“Probably all of the organic beings which have ever lived on this Earth have descended from some one primordial form,” Darwin wrote in his Origin of Species, published in 1859. Darwin had no way to peer that far back in time. But genome sequencing has given researchers hope that they can finally learn something about the ancestor of all life. In 1999, they even gave it a name, LUCA, for the last universal common ancestor1.

Yet despite the wealth of genomic data, LUCA has proven elusive. In theory, remnants of the organism from which all life evolved should be scattered around modern genomes. But so far, efforts to reconstruct LUCA's genes by building family trees from modern sequences have ended in frustration. Basic questions about LUCA's nature remain unanswered. Did it live in a hot-water environment, such as a hydrothermal vent at the bottom of the ocean, or in cooler conditions at the ocean surface? Was LUCA simple, like a bacterium, or more complex?

But there are now hints that some of these questions may be answerable. While some groups are zeroing in on LUCA's preferred temperature, others are targeting its genetic blueprint. From all this work, one of the more surprising theories to emerge may also help to explain why LUCA has been so hard to find. Perhaps it wasn't a single organism at all. Instead, most researchers now believe we should think of LUCA as a pool of genes shared among a host of primitive organisms.

“The naive picture that a group of organisms got all their genes from a simple last common ancestor is breaking down,” says microbiologist Gary Olsen of the University of Illinois at Urbana-Champaign. In its place, the image of a sophisticated, global community is emerging, he says. “In the past two years, it feels like it's fallen together into a coherent picture.” Rather than a last common ancestor, LUCA may have been a last common community.

Family fortunes

The search for LUCA began with the study of ribosomal RNA, an essential part of the cell's protein-making machinery. The genes for these molecules are highly conserved, meaning the sequences of their four bases (cytosine, guanine, adenine and thymine, which becomes uracil in RNA) are almost identical in all forms of life. Because these sequences have changed so little over time, they are ideal for building family trees, or phylogenies, of evolutionary splits that occurred billions of years ago.

The technique is powerful, but it has not yet given a definitive answer to the question of temperature. Some believe LUCA was ‘thermophilic’ or heat-loving. Not only was Earth probably warmer 3.5 billion years ago2, but the discovery in the 1970s of a new domain of life, the archaea, also suggested a hot-water origin3. These single-celled organisms share characteristics with the other two domains, the bacteria and the eukaryotes — more complex creatures such as fungi, plants and animals. In many phylogenies, the modern groups most closely related to the common ancestor of archaea and bacteria are hyperthermophiles that live in places where temperatures exceed 80°C—for example the hot springs of Yellowstone National Park in Wyoming and hydrothermal vents on the ocean floor.

Blowing hot and cold: genetic analysis has failed to resolve the question of whether the last common ancestor of bacteria, eukaryotes and archaeans lived in cool, shallow waters, or in a hot environment such as hydrothermal vents (above) or hot springs like those in Yellowstone National Park. Credit: VERENA TUNNICLIFFE

“No matter where you root the tree, all of the groups closest to the bottom are hyperthermophiles,” says microbiologist Michael Adams of the University of Georgia, Athens.

Others, however, think LUCA might have preferred the cooler surface waters of the ocean. They distrust phylogenies that point to hot origins because they reach so far back. Mutations build up and overwrite one another over such a span of time, erasing the phylogenetic signal. Also, phylogenetic analysis has some statistical quirks. For example, the gene sequences that evolve most quickly tend to come together on trees, even if they are only distantly related. A recent study based on slowly evolving sequences placed a non-thermophilic group at the base of the bacterial family tree4.

Looking at LUCA's ribosomal RNA could also reveal something about the temperature of its habitat. Organisms in hot environments have ribosomal RNA that is rich in pairs of the bases guanine (G) and cytosine (C). These G–C pairs are more tightly bound than the alternative, A–U, and are therefore more stable at high temperatures.

But so far the results have been split. One reconstruction of a putative ribosomal RNA sequence for LUCA suggests that it was a cool-water creature5. But this analysis omitted some thermophilic groups, says David Saul of the University of Auckland. He believes that there is a trend towards lower G–C ratios as one moves up the tree of life, showing that LUCA was a thermophile. So whether LUCA played in the surf or basked in the hot springs is still an open question.

It had been hoped that ribosomal RNA would also reveal which of the three domains of life — bacteria, eukaryotes or archaeans — LUCA was most like. It was not so simple. Building phylogenies relies on the principle that a bigger difference in sequences between two species means a more remote common ancestor. But the number of possible trees rises exponentially with each species added to the analysis. For deep phylogenies, which reach into the distant past, ribosomal RNA sequences from dozens of species are required for comparison, and the number of possible trees is astronomical. Researchers have devised mathematical techniques to find the most likely tree, but it is often difficult to choose between the many possibilities with any confidence.

Comparing several genes can make the choice easier, and over the past decade, high-throughput genome analysis has begun to make this feasible. More than a hundred organisms have now been fully sequenced, leading to a great deal of optimism that the roots of life might be traceable. After all, the genetic code and much of biochemistry is universal, so perhaps LUCA can be revealed by finding a set of genes for basic biological functions that is present in all organisms.

Universal genes

Oddly, this extensive comparison of genome sequences from widely divergent modern organisms has identified only about 60 genes that appear to be universal, and therefore probably date back to LUCA. That's nowhere near enough to sustain an organism, says Eugene Koonin, an evolutionary genomics researcher at the National Center for Biotechnology Information in Bethesda, Maryland. The majority of these genes are involved in translation6, the process of converting the sequence of bases in DNA into the sequence of amino acids in protein. “On these genes alone, LUCA would go nowhere,” Koonin says. “There is nothing for a cell membrane, or for energy metabolism, or any synthetic capabilities. There should have been several times more genes.”

According to some evolutionary biologists, the implications for LUCA are strange indeed. If a single LUCA laid the foundations for the modern diversity in membranes, metabolism and so on, it must have had several different versions of many important genes, in addition to the universal 60. Later lineages would each have pruned all but one from this set, giving rise to the current diversity in basic biochemical pathways. The idea that organisms become more complex rather than less as you get closer to the root of the tree of life is impossible to swallow, says Saul. A single LUCA “would have to have had the most bizarre biochemistry imaginable”.

At the same time, evidence was mounting that early life forms were particularly promiscuous in sharing their genes around, in a process called horizontal transfer. Among the genes that should be highly conserved—and therefore good for phylogenies — are those involved in handling genetic information, such as DNA polymerase, which copies DNA, and topoisomerase, which controls the structure of DNA. But the surprise result from this work was that the patterns of ancestry vary depending on which gene you look at7. In other words, the phylogenies revealed only the ancestry of the genes themselves, not the relatedness of the species that housed them. This showed genes had hopped between lineages.

For those trying to trace the tree of life back to its roots, the evidence that horizontal transfer had scrambled phylogenies was a serious blow. The impact on evolutionary biology was like “a fox in a hen house”, says Carl Woese, an evolutionary biologist at the University of Illinois in Urbana-Champaign.

Some biologists believe that horizontal gene transfer makes LUCA unknowable8. “Four billion years is plenty of time to scramble the phylogenetic record,” says evolutionary biologist Ford Doolittle of Dalhousie University in Halifax, Canada. “Life had achieved its modern cellular status much earlier than anything we can trace back.”

Olsen counters that, while trees vary from gene to gene, the average tree that emerges from comparing many genes at once is consistent with the results obtained for ribosomal RNA alone. In other words he believes the phylogenetic record has been eroded, but not erased.

Carl Woese: “I picture genes and their products flowing through a sea of cells” Credit: JASON LINDSEY

In 1998, the puzzles surrounding the nature of the first life forms led Woese to propose that the universal common ancestor was actually a community of organisms sharing genes9. In this communal world, the various primordial organisms had independently come up with solutions to similar problems, such as how to build and expand a membrane, or how to convert organic molecules into useful energy. All of the genes encoding these solutions were available to all the cells in the commune. “I picture genes and their products flowing through a sea of cells,” Woese says.

Plug-and-play genes

The genes in this early world would have been ‘modular’ — their products able to function on their own. Whereas today many cellular functions, such as DNA translation and replication, rely on complex machines encoded in several genes, nearly all the genes in LUCA's community would have been able to function by themselves, like cassettes that can be loaded, removed and replaced. Antibiotic-resistance genes are like that today. A bacterium need only acquire one to fend off an antibiotic.

Ultimately, around 3.5 billion years ago, the modern domains of life would have emerged from the gene-swapping mêlée with many of the genes from the last common community riding on their coat-tails. Inheritance and mutation would then have replaced gene transfer as the most important source of biological novelty as cells became more complex and their functions became less interchangeable. This point, says Woese, was the true origin of species, and so he has christened it the darwinian threshold10.

Woese's genetic and biochemical commune has become a leading theory of early life, but it has not entirely quashed the concept of LUCA. Koonin, for instance, is still working to provide LUCA with a coherent genome by expanding the list of genes it contained beyond the universal 60. He believes modern genes that are almost universal probably also date back to LUCA, and have simply been replaced in a few organisms. So these should be added to the list. It's trickier to tell with less-common genes, and bio-informatic techniques are not yet sophisticated enough to construct a full gene set for LUCA, Koonin says. But it will get easier as sequencing data continue to accumulate.

“The ultimate goal is to arrive at a most defendable reconstruction in terms of a set of genes,” he says. This set will number about 600, Koonin estimates, based on what contemporary genomes tell us about the minimum number of genes needed by a self-sufficient organism6.

Building a microbe

Once that set of genes is known, it might even be possible to create LUCA in a dish. Building a microbe may seem outlandish, but just such a project is already under way. In 2002, human genome sequencer Craig Venter announced his plan to build an artificial cell with a minimal genome based on modern genes at his Institute for Biological Energy Alternatives in Rockville, Maryland. And last year, a team led by Steven Benner of the University of Florida, Gainesville, used phylogenetic analysis to resurrect a protein from an ancient bacterium that lived around a billion years ago11.

If an artificial LUCA could survive in a laboratory dish, it might deal a blow to the notion of a communal ancestor that depended for survival on neighbours exchanging genes. That would not surprise Patrick Forterre of the Paris-Sud University in Orsay and the Pasteur Institute in Paris, who says the communal LUCA notion doesn't fit with the way evolution works. “To think of LUCA in terms of a community is to remove the idea of darwinism from early evolution,” he says. Although LUCA undoubtedly swapped genes with its neighbours, Forterre argues that it would also have competed with them and ultimately triumphed through some key innovation.

Late last year, Canadian mathematicians provided support for Forterre's case with a model suggesting that a ‘commune’ of protocells could not persist. Peter Antonelli and Solange Rutz at the University of Alberta in Edmonton wrote equations describing the competition for resources and the sharing of genes and biochemicals that would have gone on in such a world, and found them to be mathematically unstable. In other words, they say, the commune would have fallen apart12.

But Woese believes the critique says nothing about his hypothesis. The world of a communal LUCA has so far only been described in qualitative terms, he points out, so there are many mathematical models that could produce an approximate version of it. Also, the modern microbial world is full of bacterial communities that share resources, if not genes.

Of course, finding LUCA would not solve the puzzle of how life began. The idea of a last common community, with a communally sophisticated biochemistry, raises another question: how did all this evolve? This is for someone else to answer, says Woese. “We don't understand how to create novelty from scratch — that's a question for biologists of the future.”