Growth, Secondary Metabolites and Enzyme Activity Responses of Two Edible Fern Species to Drought Stress and Rehydration in Northeast China

The drought resistance mechanism of Matteuccia struthiopteris (L.) Todar. and Athyrium multidentatum (Doll.) Ching were measured under natural drought exposure. The results showed that the two edible fern species showed stronger resistance in the early stages of drought, mainly expressed as the decrease of relative leaf water content (RLWC), increase of osmotic substances, secondary metabolites such as flavonoids (FC), total phenols (TPC), proantho cyanidins (PCC) content and enzyme activity (superoxide dismutase (SOD), peroxidase (POD), catalase (CAT) and ascorbate peroxidase (APX)). The higher RLWC, FC, TPC, PCC and abscisic acid (ABA) content and lower H2O2 content indicates the stronger non-enzymatic antioxidant system and drought resistance of A. multidentatum. However, the proline (Pro) content changed slowly, and the synthesis of soluble protein (SP), total phenols, proantho cyanidins and ABA, SOD activity of two fern species were inhibited in the late stages of drought stress. This study can provide a scientific basis for the cultivation and utilization of edible fern species under forest in Northeast China.


Introduction
Global warming has already led to frequent and extreme droughts in many areas [1,2]. The northeastern region of China is located in the middle and high latitudes of the northern hemisphere and is an important grain production base in China [3]. In recent years, the climate has been warming and drying in Northeast China. Since 1960, the drought and flood situation in the northeast has alternated, and the frequency of drought has gradually increased with the decadal changes [4], and the degree of drought has gradually increased. Moreover, the drought range is the largest and the intensity is strongest in summer [5].
Drought stress is one of the most serious abiotic stresses, which can lead to serious damage to plants including stomata closure, limitation of photosynthesis and inhibition of physiological functions such as antioxidant functions [6][7][8]. Under drought stress, plants will regulate the morphology, physiology and metabolism to adapt to a water deficit environment [9]. Generally, plants accumulate osmotic substances such as soluble sugars (SS), soluble proteins (SP) and proline (Pro) to regulate water potential and encourage plants to absorb more water from the environment to maintain normal

Experiment and Stress Treatment
The experiment was carried in the Arboretum (41 • 46 N, 123 • 27 E) of the Institute of Shenyang Applied Ecology, Chinese Academy of Sciences, China. The two fern species were placed under a canopy in a natural environment. Twenty-five of the two fern species were amply watered and then the irrigation was stopped, and the plants were exposed to the gradually drying soil environment after 1 day. The plants watered for 1 day were used as plant materials of day 0. Five pots were randomly selected for the measurement of growth indexes and three pots were randomly selected for the measurement of physiological indexes every 5 days, rehydration was conducted at day 20 and various indexes of plants were determined at day 21 after rehydration. TDR 200 Soil Moisture Sensor was used to monitor soil moisture, the soil moisture content was 31%, 17%, 8%, 5%, 2% at day 0, 5, 10, 15 and 20 of drought exposure, respectively, and then rose to 34% after rehydration.

Growth Traits
Five leaves were collected from each plant, the fresh weight (FW) was weighed, and the leaf area (LA) was measured according to the description of Xiao et al. [36]. The specific leaf area (SLA) was calculated as leaf area per dry weight (cm 2 g −1 ). The leaves were weighed (FW) and placed in distilled water for 8 h, the total weight (TW) was weighed, then the leaves were put in paper bags and dried at 105 • C for 30 min, then dried at 80 • C to a constant weight, and the dry weight (DW) was recorded. The relative leaf water content (RLWC) was calculated according to Li et al. [37] using the formula: RLWC (%) = [(FW − DW)/(TW − DW)] × 100.

Hydrogen Peroxide (H 2 O 2 ), Malondialdehyde (MDA) Content and Ion Leakage (IL)
The H 2 O 2 concentration was measured as described by Mukherjee and Choudhuri [38]. Leaf samples (0.2 g) were ground in liquid nitrogen, extracted in ice-cold acetone, then centrifuged at 12,000 rpm for 20 min at 4 • C. The extracted solution (1 mL) was mixed with 0.1 mL 10% titanium chloride in HCl, 0.2 mL strong ammonia. After precipitate was formed, the mixture was centrifuged at 10,000 rpm for 10 min. The precipitate was washed repeatedly with acetone 3 to 5 times until the plant pigments were removed. H 2 SO 4 (2 mol L −1 10 mL) was added to the precipitate after washing. When the precipitate was dissolved completely, the solution was used to measure the H 2 O 2 content at 415 nm by using UV-1800 spectrophotometer (Shimadzu, Kyoto, Japan).
The MDA concentration as the marker of lipid peroxidation was determined as described by Heath and Packer [39] with modifications. Leaf samples (0.2 g) were powdered with liquid nitrogen and homogenized with 5 mL 20% TCA. The mixture was centrifuged at 7000 rpm for 10 min and the supernatant was taken. Two milliliters of 20% TCA containing 0.5% thiobarbituric acid (TBA) was added to the supernatant and heated in a 100 • C water bath for 15 min and then rapidly cooled down in crushed ice. After centrifugation at 10,000 rpm for 10 min, the absorbance of the supernatant was measured at 450 nm, 532 nm and 600 nm by using a UV-1800 spectrophotometer (Shimadzu, Kyoto, Japan).
Ion leakage (IL) of leaves was measured as described by Liu et al. [40] with modifications. Leaves (0.2 g) from different plants were placed into test tubes, and 50 mL distilled water was added. The test tubes were stood for 8 h before measuring the electrical conductivity (S1). The test tubes were placed in a boiling water bath for 15 min and then cooled to room temperature, and the electrical conductivity was measured (S2). At the same time, the electrical conductivity of distilled water was measured (S3). The IL formula is as follows: IL (%) = (S1 − S3)/(S2 − S1) × 100.

Osmotic Regulators
Soluble protein (SP) was determined according the method of Marion [41]. Leaf samples (0.2 g) were powdered with liquid nitrogen and homogenized with 5 mL 0.05 mol L −1 phosphate buffer (pH = 7.8), and then centrifuged at 3000 rpm for 10 min. Supernatant (1 mL) and coomassie brilliant blue Agronomy 2019, 9, 137 4 of 17 (5 mL) were added to a 10 mL centrifuge tube for 2 min, then the absorbance of the reaction mixture was measured at 595 nm by using a UV-1800 spectrophotometer (Shimadzu, Kyoto, Japan).
Proline (Pro) was extracted and measured as described by Irigoyen et al. [42] with modifications. Leaf samples (0.5 g) were placed in a large test tube, and 3% 5 mL sulfosalicylic acid solution was added, and extracted was heated in a boiling water bath for 10 min (shaken frequently during the extraction process). After cooling, the mixture was filtered in a clean test tube, and the filtrate was the extract of Pro. Extracts (2 mL) were transferred into a centrifuge tube with a lid, 2 mL glacial acetic acid and 2 mL 2.5% ninhydrin were added, and then the centrifuge tubes with reaction solution were heated in a boiling water bath for 30 min. After cooling, 4 mL of toluene was added. The sample was shaken for 30 s and stood for a while, and the upper layer was taken and centrifuged for 5 min at 3000 rpm. The upper layer of the Pro red toluene solution was taken to measure the absorbance at 520 nm by using a UV-1800 spectrophotometer (Shimadzu, Kyoto, Japan).
The ABA Elisa kit (ZC BIO ZC-53364) (http://www.zcibio.com) was used to measure the ABA content. Leaf samples (0.2 g) were ground in liquid nitrogen, 1 mL 80% methanol was added and was set at −20 • C overnight. The extract was centrifuged for 20 min at 4 • C, 8000 rpm. The supernatant was passed through a C-18 solid phase extraction column. After passing through the column, the sample was vacuum dried. Phosphate buffer (pH = 7.4 1 mL) was added into the tube containing sample and mixed, and stood at room temperature for 30 min, which was centrifuged at 4 • C, 8000 rpm for 15 min. The supernatant was stored at 4 • C for determining ABA content by using a Microplate reader (InterMed, South Portland, ME, USA). Standard curves and calculation formulas were provided in the kit instructions.

Secondary Metabolites
Secondary metabolites were determined according to Ali et al., Sarker and Oba, and Zhang et al. [43][44][45]. The samples were dried to constant weight, pulverized, and passed through a 40 mesh sieve. About 0.1 g dry sample was weighed to extract secondary metabolites by using 60% ethanol. Extraction was performed using ultrasonic extraction, and the power was 300 W and the temperature was 60 • C. After extracting for 30 min, centrifugation at 12,000 rpm at 25 • C for 10 min, the extraction was used to test these three indexes by using a Microplate reader (InterMed, South Portland, ME, USA). The reagents from a biochemical kit (ZCI BIO ZC-S0358, ZC-S0356, ZC-S0357) (http://www.zcibio.com) were used to determine Proantho cyanidins, flavonoids and total phenols content. Standard curves and calculation formulas were provided in the kit instructions.

Antioxidant Enzyme Activities
Leaf samples (0.2 g) were powdered with liquid nitrogen and ground with 5 mL 0.05 mol L −1 phosphate buffer (pH = 7.8). The mixture was centrifuged at 8000 rpm at 4 • C for 15 min, and the supernatant was used to analyze the CAT, POD and SOD activities. The CAT activity was measured as described by Azevedo et al. [46]. The reaction mixture contained 0.05 mol L −1 phosphate buffer (pH = 7.8) and 0.1 mol L −1 H 2 O 2 , distilled water and enzyme extract was used to measure the CAT activity at 240 nm by using a UV-1800 spectrophotometer (Shimadzu, Kyoto, Japan). One unit of CAT was defined as the amount of enzyme that reduced by 0.1 under A 240 in 1 min. The POD activity was measured as described by Zhou and Leul [47] with some modifications. The reaction mixture included 0.05 mol L −1 phosphate buffer (pH = 7.0), 0.2% (v/v) guaiacol and 30% H 2 O 2 . The enzyme extract was added to reaction mixture (3 mL), and the mixture was used to measure POD activity at 470 nm by using a UV-1800 spectrophotometer (Shimadzu, Kyoto, Japan). One unit of POD was defined as 0.01 absorbance increase per minute at 470nm. The SOD activity was measured as described by Giannopolitis and Ries [48]. The reaction system included 0.05 mol L −1 phosphate buffer, 130 mmol L −1 methionine, 630 µmol L −1 NBT, enzyme extract and 13 µmol L −1 riboflavin. The SOD activity was measured at 560 nm by using a UV-1800 spectrophotometer (Shimadzu, Kyoto, Japan). One unit of SOD was defined as the amount of enzyme added by 50% inhibition of NBT reduction. The APX was measured according to Nakano and Asada [49]. Leaf samples (0.2 g) were powdered with liquid nitrogen and ground with 3.5 mL 0.1 mol L −1 phosphate buffer (pH = 7.0) containing 1 mmol L −1 EDTA, 1 mmol L −1 ascorbic acid (ASA), 4% PVP and 0.25% Triton X-100, the mixture was transferred into a centrifuge tube, then, 1.5 mL saturated (NH 4 ) 2 SO 4 was added and the centrifuge tube was shaken well. The mixture was centrifuged at 13,000 rpm, at 4 • C for 20 min, and the supernatant was taken. The reaction mixture containing 0.05 mmol L −1 ASA, 2 mmol L −1 H 2 O 2 and enzyme extract was used to analyze the APX activity at 290 nm by using a UV-1800 spectrophotometer (Shimadzu, Kyoto, Japan). One unit of APX was defined as the amount of enzyme that oxidizes 1 µmol of ASA per minute at room temperature (extinction coefficient 2.8 mM −1 cm −1 ).

Statistical Analysis
All the data were analyzed with Microsoft Excel 2016 and SPSS 22.0 software, graphs were edited with GraphPad Prism 5 software. The data were statistically analyzed by the ANOVA program in SPSS 22. Comparisons of the means used the least significant difference (LSD) at p ≤ 0.05. The data in the chart is the mean ± standard deviation.

Growth Traits
The RLWC of M. struthiopteris and A. multidentatum decreased slowly to 91.48% and 92.10% at 5% soil moisture, respectively, and then at 2% soil moisture, the RLWC fell sharply to 79.76% and 83.14%, respectively. After rehydration, the RLWC rose to 94.22% and 93.34%, respectively ( Figure 1A). containing 1 mmol L EDTA, 1 mmol L ascorbic acid (ASA), 4% PVP and 0.25% Triton X-100, the mixture was transferred into a centrifuge tube, then, 1.5 mL saturated (NH4)2SO4 was added and the centrifuge tube was shaken well. The mixture was centrifuged at 13,000 rpm, at 4 °C for 20 min, and the supernatant was taken. The reaction mixture containing 0.05 mmol L −1 ASA, 2 mmol L −1 H2O2 and enzyme extract was used to analyze the APX activity at 290 nm by using a UV-1800 spectrophotometer (Shimadzu, Kyoto, Japan). One unit of APX was defined as the amount of enzyme that oxidizes 1 μmol of ASA per minute at room temperature (extinction coefficient 2.8 mM −1 cm −1 ).

Statistical Analysis
All the data were analyzed with Microsoft Excel 2016 and SPSS 22.0 software, graphs were edited with GraphPad Prism 5 software. The data were statistically analyzed by the ANOVA program in SPSS 22. Comparisons of the means used the least significant difference (LSD) at p ≤ 0.05. The data in the chart is the mean ± standard deviation.
The change in SLA of the two fern species is shown in Figure 1B. The SLA of M. struthiopteris showed an upward trend. The SLA increased by 12% at 2% of soil moisture compared to the SLA at 31% of soil moisture. However, the change in the SLA of A. multidentatum is complicated, the SLA at 2% of soil moisture increased by 8% compared to the SLA at 31% of soil moisture. After rehydration, the SLA of M. struthiopteris increased by 4%, while the SLA of A. multidentatum decreased by 9% compared to the SLA at 2% of soil moisture. The SLA of M. struthiopteris was significantly higher than A. multidentatum (p ≤ 0.05).

H2O2, MDA Content and IL
The H2O2 content of M. struthiopteris increased during drought stress, and continued to increase to the maximum (1.55 μmol g −1 FW) after rehydration. However, the H2O2 content of A. multidentatum reached the maximum (0.29 μmol g −1 FW) at 5% of soil moisture and then decreased ( Figure 2A). The The change in SLA of the two fern species is shown in Figure 1B. The SLA of M. struthiopteris showed an upward trend. The SLA increased by 12% at 2% of soil moisture compared to the SLA at 31% of soil moisture. However, the change in the SLA of A. multidentatum is complicated, the SLA at 2% of soil moisture increased by 8% compared to the SLA at 31% of soil moisture. After rehydration, the SLA of M. struthiopteris increased by 4%, while the SLA of A. multidentatum decreased by 9% compared to the SLA at 2% of soil moisture. The SLA of M. struthiopteris was significantly higher than A. multidentatum (p ≤ 0.05).

H 2 O 2 , MDA Content and IL
The H 2 O 2 content of M. struthiopteris increased during drought stress, and continued to increase to the maximum (1.55 µmol g −1 FW) after rehydration. However, the H 2 O 2 content of A. multidentatum reached the maximum (0.29 µmol g −1 FW) at 5% of soil moisture and then decreased (Figure 2A). The H 2 O 2 content of A. multidentatum was significantly higher than M. struthiopteris except at 31% Agronomy 2019, 9, 137 6 of 17 and 5% of soil moisture. The MDA content of M. struthiopteris increased from 6.68 to 20.16 µmol g −1 FW, and then decreased to 9.71 µmol g −1 FW after rehydration. However, the MDA content of A. multidentatum reached its maximum (29.65 µmol g −1 FW) at 17% of soil moisture, and then increased from 14.39 (8% of soil moisture) to 28.26 µmol g −1 FW (2% of soil moisture). After rehydration, it decreased to 22.00 µmol g −1 FW ( Figure 2B). In addition, the MDA content of A. multidentatum was significantly higher than M. struthiopteris (p ≤ 0.05) except at 8% of soil moisture. The IL showed similar variation ( Figure 2C). The IL reached their maximum (29.47% and 33.31%) when the soil moisture dropped to 2%, respectively. Then the IL decreased after rehydration.  Figure 2B). In addition, the MDA content of A. multidentatum was significantly higher than M. struthiopteris (p ≤ 0.05) except at 8% of soil moisture. The IL showed similar variation ( Figure  2C). The IL reached their maximum (29.47% and 33.31%) when the soil moisture dropped to 2%, respectively. Then the IL decreased after rehydration.

Osmotic Regulators
The SP of M. struthiopteris and A. multidentatum showed same trend ( Figure 3A), which reached their maximum (1.75 mg g −1 FW) at 17% of soil moisture and 8% of soil moisture (1.66 mg g −1 FW) during the exposure to drought stress, respectively, and then decreased until rehydration. The SP of the two fern species increased to 1.66 and 1.73 mg g −1 FW after rehydration, respectively. However, the Pro content did not express an obvious change except for the SP content at 2% of soil moisture. The maximum values of Pro content of the two fern species were reached at 2% of soil moisture (21.12 and 19.21 mg g −1 FW) and then decreased after rehydration ( Figure 3B).
In the early stages (31% of soil moisture) of drought exposure, the ABA content of M. struthiopteris was 577.4 ng g −1 , and then decreased. There were no differences at 17%, 8% and 5% of soil moisture and rehydration (34% of soil moisture), but the ABA content at 2% of soil moisture was significantly lower than others (p ≤ 0.05). However, the ABA content of A. multidentatum reached the maximum (564.5 ng g −1 ) when the soil moisture was 8%, but there were no significant differences at 31%, 17%, 8% and 5% of soil moisture. The minimum ABA content (41.02 ng g −1 ) appeared when the soil moisture dropped to 2%, and then increased after rehydration ( Figure 3C). The ABA content of  (Figure 2B), the significant mark of M. struthiopteris is on the upper side, and the significant mark of A. multidentatum is on the lower side. In (Figure 2C), the significant mark of M. struthiopteris is on the lower side, and the significant mark of A. multidentatum is on the upper side. Error bars are ±SD (n = 3, 3 and 5).

Osmotic Regulators
The SP of M. struthiopteris and A. multidentatum showed same trend ( Figure 3A), which reached their maximum (1.75 mg g −1 FW) at 17% of soil moisture and 8% of soil moisture (1.66 mg g −1 FW) during the exposure to drought stress, respectively, and then decreased until rehydration. The SP of the two fern species increased to 1.66 and 1.73 mg g −1 FW after rehydration, respectively. However, the Pro content did not express an obvious change except for the SP content at 2% of soil moisture. The maximum values of Pro content of the two fern species were reached at 2% of soil moisture (21.12 and 19.21 mg g −1 FW) and then decreased after rehydration ( Figure 3B).
In the early stages (31% of soil moisture) of drought exposure, the ABA content of M. struthiopteris was 577.4 ng g −1 , and then decreased. There were no differences at 17%, 8% and 5% of soil moisture and rehydration (34% of soil moisture), but the ABA content at 2% of soil moisture was significantly lower than others (p ≤ 0.05). However, the ABA content of A. multidentatum reached the maximum (564.5 ng g −1 ) when the soil moisture was 8%, but there were no significant differences at 31%, 17%, 8% and 5% of soil moisture. The minimum ABA content (41.02 ng g −1 ) appeared when the soil moisture dropped to 2%, and then increased after rehydration ( Figure 3C). The ABA content of A. multidentatum was significantly higher than M. struthiopteris (p ≤ 0.05) except at 31% of soil moisture and after rehydration. A. multidentatum was significantly higher than M. struthiopteris (p ≤ 0.05) except at 31% of soil moisture and after rehydration.

Secondary Metabolites
The FC of M. struthiopteris reached the maximum (9.40 mg g −1 FW) at 17% of soil moisture, and then increased from 6.53 (8% of soil moisture) to 9.48 mg g −1 FW after rehydration. The FC of A. multidentatum increased from 5.74 (31% of soil moisture) to 12.23 mg g −1 FW after rehydration ( Figure  4A). Furthermore, the FC of A. multidentatum was significantly higher than M. struthiopteris when the soil moisture was 8%, 5%, 2% and 34% (rehydration) (p ≤ 0.05). The TPC of M. struthiopteris reached the maximum (0.62 mg g −1 FW) when the soil moisture was 2%. There were no significant differences among 31%, 17% and 8% of soil moisture. After rehydration, the TPC decreased to the same level of 31%, 17% and 8% of soil moisture. However, the TPC of A. multidentatum increased from 0.39 (31% of soil moisture) to 0.63 mg g −1 FW (5% of soil moisture), and then decreased, but turned to increase again after rehydration ( Figure 4B). Otherwise, the TPC of A. multidentatum at 17% and 8% of soil moisture was significantly higher than M. struthiopteris (p ≤ 0.05). The PCC of M. struthiopteris increased from 1.63 (31% of soil moisture) to 3.47 mg g −1 FW (5% of soil moisture), and then decreased, while the PCC of A. multidentatum had the minimum (1.91 mg g −1 FW) at 31% of soil moisture, and kept at a higher level afterwards ( Figure 4C). The PCC of A. multidentatum was higher than M. struthiopteris except at 8% and 5% of soil moisture.

Secondary Metabolites
The FC of M. struthiopteris reached the maximum (9.40 mg g −1 FW) at 17% of soil moisture, and then increased from 6.53 (8% of soil moisture) to 9.48 mg g −1 FW after rehydration. The FC of A. multidentatum increased from 5.74 (31% of soil moisture) to 12.23 mg g −1 FW after rehydration ( Figure 4A). Furthermore, the FC of A. multidentatum was significantly higher than M. struthiopteris when the soil moisture was 8%, 5%, 2% and 34% (rehydration) (p ≤ 0.05). The TPC of M. struthiopteris reached the maximum (0.62 mg g −1 FW) when the soil moisture was 2%. There were no significant differences among 31%, 17% and 8% of soil moisture. After rehydration, the TPC decreased to the same level of 31%, 17% and 8% of soil moisture. However, the TPC of A. multidentatum increased from 0.39 (31% of soil moisture) to 0.63 mg g −1 FW (5% of soil moisture), and then decreased, but turned to increase again after rehydration ( Figure 4B). Otherwise, the TPC of A. multidentatum at 17% and 8% of soil moisture was significantly higher than M. struthiopteris (p ≤ 0.05). The PCC of M. struthiopteris increased from 1.63 (31% of soil moisture) to 3.47 mg g −1 FW (5% of soil moisture), and then decreased, while the PCC of A. multidentatum had the minimum (1.91 mg g −1 FW) at 31% of soil moisture, and kept at a higher level afterwards ( Figure 4C). The PCC of A. multidentatum was higher than M. struthiopteris except at 8% and 5% of soil moisture.

Antioxidant Enzyme Activities
The SOD activity of two fern species increased to the maximum (6.29 × 10 6 and 6.33 × 10 6 U g −1 h −1 FW) at 8% of soil moisture, respectively, and then decreased to the minimum (1.03 × 10 6 and 0.45 × 10 6 U g −1 h −1 FW) after rehydration ( Figure 5A). The POD activity of M. struthiopteris decreased from 8.72 × 10 2 U g −1 min −1 FW (31% of soil moisture) to 6.25 × 10 2 U g −1 min −1 FW (5% of soil moisture), and then increased to the maximum (11.10 × 10 2 U g −1 min −1 FW) after rehydration. However, the POD activity of A. multidentatum decreased from 8.32 × 10 2 U g −1 min −1 FW (31% of soil moisture) to 5.00 × 10 2 U g −1 min −1 FW (17% of soil moisture), and then increased to the maximum (12.26 × 10 2 U g −1 min −1 FW) when the soil moisture dropped to 2%. After rehydration, the POD activity of A. multidentatum decreased ( Figure 5B). The CAT activity of M. struthiopteris increased to the maximum (13.39 U g −1 min −1 FW) at 8% of soil moisture, and then continued to decline. The trend of the CAT activity of A. multidentatum was similar to POD activity, which decreased to the minimum (8.32 U g −1 min −1 FW) at 17% of soil moisture, and then increased to the maximum (17.36 U g −1 min −1 FW) at 2% of soil moisture. The CAT activity of A. multidentatum decreased once again after rehydration ( Figure 5C). The APX activity of M. struthiopteris reached its maximum when the soil moisture was 2%, and then decreased after rehydration. The APX activity of A. multidentatum at 5% of soil moisture was significantly higher than the others (p ≤ 0.05), although which increased a little after rehydration compared to the APX activity at 2% of soil moisture ( Figure 5D).

Antioxidant Enzyme Activities
The SOD activity of two fern species increased to the maximum (6.29 × 10 6 and 6.33 × 10 6 U g −1 h −1 FW) at 8% of soil moisture, respectively, and then decreased to the minimum (1.03 × 10 6 and 0.45 × 10 6 U g −1 h −1 FW) after rehydration ( Figure 5A). The POD activity of M. struthiopteris decreased from 8.72 × 10 2 U g −1 min −1 FW (31% of soil moisture) to 6.25 × 10 2 U g −1 min −1 FW (5% of soil moisture), and then increased to the maximum (11.10 × 10 2 U g −1 min −1 FW) after rehydration. However, the POD activity of A. multidentatum decreased from 8.32 × 10 2 U g −1 min −1 FW (31% of soil moisture) to 5.00 × 10 2 U g −1 min −1 FW (17% of soil moisture), and then increased to the maximum (12.26 × 10 2 U g −1 min −1 FW) when the soil moisture dropped to 2%. After rehydration, the POD activity of A. multidentatum decreased ( Figure 5B). The CAT activity of M. struthiopteris increased to the maximum (13.39 U g −1 min −1 FW) at 8% of soil moisture, and then continued to decline. The trend of the CAT activity of A. multidentatum was similar to POD activity, which decreased to the minimum (8.32 U g −1 min −1 FW) at 17% of soil moisture, and then increased to the maximum (17.36 U g −1 min −1 FW) at 2% of soil moisture. The CAT activity of A. multidentatum decreased once again after rehydration ( Figure 5C). The APX activity of M. struthiopteris reached its maximum when the soil moisture was 2%, and then decreased after rehydration. The APX activity of A. multidentatum at 5% of soil moisture was significantly higher than the others (p ≤ 0.05), although which increased a little after rehydration compared to the APX activity at 2% of soil moisture ( Figure 5D).

Discussion
Drought stress has a lot of negative effects on plants, such as inhibiting photosynthesis and growth, destroying the stability of membranes and leading to a series of physiological reactions [50][51][52][53]. In addition, severe drought stress can cause plant deaths and crop yield reduction [15,54]. Previous studies have focused on crops, ornamental plants, etc. [10,15,51,53], however, little research has been done on the drought resistance of wild vegetables. M. struthiopteris and A. multidentatum as important wild vegetables and Chinese medicine in Northeast China, are sensitive to water condition [35,55]. Thus, understanding the drought resistance of these two fern species is important for the cultivation, management and utilization.

The Responses of Growth Traits to Drought Stress
Phenotypic plasticity is crucial in plants' responses to environment stress [56,57]. The plasticity of plants is negatively correlated with the resistance activity to stress [58,59]. RLWC reflects the ability of plant leaves to maintain water balance, and the RLWC of plants with strong drought resistance would decrease with decreasing soil moisture in order to maintain normal physiological functioning of plants [60,61]. In this study, the RLWC of the two fern species showed a downward trend throughout the entire drought stress. The change of RLWC of the two fern species was slow between 31% and 5% of soil moisture, but decreased sharply at 2% of soil moisture. The results were similar to what was reported for Elymus elongates subsp. ponticus (Podp.) Melderis [62], Phragmites australis (Cav.) Trin. ex Steud. [63] and Leucaena leucocephala (Lam.) de Wit [60], indicating that the two fern species have the ability to maintain water content at the early stages of drought, but lost the ability at the late stages of drought [64,65]. Similar to the results of Deng et al. [66], the RLWC of the two fern species increased after rehydration. This change of RLWC indicates that the two fern species have the ability to resist drought to some extent. In addition, the RLWC of A. multidentatum was higher than

Discussion
Drought stress has a lot of negative effects on plants, such as inhibiting photosynthesis and growth, destroying the stability of membranes and leading to a series of physiological reactions [50][51][52][53]. In addition, severe drought stress can cause plant deaths and crop yield reduction [15,54]. Previous studies have focused on crops, ornamental plants, etc. [10,15,51,53], however, little research has been done on the drought resistance of wild vegetables. M. struthiopteris and A. multidentatum as important wild vegetables and Chinese medicine in Northeast China, are sensitive to water condition [35,55]. Thus, understanding the drought resistance of these two fern species is important for the cultivation, management and utilization.

The Responses of Growth Traits to Drought Stress
Phenotypic plasticity is crucial in plants' responses to environment stress [56,57]. The plasticity of plants is negatively correlated with the resistance activity to stress [58,59]. RLWC reflects the ability of plant leaves to maintain water balance, and the RLWC of plants with strong drought resistance would decrease with decreasing soil moisture in order to maintain normal physiological functioning of plants [60,61]. In this study, the RLWC of the two fern species showed a downward trend throughout the entire drought stress. The change of RLWC of the two fern species was slow between 31% and 5% of soil moisture, but decreased sharply at 2% of soil moisture. The results were similar to what was reported for Elymus elongates subsp. ponticus (Podp.) Melderis [62], Phragmites australis (Cav.) Trin. ex Steud. [63] and Leucaena leucocephala (Lam.) de Wit [60], indicating that the two fern species have the ability to maintain water content at the early stages of drought, but lost the ability at the late stages of drought [64,65]. Similar to the results of Deng et al. [66], the RLWC of the two fern species increased after rehydration. This change of RLWC indicates that the two fern species have the ability to resist drought to some extent. In addition, the RLWC of A. multidentatum was higher than M. struthiopteris throughout the drought stress, indicating that A. multidentatum has stronger ability to maintain water content.
SLA is an indicator of plant leaf thickness [67]. Previous studies suggested that the SLA decreased with drought stress, and the drought resistance of plants was negatively correlated with SLA [68]. However, we observed that the SLA increased slowly in general during exposure to drought stress but not significantly, which is similar to the observations by Díaz-Lòpez et al. [69] and Ludewing et al. [70]. Additionally, the SLA of A. multidentatum was higher than M. struthiopteris, suggesting that A. multidentatum has a stronger ability to accumulate more dry mass per unit leaf area than M. struthiopteris under drought stress [44].

The Oxidative Damage of Drought Stress to the Two Fern Species
As a ROS, H 2 O 2 induces different degrees of damage to plant membrane lipids, proteins and DNA [9,15,71]. MDA is the end product of membrane lipid peroxidation, and its accumulation means that plant cells have high levels of lipid peroxides [9,[72][73][74]. Membrane lipid peroxidation causes ion leakage by breaking the permselectivity of membranes [75]. The H 2 O 2 and MDA content of Vitis vinifera L. [9], Amaranthus tricolor L. [44], Achillea species [76], Scutellaria baicalensis Georgi [6] and Cucumis sativus L. [77] increased under drought stress. In our experiment, the H 2 O 2 content of M. struthiopteris increased, while the H 2 O 2 content of A. multidentatum did not increase like M. struthiopteris, which means M. struthiopteris is more sensitive to drought stress than A. multidentatum. However, the MDA content of the two fern species increased during exposure to drought and then decreased after rehydration, which reflected that the two fern species suffered from severe lipid peroxidation under drought. However, the MDA content of M. struthiopteris changed immediately, indicating M. struthiopteris is more sensitive to drought stress [6,44,78,79]. Enhanced IL is regarded to be the sign of destruction of the membrane under water deficit [44,75,78]. In this research, IL increased remarkably with the drought exposure. This indicates that the membrane has been damaged and reflects the increase in MDA and H 2 O 2 content [44,80,81].

The Regulation of Osmotic Regulators to Drought Stress
SP and Pro are important osmotic regulators, which can reduce the water potential of plant cells and maintain tissue moisture to make plants perform normal physiological function. Additionally, Pro also improves the micro-environment in plant tissues [82]. Research on drought stress indicated that plants would accumulate SP and Pro when exposed to drought stress [6,9,44], but some studies showed a different trend [10,83]. Similar to the results of Li et al. [10], in our research, the content of SP in two fern species increased rapidly after exposure to drought stress, indicating that plants accumulate SP immediately to regulate water content [44]. After rehydration, the SP content increased again, which means rehydration was helpful to protein synthesis after drought stress [83]. Although, the change in Pro content was lagging behind, probably because it is not suitable for the water regulation in rapid short-term drought [10].
ABA plays an important role in the response of plants to abiotic stress, such as drought stress and high temperature stress [84,85]. ABA stimulates short-term responses such as stomatal closure, so as to regulate water balance and plants growth under osmotic stress [86][87][88]. Studies have shown that the ABA content increases sharply under drought stress to help plants maintain water status [86,89]. In our study, the ABA content of M. struthiopteris showed a downward trend throughout the exposure to drought stress, but the ABA content of A. multidentatum maintained at a relative stable high level. In addition, the ABA content of the two fern species is negatively correlated with H 2 O 2 content and IL (r = −0.52 and 0.49, p ≤ 0.01, Supplementary Table S1), suggesting that the two fern species can enhance drought tolerance by increased ABA content [15,90]. Furthermore, the ABA content of A. multidentatum was higher than M. struthiopteris, which indicates that A. multidentatum has higher drought resistance by synthesizing ABA in time to regulate the physiological functions under drought stress.

The Effects of the Antioxidant System
Plants' antioxidant systems include an enzymatic antioxidant system and non-enzymatic antioxidant system [6,17,91]. The non-enzymatic antioxidant system contains some secondary metabolites such as glutathione, ascorbic acid, phenols, flavonoids [92,93]. Flavonoids not only interact directly with ROS, but also serve as substrates for different peroxidases [94,95]. Similarly, phenols play an important role in the mitigation of oxidative stress, because they are related to the ROS scavenging [50,96]. Many studies showed that plants would accumulate flavonoids and phenols with decreasing soil moisture, such as Chrysanthemum morifolium L. [97], Gapsicum species [50], Citrus species [98] and Achillea species [76]. High levels of phenols and flavonoids in plants under drought stress means that the plants adapt to the drought environment [99]. In agreement with this, the two fern species had high levels of FC and TPC under drought exposure, which reflected that the two fern species adapt to drought stress by scavenging ROS by FC and TPC [50,100]. In addition, the FC and TPC are positively correlated with MDA content (r = 0.54 and 0.50, p ≤ 0.01, Supplementary Table S1). This may indicate that flavonoids and phenols are involved in the repair of membrane lipid peroxidation. However, the FC and TPC of M. struthiopteris are generally lower than that of A. multidentatum, reflecting the inherently higher non-enzymatic antioxidant system of A. multidentatum.
Proantho cyanidins are secondary metabolites with antioxidant effects, and their accumulation is greatly regulated by the environment. The research about the secondary metabolites of Korean Pine (Pinus koraiensis Sieb. et Zucc.) under drought indicated that the PCC decreased under severe drought stress [45]. However, in our research, the PCC of two fern species increased at the early drought stage. Similarity, the PCC was positively correlated with MDA (r = 0.61, p ≤ 0.01, Supplementary Table S1), which indicates it can improve the drought tolerance of the two fern species [45]. However, there was no significant difference in PCC between the two fern species in the early drought stage. In the late stages of drought, the higher PCC of A. multidentatum reflects the higher non-enzymatic antioxidant system of A. multidentatum. The enzymatic antioxidant system of plants includes SOD, POD, CAT, APX and GR, acting as scavengers for ROS to protect plants from oxidative damage [101,102]. Some studies have shown that plants with stronger drought tolerance have a stronger antioxidant system, which enables plants maintain the dynamic balance of ROS under drought stress [103]. In our research, although the SOD, POD, CAT and APX activity of the two fern species changed differently, the antioxidant system could respond to drought to reduce ROS accumulation and maintain the stability of the membrane over time [15]. However, the SOD activity of the two fern species and CAT activity of M. struthiopteris decreased at the late stages of drought, which are consistent with the findings of Ju et al. [9] and Ji et al. [104], probably because the antioxidant system was impaired under severe drought stress [9,10,99,105,106]. The POD and CAT activity of A. multidentatum increased rapidly for drought stress, while the POD activity of M. struthiopteris changed slowly, indicating that A. multidentatum activates the POD and CAT rapidly to scavenge H 2 O 2 [10,16,107].

Conclusions
In the present study, the drought resistance of two fern species in Northeast China was compared based on growth traits, secondary metabolites and enzyme activities. The two fern species displayed stronger drought resistance features at the early stages of drought. In addition, the growth, osmotic regulators and antioxidant system reacted immediately to drought stress. However, in the late stages of drought, the two fern species did not express stronger drought resistance. A. multidentatum showed higher drought resistance, mainly expressed as higher ABA content and non-enzymatic antioxidant system. In addition, our results suggested that secondary metabolites may play a preponderant role in the two fern species in drought resistance. Future studies may rely on these results to further study the relationship between secondary metabolites and drought resistance of fern species.
Author Contributions: Y.W., W.C. and X.H. conceived the study; Y.W. and S.G. performed the experiment; Y.L., P.L. and Y.Z. contributed reagents and materials; Y.W. and S.G. analyzed the data; Y.W. led the writing of the paper with substantial input from all co-authors.