Walnut seeds of ‘Liaohe No. 1’ variety were pre-disinfected with 75% ethanol and germinated in autoclaved sands at room temperature. Subsequently, the seedlings having four leaves were transplanted into 2.4-L plastic pots containing 2.05 kg autoclaved mixture of sand and soil in the volume ratio of 1: 3. Mycorrhizal fungal inoculums were applied to the rhizosphere of walnut seedlings at the time of plant transplanting. The AMF-inoculated treatment (+AMF) received 150 g inoculum (23 spores/g) of D. spurca per pot, and the non-AMF-inoculated treatment (-AMF) received both 2 mL of 25 μm of inoculum filtrates and 150 g of autoclaved mycorrhizal inoculum per pot. The origin and propagation of the D. spurca strain were described in detail by Huang et al. (2020).

After plant transplanting, two soil moisture regimes (75% and 50% of the maximum soil water capacity) were performed according to the result of Li et al. (2020). The water content of potted soil was controlled at 75% of the maximum soil water capacity (well-watered, WW). After 7 weeks, half of the treated plants continued to maintain under WW conditions, and the other half of the treated plants was adjusted to 50% of the maximum soil water capacity (DS). The soil moisture was monitored by daily weighing, and the reduced water was supplemented immediately, so as to maintain the designed soil moisture condition. Such DS treatment was maintained for 6 weeks, and the seedlings were harvested. All the plants were grown in a greenhouse with a light density of 1360 lux, a relative air humidity of 66%, and a temperature of 28°C/22°C (day/night).

The experiment was a completely randomized block design consisting of two factors: (i) D. spurca inoculation (+AMF) and non-inoculation (-AMF); and (ii) soil moisture regimes with WW and DS. A total of four treatments in the experiment were arranged, coupled with five replicates (two pots as a replicate) per treatment.

Stem diameter, plant height, and leaf number per plant were measured before harvesting. Shoot and root biomass was weighed after the harvest. Root mycorrhizas were stained using the trypan blue described by Phillips and Hayman (1970). Mycorrhizal fungal colonization degree (%) = (mycorrhizal colonized root length/total length of root segments examined) × 100. Hyphal length in the soil was determined as per the method of Bethlenfalvay and Ames (1987).

Malondialdehyde (MDA, an indicator of the degree of membrane lipid peroxidation) concentrations in leaves were assayed by the thiobarbituric acid colorimetry (Sudhakar et al., 2001). H₂O₂ and O₂ ^•─ levels were assayed by the 1 mol/L KI colorimetric method and the hydroxylamine reaction, respectively (Li et al., 2022).

Soluble protein concentrations in leaves were measured as per the protocol described by Bradford (1976). Ascorbic acid (ASC) and glutathione (GSH) in leaves were extracted by grinding 0.15 g of leaf samples with 6 mL of 5% trichloroacetic acid into a homogenate and centrifuging at 15,000×g for 10 min (Wu et al., 2006). ASC and GSH concentrations in the supernatant were measured according to the method described by Li (2009). In addition, the 1 mL supernatant was incubated with 0.5 mL of 60 mmol/L dithiothreitol for 10 min to reduce the dehydroascorbic acid (DHA). The reaction solution was then incubated with 5% trichloroacetic acid, 0.4% phosphoric acid, 0.5% bathophenanthroline, and 0.03% FeCl₃, and the absorbance at 534 nm was measured for total ascorbic acid (TASC) concentrations. The DHA content was obtained by subtracting ASC from TASC.

Extraction and activity of CAT were carried out by UV spectrophotometry (Li et al., 2022). POD activity was determined using the guaiacol (0.05 mol/L) method (Li et al., 2022). Ascorbate peroxidase (APX) and glutathione reductase (GR) activities were assayed by the protocol outlined by Wu et al. (2006). Fe-SOD, Mn-SOD, and Cu/Zn-SOD activities were measured by the Enzyme-Linked Immunosorbent assay using the corresponding kit (ml902210, mll614100, and ml201168) (Shanghai Enzyme-link Biotechnology Co., Ltd., Shanghai, China), on the basis of the user manual.

Based on the identification of Hsfs in walnuts by Liu et al. (2020), the sequence of walnut Hsfs genes (JrHsf03, JrHsf05, JrHsf20, JrHsf22, and JrHsf24) was extracted from the walnut genome (https://www.ncbi.nlm.nih.gov/genome/?term=txid2249226 [orgn]). Primer sequences ( Table 1 ) were designed using Primer premier 5.0. The TaKaRa MiniBEST plant RNA kits (9769; Takara, Dalian, China) were used to extract leaf total RNA according to the user manual. After checking the integrity and concentration, the RNA was reversely transcribed into cDNA using the PrimeScript™ RT reagent kits with gDNA Eraser (RR047A; Takara, Dalian, China). The cDNA was used as a template for qRT-PCR amplification using 18S-rRNA as a house-keeping gene. Prior to performing qRT-PCR, the selected primers and melting curves had been checked to determine the reliability of the relative quantification results. Real-time fluorescence quantitative expression analysis was performed using a fluorescent dye method with three biological replicates of each treatment, and relative expression of genes was calculated using the 2^-ΔΔCt method (Livak and Schmittgen, 2001).

Statistical analysis was performed with two-factor analysis of variance, based on the SAS software 8.1v (SAS Institute Inc., Cary, NC, USA). The Duncan’s multiple range test at the 0.05% level was used to compare the significant difference among treatments.

No mycorrhiza was observed in the roots of walnut inoculated without D. spurca, and the degree of mycorrhizal colonization on the roots of walnut inoculated with D. spurca ranged from 62.9% to 73.5% ( Table 2 ). Soil water deficit significantly inhibited the degree of root mycorrhizal colonization by 14.4%, compared to WW treatment. Soil drought treatment and AMF inoculation significantly interacted on root mycorrhizal colonization.

Drought treatment obviously inhibited the growth of walnut seedlings, while mycorrhizal introduction improved plant growth ( Table 2 ). D. spurca inoculation only significantly increased plant height under WW by 29.3%, whereas it increased plant height, stem diameter, and leaf number per plant under DS significantly by 36.1%, 47.2%, and 14.4%, respectively. AMF-inoculated seedlings exhibited 25.2% significantly higher biomass under WW and 29.4% higher biomass under DS, compared with non-inoculated seedlings. A significant interaction between drought treatment and mycorrhizal inoculation occurred on biomass production.

Drought treatment significantly induced an increase in leaf H₂O₂ and O₂ ^•─ concentrations by 20.3% and 152.6% in uninoculated plants and by 32.2% and 98.7% in inoculated plants ( Figures 1A, B ). However, the inoculated plants with D. spurca recorded significantly lower leaf H₂O₂ and O₂ ^•─ concentrations by 44. 5% and 41.7% under WW and by 27.7% and 80.2% under DS, respectively, compared with the uninoculated plants. A significant interaction between drought treatment and AMF inoculation occurred on O₂ ^•─ concentrations ( Table 3 ).

The DS treatment significantly promoted MDA levels in both inoculated and uninoculated plants by 29.7% and 157.6%, respectively, relative to the WW ( Figure 2 ). On the other hand, inoculated walnut plants showed significantly lower MDA levels by 13.8% under WW and 126.1% under DS. There was a significant interaction between drought treatment and AMF inoculation for MDA levels ( Table 3 ).

Compared to the WW treatment, the DS treatment triggered a distinct decrease in soluble protein and DHA concentrations in inoculated and uninoculated plants, but induced an increase in GSH, ASC and TASC concentrations in inoculated and uninoculated plants ( Figures 3A –E ). Under WW conditions, soluble protein, ASC and TASC concentrations were increased by 40.54%, 141.57% and 3.79% in inoculated plants, compared to uninoculated plants ( Figures 3A, C, E ). Under DS conditions, soluble protein, GSH, ASC and TASC concentrations of inoculated plants were increased by 112.50%, 9.52%, 91.89% and 4.17%, compared to that of uninoculated plants ( Figures 3A, B, C, E ). AMF inoculation caused the decrease in DHA concentrations by 52.46% and 110.78% under WW and DS, respectively, compared with non-AMF inoculation ( Figure 3D ). No significant interaction between drought treatment and AMF inoculation occurred on non-enzymatic antioxidants ( Table 3 ).

The DS treatment significantly increased various antioxidant enzyme activities compared to WW treatment, independent of AMF inoculation or not ( Figures 4A –E ). In addition, under WW, Mn-SOD, Fe-SOD, CAT, POD, APX and GR activities were increased by 8.8%, 8.9%, 570.2%, 142.3%, 98.7% and 76.0% in inoculated plants compared to uninoculated plants, respectively; under DS, Mn-SOD, Cu/Zn-SOD, CAT, POD, APX and GR activities were increased by 13.8%, 1.7%, 340.4%, 80.5%, 106.3% and 77.2% in inoculated plants compared to uninoculated plants, respectively. A significant interaction between DS and AMF treatment occurred on Fe-SOD, CAT, and APX activities ( Table 3 ).

Drought treatment up-regulated expressions of JrHsf03, JrHsf20, Jrhsf22 and JrHsf24 in inoculated and uninoculated walnut plants, compared to WW treatment ( Figure 5 ). However, DS also induced JrHsf05 expressions in uninoculated plants, but down-regulated JrHsf05 expressions in inoculated plants. AMF inoculation significantly up-regulated expressions of JrHsf03, JrHsf05, JrHsf20, JrHsf22, and JrHsf24 under WW by 4.42-fold, 2.57-fold, 3.15-fold, 2.60-fold, and 1.95-fold, respectively, compared to non-AMF treatment; under DS, AMF up-regulated expressions of JrHsf03, Jrhsf22, and JrHsf24 by 1.32-fold, 1.96-fold, and 1.39-fold, respectively, compared to non-AMF inoculation, with no effect on the expression of JrHsf05 and JrHsf20. A significant interaction between drought treatment and AMF inoculation occurred on JrHsf05 and JrHsf22 expression ( Table 3 ).

Our study indicated that the DS treatment reduced the degree of root colonization by D. spurca in walnut seedlings. Similar result was reported in trifoliate orange (Liang et al., 2021) and peanut (Bi et al., 2021). Such reduction of mycorrhizal colonization under DS is due to the decrease in roots, host’s carbohydrates, and root exudates by DS, thus inhibiting spore germination and mycorrhizal colonization (Tyagi et al., 2017). Inoculation with D. spurca significantly promoted plant growth performance of walnut seedlings, which is attributed to the fact that AMF enhanced the uptake of mineral nutrients such as P, Zn and Cu, along with water absorption by mycorrhizal extraradical hyphae (Cheng et al., 2021).

Under stress, plants produce electron overflow in chloroplasts, mitochondria, peroxisomes, and plasma membranes, and thus lead to excess accumulation of ROS such as H₂O₂ and O₂ ^•─, which thus triggers oxidative damage (Cao et al., 2022). This study showed that the soil drought triggered oxidative burst (H₂O₂ and O₂ ^•─) in leaves of mycorrhizal and non-mycorrhizal walnut plants, thus increasing the degree of membrane lipid peroxidation, in accordance with increased MDA levels. However, D. spurca-inoculated walnut plants presented significantly lower ROS and MDA levels than uninoculated plants, suggesting that inoculated plants suffered relatively lower oxidative damage than uninoculated plants (Abd_Allah et al., 2015). This finding is consistent with that on trifoliate orange (Zhang et al., 2020) and lettuce (Kohler et al., 2009).

Soil drought causes oxidative damage to plants, while plants also have enzymatic (e.g., SOD, POD, and CAT) and non-enzymatic (e.g., soluble protein, ASC, and GSH) antioxidant defense systems to reduce ROS levels (Verma et al., 2019). Soluble proteins are involved in the metabolic process and are related to the water holding capacity of cells and the protective role in cell membranes (Lu et al., 2012). Our study showed that soil drought treatment inhibited soluble protein concentrations in walnut leaves, while AMF inoculation promoted soluble protein concentrations, suggesting that the inoculated plants had stronger water retention of cells and protective effect on cell membranes than uninoculated plants. Cao et al. (2022) found that Piriformospora indica, but not G. versiforme, also significantly increased leaf and root soluble protein concentrations in Satsuma mandarin, under cold temperature, but not favorable temperature conditions, implying that AMF-mediated changes in soluble protein depend on AMF species, host genotypes, and environmental stresses.

ASC-GSH cycle regulates the balance of redox state of plant cells and is an important pathway for ROS removal (Miller et al., 2010). Meanwhile, GSH maintains cell function and regulates the state of sulfhydryl groups, and ASC as an electron donor participates in substance transformation (Noctor, 2006). In this cycle, ASC is first oxidized to monohydroascorbic acid (MDHA), in which APX utilizes ASC as an electron donor to remove H₂O₂ (Wang et al., 2012). In our study, inoculated walnut plants under two soil moisture conditions showed significantly higher ASC and TASC concentrations and stronger APX activities than uninoculated plants, implying that AMF activates ASC to scavenge more H₂O₂ of the host caused by drought. In addition, AMF inoculation also increased GSH concentrations of walnut plants while decreased DHA concentrations under DS. It is known that MDHA can undergo disproportionation reaction to produce ASC and DHA (Qadir et al., 2022). DHA uses GSH as the substrate to generate GSSG and ASC under the action of dehydroascorbate reductase, and GSSG further combines with NAD(P)H as an electron donor to generate GSH under the action of GR (Li et al., 2010). Lower DHA levels and higher GSH levels and GR activity in mycorrhizal plants under DS mean that mycorrhizal plants convert more DHA to ASC and modulate more accumulation of GSH under the action of GR, as compared with non-mycorrhizal plants. Al-Arjani et al. (2020) also reported the elevation in GR and APX activities in Ephedra foliata plants after inoculated with AMF under DS. Saroy and Garg (2021) also observed that Rhizoglomus intraradices distinctly increased APX and GR activities and GSH, TASC, ASC, and GSSH concentrations in two pigeon pea genotypes under Ni stress. Glomus viscosum-inoculated Cynara scolymus plants also exhibited higher ASC and GSH concentrations than non-inoculated plants, along with elevated APX activities, for resisting the fungal pathogen Verticillium dahliae (Villani et al., 2021). These results indicate that mycorrhizal plants have a stronger ASC-GSH cycle to remove more H₂O₂ induced by stresses than non-mycorrhizal plants, thus maintaining lower oxidative damage.

In ROS scavenging enzymes, SOD catalyzes O₂ ^•─ to H₂O₂; generated H₂O₂ is then removed by POD and CAT (Qadir et al., 2022). In our study, three SODs, CAT, and POD activities were enhanced by DS, indicating that the enzymatic antioxidant defense system in walnut plants was activated in response to drought. Additionally, mycorrhizal walnut plants recorded higher POD, CAT, Mn-SOD, and Zn-SOD activities than non-mycorrhizal plants under two soil moisture regime conditions. Similar results were reported in Citrus sinensis inoculated with three different AMF species (Li et al., 2022). He et al. (2020) further found the induced expression of PtMn-SOD, PtCAT1, and PtPOD genes in trifoliate orange by F. mosseae under DS. These results suggest that AMF enhanced enzymatic antioxidant defense system to mitigate oxidative burst in response to DS.

Our study revealed that soil drought induced transcriptional levels of JrHsf03, JrHsf20, Jrhsf22, and JrHsf24 in walnut plants, independent on mycorrhizal presence. It suggests that Hsfs of walnut can respond to DS, not limited to heat stress, which is consistent with the results of Liu et al. (2020) in Hsfs of walnut under DS, heat stress, and salt stress. Similar responses of Hsfs to DS were also observed in mulberry (Zhai and Zhu, 2021) and arabidopsis (Tan et al., 2015). Moreover, We also firstly observed that JrHsf03, JrHsf05, JrHsf20, JrHsf22, and JrHsf24 were up-regulated by AMF inoculation under WW, while only JrHsf03, JrHsf22, and JrHsf24 were induced by AMF inoculation under DS, indicating that AMF-mediated response of JrHsfs depends on Hsfs types. It is not clear whether JrHsf03, JrHsf22, and JrHsf24 are specifically induced by AMF, which needs to be confirmed by additional studies. However, Gaude et al. (2012) reported the inhibited expression of HsfB3 in arbuscule-containing cortical cells of mycorrhizal roots versus cortical cells of non-mycorrhizal roots with 3.2-fold after three weeks of inoculation in roots of Medicago truncatula plants. These heat shock factors were suppressed during the initial mycorrhizal colonization (Gaude et al., 2012). Nevertheless, our study was performed for 13 weeks along with soil drought, mycorrhizal colonization had already been established, and thus this suppression may be relieved. In addition, Hsfs members are redox-sensitive transcription factors sensing ROS, transducing and amplifying the ROS signal by various proteins and transcription factors (e.g., WRKY) (Miller et al., 2008). In walnut plants, Hsfs may be associated with the signaling pathways of abscisic acid and Ca²⁺ that regulate ROS production (Mohanta et al., 2018; Liu et al., 2020). Hence, it is concluded that AMF activated some Hsfs members such as JrHsf03, JrHsf22, and JrHsf24 to regulate ROS production, but additional evidence needs to be presented. In addition, most of Hsfs members are expressed highly in roots than other tissues (Dossa et al., 2016), and mycorrhizal colonization firstly occurs in roots. More work needs to focus on the responsive pattern of root Hsfs to AMF colonization under drought, how AMF-initiated Hsfs trigger the antioxidant defense system, and whether AMF’s Hsfs are also involved in this response.