Suppression of Arbuscular Mycorrhizal Fungi Aggravates the Negative Interactive Effects of Warming and Nitrogen Addition on Soil Bacterial and Fungal Diversity and Community Composition

ABSTRACT

We examined the impacts of warming, nitrogen (N) addition, and suppression of arbuscular mycorrhizal fungi (AMF) on soil bacterial and fungal richness and community composition in a field experiment. AMF root colonization and the concentration of an AMF-specific phospholipid fatty acid (PLFA) were significantly reduced after the application of the fungicide benomyl as a soil drench. Warming and N addition had no independent effects but interactively decreased soil fungal richness, while warming, N addition, and AMF suppression together reduced soil bacterial richness. Soil bacterial and fungal species diversity was lower with AMF suppression, indicating that AMF suppression has a negative effect on microbial diversity. Warming and N addition decreased the net loss of plant species and the plant species richness, respectively. AMF suppression reduced plant species richness and the net gain of plant species but enhanced the net loss of plant species. Structural equation modeling (SEM) demonstrated that the soil bacterial community responded to the increased soil temperature (ST) induced by warming and the increased soil available N (AN) induced by N addition through changes in AMF colonization and plant species richness; ST directly affected the bacterial community, but AN affected both the soil bacterial and fungal communities via AMF colonization. In addition, higher mycorrhizal colonization increased the plant species richness by increasing the net gains in plant species under warming and N addition.

IMPORTANCE AMF can influence the composition and diversity of plant communities. Previous studies have shown that climate warming and N deposition reduce the effectiveness of AMF. However, how AMF affect soil bacterial and fungal communities under these global change drivers is still poorly understood. A 4-year field study revealed that AMF suppression decreased bacterial and fungal diversity irrespective of warming or N addition, while AMF suppression interacted with warming or N addition to reduce bacterial and fungal richness. In addition, bacterial and fungal community compositions were determined by mycorrhizal colonization, which was regulated by soil AN and ST. These results suggest that AMF suppression can aggravate the severe losses to native soil microbial diversity and functioning caused by global changes; thus, AMF play a vital role in maintaining belowground ecosystem stability in the future.

INTRODUCTION

Soil microorganisms play an essential role in terrestrial ecosystem processes as primary producers and decomposers, influencing organic matter decomposition and nutrient exchanges (1, 2). It has been widely reported that the structure and activities of microbial communities are strongly regulated by global warming and nutrient enrichment (3–6).

Global land surface temperatures may increase by 2.6 to 4.8°C by the end of this century (7). Elevated temperatures have been shown to affect the diversity and composition of soil microbial communities via changing soil moisture as well as soil nutrients in a tallgrass prairie in the United States (8), a temperate grassland in New Zealand (9), a paddy field in China (10), and a forest in the Swiss Alps (11). Concerning N deposition, it can influence soil microbial communities by alleviating resource limitations (12) and altering soil physiochemical properties such as soil acidity (13). Soil bacterial diversity was negatively and fungal diversity positively correlated with soil available nutrients in alpine grassland (14), while in a semiarid grassland, the effects of N addition on fungal diversity were related to water variation, which is a determinant of soil pH (15). In addition, global warming and N deposition has been demonstrated to affect plant species richness and diversity (16–18), which will alter the quantity and quality of litter inputs to soil (19) via contributions of net gains or net losses of plant species, a carbon resource for microbial growth, and, thus, influence soil microbial communities (19–22). Previous studies suggested that bacterial and fungal richness and community composition were mainly determined by plant species richness (23–25); shifts in soil microbial community structure were related to the changes in plant inputs (26). Therefore, global warming and N deposition potentially can have both direct and indirect impacts on soil microbial communities, and their relative contributions have been rarely addressed. Until now, few studies have reported soil microbial responses to the combined effects of global warming and N addition, and much more attention should be paid to multifactor research to fully understand how global warming and N addition affect soil microbial communities.

Arbuscular mycorrhizal fungi (AMF) are some of the most important soil microorganisms and are associated with more than 80% of terrestrial plant species (27). They can greatly increase plant nutrient uptake through an extensive network of extraradical hyphae (ERH), and the host plant in return provides photosynthate needed by the fungus (27). Therefore, AMF represent a large carbon (C) sink because of the large number of ERH (28) that supply C to other soil microorganisms. In addition, ERH exudates can stimulate or otherwise affect the growth of soil bacteria, and the large surface area of the ERH that interacts with the surrounding soil environment (the hyphosphere) provides important niches for bacterial colonization and growth (29). Meanwhile, AM associations, which represent vital links between aboveground and belowground biotic communities in ecosystems, are influenced by global warming and N deposition (30–33). Hence, the suppression of AMF would reduce microbial diversity and alter their community composition because of the lack of C resources. However, little is known about how AMF suppression influences soil microbial communities in the context of future global changes.

Semiarid grasslands account for 78% of the total grassland area in China and provide important functions and services for people living in northern China (34). Both climate warming and atmospheric N deposition are predicted to increase in this important ecosystem in the coming years, which will have potentially important but largely unknown ecological consequences (7, 35). The average air temperature in Chinese semiarid grasslands has increased by 2°C in the last 20 years, and N deposition has increased by 10.5 g m⁻² year⁻² (36, 37). Previous studies found that warming and N addition had different effects on plant species biodiversity; N addition reduced plant species diversity, while warming did not affect it (38). Moreover, warming and N addition suppressed the development of AMF and changed the fungal community composition (39). However, how AMF suppression interacts with microbial communities under warming and N addition in semiarid grasslands in northeastern China is still unclear.

Here, we examined the extent to which AMF suppression may affect microbial communities under warming and N addition. We applied benomyl (a fungicide known to reduce AMF growth and colonization) in a 4-year field experiment with elevated temperatures and N addition. We hypothesized that (i) warming and N addition would interactively affect bacterial and fungal communities; (ii) the suppression of AMF might reduce bacterial and fungal richness under warming and N addition; and (iii) the changes in plant species richness due to plant species net gains or net losses would help to explain bacterial and fungal community response to the suppression of AMF under warming and N addition.

RESULTS

Soil microclimate.

Warming and N addition increased the soil temperature by 1.7°C (F_{1, 35} = 392.55, P < 0.001; see Fig. S1a in the supplemental material) and 0.5°C (F_{1, 35} = 100.12, P < 0.001; Fig. S1a), respectively, but benomyl addition reduced soil temperature by 0.4°C (F_{1, 35} = 133.83, P < 0.001; Fig. S1a). While warming increased soil temperature, N addition canceled out the increment, indicating interactions between these two treatments (Table 1). Warming significantly decreased soil moisture by 15% ± 1% (Fig. S1b).

TABLE 1.

Results (F values) of LMM testing^a

Parameter	Species richness	Species gain	Species loss	Mycorrhizal colonization	ST	SM	pH	An	Bacterial OTU richness	Bacterial Shannon index	Fungal OTU richness	Fungal Shannon index
W	0.44	0.74	12.93***	2.06	392.55***	247.04***	0.26	1.04	0.88	0.61	0.00	0.10
N	7.00**	0.74	2.50	4.36*	100.12***	1.64	1.35	12.28**	0.04	0.15	0.76	0.05
B	33.52***	8.61**	7.47**	40.86***	133.83***	8.78	3.77	5.82*	1.94	21.03***	3.75	4.93*
W × N	0.99	3.60	3.50	0.35	43.49***	1.82	0.16	1.69	0.02	0.43	6.96*	0.42
W × B	4.62*	0.74	1.01	0.49	0.69	0.04	2.09	0.53	0.89	0.67	1.67	0.11
N × B	0.03	0.74	0.02	3.94*	0.92	0.10	0.00	2.38	1.41	1.46	5.09*	0.86
W × N×B	0.03	1.46	0.19	3.28	1.22	0.14	1.61	0.97	4.65*	4.01	0.13	0.00

^aLMMs tested the effects of warming (W), nitrogen addition (N), benomyl addition (B), and their interactions on plant species richness, net gain of plant species, net loss of plant species, mycorrhizal colonization, soil temperature (ST), soil moisture (SM), soil pH, soil available N (AN), bacterial OTU richness, bacterial Shannon-Wiener index, fungal OTU richness, and fungal Shannon-Wiener index. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

AMF suppression.

N addition and benomyl addition interactively affect mycorrhizal colonization and AMF PLFA concentrations (Table 1). Under no-N addition, benomyl addition decreased mycorrhizal colonization and AMF PLFA concentrations by 60% ± 4% (F_{1, 35} = 5.91, P < 0.001; Fig. S2a) and 42% ± 12% (F_{1, 14} = 3.55, P = 0.05; Fig. S2b), respectively. Under the N addition condition, benomyl addition decreased mycorrhizal colonization and AMF PLFA concentrations by 36% ± 6% (F_{1, 35} = 3.11, P = 0.01; Fig. S2a) and 46% ± 6% (F_{1, 14} = 4.14, P = 0.001; Fig. S2b), respectively. The concentrations of the other bacterial and fungal PLFAs were not influenced by benomyl addition (F_{1, 14} = 2.79, P > 0.05; F_{1, 14} = 3.36, P > 0.05, respectively), warming (F_1,14 = 1.68, P > 0.05; F_1,14 = 1.67, P > 0.05, respectively), or N addition (F_1,14 = 0.53, P > 0.05; F_1,14 = 0.42, P > 0.05, respectively), and no signs of pathogenic fungi or their effects were observed through visual inspection of the roots (Table S1).

Effect of AMF suppression on net changes in plant species under warming and N addition.

N addition and benomyl addition decreased plant species richness by 11% ± 2% (F_{1, 77} = 7.01, P < 0.01; Fig. 1a) and 21% ± 3% (F_{1, 77} = 33.52, P < 0.001; Fig. 1a) on average compared with the nonaddition treatments across the 4 years. Although warming did not affect plant species richness, benomyl addition plus warming interactively decreased plant species richness (F_{1, 77} = 4.62, P = 0.03; Fig. 1, Table 1). Warming alone only decreased the net loss of plant species by 32% ± 6% on average (F_{1, 56} = 12.93, P < 0.001; Fig. 1c). Benomyl addition decreased the net gain of plant species by 31% ± 7% (F_{1, 56} = 8.61, P < 0.01; Fig. 1b) and increased the net loss of plant species by 28% ± 5% (F_{1, 77} = 7.47, P < 0.01; Fig. 1c).

FIG 1.

Effects of benomyl addition on plant species richness (a), net gain of plant species (b), and net loss of plant species (c) under warming (W), nitrogen addition (N), and their interaction (WN; warming plus N addition). Values are means ± standard errors (SE). Statistically significant changes are reported in Table 1.

Effect of AMF suppression on soil microbial diversity and community composition under warming and N addition.

Bacterial operational taxonomic unit (OTU) richness was interactively affected by warming × N addition × benomyl addition (Table 1). Under warming plus N addition conditions, benomyl addition decreased bacterial OTU richness by 6% ± 1% (F_{1, 14} = 4.65, P = 0.04; Fig. 2a). Benomyl addition reduced bacterial diversity (Shannon index) by 4% ± 1% on average (F_{1, 14} = 21.03, P < 0.001; Fig. 2b) across all treatments. Significant interactive effects of warming × N addition and N addition × benomyl addition on fungal OTU richness were observed (Table 1). Without benomyl addition, warming increased fungal OTU richness, but warming plus N addition decreased fungal species richness by 16% ± 8% (F_{1, 14} = 6.96, P = 0.01; Fig. 2c) compared with N addition without warming. In addition, benomyl addition slightly increased fungal OTU richness, but under N addition, benomyl addition decreased fungal OTU richness by 21% ± 6% (F_{1, 14} = 5.10, P = 0.04; Fig. 2c). Benomyl addition alone decreased fungal diversity by 20% ± 6% on average (F_{1, 14} = 4.93, P = 0.04; Fig. 2d) across all treatments. Community dissimilarity patterns also revealed differences among communities that were generated by benomyl addition per the nonmetric multidimensional scaling (NMDS) ordination (Fig. 3), which showed a clear separation between benomyl addition and nonbenomyl addition plots for both bacterial and fungal communities (Table S2).

FIG 2.

Effects of benomyl addition on bacterial richness (a), fungal richness (b), bacterial Shannon-Wiener index (c), and fungal Shannon-Wiener index (d) under warming (W), nitrogen addition (N), and their interaction (WN; warming plus N addition). Values are means ± SE. Statistically significant changes are reported in Table 1.

FIG 3.

NMDS ordination based on Bray-Curtis similarities depicting bacterial (top) and fungal (bottom) community composition under each treatment. C, control; W, warming; N, N addition; WN, warming plus N addition; B, benomyl addition; WB, warming plus benomyl addition; NB, N addition plus benomyl addition; WNB, warming plus N addition plus benomyl addition.

Across all treatments, bacterial communities were consistently dominated by certain phyla (Fig. 4a), including Proteobacteria (45.2%), Acidobacteria (14.0%), Actinobacteria (12.4%), Gemmatimonadetes (10.8%), and Chloroflexi (6.5%). The dominant fungal taxa included (Fig. 4b) Ascomycota (31.2%), Basidiomycota (8.4%), Mortierellomycota (7.0%), and Chytridiomycota (6.9%). For bacteria, of all taxa with a mean relative abundance greater than 1%, benomyl addition increased the abundance of two out of eight phyla (Table S3 and Fig. S3, Proteobacteria and Saccharibacteria), three out of nine classes (Table S3 and Fig. S4, Alphaproteobacteria, Gammaproteobacteria, and Actinobacteria), three out of nine orders (Table S3 and Fig. S5, Sphingomonadales, Rhizobiales, and Xanthomonadales), and two out of four families (Table S3, Fig. S6, Sphingomonadaceae and Xanthomonadaceae) but decreased three out of seven phyla (Table S3, Fig. S3, Acidobacteria, Actinobacteria, and Gemmatimonadetes), four out of nine classes (Table S3, Fig. S4, Acidimicrobiales, Gemmatimonadetes, Blastocatellia, and Deltaproteobacteria), four out of nine orders (Table S3, Fig. S5, Acidimicrobiales, Gemmatimonadales, Blastocatellales, and Rhodospirillales), and two out of four families (Table S3, Fig. S6, Gemmatimonadaceae and Longimicrobiaceae). One out of nine classes (Table S3, Fig. S4, Blastocatellia) and one out of nine orders (Table S3, Fig. S5, Blastocatellales) were affected by interactions between warming and N addition. Two out of eight phyla (Table S3, Fig. S3, Actinobacteria and Nitrospirae), one out of nine classes (Table S3, Fig. S4, Actinobacteria), and one out of nine orders (Table S3, Fig. S5, Rhizobiales) were affected by interactions between warming and benomyl addition. One out of nine classes (Table S3, Fig. S4, Blastocatellia) and one out of nine orders (Table S3, Fig. S5, Blastocatellales) were affected by interactions between N addition and benomyl addition. One out of eight phyla (Table S3, Fig. S3, Chloroflexi), one out of nine classes (Table S3, Fig. S4, Betaproteobacteria), and one out of nine orders (Table S3, Fig. S5, Burkholderiales) were affected by interactions among warming and N addition and benomyl addition. For fungi, of all taxa with a mean relative abundance greater than 1%, benomyl addition increased the abundance of one out of 10 orders (Table S4, Fig. S9, Capnodiales) and 3 out of 10 families (Table S4, Fig. S10, Pleosporaceae, Myrmecridiaceae, and Cladosporiaceae) but decreased 1 out of 10 families (Table S4, Fig. S10, Phaeosphaeriaceae). One out of seven classes (Table S4, Fig. S8, Sordariomycetes), 1 out of 10 orders (Table S4, Fig. S9, Coniochaetales), and 1 out of 10 families (Table S4, Fig. S10, Coniochaetaceae) were affected by interactions between warming and N addition. One out of four phyla (Table S4, Fig. S7, Ascomycota) and 1 out of 10 families (Table S4, Fig. S10, Pleosporaceae) were affected by interactions between warming and benomyl addition. One out of seven classes (Table S4, Fig. S8, Spizellomycetes) and 1 out of 10 orders (Table S4, Fig. S9, Spizellomycetales) were affected by interactions between N addition and benomyl addition.

FIG 4.

Relative abundance of the bacterial (a) and fungal phyla (b). C, control; W, warming; N, N addition; WN, warming plus N addition; B, benomyl addition; WB, warming plus benomyl addition; NB, N addition plus benomyl addition; WNB, warming plus N addition plus benomyl addition.

Mechanisms underlying soil microbial community responses to AMF suppression under warming and N addition.

Pearson correlation analysis showed that soil, mycorrhizae, net changes in plant species, and microbial community composition had strong correlations (Fig. 5). Mycorrhizal colonization was positively correlated with soil temperature, plant species richness, plant species net gains, bacterial diversity, bacterial community composition (NMDS1B), and fungal community composition (NMDS1F) but negatively correlated with soil available N and plant species net losses. Soil available N was negatively correlated with plant species richness and plant species net gains but positively correlated with plant species net losses. Soil temperature was positively correlated with plant species richness, plant species net gains, and bacterial community composition (NMDS1B) but negatively correlated with plant species net losses.

FIG 5.

Pearson correlations among soil (ST, SM, pH, and AN), bacteria (OTUs, Shannon), fungi (OTUs, Shannon), and plant characteristics (richness, NG, and NL). ST stands for soil temperature; SM stands for soil moisture; AN stands for soil available N; OTUsB stands for bacterial richness; OTUsF stands for fungal richness; ShannonB stands for OTUsB diversity; ShannonF stands for OTUsF diversity; richness stands for plant species richness; NG stands for net gain of plant species; and NL stands for net loss of plant species. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

Structural equation modeling (SEM) demonstrated that soil temperature had direct effects only on bacterial communities and not on fungal communities (Fig. 6). Soil temperature and soil available N had indirect effects on fungal communities through mycorrhizal colonization (Fig. 6b) but had indirect effects on bacterial communities through mycorrhizal colonization and the subsequent plant species net gains or increased plant species richness (Fig. 6a). Overall, mycorrhizal colonization, which was affected by soil temperature and soil available N, directly affected bacterial and fungal community structure and indirectly affected bacterial community structure through plant species net gains and plant species richness (Fig. 6).

FIG 6.

SEM depicting the direct and indirect effects of environmental conditions on the microbial community composition. (a) Bacterial community, represented by the first axis of the NMDS. (b) Fungal community, represented by the first axis of the NMDS. Significant positive and negative paths (P < 0.05) are shown as black and red arrows, respectively, and gray dashed arrows indicate nonsignificant pathways. The width of the paths (arrows) is proportional to the strength of the path coefficients. The path coefficients are the numbers on the arrows, and R² indicates the proportion of variance explained. ST, soil temperature. AN, soil available N. *, P < 0.05; **, P < 0.01; ***, P < 0.001.

DISCUSSION

Our results showed that AMF suppression significantly decreased soil microbial diversity, plant species richness, and plant species net gains but increased plant species net losses. Meanwhile, AMF suppression interacted with global change drivers to reduce microbial richness. Climate warming and N addition, which were represented by soil temperature and soil available N, indirectly affected the soil microbial community composition through mycorrhizal colonization. Moreover, the soil temperature directly affected the soil bacterial community composition but had no impact on the soil fungal community composition. In addition, AMF suppression altered the soil bacterial community composition via plant species richness associated with plant species net gains instead of net losses; this did not occur for fungi. The changes in the soil microbial community composition were due to the increase or decrease in the relative abundance of certain bacterial or fungal taxa caused by AMF suppression. These results suggest that AMF suppression can exacerbate soil microbial diversity losses caused by global changes and can shift the soil microbial community composition and functions under the anticipated global changes.

Effects of benomyl addition on the AMF functions and plant species richness under warming and N addition.

Benomyl addition reduced AMF root colonization by approximately 60% under no-N addition, which was consistent with previous studies in grasslands (40–42). Meanwhile, AMF PLFA concentrations were decreased with benomyl addition, and no changes were found for that of other soil fungi and bacteria, similar to 8-year and 3-year benomyl application research (40, 42). Hyphal networks for most of the AMF PLFA concentrations can connect plants to transport and redistribute mineral nutrients (e.g., N and P); inhibition of these networks would suppress AMF functions (42). Moreover, in this study, benomyl addition reduced soil bacterial and fungal diversity to an extent of 4% to 20%, while plant species richness was decreased with benomyl addition to an extent of 21% to 31%. The degree of reduction is less than that for AMF functions, indicating the primary effect of benomyl in semiarid grassland soils is suppression of mycorrhizal fungi and, thus, indirectly affects soil biota and nutrient cycling. Smith et al. (43) also found that the magnitude of benomyl effects on soil biota and processes was small (e.g., 12 to 33% reduction) relative to the intended target effects of benomyl on mycorrhizal fungi (e.g., an average of 80% reduction in mycorrhizal root colonization) (43).

Warming decreased plant species net loss and N addition decreased plant species richness. In temperate grasslands, plant communities are rather resistant to increased temperature alone (17). C₄ grasses dominated our studied system and are considered well-adapted to heat and drought conditions and, thus, are not susceptible to exclusion. Soil acidification due to accumulation of N would exclude species adapted to N-poor conditions (44), and eutrophication caused by N enrichment causes plant diversity losses through enhanced light competition (45). Benomyl addition decreased plant species richness and plant species net gain but increased plant species net loss. This may be because plants with low mycorrhizal dependence were dominant in this semiarid grassland, and AMF suppression promoted the growth of dominant plant species and suppressed that of rare species with high mycorrhizal dependence, thereby reducing plant species richness. Hence, benomyl addition indirectly changes plant species richness via affecting plant species net gain and net loss differently based on their mycorrhizal dependence, which has been shown before (46, 47).

Effects of warming and N addition on microbial community structure.

In this study, we found that warming did not affect soil bacterial and fungal species richness or diversity (Fig. 2). A similar finding was obtained from an alpine grassland in which warming had no impact on the species richness of soil bacteria and fungi (39). This may be due to warming having no direct influence but indirectly affecting soil microbes due to its side effects, which include reductions in soil moisture (48), lowered labile carbon substrate availability (49, 50), and shifts in above- or belowground plant biomass (51). The present results suggest that warming alone without other related environmental changes has limited or no effects on the soil microbial community. These results imply that the effect of warming on the soil microbial community is constituted by direct or indirect mechanisms and that the latter play more important roles in semiarid grasslands. Similarly, the species richness and diversity of soil bacteria and fungi were not affected by N addition, which is in agreement with a previous result showing that the soil bacterial community composition was predominantly driven by the indirect effects of N addition, i.e., altering soil pH and plant community structure, rather than by direct effects on soil resource availability (24). However, warming in combination with N addition decreased soil fungal richness compared with individual warming or N addition, illustrating that the effects of warming and N addition were antagonistic, which subsequently altered the soil microbial community structure. Previous research found that warming combined with high N deposition significantly limited the soil microbial biomass in a subalpine forest, and warming-induced evaporation strengthened the toxicity of the added N in soil (52). In addition, the present results suggest that soil bacterial richness and diversity are highly resistant to warming and N addition and, thus, maintain ecosystem stability in the studied semiarid meadow ecosystem.

The relative abundance of particular microbial taxa also yields insights into microbial community differences under warming and N addition. At the higher taxonomic level (phylum, class, order, and family), some consistent response patterns emerged. Few variations in these dominant components of soil bacterial and fungal communities under warming and N addition were observed in this study, indicating that soil bacteria and fungi were stable under climate warming and N deposition. However, compared with warming or N addition alone, the combination of warming and N addition reduced the relative abundance of Blastocatellia class, Blastocatellales order within the phylum Acidobacteria, as well as Sordariomycetes class, Coniochaetales order, Coniochaetaceae family within the phylum Ascomycota; these changes suggest a shift in the microbial community from oligotrophic to copiotrophic taxa, since Acidobacteria and Ascomycota are both favored in less fertile environments (11, 53). At the OTU level, warming and N addition exerted no influence on soil microbial communities, suggesting a high degree of functional redundancy and a capacity for soil community resilience to warming and N addition. This result implies that belowground biological systems in harsh habitats are more resistant to environmental changes than systems in more favorable habitats.

Effects of AMF suppression on the microbial community under warming and N addition.

While warming plus N addition reduced soil bacterial richness, AMF suppression aggravated this negative effect. Meanwhile, N addition induced an increase in the fungal richness that was diluted by AMF suppression and, thus, led to a reduction in soil fungal richness. However, AMF suppression reduced the diversity of soil bacteria and fungi across these two global change drivers. Similarly, Xu et al. (54) found a significant increase in the diversity of bacteria and fungi caused by inoculation with a mixed AMF inoculum due to the greater nutrient transportation from the AMF to the host plant in exchange for greater carbon allocation from the host to the AMF hyphae, which may have affected the associated microbial community (54). Other studies also found that carbohydrates from the extraradical mycelium of an AMF have a positive effect on soil microorganisms, resulting in elevated numbers and vitality (29); this attracts specific microbial species to colonize the hyphae, increasing their abundance or diversity in the community and enhancing the functioning of these microbes (55, 56).

The four most abundant bacterial phyla were Proteobacteria, Acidobacteria, Actinobacteria, and Gemmatimonadetes, which showed various responses to AMF suppression in this research. Specifically, Actinobacteria play important roles in soil nutrient cycling (57, 58) and enhance plant growth by producing auxin and controlling various plant pathogens (59). In the present study, while warming increased the relative abundance of Actinobacteria, AMF suppression reversed this trend and led to a decline, which is in accordance with the findings of other studies (60, 61). Proteobacteria have been viewed as fast-growing bacteria that respond positively to AMF, and AMF suppression decreased the relative abundance of Deltaproteobacteria in this study. However, the relative abundance of Alphaproteobacteria and Gammaproteobacteria showed the opposite trend under AMF suppression, indicating that AMF suppression affects soil bacterial competition in the rhizosphere. AMF suppression decreased the relative abundance of Acidobacteria (Blastocatellia class), which could slow down nutrient cycling, because the family Blastocatellaceae within this phylum is highly efficient at using the recalcitrant organic pool (62). AMF suppression reduced the relative abundance of Gemmatimonadetes, which have been documented to positively respond to the presence of AMF hyphae in the litter (57) and to possess genes that can utilize exogenous N₂O (63, 64). This result suggests that the suppression of AMF induces an increase in N₂O emissions and exacerbates the greenhouse effect. In addition, the adonis analysis suggested that AMF suppression strongly affected soil bacterial communities. The present results suggest that AMF suppression hampers certain bacterial taxa that are beneficial for plant growth and nutrient cycling but promote other competitors in a global change context; therefore, AMF suppression appears to influence bacterial community composition, which is in agreement with previous findings (60).

The present results suggest that AMF suppression significantly affected the soil fungal community structure. Among the fungal phyla, although warming increased the relative abundance of Ascomycota, AMF suppression plus warming had the opposite effect. As the most diverse groups of saprotrophic fungi in the soil, changes in Ascomycota abundance and composition suggest functional shifts. At the family level, AMF suppression significantly increased the relative abundance of Pleosporaceae, Cladosporiaceae, and Myrmecridiaceae, which belong to Ascomycota, but reduced that of Phaeosphaeriaceae. Since these families include numerous saprobic and opportunistic human or plant-pathogenic taxa, stimulation or suppression of these communities by AMF suppression can change the health of their host plant. Meanwhile, N addition slightly increased the relative abundance of class Spizellomycetes, order Spizellomycetales, which possess a wide range of tolerances to harsh soil environments (65); N addition combined with AMF suppression decreased the relative abundance of both taxa, suggesting that AMF suppression hampered the growth of certain fungal taxa, changing the fungal community structure and functioning. In addition, the results of the NMDS and adonis analyses showed that AMF suppression was the main factor shaping the fungal community. These results illustrate that soil fungi exhibited resistance to environmental factors, showing only a few taxon changes, and that AMF suppression could induce variation in the fungal community composition and functions under a global change scenario. Thus, AMF are important biotic soil components, and their absence or depletion due to, e.g., anthropic inputs or climate change can lead to less efficient ecosystem functioning.

Mechanisms shaping the microbial community.

Previous studies have shown that climate warming and N addition can directly affect soil microbial communities or indirectly affect them by affecting the plant community composition (24, 66). In this study, warming increased the soil temperature but decreased the soil moisture, N addition increased the soil available N, and neither warming nor N addition changed the soil pH. Moreover, AMF suppression decreased mycorrhizal colonization. Mycorrhizal colonization showed positive correlations with soil temperature, and the bacterial and fungal communities of the first axis in the NMDS showed negative correlations with soil available N (Fig. 5). Our SEM thus suggested that the indirect mycorrhizal colonization effects of soil temperature and soil available N accounted for the variation in bacterial and fungal communities (Fig. 6). While warming reduced plant species net losses and N addition decreased plant species richness, AMF suppression decreased plant species richness and plant species net gains but increased the plant species net losses. SEM showed an indirect effect of AMF suppression through plant species richness on the bacterial community but not on the fungal community. Meanwhile, mycorrhizal colonization impacted plant species richness directly and indirectly by providing plant species net gains without species net losses (Fig. 6).

Overall, warming and N addition influenced the soil bacterial community by affecting AMF colonization and the richness of the associated plant species community, whereas AMF colonization had a direct effect on the fungal community. Our results suggest that mycorrhizal colonization plays vital roles in maintaining microbial diversity and stabilizing belowground communities under future global changes; these roles are also essential for maintaining aboveground ecosystem structure and functioning since above- and belowground biota are closely connected.

MATERIALS AND METHODS

Study site.

The experiment was conducted at the Songnen Grassland Ecological Research Station (44°45′N, 123°45′E), Northeast Normal University, Jilin Province, Northeast China. The grassland is located at the eastern edge of the Eurasian steppe and is characterized as a temperate meadow. The mean annual precipitation is approximately 400 mm, 90% of which occurs in the growing season from May to October. The annual average air temperature is 4.9°C, and the annual average land surface temperature is 6.2°C. The soil in this area is soda-saline soil, and the soil pH is approximately 9.0, with 3 to 4% organic matter in the soil surface layer. The plant species Leymus chinensis and Puccinellia tenuiflora (Trin. Tzvel, Poaceae) are the dominant species at the experimental site.

Experimental design.

This study involved a completely randomized block factorial experimental design including two major global changes: warming (with and without) and N addition (with and without). Each global change driver included two AMF conditions: AMF presence (without benomyl) and AMF suppression (with benomyl). Overall, there were a total of eight treatments replicated in three blocks: control (C), warming (W), N addition (N), benomyl addition (B), warming plus N addition (WN), warming plus benomyl addition (WB), N addition plus benomyl addition (NB), and warming plus N addition plus benomyl addition (WNB). Plots were 2 m by 1.5 m in area and separated by 1 m within each block, and the blocks were separated by 2 m. In the warming treatment, infrared heaters (MSR-2420; Kalglo Electronics Inc., Bethlehem, PA, USA) were suspended over the plot center at a height of 2.25 m to maintain a temperature of approximately 1.7°C (0 to 30 cm average soil temperature; CR1000 Datalogger; Campbell Scientific Inc., Logan, UT, USA). Dummy heaters of the same shape and size were installed above untreated plots. Studies in northern temperate grasslands have suggested that changes in aboveground biomass, species richness, and plant functional group composition saturate at an N deposition rate of approximately 10 g m⁻² year⁻¹ (37). Therefore, we added ammonium nitrate 10 times at 1 g N m⁻² (in 10 liters of water) each time from May to September every 2 weeks each year for a total of 10 g m⁻²year⁻¹ in the N addition treatment. An identical amount of water (without N) was added to the other plots. Mycorrhiza-suppressed plots received the fungicide benomyl as a soil drench (18 g m⁻² active ingredient in 10 liters of water), and an identical amount of water (without benomyl) was added to the other plots. This fungicide has been used to suppress AMF activity, since it is effective in reducing AMF colonization and has negligible side effects on nontarget fungi and plant growth (40, 41, 67). The experiment started in May 2015.

Soil sampling and plant community survey.

In each plot in mid-September 2018, we collected, sieved (2 mm), and mixed three soil cores (15 cm in depth, 5 cm in diameter) to create a composite soil sample. The fresh soil samples from 2018 were frozen and stored at –80°C for phospholipid fatty acid (PLFA) analysis and DNA extraction. The roots from each species in each plot were cut into approximately 1-cm lengths and thoroughly mixed to estimate root mycorrhizal colonization. A permanent quadrant of 1 m by 1 m was established in each subplot in May 2015. From 2015 to 2018, we recorded each plant species within each quadrant in mid-July.

Calculations of plant species changes.

Plant species gains and losses were calculated for each year of the experiment. First, the total number of plant species present in each plot (the species richness) was calculated for every year of the experiment. Second, we determined the number of species gained and lost within each plot relative to the number in 2015. A species gain (colonization) was defined as any species that was previously absent but subsequently appeared in the community. Species loss (local extinction) was defined as the disappearance of a species from the community. The plant species richness was calculated for every year of the experiment, from 2015 through 2018, for each plot. We repeated the above-described calculations for species gains and losses over 4 years relative to the number of species in the first year of the study (2015). The results calculated are referred to as the net gain and net loss during the experimental period.

Soil DNA extraction and Illumina HiSeq sequencing.

Soil DNA was extracted from the samples using the Power Soil DNA isolation kit (MO BIO Laboratories) according to the manufacturer's protocol. DNA quality and quantity were assessed by the ratios of 260 nm/280 nm and 260 nm/230 nm. DNA then was stored at –80°C until further processing. The V3-V4 region of the bacterial 16S rRNA gene was amplified with the primers 338-F (5′-ACTCCTACGGGAGGCAGCA-3′) and 806-R (5′-GGACTACHVGGGTWTCTAAT-3′) combined with adapter sequences and barcode sequences (68). The ITS1 region of the rRNA gene of fungi was amplified using the primers ITS1-F (5′-CTTGGTCATTTAGAGGAAGTAA-3′) (69) and ITS2-R (5′-GCTGCGTTCTTCATCGATGC-3′) (70). PCR amplification was performed in a total volume of 50 μl that contained 10 μl buffer, 0.2 μl Q5 high-fidelity DNA polymerase, 10 μl high GC enhancer, 1 μl deoxynucleoside triphosphate, 10 μM each primer, and 60 ng genome DNA. The thermal cycling conditions were an initial denaturation at 95°C for 5 min followed by 15 cycles at 95°C for 1 min, 50°C for 1 min, and 72°C for 1 min, with a final extension at 72°C for 7 min. The PCR products from the first step of PCR were purified through VAHTS DNA clean beads. The second round of PCR was then performed in a 40-μl reaction mix that contained 20 μl 2× Phμsion HF MM, 8 μl double-distilled H₂O, 10 μM each primer, and 10 μl PCR products from the first step. The thermal cycling conditions were an initial denaturation at 98°C for 30 s followed by 10 cycles at 98°C for 10 s, 65°C for 30 s, and 72°C for 30 s, with a final extension at 72°C for 5 min. Finally, all PCR products were quantified by Quant-iT double-stranded DNA HS reagent and pooled. High-throughput sequencing analysis of the bacterial and fungal rRNA genes was performed on purified, pooled samples using the Illumina HiSeq 2500 platform (2 × 250 paired ends) at Biomarker Technologies Corporation, Beijing, China.

Sequence analyses.

The paired-end reads of the fungal internal transcribed sequence (ITS) region amplicons and bacterial 16S rRNA gene were processed, and the ITS region and 16S sequences were quality screened according to previously described methods (71). The tags of >97% identity were clustered into operational taxonomic units (OTUs) using QIIME (version 1.8.0). The tags were classified into different taxonomies according to the Silva and UNITE databases for soil bacterial and fungal communities, respectively. There were 17,822 OTUs of soil bacteria and 4,986 OTUs of soil fungi after removing those OTUs that did not belong to the soil bacterial and fungal communities.

Laboratory analysis.

The soil temperature and soil moisture were recorded by a data logger that was inserted into the soil. The soil ammonium and nitrate were extracted with 2 M KCl (a soil to water ratio of 1:5) and measured using a continuous flow analyzer (Futura II; Alliance Instruments Ltd., France). The soil inorganic N is presented in milligrams per kilogram.

The mixed roots were cleaned using 10% KOH and stained in 0.05% trypan blue in a lactic acid-glycerin solution (72). Mycorrhizal colonization was then examined microscopically (73).

PLFAs were used to test the effects of benomyl on soil bacteria and fungi according to a modified Bligh and Dyer methodology (74). PLFA analyses were performed using an Agilent 6890 gas chromatograph (Agilent Technologies, Palo Alto, CA, USA) and SHERLOCK software (MIDI Inc., Newark, NJ, USA). The fatty acids i14:0, i15:0, a15:0, 15:0, i16:0, 16:1ω7c, i17:0, a17:0, 17:0cy, 17:0, 18:1ω7c, and 18:1ω5c were chosen to represent bacterial PLFAs. 16:1ω5c was selected to represent the AMF PLFAs, and 18:2ω6.9c and 18:1ω9c were used to represent the PLFAs of other fungi (75).

Statistical analyses.

All analyses were conducted using R version 3.6.0. (76). Plant species net gain and loss, plant species richness, and mycorrhizal colonization were analyzed using linear mixed-effects models, with warming, N addition, and benomyl addition as fixed factors and block and year as random factors. We also generated linear mixed-effects models with warming, N addition, and benomyl addition as fixed factors and the block as the random factor to test the effects of warming, N addition, benomyl addition, and their interactions on soil temperature, soil moisture, soil pH, soil available N, bacterial richness, bacterial diversity, fungal richness, fungal diversity, and taxa with a mean relative abundance higher than 1%. Analyses were performed using the “lme” function from the “nlme” package (77). A correlation matrix was constructed to look for relationships between soil properties (soil temperature, soil moisture, pH, and soil available N), mycorrhizal colonization, plant species changes (net gain, net loss, and richness), and soil microbes. Analyses were performed using the “rcorr” function from the “Hmisc” package (78).

We analyzed the changes in the composition and structure of bacterial and fungal communities through permutational analyses of variance (PERMANOVA; 999 permutations) using the “adonis” function in the “vegan” package (79). QIIME software was used to statistically examine differences among bacterial and fungal communities across all treatments and to calculate the two first components of the NMDS for bacterial and fungal communities.

The “lavaan” package (80) in R was used to implement structural equation modeling (SEM) based on an a priori hypothesis of both direct and indirect effects of treatments on bacterial and fungal community structure, informed by the results of the PERMANOVA and correlation analyses. The model fit was considered good when the χ² test and its associated P values were low (<2) and high (>0.05), respectively. The root mean square error of approximation (RMSEA) was also used to evaluate the goodness of fit. A model has a good fit when the RMSEA is <0.05 and its associated P value is >0.05

Data availability.

The sequence reads for all samples have been deposited in the National Center for Biotechnology Information Sequence Read Archive with the accession numbers PRJNA638025 and PRJNA638076.