Here’s what you’ll learn when you read this story:

• All known forms of life need at least 20 amino acids to survive, but there is speculation about whether or not this was true in the earliest phases of evolution.
• Determined to see if E. coli bacteria could survive with fewer amino acids, a team of researchers attempted to find viable ways to delete one using AI algorithms.
• The researchers managed to create E. coli bacteria with one amino acid missing from its ribosomal proteins, and it was able to reproduce for hundreds of generations.

Life—at least, as we know it—needs 20 amino acids, which it combines into the proteins that build living tissues. How life actually arrived at a minimum of 20 canonical amino acids (CAAs) in its journey from primordial ooze, however, is still a mystery. Some microbial species use up to 22 amino acids, but no species on Earth uses fewer than 20.

However, that may not have always been the case. Curious about how the precursors of life may have made it on a hostile young Earth before LUCA (the Last Universal Common Ancestor of all organisms) finally came into being, a team of researchers from Columbia University set out to see whether microbes emerging back then could have possibly run on fewer than 20 amino acids.

“Multiple lines of evidence suggest the feasibility of life with a more restricted amino acid alphabet,” the researchers said in a study recently published in the journal Science. “Several groups of amino acids have a high degree of biochemical similarity and potential functional redundancy […] [and] computational protein structure modeling suggests that only 9 to 12 amino acids are needed to construct nearly all protein folds.”

In a strand of DNA, each sequence of three nucleotides, known as a codon, encodes information for producing a particular amino acid. Adenine, cytosine, guanine, thiamine, and uracil are some of the most commonly known amino acids, but there are many others that fly under the radar in high school biology. Isoleucine, for instance, is one that usually doesn’t show up on final exams. In humans, it’s involved in both the metabolism and the immune system. It also requires an overwhelming amount of energy to synthesize, and because of this, it’s often been substituted for throughout evolution. As a result, when searching for an amino acid that E. coli bacteria might be able to live without, the research team decided that isoleucine would be a promising candidate.

Mutated E.coli bacteria deficient in a compound containing isoleucine have been found to keep themselves going with valine, which is similar enough to sometimes replace isoleucine in nature. Artificially making this replacement could create a microbe able to survive with just 19 amino acids. To make this replacement, the team had to figure out exactly which regions of the E.coli genome they needed to edit to create a modified species they called Ec19—a process that would require edits to be made at more than 81,000 sites on the genome.

With this in mind, the team decided to focus on something less intimidating and honed in on E. coli’s ribosomes, which synthesize proteins. Each ribosome is also made of 50 proteins itself. Through this process, isoleucine was replaced with with either valine or leucine in 39 E.coli genes that were either necessary or highly expressed.

Most cells that had isoleucine swapped out for valine unfortunately had trouble growing, or ended up dying. Seeing this, the researchers decided to train specialized AI algorithms (known as protein language models) to create new sequences of amino acids while predicting how those sequences would interact with each other and be incorporated into proteins. The trial was meant to test how many amino acids the researchers could factor out while keeping the organism viable.

The algorithms, according to Science, made suggestions for edits that resulted in “nonintuitive ribosomal protein sequences,” which were used to compensate for the missing amino acid. Isoleucine is naturally absent from two of the 52 ribosomal proteins in E.coli, so the researchers were hoping their AI-assisted process could offer viable restructuring ideas for the other 50 proteins. They decided to use some of these suggestions to test the limits of living E. coli bacteria in their lab. After creating several iterations that struggled to survive, they successfully designed a version of the E.coli bacteria in which 21 ribosomal proteins went without isoleucine. While the new strain grew at a slightly slower rate than your average E.coli bacteria, it managed to stay alive without “fixing” the alterations, and produced over 450 generations with the modified ribosome.

While this isn’t exactly the Ec19 they sought, as the organism still technically needs 20 amino acids—while it was systematically deleted from ribosomal proteins, most of the bacteria’s genome still needed to use isoleucine—the researchers are optimistic about redefining life with a 19-amino acid microbe in the future. It could lead to groundbreaking applications in biotechnology, synthetic biology, and materials medicine. A living example of a microbe that can survive on only 19 amino acids may also tell us more about the beginnings of life on Earth and hint at hypothetical life-forms that might exist on other moons and planets.

“Although there is still speculation regarding which amino acids were most recently fixed in the LUCA, our study suggests that the modern-day amino acid alphabet code could be reduced,” they said. “Our efforts represent the largest targeted sequence changes […] while introducing [significant] alterations.”