Biogeography for Bacteria

See allHide authors and affiliations

Science  15 Aug 2003:
Vol. 301, Issue 5635, pp. 925-926
DOI: 10.1126/science.1089242

Almost all macroscopic animal and plant species have a limited distribution range on the surface of Earth—not only because they require particular habitats or climates, but also because of historical contingencies. Over geological time, taxonomic groups have evolved within continents or even within the confines of lakes, mountain ranges, or oceanic islands. Members of these groups have remained there as a result of natural barriers to migration. Microbes seem to be different. In 1913, the Dutch microbiologist Beijerinck concluded that any bacterial species can be found wherever its environmental requirements are met: The distribution of microbes requires no historical explanation, but can be understood solely in terms of habitat properties. Most taxonomic experts have implicitly assumed that this also applies to unicellular eukaryotes (protozoa, microalgae), but the topic has recently drawn renewed interest (1). For sexual, outbreeding eukaryotes, a theoretically based species concept exists, at least in principle, and so “cosmopolitan distribution” can be given a precise meaning—something that is less obvious in the case of bacteria. The explanation for cosmopolitan distribution of microbes is numbers: Microbial population sizes are enormous. Thus, the probability of dispersal is high and the probability of local extinction is extremely low. In the absence of effective migration barriers and local extinctions, every habitat will contain a pool of bacterial species that do not thrive locally, but may grow if the environment becomes more favorable. It has been calculated that about 1018 viable bacteria annually are transported through the atmosphere between continents (2). It is also possible to isolate bacteria from places where they “should not be” such as thermophilic bacteria from cold seawater (3). Indeed, there is evidence that species of aquatic and soil microbes are found wherever their needs are met. This view is now challenged by Whitaker et al. on page 976 of this issue (4) and by Papke et al. in a recent issue of Environmental Microbiology (5) with their studies of thermophilic prokaryotes from geothermal springs.

Whitaker et al. (4) studied the archaebacterium Sulfolobus solfataricus. This organism thrives in water and mud surfaces of hot springs at temperatures between 50° and 87°C and at a pH of about 4. It makes its living oxidizing elemental sulfur to sulfate or metabolizing organic matter. Strains were isolated from geothermal springs in Kamchatka (Russia), Yellowstone National Park and Lassen National Park (North America), and Iceland. Nominally the strains belong to the same species because they have almost identical 16S ribosomal RNA (rRNA) sequences. However, when the authors sequenced nine protein-coding loci they found that genetic distances between populations increased proportionally with geographic distance. The result is not fortuitous as several individual cell lines were sequenced at each location. Although the genetic distances are slight, results indicate that there is not a continuous rapid exchange of cells between these remote habitat patches.

Papke and colleagues (5) studied two groups of thermophilic cyanobacteria: the unicellular Synechococcus and the colonial, filamentous Oscillatoria. These are green organisms that undergo photosynthesis, producing oxygen; in hot springs they form microbial mats on surfaces at temperatures up to about 70°C. Material was collected in hot springs from Yellowstone National Park (see the figure), Oregon, Japan, Italy, and New Zealand. Sequencing rRNA genes revealed a number of lineages, each consisting of a cluster of genotypes. These clusters perhaps may be considered as “species.” The results do show some geographical patterns. Notably, representatives for two of the clusters were found only at North American sites, but in other cases, strains from, for example, New Zealand and Japan were clustered together. For both studies, genotypes do not correlate with their occurrence in particular microhabitats (water chemistry, temperature, pH). Phenotypically the strains appear identical within the genotype clusters, showing that genetic variation does not necessarily indicate adaptive fine-tuning and that the recorded variation probably reflects neutral or near-neutral mutations.

Not feeling the heat.

Octopus Spring in Yellowstone National Park provides a favorable habitat for thermophilic bacteria. [Photograph courtesy of Michael Kühl]

The two reports may show that there is limited exchange of cells between the study sites. The small total area constituted by geothermal springs on Earth's surface, their mutual remoteness, and the fact that the specialized obligate thermophiles succumb under more normal ambient temperatures are unusual properties of microbial biota and may explain the limited transfer of cells from one site to the other. It may be the exception that supports the rule.

But the studies are, perhaps, not so easy to interpret. In panmictic (randomly mating) populations of outbreeding organisms, genetic variation between individuals is constrained although there are normally no two genetically identical individuals. When species populations are structured into metapopulations or have wide geographical ranges, spatial genotypic variation occurs as a result of local selection pressure and genetic drift, irrespective of limited gene flow between subpopulations. But genetic differentiation and speciation do require some degree of genetic isolation. Bacteria have clonal evolution (occasional horizontal gene transfer, notwithstanding), and there is no such constraint on genetic diversification. The descendants of a given cell will over time continuously accumulate neutral mutations, and different isolates of nominal species often show great mutual genetic distances, whereas natural selection maintains particular phenotypes on adaptive peaks. Such neutral genetic variation is occasionally purged from populations. This happens when a carrier of a favorable mutation and its descendants outcompete other lineages, resulting in a clustering of genotypes—and such clusters may be considered bacterial “species” (6). For practical purposes, the number of possible genotypes within such “species” is infinite. Locally, some particular clone may dominate at a given time, and gene sequences of bacteria sampled from different sites may therefore lead to the impression of geographical structuring.

It would be illuminating to carry out studies similar to those of Whitaker et al. (4) and Papke et al. (5) on Thermoplasma acidophilum. This thermophilic archaebacterium was first discovered in self-igniting coal refuse piles in North America; its natural habitat in sulfurous hot springs was discovered later. Coal refuse piles seem to be populated quickly by this organism—so at least some extreme thermophiles do move around (7). If these newly founded Thermoplasma populations show a substantially greater genetic homogeneity than the natural hot-spring biota, this would support the hypothesis that the patterns shown by Whitaker et al. (4) and by Papke et al. (5) reflect geographical isolation of populations over a longer time scale. But so far, a microbial counterpart of the marsupial mammals of Australia or the giant tortoises of the Galapagos is yet to be found.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.

Navigate This Article