Power, Sex, Suicide: Mitochondria and the Meaning of Life (23 page)

Although the importance of lateral gene transfer in bacteria has been recognized since the 1970s, we have only recently begun to appreciate the degree to which it can confound evolutionary trees. In some bacterial species, more than 90 per cent of observed variation in a population comes from lateral gene transfer, rather than the conventional selection of cells in clones or colonies. The transfer of genes between different species, genera, and even domains means that bacteria do not pass on a consistent core of genes by vertical inheritance, as we do to our children. This makes it embarrassingly difficult to define the term ‘species’ in bacteria. In plants and animals, a species is defined as a population of individuals that can interbreed to produce fertile offspring. This definition does not apply to bacteria, which divide asexually to form clones of supposedly identical cells. In theory, the clones drift apart over time as a result of mutations, leading to genetic and morphological differences sufficient to call ‘speciation’. But lateral gene transfer often confounds this outcome. Genes can be switched so quickly and so comprehensively that the cacophony obliterates all traces of ancestry—no gene is passed on to daughter cells for more than a few generations before being replaced by an equivalent gene from another cell with a different ancestry. The current champion is
Neisseria gonorrhoeae
: this recombines genes so quickly that it is impossible to detect any clonal groups at all: even the gene for ribosomal RNA, often claimed to represent the true phylogeny (lineage) of bacteria, is swapped so often that it gives no indication of ancestry.

Over time, such gene transfers make a big difference. Just to give a single example, gene transfer has produced two different strains of the bacterial ‘species’
E. coli
that differ more radically in their gene content (a third of their genome, or nearly 2000 different genes) than all the mammals put together, perhaps even all the vertebrates! The importance of vertical inheritance, descent with modification, in which the genes are
only
passed on to the daughter cells during cell division, is often ambivalent among bacteria. Imagine trying to work out our own provenance by examining the heirlooms passed
down in the family, only to discover that our ancestors were compulsive kleptomaniacs, forever pilfering each other’s family silver. As the branching ‘tree of life’ is based strictly on vertical inheritance—the erroneous assumption that the heirlooms only pass from parents to children—its veracity is open to question. Among the bacteria at least a network may be a better analogy. As one despairing expert put it, reflecting on the troubles of constructing a tree of life, ‘only God can make a tree’.

So why are bacteria so open-handed with their genes? It might sound like altruistic behaviour, sharing genetic resources for the good of the population as a whole, but it is not; it is still a form of selfishness, what Maynard Smith described as an ‘evolutionarily stable strategy’. Compare lateral transfer with conventional ‘vertical’ inheritance. In the latter case, if an antibiotic threatens a population of bacteria, and only a few cells have retained the genes needed to save their lives, then the rest of the unprotected population will die, and only the offspring of the tattered survivors can thrive to replenish the population. If conditions then change again, favouring a different gene, this surviving population too may be decimated. In swiftly changing conditions, only the cells that retain an enormous repertoire of genes will survive most exigencies, and they will be so large and unwieldy that they can be out-competed by bacteria able to replicate faster in the interim. Such streamlined bacteria, of course, may be threatened by any exigencies at all—but not if they are able to pick up genes from the environment; then they can combine speedy replication with the genetic resilience to cope with almost anything thrown at them. Bacteria that lose and gain genes in this way will thrive in place of either lumbering genetic giants, or bacteria that refuse to pick up any new genes at all. Presumably, the most effective way of picking up new genes is by conjugation, rather than from the dead bacteria whose genes may be damaged, so ultimately an apparently altruistic, though individually selfish, sharing of genes is favoured. Overall, then, we see the dynamic balance of two different trends in bacteria—the tendency to gene loss, which reduces the bacterial genome to the smallest possible size in the prevailing conditions; and the accumulation of new genes by means of lateral gene transfer, according to need.

I have cited examples of gene loss in bacteria like
Rickettsia
, and in the lab, but beyond the sparseness of their genome (the small number of genes and the lack of junk DNA), it is difficult to prove that gene loss is important in bacteria in ‘the wild’. But the importance of lateral gene transfer among bacteria also testifies to the strength and pervasiveness of the selection pressure for bacteria to lose any superfluous genes—otherwise they would not be under such an obligation to pick them up again. Despite taking up new genes, bacteria don’t expand their genomes, so presumably they must lose genes at the same rate. And they lose genes at this rate because the competition between cells within a
species (and between cells in different species) must continually reduce the genome to the smallest size possible in the prevailing conditions.

The upper limit of any known bacterial genome is about 9 or 10 million letters, encoding some 9000 genes. Presumably, any bacteria that acquire more genes than this tend to lose them again, as the time needed to copy the extra genes slows down replication without providing any countering benefits. This is a stark contrast between bacteria and eukaryotes. The more we learn about bacteria, the harder it becomes to make valid generalizations about them. In recent years, we have discovered bacteria with straight chromosomes, with nuclei, cytoskeletons, and internal membranes, all traits once considered to be unique prerogatives of the eukaryotes. One of the few definitive differences that hasn’t evaporated on closer inspection is gene number. Why is it that there are no bacteria with more than 10 million DNA letters, when, as we noted in
Chapter 1
, the single-celled eukaryote
Amoeba dubia
has managed to accumulate 670
billion
letters—67 000 times more letters than the largest bacteria, and for that matter 200 times more than humans? How did the eukaryotes manage to evade the reproductive constraints imposed on bacteria? The answer that I think gets to the heart of the matter was put forward by Tibor Vellai and Gábor Vida in 1999, and is disarmingly simple. Bacteria are limited in their physical size, genome content, and complexity, they say, because they are forced to respire across their
external
cell membrane. Let’s see why that matters.

Recall from
Part 2
how respiration works. Redox reactions generate a proton gradient across a membrane, which is then used to power the synthesis of ATP. An intact membrane is necessary for energy generation. Eukaryotic cells use the inner mitochondrial membrane to generate ATP, while bacteria, which do not have organelles, must use their external cell membrane.

The limitation for bacteria is geometric. For simplicity, imagine a bacterium shaped like a cube, then double its dimensions. A cube has six sides, so if our cubic bacterium had dimensions of one thousandth of a millimetre each way (1 µm), doubling its size would quadruple the surface area, from 6 μm
²
(1 × 1 × 6) to 24μm
²
(2 × 2 × 6)μm
²
. The volume of the cube, however, depends on its length multiplied by its breadth by its depth, and this rises eightfold, from 1μm
³
(1 × 1 × 1) to 8μm
³
(2 × 2 × 2). When the cube has dimensions of 1 μm each way, the surface area to volume ratio is 6/1 = 6; with dimensions of 2 μm each way, the surface area to volume ratio is 24/8 = 3. The cubic bacterium now has half as much surface area in relation to its volume. The same thing happens if we double the dimensions of the cube again. The surface area to volume ratio
now falls to 96/64 = 1.5. Because the respiratory efficiency of bacteria depends on the ratio of surface area (the external membrane used for generating energy) to volume (the mass of the cell using up the available energy) this means that as bacteria become larger their respiratory efficiency declines hyperbolically (or more technically, with mass to the power of 2/3, as we’ll see in the next Part).

This decline in respiratory efficiency is coupled to a related problem in absorbing nutrients: the falling surface area to volume ratio restricts the rate at which food can be absorbed relative to the requirement. These problems can be mitigated to some extent by altering the shape of the cell (for example, a rod has a larger surface-area-to-volume ratio than a sphere) or by folding the membrane into sheets or villi (as in our own intestinal wall, which is subject to the same need to maximize absorption). Presumably, however, there comes a point when complex shapes are selected against, simply because they are too fragile, or too difficult to replicate with any accuracy. As any spatially challenged plasticine modeller knows, an imperfect sphere is much the most robust and replicable shape. We aren’t alone: most bacteria are spherical (cocci) or rod-like (bacilli) in shape.

In terms of energy, a bacterial cell with double the ‘normal’ dimensions will produce half as much ATP per unit volume, while being obliged to divert more energy towards replicating the cellular constituents, such as proteins, lipids, and carbohydrates, that make up the extra cell volume. Smaller variants, with smaller genomes, will almost invariably be favoured by selection. It is therefore hardly surprising that only a handful of bacteria have achieved a size comparable with eukaryotes, and these exceptions merely prove the rule. For example, the giant sulfur bacterium
Thiomargarita namibiensis
(the ‘sulfur pearl of Namibia’), discovered in the late 1990s, is eukaryotic in size: 100 to 300 microns in diameter (0.1 to 0.3 mm). Although this caused some excitement, it is actually composed almost entirely of a large vacuole. This vacuole accumulates raw materials for respiration, which are continually washed up and swept away by the upwelling currents off the Namibian coast. Their giant size is a sham—they amount to no more than a thin layer covering the surface of a spherical vacuole, like the rubber skin of a water-filled balloon.

Geometry is not the only stumbling block for bacteria. Think again about proton pumping. To generate energy, bacteria need to pump protons across their external cell membrane, into the space outside the cell. This space is known as the periplasm, because it is itself bounded by the cell wall.
¹
The cell
wall presumably helps to keep protons from dissipating altogether. Peter Mitchell himself observed that bacteria acidify their medium during active respiration, and presumably more protons are free to disperse if the cell wall is lost. Such considerations may help to explain why bacteria that lose their cell wall become fragile: they not only lose their structural support but also lose the outer boundary to their periplasmic space (of course they retain the inner boundary, the cell membrane itself). Without this outer boundary, the proton gradient is more likely to dissipate, at least to some extent—some protons appear to be ‘tethered’ to the membrane by electrostatic forces. Any dispersal of proton gradient is likely to disrupt chemiosmotic energy production: energy is not produced efficiently. As energy production runs down, all other aspects of a cell’s housekeeping are forced to run down too. Fragility is the least of what we would expect; it’s more surprising that the denuded cells can survive at all.

While many types of bacteria do lose their cell wall during parts of their life cycle only two groups of prokaryotes have succeeded in losing their cell walls permanently, yet lived to tell the tale. It’s interesting to consider the extenuating circumstances that permitted them to do so.

One group, the
Mycoplasma
, comprises mostly parasites, many of which live inside other cells.
Mycoplasma
cells are tiny, with very small genomes.
M. genitalium
, discovered in 1981, has the smallest known genome of any bacterial cell, encoding fewer than 500 genes. Despite its simplicity, it ranks among the most common of sexually transmitted diseases, producing symptoms similar to
Chlamydia
infection. It is so small (less than a third of a micron in diameter, or an order of magnitude smaller than most bacteria) that it must normally be viewed under the electron microscope; and difficulties culturing it meant its significance was not appreciated until the important advances in gene sequencing in the early 1990s. Like
Rickettsia
,
Mycoplasma
have lost virtually all the genes required for making nucleotides, amino acids, and so forth. Unlike
Rickettsia
, however,
Mycoplasma
have also lost all the genes for oxygen respiration, or indeed any other form of membrane respiration: they have no cytochromes, and so must rely on fermentation for energy. As we saw in the previous chapter, fermentation does not involve pumping protons across a membrane, and this might explain how
Mycoplasma
can survive without a cell wall. But fermentation produces up to 19 times fewer ATPs from a molecule of glucose than does oxygen respiration, and this in turn helps to account for the regressive character of
Mycoplasma
—their tiny size and genome content. They live like hermits, with little to call their own.

The second group of prokaryotes that thrive without a cell wall is the
Thermoplasma
, which are extremophile archaea that live in hot springs at 60°C and an optimal acidity of pH 2. They would probably fare well in a British fish and chip shop, as their preferred living conditions are equivalent to hot vinegar. Lynn Margulis once argued that
Thermoplasma
may be the archaeal ancestors of the eukaryotic cell, on the grounds that they can survive ‘in the wild’ without a cell wall; but, as we saw in
Part 1
, stronger evidence supports the methanogens as the putative original host. When the complete genome sequence of
Thermoplasma acidophilum
was reported in
Nature
in 2000, it provided no evidence of a close link to the eukaryotes.