Cryptic microbial communities in Antarctic deserts

The Antarctic Dry Valleys, situated in the Ross Sea region of Eastern Antarctica, are often referred to as one of the most extreme environments on Earth. Although this statement is subject to some argument, largely surrounding the issue of what “extreme” actually means, there is no arguing that the combination of macroenvironmental and microenvironmental conditions to which organisms living in the Antarctic Dry Valleys are exposed represent a severe threat to organismal survival. The combination of extreme cold and desiccation, high soil salinity, low nutrient levels, high summer UV radiation levels, and physical instability caused by strong katabatic winds all contribute to the visual appearance of a sterile environment (1–3).

Despite the apparent severity of the environment, life does exist in the Antarctic Dry Valleys. The article by Pointing et al. in this issue of PNAS (4) focuses on three types of cryptic microbial communities that are widespread in this environment. These communities (see Fig. 1) all represent adaptive community responses to the extreme conditions of the polar climate. The lithic environment can provide protection against some (but not all) of these conditions: i.e., protection from wind scouring and surface mobility (5), a reduction in UV exposure (6), reduced desiccation and enhanced water availability (7), and thermal buffering (5). Mean annual temperatures are generally unchanged, unlike hot desert lithic habitats where the rock provides some protection from solar heating (8). Interestingly, Pointing et al.'s study suggests that soil salinity is a major factor in community structure (but not necessarily community survival). In desert ecosystems, even modest levels of soil salinity may be an important determinant, because water activity [a_w (9)], which determines biological water availability, is reduced in the presence of soluble salts (10).

Fig. 1.

Lithic habitats in the cold hyperarid deserts of the Eastern Antarctica Dry Valleys. (A) Desert pavement in the high altitude Beacon Valley. (B) A typical quartz hypolith showing no external evidence of the underlying microbial community. (C) Casmolithic lichenised microbial communities protruding from a granite rock fracture. (D) Endolithic community layer in fractured sandstone. Photograph courtesy of S. Pointing (School of Biological Sciences, University of Hong Kong, Hong Kong).

The study by Pointing et al. (4) shows that Dry Valley lithic communities contain a diverse range of prokaryotes and lower eukaryotes at high biomass levels (see table 1 in ref. 4), exist in surprisingly complex assemblages, and harbor a significant number of previously un reported microbial signatures. These observations raise a number of very interesting issues. First, Pointing et al. rightly suggest that an upward revision of standing biomass (and by implication, productivity) in these Antarctic soils is probably warranted. This view is supported by an earlier study (11), albeit in the low-altitude, maritime-influenced Miers Valley, which demonstrated using ATP, lipid, and DNA quantification that standing biomass was in the range of 10⁶ to 10⁸ cells·g⁻¹ soil, orders of magnitude higher than determined by microscopic and culture-dependent estimates (see, for example, ref. 12).

Second, the presence of these complex cryptic communities appears to strongly contradict the dogma that microbial diversity is inversely proportional to the severity of the climatic conditions (13). Pointing et al. (4) also note that open soil microbial diversity was much lower in lithic communities and lacked a photoautotrophic component. This highlights one of the key elements of such lithic communities. In a very nutrient poor (oligotrophic) environment, photoautotrophic carbon input is critical to community development. They also note that many of the phylotypes identified are putative nitrogen fixers and that these organisms are likely to be responsible for complementing the nitrogen deficiency of the Dry Valley soils. Recently, hypolithic communities in the southern low altitude Dry Valley regions have been shown to be capable of acetylene reduction, the accepted functional proxy for dinitrogen fixation. The observation that cyanobacterial Chroococcidiopsis-like phylotypes in endoliths and casmoliths and Leptolyngbia in endoliths are dominant members of the respective communities is entirely consistent with previous microscopic (7, 14) and phylogenetic observations (15).

The separate clustering of phylotypic signals (see figure 2 in ref. 4) for open soils, hypoliths, and chasmoliths/endoliths is likely to reflect some fundamental difference in the factors controlling the development of the communities. It may be relevant that neither the chasmolithic nor the endolithic communities have any direct interaction with the desert pavement, and some soil-related factor may account for this degree of discrimination. It is noted, however, that the most obvious candidate, salinity (K/Na and Cl), was identified as a key driver for all communities. It is highly likely that the separate grouping of the two soil diversity datasets (hypolith and open soil) is driven by the dominance of cyanobacteria in the former, and their apparent absence in the latter. It follows therefore that survival of such “refuge” communities is probably linked intimately to survival of the keystone primary producers. Interesting, several studies have suggested that photosynthetic processes are more sensitive to desiccation than other central carbon metabolic pathways (16).

In detailed phylogenetic surveys, based on 16S rRNA gene sequences, the presence of a substantial number of unknown bacteria is not surprising; virtually any phylogenetic survey of any environmental sample will give unidentifiable phylotypic signals. Nevertheless, these observations raise the interesting issue of microbial endemism. It is commonly believed that the more extreme the environmental conditions and the more isolated the habitat, the higher the level of microbial endemism. The Dry Valley soils, although open to seeding from particles transported by high-altitude winds (17), are both remote and environmentally stringent. A high level of both continental (and/or bipolar) endemism might thus be expected.

A cautionary note on the significance of phylogenetic surveys is always necessary. It is widely recognized by microbial molecular ecologists that a phylotypic signal (for example, a partial 16S rRNA gene sequence amplified by PCR) is no guarantee that the organism (phylotype) identified from that signal is a functional member of the microbial community. Extracted metagenomic DNA is typically derived from live, dormant, and dead cells and even from exogenous DNA from dead and lysed cells. The latter is typically assumed to be a small contribution because of the high turnover (adsorption, degradation, and reutilization) of polynucleic acids in moist environments (18). However, it is noted that the cold arid conditions prevailing in Antarctic desert soils are well suited to the preservation of cells and their constituents and closely parallel those typically used in laboratories for cell preservation, i.e., freeze-drying. The issue of which members of the community are metabolically active is typically resolved by a parallel analysis using methods such as stable isotope probing (19) and functional gene expression analysis. Although Pointing et al.'s article (4) does not carry the analysis this far, it is fair to say that the basic tenets of this work are largely unaffected by our current ignorance of the metabolic state of community members.

Dry Valley lithic communities contain a diverse range of prokaryotes and lower eukaryotes.

It is commonly assumed that Antarctic lithic microbial communities are close to the cold-arid limit for life (20) and, as such will be highly sensitive to environmental perturbation. The corollary to this assumption is that these communities will be sensitive indicators of changes in climatic conditions, particularly those relating to the key determinants of survival: water availability and austral summer temperatures (i.e., the photosynthetic window). This assumption is yet to be tested. However, as a working hypothesis the following can be predicted: (i) where communities are surviving close to the cold-arid limit of life, a small negative change in mean temperature and the associated reduction in water availability should have a deleterious effect on community survival, and (ii) “positive” microenvironmental changes, such as those that might be associated with regional temperature increases (21), could increase the physical extent and standing biomass of lithic communities. Conversely, the same positive changes might reduce factors that provide positive community structure discrimination and lead to the homogenization of community profiles. It should not be forgotten that even with static mean temperatures, changes in the amplitude of temperature fluctuations may also impact dramatically on community composition and/or survival.

Finally, it is worth emphasizing that functional criteria (see above) are probably a more sensitive determinant of climate change impact than phylogenetic surveys.