Drought Resistance of 10 Ground Cover Seedling Species during Roof Greening

10 species’ drought resistance cases have been studied, including Paeonia lactiflora, Hemerocallis dumortieri, Physostegia virginiana, Iris lacteal, Hylotelephium erythrostictum, Sedum lineare, Iris germanica, Cosmos bipinnata, Hosta plantaginea and Dianthus barbatus. By researching these drought resistance cases, a suggestion can be given for roof greening. This research sets 3 drought stress scenarios by controlling the soil relative water content (RWC), including moderately drought stress (40%±2% < RWC < 45%±2%), strong drought stress (RWC < 30%±2%) and control group (RWC > 75%±2%). After the seedlings survived the drought stress, the damaging rate of permeability (DRP), total chlorophylls concentrations (Chl), superoxide dismutase (SOD), peroxidase (POD) and ascorbate peroxidase (AsAPOD) of seedlings will be measured. Finally, a subordinate function method was applied to assess these species’ drought resistance. Cosmos bipinnata and Physostegia virginiana was dead after having suffered with moderately drought stress and strong drought stress, respectively. Although other species survived, the individual variation was huge especially for physiological and biochemical index. Hemerocallis dumortieri, Iris lactea and Hosta plantaginea’s DRP had little change when they lived in the normal water condition and suffered with drought stress. Most of the species (except Paeonia lactiflora and Sedum lineare) showed a lower SOD activity during moderately drought stress compared with the sufficient soil water condition and strong drought stress condition. The changes of plants’ POD activity and AsAPOD activity are very similar: when drought stress enhanced, the activity of protect enzyme reduced. According to the subordinate function method, the order of plants’ resistance to the drought is as follow: Hosta plantaginea > Sedum lineare > Iris germanica > Hemerocallis dumortieri > Iris lactea >Hylotelephium erythrostictum > Dianthus barbatus > Paeonia lactiflora > Physostegia virginiana > Cosmos bipinnata.


Introduction
Roof greening, regarded as "Fifth surface greening", is one of the fundamental measures for sponge city, a significant national policy to improve relationships between city development and nature protection, in order to keep hydro-ecology balance (Wang and Zhang, 2015;Yu et al., 2015). As an important supplement of urban landscaping, roof greening is good for mitigating Urban Heat Island Effect (UHI) (Ondimu and Murase, 2007), improving air quality (Baik et al., 2012) and enriching biodiversity of city (Köhler and Clements, 2013), hence the reason to extend such landscaping style to all of China. Green roof can be roughly categorized into two types: support diverse plants (shrubs and trees, grass and flowers), namely intensive green roof (IGR) and simple herbaceous plant species, namely extensive green roof (EGR) (Peng and Jim, 2015). The plant has many challenges to grow on the roof. Take Beijing as an example, plants applied to the roof green are suffered with restricted rainfall in the winter, spring and autumn; the evaporation always increased by summer high temperature. Drought is considered one of the most common environmental stresses presently affecting plant growth (Nahar et al., 2015;Romain et al., 2006). When plants are suffering drought stress, the reactive oxygen species (ROS) would produce in the plant (Sharma and Dubey, 2005). Different ROS, including singlet oxygen ( 1 O 2 ), superoxide radical (O 2 -), hydroxyl free radical (•OH), and hydrogen peroxide (H 2 O 2 ) (Smirnoff, 1993) will reduce the productivity in crop and depress viability in the plant, as they would cause oxidative damage to proteins, DNA, and lipids (Apel and Hirt, 2004). Accordingly, drought stress will disturb leaf membrane permeability (MP) (Bai et al., 2006), the total chlorophyll concentrations (Chl) (Rulcová and Pospíšilová, 2001), superoxide dismutase (SOD) (Bedard and Krause, 2007), peroxidase (POD) (Bahari et al., 2015), and ascorbate peroxidase (AsAPOD) (Ghozlene et al., 2014). Actually, these indicators mentioned above that measure the degree of plants bearing drought stress include MP, Chl, SOD, POD and AsAPOD usually are analyzed as a whole. That is because these indexes have strong relations with each other, for instance SOD, POD and AsAPOD as anti-oxidative enzymes (Chang et al., 2006) can eliminate ROS to protect the cells from being damaged.
According to the "Beijing local standards, roof greening specification (DB11/T 281-2005)" (Beijing Municipal Administration of Quality and Technology Supervision), there are more than 20 ground cover seedlings included Paeonia lactiflora Pall., etc., that recommended to applying during the roof greening. However, these plants' drought resistance has not sequenced. To figure out the plant with strong resistance to the drought environment can improve the surviving rate of these plants and save maintenance cost during roof greening.

Study Area Overview
Milu Park is situated 2 km to the south of the South-5-Ring Road in Beijing, and surrounded by Nan-Haizi Suburb Park. The drought stress test was held in a rain protection shed, located in the core-protection area of David's Deer in  Table 1. These seedlings were provided by Yu-quanying flowers market, a big and famous market in Beijing.

Field soil collection
The plant seedlings need replant into the special flowerpot, that was plastic made, with 20 cm high and 13 cm diameters, and there are 3 little holes at the bottom of the pot for pervious to water for the drought experiment. The field soil in Tao-huadao, a small wetland already dried up entirely in Milu Park was used as soil for transplanting.
The soil is belonging to the medium salty soil, with mean value of pH are 7.89, mean content rapidly available nitrogen are 24.73 mg·kg -1 , mean content rapid available phosphorus are 18.87 mg·kg -1 , mean content rapidly available potassium are 322.18 mg·kg -1 (Zhu et al., 2016). The field soil was pulverized to powders by rolling stick after they dried out naturally.

Plants transplanting
The program of transplantation is as follows: firstly, about 400 g soil powder was put in the pot. Secondly, tear off the one-off plastic wrap surrounded on the seedlings' roots. Then pick up the seedlings very carefully, and put them in the pot. Place the roots as close to the center of the pot as possible. Thirdly, fill other about 300 g soil powder into the pot and bury the root. Those soil powders need compressed tightly with fingers. At last, water seedlings once every 10 min and repeat 3 times, for keeping there are enough water support the plant rooting. The success of transplanting a plant can be judged by observing whether it grows new leaves, and whether the stem is fresh. After a week of replanting, the seedlings' surviving rate achieved 99%.

Drought stress design
Three drought stress degrees have been set, including moderately drought stress (Treatment 1 is short for T1, the water content in soil varies from 40%±2% ~ 45%±2%), strong drought stress (Treatment 2 is short for T2, the water content in soil are less than 30%±2%) and sufficient soil water condition (Control group is short for Cg, the water content in soil are higher than 75%±2%) (Bai et al., 2015;Bu et al., 2010). For every drought stress degrees there were 30 seedlings with 3 replicates for each plant species have been tested.
For Cg, water plant seedlings every 4 days. The drought stress was dependent on natural evaporation. During the drought stress period, the W.E.T Sensor type WET-2 TM made by Delta-T Devices Limited Company was used to measure the water content. Once the relative water content of soil meets the requirements of experiment, keep the plant seedlings lived in such environment about 2 days to make sure the physiology and biochemistry have been changed.

Leaves sampling
The details of sampling method in VDI-Guideline 3975 Part 11 (2007) are referenced. Leaves of 10 plant species were collected during April 27 th to May 9 th , 2015. Triplicate samples were collected for each plant species of each treatment. At least 15 g of varies numbers of leaves has been collected for each sample in order to have enough quantity for analysis. Healthy-looking leaves were chosen as possible. Leaves were placed into sealed plastic bags and kept in a portable ice-box at 0 ~ 4 ℃ before being transferred to the lab for further physiological and chemical analysis.

MP, Chl, SOD, POD and AsAPOD determination
Membrane Permeability of leaves was determined as DRP (%) = . Where L t is the relative electrical conductance of drought stress treatments; L Cg is the relative electrical conductance of control group. The relative electrical conductance followed L . Where S 1 is the original conductivity value of the deionized water with = 1 -0 2 -0 fractured fresh leaves; S 2 is the conductivity value of the boiled deionized water with fractured leaves; S 0 is the conductivity value of deionized water (Zhang et al., 2011). Leaves membrane permeability was determined using a model Thermo TM Scientific Orion 3-star inductivity measurer. Before the test, all samples were flushed by de-ionized water for 3 times and wipe off water on the leaves' surface.
The total chlorophyll content estimation was carried out by the method of Arnon (1949). The method details were described in a published paper .
Treating the crude enzyme extracted from leaves is the impurity to measure the SOD, POD and AsAPOD. About 0.5 g fresh leaves with a small number of CaCO 3 and high purity quartz sand were crashed into powder under freezing environment, with 5 mL phosphate buffer (0.05 mol·L -1 ) in a mortar. The paste were breathed into a 10 mL centrifuge tube, and then diluted with de-ionized water to 10 mL. By centrifuging at high speed (F=1 3000 g) for 20 min at 0 ℃ to 4 ℃ (Zhang et al., 2011). SOD and POD reaction system was described by Zhang et al (2011), AsAPOD reaction systems was described by Tang et al (2012) as follows: when all the tubes shined with the light, the check tube is wrapped in the tinsel avoid be shined; A s is the is the absorbance of samples; V T (mL) is the total volume of samples; V 1 (mL) is the volume of participation in the reaction and W (g) is the weight of the fresh leaves.
All the solution were put into the test tube, which followed by the POD reaction system (Table 2-POD). Then record the solution's light absorption value of each tube (the wave length was regulated at 470 nm). Read every 1 minute, and each solution was recorded 5 times in 5 minutes.
Where A 470 is the change of absorbance during the reaction period; W (g) is the weight of the sample; t (min) is the reaction time; Vs (mL) is the volume of participation in the reaction and the V T (mL) is the total volume of samples.
According to the AsAPOD reaction system (Table 2-AsAPOD), the mixture was put into test tubes, and then records the light absorption values at 290 nm at every minute. And such the step repeats 5 times. The formula to account the AsAPOD as follows: Where A 290 is the change of absorbance during the 5 min; V 1 (mL) is the volume of crude enzyme; V 2 (mL) is the volume of crude enzyme to take part in the reaction (0.1 mL in this test); t (min) is the reaction time (5 min in this test) and the W (g) is the weight of the fresh leaves. FW is the short for leaves fresh weight. A drought resistance assessment method (Wu et al., 2013;Chen et al., 2002) for the plant species based on the subordinate function value belong to the fuzzy mathematics is applied in the research. By the method, the DRP, Chl, SOD, POD and AsAPOD's results can be synthesized as a single value of each species. The subordinate function value formula is as follows:

Data analysis
Where the lowercase "i" and "j" is represents plant species and determination index respectively. Therefore, the "X ij " is the mean value of "j" index of the "i" species. "X imax " and "X imin " is represents the maximums and minimums of the "j" index of the "i" species. " " is the subordinate function value. " " represents the drought resistance of ij ij plant seedlings. Then the average of subordinate function value is applied to estimate the adaptive capacity of plants that bear the drought stress. The average formula is as follows ("n" represents the amount of indexes and " " represents the average):

Results and discussions 4.1 Soil moisture content
The water content of soil sees Table 3. Although the water content of soil has standard deviation, the 3 drought stress levels are distinguished, well agreeable to the designed levels.  (n=30). Cg is short for control group (with normal watering condition); T 1 represents the "moderately drought stress"; T 2 represents the "strong drought stress". Greek alphabet "I, Ⅱ and Ⅲ " represents the first, second and third drought stress tests.

Membrane permeability
The values of MP represent how much the electrolyte seeping from the cell of plant. According to Sanchez-Rodriguez et al. (2010), the maintenance of sufficient water in plant tissue protects plant from dehydration and carboxylation, and their other enzymes from inactivation. The low-temperature environment (Shen, 2005), drought stress environment (Dai et al., 2006), salt stress environment (Xu et al., 2002;Du et al., 2012) and heavy metal stress (Vos et al., 1989) has been proven to achieve a higher level of damaging rate of permeability.. In other words, the higher value of MP is, the more the cell damage is. With the same water condition, the lower MP value implies the stronger adaptation of plant species to the environment (Wang et al., 2014).

Figure 1
Membrane permeability changes of 10 species in the drought stress tests.
Note: The histogram shows the mean values. Cg is short for control group (with normal watering condition); T1 represents the "moderately drought stress"; T2 represents the "strong drought stress". The lowercase letter before the comma shows the statistical significance among different plant species, the lowercase letter behind the comma shows the statistical significance among Cg, T1 and T2.

Total chlorophylls contents
Usually, when the leaves of plants lost too much water, it would impede the chlorophyll production and may even resolve the chlorophyll (Yang et al., 2004;Zhang et al., 2003). That's because the ROS ( 1 O 2 , O 2and ·OH) can make the lipid peroxidation directly and indirectly, thus damage the chlorophyll, and reduce the ratio of chlorophyll a and b (Zhang and Tan, 2001).

Figure 2
Total chlorophylls contents changes of 10 species in the drought stress tests.
Note: The histogram shows the mean values. Cg is short for control group (with normal watering condition); T1 represents the "moderately drought stress"; T2 represents the "strong drought stress". The lowercase letter before the comma shows the statistical significance among different plant species, the lowercase letter behind the comma shows the statistical significance among Cg, T1 and T2. Different lowercase letters indicate a significant difference at P < 0.05.
According to Figure 2

Superoxide dismutase activity
The superoxide dismutases (SODs) constitute a first line of defense at the cellular level (Noctor and Foyer, 1998) against ROS (Alscher et al., 2002). SODs are typically classified into three different groups depending on the prosthetic metal present in the active site, and they are designated as CuZn-SODs, Mn-SODs and Fe-SODs (Fridovich, 1986). McCord and Fridovich (1969) (Xu et al., 2014). Note: The histogram shows the mean values. Cg is short for control group (with normal watering condition); T1 represents the "moderately drought stress"; T2 represents the "strong drought stress". The lowercase letter before the comma shows the statistical significance among different plant species, the lowercase letter behind the comma shows the statistical significance among Cg, T1 and T2. Different lowercase letters indicate a significant difference at P < 0.05. According to Figure 3, the SOD activity of plant species from high to low as Cosmos bipinnata, Hemerocallis was the lowest only with 0.04 U·mg -1 ., that was significant lower (P<0.05) than the other 10 plant seedlings.
The SOD activity of plant species from high to low as Paeonia lactiflora, Hosta plantaginea, Hemerocallis Paeonia lactiflora and Sedum lineare was increased caused by drought.
The SOD activity of these plants that survived during strong drought stress (T2) was further enhanced compared with those under T1. The SOD activity of these plant seedlings from high to low as Paeonia lactiflora, Iris lactea, Hemerocallis dumortieri, Hosta plantaginea, Hylotelephium erythrostictum, Sedum lineare, Iris germanica and Dianthus barbatus. In there, species Sedum lineare and Iris germanica was both 0.79 U·mg -1 . Species Paeonia lactiflora's SOD activity was the highest with 1.04 U·mg -1 , that was significant higher (P<0.05) than the other 8 surviving plant seedlings. Plants seedlings' SOD activity was enhanced under strong drought, which indicates there is more protective enzyme produced by plants to eliminate ROS.

Peroxidase activity and Ascorbate peroxidase activity
Peroxidase is a kind of antioxidant enzymes that it could scavenge and decompose ROS (Mohammadi et al., 2016).
The mechanism of POD scavenging ROS was described as follows: RH 2 +H 2 O 2 →2H 2 O+R, then H 2 O 2 were transformed into H 2 O thoroughly Qin et al., 2005). Survila et al (2016) demonstrates that increased peroxidase activity increases permeability of the leaf cuticle. When Chen et al (2017) and Wu et al (2016) researched the drought resistance of Cucumber and dendrobium moniliforme, they found that POD increased activity helps alleviate oxidative damage.
According to the Figure 4, POD activity kept a relatively high level in the Cg. Hosta plantaginea achieved 246.33 U·g -1 ·min -1 , and it was significantly higher (P<0.01) than the other species except Hemerocallis dumortieri. The lowest species was Physostegia virginiana, only 14.36 U·g -1 ·min -1 . When the seedlings suffered with T1, all the species' POD activity reduced sharply. The highest POD activity was Hemerocallis dumortieri with 22.08 U·g -1 ·min -1 , thus it have no significant difference with the other species. In addition, the higher peroxidase enzyme activity was obtained under drought stress and can be attributed to the plant defense mechanisms against free radical formation resulting from water defcit (Ruppenthal et al., 2016). In the case of T1, Dianthus barbatus, Iris germanica and Iris lactea showed stronger POD activity compared with the others. This indicates these 3 seedlings have better drought resistance to the moderately drought stress. Similarly, when the plants suffered with the strong drought stress, the POD activity of seedlings were further decreased. POD activity of Paeonia lactiflora, Iris germanica and Dianthus barbatus showed more sensitivity to the strong drought stress. They were significantly lower (P<0.01) than the T1.
In general, POD activity of all the species showed an obviously decreasing trend when the seedlings suffered with drought stress. The reason for the POD activity reduction in this research may be the different protective enzymes work as a whole eliminating the ROS. Such that when the SOD eliminates O 2 ─ , the H 2 O 2 would increase during the physiological reaction processes. However, with the drought stress enhancing, the reduction of POD activity impeded scavenge H 2 O 2 (Sun et al., 2003). This indicates the plant seedlings defensive system of eliminating the ROS was too weak.

Figure 4
Peroxidase activity changes of 10 species in the drought stress tests.
Note: The histogram shows the mean values. Cg is short for control group (with normal watering condition); T1 represents the "moderately drought stress"; T2 represents the "strong drought stress". The lowercase letter before the comma shows the statistical significance among different plant species, the lowercase letter behind the comma shows the statistical significance among Cg, T1 and T2. Different lowercase letters indicate a significant difference at P < 0.05.
Ascorbate peroxidase belongs to the class I heme-peroxidases and is found in most eukaryotes including higher plants (Anjum et al., 2016). Ascorbate peroxidase exists as isoenzymes and plays an important role in the metabolism of H 2 O 2 in higher plants Shigeoka et al., 2002;Sukweenadhi et al., 2017). Meanwhile, AsAPOD also have the ability to scavenge the ROS (Li et al., 2013). That means the higher AsAPOD activity in plants resulted in faster removal of H 2 O 2 , which leads to alleviation of oxidative damage (Wu and Xia, 2006).
According to the Table 7, the highest values of AsAPOD activity was happened in the Cg, that they were significantly higher (P<0.05) than T1 and T2. The AsAPOD activity is very similar to the change of POD activity. In the Cg, Paeonia lactiflora obtained the highest value as 17.86 U·min -1 ·g -1 FW . The others present an equally same trend, that the values were during the 9.50 U·min -1 ·g -1 FW to 7.06 U·min -1 ·g -1 FW . When the seedlings suffered with T1, the AsAPOD activity languished from the drought stress observably, and all of them were significantly lower (P<0.01) than the Cg. All the plant's AsAPOD activity was less than 1.00 U·min -1 ·g -1 FW , except for Paeonia lactiflora in the T1. All the plant's AsAPOD activity shows a further decline in the T2, none of them could exceed 0.50 U·min -1 ·g -1 FW . The record was still Paeonia lactiflora in the T2. The change of AsAPOD activity was intense during the drought stress test, thus indicating AsAPOD was sensitive to the drought stress. As for AsAPOD activity of the 10 plant species, Paeonia lactiflora showed a better resistance to ROS.

Figure 5
Ascorbate peroxidase activity changes of 10 species in the drought stress tests.
Note: The histogram shows the mean values. Cg is short for control group (with normal watering condition); T1 represents the "moderately drought stress"; T2 represents the "strong drought stress". The lowercase letter before the comma shows the statistical significance among different plant species, the lowercase letter behind the comma shows the statistical significance among Cg, T1 and T2. Different lowercase letters indicate a significant difference at P < 0.05.

Assessment of drought resistance to the plants
The drought resistance of botany is a complex process, because the physiological and biochemical of plants changed as the ROS changed. A subordinate function method can be used for the assessment of drought resistance to the plants. The method's essence was to apply fuzzy mathematics to the resistance indexes, such as DRP, Chl, SOD, POD and AsAPOD in this research. According to the formula of subordinate function, the results showed as following  Note: Cg is short for control group (with normal watering condition); T1 represents the "moderately drought stress"; T2 represents the "strong drought stress"; "-" represents the species extinct during the test period.
Through the calculation, the drought resistance of 10 plant species was very clear. Iris germanica was the best in drought resistance. Cosmos bipinnata and Physostegia virginiana died when they suffered with moderately drought stress and strong drought stress, respectively. Paeonia lactiflora survived the weakest drought resistance.

Conclusions
The drought stress disturbed the plant growth, the MP, Chl, SOD, POD and AsAPOD have been changed compared to the control group. The result shows Cosmos bipinnata and Physostegia virginiana died after having suffered with moderately drought stress and strong drought stress, respectively. Although other species survived, the individual variation was huge especially for physiological and biochemical index. Hemerocallis dumortieri, Iris lactea and Hosta plantaginea's MP had little change when they lived in the normal water condition and suffered with drought stress. The change of SOD activity is sensitive to drought stress. The most of the species (except Paeonia lactiflora and Sedum lineare) showed a lower SOD activity during moderately drought stress compared to the sufficient soil water condition and strong drought stress condition. The changes of plants' POD activity and AsAPOD activity are very similar: when drought stress enhanced, the activity of protect enzyme reduced.
As a kind of local flowering plant, Hosta plantaginea shows good drought resistance. It is suggested to apply to the roof greening in Beijing and other northern cities in China widely. Cosmos bipinnata and Physostegia virginiana are too sensitive to drought stress, so it is not suggested to apply them to the roof greening, especially in the arid region.
Other species can be applied to the roof greening for creating the landscape with no limitations. In order to increase the plants' rate of survival and control the cost of roof greening, those plants with good drought resistance should be considered first.