Keywords

16.1 Introduction

A range of biotic, abiotic and historical variables shape the structure and species richness of ectomycorrhizal (EcM) fungal communities. Host and fungal compatibility (i.e. host preference or specificity) that may vary widely across host taxa, has been increasingly shown to influence the structure and richness of EcM fungal assemblages at various taxonomic levels of plants (Ishida et al. 2007; Morris et al. 2008; Tedersoo et al. 2010b, 2013; Bahram et al. 2012; Põlme et al. 2013). To date, the influence of environment on EcM fungal richness and composition has received relatively limited attention and the biodiversity of several tropical plant groups has remained unknown. Unlike in most other organisms and fungi overall, EcM fungi exhibit greater diversity in temperate compared with tropical and arctic ecosystems (Tedresoo et al. 2012, 2014; Chap. 18). This has been ascribed to historical factors (the lack of earliest evolving Pinaceae hosts in lowland tropical habitats), rapid turnover of organic matter in tropical soils and low relative abundance of hosts.

The genus Coccoloba (Polygonaceae) comprises ca. 170 species of shrubs and trees with neotropical distribution (Howard 1960; Chap. 19). Coccoloba spp. are probably of South American origin, with greatest richness in northern Amazonia and the Atlantic rain forest (Howard 1961; Chap. 20). EcM associations were first described from sea grape (Coccoloba uvifera (L.) L.) in Cuba (Kreisel 1970) and thereafter in South American rain forest species (Moyersoen 1993; Bereau et al. 1997). The associated EcM fungi are poorly known, because host plants have remained unsettled in most South American mycological studies (Pegler 1983; Singer et al. 1983; Roy et al. 2016; but see Tedersoo et al. 2010b).

Sea grape is a tropical tree species with a native distribution from southern Florida to eastern Mexico to northeastern Brazil. Sea grape grows in immediate proximity of seashores, representing one of the first colonizers of sandy and rocky coastal areas. These habitats are characterized by high salinity, steady wind, seasonal drought and low soil nutrient availability. EcM symbiosis could potentially enhance plant tolerance of those harsh conditions, particularly of high salinity (Bandou et al. 2006). Séne et al. (2015) conducted a profound study sampling sporocarps and EcM root tips in sea grape communities and reported very limited EcM fungal diversity in Guadeloupe Island. The authors postulated four nonexclusive hypotheses to explain the low EcM fungal diversity: (1) isolation from mainland, (2) overall lower EcM fungal diversity at lower latitudes, (3) environmental filtering due to stressful conditions, (4) recent origin of EcM symbiosis within the Polygonaceae. In spite of profound local sampling effort in a small volcanic island, their study cannot be used to generalize about sea grape EcM communities over a wider geographic context.

In order to test the above hypotheses in a wider geographical and historical context, we sampled six additional sea grape communities around the Caribbean basin and compared these in the biogeographic perspective. We further addressed the potential specificity and origin of EcM fungi of C. uvifera by comparing the associated Tomentella species—the dominant group of mycobionts—to these from other hosts in North and South America.

16.2 Approaches

We performed root sampling in six study sites in the following locations: USA: Miami (June 2013; 26.0408°N; −80.1145°E), Mexico: Celestún (October 2015; 20.9336°N; −90.3748°E), Costa Rica: Cahuita (June 2013; 9.8590°N; −82.9458°E), Cuba: Cayo Santa Maria (December 2008; 22.6581°N; −79.0413°E), French Guiana: Montabo (November 2013; 4.9436°N; −52.2973°E) and Colombia: Los Naranjos (November 2014; 11.2973°N; −73.8946°E). From each site, roots from ten sea grape individuals were collected (except for the Cuba plot where 14 samples were collected). Randomly selected samples (15 × 15 cm to 10 cm depth) were situated at least 10 m apart.

Roots were cleaned from adhering soil in tap water and morphotyped under a stereomicroscope. EcM morphotypes were distinguished based on colour and roughness of mantle, presence of emanating hyphae and rhizomorphs. At least two EcM root tips from each morphotype per soil sample were stored in CTAB buffer (1% cetyltrimethylammonium bromide, 100 mM Tris–HCL (pH 8.0), 1.4 M NaCl, 20 mM ethylenediaminetetraacetic acid) for molecular analyses.

DNA was extracted from EcM root tips using Thermo Scientific Phire Plant Direct PCR Kit (Thermo Scentific, Waltham, MA, USA) according to the manufacturer’s instructions. In the course of the study, PCR was performed by use of 5× HOT FIREPol Blend Master Mix Ready to Load (Solis BioDyne, Tartu, Estonia). In EcM root tips, fungal rDNA Internal Transcribed Spacer (ITS) region was amplified with a forward primer ITSOF-T (5′-acttggtcatttagaggaagt-3′) in combination with reverse primers LB-W (5′-cttttcatctttccctcacgg-3′) or TW13 (5′-ggtccgtgtttcaagacg-3′). In case of PCR failure, we combined ITSOF-T with universal primers ITS4 (5′-tcctccgcttattgatatgc-3′), or basidiomycete-specific primer ITS4B (5′-caggagacttgtacacggtccag-3′) and LR0B (5′-acccgctgaacttaagc-3′) in order to amplify a shorter fragment of fungal DNA. To improve sequence quality, some root tip extracts were re-amplified with taxon-specific primers (Tedersoo et al. 2008). PCR and sequencing were run following Põlme et al. (2013). Sequences were assembled, checked, trimmed and manually corrected in Sequencher 5.1 software (GeneCodes Corp., Ann Arbor, MI, USA).

Sequences were confirmed to belong to EcM fungal lineages (cf. Tedersoo et al. 2010a; Chap. 6) by use of BLASTn searches against the International Sequence Databases (INSD) or UNITE (Abarenkov et al. 2010a). Sequences were partitioned into operational taxonomic units (OTUs), defined as a group of sequences sharing at least 97% pairwise similarity. Sequences with sufficient length and quality were assigned to UNITE species hypothesis (SHs; Kõljalg et al. 2013) with 3% dissimilarity threshold. We also included the recently published data of Séne et al. (2015) from Guadeluope. This study covered four study plots and had a much higher per-site sampling effort.

Using the PlutoF web platform (Abarenkov et al. 2010b), we downloaded all Tomentella sequences originating from South, Central and Northeast America. Identical sequences from the same sampling plots and hosts were removed prior to phylogeny construction. All sequences were aligned using MAFFT software (Katoh and Standley 2013). The alignment was manually adjusted in AliView software (Larsson 2014) and Maximum likelihood analysis was performed in FastTree 2.1 (Price et al. 2010) using default settings, with Odontia fibrosa (UDB000284) as an outgroup (Tedersoo et al. 2015).

16.3 Fungal Diversity

Out of 147 EcM root tips subjected to molecular analyses, 131 (89%) yielded good quality sequences. These sequences (including material collected by Séne et al. 2015 from Guadeloupe) were clustered into 42 OTUs (Table 16.1). Altogether 33 of OTUs were accommodated to existing SHs at the 97% sequence similarity cutoff level. Of these SHs, 26 (78.8%) were exclusively associated with sea grape. Only five fungal OTUs found from newly sampled sites overlapped with the Guadeloupe study, whereas eight OTUs remained exclusive to Guadeloupe.

Table 16.1 Distribution of fungal species hypothesis and OTUs at 97% cutoff level, associating with Coccoloba uvifera individuals in seven sampling areas
Full size table

The /tomentella-thelephora was by far most taxon-rich phylogenetic lineage of EcM fungi comprising 21 OTUs. Other sea grape-associating lineages were represented with the following number of OTUs: /boletus (four), /inocybe (four), /clavulina (three), /pisolithus-scleroderma (two), /sebacina (two), /paxillus-gyrodon (two), /russula-lactarius (one), /cantharellus (one), /serendipita (one; questionable mycorrhizal status) and /cenococcum (one). In spite of taxonomical richness of the /tomentella-thelephora lineage, Scleroderma bermudense (SH003700.07FU) and Scleroderma sp. (SH003702.07FU) were the most common individual taxa that were present in six and four sites out of seven, respectively.

Florida constituted the most OTU-rich site harbouring 13 EcM fungal taxa, followed by Cuba (11 OTUs), Colombia (10) and Costa Rica (9). Interestingly, Mexico and French Guiana sites harboured only two and one EcM fungal OTU, respectively (Fig. 16.1). In comparison, altogether 14 EcM OTUs were identified from root tips from four plots in Guadeloupe, with 4–9 EcM fungal OTUs per plot (Séne et al. 2015), but this can be ascribed to more extensive sampling effort on a local scale. None of the sea grape-associated OTUs overlapped with EcM fungal taxa associated with the five sampled Coccoloba spp. in Ecuador (Tedersoo et al. 2010b).

Fig. 16.1
figure 1

Map of study sites indicating number of fungal OTUs in each sampling area. Shared OTUs between areas are shown with arrows. Note that inland Coccoloba species from Ecuador (Tedersoo et al. 2010b) do not share any fungal OTUs with Coccoloba uvifera communities from the coastal areas

Full size image

The overall OTU richness detected over a broader geographic scale increased nearly three-fold compared to the single study conducted in Guadeloupe (Séne et al. 2015). In addition to formerly reported EcM lineages, we found /sebacina, /clavulina and /boletus lineages from our sampling sites, which were not found in Guadeloupe. Interestingly, several EcM fungal lineages such as /paxillus-gyrodon, /serendipita, /russula-lactarius and /cantharellus were not found outside Guadeloupe. The two Guadeloupean species of Melanogaster belonging to the /paxillus-gyrodon lineage are remarkable, because this genus and the entire EcM lineage is not known to associate with hosts of tropical origin. Species of Cantharellus fruited in abundance in another C. uvifera stand ca. 3 km distant from the Colombian site, but ITS sequencing of these fruit-bodies failed, indicating either primer bias or extensive length of the ITS region (Tedersoo et al. 2016). It is likely that additional sampling effort would have revealed these taxa from the mainland as well and that the true EcM fungal species richness associated with sea grape is considerably higher than the currently reported 42 fungal OTUs (Fig. 16.2). However, the coarse structure of EcM fungal communities was relatively similar to that previously reported in Guadeloupe i.e. /tomentella-therephora being most taxon rich lineage and Scleroderma bermudense being the most abundant fungal taxon. Sites in French Guiana and Mexico were extremely species-poor, comprising only one and two fungal OTUs respectively. In French Guiana, S. bermudense colonized sea grape root systems in all samples. The French Guiana site was characterized by intense anthropogenic disturbance in addition to a small host tree population. However, the Florida site with the highest OTU richness was also characterized by substantial anthropogenic impact but with considerably larger host population.

Fig. 16.2
figure 2

Rarefied OTU accumulation curve of Coccoloba uvifera associated EcM fungi found in Caribbean basin. Closed circles and open circles represent the rarefied curve and its 95% confidence intervals, respectively. Triangles and squares represent the values of Chao2 and Jacknife2 minimum richness estimators, respectively. The values were calculated based on 999 permutations using EstimateS 9 (Colwell 2013)

Full size image

16.4 Environmental Filtering and Host Specificity

Séne et al. (2015) proposed that the impoverished EcM fungal richness detected in Guadeloupe could partly result due to a founder effect and isolation from mainland. Although our sampling intensity is too low for comprehensive statistical comparison, the mainland and Guadeloupe sea grape communities harboured comparable EcM fungal richness, largely refuting this hypothesis. In spite of phylogenetic and geographical proximity, the absence of shared OTUs with inland Coccoloba species from neotropical forest, which harboured higher EcM diversity (Tedersoo et al. 2010b), supports the hypothesis of environmental filtering, also proposed by Séne et al. (2015) as an alternative. This explanation coincides well with the fact that vast majority of the SHs detected were exclusively associated with the sea grape. Similarly, Pisonia grandis (Nyctaginaceae) inhabiting small and often guano-rich Indian Ocean and Pacific islands harbours species poor and highly specific EcM assemblage (Suvi et al. 2010; Hayward and Horton 2012). Therefore, putative host specificity in such stressful habitats is most likely bounded with environmental filtering (but see Hayward and Horton 2012). The strong intrageneric ecological specificity in EcM fungi associated with Coccoloba contrast to the largely genus-level specific EcM fungi of Alnus (Põlme et al. 2013). Interestingly, environmental filtering is likely to have an important role in driving specificity of EcM interactions in both cases (Huggins et al. 2014). Nevertheless, we are unable to confidently disentangle the cause and consequence between the ecological host specificity and environmental filtering, because genetic and physiological compatibility between host and symbiont is likely to evolve mutually in extreme habitats over extended periods of time. Séne et al. (2015) pointed out that EcM origin in Coccoloba is relatively recent and this also holds true for Pisonia (Chap. 19), possibly explaining low EcM diversity in both groups. The fact that numerous EcM host taxa, with much broader range of mycobionts, have diverged in a comparable time frame (Chap. 19), makes this hypothesis disputable. Taken together, the hypothesis of strong environmental filtering seems the most plausible explanation for the low fungal diversity in C. uvifera.

16.5 Biogeography of Thelephoraceae

Using the PlutoF workbench, we were able to recover 832 sequences belonging to the /tomentella-thelephora lineage originating from eastern North America, northern South America and Central America. After removal of redundant sequences originating from the same host and study site, 525 sequences were subjected to a phylogenetic analysis (Fig. 16.3). In the large Thelephoraceae phylogram, sea grape-associated sequences clustered together more commonly with sequences originating from Central America and North America rather than with those from South America. This conflicts with the putative South American origin of the genus (Raven and Axelrod 1974; Chap. 20) and previously established patterns in Russulaceae (De Crop et al. 2017) and Sclerodermataceae (Wilson et al. 2012) putatively associating with Coccoloba spp. In terms of host identity, Thelephoraceae sequences associated with C. uvifera most often clustered together with sequences from Pinus spp. and to a lesser extent with those from Quercus spp. This suggests potential host shifts from phylogenetically distant host taxa. Previous studies focusing on the mycobionts of introduced plants support that host shifts for EcM fungi may occur between distantly related taxa such as Fagales and Pinales in a very short time frame (Bahram et al. 2013). Currently, the distribution of Coccoloba spp. overlaps with that of Pinus spp. and Fagales from Central Mexico to Nicaragua and in Cuba (Chap. 20). Furthermore, the fossil record indicates much greater overlap among the geographic range of Pinales, Fagales and Coccoloba during the Oligocene and Eocene (Gray 1960; Graham and Jarzen 1969), when Coccoloba spp. were distributed up to Central USA in the north, whereas Fagales spp. were more widely distributed in the Caribbean islands. In particular, C. uvifera shares its habitat with Pinus spp. in coastal sand dunes, although these species do not co-exist in present-day communities, except perhaps in the Bahamas. Nevertheless, the results of phylogenetic analysis should be interpreted with caution, because South America remains relatively under sampled compared to Central America and especially North America, which may slightly overestimate the links between Pinaceae and C. uvifera and particularly the genus Coccoloba in general.

Fig. 16.3
figure 3figure 3

Maximum likelihood phylogram of Thelephoraceae sequences from South, Central and Northeast America. Clades that do not contain Coccoloba associated Thelephoraceae sequences have been collapsed. Clades highlighted in purple, red and green represent Northeast, Central and South American groups, respectively. Blue, yellow and orange represent the mixture of North and Central American groups, North and South American groups, and Central and South American groups, respectively. Transparent clades represent a mixture of all three regions. Bold names indicate Thelephoraceae sequences associating with various Coccoloba species. The sequence name includes accession number, country, associating host species and species hypothesis (if available)

Full size image

We also tested, whether the habitats historically influenced by Pinaceae and Fagales (Florida, Costa Rica, Mexico, Cuba) harbour more OTUs of Thelephoraceae shared with northern hosts than the sites far from natural Pinaceae and Fagales habitats (Guadeloupe, Colombia, French Guiana). Contrary to our expectations, there were no differences between the sites with shared and unshared habitats (n = 34; χ 2 = 0.123; P = 0.726). The wide distribution of EcM fungi of potentially northern temperate origin indicates their effective dispersal and good adaptation to tropical climate.

16.6 Conclusions

Our regional-scale study provides strong evidence that sea grape harbors distinct and relatively species-poor EcM fungal communities throughout its range. These associated fungi are highly distinct from the mycobionts of other Coccoloba spp., suggesting that strong environmental filtering due to salinity and perhaps high pH may cause the high observed ecological specificity. Close affinities of C. uvifera-associated Tomentella spp. with those from the phylogenetically distant Pinaceae and Fagaceae hosts supports multiple host shifts and/or broadening of host range due to more commonly shared habitats in the past. Future phylogeographic studies including more samples from South America are needed to enlighten the evolutionary history of EcM symbiosis in Central and South America.