Species richness and root colonization of arbuscular mycorrhizal fungi in Syngonanthus elegans, an endemic and threatened species from the Cerrado domain in Brazil

Syngonanthus elegans is an endangered plant species occurring in the Brazilian Cerrado whose interaction with arbuscular mycorrhizal fungi (AMF) is poorly understood. The aim of this work was to evaluate the occurrence of AMF species and mycorrhizal colonization of S. elegans in two sampling areas named “Soberbo” stream (Soberbo) and “Parque Nacional das Sempre-Vivas” (Park), both found in Diamantina-MG, Brazil. In each area, one plot (100 x 100 m) was established, and roots and soil samples near the roots were collected from 10 plants in each plot. Further sampling included three specimens each of Loudetiopsis chrysothrix and Xyris sp.. Typical mycorrhizal colonization structures were observed in S. elegans roots, and colonization was measured at 75%. Considering both sites and all three hosts, 26 AMF species were recovered, 8 of which were identified only at the genus level. Glomus sp. 1, Scutellospora pernambucana, Acaulospora cavernata and Acaulospora mellea were classified as dominant in both areas. Other species were also considered dominant, including Glomus sp. 4 in Soberbo and Dentiscutata biornata and Gigaspora albida in Park. Trap cultures revealed the presence of seven additional species. For S. elegans, AMF species richness was slightly higher in Park than in Soberbo. Simpson diversity and evenness were slightly higher in Soberbo for S. elegans-associated AMF communities. Overall, S. elegans is highly colonized by arbuscular mycorrhizal fungi and is associated with a wide range of AMF species in the field, suggesting that this association is important for the establishment and survival of this threatened species. Some of the observed species may be new to science.


INTRODUCTION
Syngonanthus species (Eriocaulaceae) are known as star flowers or everlasting plants because their inflorescence and scapes maintain their form and color even long after harvesting (Borba;Amorim, 2007;Miranda;Giulietti, 2001).These properties contribute to star flowers being harvested and exported for ornamental purposes (Giulietti et al., 1996;Nunes et al., 2008).Syngonanthus elegans (Bong.)Ruhland, the gold feet everlasting plant, is one of the most valuable star flowers species due to its long golden scape, which is used for handmade crafts.Plants of this species are harvested by local communities in an extractivist way as an alternative source of revenue, leading this species to be considered at risk of extinction (Borba;Amorim, 2007;Menezes;Giulietti, 2000).
S. elegans is an endemic of the Rupestrian fields or Rupestrian grasslands, a unique ecoregion that covers portions of the southeastern Brazilian mountains in the transition between the Atlantic rain forest and the Cerrado domains.This ecoregion is one of the most endangered ecosystems in the world, occurring in altitudes above 900 m with rocky areas, shallow soils, sandy plateaus and experiencing drought periods alternated with hydric soil saturation (Joly, 1970;Giulietti;Pirani;Harley, 1997;Menezes;Giulietti, 2000;Conceição;Pirani, 2007).Estimates suggest that 30% of the Rupestrian fields plant taxa are restricted to this ecoregion, which is mainly dominated by Eriocaulaceae, Poaceae, Cyperaceae, Xyridaceae, Velloziaceae, Orchidaceae, and Melastomataceae (Conceição, Pirani, 2005;Costa, 2005;Carvalho et al., 2014;Zappi;Taylor, 2003).Some species belonging to the Eriocaulaceae family, such as S. elegans, S. aurifibratus Silveira and S. lineares Ruhland, have developed strategies to minimize water loss through mucilage production in the root system (Scatena;Menezes, 1996).This mucilage possibly favors the relationship of these plants with soil microorganisms such as associative bacteria and fungi including arbuscular mycorrhizal fungi (AMF).
Considering the well-established fact that AMF enhances plant nutrient uptake in low-fertility soils when associated with plant roots, mycorrhizal association in S. elegans might represent a strategy to cope with the harsh and poor soil conditions of Rupestrian fields.Mycorrhizal associations of native species growing in this region are poorly understood, but recently, studies focusing on AMF diversity in distinct habitats within Rupestrian fields (Carvalho et al., 2012) and assessing the mycorrhizal status of some plant species (Pagano;Cabello, 2012) have been developed, contributing to management conservation programs for this ecoregion (Fernanda;Echeverría;Pagano, 2012).
The occurrence of mycorrhizal association in endemic plants has been reported previously in the literature (Bashan et al., 2007;Koske;Gemma;Flynn, 1992;Koske;Gemma, 2002;Silva et al., 2001), however, reports on the mycorrhizal status of Eriocaulaceae plants are scarce.Mycorrhizal colonization up to 20% was registered in Tonina fluviatilis Aubl.(Silva et al., 2001), and the occurrence of colonization by six Glomus species was observed in Syngonanthus mucugensis Giulietti and Syngonanthus curralensis Moldenke (Borba;Amorim, 2007).Although some progress has been achieved regarding knowledge of biological systems in Rupestrian fields, there is a general lack of information regarding interactions between soil microorganisms and plants in this ecoregion.To the best of the authors' knowledge, no information regarding AMF occurrence and colonization associated with S. elegans has been published.In this study, our goal was to evaluate root mycorrhizal colonization the occurrence and species richness of AMF associated with S. elegans occurring in two Rupestrian field areas.

Study area
The study area was localized in the meridional portion of the "Serra do Espinhaço", Minas Gerais state, southeastern Brazil and is characterized by the presence of Rupestrian fields.Climate in this area is mesothermic (Cwb according to Köppen) and is characterized by moderate and humid summers and cold and dry winters (Neves;Abreu;Fraga, 2005).The average annual temperature is 18.3 °C, and average annual precipitation is 1.404 mm, with 88% of the total annual precipitation occurring during the rainy season from October to March (Vieira et al., 2010).To increase the detection capability of AMF diversity, two sampling sites that were 40 Km apart from one another were selected for this study in the municipality of Diamantina, Minas Gerais.One site was located near the headspring of the Soberbo stream, here identified as Soberbo, within the Juscelino Kubitschek campus of the Universidade Federal dos Vales do Jequitinhonha e Mucuri (18º 12' 21.1" S 43º 33' 47.6" W) at an altitude of 1.368 m.The other site was in the National Park "Parque Nacional das Sempre-vivas", here referred to as Park (17º55`02.9"S 43º47`11.2"W) with an altitude of 1.310 m.Soils in both sites were classified as typic quartzipsamment according to Empresa Brasileira de Pesquisa Agropecuária -Embrapa (2006), which correspond to Entisols according to Soil Survey Staff (1999).

Study design and sampling
One plot (100 x 100 m) was established within each site, and roots along with soil samples near the roots were collected from 10 specimens of S. elegans during the dry season.Three specimens each of Loudetiopsis chrysothrix (Nees) Conert (Poaceae) and Xyris sp.
COSTA, H. A. O. et al (Xyridaceae) were also sampled for comparison (control).Soil samples were stored in plastic bags and kept in thermo boxes until transportation to the laboratory and stored at 4 o C until analysis.A composite soil sample from each area was sent to the laboratory for chemical and physical analysis (Table 1).
Roots were washed under tap water, cut in 1-2 cm fragments and stored in a 50% alcohol solution.Roots were stained according to Phillips and Hayman (1970) with modifications in the bleaching step.For this, roots were placed in a 10% KOH solution for 30 min.This procedure was repeated three times; the roots were subjected to a 70 o C water bath for two min during the third repeat of the procedure.The percentage of colonized root length by AMF was evaluated according to the grid-line method proposed by Giovannetti and Mosse (1980).
From each soil sample, a 50 g aliquot of soil was used to extract AMF spores using the wet sieving method (Gerdemann;Nicolson, 1963) and two stacked sieves with 710 µm and 45 µm openings, followed by sucrose gradient centrifugation (20% and 60%).Spores were then separated by morphotypes and counted under a dissecting microscope.For identification, spores were placed on glass slides in polyvinyl alcohol-lactic acid glycerin (PVLG) with PVLG + Melzer´s reagent as the mounting media.AMF species were identified based on the subcellular structures of asexual spores following descriptions of living reference cultures at INVAM (http://invam.caf.wvu.edu), and by comparison with original species description protocols.Classification of AMF into families and genera followed that of Redecker et al. (2013).
To recover AMF species that were not sporulating at the sampling time, trap cultures were established according to Stutz and Morton (1996).For each plant species, three pots of trap cultures were established by mixing 750 g of field soil with sterilized sand (3 × at 121 °C for 60 min).This mixture was placed in 1.6 L pots and sown with 20 seeds of Brachiaria brizantha (superficially disinfected with 70% alcohol for 2 min and 1% sodium hypochlorite for 1 min).Plants were kept under greenhouse conditions and watered daily if necessary.Every two weeks each pot received 20 mL of Hoagland's (Hoagland;Arnon, 1950) nutrient solution without phosphorus.After 3 months, plants were allowed to dry out in situ by suspending watering.Then, shoots and the top 2 cm of substrate from each pot were discarded.The remaining content of each pot was thoroughly homogenized and 50 g used to extract and identify AMF spores as previously described.
Frequency of occurrence (F i ) for each AMF species was calculated according to the equation F i = (J i /K)*100, where F i is the frequency of occurrence of species "i", J i represents the number of samples that species "i" was present, and K is the total number of samples for each site (Brower;Zar;Von Ende, 1998).According to the frequency of occurrence, AMF species were classified as dominant (D) (F i > 50%), most common (MC) (30% < F i ≤ 50%), common (C) (10% < F i ≤ 30%), and rare (R) (F i ≤ 10%) (Zhang;Gui;Liu, 2004).Species richness was obtained from the number of AMF species, and spore abundance of each species was used to calculate diversity indices according to Magurran (1988).Simpson diversity indices were calculated using the equation 1-Is, where: Is = Σ n i (n i -1)/N(N-1), with n i being the number of individuals of each species and N the total number of individuals.Simpson dominance indices were calculated using the equation Σ (n i /N) 2 .The Pielou evenness index was obtained using the equation H`/ln (S), wherein H` = -Σ p i ln p i e p i = n i /N and S is the total number of species.

Statistical analysis
The number of spores and percentage of colonized root length by AMF were transformed using ln (x + 1) because they do not follow a normal distribution.The number of spores between areas was submitted to an analysis of variance (ANOVA).

RESULTS AND DISCUSSION
Roots of S. elegans were colonized by typical mycorrhizal structures such as hyphae, vesicles and arbuscules.Percentage of mycorrhizal root lengths in S. elegans averaged 75%, and no differences were detected between plants in Soberbo and Park (P < 0.05).Root colonization of L. chrysothrix and Xyris sp.averaged 87% and 76%, respectively.This work represents the first report in the literature relating the occurrence of arbuscular mycorrhiza in S. elegans, L. chrysothrix and Xyris sp.. Analysis of the root system showed typical arbuscular mycorrhizal structures including arbuscules, vesicles and internal mycelium and a percentage of mycorrhizal root colonization ranging from 75 to 87%.Based on these high levels of mycorrhizal colonization in the field, these plants can be considered mycotrophic.Levels of mycorrhizal colonization in all three hosts were relatively high compared with other species in the Eriocaulaceae.For example, 20% of root colonization was measured in Tonina fluviatilis (Silva et al., 2001) and up to 67% in Paepalanthus bromelioides (Pagano;Scotti, 2009).The low nutrient content in soils (Table 1) and the mild climate (average annual temperature between 19 and 18 °C) of the studied areas may contribute to a low rooting turnover that results in high colonization levels.In low-fertility soils, plants usually maintain high levels of arbuscular mycorrhizal colonization (Powers;Treseder;Lerdau, 2005;Saggin Júnior et al., 1995;Treseder, 2004).The similar values of root mycorrhizal colonization between Soberbo and Park may be explained by their similar soil properties, climate and altitude.
A total of 23 AMF species were detected in the soil samples near roots of S. elegans, and the number of species was slightly higher at Park (20 species) than at Soberbo (16 species) (Table 2).AMF species richness in soil under S. elegans tended to be higher than under L. chrysothrix and Xyris sp..This AMF species richness associated with S. elegans can be considered relatively high (Table 2) compared with other Syngonanthus species.Only six AMF species were associated with S. mucugensis and four species with S. curralensis, both of which also occur in the "Serra do Espinhaço Meridional" in Bahia state, Brazil (Borba, Amorim, 2007).Comparing our results with those found in other endemic plants species, AMF species richness associated with S. elegans was of the same magnitude as that detected in Fouquieria columnaris (Kell) Kell.(23 species) endemics of Mexico (Bashan et al., 2007), and 5-6-fold that of other registered endemic plants such as Dubautia menziesii (A.Gray) Keck D (4 species) (Koske;Gemma, 2002) and P. bromelioides (3 species) (Pagano;Scotti, 2009).
We attributed this high species richness associated with S. elegans to our sampling strategy (ten plants in two locations) and the use of trap cultures.Trap cultures are important to trigger AMF sporulation, especially in environments with hydric scarcity such as Rupestrian fields.Stutz and Morton (1996) observed that AMF species richness from a desert environment was comparable to that detected from mesic environments and hypothesized that trap cultures provide a more homogenous and constant environment that favors AMF sporulation (Stutz;Morton, 1996).Considering that Rupestrian fields experience periods of water scarcity and that our sampling was conducted during the dry season, the trap culture environment certainly acted as a promoter of AMF sporulation.
Across both sites and all three hosts, a total of 26 AMF species was recovered (Table 2).Eight of these species could be identified only to the genus level and putatively represent new species for science.Some of the nonidentified species produced very few spores, precluding a more thorough comparison to delimit species.The presence of unknown species in Brazilian ecosystems is common, where 20-26% of AMF species in many surveys are not identified (Stürmer;Siqueira, 2006).Rupestrian fields represent an ecosystem with a high endemism for plant species (Conceição;Pirani, 2007) that could be associated with unique AMF species and may be considered a hotspot for this group of fungi (Carvalho et al., 2012).
In the field, S. elegans was associated with AMF belonging to Gigasporaceae, Acaulosporaceae, Glomeraceae and Ambisporaceae, with the last family detected only in Soberbo (Table 2).The proportion of species per family differed between both areas (Figure 1a).In Soberbo, Acaulosporaceae represented 47% of the total number of species recovered, while in Park, species were evenly distributed among Gigasporaceae (33%), Acaulosporaceae (29%) and Glomeraceae (29%).Despite the higher number of Acaulosporaceae in Soberbo, the AMF community associated with S. elegans was evenly dominated by the three main families of Glomeromycota (Table 2, Figure 1a), which contrasts with the AMF communities detected under S. mucugensis and S. curralensis that were dominated by members of Glomeraceae (Borba;Amorim, 2007).However,  this does agree with a survey of a mosaic of habitats in Rupestrian fields that were dominated by Glomeraceae, Acaulosporaceae, and Gigasporaceae (Carvalho et al., 2012).Soil pH in Soberbo and Park were 4.6 and 5.0, respectively (Table 1), and this soil factor could have favored dominance of Acaulosporaceae and Glomeraceae, as organisms pertaining to both families are often found in soil with a pH from 4.0 to 6.0 (Oehl et al., 2006;Silveira, Souza;Koller, 2002;Stürmer, 1999).Other studies in Rupestrian fields mentioned above recorded either zero (Borba;Amorim, 2007) or 12 species of Gigasporaceae (Borba;Amorim, 2007;Carvalho et al., 2012), while our study detected four genera and 10 species from this family (Table 2), suggesting that members of this family are also important components of the mycorrhizal community associated with S. elegans.Genera associated with S. elegans included Dentiscutata, Racocetra, Scutellospora, Gigaspora, Acaulospora, Rhizophagus, Glomus and Ambispora (Figure 1b).Three and five S. elegans-associated species were found exclusively in Soberbo and Park, respectively (Table 2).Acaulospora and Glomus accounted for most of the AMF species in both areas (Figure 1b).It is interesting to note that AMF species associated with S. elegans came from four families (Table 2, Figure 1a) and eight genera (Table 2, Figure 1b) of Glomeromycota, spanning all major monophyletic clades within the phylum, suggesting that the environmental conditions of Rupestrian fields do not restrict fungal taxa establishment once they arrive at these sites.Considering the consensus classification of (Redecker et al., 2013), eight out of 18 genera and four out of 9 families of AMF were detected herein.
The number of AMF spores observed in soil samples near roots of all three hosts varied from 11 to 390 in 50 g of soil, with the largest number found under S. elegans at Soberbo and the smallest value in Xyris sp. at Park (Figure 2).AMF spores number recovered from soil samples near roots of S. elegans was significantly higher in Soberbo (390 spores per 50 g of soil) compared to Park (245 spores per 50 g soil) (P < 0.05) (Figure 2).The same tendency was observed for L. chrysothrix and Xyris sp..This trend of recovering a higher number of spores in Soberbo compared to Park (Table 3, Figure 2) is intriguing considering that the host identity and sampling time were the same in both areas.We have no explanation for this pattern, but we speculate that the higher humidity in Soberbo (Table 1) could have favored AMF sporulation.The number of AMF spores associated with S. elegans (Table 3, Figure 2) can be considered high compared to other plants from the same family occurring in the same ecoregion.A total of 121 spores per 50 g soil were recovered under S. curralensis growing in the "Serra do Espinhaço Meridional" in Bahia state (Borba;Amorim, 2007) and 39-70 spores per 50 g soil were detected in P. bromelioides in the "Serra do Cipó" Minas Gerais state (Pagano;Scotti, 2009).We acknowledge that our results are from one sampling time only, but this indicates that S. elegans supports high sporulation and diversity of AMF.Most field-produced spores in both sites pertained to Glomus sp. 1, Acaulospora cavernata, and Scutellospora pernambucana, which together accounted for 58.5% and 66.2% of the total number of spores in Soberbo and in Park, respectively (Table 3).This result indicates that host identity and environmental conditions work together to influence sporulation of AMF species in Rupestrian fields, and comparative studies such as this one are important to elucidate biotic and abiotic factors that influence the auto-ecology of AMF species.This dominance by three species in the community structure has been identified in other Brazilian ecosystems in different habitats such as sand dunes, tropical forest, Cerrado, and agro-ecosystems and is also based on field-collected spores (Stürmer;Siqueira, 2006).Molecular data also indicates that species abundance of AMF in natural communities follows the log normal and broken stick model, where communities are dominated by a few species (Dumbrell et al., 2010).Aside from those species, others dominant species were Acaulospora mellea in both sites, Glomus sp. 4 in Soberbo, and Dentiscutata biornata and G. albida in Park (Table 3).Most species were classified as common in both sites, and rare species occurred mostly in Park (Table 3).
For S. elegans, AMF species richness was slightly higher in Park than in Soberbo.Simpson diversity and evenness were slightly higher in Soberbo for AMF communities associated with S. elegans, while both indices were higher in Park for L. chrysotrhis and Xyris sp.(Figure 3).Simpson dominance was higher in Park for S. elegans and in Soberbo for the other hosts.
Rupestrian fields are extremely fragile environments with low or no resilience to recover after disturbance (Le Stradic et al., 2014;Le Stradic;Buisson;Fernandes, 2014).Surveys of soil microorganisms such as diazotrophic bacteria (Marra et al., 2012) and arbuscular mycorrhizal fungi as reported here are the first step to understanding microbial diversity that provides ecosystem services important for the conservation and maintenance of Rupestrian fields.This work contributed to our knowledge of mycorrhizal status and AMF species richness associated with S. elegans, an important endemic and threatened species of Rupestrian fields.In Brazil, despite the large number of endemic plant species, there is a general lack of information concerning the association of these species with arbuscular mycorrhizal fungi.Rupestrian fields shelter 23% of AMF species described across the world and 41% of identified species in Brazilian ecosystems (Stürmer;Siqueira, 2008;Souza et al., 2010).Efforts to isolate and cultivate AMF species from Rupestrian fields and further tests on their effects on growth of S. elegans will contribute to the improvement commercial cultivation of this economically important plant and help preserve its population in the ecosystem.Classification according to the frequency of occurrence of AMF species: dominant (D) = Fi > 50%, most common (MC) = 30% < Fi ≤ 50%, common (C) = 10% < Fi ≤ 30% and rare (R) = Fi ≤ 10% (Zhang, Gui, Liu, 2004).

CONCLUSIONS
Based on the high levels of mycorrhizal colonization, Syngonanthus elegans is considered a mycotrophic host.Members of the main families and genera of Glomeromycota are associated with S. elegans in Rupestrian fields, although the identity of AMF species associated with this plant varied according to the area sampled.Trap cultures established using soils from Rupestrian fields represent an important strategy to complement AMF diversity assessment and help in the identification of field collected spores, as 30% of the total diversity was detected with this technique.

Figure 1 :
Figure 1: Families and genera of arbuscular mycorrhizal fungi associated with S. elegans at the Soberbo and Park sites.Proportion of AMF species per family (a), proportion of AMF species per genus (b).

Figure 2 :
Figure 2: AMF spores (in 50 g soil) recovered from soil samples near roots of S. elegans (n = 10), L. chrysothrix (n = 3), and Xyris sp.(n = 3) in the Soberbo and Park sites in Diamantina, MG. * = Differences between areas were significant at 5% as determined by an "F" test

Table 1 :
Soil characteristics from sampling sites Soberbo and Park in Diamantina, MG, Brazil.
1 P and K (extracted by Mehlich I).2 Exchangeable Ca, Mg and Al (extracted with KCl 1,0 mol L -1 ) and analyzed by titulation.

Table 2 :
Occurrence of AMF species detected in soil samples near roots and trap cultures (*) of S. elegans, L. chrysothrix and Xyris sp. at the Soberbo and Park sites in Diamantina, MG, Brazil.

Table 3 :
Number of spores (50 g soil) and frequency of occurrence (Fi) of AMF species from soil samples near roots of S. elegans in Soberbo and Park.