Hortal, J. (2011) In Biogeography of micro-organisms. Is everything small everywhere? (ed. by D.Fontaneto) in press
Chapter 17
Geographical variation in the diversity of microbial communities: research directions and prospects for experimental biogeography
Joaquín Hortal1,2*
1Departamento de Biodiversidad y Biología Evolutiva, Museo Nacional de Ciencias Naturales (CSIC), C/José Gutiérrez Abascal 2, 28006 Madrid, Spain.
2Azorean Biodiversity Group – CITA-A, Universidade dos Açores, Terra-Chã, 9700-851 Angra do Heroísmo, Terceira, Açores, Portugal.
* e-mail:
Accepted for publication:
Hortal, J. (2011). Geographical variation in the diversity of microbial communities: research directions and prospects for experimental biogeography. In Biogeography of micro-organisms. Is everything small everywhere? (ed. by D. Fontaneto), pp. in press. CambridgeUniversity Press.
17.1 Introduction
Traditionally, most ecologists understand the world from a human scale. Ecosystems are often understood as large visible units of the landscape[1], usually homogeneous land patches or a series of adjacent patches with intense flows of individuals, energy or biomass and nutrients. However, there is more in a landscape than meets the eye. An arguably homogeneous land patch within a landscape hosts many small ecosystems, or microhabitat patches, where many different communities of microbes[2]dwell and interact. For example, imagine you are standing in a clearing of an open forest in a temperate region. A terrestrial ecologist studying macroscopic organisms would think he is looking at part of a single ecosystem. On the contrary, a microbial ecologist will identify a plethora of different ecosystems, including leaf litter of different degrees of humidity, the bark of each different tree and shrub species, treeholes, temporary puddles and pools, moss cushions of different life forms growing over different substrates, etc. Not to mention soil communities. In other words, a 1 ha clearing within a forest could be considered a whole landscape for many groups of microbes.
A key question in microbial ecology is thus whether the patterns and organization of microbial communities differ from those of macroscopic organisms just in terms of scale or they are so radically different that the rules affecting macrobes cannot be extrapolated to microbes. The debate on this question extends to the biogeography of microorganisms. Strikingly, it has been argued that most microorganisms do not have biogeography; that is, that contrary to macroorganisms, the distributions of microorganism species are just limited by local environmental conditions (e.g. Fenchel & Finlay, 2003, 2004 and below). But, are microbes so different from their larges relatives than they follow different ecological and biogeographical rules?
Here I will argue that when it comes to the spatial distribution (and basic ecology) of their communities, many microbes (especially multicellular ones) are just smaller than large organisms, rather than radically different in their ecological organization and biogeographical responses. More precisely, I will first argue that not everything small is everywhere, and then provide a brief account of current evidence on the spatial distribution of microbe diversity at the community level. Given the purpose of this chapter, this review will be argumentative rather than exhaustive. Based on such review, I will propose the study of microbes as a way of advancing current biogeographical and macroecological theory[3], under the hypothesis that some biogeographical principles can be evaluated with success on microorganisms, controlling for many of the confounding factors acting at large scales, or even allowing to develop experiments in biogeography.
17.2 Spatial variations in the diversity of microscopic organisms
17.2.1 Is everything small everywhere?
Perhaps the most striking difference between the known spatial distributions of macrobe and microbe species is that while restricted distributions are the rule for the former, it has been argued that they may be exceptions for the latter (Fenchel & Finlay, 2003, 2004; Kellogg & Griggin, 2006; Fontaneto & Hortal, 2008). However, the level of knowledge about the spatial distribution of microbial diversity is not comparable to that of macroscopic organisms. Despite the causes of the spatial distribution of diversity are still under debate (see section 17.4 below), the current degree of knowledge on macrobes is rather good. Although most of the groups with well-known diversity patterns at the global scale are vertebrates (Grenyer et al., 2006; Schipper et al., 2008), the variations in the numbers of species of plants (Kreft & Jetz, 2007) or insects (Dunn et al., 2009) throughout the globe are also starting to be well-known, and at least partly understood (Lomolino et al., 2006). In contrast, the level of knowledge on the spatial distribution of most (if not all) groups of microorganisms is quite limited (Foissner, 2008; Fontaneto & Hortal, 2008; and many chapters of this book). Whether such deficit in knowledge is the cause of the apparent lack of biogeography of most microorganisms or not is perhaps the hottest debate in current microbial ecology.
The realization that, apparently, many microbial species are found in quite distant localities led to the proposition of the Everything is Everywhere (EiE) hypothesis at the beginning of the twentieth century (Beijerinck, 1913; Baas-Becking, 1934). This hypothesis is further supported by the high dispersal potential (sensu Weisse, 2008) of most microbes (Finlay, 2002; Kellogg & Griffin, 2006). Their small size, large population numbers and, especially, the ability either to enter dormant states or produce small spores allow many microorganisms to produce vast numbers of propagules that are easily dispersed in a passive way (i.e. ‘ubiquitous dispersal’: Fenchel, 1993; Finlay et al., 1996a, 2006; Cáceres, 1997; Wilkinson, 2001; Fenchel & Finlay, 2004). Arguably, this would permit many microbes to maintain cosmopolitan distributions. Although the EiE hypothesis has been the dominant paradigm for microbial biogeography until relatively recently (O’Malley, 2007, 2008), it has been hotly debated during the last decade, dividing microbial ecologists into two factions (Whitfield, 2005). Some argue that the rule for microorganisms is ‘everything is everywhere, but the environment selects’ (Finlay, 2002; de Wit & Bouvier, 2006; Fenchel & Finlay, 2006). Others counter that many apparently cosmopolitan ranges are actually artifacts of the deficient taxonomy of microbes, which does not permit distinguishing between morphologically similar but spatially and genetically isolated lineages (Coleman, 2002; Foissner, 2006, 2008; Taylor et al., 2006).
In my opinion there is now enough information to develop a theoretical (and analytical) framework that will resolve the EiE debate and lay the foundations for a general theory of microbial biogeography. However, this is beyond the intended scope of the chapter; more information can be found in several chapters of this book, or by consulting the references above (see also Martiny et al., 2006; Green & Bohannan, 2006; Telford et al., 2006; Green et al., 2008). Having said this, any study on the spatial distribution of microorganism communities shall necessarily address the question of whether everything small is everywhere. Should microbes be locally abundant and extremely widespread, their local diversity would be the result of random colonization processes, as argued by, e.g. Finlay et al. (1999, 2001). Here, differences among communities would be determined only by local environmental conditions. However, such cosmopolitanism seems to be far from universal. Many microbe species have been found to have restricted distributions (Mann & Droop, 1996; Smith & Wilkinson, 2007; Frahm, 2008; Segers & De Smet, 2008; Vanormelingen et al., 2008; Spribille et al., 2009). Hence, the dependence of range size on body size hypothesized by Finlay and colleagues (e.g. Finlay et al., 1996b; Finlay, 2002; Finlay & Fenchel, 2004) is not as general as they argue (Valdecasas et al., 2006; Pawlowski & Holzmann, 2008; but see Martiny et al., 2006). More importantly, the EiE hypothesis is challenged in its assumption that the large dispersal potential of microbes necessarily results in high rates of effective dispersal (i.e. successful dispersal events, see Weisse, 2008). Rather, the propagules of many (but not all) microorganisms are not ‘universally successful’ in maintaining significant levels of gene flow between geographically remote populations (Jenkins, 1995; Jenkins & Underwood, 1998; Bohonak & Jenkins, 2003; Foissner, 2006, 2008; Jenkins et al., 2007; Weisse, 2008; Frahm, 2008).
As a direct consequence of the total or partial isolation of populations in relation to distance, phylogeographic variations (i.e. geographically structured genetic differences) have been found for many microbial taxa. Increasing spatial distance between populations results in genetic divergence and isolation even for prokaryotes (Whitaker et al., 2003; Prosser et al., 2007; Vos & Velicer, 2008). This may emerge as the common rule for many protists and multicellular microbes, once traditional approaches to their taxonomy are complemented with more detailed molecular studies (Foissner, 2008; Weisse, 2008; Pawlowski & Holzmann, 2008). In fact, many recent studies finding significant hidden genetic divergence within morphologically-based microbial species find also that these genetically different populations or species occupy geographically distant areas (Gómez et al., 2007; Mills et al., 2007; Weisse, 2008; Fontaneto et al., 2008a; Xu et al., 2009). Therefore, it could be expected that as knowledge of the phylogeny of microorganisms and their taxonomy improves, the number of microbe with restricted distributions will increase as well. The actual proportion of microbe species with reduced range sizes remains as a mystery, although some estimates indicate that at least one third of protist species will show restricted distributions (according to the moderate endemicity model; see Foissner, 2006, 2008). Nevertheless, such hidden microbial diversity will have an impact on the patterns of diversity observed at the community level.
17.2.2 Spatial variations in microbe communities
If the distributions of microorganisms are not cosmopolitan, microbe communities in similar substrates situated in geographically distant areas ought to show significant differences in their diversity and species composition. The spatial replacement of species in macrobial communities is typically the result of both environmental differences and geographical distance, no matter whether they are lake fishes (Genner et al., 2004), mammals (Hortal et al., 2005), birds or land snails (Steinitz et al., 2006). Microbes are to some extent similar to macrobes in this particular aspect. Although environmental heterogeneity is the main driver of the decay of compositional similarity with distance in microorganisms (see Martiny et al., 2006; Green & Bohannan, 2006), it is not the only source of spatial variation. Using an array of studies on lake diatoms encompassing several continents, Verleyen et al. (2009) found that, although environment accounts for the larger part of the spatial replacement of species, connectivity[4]also explains a large proportion of the compositional similarities between lakes: all else being equal, the closer the lakes, the more similar their species composition. Similar patterns were found in the phytoplankton communities of the Swedish lakes studied by Jankowski & Weyhenmeyer (2006). Interestingly, the strength of such replacement may vary according to the kind of habitat for both micro- and macrobes. Macrobial communities often show different patterns of distance decay of similarity[5]in different kinds of habitats (e.g. palm trees, Bjorholm et al., 2008; birds and land snails, Steinitz et al., 2006). Similarly, the degree of compositional similarityof the communities of bdelloid rotifers in a valley of northern Italy varies from one ecological systems to another: whereas stream communities were relatively similar to one another throughout the valley, the species integrating the communities from terrestrial habitats and lakes were highly variable (Fontaneto et al., 2006).
In contrast with these qualitative similarities between small- and large-sized organisms, the magnitude of the distance-driven changes in species composition across similar habitats may not be comparable. The slope of the taxa–area relationship in contiguous habitats can be used as a raw measure of the accumulation of species or other taxonomic units with area, and hence of compositional changes in space (Prosser, 2007; Santos et al., 2010; see also Rosenzweig, 1995 and Whittaker & Fernández-Palacios, 2007 for extensive reviews on the species–area relationship). To date, the slopes recorded for microbes are typically smaller than those of macrobes; while the former may range from 0.043 to 0.114 in natural systems, the latter are typically larger than 0.15 (in a power model; Finlay et al., 1998; Azovsky, 2002; Green et al., 2004; Smith et al., 2005; Bell et al., 2005; Green & Bohannan, 2006; Prosser et al., 2007). This indicates that in the absence of environmental differences, the spatial replacement in the composition of local communities occurs at much larger scales for microbes, varying in the range of a few hundreds to thousands of km, instead of the hundreds or even tens of km usually found for macrobes.
Strikingly, however, when taxa–area relationships are calculated for habitat islands (i.e. separate territories/habitat patches instead of contiguous habitats), the slopes may reach values well over 0.2 for bacteria, which are similar to those for large-sized organisms (Bell et al., 2005; van der Gast, 2005; see also Green & Bohannan, 2006; Prosser et al., 2007). This indicates that in spite of their minute size, habitat area plays a critical role in determining the number of bacterial species that can coexist in a given place, as it does for macroorganisms. In other words, although microbial communities change with distance at a lower pace than macrobes, the increase in the number of species with area is similar for both groups. Further study is required to determine whether this dependence on area is due to the carrying capacity of the locality (which increases with its area), the increase of habitat diversity with increasing area, or to passive sampling mechanisms (i.e. larger areas receive more propagules and therefore may be colonized by more species).
The similarities between the macroecological patterns of micro- and macroorganisms do not end with their shared dependence on area. As with their larger counterparts, the diversity of microbe communities varies along environmental gradients. One of the most studied is the altitudinal gradient: species richness varies with altitude, increasing or decreasing for localities closer to or farther from an optimal altitudinal band (Rahbek, 1995, 2005). These changes in richness are often accompanied by changes in species composition (e.g. the Alpine dung beetles studied by Jay-Robert et al., 1997). This macroecological pattern has been also observed in many microorganisms. A good example is the altitudinal variations in richness shown by rotifers in the Alps (Fontaneto & Ricci, 2006; Fontaneto et al., 2006; Obertegger et al., 2010). The diversity of tardigrades inhabiting moss and leaf litter at the Guadarrama mountain range also shows a typical hump-shaped relationship with elevation (Guil et al., 2009a). Alike macrobes, altitudinal variations in microbe species richness are caused by the varying environmental conditions at different elevations. Sometimes the decrease in productivity with altitude limits local diversity, as occurs for phytoplankton richness in a series of Swedish lakes (Jankowski & Weyhenmeyer, 2006). In other cases, the climatic variations associated with elevational changes cause spatial gradients in local richness, as with epiphytic bryophyte communities in Guiana (Oliveira et al., 2009) or NW Spain (N.G. Medina, B. Albertos, F. Lara, V. Mazimpaka, R. Garilleti, D. Draper & J. Hortal, unpublished manuscript). Such climate-driven diversity variations in microbes may arise simply because of habitat sorting (i.e. due to the differences in niche requirements of each of the species regionally available; see e.g. Whittaker, 1972). In fact, in the two examples given above many species of both bdelloid rotifers and tardigrades show strong habitat selection (i.e. habitat sorting; Fontaneto & Ricci, 2006 and Guil et al., 2009b, respectively). This is consistent with a high potential for dispersal and local environmental selection (although this process may involve an important degree of stochasticity, see Fontaneto et al., 2006).
However, geographical changes in the diversity of microbe species inhabiting similar microhabitats are not only the result of local productivity and/or carrying capacity and environmental variations. The pool of species that can colonize each microhabitat varies also in space, therefore limiting the number and identity of the colonizing species. The regional differences produced by changes in the species pool are one of the major determinants of the geographical variations of assemblage diversity for macroorganisms (Ricklefs, 1987, 2004, 2007; Huston, 1999; Hawkins et al., 2003a; Hortal et al., 2008; Hawkins, 2010). Some environments or particular habitats may host fewer species simply because fewer of the available species have evolved adaptations to these environments, either because these environments are rare in nature, they are too recent in the region, or they were affected by important changes in the past, like glaciations. In fact, although the dispersal distances and the location of refugia may differ, the diversity of several microbial groups has been shaped by post-glacial dispersal (e.g. Gómez et al., 2007; Smith & Wilkinson, 2007; Smith et al., 2008). More importantly, the composition of the regionally available species pool varies among continents, at least for lake diatoms (Verleyen et al., 2009). As with macrobes, these differences in the species pool produce different responses to elevational or climatic gradients, as shown also for lake diatoms by Telford et al. (2006).
The ultimate consequence of the spatial variations in microbe communities is the existence of geographic differences in ecosystem functioning. Community richness, composition, assembly and functional dissimilarity affect ecosystem productivity and functioning (Laakso & Setälä, 1999; Fukami & Morin, 2003; Heemsbergen et al., 2004; Sánchez-Moreno et al., 2008). Given that microbes perform many ecosystem services, geographic variations in community composition are likely to have important functional consequences (Naeslund & Norberg 2006; Green et al. 2008). However, the impact of these geographic differences in the functions provided by microbial communities in ecological processes at regional and global scales is, to date, poorly known.
17.3 A frontier of biogeography
An overall insight from the short review above is that the biogeography and macroecology of microbial assemblages presents both similarities and differences with those of macroorganisms. To synthesize, microbes and macrobes both show distance decay in community similarity. It follows that not everything is everywhere. Hence, although the decay of similarity and the associated compositional changes are mainly caused by environmental gradients, they are also driven by geographical variations in the composition of the species pool and the degree of connectivity or the presence of barriers to dispersal between localities. The geographical variations in microbe assemblages typically occur at larger distances than for macrobes, and thus they show a shallower increment of species with increasing area. In spite of this, the relationships between local richness and area and environmental gradients are comparable to the ones found in macrobes. The differences between micro- and macroorganisms seem thus limited to their differences in size and dispersal power, which link microbes to smaller microhabitats and make their distributions typically larger.
Some of the challenges awaiting biogeography are right in front of our eyes, rather than in distant places. Whether the similarities and differences between the diversity of macrobe and microbe assemblages arise from fundamentally similar or different processes needs further investigation. Here I outline a research agenda to help explore this particular frontier of biogeography (see also Martiny et al., 2006; Green & Bohannan, 2006; Prosser et al., 2007; Green et al., 2008). To understand the spatial variations in the diversity of microbial assemblages, research in four main areas is needed: microbial taxonomy, description of biogeographical and macroecological patterns, community ecology and assembly, and the study of functional diversity and ecosystem functioning.