We don't know what all genes do—but does it matter?

Scientists have created a "minimal cell" with just 473 genes—such work could have huge benefits beyond biology

April 04, 2016
article header image
Read more: There is no intelligence gene... singular 

How many genes does it take to make an organism? One answer has just been supplied by researchers at the J. Craig Venter Institute (JCVI) in La Jolla, California, among them biotech entrepreneur Venter himself. They have created a “minimal cell” christened JCVI-syn3.0, based on the bacterium Mycoplasma mycoides but with a tailor-made, synthetic genome stripped of what seem to be all superfluous genes to leave just 473 of them. In the popular but somewhat misleading metaphor, you might say that these cells run on just 473 instructions.

I say “one answer” because there is nothing definitive about this number: it’s just the smallest viable genome so far identified. The question is probably ill-posed, since it depends on what you mean by “life.” Many viruses have much smaller genomes than JCVI-syn3.0, but they are not really autonomous living entities, depending instead on their ability to hijack the genetic machinery of the host organisms they invade.

And when the notion of a minimal cell like this was discussed at the first international meeting on synthetic biology (the engineering of “artificial” organisms) at the Massachusetts Institute of Technology in 2004, I asked the question how far it could be expected to go before an organism would become irredeemably frail through loss of all its defences against the slings and arrows of nature. I was told that certainly there was a risk of ending up with an organism that died “the moment you looked at it."

JVCI-syn3.0 doesn’t do that. It seems to thrive, given the right nutrients and environment: a colony doubles in size every three hours. The JCVI scientists, led by the formidably gifted biologists Clyde A. Hutchinson and Hamilton Smith, stripped down the M. mycoides genome by trial and error, having first discovered that their attempts to design such a genome on the drawing board didn’t work. “Our current knowledge of biology is not sufficient to sit down and design a living organism and build it,” Venter admitted to Science, where their work is published.

So instead they developed methods for chopping up the bacterium’s genome and removing segments piece by piece, allowing them to see which genes were essential and which were not. Their starting point was not natural M. mycoides but an earlier “synthetic” version, called JCVI-syn1.0. Five years ago Venter’s team showed that they could chemically synthesize this modified version of the M. mycoides genome and insert that DNA into a related bacterial species emptied of its own genome, thereby converting the host cells into a different (and arguably “artificial”) organism.

As well as exploring fundamental questions about the minimal genetic requirements for life, the JCVI scientists hope that a minimal cell will have practical uses. One of Venter’s motivating visions is to redesign bacteria and other relatively simple organisms (such as yeast and algae) so that they generate useful chemicals and materials—hydrogen or hydrocarbon fuels, for example, or raw ingredients for making materials and medicines. Vats of such organisms might then be cultured for an environmentally friendly form of chemical processing, much as some drugs (like insulin) are already made from genetically modified bacteria. The value of a minimal cell here is that it is much easier to do design and engineering with a simplified organism rather than having to juggle with larger, ungainly genomes in which many of the genes aren’t strictly necessary for the task in hand. In this way a minimal cell could supply a general-purpose “chassis” for making bespoke organisms with all kinds of useful functions.

That’s a bold and, some would say, inspiring vision. But there’s a sobering message from JCVI-syn3.0 too. The trial-and-error method of selecting genes led to a genome with 149 genes—about a third of the total—that have a function currently unknown. That was no doubt a big part of the reason why Venter’s attempts to design from scratch failed: you can’t hope to build a device if you don’t know what all the components actually do.

On the one hand this might seem like a simple gap to plug. Just as we don’t yet know what many of the genes in the “fully decoded” human genome do, so even for a small organism like M. mycoides we haven’t yet figured out the role of every gene. It’s just a matter of getting round to looking. Perhaps that will indeed prove to be the case: we will come to understand the role of these mysterious 149 genes one by one, and will then stand back and say “Ah, of course! That’s why they’re needed.”

But it’s also possible that it won’t be quite that simple. We might find instead that, even if some kind of biochemical role can be assigned to all 473 of JCVI-syn3.0’s genes, it won’t be at all obvious why this particular combination, and not some other, is viable. We might discover that the design principles for a living organism can’t readily be deduced from a bottom-up consideration of their component parts, any more than we can understand how birds flock by studying them individually. The viability even of the simplest known cells like Mycoplasma might turn out to depend on rules operating at a higher level, among the interactions between different genes. That would make life an emergent phenomenon not deducible from even a comprehensive understanding of its separate components. Personally, I suspect this will be the case.

Now read: There is no "gene for" everything