Dendrobium officinale collected from Jinhua city (Zhejiang, China) and D. primulinum collected from Puer city (Yunnan, China) were cultivated and acclimatized in a walk-in phytotron at China Agricultural University (Beijing, China) for 2 years. K. daigremontiana, an obligate CAM plant, used as a reference, was grown from leaf-borne ramets in the same walk-in phytotron. The two Dendrobium species and K. daigremontiana were planted in 0.4 L plastic pots, three plants per pot. The cultivation substrate was moss for these two Dendrobium species, and a mixture of vermiculite and perlite (volume ratio was 3:1) for K. daigremontiana. All pots were irrigated with 70 ml tap water every 2 days, and irrigated with the same amount of nutrient solution every 7 days for 2 years. Nutrient solution formula was as follows: Ca(NO₃)₂•4H₂O 205 mg L⁻¹, MgSO₄•7H₂O 60 mg L⁻¹, KH₂PO₄ 136 mg L⁻¹, NH₄NO₃ 80 mg L⁻¹, MnSO₄•H₂O 3.6 mg L⁻¹, H₃BO₃ 2.7 mg L⁻¹, FeSO₄•7H₂O 13.4 mg L⁻¹, CuSO₄•5H₂O 0.1 mg L⁻¹, ZnSO₄•7H₂O 0.4 mg L⁻¹, and (NH₄)₆Mo₇O₂₄•4H₂O 0.1 mg L⁻¹. EC and pH of the nutrient solution were controlled at 0.6–0.7 ms cm⁻¹ and 6.0–6.5, respectively. The environmental parameters of the phytotron for 2 years of long-term cultivation were as follows: artificial light source was T5 tricolor fluorescent lamp (28 W, Beijing Lighting Valley Co., Ltd., Beijing, China); the photosynthetic photon flux density (PPFD) at the plant canopy was 150 μmol m⁻² s⁻¹; light/dark cycle was 12/12 h (light period, 0800–2000 hours; dark period, 2000–0800 hours); temperature during the light and dark periods was 26°C ± 1 and 22°C ± 1°C, respectively; relative humidity was 65% ± 5%; CO₂ concentration was 500 ± 50 μmol mol⁻¹.

Different light/dark cycle treatments began after acclimatization in the environment of phytotron for 3 months, when both D. officinale and D. primulinum plants had 8–10 expanded leaves, and K. daigremontiana plants had 16 expanded leaves. All the plants used in this experiment were in a vegetative stage when they were subjected to different light/dark cycle treatments. Twelve similar-sized healthy plants were selected from each species treating with a light/dark cycle of 12/12 h (light period, 0800–2000 hours; dark period, 2000–0800 hours) for five cycles (5 days). At the end of the fifth dark period, the light/dark cycle was changed to 4/4 h for 15 cycles (5 days). After the 15th dark period (the end of the fifth day), the light/dark cycle was changed back to 12/12 h for another five cycles (5 days) as mentioned above. The light intensity, temperature, relative humidity, and CO₂ concentration during the light/dark cycle treatment were maintained at the same levels as mentioned above.

The photosynthetic continuous measurement system (Zhang et al., 2014) used in this study consists of four cuvettes (25 cm × 15 cm × 6 cm), a host computer, and an IRGA CO₂ analyzer (LI-7000, LICOR, Lincoln, USA). One shoot of D. officinale and one shoot of D. primulinum, each with 8–10 leaves, as well as two fully expanded mature leaves from two different K. daigremontiana plants were selected for this measurement. Each of the four different samples was enclosed into each of the different cuvettes. All leaves were held flat by several horizontal nylon wires. All plants were kept intact and irrigated with 70 ml tap water once a day during the measurement. The measurement was repeated for three times. The temperature and relative humidity of the cuvettes was the same level as that of the phytotron as mentioned above. PPFD at the top of the cuvettes was 150 μmol m⁻² s⁻¹. The air flow rate of each cuvette was 1.0 L min⁻¹. The difference of CO₂ concentration between reference and sample gas of each cuvette recorded every 10 min throughout the treatment period. The measurements for these three species were carried out concurrently. Leaf area of each cuvette was determined according to Yang et al. (2002) after 15 days. Then net CO₂ exchange rate of each cuvette was calculated according to Zhang et al. (2014). Dark net CO₂ exchange amount, light net CO₂ exchange amount, and daily net CO₂ exchange amount were integrated based on dark, light, and daily net CO₂ exchange rates everyday (24 h). The dark net CO₂ exchange percentage was defined as dark net CO₂ exchange amount divided by daily net CO₂ exchange amount times 100%. Dark, light, and daily net CO₂ exchange amount, as well as dark net CO₂ exchange percentage measurement data for these three species were collected at the last 2 days of each light/dark cycle (days 4 and 5, days 9 and 10, days 14 and 15). Three replicates were conducted in this measurement for each species.

A leaf porometer (SC-1, Decagon, Washington, USA) was used to measure the stomatal conductance of abaxial surface of the leaves of other plants outside the cuvettes for D. officinale, D. primulinum, and K. daigremontiana. Daily desiccant replacement and the leaf porometer calibration were done before the measurement. Automatic mode was used to obtain each value in 30 s. Measurements were conducted at 0700, 0900, 1,300, 1,600, 1900, 2100, and 2300 hours on day 5 (0700, 2100, and 2300 hours were in dark period), and at 0200, 0600, 1000, 1400, 1800, and 2,200 hours on day 9 (0600, 1400, and 2200 hours were in dark period). Four plants per species were measured at each time. The stomatal conductance were measured for the third leaf from top of each plant (n = 4).

Statistics analysis was performed using the IBM SPSS Statistics 21 (IBM, Inc., Armonk, NY, USA). The average dark, light, and daily net CO₂ exchange amount, as well as dark net CO₂ exchange percentage for each species (D. officinale, D. primulinum, and K. daigremontiana) were compared respectively between different light/dark cycles by Duncan’s multiple range test at p < 0.05. The average stomatal conductances at each time of day 5 and day 9 were compared for each species by the same method as mentioned above, respectively.

The net CO₂ exchange exhibited obvious trailing phenomenon during the conversion between the light and dark period. Thus its influence had been considered in the subsequent results analysis. When light/dark cycle was 12/12 h, net CO₂ exchange rates of D. officinale, D. primulinum, and K. daigremontiana exhibited multiple periodic variations. The net CO₂ exchange rates of D. officinale, D. primulinum, and K. daigremontiana increased successively after onset of the dark period. When switched to light period, the net CO₂ exchange rate of K. daigremontiana first rose to the peak and then fell to near zero at 0900 hours; that of D. officinale and D. primulinum first increased and then decreased until 1600 hours before falling to near zero. The net CO₂ exchange rates of D. officinale exhibited the shortest time of fluctuation around zero, followed by D. primulinum, that of K. daigremontiana exhibited the longest duration near zero. The net CO₂ exchange rates of all these three species would rise at the end of the light period (Figure 1A). After light/dark cycle was changed to 4/4 h for 3 days (on day 9), net CO₂ exchange of D. officinale and D. primulinum showed C3-like pattern, with net CO₂ uptake in light period and net CO₂ release in dark period, whereas net CO₂ uptake in dark period maintained in K. daigremontiana. Net CO₂ exchange rate of K. daigremontiana at 0200 hours was higher than that at 1000 and 1800 hours (Figure 1B). After the light/dark cycle was changed back to 12/12 h for 1 day, net CO₂ exchange of these three species also switched back to the similar pattern of earlier light/dark cycle of 12/12 h (Figure 1C).

For D. officinale, daily net CO₂ exchange amount increased significantly from 47 to 69 mmol m⁻² day⁻¹ and then further increased significantly to 85 mmol m⁻² day⁻¹ when the light/dark cycle was changed from 12/12 h to 4/4 h and then back to 12/12 h (Figure 2A). After the light/dark cycle was changed from 12/12 h to 4/4 h, the increase of daily net CO₂ exchange amount mainly resulted from the increase of light net CO₂ exchange amount. Compared to 4/4 h light/dark cycle, daily net CO₂ exchange amount significantly increased after light/dark cycle was changed to 12/12 h, due to the increase of dark net CO₂ exchange amount was more than the decrease of light net CO₂ exchange amount. Compared to the previous 12/12 h light/dark cycle, daily net CO₂ exchange amount significantly increased after light/dark cycle was changed back to 12/12 h, due to both the increase of light net CO₂ exchange amount and dark net CO₂ exchange amount.

For D. primulinum, daily net CO₂ exchange amount decreased significantly from 76 to 40 mmol m⁻² day⁻¹ and then remained no significantly changed when the light/dark cycle was changed from 12/12 h to 4/4 h then back to 12/12 h (Figure 2B). After the light/dark cycle was changed from 12/12 h to 4/4 h, the decrease in daily net CO₂ exchange amount mainly resulted from the large decrease of dark net CO₂ exchange amount, which offset against the increase of light net CO₂ exchange amount. Compared to 4/4 h light/dark cycle, there was no significant change in daily net CO₂ exchange amount after the light/dark cycle was changed to 12/12 h. During this light/dark cycle switch, the increase of dark net CO₂ exchange amount was offset by the decrease of light net CO₂ exchange amount. Compared to the previous 12/12 h light/dark cycle, daily net CO₂ exchange amount significantly decreased after light/dark cycle was changed back to 12/12 h, due to the decrease of both light net CO₂ exchange amount and dark net CO₂ exchange amount.

For K. daigremontiana, there were no significant changes in daily net CO₂ exchange amount during the conversion of different light/dark cycles (Figure 2C). Compared to 12/12 h light/dark cycle, the daily net CO₂ exchange amount did not change significantly after the light/dark cycle was changed from 12h/12 h to 4/4 h. During this light/dark cycle switch, the increase of light net CO₂ exchange amount was almost equivalent to the decrease of dark net CO₂ exchange amount. Compared to the previous 12/12 h light/dark cycle there was no significant change in daily net CO₂ exchange amount when light/dark cycle was changed back to 12/12 h, due to no significant change in light and dark net CO₂ exchange amount.

The dark net CO₂ exchange percentage of D. officinale, D. primulinum, and K. daigremontiana showed an increasing trend during 12/12 h light/dark cycle (Table 1). When the light/dark cycle was changed from 12/12 h to 4/4 h and then back to 12/12 h, the dark net CO₂ exchange percentage of D. officinale and D. primulinum changed from a positive value to a negative value and then back to a positive value. Compared to that of the original 12/12 h light/dark cycle, the positive value was significantly decreased for these two Dendrobium species after the light/dark cycle was changed back to 12/12 h light/dark cycle from 4/4 h. The dark net CO₂ exchange percentage of K. daigremontiana always remained positive and showed a trend of decreasing first and then increasing during the different light/dark cycles conversion. For K. daigremontiana, there was no significant difference in the dark net CO₂ percentage of 12/12 h light/dark cycle before (85.5%) and after (91.3%) 4/4 h light/dark cycle.

Statistically, there was a significant decline in stomatal conductance of D. officinale (Figure 3A), D. primulinum (Figure 3B), and K. daigremontiana (Figure 3C) from 0700 (dark period) to 1300 hours (light period); however, stomatal conductance of all these three species increased significantly from 1300 to 1600 hours during the light period of the 12/12 h light/dark cycle. After light/dark cycle was changed to 4/4 h for 3 days (day 9), the stomatal conductance of D. officinale decreased in the dark period and increased in the light period significantly. Although the stomatal behavior of D. primulinum exhibited the similar pattern as those of D. officinale, the amplitude was much smaller. For K. daigremontiana, it showed a gradual decrease in the stomatal conductance from 0200 (light period) to 2200 hours (dark period) 3 days after changing the light/dark cycle from 12/12 h to 4/4 h. Compared to the previous 12/12 h light/dark cycle. After light/dark cycle was changed from 12/12 h to 4/4 h, for D. officinale and D. primulinum, stomatal conductance of the light period significantly increased, whereas it was not affected for K. daigremontiana.

Based on the percentage of daily carbon gained by dark fixation, different extent of CAM plants could be easily distinguished (Winter and Holtum, 2002; Zhang et al., 2014; Winter, 2019). Therefore, D. officinale was identified as a C3-CAM plant, whereas D. primulinum was a CAM plant between the C3-CAM and full CAM. After the light/dark cycle was changed from 12/12 h to 4/4 h for 3 days, net CO₂ exchange pattern of D. officinale and D. primulinum switched from CAM defined by Osmond (1978) to C3-like (net CO₂ uptake in the light and net CO₂ release in the dark, Figure 1). Stomatal behavior of these two Dendrobium species also switched from the CAM pattern to the C3-like pattern (increasing in light period and decreasing in dark period, Figure 3). Tallman (2004) suggested that in phase III of CAM, the photosynthesis of guard cells can obtain a large amount of CO₂ from the mesophyll and build a strong sink for NADPH, thereby inhibiting the degradation of endogenous ABA in guard cells and promoting stomatal closure (Lind et al., 2015). Therefore, it can be speculated that when the dark CO₂ absorption decreases to a certain extent, the mesophyll cells will not have enough CO₂ supply to the guard cells in the light period, so as not to inhibit the opening of the stomata. Some studies related to facultative CAM plants have concluded that light regulates stomatal conductance of these plants only when they are in C3 pattern (Lee and Assmann, 1992; Tallman et al., 1997). Our results were consistent with these previous studies. Net CO₂ uptake in dark period generally existed in K. daigremontiana regardless of light/dark cycle (Figure 2C). However, there were no significant changes in the stomatal conductance during both dark and light periods from 0600 to 2200 hours after changing the light/dark cycle from 12/12 h to 4/4 h (Figure 3). Dark CO₂ fixation was almost exclusively catalyzed by PEPC in CAM species. The allosteric performance of PEPC was regulated posttranslationally by a circadian clock controlled protein kinase called phosphoenolpyruvate carboxylase kinase (PPCK) (Nimmo et al., 1984, 1986, 1987; Hartwell et al., 1999; Owen and Griffiths, 2013; Boxall et al., 2017; Ohara and Satake, 2017). The activity of PPCK of Bryophyllum fedtschenkoi reach appeared several hours after the onset of darkness (Carter et al., 1991). It was possible that when the dark period was too short, the activities of PPCK might not be high enough, which might inhibit the dark net CO₂ fixation in D. officinale, D. primulinum, and K. daigremontiana to different extents (Figure 2). However, this study did not measure the activity of PPCK of these three species during the conversion of different light/dark cycles, which needs to be studied further in the future.

After the light/dark cycle was changed back to 12/12 h light/dark cycle, dark net CO₂ exchange percentage of two Dendrobium species decreased significantly compared to the original 12/12 h light/dark cycle, but that of K. daigremontiana did not (Table 1). However, it did not imply that the 4/4 h light/dark cycle induction suppressed the CAM activity of D. officinale. Compared to the original 12/12 h light/dark cycle, dark net CO₂ exchange amount increased significantly for D. officinale after the light/dark cycle was changed back to 12/12 h (Figure 2A).

After the light/dark cycle was changed from 12/12 h to 4/4 h, daily net CO₂ exchange amount increased by 47% for D. officinale (Figure 2A). It might be due to the fact that sharp increased stomatal conductance (Figure 3) were favorable to the diffusion of CO₂ into the leaf, eventually promoting the CO₂ absorption during the light period of 4/4 h light/dark cycle. After the light/dark cycle was changed from 4/4 h back to 12/12 h, daily net CO₂ exchange amount of D. officinale reached a higher value compared to 4/4 h light/dark cycle (Figure 2A) due to a little down regulation of C3 activity (represented by light net CO₂ exchange amount) and large up regulation of CAM activity (represented by dark net CO₂ exchange amount). However, for D. primulinum, the daily net CO₂ exchange amount decreased by 38% after the light/dark cycle was changed from 12/12 h to 4/4 h (Figure 2B), it was possibly due to the fact that its stomatal conductance increased much less than that of D. officinale limited the diffusion of CO₂ into the leaf (Evans and Loreto, 2000), eventually inhibiting the substantial increase of CO₂ absorption during the light period of 4/4 h light/dark cycle. The contrary responses of daily net CO₂ exchange amount of D. officinale and D. primulinum on 4/4 h light/dark cycle may also be related to their different mesophyll conductance (g_m). A strong correlation has been found between g_m and photosynthetic capacity in two species of Orchidaceae, and g_m was mainly determined by surface area of mesophyll cells, chloroplasts exposed to intercellular airspace per unit of leaf area and cell wall thickness (Yang et al., 2018). Reduced intercellular air space (IAS) and reduced surface of mesophyll exposed to IAS (L_mes/area) were positively related to CAM function, and negatively related to C3 function (Nelson and Sage, 2008). Structure features of leaf may affect CO₂ concentration in chloroplast stroma (Terashima et al., 2011). It was reported that increased O₂/CO₂ eventually increased photorespiration. Photorespiration had an especially high demand for energy (Osmond and Grace, 1995; Heber et al., 2001; Heber, 2002). After the light/dark cycle changed from 12/12 h to 4/4 h, the daily net CO₂ exchange amount of D. primulinum decreased significantly, which suggested that D. primulinum might subject to severe photorespiration during the light period of 4/4 h light/dark cycle. Therefore, the changes of daily net CO₂ exchange amount of D. officinale and D. primulinum were reversed after the photosynthetic pathway was switched from CAM to C3. This may be related to stomatal conductivity, g_m, cell wall conductance, cytosol conductance, stromal conductance, etc., which require further study to clarify the relationship between leaf functional structure and photosynthesis of Dendrobium plants.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.