Leaf material for characterization of key photosynthetic parameters was sampled from a 4-year old T. distichum (L.) Rich at the Institute of Botany, Jiangsu Province and Chinese Academy of Sciences (32°05’ N, 118°83 ‘E) during midmorning from 9 to 10 a.m. The plant height was 2.6 m high and the diameter at breast height was 2.32 cm, which was in a rapid growth phase. Young but fully expanded and healthy branchlets with green needle-like leaves of the linear-lanceolate type were chosen for these experiments in June–July, whereas scale-like and appressed needles were disregarded.

For the long-term submergence experiment, seeds of T. distichum were sown in a seedling tray and after germination transferred to pots (23 cm upper diameter, 15 cm basal diameter, and 22 cm high) filled with a mixture of potting soil and sandy clay. The seedlings were grown in a greenhouse (temperature: 23 ± 2°C; humidity: 60%–70%) for 2 months and then moved outdoors for another 4 months, and the pots were watered daily with tap water. 30 healthy seedlings with an average height of 45 cm were selected for the experiment.

Underwater P_N was measured following the approach of Pedersen et al. (2013). In brief, artificial floodwater was prepared according to Smart and Barko (1985) with a final alkalinity of 2.0 mol H⁺ equivalent m⁻³. Considering that the underwater photosynthetic CO₂ saturation concentration of submerged leaves of terrestrial wetland plants is approximately 20–75 times or even higher than the atmospheric equilibrium concentration (~18 mmol m⁻³ free CO₂) (Pedersen et al., 2009), a CO₂ concentration range of 10–2,000 mmol m⁻³ was set. In addition, compared with high CO₂ concentration, underwater P_N changes more significantly under low CO₂ concentration; thus, six concentration gradients (10, 25, 50, 100, 200, and 500 mmol m⁻³) were set for low concentration and three concentration gradients (1,000, 1,500, and 2,000 mmol m⁻³) were set for high concentration. Prior to the pH adjustment, the solution was purged with N₂ to reduce the O₂ concentration approximately 30%–50% of air equilibrium to prevent photorespiration during incubation (Setter et al., 1989). The artificial floodwater was then siphoned into 44-mL glass vials and two pieces of 3-mm glass beads were added to each vial to ensure the mixing during incubation. One branchlet with needles (approximately 1 cm² or approximately 15.8 mg fresh mass) before the vial was sealed with a glass lid (no headspace or gas bubbles present). The vials were mounted on a vertically rotating disk (10 rpm) and inundated in a constant temperature bath at 25°C and illuminated with a photon flux of 1,000 µmol m⁻² s⁻¹ (see below). Vials without tissue served as blanks.

After 60 min, the vials were retrieved and the O₂ concentration was measured using an O₂ optode (OPTO-MR, Unisense, Denmark) inserted into the vial. The needles in the vial were then neatly placed on a clean white background board while a ruler is placed to take the photo. Make sure the leaves do not overlap each other when taking the photo; ImageJ software was used to measure the exact leaf area (Schneider et al., 2012). The photosynthetic rate was calculated using the following equation:

where ΔO₂ is the difference in O₂ concentration in vial with tissue and blanks, V _vial is the volume of the vials, i.e., 0.044 L, t is the incubation time, and A is the area of the needles.

Floodwater and tissue were prepared as above but with a fixed CO₂ concentration of 500 mmol m⁻³. Compared with high light intensity, underwater P_N changes more significantly under low light intensity; thus, six gradients (0, 50, 100, 200, 300, and 500 µmol photons m⁻² s⁻¹) were set for low light intensity and three gradients (1,000, 1,500, and 2,000 µmol photons m⁻² s⁻¹) were set for high light intensity. Different light intensities were established by using a high-pressure Na lamp at various distances and by regulating the voltage. For zero light, the vials were wrapped in aluminum foil. The light intensities were measured using a spherical PAR (photosynthetically active radiation) sensor (QSL2101, Biospherical Instruments Inc., USA). As for the CO₂ response (see above), samples were incubated for 60 min.

Floodwater and tissue were prepared as above but with a fixed CO₂ concentration of 500 mmol m⁻³ and PAR at 1,000 µmol photons m⁻² s⁻¹. The temperature was set to 10, 15, 20, 25, 30, or 35°C using a combination of an immersion heater (300W, SUNSUN, Zhejiang, China) and a water cooler (TECO, Taiwan, China) to achieve a stable temperature, which was monitored in real time by a temperature electrode (Temp-UniAmp thermosensor, Unisense, Denmark). As for the CO₂ response (see above), samples were incubated for 60 min.

A 30-day submergence experiment with two treatments (drained controls or completely submerged) was conducted using 6-month-old T. distichum seedlings. Fifteen pots with one plant in each were transferred to three plastic tanks (depth, 69 cm; volume, 122 L) filled with tap water, with five seedlings (technical replicates) in each tank (true replicates). The water was changed every 2 days, and the hose was placed at the bottom of the bucket to ensure that the water was completely replaced. Another 15 plants served as controls, and these were unsubmerged and kept under conditions (photoperiod and temperature) similar to those for submerged plants and irrigated every 2 days. Underwater P_N was measured on healthy needles on days 7, 14, 21, and 31 using a PAR of 1,000 µmol photons m⁻² s⁻¹ and a CO₂ concentration of 500 mmol m⁻³ (see experimental procedure above) and leaves were transported in water back to the laboratory to minimize damage and exposure to air. It should be noted that the underwater P_N of the submerged leaves of terrestrial wetland plants will be severely limited by the CO₂ availability at atmospheric equilibrium CO₂ concentrations. In order to more accurately evaluate the underwater photosynthetic capacity of submerged leaves, a CO₂ concentration higher than atmospheric equilibrium was required, and 500 mmol m⁻³ is set for this experiment.

During the first 8 days of the submergence experiment, submerged leaves were harvested every other day. These were observed and photographed with scanning electron microscopy (SU8100, Hitachi Scientific Instruments, Japan). Subsequently, stomatal density (number of stomata per mm⁻²) and stomatal index (ratio of number of stomata to the total number of epidermal cells including stomata) were calculated from these images (Hegde and Krishnaswamy, 2021; Li et al., 2022).

At the end, healthy leaves were sampled and then transported into water to minimize damage and re-exposure to air. Chlorophyll measurements were conducted on submerged leaves as well as controls using ethanol extractions and absorbance of the extract was measured on a spectrophotometer (UV-1800, Shanghai Mepuda Instrument Co., LTD, China). Chlorophyll was calculated using the equations in Zhang et al. (2020).

Finally, cross-sections for microscopy were prepared from paraffin-embedded needles and later studied using visible light microscopy (BX53F, Olympus, Tokyo, Japan).

To investigate the effects of leaf gas films on underwater P_N , four healthy 6-month-old T. distichum seedlings were selected. Two fully unfolded branchlets were sampled from each plant and then they were divided into two groups. To remove hydrophobicity and thereby prevent formation of leaf gas films, one group was brushed five times, on both sides, with a fine paintbrush dipped into 0.01% (v/v) Triton X. After that, they were washed for 5 s, three times, in artificial floodwater without Triton X (Teakle et al., 2014; Winkel et al., 2014). The other group was untreated and served as control with each group having four replicates. Underwater P_N was measured as described above with 500 mmol CO₂ m⁻³ under a photon flux of 1,000 µmol photons m⁻² s⁻¹ at 25°C.

Leaf gas film thickness was measured following the approach of Raskin and Kende (1983). In brief, the buoyancy of a branchlet was measured on five replicates using a four-digit balance with a hook underneath before and after removal of hydrophobicity using 0.01% Triton X; see above. Next, the area of the needles was determined (see above), and the gas film thickness (m) was calculated as gas film volume (m³) divided by needle area (m²).

Surface hydrophobicity was assessed by measuring the contact angle of a 1-mm³ droplet of water on the needle surfaces following Sikorska et al. (2017). Branchlets with needles were held horizontal using a glue stick. Water droplets were applied to the lamina of 10 replicate needles (5 on the adaxial side and 5 on the abaxial side), and photographed at ×35 magnification using a horizontally positioned dissecting microscope (MZ62, Mshot, China) and a digital camera. The droplet contact angles were measured using ImageJ (ImageJ v.1.43U, National Institutes of Health, Bethesda, MD, USA).

Finally, specific leaf area (SLA) was measured by determining the area (see above) and dry mass of needles, where the needles were dried for 48 h at 60°C. SLA was calculated as area (m²) divided by dry mass (kg).

We used non-linear regression to fit models derived from FvCB (CO₂ response) (Liang and Liu, 2017) and light response was fitted according to the Ye model (Ye et al., 2013). The temperature optimum was modeled using a Gaussian function and a standard exponential function was used to predict temperature coefficient Q ₁₀ (Pedersen et al., 2016). The temperature coefficient Q ₁₀ represents the relative change of underwater P_N with every 10°C change in temperature. The data were processed using Excel 2016 and graphed with Origin software (2021 64Bit, Electronic Arts Games, USA). Results were expressed as means ± standard deviation. All statistical tests were conducted using SPSS 16.0s (SPSS Inc., USA) including that of the Duncan’s multiple range test. Unless otherwise stated, a probability level of 0.05 was used.

Waterlogging, partial submergence, and even complete submergence of Taxodium is a recurrent phenomenon during times of high water levels. When the water first starts rising, the soil becomes flooded, resulting in waterlogging of vast areas of trees ( Figure 1A ), and as the water continues to rise, the lower branches become submerged ( Figures 1B, C ). Ultimately, the entire canopy is under water ( Figure 1D ), and gas exchange with the atmosphere is no longer possible. The floodwater in the Yangtze River is murky as a result of suspended materials ( Figure 1D ), and therefore the light environment also changes upon submergence, resulting in reduced photosynthesis due to the combination of low light and restricted CO₂ availability. The habitat photos in Figure 1 clearly demonstrate the relevance of our study, where we are aiming to characterize the ability of T. distichum to continue photosynthesizing under water with emphasis on response to CO₂ and light availabilities, and temperature.

We characterized the response of underwater net photosynthesis (P_N ) to dissolved CO₂, light availability, and temperature, which are relevant environmental parameters during submergence of T. distichum. Submerged branchlets with needles produced O₂ when incubated in the artificial floodwater in the light visible as bubble formation on the needles ( Figure 2A ). The response of P_N (i.e., net O₂ consumption) to dissolved CO₂ at 25°C and a photon flux of 1,000 µmol photons m⁻² s⁻¹ showed a typical saturation response, increasing as CO₂ was raised to 2,000 mmol m⁻³ ( Figure 2B ). The FvCB model estimated the maximum carboxylation rate (V _cmax), maximum electron transfer rate (J _max), and day respiratory rate (R _day) to be 19.73 µmol m⁻² s⁻¹, 39.34 µmol m⁻² s⁻¹, and 0.3 µmol m⁻² s⁻¹, respectively. However, the CO₂ compensation point (Γ*) and CO₂ saturation point (C _i,TUP) are 37.43 mmol m⁻³ and 2,334 mmol m⁻³, which are approximately 2-fold and 130-fold atmospheric equilibrium (~18 mmol m⁻³ free CO₂), respectively ( Table 1 ).

We diagnosed P_N at contrasting light availabilities with 500 µmol CO₂ L⁻¹ in the floodwater. As expected, underwater P_N also followed a saturation response with increasing light availability. In darkness, the dark respiration (R _dark) of branchlets with needles was 1.31 µmol O₂ m⁻² s⁻¹ ( Figure 2C ). Using the equation from Ye et al. (2013), the light compensation point (I _c) and light saturation point (I _m) were 65.38 µmol photons m⁻² s⁻¹ and 1,666.67 µmol photons m⁻² s⁻¹, respectively. P_{N max} was estimated to 10.97 µmol O₂ m⁻² s⁻¹ at the given environmental conditions, i.e., dissolved CO₂ at 500 µmol L⁻¹ at 25°C ( Table 1 ).

Underwater P_N followed an exponential increase with increasing temperature in the tested interval from 10 to 25°C whereafter it steeply decreased with increasing temperature. Using a Gaussian model, we estimated the temperature optimum for underwater P_N in T. distichum to 25°C with 500 µmol L⁻¹ dissolved CO₂ in the floodwater at a photon flux of 1,000 µmol photons m⁻² s⁻¹ ( Figure 2D ). Using the same dataset, but without considering temperatures exceeding the optimum for underwater P_N , we estimated the Q ₁₀ of P_N to 1.84 demonstrating the strong dependence of underwater P_N on environmental temperature.

At the onset of submergence, T. distichum forms a thin gas film on its needles and therefore we aimed at characterizing hydrophobicity, gas film thickness, and other key features known to influence underwater P_N . Macroscopically, the needles are very similar on their adaxial and abaxial sides, but stomatal density differs with more than fourfold higher density of stomata on the abaxial side ( Figures 3A–D, G ). However, the water-repellent traits were similar, showing a contact angle of 146° on both sides, and although these angles only render the needles hydrophobic (and not superhydrophobic) (Koch and Barthlott, 2009), the hydrophobicity was nevertheless sufficient to initially retain a 35-µm-thick gas layer upon submergence ( Figures 3E–G ). The needles of T. distichum in air have been observed to repel water, and the gas film formed underwater is directly visible as a silvery sheen ( Figures 3H, I ), and its well-described facilitation of underwater photosynthesis is evident from the bubble formation when the branchlets are submerged in CO₂-rich water in the light ( Figure 3J ). We also manipulated needle hydrophobicity to enable a direct comparison of underwater P_N of needles with or without a gas film, and we found that the gas film increased underwater P_N 2.1-fold as compared with needles where gas film formation was prevented ( Figure 3K ).

The fact that Taxodium can become completely submerged for several months prompted us to conduct a controlled laboratory experiment where we submerged 6-month-old plants for 30 days. During the first 24 days of submergence, the hydrophobicity was lost and the leaf cuticle gradually became colonized with bacteria with the first bacterial cells appearing already after 2 days of submergence ( Figure 4A ). The colonization of bacteria was accompanied by a decline in hydrophobicity, and leaf gas films dramatically decreased during the first few days of submergence ( Figure 4B ). This loss of gas films resulted in a significant decline in underwater P_N as indicated by the significant positive correlation between gas film thickness and under P_N ( Figure 4C ). Key stomatal features (stomatal density and stomatal index) did not change during the first 8 days of submergence ( Figure 4D ).

However, over the entire submergence period of 30 days, significant changes took place at the needle level. The needles started yellowing ( Figure 5A ), and the yellow was also reflected in a significant decline in chlorophylls ( Figure 5B ). The combined effect of loss of hydrophobicity and the decline in chlorophylls resulted in a steep decline in underwater photosynthesis already within the first 8 days of submergence with a predicted T _½ in P_N of 1.85 days ( Figure 5C ). In addition to the biochemical changes in chlorophyll concentration, the needles also underwent anatomical changes during the 30 days of submergence as all of the palisade cells degraded ( Figures 5D, E ). Importantly, even with the loss of palisade tissues and the significant declines in chlorophylls, the needles maintained some capacity for underwater P_N during the entire submergence period as underwater P_N never declined below 1 µmol O₂ m⁻² s⁻¹.

Terrestrial leaves generally perform poorly under water due to the restricted gas exchange in water compared with air resulting in restricted O₂ uptake for respiration or CO₂ uptake for photosynthesis. The poor performance has been clearly demonstrated using the model plant, Rumex palustris, showing that underwater P_N of aerial leaves was only 0.5% of the rate in air (Mommer et al., 2006). In stark contrast, leaves of R. palustris formed under water could attain rates of underwater P_N at 35% of that in air showing the great benefit of leaf acclimation to underwater gas exchange. However, leaf acclimation is only a feasible strategy for long-term submergence, as production of new aquatic leaves requires reallocation of carbohydrates to fuel leaf. Instead, superhydrophobic leaves that retain a gas film under water have been shown to be a very competitive solution to enhance gas exchange without further investment in leaf acclimation (Colmer and Pedersen, 2008).

Hydrophobicity can be characterized using the contact angle of a microscopic water droplet. Accordingly, leaf cuticles with contact angles exceeding 150° are classified as superhydrophobic (with rice being a typical example) (Kwon et al., 2014), whereas those with contact angles less than 150°—but larger than 90° (Koch and Barthlott, 2009)—are hydrophobic. In the case of T. distichum, the contact angles were just below the 150° cut (146°, Figure 3G ), but the needles nevertheless retained a gas film upon submergence ( Figure 3I ). The gas films forming on the needles of T. distichum were of similar thickness (35 µm, Figure 3G ) to those formed by rice leaves (30–60 µm) (Winkel et al., 2014; Kurokawa et al., 2018; Mori et al., 2019) or by wheat leaves (20–40 µm) (Konnerup et al., 2017). Consequently, it is not surprising that the beneficial effects on gas exchange and underwater P_N were significant.

In water, gas diffusion is slow and therefore physiological processes relaying on gas exchange can become restricted by slow substrate supply such as CO₂ for photosynthesis. However, leaf gas films greatly enhance gas exchange (Verboven et al., 2014), and we found that needles with gas films achieved twofold higher photosynthetic rates compared with needles that had the hydrophobicity removed and where gas films therefore did not form ( Figure 3K ). It has previously been found that leaf gas films can increase underwater P_N up to three- to sevenfold (Colmer and Pedersen, 2008), but the photosynthesis in these experiments was assessed at lower (200 µM) external CO₂ concentrations, where the beneficial effect of gas films on gas exchange is more pronounced (Winkel et al., 2017). Interestingly, the underwater P_N obtained at saturating light and CO₂ levels in the present study matched those of P_N in air; i.e., in both environments, the rates were approximately 10–12 µmol O₂ m⁻² s⁻¹ ( Figures 2B, C ) (Neufeld, 1983; Mommer et al., 2006), underlining the significant effect of gas films on the needles of T. distichum. Consequently, we propose that a key reason for the stunning flood tolerance of T. distichum is its ability to maintain a substantial photosynthetic activity during submergence resulting in both carbohydrate and O₂ production. However, the realized photosynthetic rates greatly depend on both light and CO₂ availability under water and, to a large extent, temperature as well.

CO₂ uptake by T. distichum followed a classical FvCB response curve. However, underwater P_N remained negative until the CO₂ compensation point at 37.43 mmol m⁻³ was reached; below, the needles of T. distichum consumed more O₂ than they produced. The CO₂ compensation point is equivalent to approximately twofold that of atmospheric equilibrium (~18 mmol m⁻³ free CO₂), demonstrating the importance of net CO₂ production in the floodwater in order for the underwater P_N to become positive and therefore result in significant carbohydrate production. The C _i,TUP (CO₂ saturation point) was estimated to 2,334 mmol m⁻³, or 130-fold atmospheric equilibrium, which is higher than that of submerged rice (Winkel et al., 2013) and submerged wheat (Winkel et al., 2017), and even higher than that of submerged Hordeum marinum (Pedersen et al., 2010). It has been demonstrated that the underwater P_N capacity of submerged plants was severely limited at atmospheric equilibrium CO₂ concentrations (Pedersen et al., 2009). Although some studies found that the CO₂ concentration recorded in flooded rice fields was 20–180 times the atmospheric equilibrium concentration (360–3,240 mmol m⁻³) (Setter et al., 1987), the diffusion rate of the gas in water is 10,000 times lower than in air, resulting in underwater P_N being still limited (Pedersen et al., 2013). CO₂ availability limitations are a long-standing challenge for submerged plants, but interestingly, rice, wheat, and H. marinum retain leaf gas films when submerged, which has been shown to significantly enhance gas exchange. The overall similarity of the CO₂ response in T. distichum to the other three terrestrial species with leaf gas films is likely due to the physical effect of the gas films facilitating the exchange between needles and floodwater rather than physiological similarities among these distantly related species.

Light utilization by T. distichum under water also showed a saturating response to light when assessed with 500 µmol CO₂ L⁻¹ in the floodwater. As for CO₂, underwater P_N was initially negative at low light levels and only reached positive values (I _c) at PAR > 65.38 µmol photons m⁻² s⁻¹ ( Figure 2C ). When PAR is higher than the saturation point (I _m) of 1,666.67 µmol photons m⁻² s⁻¹, underwater P_N reaches a maximum value (P_{N max}) of 10.97 µmol m⁻² s⁻¹, which is very difficult to achieve due to light loss; thus, underwater P_N is generally limited by PAR. The initial quantum efficiency (α) of 0.0221 absorbed was in the same order of magnitude as that of submerged wheat (Winkel et al., 2017) and Phalaris arundinacea (a terrestrial wetland species also forming leaf gas films upon submergence) (Vervuren et al., 1999; Winkel et al., 2017), whereas two other species without gas films utilized light much less efficiently (Rumex crispus and Arrhenatherum elatius) (Vervuren et al., 1999). These findings emphasize the importance of gas films also for light use efficiency as the light use relies not only on incident light reaching the leaf surfaces but also on entry of CO₂.

In addition to CO₂ and light, underwater P_N of T. distichum was also strongly affected by the environmental temperature. In the temperature range tested, i.e., 10 to 35°C, there was a strong positive relationship between temperature and underwater P_N until the temperature optimum was reached at 25°C after which P_N declined with increasing temperature ( Figure 2D ). This is a common response of underwater P_N to rising temperature as also demonstrated for two tidal seagrass species, Thalassia hemprichii and Enhalus acoroides (Pedersen et al., 2016). Being tropical species, these seagrasses showed a temperature optimum for underwater P_N at 33°C, i.e., 8°C above that of T. distichum. The Q ₁₀ of underwater P_N in T. distichum was somewhat lower (1.8) than that of T. hemprichii (2.0) and E. acoroides (2.8) (Pedersen et al., 2016), but it nevertheless show the strong dependency of temperature for underwater P_N also in T. distichum. This is an important point to consider when extrapolating the current laboratory findings to the field situation since the water temperature in the Yangtze River can fluctuate from 11 to 22°C during the time of the year when the trees on the banks become submerged (Yu et al., 2021).

The trees growing on the banks of the Yangtze River can become partially or completely submerged for up to 120 days and still survive (Yang et al., 2023), and we therefore tested the response of T. distichum to long-term submergence. Six-month old seedlings were completely submerged for a period of 30 days with sampling of leaf tissue during the period at discrete time points. Leaf gas films quickly diminished, and the decline in gas film thickness was accompanied by a decline in underwater P_N ( Figure 4C ). The strong relationship between gas film thickness and underwater P_N during long-term submergence has previously been observed in rice (Winkel et al., 2014), but our study represents the first to demonstrate this relationship for a tree species.

The gas films forming on the surfaces of the submerged needles of T. distichum persist longer than in other species tested with hydrophobic cuticles. In the present study, the gas films were detectable up to 24 days of submerged after which they had totally vanished ( Figure 4B ). This is longer than observed for rice, which represents the only other species with superhydrophobic leaves, where gas film thickness have been followed during a submergence event. Here, it was found that gas film thickness was below the detection limit already after 7 days of submergence (Winkel et al., 2014). In T. distichum, the needles maintained their ability to photosynthesize also after the gas films were lost at a rate of approximately 1 µmol O₂ m⁻² s⁻¹, which is similar to the photosynthetic rate of rice and wheat once the leaf gas films of these species have also disappeared due to long-term submergence (Winkel et al., 2014; Konnerup et al., 2017). It is still not fully understood why the loss in hydrophobicity occurs during submergence. However, the present study as well as one on wheat (Konnerup et al., 2017) clearly demonstrated that a biofilm was established during submergence ( Figure 3I ), but if this biofilm is the result or the cause of loss of hydrophobicity remains unknown.

Distinct anatomical and biochemical changes occurred during the long-term submergence event. The decline in total chlorophylls was significant with initial values at 2.1 mg g⁻¹ DM and only 1.0 mg g⁻¹ DM after 30 days of submergence ( Figure 5B ). While this decline will have significant consequences for the light capturing capabilities, the decline in submerged rice was even more pronounced where the initial levels are at approximately 18 mg g⁻¹ DM to less than 5 mg g⁻¹ DM in only 2 weeks. The decrease in chlorophyll concentration observed in the present study may be one of the important reasons for the decrease in underwater P_N , as it has been shown that underwater P_N of rice is positively correlated with leaf chlorophyll concentration (Winkel et al., 2014). The significant decrease in chlorophyll concentration is likely a result of palisade tissue loss in needles ( Figures 5D, E ). To our knowledge, similar observations are missing in the literature as previous studies have focused on anatomy of leaves formed during the submergence event (Mommer et al., 2007) and not on acclimation of already existing leaves. The lysis of palisade tissues observed towards the end of the 30-day submergence event might be accompanied by water infiltration in the newly formed cavities, and such water-filled cavities would slow down intra-tissue diffusion of O₂ and CO₂ (Armstrong, 1980). Interestingly, the parallel decline in chlorophyll concentration and leaf gas film thickness makes it difficult to identify the primary causal effect of the observed decline in underwater P_N .