Ann Donnelly was utterly confused the first time she examined her protein. On all counts, it behaved like an enzyme—a protein catalyst that speeds up biological reactions in cells. One could argue that enzymes, sculpted by eons of evolution, make life possible.
There was just one problem: her protein wasn’t evolved. It wasn’t even “natural.” It was, in fact, a completely artificial construct made with random sequences of DNA—something that’s never existed in nature before.
Donnelly was looking at the first artificial enzyme. An artificial protein that, by all accounts, should not be able to play nice with the intricate web of biochemical components and reactions that support life.
Yet when given to a mutant bacteria that lacks the ability to metabolize iron, the enzyme Syn-F4 filled in the blank. It kickstarted the bacteria’s iron processing pathways, naturally replacing the organism’s missing catalyzer—even though it was like nothing seen in life.
“That was an incredible and unbelievable moment for me—unbelievable to the point that I didn’t want to say anything until I had repeated it several times,” says Donnelly, who published her results in Nature Chemical Biology.
The big picture? We are no longer bound by the chemical rules of nature. In a matter of months, scientists can engineer biological catalysts that normally take millions of years to evolve and fine-tune.
And with brand new enzymes comes the possibility of brand new life.
“Our work suggests that the construction of artificial genomes capable of sustaining cell life may be within reach,” says Dr. Michael Hecht at Princeton University, who led the study.
Cogs in the Machine
In 2011, Hecht was examining the limits of artificial biology.
At the time, many synthetic biologists had begun viewing biological processes as Lego blocks—something you could deconstruct, isolate, and reshuffle to build new constructs to your liking.
But Hecht was interested in something a little different. Rather than copy-and-pasting existing bits of genetic code across organisms, could we randomly build brand new molecular machines—proteins—from scratch?
Ultimately, it comes down to numbers. Like DNA, proteins are made up of a finite selection of chemical components: 20 amino acids, which combine in unique sequences into a chain.
For an average protein of 100 “letters,” the combinations are astronomically large. Yet an average cell produces only about 100,000 different proteins. Why is this? Do known proteins have some fundamental advantage? Or has evolution simply not yet had the chance to fashion even better workers? Could we tap into all that sweet potential?
A New Toolkit
Hecht and his group used a computer program to randomly generate one million new sequences. The chains were then folded into intricate 3D shapes based on the rules of biophysics.
The litmus test: one by one, the team inserted this new library of artificial proteins into mutant strains of bacteria that lacked essential genes. Without these genes, the mutants couldn’t adapt to harsh new environments—say, high salts—and died.
Remarkably, a small group of artificial proteins saved these mutants from certain death. It’s like randomly shuffling letters of known words and phrases to make new ones, yet somehow the new vocabulary makes perfect sense in an existing paragraph.
“The information encoded in these artificial genes is completely novel—it does not come from, nor is it significantly related to, information encoded by natural genes, and yet the end result is a living, functional microbe,” Michael Fisher, a graduate student in Hecht’s lab said at that time.
A series of subsequent studies showed that many of these artificial proteins worked by boosting the cell’s backup biological processes—increasing the expression of genes that allows them to survive under selection pressure, for example.
The lab thought they had it nailed, until one protein came along: Syn-F4.
The New Catalyst
Syn-F4 is a direct descendant of one of the original “new” proteins. Earlier, the team discovered that the protein could help mutant bacteria thrive in a low-iron environment—just not very well.
Mimicking evolution, they then randomly mutated some of the “letters” of the protein into a library, and screened them for candidates that worked even better than the original for supporting low-iron life. The result? Syn-F4.
Donnelly took on the detective work. Normally, scientists can scour the sequence of a newly discovered protein, match it up to similar others and begin guessing how it works. This was obviously not possible here, since Syn-F4 doesn’t look like anything in nature.
The protein also escaped all attempts at crystallizing it, which would freeze it in 3D and allow scientists to figure out its structure.
In a clever series of experiments, Donnelly cracked the mystery. Like baking soda, Syn-F4 sped up iron-releasing reactions when mixed with the right ingredients. What’s more, it’s also extremely picky about its “clients”: it would only grab onto one structural form of an ingredient (say, a form that looks like your left hand) but not its mirror image (the right hand)—a hallmark of enzymes.
Several more years of digging unveiled a true gem: Syn-F4’s catalytic core, a short sequence hidden in the protein’s heart that makes its enzyme activity possible.
Mutating the protein’s letters one by one, Donnelly tenaciously picked those that rendered the protein inactive. This process eventually identified key letters that likely form the protein’s so-called “active site,” splattered across Syn-F4’s sequence.
Like petals on a rosebud, the process of folding brings these active letters together into a 3D core. And like the Syn-F4 itself, its structure looks completely different than that of any native enzyme.
The team explains, “We don’t think Syn-F4 is replacing the mutant bacteria’s missing enzymes; we think it’s working through a completely different mechanism.”
“We have a completely novel protein that’s capable of sustaining life by actually being an enzyme—and that’s just crazy,” says Hecht.
A New Life?
The implications are huge, says Dr. Justin Siegel at the UC Davis Genome Center, who wasn’t involved in the study.
Biotechnology routinely relies on enzymes for industrial applications, such as making drugs, fuel, and materials.
“We are no longer limited to the proteins produced by nature, and that we can develop proteins—that would normally have taken billions of years to evolve—in a matter of months,” he says.
But even more intriguing is this: the study shows that enzymes made naturally aren’t the solution to life. They’re just one solution.
This means we need to broaden our search for biochemical reactions and life, on Earth and elsewhere. After all, if multiple solutions exist for a biological problem, it makes it much more likely that one has already been found elsewhere in the universe.
Back on Earth, Hecht is extremely excited for the future of artificial life.
“We’re starting to code for an artificial genome,” he says. Right now we’ve replaced about 0.1 percent of genes in a bacteria, so it’s just a weird organism with some funky artificial genes.
But suppose you replace 20 percent of genes, 30 percent, or more. Suppose a cohort of completely artificial enzymes runs the bacteria’s metabolism.
“Then it’s not just a weird E. coli with some artificial genes, then you have to say it’s a novel organism,” he says.
Image Credit: Ann Donnelly/Hecht Lab/Princeton University