The Amazon region extends 8 million km² of intact tropical rainforest where the Amazon River and its tributaries transport approximately 20% of the Earth’s freshwater, from the Andes, crossing nine countries in South America all the way down to the Atlantic Ocean. The Amazon is estimated to comprise 50% of the remaining global tropical forests on our planet and is considered a biodiversity hotspot with endemic and endangered flora and fauna (Guayasamin et al., 2021). The Amazon rainforest can be divided in 4 main forest habitats: non-flooded terra firme forests which dominate the Amazon region, followed by large extensions of várzea which are seasonally flooded by white water rivers. Seasonally flooded forest next to black water rivers are called igapó, and natural areas with white sandy soils are known as campina, but both have smaller distributions (Ritter et al., 2019). Várzea forests are rich in eroded sediments resulting in more fertile soils as they are seasonally flooded from rivers that originate high up in the Andes, and typically host fewer tree species compared to non-flooded terra firme forests (Wittmann et al., 2006). The elevated terra firme forests have relatively well-drained soils, and their vegetation structure is commonly divided by strata, whereas várzea forests present a deeper layer of organic matter, with less clay content, and a higher capacity to capture soil moisture in comparison to terra firme sites (Ensley-Field, 2016). Although in the Amazon rainforest the nutrient cycle is incredibly fast, yet the bioavailability of essential nutrients like phosphorus (P) and nitrogen (N) is generally low. Although high temperature and precipitation in Neotropical forests promote leaf decomposition and subsequent nutrient release, over time, lixiviation will leach nutrients to deeper layers where plant roots encounter difficulties absorbing those nutrients.

To explore the concept of nutrient availability in tropical ecosystems, and to comprehend how the Amazon soils can sustain the highest plant diversity on the planet despite low P quantities, requires a thorough understanding of soil characteristics and biogeochemical processes, and the role microorganisms play within these processes (Ensley-Field, 2016). Tropical ecosystems like the Amazon provide the broadest ecological niche for the development of microbial interactions, and it is thought that plant-associated soil fungi play a key role in promoting and maintaining this high plant diversity in tropical ecosystems (Bennett et al., 2017; Schappe et al., 2017; Sheldrake et al., 2017). One of the most important terrestrial mutualistic relationships are associations between mycorrhizal fungi and approximately 85% of all vascular plants (Van Der Heijden et al., 2017; Brundrett and Tedersoo, 2018; Field and Pressel, 2018). Specifically, the arbuscular mycorrhizal fungi (AMF) are considered the most diverse and widespread fungal partners of land plants, belonging to a group of specialized fungi that colonize and inhabit the rhizosphere (Smith and Read, 2008; Strullu-Derrien et al., 2018). For more than 450 million years, the coevolution between terrestrial plants and AMF has provided an important selective advantage for both partners and is even thought to have facilitated the earliest plant colonization of terrestrial ecosystems (Smith and Smith, 2011; Tedersoo et al., 2017; Field and Pressel, 2018; Vasco-Palacios et al., 2018). AMF are obligate mutualistic fungi, with around 50,000 species known to form associations with over 250,000 vascular and non-vascular plants in all terrestrial ecosystems (Chen et al., 2018). This wide distribution points towards crucial roles in the maintenance and structure of global plant communities (Walder et al., 2012; Marinho et al., 2018).

AMF are aseptate filamentous fungi which hyphae enter the root, where they construct their mycelium in the intercellular apoplastic spaces, until they build intracellular fungal structures called “arbuscules”, which serves as an interface for the exchange of nutrients (Field and Pressel, 2018; Strullu-Derrien et al., 2018). The bidirectional exchange of fungal-acquired nutrients like P and N through the fungal periarbuscular membrane for plant-fixed carbon (C) is an integral part of the AMF mutualism (Schappe et al., 2017; Field and Pressel, 2018). Up to 20% of the plant photosynthates are transferred from the host plants to AMF, essential to complete their biological cycle (Tang et al., 2016). In return for carbon-based compounds, AMF transfer the equivalent of up to 90% of the phosphorus and 80% of the nitrogen requirements for plant growth and development (Smith and Read, 2008; Marinho et al., 2018). Although this exchange is considered a costly investment for the host plant, it is through this reciprocal exchange that both partners can meet their important physiological needs. Physical-chemical soil properties on the one hand, and edaphic factors like soil pH, water content, acidity, and soil aeration on the other hand determine AMF species richness and community structure (Wang et al., 1993; de Oliveira Freitas et al., 2014; Lugo and Pagano, 2019).

AMF perform a variety of ecosystem services, as their hyphal networks physically extend the plant’s rhizosphere, allowing plants to access a larger soil volume to promote the overall absorption of water and nutrients (Marinho et al., 2018). At a local level, AMF associations shape the plant community composition and overall plant productivity, promoting plant fitness and boosting reproductive capacities (Asghari and Cavagnaro, 2011; Cofré et al., 2019; Tedersoo et al., 2020). But mycorrhizal partnerships also play an important role in the global carbon cycle by augmenting soil C sequestration processes (Averill et al., 2014; Van Der Heijden et al., 2017; Field and Pressel, 2018).

Although the arbuscular mycorrhizae type is the most studied plant-fungus association, there has been a bias towards AMF research in temperate regions, with considerably less AMF research being conducted in the tropics (Cofré et al., 2019; Peña-Venegas and Vasco-Palacios, 2019). AMF communities in primary tropical forest ecosystems can be complex, and so are the factors that drive the community establishment. Moreover, the proportion of mycorrhizal plants increases towards the equator, which may reflect the arbuscular mycorrhizal ancestral state and tropical origin of major plant clades (Delavaux et al., 2019). To generate a better understanding of these aspects at an ecological level, there is a need to generate greater knowledge about this group at a taxonomic level and their specific ecological and evolutionary importance (de Oliveira Freitas et al., 2014).

For a long time, it has been a generally accepted dogma that plants in tropical rainforest ecosystems, with their exceptionally high turnover of nutrients and optimal climatic conditions around the equator, simply do not need to associate themselves with beneficial fungi. However, in recent years, there is growing evidence that AMF diversity in fact is high in tropical ecosystems (Peña-Venegas and Vasco-Palacios, 2019). As previously mentioned, many Neotropical rainforest soils in fact feature P deficiencies, and there might be an increased selective pressure on tropical plants to invest in the establishment of AMF associations. According to Marinho et al. (2018), tropical forests harbor 75% of all Glomeromycotina species known to date and might therefore be considered hotspots for AMF diversity. Furthermore, it has been proposed that the highest number of undescribed fungal species could be found in tropical biodiversity hot spots (Lugo and Pagano, 2019).

Given this extraordinary prediction about the AMF diversity in South America, there is a great potential to explore AMF in the Amazon rainforest. Although classified as one of the world’s important biodiversity hotspots regarding the macrobiodiversity, it could consequently be considered an important region to study the microbiodiversity that functionally supports the visible flora and fauna. This microbial biodiversity has been relatively well described for the Amazon regions of Colombia, Brazil, Venezuela, and Peru, whereas many microorganisms, including bacteria and fungi, but also insects are still largely unknown for the Amazon rainforest of Ecuador (Cañarte et al., 2020). Our research team recently published that the mycorrhizal diversity as mentioned in literature studies at the Ecuadorian Amazon is similar to other studies in the Amazon region but concluded that the number of undescribed mycorrhizal species in the tropical forests of Ecuador need more attention as patterns of occurrence and diversity remain largely unknown (Duchicela et al., 2022).

The present study investigated the fungal abundance and diversity in terra firme and várzea soils associated to the roots of tree seedlings of the genus Inga, an exclusively Neotropical genus belonging to the Fabaceae, a diverse family known to invest in associations with beneficial bacteria and fungi (de Oliveira Freitas et al., 2014; Antunes et al., 2019). Members of the genus Inga are omnipresent in both forest ecosystems within the study area at the Tiputini Biodiversity Station (TBS), a pristine and undisturbed tropical research area situated in the Yasuní Biosphere Reserve in the Orellana Province of Ecuador. We measured the roots and shoots of Inga seedlings and collected rhizosphere soil samples to measure soil physical and chemical properties, and to extract DNA to perform metabarcoding methods to assess what fungi, with a special focus on AMF, were present in the rhizosphere of Inga sp. Moreover, we were interested whether the abundance and composition of the Inga rhizosphere mycobiome differed between terra firme and várzea sampling sites, and what factors could influence these possible differences. Furthermore, we used microscopy to morphologically identify AMF spores to compare them with the metabarcoding results. Considering the differences in forest structure and abiotic factors within each ecosystem, we expected to find AMF to be present in both ecosystems, but the AMF community structure to differ in response to these conditions. Hence, in terra firme forests, where soils tend to be poorer in nutrients in comparison to várzea soils, a higher degree of AMF colonization is expected, and possibly a higher abundance and diversity of AMF in close proximity to the roots of Inga seedlings.

This study explored the composition of the mycobiome in the rhizosphere of Inga seedlings in two different but neighboring forest ecosystems in the undisturbed tropical Amazon rainforest of the Tiputini Biodiversity Station in Ecuador. In terra firme plots, which were situated higher up and therefore typically outside of the influence of river floods, and in várzea plots, the lower part of the forest located near the riverbanks and therefore seasonally flooded, tree seedlings of the genus Inga were randomly collected and measured, and we sampled the rhizosphere soils surrounding the root systems. We used an array of techniques, including physical-chemical soil analysis, metabarcoding using the molecular marker ITS to characterize the fungal rhizosphere community structure, and microscopy to morphologically identify AMF spores.

Within the TBS, members of the Fabaceae family and specifically the genus Inga demonstrate a high abundance, in particular in várzea habitats where 14.2% of all adult and subadult trees were identified as members of the genus Inga (personal comment Gonzalo Rivas-Torres). A factor that could explain the high abundance of this group might be their ability to form rhizospheric associations with both nitrogen fixing bacteria and AMF. These plant-associated microorganisms promote plant health by positively stimulating plant growth, boosting physiological responses towards biotic and abiotic stress factors, and supporting the plant’s adaptation to environmental circumstances (Mendes et al., 2011).

Temperature is an environmental factor that can have major effects on both plant growth and root exudation, thereby directly influencing AMF vesicle formation and colonization (Graham et al., 1982; Fitter et al., 2004). Furthermore, studies have shown that AM fungi can also directly be affected by temperature (Hawkes et al., 2008). This stimulating effect of elevated temperature on AMF growth and inoculation could explain why in tropical ecosystems, characterized by high temperatures, there is a greater abundance and diversity of AM fungi. It is therefore expected that these climatic conditions provide a selective force on AMF populations (Hawkes et al., 2008). The average temperatures as logged by the TBS weather station during the timeframe of our study are consistent with the normal average annual temperature (25.1°C) and within the range of common seasonal changes in temperature (personal communication TBS STAFF). As our study is a baseline mycobiome exploration study and should be considered as the first one of its kind in the Yasuní National Park area in Eastern Ecuador, the current study design does not allow to correlation prevailing temperature regimes and possible shifts in the belowground mycobiome composition. However, the period January-March 2020 was considerably drier compared to the same period in 2018 and 2019, which is directly correlated with the local flooding incidence. It has been reported that roots of plants in flooded soils present a lower AM fungi diversity (Deepika and Kothamasi, 2015). In tropical ecosystems like the current study site (TBS), this parameter could therefore function as an environmental filter, which might result in a reduced number of AMF species which can tolerate these prevailing habitat conditions, due to low prevailing oxygen levels and soil nutrient inputs during floods. but these conditions could put evolutionary pressure towards flood-tolerant AMF phylotypes, specifically evolved to survive extended periods with low oxygen contents in the soils during such a flood

As a bioindicator whether Inga seedlings in terra firme and várzea invest more of their resources aboveground or belowground, we measured the shoots and primary roots of randomly sampled Inga tree seedlings. The shoot:root ratio of Inga seedlings growing in terra firme plots had the tendency to present lower ratios compared to Inga seedlings developing in várzea plots ( Figure 4 ). Hence, Inga tree seedlings seemed to use relatively more of their acquired nutritional energy in the development of their root systems in terra firme, compared to várzea, where they invested more in vegetative growth. Although the tendencies of these phenotypic observations in Inga seedling development could be explained by different nutritional regimes or edaphic factors in both forest ecosystems or could be influenced by differences in the level of nutritional outsourcing provided by the establishment of beneficial associations with microorganisms, for instance by the residing fungal communities or specifically through intimate relationships with AMF. These differences as shown in Figure 2 were not significantly different but are mentioned as an interesting observation which requires further investigation.

In this study, macro and micronutrients concentrations were quantified for the rhizosphere samples collected from the root systems of Inga seedlings from terra firme and várzea forest sites to explore the soil nutritional status. As a result of this soil analysis, half of the assessed soil parameters did not present significant statistical differences between várzea and terra firme plots. This could be related to rapidly changing river geography within the Amazon Basin (Wittmann et al., 2006). Thus, areas that are considered drier parts, could become inundated in a short geological timeframe. Furthermore, a higher sample number could be a statistically stronger basis to detect potential differences, as small changes in soil composition are known to have strong impacts on soil microorganism composition, thereby exerting effects not only on AMF populations but even on the plant community structure aboveground. It has been previously recorded that a strong influence of edaphic factors exists on AMF formation and development in acid and nutrient-poor soils in Amazon forest ecosystems (Oliveira and Oliveira, 2010). Therefore, we will discuss both biological trends and significant differences as revealed by the physical and chemical property assessment of terra firme and várzea soils surrounding the roots of Inga individuals.

Results of the physical-chemical properties showed that várzea soils, as compared to terra firme soils, presented higher concentrations of all chemical elements except N, Mo, and V, which confirmed that periodic flooding of white-water rivers like the Tiputini river led to increased soil fertility of the várzea floodplains flanking these rivers. Furthermore, the relatively lower N levels in várzea forests could be a result of lixiviation, as nitrogen is a mobile compound and could easily be drained from the forest topsoil layers in várzea forests. Nitrogen is a major nutrient for plant development that frequently limits primary productivity in terrestrial ecosystems (Veresoglou et al., 2012). In highly productive systems which sustain high levels of plant biomass and are limited by nitrogen, such as those observed at the TBS, even small amounts of N gained by trees may give them a competitive advantage. Under these conditions, investing in natural mutualistic associations proves beneficial and could result in AMF recruitment by plants. AMF colonization, temporal and spatial abundance and AMF spore density are positively correlated with soil organic matter and total soluble N (He et al., 2002). AMF have the capacity to acquire N from both inorganic and organic sources, and partially transfer this N to their host plant (Hodge and Storer, 2015). Although overall AMF abundances did not significantly differ between terra firme and várzea soils, there were significant differences in the AMF composition between both forest ecosystems. Whether there is a correlation between the significantly lower N concentrations in várzea, and an AMF composition shift towards species which potentially can exchange higher amounts of N, is difficult to show due to the many unidentified ASVs as shown in Figure 4 . Furthermore, it would be very interesting to explore whether nitrogen-fixing bacteria show higher abundances in association to the roots of Inga in várzea soils. Cobalt (Co) has a positive effect in promoting both shoot and root biomass, particularly under low P levels. It is possible that high concentrations of this element in várzea forests could have led to a higher shoot to root ratio as observed in Figure 2 . In Neotropical forests, other elements like Al, Ca, Mg, K, Fe and Mn have been reported to affect colonization and spore numbers of AM fungi (Oliveira and Oliveira, 2010), which is in line with our metabarcoding results as these soil chemical elements (K, Al, Fe, and Cu) explained the highest amount of variation between the composition of the mycobiome and AMF communities of Inga in terra firme or várzea forest soils (see PCoA showed in Figure 6 ). Ca is known to be a vital element for the neutralization of soil or water acidity after floodings (Katutis and Rudzianskaite, 2015), but is also linked to both AMF richness and diversity (Guo et al., 2022). The beneficial effects of the AMF for host plants are directly related to the fungus’ ability to supply nutrients by extending the effective root absorption surface areas by incorporating its hyphal network. By doing this, the fungus can explore larger soil volumes and overcome nutrient and water depletion zones near the plant root (Clark and Zeto, 2000). Finally, pH levels in both ecosystems were observed to be acidic (pH range from 3.8 to 5), which could act as a selective force for AMF community structure. Winagraski et al., 2019 reached the conclusion that the important factor shaping the AMF community structure are environmental factors rather host-plant characteristics.

As showed in Supplementary Table 2 , the richness and homogeneity of each Inga rhizosphere sample suggested that there was no difference between either the terra firme and várzea forest ecosystems, nor between the two sampling campaigns. According to Ritter et al. (2020), this is because different areas in the Amazon can present similar levels of species richness but could display different community compositions. There was the exception for homogeneity and richness, measured through the Evenness and Shannon index, respectively, for the sampling period of the ASVs only from the Fungi kingdom, where November 2018 presented greater diversity and richness than in August 2019. The soils of occasionally flooded forests like várzea tend to have a slightly more acidic pH (Randall, 2017). The slight differences in pH levels of terra firme (4.92) and várzea soils (4.65) as encountered in this study, could be responsible for different fungal communities ( Supplementary Table 1 ). Analysing the β-diversity of the ITS sequences, both ecosystems clustered differently when analysing the Jaccard, Weighted and Unweighted UniFrac indices. Although fungi are generally less sensitive to changes in pH and have a wide range of pH in which they grow without inhibition of their development, it can be said that the pH and other conditions influence the community structure. Rousk et al. (2010) published that the rhizosphere provides a buffer environment in which changes in the soil environment did not result in variations in the fungal community associated with the rhizosphere. This seemed to be the case in our study, as the results of the beta diversity analysis on Glomeromycotina ASVs showed similar AMF abundances in both ecosystems.

Within the Fungi kingdom, several phyla were identified ( Figure 3 ), with the phylum Ascomycota as the most abundant, then followed by Basidiomycota, Rozellomycota, Glomeromycota (now subphylum Glomeromycotina), Zygomycota, and Chytridiomycota, which were present in smaller proportions, in accordance with Zhao et al. (2020). They indicated that Ascomycota portray a faster evolutionary rate and a greater overall diversity of species, which is why they manage to adapt faster to different environments, and they successfully survive in the litter or humus of both terra firme and várzea soils (Richardson et al., 2001). A fungal phylum which was present in higher abundances in the terra firme samples compared to the várzea samples were members of the Blastocladiomycota. They live in muddy soils where they portray a saprotrophic lifestyle, decomposing plant and animal organic matter, or parasitize arthropods (Money, 2016). It is interesting to observe this “aquatic” fungi to be more present in terra firme soils, as based in their lifestyle you might expect them to be more present in seasonally flooded várzea soils. Moreover, members belonging to the Glomeromycotina (former Phylum Glomeromycota) were present in small proportions, and included 13 families, including Glomeraceae, Gigasporaceae, Acaulosporaceae, Diversisporaceae, and 9 previously unidentified families. The Glomeraceae family was the predominant family with a relative abundance of 72.23% of the residing AMF community. Similar trends have been found by Peña-Venegas and Vasco-Palacios (2019), which identified Glomeraceae and Acaulosporaceae as the most common arbuscular mycorrhizal fungi in Amazonian soils due to their associations with species like the members of the genus Inga, and as they thrive in warm and humid conditions. As one of the few other mycobiome studies in the Ecuadorian Amazon, which explored the AMF diversity associated to native plants growing in petroleum-polluted sites, Garcés-Ruiz et al., 2019 discovered that Acaulosporaceae dominated the AMF community composition in these oil pits. We show contrasting results for undisturbed sites at the TBS, as we identified members of this family to represent a smaller proportion of the AMF mycobiome. We encourage further studies to confirm whether this composition shift is due to the oil contamination, or for instance the difference in host plants in both studies. The higher presence of species of Glomerceae species which have been recorded to present higher root hyphal densities compared to for example Acaulosporaceae species, could be due to their capacity to provide profound P benefits to their host plants like Inga (Deepika and Kothamasi, 2015). Acaulosporaceae are considered stress tolerant and have been reported in acidic soils and under harsh environmental conditions, whereas Gigasporaceae have a slower growth and longer colonization period, once established, they provide nutrients in higher amounts than Glomeraceae or Acaulosporaceae (Goss et al., 2017). Although it requires some expertise to be able to identify AMF spores to the species level, the results of the AMF spore morphological identification using microscopy encountered AMF spores from all three families: Glomeraceae, Acaulosporaceae, and Gigasporacea. This confirmed the outcome of the molecular AMF characterization and strengthened the importance of these groups of mycorrhizae-forming fungi in Neotropical forest soils. One out of four ASVs was annotated to 9 AMF families of unidentified origin, which confirms that the percentage of molecularly identified arbuscular mycorrhizae in the rhizosphere of plants is still very limited and supports the idea to promote to archive more AMF sequences to public databases (Buee et al., 2009; Garcés-Ruiz et al., 2019). Interestingly, results of recent studies in Brazilian forest ecosystems indicate that some arbuscular mycorrhizal fungi are considered omnipresent, whereas others are relatively rare to find (Winagraski et al., 2019). When Fungal ASVs were molecularly analyzed using GNEISS metrics, it was identified that the várzea ecosystem presented a slightly higher Sordariomycetes proportion, while the terra firme ecosystem revealed higher incidences of the Scedosporium genus and two taxa of the Nectriaceae family. Furthermore, various taxa of the Penicillium genus were present in higher proportions in samples collected in August 2019, compared to samples of November 2018.

In conclusion, the present study describes for the first time the mycobiome associated to the rhizosphere of seedlings of the genus Inga in the Tiputini Biodiversity Station in the Ecuadorian Amazon. The rhizosphere of Inga seedlings is surrounded by AMF in relatively lower but similar proportions compared to other fungal phyla, but AMF were found in both terra firme and várzea ecosystems and in both sampling seasons. Understanding and evaluating the spatio-temporal dynamics of the root mycobiome, with a specific focus on the role of AMF, could be crucial to understand how undisturbed primary tropical rainforest ecosystems function, and how they influence the underlying mechanisms.

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.ebi.ac.uk/ena,PRJEB57077.

The study was designed by AH-D, PH, and JD. Research permits were managed by SZ and PH. Samples were taken by AH-D, PH, and JD. AH-D and PH prepared the samples for physical chemical soil analysis and amplicon sequencing. VA-G and BP-V conducted the bioinformatics analyses. JD prepared the soil, root, and spore samples and executed the mycorrhizal potential experiments. PH and JD complemented the analyses with their specific expertise and interpreted the overall data. AH-D, PH, JD, VA-G, and BP-V visualized the results and wrote the manuscript. All authors read and approved the final version of the manuscript.