Differential soil fungus accumulation and density dependence of trees in a subtropical forest

See allHide authors and affiliations

Science  04 Oct 2019:
Vol. 366, Issue 6461, pp. 124-128
DOI: 10.1126/science.aau1361

Fungal influence on density dependence

Tree species in highly diverse tropical forests tend to exhibit conspecific negative density dependence, a phenomenon whereby individuals of the same species tend to grow at a distance from one another. This is understood to be a key driver of species coexistence. The strength of negative density dependence varies between species, but the mechanisms driving this variation are unknown. Chen et al. studied tree species in a subtropical forest in China and found an important role for soil-dwelling fungi in this variation. Elevated accumulation of pathogenic fungi leads to stronger negative density dependence, whereas elevated accumulation of mutualistic fungi leads to weaker negative density dependence.

Science, this issue p. 124


The mechanisms underlying interspecific variation in conspecific negative density dependence (CNDD) are poorly understood. Using a multilevel modeling approach, we combined long-term seedling demographic data from a subtropical forest plot with soil fungal community data by means of DNA sequencing to address the feedback of various guilds of soil fungi on the density dependence of trees. We show that mycorrhizal type mediates tree neighborhood interactions at the community level, and much of the interspecific variation in CNDD is explained by how tree species differ in their fungal density accumulation rates as they grow. Species with higher accumulation rates of pathogenic fungi suffered more from CNDD, whereas species with lower CNDD had higher accumulation rates of ectomycorrhizal fungi, suggesting that mutualistic and pathogenic fungi play important but opposing roles.

Conspecific negative density dependence (CNDD) is thought to explain species coexistence in diverse tree communities (1, 2). Coexistence theories suggest that specialist natural enemies accumulate near dense patches of their hosts and attack seeds and seedlings of the same species, ultimately lending an advantage to locally rare species (3, 4). Experimental applications of fungi (5, 6) or fungicides and soil sterilization (7, 8) and reciprocal transplants (9, 10) provide compelling evidence for the diversity-promoting effects of pathogens on local tree neighborhoods. Tree species vary greatly in how they respond to conspecific density within a community (11, 12), yet few studies have focused on the underlying mechanisms that lead to such variation among species in CNDD and have inspired much debate. Key studies of correlations between the strength of CNDD and species abundance in tropical forests suggest that at finer spatial scales (<10 to 20 m), species that are more abundant actually suffer less from CNDD than rarer species do (i.e., a rare-species disadvantage) (9, 12).

Soil fungi are one of the most diverse groups of microorganisms in the world and can interact intimately with plants as pathogens, symbiotic mutualists, and decomposers (13). The accumulation of fungal pathogens could promote and maintain tree community diversity by reducing the recruitment and survival of dominant species (14), whereas fungal mutualists can benefit plants by offering nutrients and protection from pathogens, resulting in monodominance through positive plant-soil feedback (15). Saprotrophic fungi are major agents of litter decomposition and nutrient cycling and might also have an effect on species coexistence through altering nutrient availability (16, 17). Recent work investigating the strength of negative density dependence in North American trees has found that tree mycorrhizal association could explain plant responses to their conspecific neighbors, where ectomycorrhizal (EcM) species experience weaker CNDD than do arbuscular mycorrhizal (AM) species, potentially because, compared with AM fungi, EcM fungi act as superior protectors from antagonists (18). Thus, a current research challenge is to understand how different functional types of fungi mediate conspecific neighbor effects and whether feedback with soil fungi contributes to a better mechanistic understanding of variation among tree species in their patterns of CNDD.

Here, we present a study combining long-term seedling survival measurements and an assay of the fungal community to address the feedback of various guilds of soil fungi on the CNDD of tree species in a 24-ha Gutianshan subtropical forest dynamics plot. Species in the plot can be classified into three different groups: AM, EcM, and ericoid mycorrhizal (ErM) plants, and more than half (56.4%) of the total basal area was contributed by EcM trees. We test two hypotheses with respect to tree-fungus associations and their roles in driving interspecific variation in the strength of CNDD. First, we hypothesize that species that accumulate pathogenic fungi through tree ontogeny more quickly are likely to suffer more negative demographic impacts when conspecific densities are higher. Second, we hypothesize that the variation in functional types of soil fungi (pathogenic, AM, EcM, and saprotrophic) accumulated through tree ontogeny is itself related to variation among species in the strength of CNDD.

To test these hypotheses, we first collected soil from the rooting zone of 322 individual trees from 34 species (23 AM, 8 EcM, and 3 ErM) once at the end of the seedling census in the same plot. High-throughput DNA sequencing was used to measure the soil fungal taxa associated with each focal tree. Each taxon was identified as a pathogenic, AM, EcM, or saprotrophic fungi on the basis of the most updated knowledge of functional annotations of fungal genera (19). We quantified the accumulation rate of fungal species in each tree species by characterizing the abundance and richness of each fungal guild of conspecific individuals in different tree size classes {i.e., using individuals with varying diameters at 1.3 m above ground [diameter at breast height (DBH)]} (19). We used this approach because larger trees within a species are generally older than smaller individuals in environments that are relatively homogeneous spatially and temporally (20) and because our study organisms are long lived, making repeat sampling through the life span of an individual not feasible. Second, we used a neighborhood modeling approach (21, 22) to measure conspecific neighborhood effects on seedlings from natural seed-fall recruitment in nine censuses from 2006 to 2014, following a standard census protocol (23). For 28 of the 34 plant species that had a sufficient number of seedlings, two-level mixed models were then used to test whether species-level variation in density-dependent seedling survival can be predicted by the rate at which each tree species accumulated various guilds of soil fungi. Lastly, we also tested whether fungi were host specific by tallying the number of tree species on which each taxon was found, as host specificity for soil fungi in plant rhizospheres was an essential component in generating differential fungal effects on plants.

We used ITS1 and 18S barcoding primers to sequence the total fungal community and AM fungal community, respectively, at a 97% sequence similarity level. Clustering at a cutoff sequence similarity of 97% is often considered a compromise between natural intraspecific and interspecific sequence variation and random sequencing errors (24). We obtained 6261 operational taxonomic units (OTUs) affiliated with the ITS1 region, of which 70.7% could be assigned to different fungal functional groups. The largest group was saprotrophs, which accounted for 2705 (43.2%) of the taxa, followed by EcM fungi (1304; 20.8%) and then plant pathogens (242; 3.9%) (fig. S1). However, EcM fungi contributed to the largest proportion of sequences (48.4%), reflecting the dominance of EcM tree species in the plot. For the 18S primer system, the AM fungal dataset resulted in 141 OTUs after excluding the non-AM fungal OTUs, ranging from 21 to 65 per sample. These OTUs covered all four AM fungal orders (27.0% to Archaeosporales, 12.1% to Diversisporales, 54.6% to Glomerales, and 5.0% to Paraglomerales), and Glomerales contributed 85.0% of sequences (fig. S2). Compared with tropical forests in northern South America (25), the proportion of EcM fungal taxa was much higher (20.8% versus 8.0%), and plant pathogens were relatively less abundant and diverse than in lowland AM-dominated tropical forests.

Most fungal taxa were associated with several tree hosts, and only 609 of the taxa detected by using ITS1 sequencing and 18 of the taxa detected by using 18S sequencing were found from a single host species (fig. S3). Of the major functional groups, pathogens were the most specialized guild on the basis of their having the highest mean species specialization indices, followed by saprotrophs, EcM, and then AM fungi (fig. S4). Similarity in a fungal community between any pair of hosts was negatively but weakly correlated with phylogenetic distance (Mantel test, P > 0.1). Adonis permutation analysis revealed that host phylogeny explained only a small but statistically significant portion of the variation (<2.7%) in fungal communities after accounting for the effect of the host species (<3.6%) (table S1). However, it remains possible that the impacts of plant-fungus interactions will follow a strong phylogenetic signal, as revealed by inoculation tests (26, 27).

A wide array of antagonists is thought to accumulate in both diversity and density over tree ontogeny and can help inhibit the establishment of conspecifics near larger adults (1). We quantified the fungus accumulation rate by examining how the sequences and OTU richness of different fungal guilds affiliated with focal trees change with focal tree size (Fig. 1). In an average species, larger trees had a significantly higher proportion of pathogens and saprotrophs in both the number of sequences and the number of OTUs. Conversely, AM fungal OTU richness and the proportion of EcM fungi in sequences and OTUs decreased with focal tree size. As size class increased, there was significant variation among tree species in the rate at which the pathogens, EcM fungi, and saprotrophs accumulated (Fig. 1, A and B). Although AM fungi appear ubiquitous, the effect of tree size on AM fungal OTU richness varied little among species, likely because of their low host specificity (Fig. 1C). Tree species with lower EcM fungus accumulation rates of sequences generally had higher saprotroph or pathogen accumulation rates (table S2). Pathogen accumulation rate also depended on mycorrhizal association of tree species. For example, AM tree species accumulated pathogen sequences faster through ontogeny than did EcM tree species (fig. S5), implying greater pathogen pressure under an adult AM tree than under an adult EcM tree.

Fig. 1 Tree size effects on the abundance and richness of each fungal functional group.

The GLMM fits the log odds of (A) fungal sequence and (B) OTU richness for plant pathogens, EcM fungi, and saprotrophs by a linear function of log-transformed tree DBH, whereas a linear mixed model fits the tree size effects on (C) AM fungal OTU richness. Overall curves (solid black lines) are all significant compared with no relationship at significance level α = 0.05, and dashed black lines show 95% bootstrapped credible intervals. Gray lines represent the predicted species-specific responses if tree size effects varied significantly across species (LRT test, P < 0.05). There were 319 individual trees from 34 species included in these models.

To evaluate the strength of CNDD, we modeled annual seedling survival from 2006 to 2014 as a function of the density and identity of neighbors using generalized linear mixed models (GLMMs) for all species combined and for seedlings belonging to different mycorrhizal types. At the community level, we found distinct effects between conspecific and heterospecific neighbors. Seedling survival significantly declined with the density of conspecific seedling and adult neighbors (Fig. 2A), confirming the prevalence of CNDD (28). However, both heterospecific seedling and conspecific sapling neighbors had significant positive effects on seedling survival (Fig. 2, A and B). These are in line with previous findings that seedling survival was enhanced by increased heterospecific crowding as a result of fewer encounters with species-specific natural enemies (i.e., species herd protection) (22, 29), whereas the positive seedling response to conspecific saplings reflects higher seedling survival in preferred habitats (2).

Fig. 2 Neighborhood effects (odds ratios ± SE) on seedling survival for all, AM, and EcM tree species.

(A to D) Seedling density in a 1-m2 seedling plot and sapling and adult tree basal area in a 10-m radius were included in GLMMs, and each of them were standardized before entering the model. Odds ratios >1 indicate a positive effect on seedling survival, whereas values <1 indicate a negative effect. Estimated coefficients and additional results for ErM species are presented in table S3.

Turning to the fungal guild level, seedling survival responses to the neighborhood variables varied considerably among mycorrhizal types, and density-dependent mortality was most common for AM seedlings. Compared with EcM and ErM species, AM seedlings near other conspecific seedlings and adult neighbors were more negatively affected than those near heterospecific seedling and conspecific sapling neighbors (Fig. 2, A and B, and table S3). By dividing heterospecific neighbors into AM and EcM species, AM seedlings benefited (i.e., higher survival rates) from the presence of both heterospecific AM and EcM seedling neighbors, and EcM seedling survival was significantly enhanced by the density of heterospecific AM seedlings and EcM adults (Fig. 2, C and D), supporting the safe-site hypothesis (30). That is, heterospecific neighbors may improve survival by offering AM and EcM tree seedlings a site safe from natural enemies, with the exception of ErM species (table S3). A recent study of spatial associations of AM and EcM trees predicted that opposing mycorrhizal associations had significant inhibition (31). However, we only found that EcM seedling survival was significantly inhibited by the occurrence of AM saplings and adults in the vicinity, whereas AM seedling survival was largely enhanced by EcM neighbors (Fig. 2). Our results, therefore, implicate mycorrhizal association of species as important drivers of tree neighborhood interactions at the community level (18, 31, 32).

For 28 species, sufficient seedlings were available to evaluate the contributions of various guilds of soil fungi accumulated through tree ontogeny to interspecific variation in CNDD. Seedling response to neighbors varied widely among species [likelihood ratio test (LRT) for the random effect of neighborhood × species interaction, P < 0.001]. The strength of the negative impact of conspecific seedling neighbors on survival was negatively correlated with the rate at which trees accumulated pathogen sequences as they grew, but it was positively correlated with the rate at which EcM fungi accumulated (Fig. 3, A and B). The effect of neighboring conspecific adults, however, correlated only with EcM sequence accumulation, not pathogen accumulation (Fig. 3, C and D). These results suggest that exposure to EcM fungi protects seedlings from pathogen damage (33, 34) and that mutualists, as well as pathogens, are important players in driving interspecific variation in CNDD. Additionally, EcM and saprotrophic fungus accumulation rates were significantly related to interspecific variation in conspecific positive sapling effects (table S4). Most species experiencing significant positive conspecific sapling effects (fig. S6) have strong habitat associations, as revealed by a previous study in this same forest (33), and soil fungi may play an essential role in nutrient exploitation during early seedling development (16, 17). However, consistent with a recent study (18), we found that mycorrhizal type explained little of the variation among species in the strength of density dependence (table S5).

Fig. 3 Relationships between CNDD and pathogenic and EcM fungus accumulation rates over tree ontogeny.

(A to D) The overall relationships fitted by two-level mixed models suggest that increasing pathogenic fungal density caused lower seedling survival, but increased EcM fungal density favored seedling survival by means of density-dependent effects. AM tree species are in green, EcM tree species are in red, and ErM tree species are in blue (n = 28 tree species). The black solid and dashed lines indicate that the fitted regressions were significant at P < 0.05 or not statistically significant, respectively. Parameter estimates are available in table S4.

The presence of EcM fungi in the soil and/or on the root tips of EcM plants with a fungal mantle may be considered important for only the EcM plants in a neighborhood. However, we found that both AM and EcM tree seedlings benefit from the accumulation of EcM fungi and that the accumulation rates of EcM fungi in some AM tree species over ontogeny are even higher than those of EcM species (Fig. 3 and fig. S6). The accumulation of EcM fungi in AM tree species through ontogeny (e.g., Acer cordatum, Eurya muricata, and Loropetalum chinensis) may arise because of an overlapping spatial distribution with EcM species (fig. S7). Although AM species cannot directly affect the EcM fungal composition in the soil, they may alter soil conditions [physical, (bio)chemical, or biological], which may indirectly influence the EcM fungal community. Recent empirical studies provide strong evidence that an EcM fungal community could be indirectly affected by non-EM plants (34, 35), and our results suggest that, if such impacts occur, they may decrease the impacts of CNDD on non-EcM tree species. However, more experimental tests using controlled inoculations would be required to test this possibility and to further elucidate the general role of changes in the soil fungal composition on plant communities (36).

In AM-dominant tropical forests, where many of the concepts of CNDD were originally developed, previous work has indicated that common tree species suffered less from CNDD than rare species, potentially because of a higher resistance to pathogens (9, 12). Nonetheless, we found that tree species abundance was not a good predictor of neighborhood effects (i.e., CNDD) (table S4). This discrepancy may arise because AM fungi offer less protection from antagonists than EcM fungi do (37). Trees may maintain an extramatrical mycelial network from which EcM fungi could decrease pathogen colonization and damage to seedlings of rare species in EcM-dominant forests. Our findings that heterospecific adult EcM trees enhanced EcM seedling survival and that AM seedlings responded positively, or less negatively, to adjacent EcM neighbors provide further evidence supporting this inference (Fig. 2). Our work is congruent with the results of study on the strength of CNDD at the global scale, which shows that the CNDD-species abundance relationship varies systematically among sites, from positive in the tropics to neutral (or negative) in subtropical and temperate forests (38). Combined, these results provide an extra dimension to the Janzen-Connell hypothesis that pathogen accumulation rates may play a key role in driving the strength of CNDD, but EcM fungi may overrule them (39). We speculate that locating communities along the continuum from pathogen driven to EcM driven could help elucidate the mechanism underlying the latitudinal gradients in CNDD and its relationship with species abundance observed in natural forests (38, 40).

The negative response of seeds and seedlings to conspecific density is widely observed in forest trees (28) and likely plays a major role in maintaining diversity patterns at local community and landscape scales (14, 23, 38, 40). However, until now, studies have not examined whether species’ differential responses to soil fungi can cause variation among tree species in the magnitude of CNDD and rarely have examined the roles of different functional groups simultaneously, despite increasing recognition of the contrasting roles of mutualistic and pathogenic fungi (14, 41, 42). Our community-level analyses, relying on standardized long-term seedling census datasets and sequenced natural populations of fungal functional categories, represent a step toward a better understanding of the feedback of various guilds of soil-inhabiting fungi on the density dependence of trees. Our results highlight the critical role of EcM fungi on conspecific neighbor effects, and we suggest that these fungi might contribute to the dynamics and assembly of tree communities more than has been appreciated. Ultimately, models of tree diversity should incorporate the role of both plant pathogens and mutualists (39, 41, 42).

Supplementary Materials

Materials and Methods

Figs. S1 to S8

Tables S1 to S6

References (4476)

References and Notes

  1. See the supplementary materials and methods.
Acknowledgments: We acknowledge the hard work of the hundreds of field assistants who were involved in the collection of tree census and seedling datasets in the Gutianshan forest dynamics plot. We especially thank R. Condit for statistical advice and comments on the manuscript and C. Gao for suggestions on molecular analysis. Funding: The study was financially supported by the National Key Research and Development Program of China (2017YFA0605103), the Strategic Priority Research Program of the Chinese Academy of Sciences (XDB31030000), the NSFC (31270495), and the Youth Innovation Promotion Association of CAS (2013058). Collaboration and manuscript preparation were facilitated by NSF US-China Dimensions of Biodiversity grants (DEB-1046113 and DEB-1241136) to Stuart Davies of the Smithsonian Institution and N.G.S. Author contributions: L.C. and K.M. designed the study. L.C., X.M., and H.R. conducted fieldwork, and L.C., L.G., and N.J. collected soil samples and carried out the molecular analysis. L.C., K.M., and N.G.S. performed statistical analysis and wrote the draft of the manuscript. All authors contributed to revisions and gave final approval for publication. Competing interests: We declare no conflicts of interest. Data and materials availability: The DNA sequencing datasets have been deposited in (43).The full census dataset used in this study is available at the CTFS-ForestGEO database portal:

Stay Connected to Science

Navigate This Article