Sex-specific responses to winter flooding, spring waterlogging and post-flooding recovery in Populus deltoides

Winter flooding events are common in some rivers and streams due to dam constructions, and flooding and waterlogging inhibit the growth of trees in riparian zones. This study investigated sex-specific morphological, physiological and ultrastructural responses to various durations of winter flooding and spring waterlogging stresses, and post-flooding recovery characteristics in Populus deltoides. There were no significant differences in the morphological, ultrastructural and the majority of physiological traits in trees subjected to medium and severe winter flooding stresses, suggesting that males and females of P. deltoides were winter flooding tolerant, and insensitive to winter flooding duration. Males were more tolerant to winter flooding stress in terms of photosynthesis and chlorophyll fluorescence than females. Females displayed greater oxidative damage due to flooding stress than males. Males developed more efficient antioxidant enzymatic systems to control reactive oxygen species. Both sexes had similarly strong post-flooding recovery capabilities in terms of plant growth, and physiological and ultrastructural parameters. However, Males had better recovery capabilities in terms of pigment content. These results increase the understanding of poplars’s adaptation to winter flooding stress. They also elucidate sex-specific differences in response to flooding stress during the dormant season, and during post-flooding recovery periods.


Results
Comparative analysis of survival rates and morphological traits. Generally, all flooding-stressed P. deltoides seedlings survived, and epicormic shoot germination from nodes and new leaf emergence occurred almost simultaneously with controls. With increasing flooding duration, significant visible damages, such as mortality, leaf chlorosis, leaf necrosis, or leaf abscission were not observed during the plant growth stage of development. Morphological adaptations such as hypertrophied lenticels, aerenchyma tissues and adventitious roots, which often occur when plants are exposed to summer waterlogging stress, did not appear in the present study.
When flooding was combined with waterlogging, shoot height (Fig. 1A) and basal stem diameter (Fig. 1B) were significantly inhibited in comparison to controls. Greater inhibition was observed with increasing flooding duration. Significant differences were not found between the two flooding treatments (W-90d and W-140d). In Values are means ± SE (n = 5). Letters above the columns indicate significantly differences at P < 0.05 according to Duncan's test. The significance values of the factorial analysis (ANOVA) for: F S , F W , and F R refer different effects of species, watering, and recovery growth, respectively; F S×W , F S×R , F W×R , and F S×W×R refer species × watering effects, species × recovery growth effects, watering × recovery growth effects, and species × watering × recovery growth effects, respectively. addition, significant differences in the shoot height and basal stem diameter of females and males were not found within each flooding treatment.
During the recovery stage, the shoot heights and basal stems of flooding stressed females and males both recovered well. For example, shoot height growth rates in females and males were 38.8% and 42.4% in CK, respectively, whereas they were 49.8% and 51.8% in W-90d, and 57.9% and 54.4% in W-140d during the recovery stage. The basal stem growth rates in females and males were 21.0% and 27.6% in CK, respectively, whereas they were 29.3% and 28.8% in W-90d, and 40.8% and 32.0% in W-140d during the recovery stage. In addition, significant differences were not found between the two post-flooding recovery treatments (W-90d-R and W-140d-R). In addition, there were no significant differences in shoot height and basal stem during the recovery stage between females and males within each post-flooding recovery treatment.
Winter flooding stress affected the physiological traits of two-year-old P. deltoides trees at an early stage of development. Severe flooding stress (W-140d) significantly decreased A (Table 1), Fv/Fm  (Table 1), WUEi (Table 1) (Table 2), Chl b (Table 2), Caro (Table 2), Total Chl (Table 2), Chl a/Chl b (Table 2), gs (Table 1) 6E) were absent in both sexes of P. deltoides exposed to W-140d conditions (Supplementary Table 1 . − (Fig. 5D), POD, SOD, CAT, and GR in both sexes of P. deltoides did not vary significantly between medium flooding stress (W-90d) and severe winter flooding stress (W-140d). However, significant variations in Ci (Table 1)  . − , POD, SOD, CAT, and GR in previously severe flooded females and males of P. deltoides after 15 d recovery growth (W-140d-R) recovered to normal or even better levels than in individual unflooded plants (Supplementary Table 1). In addition, previously flooded plants subjected to different flooding durations (W-90d-R and W-140d-R) exhibited insignificant differences in terms of A, Chl a, Caro, Total Species Treatment A (μmol · m −2 s −1 ) gs (mol · m −2 s −1 ) Ci (μmol · CO 2 mol −1 ) E (mmol · m −2 s −1 ) WUEi . − , and decreases in ·OH, and insignificant variation in MDA were observed in males. In addition, compared with flooded males, the flooded females had higher levels of H 2 O 2 , O 2 . − , and MDA. Although both sexes of P. deltoides exposed to winter flooding stress displayed similar trends and activity levels of POD, SOD, and GR, there were differences in the trends of APx and CAT activities (Supplementary Table 1). Males showed significant increases in APx activity but no variation in CAT activity. In contrast, female APx activity did not vary, while CAT activity decreased.
In general, previously-flooded females and males had similar recovery capabilities in terms of majority physiological parameters during the post-recovery stage. However, when the flood water drained away, males showed significant higher levels in Ci, E, Chl a, Chl b, and Total Chl, and significant lower levels in WUEi and MDA in comparison to females under W-90d-R conditions. In addition, compared with females, males showed significant higher levels in Ci, E, Chl a, Total Chl, and APx, and significant lower levels in WUEi and MDA under W-140d-R conditions.
Comparative analysis of ultrastructural morphology. We next asked how the ultrastructural morphology of P. deltoides females and males varied in response to winter flooding stress and post flooding recovery. Thus, we examined their cellular ultrastructure. As shown in Fig. 7, chloroplasts of both sexes were well arranged. Values are means ± SE (n = 5). Letters above the columns indicate significantly differences at P < 0.05 according to Duncan's test. The significance values of the factorial analysis (ANOVA) for: F S , F W , and F R refer different effects of species, watering, and recovery growth, respectively; F S×W , F S×R , F W×R , and F S×W×R refer species × watering effects, species × recovery growth effects, watering × recovery growth effects, and species × watering × recovery growth effects, respectively. Values are means ± SE (n = 5). Letters above the columns indicate significantly differences at P < 0.05 according to Duncan's test. The significance values of the factorial analysis (ANOVA) for: F S , F W , and F R refer different effects of species, watering, and recovery growth, respectively; F S×W , F S×R , F W×R , and F S×W×R refer species × watering effects, species × recovery growth effects, watering × recovery growth effects, and species × watering × recovery growth effects, respectively.
The majority had distinct granum regions, and some small plastoglobules were found, especially in females. Under medium flooding stress (W-90d), plastoglobules became enlarged in both sexes, and their numbers decreased in females. However, under severe flooding stress (W-140d), disintegrated chloroplasts and numerous tilted granal stacks were found, and plastoglobules gradually disappeared. Some large vesicles, starch grains, and swollen mitochondria were found, especially in females. With the growth of development, more distinct granum and small starch grains developed. During the post-flooding recovery stage, the percentage of vesicles decreased, while distinct granum regions increased. The number of mitochondria, small starch grains, and small Values are means ± SE (n = 5). Letters above the columns indicate significantly differences at P < 0.05 according to Duncan's test. The significance values of the factorial analysis (ANOVA) for: F S , F W , and F R refer different effects of species, watering, and recovery growth, respectively; F S×W , F S×R , F W×R , and F S×W×R refer species × watering effects, species × recovery growth effects, watering × recovery growth effects, and species × watering × recovery growth effects, respectively.
plastoglobules increased. In addition, no significant differences in the ultrastructural morphology of P. deltoides females and males were observed.

Discussion
Winter flooding is a special event in some rivers due to the artificial water level regulation. Kozlowski 3 suggested that flooding during the growing season can have greater negative effects on deciduous species than flooding Values are means ± SE (n = 5). Letters above the columns indicate significantly differences at P < 0.05 according to Duncan's test. The significance values of the factorial analysis (ANOVA) for: F S , F W , and F R refer different effects of species, watering, and recovery growth, respectively; F S×W , F S×R , F W×R , and F S×W×R refer species × watering effects, species × recovery growth effects, watering × recovery growth effects, and species × watering × recovery growth effects, respectively. The significance values of the factorial analysis (ANOVA) for: F S , F W , and F R refer different effects of species, watering, and recovery growth, respectively; F S×W , F S×R , F W×R , and F S×W×R refer species × watering effects, species × recovery growth effects, watering × recovery growth effects, and species × watering × recovery growth effects, respectively. during the dormant season. However, information about morphological and physiological responses to winter flooding stress is still scarce. Therefore, some researchers have suggested that the mechanisms of winter flooding adaptation in flooding tolerant species should be investigated to better understand their survival responses to reversed flooding patterns [1][2][3] All plants survived 140 days of winter flooding without mortality, leaf chlorosis, leaf necrosis, or leaf abscission, which suggests that, in terms of seedling survival, both female and male P. deltoides are winter flood tolerant. Therefore, both female and male P. deltoides are superior candidates for the construction of protective riparian forests in winter flooding areas. Flooding or waterlogging during the growing season can induce aerenchyma tissues and adventitious roots in a wide variety of both flood-intolerant and tolerant angiosperms and gymnosperms, especially in flood tolerant species, as concluded by Kozlowski 3 . However, some often-visible morphological adaptations, including the formation of hypertrophied lenticels, aerenchyma tissues and adventitious roots, were not observed in winter flooding-stressed female and male P. deltoides, which might be a result of low oxygen demands during the dormant season.
The majority of previous studies have demonstrated that flooding or waterlogging stresses can decrease plant growth and development, pigment content, photosynthetic capacity, stomatal conductance, and chlorophyll fluorescence. They may also increase ROS and MDA levels, increase the contents of some cellular non-enzymatic components, and increase the activity of antioxidant enzyme systems 2-4, 6-8, 23, 24 . Compared with summer waterlogging during growth and development, winter flooding stress during the dormant season caused fewer physiological dysfunctions in P. deltoides 5,6 . The duration of flooding significantly affects survival rates, plant growth and yield, and physiological responses 1, 3, 9 . Our results suggest that female and male of P. deltoides were not very sensitive to flooding duration. The duration of winter flooding had little effect on plant growth, physiology, and biochemistry due to the low levels metabolic activity occurring during the dormant season.
Normally, although the stomata reopen slowly and the rate of photosynthesis increases when flood water drains away, previously-flooded plants may have difficultly recovering to normal levels of growth, photosynthetic capacity, and other physiological responses, because absorption of water by their small root systems can't adequately replenish transpirational losses within a short period 3,25 . In present study, the ultrastructural morphology of mesophyll cells and majority physiological parameters in previously flooded plants recovered to normal after 15 d recovery growth. These phenomena may be because the winter flooding stress caused little damage to the one-year-old root systems, and then the previously flooded plants could absorb water and nutrients normally after flood water drained away. These results also suggest that P. deltoides has strong self-repairing capabilities during post-flooding recovery.
Some studies of trees have observed sex-specific responses to a number of biotic stresses 26,27 and environmental stresses such as waterlogging stress 6 , water deficit 16 , chilling 17 , salinity 21 , enhanced UV-B radiation 28 , atmospheric CO 2 enrichment 29 , nutrient deficiency 19, 30 , excess manganese, and a combination of different stresses 31,32 . These studies concluded that females seem to be more sensitive to environmental stress and usually experience greater  negative effects. In other cases, female plants exhibited better tolerance of adverse conditions than males, females showed a more conservative strategy of water use or higher photosynthetic rates compared with males 4, 33-35 . For example, Male and female P. angustifolia genotypes grew similarly with favorable water levels, but males tended to be more inhibited by flooding, while females were more flood tolerant 4 . Females of some perennial herbs are more resistant to drought stress than males 35 . Therefore, the information available thus far is insufficient for establishing generalizations about whether male and female dioecious plants show differential abiotic stress resistance. Furthermore, most comparisons of male and female performance have mainly focused on short-term physiological measurements during the growth season. In this paper we highlight the need for physiological studies of sexual dimorphism in dioecious plants in response to different durations of winter flooding stress during the dormancy season and during post-flooding recovery. Both sexes had similarly strong flooding tolerances in terms of the majority of physiological parameters and ultrastructural variations at an early stage of development. Differences between sexes do not always exist. For example, the female and male P. angustifolia did not display significant differences in photosynthetic gas exchange, leaf reflectance, chlorophyll fluorescence, and WUE in response to groundwater availability, and soil water typically declines between May and September 7 . The similar sexual responses of P. deltoides to winter flooding stress may be the result of low metabolic activity during the dormancy season and early growth developmental stage.
The significant variations in A, Fv/Fm, Yield, and ETR (Supplementary Table 1) only in females under medium winter flooding stress conditions (W-90d) suggest that female P. deltoides are more sensitive than males to winter flooding stress in terms of photosynthesis and chlorophyll fluorescence. Our results are consistent with previous studies. For example, female P. yunnanensis grown in China during a drought exhibited gas exchange rate depression and greater damage to cell organelles than male did 32 . Female P. cathayana grown in China were also more responsive and showed greater negative effects on net photosynthesis than males when grown under increased drought stress and elevated temperatures 16 . Additionally, although the previously-flooded females and males had similar recovery capabilities in the post-flood period, males had better recovery capabilities than females in terms of pigment content, especially chl a.
Root oxygen deficiency resulting from flooding stress caused photoxidative damage to leaves via increased generation of ROS 2, 3 . These are important signaling molecules indicative of oxidative stress. They can directly attack membrane lipids, resulting in lipid peroxidation and oxidation of proteins and nucleic acids 6,[36][37][38]  , OH, MDA, APx, and CAT in response to winter flooding stress, which suggest that female plants encountered more serious oxidative damage during winter flooding stress than males. Plants can protect cellular and sub-cellular systems to control ROS levels and membrane lipid peroxidation with antioxidant enzymatic systems. These results indicate that males develop more efficient antioxidant enzymatic systems to control ROS accumulation than females do. Our results indicate that, in terms of antioxidant enzymatic systems, P. deltoides males are more tolerant to winter flooding stress than females. Previous studies have shown that females of P. yunnanensis, P. cathayana, and P. deltoides exhibit greater ROS accumulation and oxidative stress damage, and have less efficient antioxidant enzymatic systems than males during environmentally stressful growth season conditions 6,17,18,21,28,30 .
In conclusion, both sexes of P. deltoides are winter flood tolerant, based on seedling survival, and morphological, physiological, and ultrastructural responses. Winter flooding stress differentially affected physiological traits in P. deltoides at early growth stages of development. Significant variations in terms of chl a, chl b, caro, total chl, chl a/chl b, gs, soluble protein, reducing sugar, proline, ·OH, CAT, and POD were absent under severe winter flooding stress in both sexes. In addition, fatally-damaging ultrastructural responses were not found in P. deltoides. In both sexes, the duration of winter flooding stress had insignificant effects on morphological, ultrastructural, and the majority of physiological responses, except for A, gs, E, Fv/Fm, Yield, ETR, H 2 O 2 , and APx. When flood water was drained away, previously-flooded plants grew at a faster rate than unflooded plants. The majority of physiological parameters and ultrastructural morphology of mesophyll cells in previously-flooded plants could recover to normal levels in previously-flooded plants, whose recover was better than that of unflooded plants. Both sexes had similar responses to W-90d and W-140 conditions in terms of majortity physiological parameters. However, females were more sensitive than males to winter flooding stress in terms of photosynthesis and chlorophyll fluorescence, based on their significant declines in A, Fv/Fm, Yield, and ETR under W-90d conditions. Females encountered more serious oxidative damage than males under flooding conditions. The results indicate that males develop more efficient antioxidant enzymatic systems to control ROS accumulation than females. Additionally, although previously-flooded females and males had strong and similar post-flooding recovery capabilities, males recovered better in terms of pigment content, especially chl a. This study provides new light on the adaptation mechanisms of P. deltoides trees subject to winter flooding stress. It also increases the understanding of sexually-dimorphic responses to flooding stress during the dormant season, and to post-flooding recovery.

Materials and Methods
Plant materials and experimental design. One-year-old cuttings of P. deltoides were collected from 25 female and 25 male trees at Qianjiang (30°09′ N, 121°31′ E), Hubei Province, China. The cuttings were planted in March 2013. After sprouting and growing for about 2 months, 360 cuttings (180 females and 180 males) with similar crown sizes and equal heights were selected and each was replanted into a 10 L plastic bucket filled with 10 kg homogenized soil. Plants were placed in a natural environment with 1261 mm mean annual rainfall, 1494 mm annual evaporation, 80% annual relative humidity and 16.9 °C annual temperature at the Wuhan Botanical Garden, Chinese Academy of Sciences. After a growing season, the branches of all plants were pruned to an identical height, i.e., 10 cm above ground level, and the leaves were removed. These plants were used for subsequent winter flooding treatments.
The experimental layout was completely randomized according to the two main factors (sex and watering rate; Fig. 8). Three watering treatment regimes were employed, being well-watered (CK), 100 days winter flooding stress combined with 40 days waterlogging (severe flooding stress, W-140d), and 50 days winter flooding stress combined with 40 days waterlogging treatment (medium flooding stress, W-90d). In the well-watered treatment, all pots were watered excessively every three days and excess water was allowed to drain through drainage holes into dishes placed under the buckets. In the winter flooding treatment, pots were watered every nine days to 5 cm above the top of the plants. For 100 days and 50 days winter-flooding treatments, the plants were submerged in December and in the next February, respectively. Both flooding treatments were simultaneously transferred to waterlogging treatments on March 29, 2014 for sprouting before the start of the growing season. During the next 40 days of waterlogging treatment, pots were watered every nine days to 5 cm above the soil surface (W-140d and W-90d). At the end of the waterlogging treatment, each plant's shoot height, basal stem diameter, gas exchange rate and chlorophyll fluorescence were measured. Fresh leaves were collected for physiological analyses, and then the excess water was drained away on May 6, 2014. During following post recovery stage, all pots were watered excessively every three days for 15 days of growth recovery (CK-R, W-140d-R, and W-90d-R) as above description. At the end of the 15 day growth recovery period, each plant's shoot height, basal stem, gas exchange rate and chlorophyll fluorescence were measured, and fresh leaves were collected for physiological analyses. All plants used for fresh leaf collection were not used in the next treatment. Five replications, each with six cuttings, were used for each treatment.
The morphological traits. At the end of each experimental stage, the plant height and basal stem diameter of each tested plants were measured. The comparative observations on leaf senescence and abscission, hypertrophied lenticels, aerenchyma tissue, and adventitious roots were performed each week.
Gas exchange measurements. The net photosynthetic rate (A), stomatal conductance (gs), intercellular CO 2 concentration (Ci) and transpiration (E) were measured from 9:00 to 11:30 am on 5, 20 May with a LI-COR 6400 portable photosynthesis system (LI-COR Inc. Lincoln, Nebr.), respectively. The PAR, provided by a 6400-02 LED light source, was set to 1400 μmol m −2 s −1 . The flow rate of air through the sample chamber was set at 500 μmol m −2 s −1 , and the leaf temperature and relative humidity was maintained at 25 ± 0.8 °C by thermoelectric coolers and 50%, respectively. The methods are modified from Yang et al. 6 and Xu et al. 16 . Five cuttings from each treatment were selected for measuring. A measurement was made on each of the three terminal leaflets of the uppermost fully opened leaf of each cutting. Instantaneous water use efficiency (WUEi = A/E) was calculated by dividing photosynthetic rate by transpiration.
Determination of chlorophyll content. Chlorophylls were extracted in 80% (v/v) chilled acetone and quantified using a spectrometer (UV-1800PC, MAPADA, Shanghai) as described by Yang et al. 6 and Xu et al. 16 . The absorbances of chlorophyll a (Chl a), chlorophyll b (Chl b), and carotenoids (Caro) were determined at 663 nm, 646 nm, and 470 nm, respectively. The absorbance values were converted to concentrations as described by Lichtenthaler 39 . And the total chlorophyll (chlorophyll a + b, Total Chl) and Chl a/Chl b were calculated.
Chlorophyll fluorescence measurements. We selected the same five cuttings and same leaves that used for gas exchange measurements for chlorophyll fluorescence measurements. Chlorophyll fluorescence kinetics parameters (Fv/Fm, maximum efficiency of PSII; Yield, the effective quantum yield of PSII; qN, non-photochemical quenching coefficient; qP, photochemical quenching coefficient; ETR, photosynthetic electron transportation rate) were measured with a PAM chlorophyll fluorometer (PAM 2500, Walz, Effeltrich, Germany). The leaf samples were placed in darkness for 30 min by covering with aluminum foil followed by measurement of minimum fluorescence (Fo) at 250 μmol m −2 s −1 PPF and Fm at 2400 μmol m −2 s −1 PPF following a saturating pulse of actinic light 6,11 . Measurements were carried out between 8:30 and 11:30 on May 6 and 21, respectively.
Determination of relative water content and relative electrolyte leakage. The fourth-sixth fully expanded leaves were sampled to determine the leaf relative water content (RWC) as described by Yang et al. 6 . Five freshly cut leaf discs (1.5 cm in diameter) from the fifth fully expanded leaves were used determined the leaf relative electrolyte leakage (REL) using a conductivity instrument (FE38, Mettler-Toledo Instruments Co., Ltd, Shanghai, China) according to procedure of Zhang et al. 17 .
Determination of soluble protein content, glutathione (GSH), reducing sugar, and free proline content. About 2 g fresh samples were ground with liquid nitrogen and then homogenized in 10 ml 100 mM universal sodium phosphate extraction buffer as described by Han et al. 40 . The supernatant was stored in volumes of 0.5 ml at -80 °C until using for the determination of soluble protein, reactive oxygen species (ROS) level and antioxidant enzymes activities 40 . The soluble protein was quantified by Bradford method 41 , the soluble protein content was expressed as mg/g · FW.