Application of Morphological and Physiological Markers for Study of Drought Tolerance in Lilium Varieties

Abstract

The shortage of water resources is an unfavourable factor that restricts the production of flowers. The use of drought-resistant morphological markers is of great significance to distinguish the drought resistance of flower varieties. In this paper, we study the difference in drought tolerance of seven common lily varieties in the flower market by morphological and physiological markers. The results showed that there were differences in leaf morphological indices and anatomical structures among the seven varieties. Drought reduced the chlorophyll content, inhibited the photosynthetic rate, and increased catalase (CAT), peroxidase (POD), superoxide dismutase (SOD), malondialdehyde (MDA), proline, soluble sugar, and soluble protein. After rewatering, the activities of CAT, POD, and SOD of ‘Lyon’, ‘Royal Sunset’, and ‘Robina’ varieties decreased, which was opposite to the varieties of ‘Immaculate’, ‘Elena’, ‘Siberia’, and ‘Gelria’. According to the membership function value of physiological indices, the drought resistance of seven lily varieties from weak to strong was ‘Immaculate’, ‘Elena’, ‘Siberia’, ‘Gelria’, ‘Robina’, ‘Royal Sunset’, and ‘Lyon’. Drought resistance is related to the thickness of leaves, palisade tissue, sponge tissue, and specific leaf area. Lily leaf structure can be used as one of the indices to judge drought resistance.

1. Introduction

Global climate change intensifies the problem of drought and could expand the scope of drought in subtropical arid areas [1]. The loss caused by drought is equivalent to 60% of the loss associated with all climate disasters. Due to water shortages, the growth and development of plants in arid and semiarid areas are restricted [2], and the physiological and biochemical characteristics of plants are changed [3]. Lilies (Lilium spp.) are perennial herbaceous bulbous flowers of the family Liliaceae (Lilium). Lilies are not only famous due to their beautiful flowers but also their wide use as potted flowers and in landscaping all over the world [4]. Abiotic stress is an important factor affecting the growth and development of the lily. Abiotic stress affects the yield and ornamental quality and restricts the application of lilies in open fields [5,6]. Drought and water shortages often occur in the main lily-producing areas. Water shortage directly affects plant growth and development and then affects the quality of cut flowers [7]. The shortage of water resources has become a serious ecological problem, restricting the development of global flower production.

In response to drought stress, there could be differences between interspecific or even intra-specific differences. Understanding the response of different species to water deficit conditions will help to identify drought-resistant morphological markers and facilitate the identification of drought-sensitive and drought-tolerant species [8]. Drought-induced morphological and physiological changes can be used to find drought-resistant genotypes or develop new flower varieties to improve drought productivity [9]. Leaves are the main organs of photosynthesis and transpiration in plants, and they are also the most sensitive organs of plants to drought stress. In the process of adapting to the external environment, the anatomical characteristics of the leaf structure are related to the drought resistance ability, such as the thickness of the cuticle on the epidermis, the developed transport tissue, and the smaller stomata [10]. Therefore, drought resistance among plants has a certain relationship with leaf morphology and anatomical structure [11]. Plants use the morphological characteristics of drought adaptation to ensure maximum water absorption under drought conditions. The leaf phenotype can be used to preliminarily judge the drought resistance adaptability of the variety [12].

When the plant is under drought stress, the root system will quickly generate chemical signals to transmit to the above-ground part, causing the stomata to close to reduce water loss. The plant shrinks the stomata to prevent more water loss. Stomatal closure will reduce the transpiration rate, limit CO₂ absorption and transmission, and reduce the net photosynthetic rate of the plant [13]. The change in chlorophyll content indicates the process of photooxidation and degradation and reflects the effect of drought stress on photosynthesis. The degree of inhibition of photosynthesis in plant leaves by drought stress reflects the strength of drought resistance [14]. When photosynthesis is inhibited, the supply of organic matter required for the development of plant reproductive organs will be reduced, which will hinder reproductive development [15]. Under the influence of drought stress, plant cells, tissues and organs accumulate reactive oxygen species (ROS), and the activities of catalase (CAT), superoxide dismutase (SOD), and peroxidases (PODs) are enhanced, which helps to remove ROS and reduce electrolyte leakage and lipid peroxidation to maintain the vitality and integrity of organelles and cell membranes. An effective antioxidant system is important to provide drought tolerance [16]. The balance of water metabolism is broken, resulting in dehydration of the cell protoplasm and a decline in water potential, resulting in the oxidative modification of proteins, lipids, and DNA and damaging normal cell functions. Proline and soluble sugar, as widely distributed important osmotic regulatory substances, can prevent damage to plants caused by drought stress by maintaining cell turgor [6]. The concentration of protein and soluble sugar is considered a general indicator of drought tolerance [17].

Plant drought resistance strategies are important in morphology, physiology, and biochemistry. The evaluation of plant drought resistance mainly considers the molecular as well as epigenetic levels of drought resistance and drought resistance recovery [18]. The study of the stress/recovery response helps us to better understand the ability of plants to adapt to different environmental and climatic conditions [17]. Stress recovery is an important part of the plant response. In the special case of drought, the carbon balance depends not only on the rate and degree of photosynthesis decline under water depletion but also on the capacity supply of photosynthesis restoration after water depletion. After drought and rewatering, plants can quickly resume growth, eliminate the inhibition of drought on plant growth, and sometimes even have a super compensation effect to compensate for the loss of plants caused by drought [19]. It is necessary to have more direct indicators for evaluation of flower production. Only by better understanding the differences in the main adaptation mechanisms of these lily varieties to cope with drought stress can we determine the key adaptation traits to distinguish their drought resistance. However, the application of morphological and physiological markers to evaluate drought tolerance of lily is very few.

The drought resistance membership function method is a better comprehensive evaluation method of drought resistance. The greater the average value is, the stronger the drought resistance. Membership function analysis provides a way to comprehensively evaluate the drought resistance of plants based on multiple determination, which can avoid the one-sidedness of a single index and improve the reliability and accuracy of drought resistance identification. The drought resistance membership function method is widely used to evaluate the drought resistance of iris [20], citrus [21], and other flowers, as well as abiotic stresses such as heavy metals [22], temperature [23], and salinity [24,25]. At present, few studies have compared and evaluated the drought resistance of several lily varieties using leaf morphological structure and physiological indicators as markers [6]. We selected seven common lily varieties in the flower market as plant materials and used pot experiments to determine the differences in leaf morphology and structure of different lily varieties. According to the changes in photosynthesis, the antioxidant system and physiological indices of lily leaves from drought stress to rewatering, we determined the drought-resistance level of seven lily varieties. We hypothesize that the difference in the leaf structure of the lily can reflect the drought resistance of varieties, which was consistent with the level of drought resistance reflected by physiological indicators. So, the results of this study can provide a reference for the selection of lily varieties in areas with a relative lack of water resources and can also provide a reference for the evaluation of drought resistance of other plants.

2. Materials and Methods

2.1. Test Site and Plant Materials

The experiment was conducted in the lily breeding experimental base (25°7′33″ N, 102°45′48″ E, 1951.1 m above sea level) of the flower Research Institute of Yunnan Academy of Agricultural Sciences, China, from May 2020 to August 2020. The Seven different lily varieties used in the experiment were: Oriental Hybrids ‘Siberia’, Pollen Abortion Cultivar ‘Immaculate’, Longiflorum Hybrid ‘Gelria’, Double Petal Cultivar ‘Elena’, Oriental Trumpet (OT) Hybrid ‘Robina’, Longiflorum Asiatic (LA) Hybrid ’Royal Sunset’, and Asiatic Hybrid ’Lyon’. The test materials were healthy in appearance, no plant diseases and insect pests, and the specifications were consistent specifications (bulb perimeter diameter 12~14 cm) after vernalization treatment (Figure 1). Cultivated in a greenhouse and exposed to a natural photoperiod (from May to August), the cultivation soil was sandy soil, with field moisture capacity (FMC) of 23.20%, organic matter content of 5.34%, pH 5.66, total nitrogen of 0.18%, total phosphorus of 0.08%, total potassium of 1.93%, alkali hydrolysable nitrogen of 139.92 mg·kg⁻¹, available phosphorus of 81.28 mg·kg⁻¹, available potassium of 112.68 mg·kg⁻¹, available iron of 421.55 mg·kg⁻¹, exchangeable calcium of 988.41 mg·kg⁻¹, and exchangeable magnesium of 110.70 mg·kg⁻¹. The greenhouse temperature was 14~25 ℃, the relative humidity was 55~70%, and the greenhouse CO₂ concentration was 389 μmol·mol⁻¹.

2.2. Experimental Design

Lily bulbs were cultivated in boxes (0.6 m × 0.4 m × 0.25 m, length × width × height). The row spacing was 12 cm, the plant spacing was 15 cm, and 9 plants were planted in each box. Nine boxes were planted for each variety. One box was used for leaf morphological analysis under natural conditions, and the other 8 boxes were used to analyse the changes in leaf physiological indices under drought stress. When the plant grew to the budding stage and the top growth stopped, the morphological indices were measured, and drought stress began. The samples were collected at 0, 5, 10, 15, 20, 25, and 30 days of drought stress and 5 days (35 days) of continuous rehydration after 30 days.

2.3. Soil Water Status

The gravimetric water content (GWC) was measured by the commonly used drying and weighing method [26]. The soil water content was monitored 0–10 cm away from the soil surface of each pot. See

Table 1. Changes in gravimetric water content (GWC) and relative water content (RWC).

	0 d	5 d	10 d	15 d	20 d	25 d	30 d	35 d
GWC (%)	19.67	14.86	13.11	10.12	8.98	7.06	5.67	20.34
RWC (%)	84.78	64.05	56.51	43.62	38.71	30.43	24.44	87.67

for the GWC and relative water content (RWC) of each treatment.

$$\mathrm{GWC}=\frac{\mathrm{M}-{\mathrm{M}}_{\mathrm{S}}}{{\mathrm{M}}_{\mathrm{S}}}\times 100\%$$

(1)

$$\mathrm{RWC}=\frac{\mathrm{GWC}}{\mathrm{FMC}}\times 100\%$$

(2)

where M: original soil weight; M_S: dry soil weight; FMC: field moisture capacity.

2.4. Determination of Leaf Morphological Indices

In the first box of lily varieties used for morphological index measurement, all leaves of 9 plants of each variety were collected for total leaf area calculation. The sixth fully expanded leaf from the top branch point of 9 plants was selected for single leaf length, width, area, dry weight, and specific leaf area (SLA).

$$\mathrm{SLA}=\frac{\mathrm{Leaf}\mathrm{area}}{\mathrm{Leaf}\mathrm{dry}\mathrm{weight}}$$

(3)

Five plants were selected to collect the seventh fully expanded leaf from the top branch point for paraffin section observation. The paraffin sections were placed into 1% safranine dye solution for staining for 1–2 h, washed with distilled water, and decolorized with 50%, 70%, and 80% gradient alcohol for 60 s. Then, the sections were placed into the 0.5% solid green dye solution for staining for 30–60 s, decolorized with anhydrous ethanol for 90 s, placed into a 60℃ oven for drying, and cleared with xylene for 5 min. The sections were observed and imaged under an optical microscope (YS100, Nikon, Tokyo, Japan).

2.5. Determination of Photosynthetic Index

Among the 8 drought stress treatment groups of each variety, the plants to be tested were induced by light at 8:30~11:30 a.m. under natural light conditions. Three plants with strong growth, relatively consistent growth, and small fluctuations in photosynthesis were selected to measure the net photosynthetic rate (Anet), transpiration rate (E), stomatal conductance (gs), and intercellular CO₂ concentration (Ci) of the sixth fully expanded leaf from the growth point by the Li-6400 photosynthetic determination system (LI-COR, Lincoln, NE, USA). The readings were repeated 3 times, and the average value was taken (air chamber temperature 14~22 ℃, external CO₂ concentration 389 μmol CO₂ mol⁻¹, air velocity in air chamber 500 m·s⁻¹) [27].

2.6. Determination of Physiological Index

In the 8 drought stress treatment groups of each variety, samples were taken from 9:00~10:00 a.m., and the seventh to twelfth leaves from the growth point were sampled completely, and each 3 leaves were mixed into a sample. The content of chlorophyll a (Chla) and b (Chlb) in leaves was determined according to the Arnon method [28]. The activity of SOD was determined by the nitrogen blue tetrazole method, with 50% inhibition of photochemical reduction in nitroblue tetrazolium as an enzyme activity unit (U); the activity of POD was measured by guaiacol colorimetry, and the change in optical density (d470 nm) at 470 nm per minute was 0.1, which is an enzyme activity unit (U); the activity of CAT was determined by the ultraviolet (UV) absorption method, and the decrease of 240 by 0.1 within 1 min was taken as an enzyme activity unit (U) [29]. The content of MDA was determined by two-component spectrophotometry [30]. The content of soluble sugar was determined by anthrone colorimetry [31]. The content of free proline was determined by the acid ninhydrin method [32]. Determination of protein content was performed by the Bradford method [33]. A UV-visible spectrophotometer (UV-5800, Shanghai Metash instruments Co., Ltd, Shanghai, China) was used to complete the index determination.

2.7. Data Processing and Statistical Analysis

The data were sorted with Excel 2021 and plotted with Origin 9.0. The Duncan test was used with the statistical software SPSS 26 (SPSS Inc., Chicago, IL, USA) to test the significance of differences between treatments (p < 0.05 level). Canonical discriminant analysis (CDA) and correlation analysis were carried out. The results of morphological and physiological indices were expressed as the mean ± standard deviation (SD). Based on determining several drought resistance evaluation indices, the photosynthetic, physiological, and antioxidant index data of lily leaves were quantitatively converted through subordination function analysis (SFA), and the drought resistance of seven lily varieties was comprehensively evaluated:

X(µ) = (X − X_min)/(X_max − X_min)

(4)

X(µ) = 1 − (X − X_min)/(X_max − X_min)

(5)

where X(µ) is the subordination function value; X is the relative value of an index under drought treatment; X_max is the maximum value in the index; and X_min is the minimum value in this indicator. If a certain index is positively correlated with drought resistance, it shall be calculated with Formula (3); if there is a negative correlation, Formula (4) will be used to calculate.

3. Results

3.1. Leaf Morphology and Anatomical Structure of Seven Lily Species

There were differences in leaf length, leaf width, single leaf area, and total leaf area among the different lily varieties (Figure 2,

Table 2. Leaf characteristics of seven lily varieties.

Varieties	Leaf Length (cm)	Leaf Width (cm)	Single Leaf Area (cm2)	Specific Leaf Area (cm2·g−1)	Total Leaf Area (cm2)
‘Siberia’	10.71 ± 0.50 d	2.27 ± 0.05 d	16.33 ± 0.53 c	114.46 ± 20.63 ab	1148.30 ± 54.22 c
‘Immaculate’	10.79 ± 0.58 c	2.49 ± 0.06 c	18.56 ± 0.50 b	123.01 ± 9.82 ab	1236.83 ± 43.93 b
‘Gelria’	10.55 ± 0.62 d	1.41 ± 0.03 e	11.58 ± 0.80 d	129.48 ± 38.72 a	817.86 ± 27.39 e
‘Elena’	10.79 ± 0.58 d	2.55 ± 0.07 b	17.55 ± 0.68 c	117.04 ± 18.87 ab	1209.21 ± 49.55 b
‘Robina’	14.28 ± 0.65 a	2.61 ± 0.09 a	25.61 ± 0.50 a	103.72 ± 22.06 c	1826.84 ± 49.09 a
‘Royal Sunset’	12.44 ± 0.67 b	1.22 ± 0.04 f	11.30 ± 0.46 d	101.58 ± 15.75 c	1138.22 ± 60.03 c
‘Lyon’	9.37 ± 0.52 e	0.85 ± 0.05 g	6.38 ± 0.33 e	81.31 ± 15.03 d	992.79 ± 21.23 d

Values are the average ± standard error. Different letters within the same column indicate significant differences between treatments (p < 0.05, n = 10).

). The maximum difference multiples of leaf length, leaf width, single leaf area, and total leaf area among varieties were 1.5, 3.1, 4.0 and 2.2 times, respectively. ‘Robina’ had significantly higher leaf length, leaf width, single leaf area, and total leaf area than the other varieties, ‘Lyon’ had the lowest leaf length, leaf width, and single leaf area, and ‘Gelria’ had the lowest total leaf area. The specific leaf area of ‘Lyon’ was significantly lower than that of other lily varieties.

The leaf thickness of the 7 lily species was 330.04–638.25 μm (Figure 3,

Table 3. Anatomical structure parameters of leaves of seven lily varieties.

Varieties	Blade Thickness (μm)	Palisade Tissue Thickness (μm)	Spongy Tissue Thickness (μm)	Palisade Tissue/Spongy Tissue	Upper Epidermis Thickness (μm)	Lower Epidermis Thickness (μm)
‘Siberia’	452.65 ± 11.24 c	61.4 ± 4.29 c	306.61 ± 10.53 b	0.20 ± 0.01 cd	66.13 ± 2.39 b	46.23 ± 3.05 bc
‘Immaculate’	446.88 ± 11.38 c	46.44 ± 3.67 d	298.15 ± 12.54 b	0.16 ± 0.01 e	65.15 ± 3.36 b	49.57 ± 2.32 ab
‘Gelria’	330.04 ± 9.86 e	61.5 ± 4.19 c	182.57 ± 6.91 d	0.34 ± 0.04 a	23.87 ± 1.29 c	36.08 ± 2.81 d
‘Elena’	407.86 ± 9.69 d	53.3 ± 4.48 cd	254.80 ± 8.14 c	0.21 ± 0.02 c	63.27 ± 3.68 b	34.98 ± 1.87 d
‘Robina’	484.25 ± 8.47 b	86.14 ± 3.88 a	280.81 ± 11.17 bc	0.31 ± 0.02 b	77.10 ± 3.57 a	42.74 ± 1.92 c
‘Royal Sunset’	615.74 ± 12.28 a	77.80 ± 5.41 ab	437.49 ± 12.23 a	0.18 ± 0.02 de	64.46 ± 3.18 b	37.97 ± 3.15 d
‘Lyon’	638.25 ± 11.97 a	75.83 ± 3.96 b	453.48 ± 11.96 a	0.17 ± 0.00 e	77.98 ± 2.38 a	51.68 ± 4.85 a

Values are the average ± standard error. Different letters within the same column indicate significant differences between treatments (p < 0.05, n = 5).

). The leaf thickness of ‘Lyon’ and ‘Royal Sunset’ was significantly higher than the leaf thickness of the other varieties, and ‘Gelria’ had the largest thickness of palisade tissue. A very interesting observation for us is that the mesophyll thickness is not in line with leaf dimensions for ‘Lyon’. The thickness of spongy tissue varied from 182.57 to 453.48 μm. ‘Lyon’ and ‘Royal Sunset’ sponges have significantly higher tissue thicknesses than other varieties, and ‘Gelria’ sponges have the smallest tissue thickness. The maximum palisade tissue/spongy tissue is ‘Gelria’. ‘Lyon’ and ‘Robina’ have the largest thickness of the upper epidermis. The thickness of the upper epidermis of ‘Gelria’ was significantly lower than the thickness of the upper epidermis of other varieties. Except for ‘Gelria’, the thickness of the upper epidermis was greater than the thickness of the lower epidermis.

3.2. Effects of Drought Stress on Photosynthesis in Lily Leaves

With the extension of drought stress treatment time, the content of Chla and Chlb showed a downward trend (Figure 4). After 30 days of drought stress, the Chla of the seven lily varieties decreased by 16.6%, 16.4%, 13.7%, 16.7%, 13.2%, 12.7%, and 12.0%, and the Chlb content decreased by 27.5%, 32.2%, 24.1%, 32.0%, 26.3%, 21.5%, and 19.8%. The chlorophyll content of ‘Lyon’ was the lowest under drought stress. The chlorophyll content of ‘Elena’, ‘Siberia’, and ‘Immaculate’ was greatly affected by drought stress. Gs, E, and Anet in leaves of seven varieties decreased gradually with the extension of stress time; Ci increased gradually with the extension of stress time, and showed the opposite trend after rewatering (Figure 5). In general, 10 days of drought stress had little effect on the photosynthetic parameters of lily leaves. The variation ranges of the above parameters varied with different varieties, and they changed significantly on the 15th day. The photosynthetic parameters of ‘Royal Sunset’ and ‘Lyon’ changed significantly after 20 days of drought stress. On the 30th day, the gs of the leaves of the seven lily varieties decreased by 74.3%, 78.3%, 66.0%, 76.7%, 60.1%, 50.3%, and 21.8%, respectively. After re-watering, ‘Robina’, ‘Royal Sunset’, and ‘Lyon’ had the best recovery effects on the net photosynthetic rate, increasing by 56.0%, 43.9%, and 42.6%, respectively. A very interesting phenomenon is that the net photosynthetic rate of ‘Royal Sunset’, ‘Lyon’, and ‘Robina’ recovers faster than other varieties after rehydration.

3.3. Effects of Drought Stress on Antioxidant Enzyme Activity in Lily Leaves

Under long-term drought stress, the activities of CAT, POD, and SOD in lily leaves showed an increasing trend (Figure 6). After 10 days of stress treatment, there was no significant difference in the activity of CAT and SOD in lily leaves of all varieties (p > 0.05). After 15 days of treatment, the activity of CAT, POD, and SOD in seven lily varieties showed an increasing trend. With the extension of stress time, the activity of CAT, POD, and SOD in ‘Royal Sunset’ and ‘Lyon’ continued to increase, and the content of MDA did not change significantly. On the 20th day, the activity of antioxidant enzymes in lily leaves of ‘Lyon’ was significantly higher than the activity of other varieties. The CAT, POD, and SOD activities of ‘Siberia’, ‘Immaculate’, ‘Gelria’, and ‘Elena’ decreased significantly on the 25th day, and the MDA content increased by 141%, 176%, 136%, and 161%, respectively, on the 25th day. After 30 days of stress treatment, the antioxidant enzyme activity of ‘Robina’, ‘Royal Sunset’, and ‘Lyon’ leaves were significantly higher than the antioxidant enzyme activity of other varieties. The CAT, POD, and SOD activity of ‘Siberia’, ‘Immaculate’, ‘Gelria’, and ‘Elena’ decreased compared with the CAT, POD, and SOD activity of the 25th day. After 35 days of re-watering treatment, the CAT, POD, and SOD activity of these four varieties increased, and the MDA content decreased. In contrast, ‘Robina’, ‘Royal Sunset’, and ‘Lyon’ decreased their antioxidant enzyme activity after stress recovery.

3.4. Effects of Drought Stress on the Physiological Metabolism of Lily Leaves

After drought stress, the content of free proline, soluble sugar, and soluble protein in lily leaves increased to varying degrees (Figure 7). Within 10 days of drought stress treatment, the content of soluble sugar in ‘Lyon’ leaves was significantly higher than the content of soluble sugar of other varieties. The content of free proline in ‘Siberia’, ‘Immaculate’, ‘Gelria’, ‘Elena’, and ‘Robina’ increased by 49.9%, 45.2%, 37.2%, 46.3%, and 35.9%, respectively, after 15 days of stress treatment. The content of free proline and soluble protein in ‘Siberia’, ‘Immaculate’, ‘Gelria’, and ‘Elena’ was significantly higher than the content of free proline and soluble protein in ‘Robina’, ‘Royal Sunset’, and ‘Lyon’. The content of free proline in ‘Lyon’ leaves was significantly higher than the content of free proline of other varieties after 30 days of drought stress treatment. After stress recovery, the content of free proline, soluble sugar, and soluble protein in the leaves of the seven lily varieties decreased to varying degrees. After rehydration, the content of proline and soluble protein decreased significantly.

3.5. Evaluation of Drought Resistance in Lily Varieties

After 30 days of drought stress treatment, the physiological, photosynthetic, and antioxidant systems were affected to the greatest extent. These time point data can reflect the drought tolerance of lily. Rewatering treatment (35 d) could reflect the recovery ability of lily under drought stress. First, the specific subordinate values of each drought resistance index in each variety were obtained, and then the subordinate drought resistance values of each index of the specified variety were accumulated. The average value was obtained to evaluate the stress resistance, and the drought resistance strength was determined according to the average value of each variety. The drought resistance of lily leaves at 30 d and 35 d was evaluated by the indices CAT, POD, SOD, MDA, proline, soluble sugar, soluble protein, Chla, Chlb, gs, E, Ci, and Anet. According to the SFA value, the drought resistance of seven lily varieties was ranked from strong to weak as ‘Lyon’, ‘Royal Sunset’, ‘Robina’, ‘Gelria’, ‘Siberia’, ‘Elena’, and ‘Immaculate’ (

Table 4. Membership function value and drought resistance ranking of lily.

	Days	Siberia	Immaculate	Gelria	Elena	Robina	Royal Sunset	Lyon
Chla	30	0.209	0.148	0.401	0.177	0.479	0.464	0.697
	35	0.256	0.213	0.390	0.210	0.524	0.675	0.815
Chlb	30	0.366	0.202	0.497	0.198	0.495	0.633	0.731
	35	0.396	0.225	0.494	0.266	0.566	0.703	0.846
CAT	30	0.220	0.104	0.362	0.139	0.801	0.811	0.918
	35	0.434	0.186	0.373	0.219	0.452	0.533	0.820
SOD	30	0.042	0.087	0.108	0.107	0.707	0.806	0.941
	35	0.216	0.085	0.198	0.108	0.324	0.695	0.811
POD	30	0.066	0.030	0.200	0.048	0.691	0.741	0.918
	35	0.280	0.075	0.401	0.124	0.693	0.717	0.866
MDA	30	0.508	0.067	0.632	0.168	0.652	0.977	0.984
	35	0.503	0.035	0.588	0.193	0.625	0.978	0.978
gs	30	0.117	0.029	0.219	0.048	0.465	0.478	0.943
	35	0.302	0.169	0.088	0.089	0.924	0.144	0.778
E	30	0.444	0.171	0.496	0.217	0.757	0.853	0.847
	35	0.548	0.173	0.380	0.119	0.929	0.712	0.770
Ci	30	0.190	0.064	0.413	0.162	0.432	0.503	0.852
	35	0.358	0.069	0.176	0.084	0.850	0.744	0.909
Anet	30	0.353	0.269	0.527	0.311	0.619	0.624	0.794
	35	0.198	0.139	0.180	0.145	0.797	0.603	0.675
Proline	30	0.227	0.254	0.162	0.317	0.304	0.585	0.879
	35	0.178	0.085	0.408	0.174	0.469	0.590	0.845
Soluble protein	30	0.202	0.284	0.289	0.258	0.513	0.705	0.854
	35	0.384	0.290	0.399	0.172	0.580	0.767	0.868
Soluble sugar	30	0.432	0.454	0.503	0.397	0.645	0.869	0.159
	35	0.668	0.483	0.441	0.626	0.709	0.867	0.456
Average membership function		0.311	0.169	0.359	0.195	0.615	0.684	0.806
Rank		5	7	4	6	3	2	1

).

3.6. Canonical Discriminant Analysis

After treatment at different time points, all physiological levels of the tested lily varieties were analysed by canonical discriminant analysis (CDA) to verify the drought resistance of the seven lily varieties (Figure 8). According to the CDA results, a total of six discriminant functions (DFSs) were identified, and the first two DFSs (DF1 and DF2) could clearly identify the differences between them (Figure 8, DF1 and DF2 accounted for 55.1% and 31.8% of the independent variable information in discriminant analysis, respectively). In general, under drought stress treatment and rewatering treatment, combined with the levels of key traits, including CAT, SOD, POD, MDA, proline, soluble sugar, soluble protein, Chla, Chlb, gs, E, Ci, and Anet, the horizontal distribution position of drought tolerance of seven lily varieties analysed by CDA was consistent with the results of the SFA.

3.7. Correlation Analysis

The results showed that leaf thickness, palisade tissue, and sponge tissue thickness were positively correlated with antioxidant enzyme activity, photosynthesis, proline, and soluble protein and negatively correlated with MDA and Ci (Figure 9). The correlation between other leaf structure indices and drought resistance was low. MDA, Ci, and SLA were negatively correlated with antioxidant enzyme activity, photosynthesis, proline, and soluble protein.

4. Discussion

Leaves are a manifestation of the long-traits adaptation of plants to the environment. Leaf morphology and structure are the organs with the greatest plasticity to adapt to environmental changes in the process of evolution, reflecting the characteristics of water use. It is necessary for plants to maintain a balance between water absorbed by roots and transpiration of leaves. Drought response and resistance is one of the main limiting factors of plant growth and development. Plant response and resistance to drought stress are a combination of complex biological processes at the cellular, physiological, and biochemical levels and the whole plant level [34,35]. Drought-resistant plants can adapt to water deficient environment through morphological adaptation, physiological regulation, and molecular signals. The response to drought stress is a process from “adaptation” to “injury” [13]. Rewatering after a certain degree of drought stress will have compensatory effects on photosynthesis, antioxidation, and physiology [36].

In the process of adapting to the arid environment, the plants growing in an arid area for a long time have formed special morphological structures, including the thickening of cuticle, the development of palisade tissue, the shrinking of cells, and the subsidence of stomata. The SLA reflects the ability of plants to obtain resources. The plants with low SLA value can better adapt to resource-poor and arid environments [37]. This study found that the change in the SLA of seven lily varieties is basically consistent with the drought resistance. ‘Lyon’ showed strong resistance among the seven lily varieties. ‘Gelria’ and ‘Robina’ have the largest ratio of palisade tissue/spongy tissue, indicating that the palisade tissue is developed, which is related to variety differences. However, due to their thin leaves, drought resistance is weakened, so the ratio of palisade tissue/spongy tissue cannot be used as an indicator to evaluate the drought resistance of the lily [38]. The leaf thicknesses of ‘Royal Sunset’ and ‘Lyon’ are significantly higher than the leaf thicknesses of other lily varieties, which is consistent with their drought resistance. Narrow and thick leaves and developed palisade tissue can prevent excessive transpiration and reduce leaf surface temperature, which is a sign of strong drought resistance. The fence organization ensures the transportation and maintenance of water and nutrients [39]. ‘Lyon’ has thick leaves, a small single leaf area, and well-developed palisade tissue and sponge tissue, showing stable characteristics of drought adaptation in morphology and anatomy. These morphological characteristics have a surface area that can effectively prevent evaporation and reduce evaporation and have important ecological significance for reducing water loss and maintaining water balance in plants.

Drought stress affects the normal physiological and metabolic processes of plants. Mesophyll is the main part of photosynthesis in leaves. This study found that drought stress causes the loss of chlorophyll in lily leaves. Water deficit caused the loss of water in cells and chloroplasts, and the activity of enzymes involved in carbon fixation in chloroplasts was inhibited, thus affecting the synthesis of chlorophyll. This change can also be used as an adaptive expression of the lily. By reducing the light absorption, reactive oxygen free radicals formed by excess excitation energy could be prevented from further damaging the photosynthetic system [40].

The decrease in chlorophyll in lily varieties with strong drought resistance was low, which may be related to leaf structure. Chlorophyll is distributed mainly in the chloroplasts of mesophyll cells. The mesophyll cells of the lily differentiate into palisade tissue and sponge tissue. The palisade tissue is close to the upper epidermis and mainly carries out photosynthesis. The thickness of palisade tissue and spongy tissue was positively correlated with chlorophyll content and photosynthetic rate. Under drought stress, the membrane system in plant cells, including the membrane structure related to photosynthesis, is destroyed due to the lack of water, nutrients, and energy, resulting in the interruption of physiological processes. These factors may directly or indirectly affect the chlorophyll content. Chlorophyll content is reported to decrease significantly with decreasing soil water content [41]. Plants that can maintain high chlorophyll content under drought stress are considered able to use light energy more effectively, so they are considered to have enhanced drought resistance [42]. The results showed that the chlorophyll content of ‘Lyon’ decreased the least and that photosynthesis was inhibited the least. Through the change range of Chla and Chlb under drought stress, the change range of Chlb under drought stress was found to be large. The change range of Chlb content can be used to evaluate the degree of inhibition of lily photosynthesis. The decrease in stomatal conductance is often the main reason for the decrease in photosynthesis under drought stress [43].

Drought stress can be divided into two types: stomatal limitation and nonstomatal limitation [44,45]. Under mild drought (0–10 d), all lily varieties showed a decrease in stomatal conductance, transpiration rate, and intercellular CO₂ concentration and no change in antioxidant enzyme system activity, cell osmotic substances, or malondialdehyde content, indicating that the lily was under mild drought stress at this stage (Figure 10). There are gaps between mesophyll cells to form a ventilation system, which exchanges gas with the outside world through pores on the epidermis [46]. The decrease in the photosynthetic rate was caused by the decrease in stomatal factors.

Under moderate drought (15–20 days), the antioxidant enzyme system activity of lily varieties began to increase, preventing the accumulation of ROS in cells and protecting cells from excessive ROS. Under normal circumstances, the production and clearance of ROS in cells are in a dynamic equilibrium state. At the same time, stimulated by the signal of reactive oxygen species, the content of cell osmotic substances began to increase, such as soluble sugars, Pro, and soluble protein, improving the drought resistance of plants [47]. The same results were also found in the study of walnut (Juglans regia L.) drought resistance. Walnut varieties with strong drought resistance contain less starch in the form of soluble sugar, and the content of free proline increases with the degree of drought stress [48]. At this stage, the stomatal conductance decreased, the transpiration rate decreased, and the Ci of ‘Gelria’, ‘Siberia’, ‘Elena’, and ‘Immaculate’ began to rise, indicating that the photosynthetic rate of the above varieties decreased and began to change from stomatal factors to nonstomatal factors. At the initial stage of drought stress, the photosynthetic rate of the leaves of the seven varieties of lily did not change significantly. The stomatal conductance of leaves decreased significantly, and Ci increased. At this time, the decline in photosynthesis was caused by the decrease in stomatal conductance.

Under severe drought (25–30 d), the antioxidant enzyme activities of ‘Gelria’, ‘Siberia’, ‘Elena’, and ‘Immaculate’ began to decrease, resulting in a reduction in the ability to remove ROS in cells, the accumulation of ROS, the oxidation of intracellular lipids, the disintegration of chloroplasts and an increase in MDA accumulation. Due to the imbalance between the production and utilization of electrons under drought stress, plants produce ROS in chloroplasts, peroxisomes, mitochondria, endoplasmic reticulum, plasma membrane, and the cell wall. When plants suffer from drought stress, the dynamic balance is broken, excessive accumulation of ROS damages cells, and oxidative deterioration may eventually lead to cell death [49]. Under ROS stress, the spatial configuration of various membrane proteins or enzymes is disturbed, resulting in increased membrane permeability and ion leakage, chlorophyll destruction, metabolic disorder, and even serious injury or death of plants [50]. ROS attack the most sensitive biological macromolecules in plant cells; induce lipid peroxidation, protein carbonylation, and DNA damage; and damage the plant’s functions, leading to a disastrous series of events. To protect cells from the harmful effects of excessive ROS, plants have evolved a series of complex enzymatic and nonenzymatic antioxidant defence mechanisms to maintain the homeostasis of the intracellular redox state. At this stage, the reduction in the photosynthetic rate was transformed into nonstomatal factors. In comparison, ‘Lyon’, ‘Royal Sunset’, and ‘Robina’ continued to increase the activity of antioxidant enzymes in cells, with less ROS accumulation and slower chloroplast disintegration, and the intercellular CO₂ concentration began to rise, indicating that the photosynthetic rate of these varieties decreased at this stage and began to change from stomatal factors to nonstomatal factors. Under the stimulation of ROS, the content of osmotic substances in the cells of all varieties continued to increase, and the drought resistance of plants increased. Drought stress leads to damage to the cell membrane system, and the increase in MDA content reflects the degree of damage caused by drought to cells. Although ‘Robina’ has a large leaf area, its strong drought resistance may be related to its triploidity. Research shows that under moderate and severe drought stress, the utilization of water by triploids is better than the utilization of water by diploids [51]. ‘Robina’ and ‘Immaculate’ have similar leaf structures, but the increase in MDA content is low, indicating that the triploid population has less damage due to drought and strong tolerance to drought. This phenomenon is also found in polyploid Lonicera japonica [52]. Stomata can quickly and sensitively sense drought stress [53]. Under drought stress, the stomatal conductance of lily leaves decreased, and ‘Lyon’ stomatal conductance only decreased by 21.8% after 30 days of drought stress, which may be related to the lily having small and thick leaves and thicker palisade tissue. The same conclusion was reached in Brassica napus varieties [54]. The photosynthetic rate of leaves did not change significantly. At this time, the inhibition degree of stress on photosynthesis may be related to the role of the antioxidant system. In conclusion, Chlb is more sensitive to adverse drought environments, and the drought-resistant structure of leaves can protect the photosynthetic system of lily leaves under drought stress.

After 5 days of rewatering, except for ‘Lyon’, ‘Royal Sunset’, and ‘Robina’, the antioxidant enzyme activity of other varieties increased, the ability to clear intracellular ROS increased, ROS decreased, and the content of malondialdehyde decreased. The results showed that the arid soil environment was alleviated, and the membrane lipid oxidation caused by the accumulation of ROS was repaired after rehydration [55]. For varieties with weak drought tolerance, the physiological functions of plants are inhibited, and the activity of CAT, POD, and SOD cannot be maintained at a high level. After restoring water conditions, the protective mechanisms were enhanced by increasing enzyme activity [56]. The protective mechanisms of improving enzyme activity in drought-tolerant varieties will be relieved when the water condition is restored. Similar laws have been found in soybeans [57]. Chlorophyll synthesis began to recover, and stomatal conductance and the photosynthetic rate increased. Among these observations, the photosynthetic recovery rate of ‘Lyon’, ‘Royal Sunset’, and ‘Robina’ was higher than the photosynthetic recovery rate of ‘Gelria’, ‘Siberia’, ‘Elena’, and ‘Immaculate’. At the same time, due to the decrease in the ROS signal, coupled with the increase in water absorption, the concentration of cell osmotic substances gradually decreased.

In this study, morphological and physiological markers were used to screen the drought resistance of different lily varieties. The relationship between drought resistance physiology and leaf morphology of lily was revealed. It provides a reference for the development of water-saving, efficient, and high-quality cut-flower production and the breeding of drought-resistant varieties. In the future, we can combine the water balance mechanism of different parts of lily and use transcriptomics, proteomics, and other biotechnology to further study the drought-resistance mechanism of lily.

5. Conclusions

There were differences in leaf morphological indices and anatomical structures among the seven varieties, and the changes in the photosynthetic system, antioxidant enzyme system activity, and cell osmotic material system of the lily varieties with different leaf types were different under different drought conditions. From the analysis of the synergistic changes in the above systems, the drought resistance of ‘Lyon’, ‘Royal Sunset’, and ‘Robina’ is better than the drought resistance of ‘Gelria’, ‘Siberia’, ‘Elena’, and ‘Immaculate’. Specific leaf area may be one of the apparent indices that affect the drought resistance of the lily.