In a First, Scientists Create Tiny Multicellular Organisms That Can Replicate

The latest version of xenobots can swim, heal, record information—and create new versions of themselves

In a remarkable development, scientists at Tufts University and the University of Vermont have created a new biological organism that can self-replicate.

The discovery builds on earlier work creating tiny automatons called xenobots from the skin cells of African clawed frogs (Xenopus laevis). The bots were able to swim through liquid, navigate through tubes, work together to collect particles into piles, heal themselves when injured, and even store information from their experience.

Now the team has designed xenobots that can make copies of themselves. “Some people have said in the past that xenobots are not organisms because they don’t replicate. Well, now they replicate,” said Michael Levin, A92, Vannevar Bush Professor of Biology and director of the Allen Discovery Center at Tufts, who did the work with Joshua Bongard, a professor of computer science at the University of Vermont.

The ability to self-replicate is important, because the first generation of this latest iteration of xenobots has to be created painstakingly by hand. Being able to have them then make hundreds or thousands of copies of themselves changes the game in terms of ultimately using them for human applications, from producing insulin to repairing spinal cord injuries, and any other application where progressively tinier living constructs have to be made.

The research team first unveiled xenobots in 2020, but had been trying to find a way to lead them to replicate. Using the Deep Green supercomputer cluster at the University of Vermont, Bongard and his graduate student Sam Kriegman (who has since earned his Ph.D. and moved to Levin’s lab for a postdoctoral position) ran evolutionary algorithms to test billions of body shapes in simulations to see which one would replicate not just one generation, but many generations, over and over.

The final shape that evolved from the algorithm was a semi-torus—something like a donut with a part of it cut away, or a Pac-Man with an open mouth.

With the computer-designed blueprints in hand, Levin and Douglas Blackiston, a senior scientist at the Allen Discovery Center, collected stem cells from the skin of frog embryos. Left in a saline solution, the cells naturally clump together and transform into a sphere of epidermis covered by a layer of cilia—little hair-like projections that whip around and enable the spheres move across a surface.

The first generation of these computer-designed replicating xenobots must be made by hand. Blackiston used a combination of squashing the spheres and then cutting away tissue using tiny microsurgical tools to sculpt the xenobots into the Pac-Man shape.

The little organisms swirled around the dish as predicted, and when a slurry of individual stem cells was added to the dish, the xenobots began to gather them into clusters, which in turn spontaneously took the shape of spherical xenobots capable of making more of their kind. In all, a Pac-Man shaped bot can lead to at least five generations of spherical bots.

“Going beyond the frog is certainly a goal,” said Blackiston. “We are looking into creating bots from mammalian cell types. The latter would be important to be compatible for potential medical applications.”

For human applications, one could envision creating bots that produce insulin, regulated by built-in sensors for glucose levels. “We could also put bots into a patient to home in on areas of tissue damage such as spinal cord injury, and release regenerative compounds to help with the repair process,” said Blackiston. “These bots would decompose and be cleared by the immune system, or meld into the area to become part of the scaffold healing the injury.”

The researchers “are very much focused now on creating guidance systems for the bots—sensors to help bring them to a target, or attract them to back a collection point after their work is done,” he added.

The Hardware and Software of Organisms

How is it that the frog cells can be so flexible in creating novel proto-organisms? “The analogy that I use is that of hardware and software,” said Levin. “What the genetics does is provide the hardware. It tells every cell exactly what components it gets to have. That biological hardware has remarkable versatility—we can assemble groups of cells to solve the problem of forming a functional creature in different ways.” The rules governing how the newly arranged cells work together is akin to the software.

Evolution has been programming these components “for millions of years, and as a result, a frog became the default ‘device.’ But those cells can be recombined and reprogrammed to create a new organism—the xenobot, and that doesn’t need millions of years of selection for xenobot functionality to learn how to swim, heal, record information, and replicate in a novel way,” Levin said. “We put the cells together and the tissue can literally figure that out in 48 hours—the length of time it takes the bot to form.”

According to the researchers, the cells are like modular blocks: they have features, and, depending on how you combine them, you end up with a super-feature, a behavior.

“You can create different behaviors by putting together different types of cells—they may even come from different species,” said Blackiston. “Eventually we would like to have a library of modules where you go to the freezer, pull out the features you want, hook them together and you have a designer organism.”

A Different Way to Reproduce

Strikingly, xenobots reproduce in a way never before observed in nature. Normally, simple living things propagate by, for example, dividing in two, using asexual budding, breaking up into fragments, or casting out spores, the smaller pieces growing into a full-size organism. Sexual reproduction is common to more evolved and complex organisms, from plants to humans, and viruses commandeer the molecular machinery within host cells to reproduce.

The new xenobots are naturally inclined to replicate by actively collecting and aggregating their cellular components into new multicellular copies. It’s a form of replication found in no other animals or plants, but, interestingly, it does occur among molecules.

Simple self-replicating molecules are likely to have played a central role in the origin of life. Chemists have demonstrated that the components of RNA can assemble and act as enzymes to assist in the assembly of more RNA from its components. Other, non-biological chemicals have been shown to create copies of themselves in this manner as well.

At some point, amino acids and peptides joined the party and more complex systems of molecules, and then organisms, were created, and they continued to evolve and perpetuate themselves. The challenges of their environments and selection for survivors that better handled those challenges led to the evolution of an astounding number of functions, from movement to storing information, sensing, and cognition.

The theory of a primordial soup giving rise to molecules that spontaneously make copies of themselves by assembly is well established and widely accepted. What researchers have done now is create a primordial chowder, where cellular organisms assemble copies of themselves from free-floating cells.

The cells in this case, from the African clawed frog, have been bred for millions of years to become part of a living, swimming, croaking frog. As skin cells, they have been specialized to protect the outside of the tadpole or frog, excreting mucous and toxins to prevent attack from pathogens and predators.

Placed in a different context, and initially shaped by hand, the cells can form a new proto-organism—a spheroid, an ottoman shape with four leg buds, or Pac-Man shaped coherent creature that can now carry out specific tasks, even tasks that the original organism, a frog, is unable to do.

In other words, the genetic plan that evolved over millions of years produces cells that do not have to make a frog. Groups of cells can make a xenobot which moves, collects material, and now, replicates in an entirely new fashion unseen before in biology.

Mike Silver can be reached at

Back to Top