A More Perfect Union.

Genetic studies show how insects and bacteria within them have teamed up

Sometime between 250 to 150 million years ago, a period beginning before the dinosaurs appeared and lasting well into their reign, a bacterium took up residence inside the body of an insect. The bacterium may have been friend, foe, or neither to its host.

At first, as the bacterium reproduced, its offspring may have moved back and forth, living sometimes within the insect, sometimes outside. But somewhere along the line, its descendants made an irrevocable choice. Forsaking life in the outside world, they opted for a permanent existence in the safe, food-rich interior of insect cells. In turn, the insect reaped benefits from the bacteria.

Although many bacteria that live inside animal cells cause disease, some are endosymbionts, organisms that live within other organisms without harming them. Many insects with nutritionally unbalanced diets, such as sap-sucking aphids, wood-eating cockroaches, carpenter ants, and blood-sucking tse-tse flies, have formed such partnerships with bacteria. Often, the lifestyles of the two organisms are so entwined that neither can survive without the other--the bacteria trading freedom and nutrients for a pampered life inside the insect's cells.

Did endosymbiont bacteria begin as benign invaders or as dangerous bacteria that grew harmless and then helpful?

Until the past several years, much about the private lives and evolution of the most specialized endosymbionts remained hidden. Scientists couldn't study these bacteria in a petri dish because most can't survive life outside their particular host. "Before 1990, there was a ton of speculation," says Nancy Moran of the University of Arizona in Tucson. "People didn't really know could've come from above what organisms you had [inside insects], and they didn't have their histories." Her lab, in cooperation with Paul Baumann's team at the University of California, Davis, has worked to establish basic facts about endosymbionts' lifestyles.

The past 2 years have seen a flurry of investigations from these labs and others using genetic sequencing, analysis of gene functions, and studies of bacteria more loosely associated with insect hosts than the most sophisticated endosymbionts are. These new findings are revealing how endosymbionts originated and evolved into their current roles. Genome changes like those of pathogenic bacteria seem to have fueled the process, argue Moran and Harold Ochman, also of Arizona, in the May 11 SCIENCE. That is, endosymbionts have lost some genes and gained others to live inside other organisms.

"What you have is domestication of the bacteria. They have become little factories for essential amino acid manufacture," says Baumann. Among the scarce nutrients that these microbes excrete are amino acids that insects require for building protein but don't manufacture themselves.

In this ancient and odd story of insects' taming bacteria for their own use or perhaps vice versa, the two organisms forge an unusually close alliance. Among the most specialized and well-studied of these evolutionary partnerships is the pea aphid Acyrthosiphon pisum and its endosymbiont, Buchnera. The pea aphids, which feed on sap, have special cells clustered around their gut to house the internal livestock. The bacteria never leave the protective body of their host, and the insects pack their eggs with starter kits of bacteria that nourish the aphid.

In the Sept. 7, 2000 NATURE, a group of researchers at three Japanese institutions published the genome of Buchnera, the first complete genetic sequence of an endosymbiont. The scientists showed just how much this domesticated Buchnera has changed from free-living bacteria. Overall, the Japanese team reports, the Buchnera genome is tiny, only one-seventh the size of the genome of Escherichia coli, a free-living bacterium that can inhabit the human gut.

The dramatic differences between E. coli and Buchnera suggest that the Buchnera genome has changed substantially through its so-called domestication. By comparison, animals domesticated by people can often interbreed with their wild relatives, which suggests that the genomes remain similar.

Although the domesticated microbe gained genes to produce more and more of the amino acids that the aphids require, it also lost many of the traits its ancestors had needed to live on their own. "In Buchnera, the amazing thing is that they keep all the pathways for the essential amino acids and lose all the nonessential amino acids that there is no shortage of [in the host]," says Moran.

Not only did Buchnera keep the genes for essential amino acids, but it added extra copies of those genes, earlier studies found. Moran and Jennifer Wernegreen, formerly of Moran's lab but now at Woods Hole (Mass.) Marine Biological Laboratory, have found that Buchnera carry the many extra copies of the essential amino acid genes on self-replicating circlets of DNA called plasmids. These extra copies enable the endosymbionts to produce quantities of amino acids for their host far in excess of what they themselves would ever need.

As Buchnera began producing more essential amino acids, the bacterium lost genes for other molecules required for independent living. For example, it lost genes--such as those encoding part of its outer membrane--required to defend itself against threats from the outside world, the Japanese team and others found. The researchers confirmed that the bacteria had also lost many regulatory genes, making it incapable of turning off the flow of essential amino acids. Similarly, a domestic sheep overproduces wool, a product people want, but it can no longer defend itself from predators as its ancestors did.

Wernegreen expects that as more endosymbiont genomes are sequenced, they will show a consistent pattern of gene loss, just as disease-causing bacteria living permanently inside cells lose a consistent set of genes.

A few researchers have looked at other endosymbionts. Colin Dale of Moran's lab points out that Wigglesworthii, a bacterium within the blood-feeding tse-tse fly, underwent genome reduction as drastic as Buchnera's.

An even more degraded genome belongs to Carsonella, the endosymbiont living inside jumping plant lice, or psyllids. Marta A. Clark of Baumann's lab and her colleagues found that Carsonella not only has fewer genes than Buchnera but the genes overlap each other, so the tail end of one gene may serve as the beginning of the next one.

Within endosymbionts, many of the genes have mutations that disrupt their normal function, Wernegreen and Moran report. These bacteria may have accidentally lost genes for proteins that help correct deleterious mutations, so genetic flaws can multiply, says Baumann. Once a gene has mutated beyond its usefulness, the bacterial offspring don't suffer further from a random deletion of the gene during reproduction. So, genes may first become damaged and then are jettisoned.

In Buchnera and other endosymbionts, decay and loss of genes seems to signal more than adaptation to the easy life of a kept bacterium. It may also result from endosymbionts' isolation inside the cells of the aphid. The mixed populations of different strains and species of wild bacteria that might live in a puddle or a person's gut frequently swap bits of genetic material. By doing this, they can replace genes or compensate for faulty or lost genes.

Jonas Sandstrom, formerly of Moran's lab and now at Uppsala University in Sweden, suggests that an endosymbiont's isolation may be a one-way ticket to extinction. Once the bacterium loses genes, he points out, it has no way of getting them back. It can't, therefore, evolve away from its special function and back toward an independent life.

Sandstrom has been studying what happens as genes disappear in Buchnera. "You paint yourself into a corner," he says. "As the genome of the endosymbiont gets smaller and smaller, it can provide less and less function." Eventually, the bacterium may lose its ability to function even in its already limited endosymbiotic role.

Wernegreen and Moran saw an example of this in a species of Buchnera that has faulty copies of the gene for making tryptophan, one of the aphid's essential amino acids. Aphids housing such debilitated bacteria are thus confined to plant species that provide the missing nutrient. "I think sometimes aphids suffer because populations of their bacteria lose genes," says Moran.

"Bacteria basically weren't designed to live in this kind of niche," says Dale. "They should be where they are able to undergo recombination. They are supposed to live in communities."

Instead, endosymbionts are inextricably tied to their hosts. Daniel J. Funk, now at Vanderbilt University in Nashville, along with Moran, Wernegreen, and Louise Hebling of the University of Arizona, sampled endosymbionts from four different aphid populations of a single species. They found no sign of bacterial trading within the species.

Even within a population of aphids, Buchnera bacteria seemingly don't move around to achieve so-called horizontal transmission. Every aphid essentially carries an isolated population of bacteria, which passes only vertically--from mother to offspring. In earlier work, Moran, Baumann, and others showed that as aphids evolved into different species, their bacteria accompanied them, evolving into new species in the same branching patterns as those of their hosts.

Moran speculates that detrimental mutations might "cause the whole system to melt down." That doesn't seem to happen, however, since the aphids and their bacteria have been around so long, she points out. Evolutionary pressure on the aphids may weed out those individuals hosting unfit bacteria, such as the strain with the tryptophan mutation.

To learn how endosymbionts may persist in insects, Dale and Sandstrom have looked at the so-called secondary endosymbionts of aphids and tsetse flies.

These are bacteria whose presence isn't necessary to the survival of the insect and whose benefit to the host, if any, is not obvious. These bacteria can propagate vertically by invading the eggs or embryos of the host in the body. Many secondary endosymbionts, such as those in aphids, seem to retain considerable independence and move between different kinds of insects. Others, such as Sodalis glossinidius, a secondary endosymbiont of the tsetse fly, stay within that insect species.

The secondary endosymbionts might reflect a lifestyle that primary endosymbionts showed earlier in their evolutionary history, Dale and Sandstrom speculate.

"In secondaries, it seems as if the process is just beginning. You can see evidence for a kind of evolutionary continuum," says Dale. "As the primary's [genome] degrades more and more, the secondary endosymbionts which live in the same tissues could move into that niche."

If that's the case, the secondaries may be able to shed some light on the evolution of all endosymbionts. Any bacterium that eventually became a primary endosymbiont would first need a means for getting into the insect and then a way to invade the cells of its host, Dale says. Pathogens have these means, as do many secondaries.

For example, Dale has just found genes in Sodalis that are very similar to those in bacteria that cause the diseases typhus and dysentery. He suspects that the molecules encoded by these genes act like a crowbar permitting the disease-causing bacteria to force their way into cells. This secondary endosymbiont might therefore have arisen from a pathogen that had evolved to live as a parasite inside the insects' cells, he suggests.

Dale describes secondary endosymbionts as "learning to become mutualists," partners that both help and gain from the insect. Sometimes, he says, these mutualists-in-training show their inexperience as partners by propagating so rapidly that they rupture the host cell. A primary endosymbiont, relying completely on the host for survival, never endangers the host in this way, Dale says.

Sandstrom and Wernegreen are more dubious of the pathogen-to-endosymbiont evolutionary route. A pathogen takes the amino acids that it needs directly from the host's body, so bacteria like Rickettsia, which causes typhus, tend to lose the genes for making these nutrients on their own.

How would a pathogen become a symbiont, wonders Sandstrom, if it had already lost the genes that would enable it to offer anything of value to the host? "The fact that [endosymbionts] maintain genes that are beneficial to the host suggests that they didn't begin as pathogens," agrees Wernegreen.

She mentions one instance in the plant kingdom where this sharp distinction doesn't seem to hold, however. Rhizobium is an essential endosymbiont that fixes nitrogen for legumes. But this beneficial bacterium has virtually all the genes found in Agrobacterium, a pathogen. A plasmid carries the lifestyle-altering genes that make Rhizobium an endosymbiont instead of a parasite, says Wernegren.

The precursor of an insect endosymbiont could have been an opportunistic bacterium with a large genome and thus many lifestyle options, Sandstrom and Wernegreen speculate. With genes that help itself into the cell and then benefit the insect, this ancestor would likely have been at home in many environments.

Thus this ancestral bacterium' might then be an opportunistic pathogen, sometimes lethal to insects and sometimes benign. It would have grown more frequently benign as the useful association with the insects developed. There is genetic evidence that Buchnera evolved from an ancestor it had in common with E. coli and that of E. coli is really an opportunistic pathogen, says Dale.

"In many ways, this works like an addiction system. The bacterium initially needs to harbor [pathogenic] genes to invade, but eventually the insects become addicted to the bacteria for the nutrients [the insect] needs," suggests Dale. After that, says Dale, the insect accepts the bacterium, which can lose its pathogenic genes.

Bacteriologist Baumann says he's leery of much speculation on the evolutionary origins of the endosymbiont lifestyle. Researchers plan to tease apart and compare several more endosymbiont genomes and those of secondary endosymbionts, says Moran. These will help unravel the origins of these bacteria.

But their story may not end here. By understanding the origins of endosymbionts, is it possible to say how they and their genomes will evolve? In a review article in the Nov. 30, 2000 CURRENT BIOLOGY that followed the publication of the Buchnera genome, Jan O. Andersson of Dalhousie University in Halifax in Nova Scotia explores the controversial idea that the genomes of Buchnera and other endosymbionts could be heading toward a permanent fusion with their host cells and may eventually come to resemble organdies.

Mitochondria in animal cells and chloroplasts in plants are examples of organelles. They perform tasks essential to the cell and the organism. Most biologists agree that each of these two organelles probably arose from an ancient bacterium infecting larger single-celled organisms (SN: 4/18/98, p. 253). Andersson points out that the genomes of mitochondria and chloroplasts show the same degeneration and gene loss as those of the endosymbionts do.

Despite the attraction of this intriguing idea, it seems to be wrong for most cases, asserts Moran. Unlike organelles, endosymbionts don't co-opt the cell's reproductive machinery for their own replication or insinuate themselves in every cell of the body. These capabilities only evolved twice--with mitochondria and chloroplasts, she says.

The ultimate fate of endosymbionts may not be such immortality through incorporation into the host, Dale suggests. Rather, as secondary endosymbionts wielding vital genomes replace the shrinking genomes of the primaries, he says, it may be the individual genes for indispensable functions that achieve immortality through successive waves of their bearers.