Ectomycorrhiza (ECM) is an intimate symbiotic association between plant roots and ECM fungi. ECM associations benefit most terrestrial plants by enhancing nutrient uptake and tolerance to environmental stressors (Tedersoo et al., 2020). ECM fungal community structure is affected by a multitude of biotic and abiotic factors (Ishida et al., 2007). The community composition is shown to change across gradients of soil pH (Bahram et al., 2011), soil nitrogen and organic matter (Pellitier et al., 2021), elevation (Matsuoka et al., 2016), and forest age (Hui et al., 2017; Boeraeve et al., 2018). As biotic and abiotic components usually show strong interdependencies, it is methodologically difficult to assess the contributions made by each specific factor in the ECM fungal community and their response to environmental changes (Weißbecker et al., 2018).

The effect of different driving factors at different scales is variable. Physiographic factors such as altitude and soil properties including soil microorganisms and soil physicochemical properties are usually important drivers of ECM fungal communities in the local-scale sampling area, such as carbon content and pH (Vasco-Palacios et al., 2020), altitude and forest age (Matsuoka et al., 2016; Geml et al., 2017; Schön et al., 2018) and host genetic and phenotypic characteristics (Rúa, 2021). In large-scale ECM fungal studies, the role of host plants is becoming increasingly obvious. Host dispersal can affect the biogeographic patterns of the ECM fungal community (Matsuoka et al., 2019), and a significant proportion of ECM fungi exhibited host specificity in conifer broadleaf forests (Ishida et al., 2007). In addition, ECM fungal communities show a stronger correlation with host composition than soil properties, while pathogens and arbuscular mycorrhizal (AM) fungi are more strongly correlated with soil properties (Schappe et al., 2020). Hence, whether the range of host habitats is the main factor for ECM fungal community structure is worth further exploring.

The abundant-center hypothesis posits that species abundance peaks in the center of its distributional range and declines toward its edges (Hengeveld and Haeck, 1982). The high abundance in the center is because of the optimal conditions (i.e., the presence of suitable habitats), while the decline in abundance at edges is because of environmental suboptimality (Lira-Noriega and Manthey, 2014). A recent modeling study showed that the abundant-center relationships were strongly affected by populations and deterministic growth environment (Dallas and Santini, 2020). This theory provides an overview of species distributions. However, studies on abundant-center hypothesis have mainly focused on animals and plants (Pironon et al., 2017), while studies on ECM fungal communities are rare.

Unlike the animal and plant species, ECM fungi are typical symbionts and more than 90% of nutrient-absorbing roots of trees are colonized by ECM fungi (Smith and Smith, 2002). ECM fungi are estimated to comprise 20,000–25,000 species globally and form symbioses with only a tiny fraction of the terrestrial plants (2% of total plant species) (Tedersoo et al., 2012). In addition, most ECM fungi reproduce sexually and produce macroscopic fruit bodies, which differ from arbuscular and ericoid mycorrhizas (Tedersoo et al., 2010). Because these fruit bodies deposit most of their spores within in a short distance, the spatial distribution of the ECM community is probably more dependent on the host plants’ distribution than environmental factors. Hence, the distribution models between ECM fungi and host plants may be the same, or at least very similar. Most studies have been performed on the relationship between host plant species composition and ECM fungal community (Durall et al., 2006; Mucha et al., 2018), while fewer studies have addressed the effects of host plants’ abundance. The distribution of some Pinaceae species, which is certainly the oldest extant plant family symbiotic with ECM fungi (Hibbett and Matheny, 2009), have found evidence for the abundant-center hypothesis (Mimura and Aitken, 2007; Gugerli et al., 2009). Previous studies have found that ECM communities are more concentrated on species of Pinaceae and showed stronger host effects than on broadleaf species (Ishida et al., 2007; Vlk et al., 2020). However, for ECM fungal distribution modeling, it appears important to consider not only the distribution of individual species but also the ECM fungal richness and biodiversity of the whole community. Many ECM fungi may fulfill similar ecological functions and that degree of functional redundancy exists in ECM fungal communities (Dahlberg, 2001). ECM fungi are usually categorized into two major groups based on their distribution features: cosmopolitan ECM members (Cortinarius, Russula and Tomentella) and endemic ECM members (McPolin and Kranabetter, 2021). Endemic ECM fungi usually exhibit higher host specialization than cosmopolitan ECM members, which is generally driven by the host range (McPolin and Kranabetter, 2021). Picea crassifolia is one of the main tree species that belongs to the Pinaceae family in semi-arid areas of China. It forms pure stands on north-facing slopes along the northeast edge of the Qinghai-Tibetan Plateau (QTP) region and in the adjacent highlands of Helan Mountain, which has not been subject to the significant human disturbance for more than one century. The ECM fungal community associated with a single host in the pure forest of P. crassifolia was considered an excellent material for ecological studies.

In this study, we examined the ECM fungal community associated with P. crassifolia to test if the distribution of ECM fungal species fit the abundant-center hypothesis model. We set out two hypotheses: (1) Is the composition of ECM fungal communities associated with P. crassifolia generally consistent across the host’s geographical range? (2) Does the richness and diversity of these ECM fungi decrease from the center to the edge of the host’s distribution?

Next-generation sequencing technologies have enabled large-scale analyses of the complex and normally species-rich ECM communities, showing high consistency with the clone-based Sanger sequencing (Kauserud et al., 2012). High-throughput sequencing technology has greatly facilitated the acquisition of genome data (Li et al., 2020), which allows us to investigate the ecophysiology of uncultured microbes (Wagner, 2009). In our study, the number of ECM species (196 OTUs) detected that were associated with P. crassifolia from the region of Inner Mongolia (NM group) is much higher than 11 OTUs at the same site (Fan et al., 2016). A total of 10 ECM species (all except 1) were found using both methods (Table 4).

Our results clearly showed Tomentella was the most species-rich genus ahead of Cortinarius, Inocybe, and Sebacina. However, none of Tomentella species was detected using Sanger sequencing (Fan et al., 2016). Thus, relative to Sanger sequencing, high-throughput sequencing can capture a greater proportion of species, which is considered to be more effective in the field of community ecology (Paul et al., 2018). We found that the most species-rich ECM fungal genera were still the most common one in many other kinds of forest trees (Tedersoo et al., 2003; Richard et al., 2005; Bergemann and Garbelotto, 2006). These genera are often considered to be widespread, forming symbiotic relationships with more host plants than other ECM genera (Halling, 1987; Frøslev et al., 2005; Jakucs and Eros-Honti, 2008).

A number of studies have reported that the ECM fungal community was significantly affected by the host, soil, atmospheric deposition, geographical distance and climate (Ishida et al., 2007; Thompson et al., 2017; van der Linde et al., 2018; Wang et al., 2019). In addition to soil characteristics or even the disturbance regime of the system, the ECM fungal community structure and diversity could be determined by host trees (Johnson et al., 2005). The host age, photosynthetic activity, distributions, forest management and species composition affected the diversity or community structure of the ECM fungi (Goldmann et al., 2015; Erlandson et al., 2016; Hui et al., 2017; Unuk et al., 2019; Matsuoka et al., 2020). To explore the effects of abiotic factors, the most efficient way to eliminate the host effect is to focus on only one distributed worldwide host species such as Pinus tabulaeformis, Quercus liaotungensis, P. crassifolia, and Populus simonii. Consistent with our previous work (Long et al., 2016; Han et al., 2017; Wang et al., 2017), we found that site was the overriding factor affecting ECM fungal community structure compared with other factors at large scales (R² = 0.37, P = 0.009, PERMANOVA). Most species of ECM fungi are thought to have limited distributions, which are commonly shaped by those of their hosts (Tedersoo et al., 2020). In general, the dominant ECM fungal genus associated with any specific host species is relatively consistent across the plant’s range. For instance, the dominant ECM fungal genus associated with Coccoloba uvifera was Scleroderma on a global scale (Séne et al., 2018). Tomentella, Inocybe, Clavulina, and Russula were the dominant ECM fungal genera in the Alnus trees (Roy et al., 2013). Consistent with our study, we found that Inocybe, Sebacian, and Tomentella were the dominant ECM fungal genera across the whole distribution area, accounting for more than 80% of the total relative abundance (Figure 2).

Although ECM fungal communities were dissimilar between sites (R = 0.33; P = 0.006, Anosim), the composition of the ECM fungal community was generally consistent among the four study sites. The number of overlapping species (137) far exceeds the number of unique species in each area (Figure 4B). We suggest that host identity should be the major factor that determines ECM fungal community composition. However, this finding is different from the study of the host plant Castanopsis sieboldii (Matsuoka et al., 2019), which suggested the explanatory power of the host was less than for climatic filtering and/or fungal dispersal. The cause of this difference could be that seawater blocked the transmission of ECM fungal spores among the whole hosts’ distribution. Furthermore, the LEfSe analysis identified 16 indicator taxa for differences among ECM fungal communities associated with P. crassifolia trees (Figure 5). The differences in ECM fungal community structure seem to be largely responsible for the shifts in the abundances of rare ECM fungal genera (Hui et al., 2017). Weißbecker et al. (2018) found that ECM fungal communities showed the greatest overlap of two sampled communities in a fungal functional group (approximately 80% of ECM fungal community similarity), although ECM fungal communities showed the highest pairwise community dissimilarities.

The distribution models consider the factors beyond consumable resources, such as space, to play a significant role in determining community structure (VanDerWal et al., 2009; McPolin and Kranabetter, 2021). The biogeographic position of each population on the mainland is typically classified as edge, sub edge or core (De Kort et al., 2021). QH is the core distribution of P. crassifolia, while other sites are at the edge in this study. ECM fungal richness decreases with the distances increasing from the core to the edge (Figure 3B), which agrees with the abundant-center hypothesis. Our model is similar to the non-linear model for endemic ECM species about ECM fungal community dissimilarity (based on species incidence) and interplot distance (McPolin and Kranabetter, 2021). However, unlike the pattern of ECM fungal richness distribution, the relative abundances of individual ECM fungal genera are difficult to link directly to the abundant-center hypothesis (Table 3). Mixed support for the abundant-center hypothesis has also been found in population genetic studies of other plants (e.g., Dixon et al., 2013). Furthermore, the center site (QH) has the most endemic ECM fungi (Figure 4B) and different biomarkers (Figure 5). The reason for this could be the changes in the relative abundance of dominant ECM fungi might be influenced by a specific environmental factor (Wakelin et al., 2007). For instance, nitrogen enrichment can lead to an increase in the biomass and abundance of the genus Tomentella (Brearley et al., 2005). In contrast, species in the genus Sebacina were usually dominant in the high-phosphorus forest (Zavišić et al., 2016). In addition, the dominant genera usually belong to the cosmopolitan group, which displayed lower species turnover across the host distribution (McPolin and Kranabetter, 2021). However, most of the low relative abundance of ECM fungi, belonging to the endemic ECM group, was generally driven by environmental history as well as biological traits (Hobohm and Tucker, 2014). At a continental scale, the distributions of these ECM fungal genera may reflect dispersal and colonization limitations, as well as greater species turnover at a local scale (McPolin and Kranabetter, 2021). These forces could produce ECM fungal distributions that conform to the predictions of the abundant-center hypothesis.

A species would achieve higher growth rates or greater stability at the core of a host species’ fundamental ecological niche than at sites farther from these optimal conditions. Thus, at core sites, hosts may have the highest diversity of ECM fungi as host effects play a critical role in the ECM fungal community assemblage and taxonomic diversity. Our results show that ECM fungal species richness was greatest at the center of the host plant’s range, consistent with the abundant-center hypothesis, while the relative abundances of individual fungal genera shifted inconsistently across the plant’s range. We encourage other researchers to test this hypothesis on other ECM fungal host species.

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://ddbj.nig.ac.jp, LC203765-LC205656.

QH and XW carried out the experiments. QH wrote the manuscript. Both authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.