels

A Green Animal That Shouldn't Work

Pick up a lettuce-leaf-shaped animal the size of a thumbnail, wave it under a light meter, and it does something no animal ought to do: it fixes carbon. Sacoglossan sea slugs in the genus Elysia eat filamentous and siphonaceous green algae — in this study, Bryopsis, Penicillus and Halimeda — and instead of digesting the chloroplasts along with everything else in the algal cell, some species pull them intact into their own gut-lining cells and keep them working [1]. The slugs are described, in the paper's own phrase, as "solar-powered" [1]. That is not a metaphor. Illuminate one and it measurably produces oxygen.

The phenomenon has a name, kleptoplasty, and it is old news by the standards of the field: a 1972 report already described "persistence of functional chloroplasts" in a related species, Elysia viridis, and by 2000 a group working on the American species Elysia chlorotica had shown that photosynthesis, thylakoid protein maintenance and even chloroplast gene expression continue for many months after the chloroplast has been separated from the algal nucleus that used to service it [2,3]. What was missing for more than fifty years was any account of the machinery on the animal's side. A stolen organelle sitting loose in an animal cell should not survive one cell division, let alone a slug's whole adult life. Something has to be holding it, feeding it, and deciding when it has stopped being useful. The 2025 paper is the first to show what that something actually is [1].

What Kleptoplasty Is, and What It Isn't

Worth being precise here, because the word gets used loosely. Kleptoplasty is not symbiosis in the textbook sense, where two organisms live inside one another across generations. It is theft with a shelf life. A juvenile Elysia hatches with no chloroplasts of its own and must feed on the right algae to acquire a working set; nothing about the arrangement is passed down through the slug's eggs, and the standard treatments of the kleptoplast state in the literature describe it exactly that way — a state each individual animal has to re-establish for itself in its own lifetime, not something it inherits [4]. That single fact is what makes kleptoplasty scientifically interesting rather than merely gross: it is intracellular organelle acquisition happening in real time, in an organism you can watch, rather than something that happened once, billions of years ago, and left only the mitochondria and chloroplasts we now take for granted as the fossil record of two such thefts that stuck.

It is also not indefinite. Even the most durable kleptoplasty on record eventually runs down; the 2025 paper's own Elysia crispata keeps chloroplasts photosynthetically active for close to a year, near the long end of what has been measured across the family, and cross-species surveys of sacoglossans from the Indo-Pacific and Mediterranean find the retention time varies enormously by species, from days to many months, depending on which algae a given slug specializes in eating [5]. The 2025 paper picked one of the champions to work with.

Where the Puzzle Actually Bites

Here is the part that should stop you. A chloroplast is not a self-sufficient machine. Its own genome encodes only a fraction of the proteins photosynthesis requires; the rest are ordinarily built by the plant cell's nucleus and imported continuously, replacing worn-out parts as fast as light and heat degrade them. Cut a chloroplast off from that supply line entirely and, by every rule of plant cell biology, it should decay within days to weeks as its machinery wears out unreplaced.

Except it doesn't, not in these slugs. A 1996 study on Elysia chlorotica found the chloroplast's own genes still being actively transcribed deep into the kleptoplast state, evidence that the organelle was not just coasting on stockpiled protein but doing ongoing biochemical work [6]. So either the sea slug has found some way to keep a foreign organelle supplied despite having none of the machinery a plant would use to do it, or the chloroplast is somehow running far longer on inherited reserves than plant biology predicts it should. Both options are strange. The 2025 paper's contribution is showing that the first one is what is actually happening, and giving the supply mechanism a name and an anatomy.

Finding the Missing Box

The evidence trail runs through a common lab technique used for an uncommon question. Allard and colleagues fed the slugs, let the kleptoplasty establish, and then looked for molecular tags that mark specific stages of an animal cell's own internal traffic system — the pathway that ordinarily handles food particles a cell swallows whole. Large cargo entering a cell by phagocytosis normally sits inside a membrane-bound phagosome, which then matures step by step into first an early and then a late endosome before its contents are broken down. Each stage carries its own molecular signature, proteins that mark "early here" or "late here" the way a shipping label marks a package's stage in transit.

The team's proteomic survey of slug tissue turned up several of those late-stage endosomal signatures sitting specifically on the compartments holding stolen chloroplasts [1]. Then a membrane dye, FM1–43, made the anatomy visible directly under the microscope: after gently lysing the slug's chloroplast-bearing cells, the dye lit up "a clear and distinct membrane" wrapped around each individual stolen chloroplast — one membrane, one chloroplast, not a shared compartment holding several at once [1]. That structure is what the paper names the kleptosome, in a coinage built the same way "kleptoplast" was: klepto-, stolen, plus -some, body. Figure 1 draws it out.

Whose Protein Is This, Anyway

A membrane studded with the slug's own trafficking proteins tells you the box is host-made. It does not by itself tell you the box is doing anything active for the chloroplast inside. For that, the team ran a pulse-labelling experiment: feed the slug a chemical tag that gets built into any protein synthesized after the tag is given, wait, then extract chloroplasts and ask whose genome the newly made proteins found inside them came from. If the chloroplast were simply coasting on a fixed stock of algal protein it arrived with, freshly made material recovered from it should be scarce, and what little there was should read as algal. Instead, the team found that nearly all of the newly synthesized peptides recovered from the chloroplasts mapped to the slug's own, metazoan genome, not to any plant or algal one [1]. Figure 2 sketches the route that finding implies.

The Channel That Keeps the Lights On

Housing and resupplying a chloroplast is not the same as tuning the chemistry around it so photosynthesis actually keeps working. That job, the paper argues, belongs mostly to one of the four marker proteins: P2X4, an ATP-gated ion channel. P2X receptors are a well-characterised family in animal physiology generally, opening a pore across a membrane specifically when extracellular ATP binds them [7]; here, one shows up sitting in the kleptosome membrane instead of on the outer cell surface, where the family usually works.

The team ran patch-clamp electrophysiology on the channel directly, using 1 mM ATP to activate it and the compound ivermectin, a known positive modulator of P2X4 in other systems, to potentiate the current further [1]. Then they asked what happens to the whole point of the exercise, photosynthesis, if the channel is blocked instead. Treating slugs with 5-BDBD, a selective P2X4 antagonist, cut net photosynthetic oxygen production by 62% relative to untreated controls, a result the paper reports as statistically significant (p = 0.0250) [1]. That is a large effect for blocking a single ion channel. In the club's own arithmetic on that number: if control slugs are producing 100% of their normal oxygen output, blocked slugs are down to only about 38% of it — call it a bit over a third — from one channel switched off.

Exactly how an ATP-gated cation channel translates into "photosynthesis works better" is not fully worked out; the authors themselves flag that the channel's precise mechanistic role remains unclear, and they note more broadly that Elysia has almost no established genetic tools, which limits how far any single lab can currently push a mechanistic story in this animal [1]. The correlational and pharmacological evidence is solid. The step-by-step biochemistry underneath it is still being filled in.

Grading the Kleptosome Against the Textbook

It is worth asking, plainly, whether "kleptosome" deserves to be called an organelle at all, or whether that is just a good name for an ordinary lysosome doing something unusual. Textbook organelle status generally rests on a handful of criteria, developed originally to make sense of how mitochondria and chloroplasts themselves went from free-living bacteria to permanent cell parts over evolutionary time [8]. Run the kleptosome through them.

Four of five, and the fifth is the one that matters most for calling this an organelle in the same sense a mitochondrion is one. A mitochondrion divides along with its host cell and is inherited down every lineage descended from that cell; a kleptosome's chloroplast is captured whole from an alga, never replicated by the slug, and eventually consumed. Every new generation of Elysia starts from zero and has to steal its own [4]. The paper's own language matches this reading closely: the authors describe the discovery as revealing mechanisms behind "the long-term acquisition and evolutionary incorporation of intracellular symbionts into organelles," language that positions the kleptosome as a working intermediate stage rather than a finished product [1]. It behaves like an organelle for as long as one animal is alive. It is not, yet, an organelle in the sense that gets inherited.

Hungry Season: Digesting the Solar Panel

Here is the finding that makes the whole system worth a second look. A structure built to keep a chloroplast alive and working is, mechanically, also a small membrane-bound sac full of lipid, protein and pigment. Starve the animal that built it, and there is no obvious reason the slug's own digestive machinery should leave that sac alone.

It doesn't. The paper's starvation experiment tracked Elysia crispata and, in parallel, the non-photosynthetic sea hare Aplysia californica as a control, both denied food and monitored over months. Within the first week, markers of the mTOR pathway — a conserved cellular sensor that shuts down growth signalling under energy stress — were already inactivated in the starved slugs, an early sign the animal had registered that it was running short [1]. By around six weeks, chloroplasts inside the kleptosomes were visibly dimmer under fluorescence imaging, the whole slug had begun to take on an orange cast as chlorophyll was lost, and expression of a core photosystem II gene, psbB, had dropped significantly (p = 0.003) [1]. Alongside the fading chloroplasts, the lysosomal marker Lysotracker began accumulating specifically on the kleptosomes undergoing that decline, while their Rab7 signal fell away and lipid droplets built up nearby — the signature, essentially, of a compartment being converted from a working organelle into a meal [1].

Elysia crispata kept this up for close to four months before finally succumbing to starvation. Aplysia californica, denied the same option because it never had chloroplasts to draw on, died within three to four weeks [1]. Figure 4 lays the two timelines side by side.

Club Arithmetic: How Much Time Does This Buy

The paper reports both survival windows as ranges rather than single numbers, so any ratio between them is necessarily an estimate, and it is entirely ours. Taking "close to four months" as roughly 16 to 17 weeks and "three to four weeks" at face value, the ratio lands somewhere between about 4× (17 divided by 4) and about 5.7× (17 divided by 3), depending which end of each range you pick. Call it, generously rounded, four to six times longer. That is the size of the metabolic reserve a slug builds simply by keeping its stolen chloroplasts around instead of digesting them the moment they arrive, and it is a genuinely large multiple for what amounts to a second food store sitting inside cells that are also doing something else entirely, all day, in the light.

One more piece of club arithmetic, smaller but worth stating plainly: a 62% reduction in oxygen output from blocking one channel means the remaining output is 100% minus 62%, or 38% — just over a third of normal. Framed the other way, two-thirds of this animal's solar income runs through a single gate that a pharmacologist can close with one compound.

Not Alone: Corals and Anemones Do It Too

Having found a Rab7-marked, arrest-and-digest mechanism in one photosynthetic animal, the authors went looking for it in others that keep living algae rather than stolen chloroplasts: the coral Lobophytum sp. and the sea anemone Exaiptasia diaphana, a standard laboratory model for coral symbiosis. Both cnidarians house their symbiotic algae inside compartments called symbiosomes, which the paper found are also marked by Rab7, and both showed a broadly convergent response under stress — described as a shared arrest mechanism operating "by modulation of lysosome biogenesis" — that parallels what the kleptosome does under starvation [1].

The comparison matters for reasons beyond tidiness. If a nearly identical solution shows up independently in molluscs and in cnidarians, two lineages that split from their last common ancestor long before either group evolved anything resembling photosynthetic partnership, that is evidence the trick is not a one-off accident of sacoglossan biology. It is more likely a fairly natural thing for an animal cell's existing endosomal toolkit to be repurposed into, whenever an animal lineage ends up needing to keep a photosynthetic passenger alive.

What's Still Missing

The authors are unusually candid about the limits of their own system. Elysia is, in their words, a wild-caught animal that is often difficult to source, with almost no established genetic tools or protocols, which limits how far the group could manipulate the animal directly rather than simply observe and pharmacologically perturb it [1]. Two consequences follow. First, the molecular switch that decides whether a given kleptosome keeps supporting its chloroplast or begins digesting it is not yet identified — the paper documents the correlates of that switch flipping (Lysotracker accumulation, Rab7 loss, lipid droplet buildup) without yet showing what flips it. Second, P2X4's mechanistic link to photosynthesis specifically, rather than to kleptosome physiology more generally, remains unclear by the authors' own account [1].

There is also the question this field note flagged early and never fully closes: kleptoplasty is not inherited, and nothing in this paper changes that. A kleptosome is real, host-built, host-regulated cellular hardware, doing real physiological work, for exactly one generation. Whether any lineage of sacoglossan is on a trajectory toward making that arrangement permanent, the way the ancestors of mitochondria and chloroplasts once did, is a question about deep evolutionary time that a study of living animals in a tank cannot answer on its own [8].

What We'd Want to See Next

Three things, roughly in the order we'd want them answered. First, the molecular trigger for the switch from maintenance to digestion — find what turns Rab7 loss and Lysotracker gain on, and you have a lever for testing whether the kleptosome's dual role is one regulated system or two independent processes that happen to share a compartment. Second, a direct test of P2X4's causal role in photosynthesis rather than a correlational one: does restoring channel activity in a 5-BDBD-blocked slug restore oxygen output on the same timescale, and does the channel's ATP sensitivity actually track the chloroplast's own energy state inside the kleptosome, as the "ATP-sensitive" framing implies? Third, and hardest: whether any related sacoglossan species has evolved even a partial route toward carrying kleptoplasty across a generation, which would be the first concrete step from "an organelle for one lifetime" toward "an organelle that stays."

References

1. 1.

   Allard, C. A. H., Thies, A. B., Mitra, R., Vaelli, P. M., Leto, O. D., Walsh, B. L., Laetz, E. M. J., Tresguerres, M., Lee, A. S. Y. & Bellono, N. W. (2025). A host organelle integrates stolen chloroplasts for animal photosynthesis. Cell 188, 5266–5277.e13. doi:10.1016/j.cell.2025.06.003

2. 2.

   Hinde, R. & Smith, D. C. (1972). Persistence of functional chloroplasts in Elysia viridis (Opisthobranchia, Sacoglossa). Nature New Biology 239, 30–31. doi:10.1038/newbio239030a0

3. 3.

   Green, B. J., Li, W.-Y., Manhart, J. R., Fox, T. C., Summer, E. J., Kennedy, R. A., Pierce, S. K. & Rumpho, M. E. (2000). Mollusc-algal chloroplast endosymbiosis: photosynthesis, thylakoid protein maintenance, and chloroplast gene expression continue for many months in the absence of the algal nucleus. Plant Physiology 124, 331–342. doi:10.1104/pp.124.1.331

4. 4.

   Rumpho, M. E., Dastoor, F. P., Manhart, J. R. & Lee, J. (2006). The kleptoplast. In The Structure and Function of Plastids (Wise, R. R. & Hoober, J. K., eds.), pp. 451–473. Springer.

5. 5.

   Evertsen, J., Burghardt, I., Johnsen, G. & Wägele, H. (2007). Retention of functional chloroplasts in some sacoglossans from the Indo-Pacific and Mediterranean. Marine Biology 151, 2159–2166. doi:10.1007/s00227-007-0648-6

6. 6.

   Mujer, C. V., Andrews, D. L., Manhart, J. R., Pierce, S. K. & Rumpho, M. E. (1996). Chloroplast genes are expressed during intracellular symbiotic association of Vaucheria litorea plastids with the sea slug Elysia chlorotica. Proceedings of the National Academy of Sciences 93, 12333–12338. doi:10.1073/pnas.93.22.12333

7. 7.

   North, R. A. (2002). Molecular physiology of P2X receptors. Physiological Reviews 82, 1013–1067. doi:10.1152/physrev.00015.2002

8. 8.

   Archibald, J. M. (2015). Endosymbiosis and eukaryotic cell evolution. Current Biology 25, R911–R921. doi:10.1016/j.cub.2015.07.055