For nearly a billion people each year, parasitic diseases like sleeping sickness and leishmaniasis represent a quiet, persistent burden that medicine has struggled to address at scale. Over four years, an international consortium led by Vamsi Mootha mapped the complete protein architecture of five parasite mitochondria — the cellular engines that sustain these organisms — seeking the molecular vulnerabilities that could be turned into weapons against multiple diseases at once. What they found was not only a set of 33 promising drug targets absent from human cells, but a deeper revelation: that
Scientists map parasite mitochondria, uncovering 33 potential drug targets for neglected diseases
Human mitochondria are actually the oddball
So the core finding is 33 protein families that could be drug targets. But how confident are we that a drug targeting one of these will actually work?
That's the honest answer—we don't know yet. These are candidates, not validated targets. The team found them by looking for proteins shared across multiple parasites but absent from human mitochondria. That's a good filter, but it's still early.
Right, and we should be clear: they identified the targets using computational methods and one experimental technique. They haven't tested whether blocking any of these proteins actually kills the parasites in a lab dish, let alone in a patient.
True. But the precedent matters. Atovaquone works by blocking a mitochondrial protein in malaria parasites, so we know the strategy is sound. This is the next step—finding more targets that could work the same way.
What surprised you most about what they found?
That some of these parasites kept ancient mitochondrial machinery that we lost. Giardia's mitochondria don't even have DNA anymore, but they still have 59 proteins doing things we don't fully understand. It rewrites what we thought mitochondria had to be.
Though we should note—they mapped the proteins, but they don't know what half of them do. That's not a criticism; it's just the frontier. The map is more complete than before, but it's not complete.
The AI model predicting proteins in 200 organisms—how reliable is that?
It's trained on real experimental data from six organisms, so it has a foundation. But predictions are predictions. The value is in narrowing the search space for future researchers, not in replacing experimental work.
And the model let them reconstruct what the earliest mitochondrion might have looked like. That's elegant, but it's a hypothesis built on a model built on six data points. Interesting, not definitive.
So what happens next?
Researchers will start testing these 33 protein families as drug targets. Some will work, some won't. But now they have a roadmap instead of guessing in the dark.
And the data is public, which means the field can move faster. That's the real win here—not the answers, but the tools to find them.
Il Polso
- Nearly a billion people annually bear the weight of diseases that existing drugs address poorly, creating urgent pressure to find new therapeutic strategies that work across multiple parasitic infections simultaneously.
- The core technical obstacle — developing a single reliable method to inventory mitochondrial proteins across five evolutionarily distant organisms — consumed more than a year before protein correlation profiling proved equal to the task.
- Thirty-three protein families shared across parasite mitochondria but absent from human cells now stand as concrete targets, raising the possibility of drugs that could strike several diseases with a single compound.
- The data shattered textbook assumptions: Giardia's mitochondria carry no DNA at all, while other parasites run backup energy systems that function without oxygen, revealing human mitochondria as the stripped-down exception rather than the standard.
- An AI model trained on the experimental data extended predictions to nearly 200 eukaryotic organisms, accelerating both drug discovery timelines and the reconstruction of how mitochondria evolved across a billion years of life.
For nearly a billion people each year, parasitic diseases like sleeping sickness and leishmaniasis represent a quiet, persistent burden that medicine has struggled to address at scale. Over four years, an international consortium led by Vamsi Mootha mapped the complete protein architecture of five parasite mitochondria — the cellular engines that sustain these organisms — seeking the molecular vulnerabilities that could be turned into weapons against multiple diseases at once. What they found was not only a set of 33 promising drug targets absent from human cells, but a deeper revelation: that these ancient parasites carry mitochondrial machinery our own cells discarded long ago, rewriting what we thought we knew about the evolution of life's most fundamental power source.
Vamsi Mootha began with a straightforward premise: to kill a parasite, cut off its power. That instinct drove a four-year international effort, involving 25 scientists across Harvard, the Broad Institute, Massachusetts General Hospital, and Boston University, to map every protein inside the mitochondria of five parasites responsible for sleeping sickness, giardiasis, leishmaniasis, and babesiosis — diseases that collectively sicken nearly a billion people each year.
The methodological challenge was steep. The team spent over a year testing approaches before settling on protein correlation profiling, a technique that breaks open cells and uses mass spectrometry to track which proteins consistently travel alongside known mitochondrial components. The result was nine published papers and the most detailed cross-species map of parasite mitochondrial architecture ever assembled.
The practical ambition was drug discovery. Researchers already knew mitochondria could be targeted — atovaquone kills malaria parasites by blocking a single mitochondrial protein. Mootha's team searched for protein families shared across multiple parasites but absent from human cells, finding 33 candidates that could theoretically anchor drugs effective against several diseases at once. He was careful to note the long road from target to treatment, but called it a concrete first step.
The deeper surprise was evolutionary. Giardia's mitochondria have shed their DNA entirely, retaining just 59 proteins. Acanthamoeba's carry a backup enzyme allowing energy production without oxygen. Rather than aberrations, these were glimpses into ancient biology that human mitochondria had long since discarded. In Mootha's framing, our own mitochondria are the specialized outlier — the parasites preserve the older, more versatile form.
An AI model trained on the dataset predicted mitochondrial proteins across nearly 200 eukaryotes and helped reconstruct what the mitochondrion of the last common ancestor of all eukaryotes may have looked like — not simple, but metabolically sophisticated. All data has been made publicly available, offered as a foundation for the next generation of researchers to build upon.
Vamsi Mootha had a simple idea: if you want to kill a parasite, cut off its power supply. That thought anchored a four-year international effort to map every protein inside the mitochondria of five parasites responsible for some of the world's most stubborn tropical diseases. The parasites—those causing sleeping sickness, giardiasis, leishmaniasis, and babesiosis—sicken nearly a billion people annually. Together with a consortium of 25 scientists from Harvard's School of Public Health, Harvard Medical School, the Broad Institute, Massachusetts General Hospital, and Boston University, Mootha's team set out to catalog the complete protein architecture of these cellular power plants, spanning a billion years of evolutionary history.
The technical challenge was formidable. The researchers needed a single, reliable method that could work across five different organisms without sacrificing accuracy. They spent more than a year testing approaches before settling on protein correlation profiling: breaking apart each organism's cells, then using mass spectrometry to track which proteins moved together with known mitochondrial components. Proteins that consistently traveled with the mitochondria, batch after batch, could be confidently added to the inventory—even ones that had never been linked to the organelle before. The payoff was nine published papers and a detailed map of mitochondrial architecture across organisms that had never been studied this way.
The immediate practical goal was drug discovery. Scientists already knew that targeting mitochondria worked: atovaquone, a malaria drug, kills the parasite by blocking a single mitochondrial protein. Mootha's team searched for protein families shared across the mitochondria of trypanosomes, Leishmania, Acanthamoeba, and Babesia—but crucially, absent from human mitochondria. Finding such targets meant a drug could theoretically kill multiple parasites without poisoning the patient. The team identified 33 protein families that fit this profile, each one a potential starting point for a medication that could treat several diseases at once. Mootha was careful to emphasize that these were early findings; the distance from a promising target to a safe, effective drug remains long. But it was a concrete first step.
What emerged from the data challenged fundamental assumptions about how mitochondria work. The textbook version—aerobic powerhouses that carry their own DNA—holds true for humans and yeast, but it tells only part of the story. Giardia's mitochondria have abandoned DNA entirely, retaining just 59 proteins, more than half of unknown function. Acanthamoeba's mitochondria, by contrast, are unexpectedly sophisticated: they run on oxygen like human mitochondria, but also carry a backup enzyme called hydrogenase that lets them produce hydrogen gas when oxygen grows scarce. These weren't aberrations. They were windows into evolutionary history. The parasites had retained ancient mitochondrial machinery that human mitochondria had discarded over millions of years. In a sense, Mootha observed, human mitochondria were the oddball—stripped down and specialized, while these parasites carried the more flexible, versatile versions from the deep past.
The team used their experimental data to train an artificial intelligence model that predicted mitochondrial proteins across nearly 200 eukaryotic organisms, most of which had never been studied directly. This allowed them to reconstruct the gains and losses that shaped modern mitochondria and to model what the mitochondrion of LECA—the last common ancestor of all living eukaryotes—might have looked like a billion years ago. The result was surprising: LECA's mitochondrion was not simple or stripped down, but sophisticated and metabolically flexible, capable of generating energy whether oxygen was present or absent. The team also found that when different lineages independently lost their mitochondrial DNA, the rest of the organelle shrank too, sometimes dramatically, suggesting that DNA loss and organellar simplification were linked processes.
Mootha and his collaborators have made all their data publicly available, hoping it becomes a lasting resource for the field. The goal, he emphasized, was not simply to create inventories for their own sake, but to enable future researchers to work faster and think deeper. The map of parasite mitochondria is now in the hands of the scientific community, waiting to be mined for the next generation of drug candidates and evolutionary insights.
Citazioni salienti
A lot of these pathogenic protists have common pathways that were, evolutionarily, part of the first mitochondria on the planet. Our human mitochondria have lost them.— Vamsi Mootha, HHMI investigator
The goal of these inventories isn't just to have inventories. It's to make really awesome science possible.— Vamsi Mootha