For more than a century, elevational gradients have yielded unique insights into the ecological and evolutionary mechanisms that generate patterns of biogeography (Grinnell, 1917). Steep elevation gradients generate rapid shifts in habitat characteristics over short spatial distances, a property that is useful in determining the degree to which dispersal is driven by spatial distance or variation in habitat characteristics (Sundqvist et al., 2013). Elevation gradients have a strong influence on the biodiversity of plants and animals (Peters et al., 2016) with many taxa exhibiting mid-elevation peaks or “hump-shaped curves” in alpha diversity (Rahbek, 2004; Moradi et al., 2020), and similar patterns have been observed for microbes (Fierer et al., 2011; Singh et al., 2012; Liu et al., 2016; Siles and Margesin, 2016). Elevation can affect biodiversity by a range of mechanisms including ecological filtering by habitat preference (Fierer et al., 2011; Wang et al., 2012; Shen et al., 2014; Cho et al., 2018), variation in carrying capacity, and historical processes linked to patterns of climate change (Badgley and Fox, 2000; Rickart, 2001; de la Giroday et al., 2011; Schai-Braun et al., 2020). In addition, taxa that occupy mountain habitats are uniquely affected by historical climate change as warming climates tend to push species distributions toward higher elevations (Flesch, 2019; Marshall et al., 2020; Neate-Clegg et al., 2021), minimizing dispersal opportunities for species of plants and animals found at high elevations (Sekercioglu et al., 2008).

Biogeographical patterns can be driven by ecological mechanisms (e.g., assembly processes driven by ecological filtering and ecological drift), evolutionary mechanisms (e.g., speciation due to selection and drift), and historical contingency (e.g., neutral processes linked to variation in geology and climate over time) (Hanson et al., 2012). Microbial biogeography is often thought to be constrained by ecological filtering, under the assumption that dispersal is largely unlimited and environmental gradients impose spatial structure on communities due to selection (Navarrete et al., 2015; Liu et al., 2018; Malard and Pearce, 2018; Malard et al., 2019). However, most evidence for unlimited microbial dispersal is obtained using highly conserved taxonomic markers (e.g., rRNA genes) that have low taxonomic resolution and are insensitive to evolutionary processes that drive diversification (Hanson et al., 2012). Studies that use higher resolution taxonomic markers often find evidence for dispersal limitation with evidence that neutral processes can play a role in shaping patterns of microbial biogeography (Whitaker et al., 2003; Polz et al., 2013; Andam et al., 2016b; Choudoir et al., 2016; Choudoir and Buckley, 2018).

To explain the mechanisms that give rise to microbial biogeography we must first understand the forces that govern microbial dispersal. Dispersal is a two-part process comprised of dissemination, the movement from one place to another, and establishment, the successful colonization of a site characterized by the ongoing production of viable offspring (Martiny et al., 2006). Dissemination can be either passive (as driven by wind, erosion, currents, and organismal vectors) or active (as driven by motility or hyphal growth) (Yang and van Elsas, 2018). It is likely that capability for dissemination varies considerably between microbial taxa. For example, windborne dissemination is likely to vary in relation to cell size (Wilkinson et al., 2012), and cell shape likely influences microbial dissemination and establishment (Young, 2006). In addition, dissemination is influenced by environmental states. For example, soil texture and temperature influence spore transport in Phytophthora fungi (MacDonald and Duniway, 1978), and weather patterns can affect aerial dissemination (de Groot et al., 2021). Successful dissemination, however, is insufficient for successful dispersal, as microbes must still establish a sustainable population in the new site. Establishment requires that the habitat be suitable for growth, and that competitive interactions (e.g., antagonism, or density-dependent blocking) do not prevent ongoing reproduction (Woody et al., 2007; Cheong et al., 2021).

We performed analysis of rpoB amplicons to investigate community assembly in Streptomyces along an elevational and spatial gradient in the Adirondacks region in New York State. The use of rpoB as a taxonomic marker for this genus improves taxonomic resolution significantly (Rong and Huang, 2012; Higgins et al., 2021) as compared to analyses made using 16S rRNA genes. The use of high-resolution taxonomic markers is essential for investigating the mechanisms that govern microbial biogeography (Hanson et al., 2012; Chase et al., 2017).

Streptomyces are bacteria that form aerial hyphae and arthrospores (Flärdh, 2003), which facilitate dissemination. They are common in soil habitats worldwide where they degrade a variety of common substrates derived from plant biomass (Yeager et al., 2017) and produce diverse antibiotics and secondary metabolites (Watve et al., 2001). Despite their high capacity for dissemination, and broad habitat tolerance, Streptomyces have been shown to exhibit endemism at regional scales, indicative of dispersal limitation (Andam et al., 2016a). Their wide distribution, ecological significance, theoretical capability for high dispersal, and their observed limited ranges make the Streptomyces genus an ideal group to understand dispersal and biogeography patterns in the soil.

We hypothesized that dispersal limitation would also occur at local scales, with dispersal limitation driven by barriers to establishment (i.e., ecological filtering) rather than barriers to dissemination. To evaluate this hypothesis, we examined Streptomyces communities at two locations in the Adirondacks region of New York State. Sites at Whiteface Mountain varied greatly in elevation, while sites at Woods Lake varied little in elevation. All sites were broadly similar in habitat characteristics other than those linked to elevation. We predicted that high rates of local dissemination, coupled with ecological filtering as driven by elevation, would produce a strong gradient of beta diversity at Whiteface Mountain, and little variation in beta diversity at Woods Lake. We also predicted that ecological filtering would cause Streptomyces taxa to exhibit phylogenetic conservation with respect to elevation preference.

We show that Streptomyces communities are dispersal limited at local spatial scales and that dispersal limitation is governed by limits to both dissemination and establishment. Streptomyces communities at Woods Lake had low rates of dispersal, resulting in ecological drift at spatial scales spanning hundreds of meters. Low dispersal could result from low rates of dissemination, but the ability of Streptomyces to make aerial hyphae and desiccation resistant spores, and the fact that long range dispersal has been observed (Andam et al., 2016b; Choudoir et al., 2016; Higgins et al., 2021), suggest that limits to dissemination are unlikely to structure communities at local scales. Given the minimal habitat variation among sites at WL, and the broad habitat suitability expected for Streptomyces, it also seems unlikely that low dispersal at WL is driven by ecological filtering. Hence, we hypothesize that low dispersal and ecological drift at WL is likely driven by biotic interactions such as antagonism or density dependent blocking (Waters et al., 2013). Streptomyces are well known to produce diverse antimicrobial compounds, and other secondary metabolites, that alter biotic interactions (Kinkel et al., 2014; Schlatter and Kinkel, 2014; Vaz Jauri and Kinkel, 2014; Essarioui et al., 2016; Otto-Hanson and Kinkel, 2020). It seems likely that biotic interactions generate barriers to establishment that govern the structure of Streptomyces communities at WL.

We see strong evidence of Streptomyces dispersal limitation at Whiteface Mountain, with elevation having significant impacts on both dissemination and establishment. Certain clades preferentially occupy either high or low elevation habitats, indicating that barriers to establishment alter community structure across the mountain. The fact that these elevational habitat preferences are also observed in a continental-scale dataset suggests that ecological filtering by habitat preference constrains Streptomyces community structure across elevation. However, we also observe that homogenizing dispersal increases toward the base of the mountain (Figure 4), and this suggests that the elevation gradient favors dissemination, likely due to movement of material down the mountain. Both alpha and beta diversity are significantly higher at WM than WL, consistent with the expectation that strong environmental gradients amplify patterns of microbial diversity despite high rates of dissemination. We expect that competitive interactions influence Streptomyces community composition at both WL and WM, but that the effect of elevation on dissemination and establishment is the main driver of community structure at WM.

Previous studies offer conflicting evidence for the effect of elevation on microbial biogeography. While research across a montane elevational gradient in Peru showed no effect of elevation on bacterial communities in soil (Fierer et al., 2011), a similar analysis on soils from the Andes Mountains found that bacterial and fungal diversity both varied with respect to elevation (Nottingham et al., 2018). Other studies have documented variation in bacterial community structure across elevation, attributing such variation to a range of factors including soil pH (Cho et al., 2018), aspect (Wu et al., 2017), soil carbon (King et al., 2008), and seasonality (Lazzaro et al., 2015; Zhu et al., 2020).

Conflicting evidence on the relationship between bacterial biogeography and elevation could result from variation in spatial scales, habitat variability, and the phylogenetic resolution of taxonomic markers. Several environmental variables can co-vary with elevation (Sundqvist et al., 2013), making it difficult to disentangle the effect of elevation as opposed to other co-varying gradients. We also know that the spatial scale and taxonomic resolution at which diversity is measured can influence our ability to observe patterns of biogeography (Bent et al., 2003; Martiny et al., 2011; van de Guchte, 2017). Most prior studies of microbial diversity across elevation gradients have been performed using the 16S rRNA gene as a taxonomic marker. The low phylogenetic resolution of this marker makes it unsuitable for assessing mechanisms of dispersal limitation (Choudoir et al., 2012). For example, common taxonomic units defined on the basis of the 16S rRNA gene encompass strains whose ancestors may have diverged 50–150 million years ago (Ochman et al., 1999), and such taxonomic units lack the resolution needed to resolve the mechanisms that underlie extant patterns of microbial biogeography (Hanson et al., 2012). Our ability to identify the effect of environmental gradients on species distributions improves in proportion to the phylogenetic resolution at which diversity is characterized (Ramirez et al., 2018). Many phenotypic traits are conserved among closely related strains (Martiny et al., 2015; Barnett et al., 2021), and so experiments that use taxon-specific, fine-scale phylogenetic markers are vital to illustrate the processes driving microbial biogeography. Several examples of non-16S gene markers already exist in the literature; dsrA, nirK, nirS, and other MLST-based schemes have been used to detect biogeographical patterns in environmental bacteria (Whitaker et al., 2003; Boucher et al., 2011; Angermeyer et al., 2016; Kou et al., 2021; Liao et al., 2021). In this study, the use of a Streptomyces-specific amplicon marker allows the exploration of phylogenetic patterns at a fine scale, sufficient for exploring the mechanisms that govern microbial dispersal.

Contemporary and historical climate variation is likely to influence patterns of Streptomyces biogeography. Elevation has a clear impact on temperature, as land temperatures decline 0.42°C for every 100 m of elevation, such that a 200 m change in elevation approximates the temperature shift associated with a 1° change in latitude (Montgomery, 2006). Phylogenetic conservation of thermal traits has been shown to influence Streptomyces dispersal across latitude (Choudoir and Buckley, 2018), and such thermal adaptation likely contributes to the latitudinal diversity gradient observed for North American Streptomyces (Andam et al., 2016b). Whiteface Mountain is one of the highest peaks in the Adirondacks (1,484 m above sea level). Geological evidence indicates that the mountain was glaciated along with the entire Adirondacks region, until glacial retreat about 10,000 years ago (Franzi et al., 2000). This geological timeline means that Streptomyces have arrived fairly recently to WM and WL and hence the time for local diversification was limited. Prior to the period of glacial retreat, about 12,000 years ago, the climate in the Adirondacks region would have been approximately 2°C cooler than current conditions (Kaufman et al., 2020). Over time, as the climate warmed, species adapted to warmer climates would have dispersed into the wider Adirondacks region while cold adapted species would have found their habitat restricted to higher and higher positions on the mountain. In this scenario, we hypothesize that ecosystem properties linked to climate variation influence microbial dispersal in soils by controlling the probability that species are able to establish at new sites. Changes in elevation influence a wide range of ecological variables both above and belowground (Sundqvist et al., 2013), and so it would be imprudent to conclude that temperature is the most important variable delimiting establishment, but it seems fair to conclude that ecological properties associated with climate variation can be expected to alter patterns of microbial establishment in soils.

Our findings indicate that a mixture of stochastic and deterministic processes govern Streptomyces dispersal. Streptomyces form aerial hyphae that produce desiccation-resistant, hydrophobic spores and their physiological traits should support broad habitat tolerance. As a result, we would generally assume that Streptomyces have a greater dispersal capacity than most other soil bacteria. However, we found high dissimilarity in communities that occupied similar habitats and similar elevations (350–450 m elevation) at both WL and WM. This result suggests local limits on dispersal, likely driven by capacity for establishment as determined by competitive interactions between existing species and new immigrants. However, we did see evidence for homogenizing dispersal at the base of Whiteface Mountain suggesting that high rates of dissemination, likely driven by mass transport down the mountain, might overwhelm the ability of deterministic processes to constrain community assembly patterns. Our ability to identify dispersal limitation was enabled by the high phylogenetic resolution of the rpoB marker that we used, since 16S rRNA analyses provide little ability to resolve patterns of dispersal in Streptomyces (Higgins et al., 2021). The existence and impact of dispersal limitation has now been documented for several microbial taxa across a range of ecosystems (Staley and Gosink, 2002; Whitaker et al., 2003; Bell, 2010; Eisenlord et al., 2012; Albright and Martiny, 2017; Bottos et al., 2018; Evans et al., 2019). Hence, it seems likely that microbial dispersal is finite and subject to change over time based on contemporary processes and historical contingencies (Hewitt, 2000; Mennicken et al., 2020; Liao et al., 2021). In the case of Streptomyces, we hypothesize that contemporary climate variation is a major determinant of establishment. We also hypothesize that historical variation in climate has contributed significantly to extant patterns of microbial biogeography in North America because rates of dispersal are low relative to rates of climate variation during the Quaternary Period.

The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found below: https://www.ncbi.nlm.nih.gov/, PRJNA790982.

JH performed the research, analysis, writing, and editing. DB supervised research and analysis and assisted with writing and editing. Both authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.