Crops are often subjected to abiotic stresses in the process of the growth and development, while abiotic stress also activates relevant functional genes and metabolic pathways to mitigate the stress damage of crops (Giordano et al., 2021). Plant stress regulation is a complex network, and it is difficult to fully analyze the responsive mechanisms of plants to stress by studying a single gene or metabolic pathway alone (Li et al., 2017; Yang et al., 2021). Metabolomics is a powerful tool to assess the whole change of organisms at the metabolite level (Alseekh et al., 2018) and also is essential for understanding the chemical signals during plant growth and development (Carrera et al., 2021). In the analysis of metabolomics, metabolites are small molecules that are formed and/or altered during metabolisms, and changes of metabolite profiles can establish a close relationship between genotype and phenotype and also reveal the plant–environment interaction comprehensively and systematically (Bernardo et al., 2019; Carrera et al., 2021).

Arbuscular mycorrhizal fungi (AMF), subphylum Glomeromycotina of the phylum Mucoromycota, are the most widely distributed mycorrhizal fungi in soil, and can colonize the root of host plants to establish arbuscular mycorrhizal symbionts (Kokkoris et al., 2020). Mycorrhizal fungi establish a well-developed network of extraradical mycelium outside the root system, which is directly involved in water and nutrient uptake and therefore responds to various environmental stresses including drought (Meng et al., 2020; Malhi et al., 2021). AMF-induced enhancement of drought tolerance is involved in many metabolic processes, such as proline, indoleacetic acid, polyamines, fatty acids, aquaporins, antioxidant defense system, H⁺-ATPase, rhizospheric microenvironment, and so on (Wu et al., 2019; He et al., 2020; Khalil and El-Ansary, 2020; Sheteiwy et al., 2021; Zou et al., 2021; Cheng et al., 2021a,b,c). Metabolomics analysis has been applied to reveal mycorrhiza-associated metabolite changes following AMF inoculation (Rivero et al., 2015). However, very few studies have attempted to use metabolomics tool to analyze metabolite changes caused by AMF inoculation under abiotic stress (e.g., drought and salinity) conditions (Rivero et al., 2018; Bernardo et al., 2019; Yang et al., 2020). For example, Yang et al. (2020) revealed that under saline conditions, inoculation with Rhizophagus intraradices induced changes in amino acids, organic acids, flavonoids, sterols, and hormones in Puccinellia tenuiflora plants to improve osmoregulation, maintain cell membrane stability, and enhance antioxidant systems. In wheat, Funneliformis mosseae mainly triggered sugars, lipids, and oleuropein lactones involved in stress mechanisms under drought stress (Bernardo et al., 2019). These limited studies still revealed that arbuscular mycorrhizal symbionts responded to stress by regulating metabolic plasticity (Rivero et al., 2018).

Terpenoids are the most diverse and widely distributed class of secondary metabolites present in plants. Terpenoids are volatile, serve as mediators of communication between plants and other organisms, and are biologically and ecologically important for the plants’ own reproduction and defense (Jiang, 2012). They guarantee the survival of plants in stressed environments by performing several functions, such as protecting plant tissues from pathogens or herbivores and helping reproduction by attracting pollinators or seed dispersers (Quan, 2013). Severe drought dramatically increased total monoterpenes and resin acids in Pinus sylvestris and Picea abies seedlings (Turtola et al., 2003). In an aromatic plant Tanacetum vulgare, root terpenoids were obviously induced by drought, and such increase in root terpenoids would serve as more storage of resources for regrowth (Kleine and Müller, 2014).

Earlier studies have shown that secondary metabolites of many hosts such as terpenoids, flavonoids, and alkaloids were regulated by AMF inoculation under drought conditions (Bernardo et al., 2019). In P. tenuiflora seedlings, AMF inoculation under saline conditions significantly altered the relative concentrations of various metabolites, including amino acids, amines, carbohydrates, polyols, phytohormone, steroids, nucleic acids, fatty acids, flavonoids, organic acids, but terpenoids were not reported (Yang et al., 2020). However, in two wheat cultivars, F. mosseae caused the change of terpenoids, along with the up-regulation of a limonene-1,2-doil and (20S)-ginsenoside Rh2 and the down-regulation of both 7-deoxyloganate and hemigossypol-6-methyl ether (Bernardo et al., 2019). Kaur and Suseela (2020) reviewed the changes in primary and secondary metabolites of plants induced by mycorrhizal fungi and concluded that host terpene metabolism is specific to the host-associated AMF and terpene species. It was shown that sesquiterpenoid cyclohexenone derivatives could be induced to increase by Glomus intraradices (Maier et al., 1997) in the member of Poaceae, but sesquiterpenoid cyclohexenone glycoside was not induced by AMF under biotic and abiotic stresses. Various volatile terpenoids in host plants were also induced by AMF, resulting in an increased emission of these compounds in response to the environment (Rapparini et al., 2007; Fontana et al., 2009). It seems that terpenoids could be regulated by AMF inoculation under stress. Because of the wide variety of terpenoids, metabolomics is the way to clearly outline the response pattern for mycorrhizal inoculation under drought conditions, while how mycorrhizas respond to drought through changes of terpenoid profile is unclear.

The aim of this study was to comprehensively analyze the changes in the metabolome of citrus (a drought-sensitive industrial crop) by AMF inoculation under soil well-watered and soil drought stress through non-targeted metabolomics, with a focus on the response of terpenoids in the differential metabolites to reveal the mechanisms by which mycorrhizas enhance plant drought tolerance.

An important function of arbuscular mycorrhizal symbiosis is to accelerate the growth of the host plant, both under normal and unfavorable environmental conditions (Malhi et al., 2021). Our study also confirmed that R. intraradices strongly promoted shoot and root biomass production in trifoliate orange regardless of soil moisture content, and the promoting effect was relatively higher under drought conditions than under well-watered conditions. Zhang et al. (2018) used a device to separate mycorrhizal extraradical hyphae from inoculated trifoliate orange rhizosphere for quantitative estimation of hyphal water absorption rate and found that the hyphal water absorption rate was much higher under soil moisture-limited conditions than under soil moisture-sufficient conditions. This suggests that the improved plant growth of AMF is more important on arid soil than on saturated soil and therefore more attention needs to be paid to mycorrhizas in arid zones (Allen, 2006).

Metabolomic analyses implied changes in root metabolites of trifoliate orange seedlings in response to AMF inoculation. Among them, in non-inoculated plants, 50 differential metabolites were identified and metabolites were predominantly down-regulated. Similarly, in quinoa (without AMF inoculation; Chenopodium quinoa), drought stress triggered 60 differential metabolites, 53 of which were down-regulated (Shi et al., 2020); in tall fescue (Festuca arundinacea), drought stress caused 282 differential metabolites, 148 of which were down-regulated (Li et al., 2017). However, in the inoculated trifoliate orange plants, 76 differential metabolites were identified, and these metabolites were predominantly up-regulated by drought stress. This result implied an elevated number of differential metabolites and a reversal of the metabolite response pattern from down-regulation (non-inoculated plants) to up-regulation (inoculated plants). In the analysis of differential metabolite pathway, drought-induced differential metabolites of non-inoculated plants and inoculated plants were annotated to 14 and 31 metabolic pathways, respectively. Among them, non-inoculated plants were dominated by amino acid (e.g., L-glutamic acid) metabolism and inoculated plants were dominated by sugars and amino acids (e.g., L-aspartic acid), showing the different metabolic responses of both. Thus, inoculated plants have stronger metabolic activity in response to drought stress than non-inoculated plants by means of predominant up-regulation, further indicating a higher drought tolerance potential of mycorrhizal plants versus non-mycorrhizal plants.

In addition, the presence of mycorrhizal symbionts strongly stimulated the changes of metabolites in roots of trifoliate orange under well-watered and drought stress, showing, respectively, 210 and 105 differential metabolites, which were predominantly down-regulated under well-watered and up-regulated under soil drought. Yang et al. (2020) also reported more up-regulated differential metabolites of P. tenuiflora seedlings under salinity after inoculation with R. intraradices. The annotated pathways of these differential metabolites were inconsistent in response to AMF inoculation under well-watered and drought stress, indicating different response mechanisms of mycorrhizal fungi promoting plants under different soil moisture conditions. Among the metabolic pathways, amino acids responded more effectively because they are the basic units of protein synthesis, and changes in them affect protein synthesis and thus interfere with normal physiological metabolism (Yang et al., 2020). AMF induced the increase in 17 amino acids (2-aminoadipic acid (L-homoglutamic acid), N-acetylmethionine, 3-(6-hydroxy-3,4-dioxo-1,5-cyclohexadien-1-yl)-L-alanine, L-glutamic acid, guanidineacetic acid (−)-3-(3,4-dihydroxyphenyl)-2-methylalanine, S-(5ʹ-adenosy)-L-homocysteine, L-pipecolic acid, L-glutamine, L-methionine methyl ester, L-alanine, N-acetyl-L-tyrosine, N-acetylaspartate, phenylalanine-phenylalanine, S-(5ʹ-adenosyl)-L-methionine, 3-hydroxy-3-methylpentane-1,5-dioic acid, and glutamic acid) and derivatives and the reduction of 3 amino acids and derivatives (3-(2-naphthyl)-D-alanine, aspartic acid-phenylalanine, and L-cysteine) under drought stress conditions. Bernardo et al. (2019) also identified 10 differential amino acids in drought-stressed wheat after AMF inoculation, 7 of which (4-amino-2-methyl-5-phosphomethylpyrimidine, 3-phospho-hydroxypyruvate, imidazole acetol-phosphate, penicillamine, L-methionine/S-ethyl-L-cysteine, O-phospho-L-tyrosine, and 3ʹ-deamino-3ʹ-oxonicotianamine) were up-regulated, and three amino acids (2-amino-3-oxobutanoate, N-acetyl-DL-methionine, and cyclogutamate) were down-regulated. In P. tenuiflora seedlings subjected to saline stress, AMF inoculation positively promoted the accumulation of 10 amino acids and amines (Yang et al., 2020). Salvioli et al. (2012) revealed the increase in the amino acid abundance in tomato fruits after inoculation with G. mosseae. In white clover, R. intraradices accelerated leaf glutamate, aspartate, arginine, and ornithine concentrations, which was associated with changes in N-assimilated enzyme activities (Xie et al., 2021). Mycorrhizal symbiosis accelerated the uptake of various amino acids in Sorghum bicolor such as phenylalanine, asparagine, arginine, tryptophan, etc., and the uptake of such neutral or positively charged amino acids was higher than that of negatively charged amino acids (Whiteside et al., 2012). In addition, the increase in amino acids under mycorrhization may originate from the synthesis of mycorrhizal fungi (Govindarajulu et al., 2005). More amino acids were accumulated in mycorrhizal plants after stress, implying more protein synthesis under adversity. This facilitates mycorrhizalized plants to further improve osmoregulation, maintain cytoplasmic membrane permeability and nutrients, and thus enhance the response to stress signals (Teixeira et al., 2020).

Soil drought strongly affected the terpenoid content of plants, usually showing an increasing trend in total terpenes (Turtola et al., 2003; Jiang, 2012). In our study, drought stress induced a significant accumulation of phytol in roots of non-inoculated plants as well as sweroside in inoculated plants, while inhibited the decrease in capsanthin, caryophyllene, and terpineol on non-inoculated plants. Similarly, Podda et al. (2019) also revealed the increase in phytol in Brassica oleracea plants in response to drought stress. It seems that phytol, a product of chlorophyll degradation, served as an important indicator of the state of chlorophyll. Other studies in C. aurantium, Eucalyptus camaldulensis, and Thymus transcaucasicus showed diverse changes in metabolites under environmental stresses (drought, salinity, and temperature stresses) conditions, along with the reduction of metabolites more commonly (Eirini et al., 2017; Manukyan, 2019; Amrutha et al., 2021). Thus, these metabolite changes are an adaptive response mechanism in trifoliate orange under drought stress.

The results of the present study showed that AMF inoculation significantly inhibited the relative concentrations of various differential metabolites in almost all cases, except for β-caryophyllene which was increased. Bernardo et al. (2019) also reported the increase of two terpenoids (a limonene-1,2-doil and (20S)-ginsenoside Rh2) and the reduction of another terpenoids (7-deoxyloganate and hemigossypol-6-methyl ether) in wheat plants under drought stress after the colonization by F. mosseae. We speculated that mycorrhizal plants possessed more biomass including roots than non-mycorrhizal plants, so there was a dilution effect, which led to the decrease in differential metabolites. Additionally, mycorrhizal fungi may have promoted effects on the amount of volatilization and release of terpenoids under drought conditions. Many terpenes are volatile organic compounds (VOCs), and a small number of studies also displayed that mycorrhizal fungal colonization increased the emission of the green leaf volatile (Z)-3-hexenyl acetate from Plantago lanceolata (Fontana et al., 2009) and the emission of limonene and artemisia ketone from Artemisia annua (Rapparini et al., 2007). The emission of these terpenoids is almost comparable to the mechanical damage and herbivore damage to plants. These suggest that mycorrhiza-induced reduction of terpenoids may lead to an increased emission of VOCs. VOCs are an indirect way for plants to attract natural enemies to defend against insect herbivores as well as for VOCs to play a signal to initiate plant defense response (Fontana et al., 2009). More work needs to be carried out around the changes in the emission of VOCs by mycorrhizal versus non-mycorrhizal plants under adversity.

Our results also displayed that AMF inoculation significantly reduced the relative concentration of phytol in roots of trifoliate orange under drought stress, but not under well-watered. In fact, in the process of abiotic stress, a certain amount of chlorophyll is disintegrated, leading to a substantial accumulation of free phytol (Podda et al., 2019). Zhang et al. (2020) also found significantly higher chlorophyll levels in drought-stressed trifoliate orange after inoculation with F. mosseae. This implies that mycorrhizal plants have a higher chlorophyll content and therefore reduced phytol levels under drought conditions. It concludes that mycorrhizal plants under drought stress are characterized with a low level of chlorophyll degradation and thus maintain a relatively high accumulation of photosynthates, which is positive for the drought tolerance of plants.

The original contributions presented in the study are included in the article/Supplementary Material, further inquires can be directed to the corresponding author.

FZ and Q-SW conceived and designed the experiment. FZ performed the experiments. S-ML, Y-NZ, Q-SW, and KK analyzed the data. S-ML, FZ, and KK prepared the Figures. S-ML wrote the paper. Q-SW and KK revised the paper. All authors contributed to the article and approved the submitted version.

This study was supported by the National Key Research and Development Program of China (2018YFD1000303) and the Plan in Scientific and Technological Innovation Team of Outstanding Young Scientists, Hubei Provincial Department of Education, China (T201604). The authors are also grateful to Excelence project PrF UHK 2011/2021-2022 for the financial support.

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.