Phylogenetic analysis
For both the mycobiont and photobiont molecular phylogenies from multi-locus sequence data (nrITS, mtSSU and RPB1 for the mycobiont (140 samples) and nrITS, psbJ-L and COX2 for the photobiont (139 samples) were inferred (Supplementary Figs. S1 and S3 online). Additionally, phylogenies based solely on the marker nrITS were calculated (Supplementary Figs. S2 and S4 online), to include samples where the additional markers were not available. Both analyses include only accessions from the study sites (Fig. 1; Table 1). The phylogenies based on the multi-locus data were congruent to the clades of the phylogenies based on the marker nrITS. Thus, in the following, the focus will be only on the latter.
Mycobiont
The final data matrix for the phylogeny based on the marker nrITS comprised 306 single sequences with a length of 550 bp. It included sequences of the families Lecanoraceae and Lecideaceae. The phylogenetic tree was midpoint rooted and shows a total of 19 strongly supported clades on species level, assigned to five genera. The backbone is not supported and therefore the topology will not be discussed. All genera are clearly assigned to their family level and are strongly supported. Only Lecanora physciella forms an extra clade as sister to the families Lecideaceae and Lecanoraeae, which is not the case at the multimarker phylogeny. L. physciella has still an uncertain status, because of morphological similarities to both sister families6. The clade of the genus Lecidea revealed seven species (L. andersonii, L. polypycnidophora, L. UCR1, L. sp. 5, L. lapicida, L. cancriformis and L. sp. 6), Lecanora five species (L. physciella, L. sp. 2, L. fuscobrunnea, L. cf. mons–nivis, L. sp. 3), Carbonea three species (C. sp. URm1, C. vorticosa, C. sp. 2), and Lecidella three species (L. greenii, L. siplei, L. sp. nov2). The samples allocated to the genus Rhizoplaca were monospecific (R. macleanii). The taxonomical assignment of the obtained sequences were based on the studies of Ruprecht et al.48 and Wagner et al.10.
Photobiont
The final data matrix for the phylogeny based on the marker nrITS comprised 281 single sequences with a length of 584 bp. The phylogenetic tree was midpoint rooted and shows six strongly supported clades, assigned to seven different OTU levels67, using the concept of Muggia et al.51 and Ruprecht et al.48. All of the OTUs belong to the genus Trebouxia (clades A, I, S), comprising Tr_A02, Tr_A04a, Tr_I01, Tr_I17, Tr_S02, Tr_S15 and Tr_S18. Photobiont sequences taken from Perez-Ortega et al.50, which were labelled only with numbers, were renamed to assign them to the appropriate OTUs48.
Analysis of spatial distribution
In general, the most common mycobionts species were Lecidea cancriformis (94 of the 306 samples), Rhizoplaca macleanii (51 samples) and Lecidella greenii (37 samples), followed by Carbonea sp. 2 (13 samples), C. vorticosa (11 samples), Lecidea polypycnidophora (10 samples) and Lecidella siplei (10 samples; see Supplementary Fig. S5 online). Nine mycobiont species were found exclusively in area 5 (MDV, 78°S): Carbonea vorticosa, Lecanora cf. mons-nivis, L. sp. 2, Lecidea lapicida, L. polypycnidophora, L. sp. 5, L. sp. 6, L. UCR1 and Rhizoplaca macleanii. On the other hand, only Lecidea cancriformis was found in all the six areas; Lecanora fuscobrunnea was present in all the areas with the exception of area 2.
The most common photobiont OTUs were Tr_A02 (165 of the 281 samples) and Tr_S02 (59 samples), both of them occurring in all the six different areas, followed by Tr_S18 (32 samples), Tr_S15 (10 samples, confined to area 5) and Tr_I01 (10 samples). However, of the 149 photobiont accessions of area 5, 134 (89.93%) were assigned to Tr_A02. This percentage is much higher than in the other areas (area 1: 44.44%, area 2: 69.23%, area 3: 21.74%, area 4a: 7.69%, area 4b: 6.67%), even if those samples with mycobionts occurring exclusively in area 5 (see above) were excluded (76.56% of the 64 remaining samples are assigned to Tr_A02).
The alpha, beta and gamma diversity values are given in Table 2. For the mycobionts, the alpha diversity of the communities was the highest in area 5 (8.93, which results in nine species) and the lowest in area 4b (two species, 1.88). In contrast, for the photobionts, the lowest alpha diversity value was found in area 5 (two OTUs, 1.50) and the highest in area 4a (four OTUs, 4.06). Thus, referring to this, area 5 plays a remarkable role: compared to the other areas, it shows the highest diversity of mycobiont species on the one hand and the lowest diversity of photobiont OTUs on the other hand.
The beta diversity values (diversity of local assemblages) for mycobiont species and photobiont OTUs are quite similar (1.69 and 1.64, respectively). This is in contrast to gamma diversity values: the overall diversity for the different areas within the whole region is much higher for the mycobionts (ten species, 9.92) than for the photobionts (three OTUs, 3.35).
For mycobionts, the overall sample coverage equals to 0.993. That means that the probability for an individual of the community to belong to a sampled species is 99.3%, or, from another point of view, the probability for an individual of the whole community to belong to a species that has not been sampled is 0.7%. The sample coverage is highest for area 4b (1.000) and lowest for area 2 (0.771). Sample coverage values of the other areas are in between (area 1: 0.895, area 3: 0.931, area 4a: 0.939, area 5: 0.981). The rarefaction/extrapolation curves for the mycobiont species (see Supplementary Fig. S6a) suggest that for any sample size up to the specified level of sample coverage of 0.95, alpha diversity within area 4b is significantly lower than alpha diversity within any other area, and alpha diversity within area 5 is significantly greater than that of area 4a and 4b (based on 95% confidence intervals).
For photobionts, the overall sample coverage as well as the sample coverages of area 1, area 2, area 3, area 4b as well as area 5 is equal 1.000. Only the sample coverage of area 4a (0.951) differs. The rarefaction/extrapolation curves for the photobiont OTUs (see Supplementary Fig. S6b) suggest that for any sample size up to the specified level of sample coverage of 0.95, alpha diversity within area 1 is significantly lower than alpha diversity of area 3 and 4a and significantly greater than that of area 5. Alpha diversity of area 5 is significantly lower than that of area 1, area 3 and area 4a.
Influence of environmental factors (elevation, precipitation and temperature)
First, the proportion of the OTU Tr_A02 samples was significantly correlated to BIO10 means of the areas (R = 0.87, p = 0.022; see Supplementary Fig. S7 online): the higher the temperature mean values of the warmest quarter of an area, the higher the proportion of samples containing photobionts that are assigned to Tr_A02.
The alpha diversity values of mycobiont species significantly positively correlated with BIO10 (R = 0.88, p = 0.021; see Supplementary Fig. S8 online): the higher the temperature mean values of the warmest quarter, the higher the mycobiont diversity within this particular area.
Furthermore, the differences in mycobiont species community composition were significantly related to BIO10 (constrained principal coordinate analysis: F = 14.7137, p = 0.001, see Supplementary Fig. S9 online), BIO12 (F = 2.7535, p = 0.012), elevation (F = 2.5108, p = 0.025) and the geographic separation of the samples (Mantel statistic r = 0.1288, p = 0.0002).
The differences in community composition of photobiont OTUs were related significantly to BIO10 (constrained principal coordinate analysis: F = 48.5952, p = 0.001, see Supplementary Fig. S10 online), BIO12 (F = 4.4848, p = 0.008), elevation (F = 6.8608, p = 0.002), and physical distance (Mantel statistic r = 0.4472, p = 0.0001).
Haplotype analysis
Haplotype networks were computed for the mycobiont species and photobiont OTUs with h ≥ 2 and at least one haplotype with n ≥ 3 (Carbonea sp. 2, Lecanora fuscobrunnea, Lecidea cancriformis, Lecidella greenii, L. siplei, L. sp. nov2 and Rhizoplaca macleanii, as well as Tr_A02, Tr_I01 and Tr_S02), in both cases based on nrITS sequence data (Figs. 2, 3). The samples of Carbonea vorticosa (11) were all assigned to a single haplotype, which was also true for Lecidea polypycnidophora (10 samples), Tr_S15 (10 samples) and Tr_S18 (32 samples). Figure 3b, c illustrate the subdivision of Tr_I0151 into Tr_I01j35,48 and Tr_I01k (in this study), and the subdivision of Tr_S02 into Tr_S0235, and Tr_S02b and Tr_S02c48.
Haplotype networks of mycobiont species with h ≥ 2 and at least one haplotype with n ≥ 3, showing the spatial distribution within the different areas, based on nrITS data. (a) Carbonea sp. 2, (b) Lecanora fuscobrunnea, (c) Lecidea cancriformis, (d) Lecidella greenii, (e) Lecidella siplei, (f) Lecidella sp. nov2, (g) Rhizoplaca macleanii. Roman numerals at the center of the pie charts refer to the haplotype IDs; the italic numbers next to the pie charts give the total number of samples per haplotype. The circle sizes reflect relative frequency within the species; the frequencies were clustered in ten (e.g. the circles of all haplotypes making up between 20 and 30% have the same size). Note: only complete sequences were included.
Haplotype networks of photobiont OTUs with h ≥ 2 and at least one haplotype with n ≥ 3, showing the spatial distribution within the different areas, based on nrITS data. (a) Tr_A02, (b) Tr_I01, (c) Tr_S02. Roman numerals at the center of the pie charts refer to the haplotype IDs; the italic numbers next to the pie charts give the total number of samples per haplotype. The circle sizes reflect relative frequency within the species; the frequencies were clustered in ten (e.g. the circles of all haplotypes making up between 20 and 30% have the same size). Note: only complete sequences were included.
The haplotype networks include pie charts showing the occurrence of the different haplotypes within the different areas. All haplotypes of Rhizoplaca macleanii are restricted to area 5, as well as Lecidella greenii mainly to area 5 and areas 1 and 4a, and Lecidella sp. 2 to areas 2 and 3. However, all other species do not suggest a spatial pattern with different haplotypes being specific for different areas. Moreover, the distribution turned out to be rather unspecific, with a great part of the haplotypes found in multiple areas. For the sake of completeness, additionally, haplotype networks based on multi-locus sequence data were computed for the most abundant mycobiont species and photobiont OTU with multi-locus data available (Lecidea cancriformis and Tr_S02). Not surprisingly, those networks show a greater number of different haplotypes, but they also do not allow conclusions concerning spatial patterns of area specific haplotypes (see Supplementary Fig. S11 online).
Diversity and specificity indices of mycobiont species and photobiont OTUs
The diversity and specificity indices for the different mycobiont species and photobiont OTUs are given in Supplementary Table S8 online.
For the sample locations of mycobiont species with n ≥ 10, BIO10 was strongly correlated to the specificity indices NRI (net relatedness index) and significantly correlated to PSR (phylogenetic species richness) and 1 – J′ (Pielou evenness index). BIO12 was significantly correlated to NRI, PSR and 1 – J′. Figure 4 illustrates these correlations: the higher the BIO10 and BIO12 mean values, the higher was the NRI (phylogenetic clustering of the photobiont symbiotic partners), the lower was the PSR (increased phylogenetically relatedness of photobiont symbiotic partners) and the higher was 1 – J′ (less numerically evenness of the photobiont symbiotic partners). Thus, for the mean values of the sample locations of a mycobiont species, a comparatively high temperature of the warmest quarter and high annual precipitation occurs with associated photobionts that are phylogenetically clustered and closer related to each other. The lowest values of NRI and the highest values of PSR were developed by Lecidea cancriformis and Lecanora fuscobrunnea, which also showed the lowest BIO10 and BIO12 mean values at their sample sites. On the contrary, the highest values of NRI and PSR were developed by Rhizoplaca macleanii, which also had the highest BIO10 and BIO12 means.
Correlation plots. Specificity indices NRI (net relatedness index), PSR (phylogenetic species richness and 1 – J′ (Pielou evenness index) against mean values of BIO10 (mean temperature of warmest quarter) and BIO12 (annual precipitation) for mycobiont species with n ≥ 10.
For the sample locations of photobiont OTUs with n ≥ 10, elevation significantly negatively correlated with h (number of haplotypes) and Hd (haplotype diversity): the higher the mean elevation of sample sites, the lower the number of haplotypes and the lower the probability that two randomly chosen haplotypes are different (Fig. 5). The highest values of h and Hd were shown by Tr_A02, Tr_I01 and Tr_S02, which occurred at sample sites with comparatively low elevations. In contrast, Tr_S15 and Tr_S18 occurred at very high elevations and showed very low values of h and Hd.
Correlation plots. Diversity indices h (number of haplotypes) and Hd (haplotype diversity) against mean elevation of sample sites for photobiont OTUs with n ≥ 10.
Analysis of mycobiont–photobiont associations
Bipartite networks were calculated for all associations between mycobiont species (lower level) and the respective photobiont OTUs (higher level) for all areas (Fig. 6). The H2′ value (overall level of complementary specialization of all interacting species) was highest in area 2 (0.921), indicating a network with mostly specialized interactions: within this network, with the exception of Lecidea andersonii, the mycobiont species are associated exclusively with one single photobiont OTU. The second highest H2′ value was developed by area 4b (0.710); in contrast, area 4a showed the lowest H2′ value (0.260), with the most abundant mycobiont species Lecidea cancriformis showing associations with five different photobiont OTUs. The H2′ values of area 1, area 3 and area 5 indicate medium specification.
Bipartite networks showing the associations between mycobiont species and photobiont OTUs for the different areas. Rectangles represent species/OTUs, and the width is proportional to the number of samples. Associated species/OTUs are linked by lines whose width is proportional to the number of associations.
In addition, the bipartite networks illustrate the different occurrence of mycobiont species and photobiont OTUs within the different areas: For example, in area 1 (and area 2), five (seven) different mycobiont species are associated with only three different photobiont OTUs. In contrast, in area 4b, only two different mycobiont species are associated with four different photobiont OTUs. In area 5, the number of associated photobiont OTUs is also four, but those four OTUs are associated with 16 different mycobiont species.
The network matrix giving all the associations between the mycobiont species and photobiont OTUs is presented in Supplementary Table S9 online.
Source: Ecology - nature.com
