**4. Discussion**

The cyanobacteria *Scytonema* spp., *Nostoc* spp. and *Tolypothrix* spp. are secondary colonizers in the ecological succession of biocrust communities [6], where they are among the most common heterocystous organisms [6,19,32,38,50], and contribute much of the nitrogen inputs to the community at this stage of development [66]. Therefore, it is logical to assume that their presence and relative abundance have direct effects on the N2-fixation capability of late successional biocrusts. Using quantitative enrichment cultures we could clearly demonstrate differential fitness in these cyanobacteria at different temperatures, in a pattern that confirms the preferences inferred in prior field [32,34] and cyanobacterial cultures thermophysiological assays [32–34], where the biocrust cyanobacteria *M. steenstrupii* complex and *Scytonema* spp. were found to be more thermotolerant than *M. vaginatus*, and *Tolypothrix* spp. and *Nostoc* spp., respectively.

Using a set of cultivated strains (12 *Scytonema* spp., 10 *Nostoc* spp. and eight *Tolypothrix* spp.) isolated from cold and hot desert locations of the Southwestern US, the temperature range for growth revealed a pattern of niche differentiation according to temperature: *Tolypothrix* spp. strains having an advantage at the lower temperatures, and *Scytonema* spp. strains at higher temperatures. *Nostoc* spp. strains occupied only the mesic part of the temperature range. This niche separation is similar to that found in non-heterocystous filamentous cyanobacteria of soil crusts [34], and parallels the much more conspicuous niche differentiation of cyanobacteria known from hot springs at temperatures

between 45-73 ◦C [67]. Similar niche separation in cyanobacterial genera are found as a function of salinity [68,69] or desiccation frequency in marine intertidal systems [70]. We could also show that the upper temperature limit for growth (and survival) under N2-fixing conditions is more constrained than that under non N2-fixing conditions (Figure 3), implicating N2-fixation as a possible driver of the effective upper limit of temperature range in nature. Although measurements of N2-fixation rates (by acetylene reduction assay or 15N isotopes) would give a much more direct result, the observed thermophysiological responses of the tested strains at 35 ◦C, coincide with more dramatic decreases in N2-fixation rates (above 30 ◦C) in cold than in hot biocrusts locations [31], and are congruen<sup>t</sup> with the fact that *Nostoc* spp. and particularly *Tolypothrix* spp. are more abundant in biocrusts from colder locations, while *Scytonema* spp. typically dominate in warmer ones [32–34].

In an effort to better understand the basis for this effect on N2-fixation we determined the ratio of heterocyst frequency at different temperatures in a selected set of strains, which were responsive to our experimental conditions (*Scytonema* sp. JS006, *Nostoc* sp. HSN008 and *Tolypothrix* pp. HSN042, Figure 3). The results sugges<sup>t</sup> that in *Nostoc* spp. and *Tolypothrix* spp., the impossibility of these strains to grow under N2-fixation conditions at temperatures above 30 ◦C may be determined by an inability to carry out the sophisticated developmental cycle leading to the differentiation of heterocysts [71]. While *Scytonema* spp. may have overcome such developmental problems (Table 1), nitrogenase denaturation, which has been reported to happen at temperatures above 39 ◦C [72] could be the basis for the observed differences in *Scytonema* spp. strains' biomass yield at 35 ◦C (Figure 3). It is also possible that the observed inability of *Nostoc* spp. and *Tolypothrix* spp. strains to differentiate heterocysts at higher temperatures is the result of a resource allocation constraint to obtain the energy required to differentiate these specialized cells. However, N2-fixation and heterocyst differentiation at temperatures above 40 ◦C is not a problem in principle, in that the freshwater thermophilic cyanobacterium *Mastigocladus laminosus* performs N2-fixation at 45 ◦C [73], and is able to grow at temperatures as high as 57 ◦C [74]. Whether the observed heterocyst frequency decrease in *Nostoc* spp. and *Tolypothrix* spp. is a direct effect of temperature rather than a side effect due to stress on other physiological processes will need further investigation.

We tested the relevance of this temperature-based niche differentiation in nature by studying the distribution of the three cyanobacterial types as a function of climate parameters in a meta-analysis of a large dataset of biocrust surveys. Indeed, we found that the maximal relative proportion of *Scytonema* spp. among all heterocystous cyanobacteria increased along the temperature gradient with increasing temperatures (Figure 4A), when the average temperatures of the growth (wet) season was considered. Clearly, however, the results point to a potential for differential sensitivity of these cyanobacteria to environmental warming, a future scenario with which biocrust will have to contend. Drylands at large will likely become warmer and drier in response to global warming. In particular, the US Southwest is predicted to experience an increase in temperature of about 1 ◦C per decade [26], accompanied by alterations in precipitation frequency [35,36].

Given the observed differential response of biocrust N2-fixing cyanobacteria to temperature, and in agreemen<sup>t</sup> with Muñoz-Martín et al., (2018), it is reasonable to forecast that a microbial replacement within biocrust heterocystous cyanobacteria may indeed be in store as a result of global warming. *Scytonema* spp. may replace more cold- and mesic-temperature adapted *Tolypothrix* spp. and *Nostoc* spp. In places such as the Colorado Plateau, the Mojave desert, the north part of the Chihuahuan Desert (Sevilleta LTER) in the USA, Alicante in Spain, Western Australia [49], temperate areas in Mexico [33], and the Brazilian savannah (Cerrado) [52], where the mean annual temperature during the growth season falls between the 17 and 23 ◦C range, this microbial replacement will likely happen faster than at those locations exhibiting mean average temperatures below 17 ◦C, that are not projected to reach sensitive temperature ranges for decades to centuries, or locations with average temperatures above 24 ◦C, which already exhibit a dominance of *Scytonema* spp. (Figure 4). This microbial replacement could have implications for drylands and biocrust nitrogen inputs beyond a mere compositional change. *Scytonema* spp. have been shown to be one of the most sensitive taxa in biocrust to changes in precipitation patterns [47]. In this scenario, the N2-fixing cyanobacteria taxa that seem to be better adapted to withstand increases in temperature, ironically, seem to be among the least adapted to withstand drought. Although it makes sense that cyanobacterial distribution patterns with increasing temperature became more apparent when mean temperature during the wettest quarter of the year was used as an explanatory variable, we were surprised by the fact that plots using MAT did not show clearer patterns (Figure 4). This highlights the need to take into account the ecophysiology of microorganisms when seeking to find important climatic drivers.

These results can also serve to improve strategies to restore biological soil crust communities, of much recent interest in conservation ecology [46,75], by providing information to optimize inoculation season and microbial inoculum formulations.

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2076-2607/8/3/396/s1, Table S1: Accession numbers for the main generic groups for *Scytonema* spp. *Nostoc* spp. and *Tolypothrix* spp. according to our taxonomic assignment using our own cyanobacterial reference tree CYDRASIL., Table S2: Outcome of enrichment cultures for nitrogen-fixing photoautotrophs (nitrogen and organic carbon free medium, in the light) using variously sourced biocrusts as inoculum as a function of the incubation temperature. Given are the number of colonies containing each cyanobacterial taxa of interest, as identified morphologically by microscopy inspection. "S" stands for *Scytonema* spp., "N" for *Nostoc* spp., and "T" for *Tolypothrix* spp., Table S3: Cyanobacterial strains and their accession number in NCBI of their partial 16S rRNA sequence. Strain denominations include coding for the site of origin (HSN: cold desert sandy clay loam soil; HS: cold desert clay loam soil; FB: warm desert loamy sandy soil; JS: warm desert clay loam soil), Table S4: Environmentalbiocrust surveys conducted at different locations around the world used in the meta-analysis, and the corresponding climate data. Raw sequences were downloaded from bacterial 16S rRNA tallies available publicly (see references). Environmental data was downloaded from WorldClim. "MAT" stands for mean annual temperature and "MTemWetQ" for mean temperature during the wettest quarter of the year (growth season)., Table S5: Full results for linear regression between relative proportions (arcsine transformed) of *Scytonema* spp. and mean annual temperature (MAT)., Table S6: Full results for linear regression between relative proportions (arcsine transformed) of *Scytonema* spp. and mean temperature during the wettest quarter of the year (MTemWetQ)., Table S7: Full results for linear regression between relative proportions (arcsine transformed) of *Nostoc* spp. and mean annual temperature (MAT)., Table S8: Full results for linear regression between relative proportions (arcsine transformed) of *Nostoc* spp. and mean temperature during the wettest quarter of the year (MTemWetQ)., Table S9: Full results for linear regression between relative proportions (arcsine transformed) of *Tolypothrix* spp. and mean annual temperature (MAT)., Table S10: Full results for linear regression between relative proportions (arcsine transformed) of *Tolypothrix* spp. and mean temperature during the wettest quarter of the year (MTemWetQ)., Figure S1. Linear regressions between the proportion of sequence reads (arcsine transformed) of each taxon among heterocystous cyanobacteria and climatic parameters (MAT and MTempWetQ).

**Author Contributions:** Conceptualization, A.G.-S. and F.G.-P.; methodology, A.G.-S., V.M.C.F. and J.B.; validation, A.G.-S and F.G.-P.; formal analysis, A.G.-S; investigation, A.G.-S.; data curation, A.G.-S. and F.G.-P.; writing—original draft preparation, A.G.-S.; writing—review and editing, A.G.-S. and F.G.-P.; visualization, A.G.-S.; supervision, F.G.-P.; project administration, A.G.-S.; funding acquisition, F.G.-P. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was partly funded by the Barret-Honors College at Arizona State University and by the Center for Bio-mediated & Bio-inspired Geotechnics (CBBG) at Arizona State University.

**Acknowledgments:** We would like to thank the late Richard W. Castenholz for his pioneering work on the role of temperature on the ecology of cyanobacteria, which showed us the path forward.

**Conflicts of Interest:** The authors declare no conflict of interest.
