*Article* **Use of Synchrotron Phase-Sensitive Imaging for the Investigation of Magnetopriming and Solar UV-Exclusion Impact on Soybean (***Glycine max***) Leaves**

**Anis Fatima 1,\*,† , Sunita Kataria 2,\*,† , Ashish Kumar Agrawal 1, Balwant Singh 1, Yogesh Kashyap 1, Meeta Jain <sup>2</sup> , Marian Brestic 3,\* , Suleyman I. Allakhverdiev <sup>4</sup> and Anshu Rastogi 5,6**


**Abstract:** The combined response of exclusion of solar ultraviolet radiation (UV-A+B and UV-B) and static magnetic field (SMF) pre-treatment of 200 mT for 1h were studied on soybean (*Glycine max*) leaves using synchrotron imaging. The seeds of soybean with and without SMF pre-treatment were sown in nursery bags kept in iron meshes where UV-A+B (280–400 nm) and UV-B (280–315 nm) from solar radiation were filtered through a polyester filters. Two controls were planned, one with polythene filter controls (FC)- which allows all the UV (280–400 nm); the other control had no filter used (open control-OC). Midrib regions of the intact third trifoliate leaves were imaged using the phase-contrast imaging technique at BL-4, Indus-2 synchrotron radiation source. The solar UV exclusion results suggest that ambient UV caused a reduction in leaf growth which ultimately reduced the photosynthesis in soybean seedlings, while SMF treatment caused enhancement of leaf growth along with photosynthesis even under the presence of ambient UV-B stress. The width of midrib and second-order veins, length of the second-order veins, leaf vein density, and the density of third-order veins obtained from the quantitative image analysis showed an enhancement in the leaves of plants that emerged from SMF pre-treated seeds as compared to untreated ones grown in open control and filter control conditions (in the presence of ambient UV stress). SMF pre-treated seeds along with UV-A+B and UV-B exclusion also showed significant enhancements in leaf parameters as compared to the UV excluded untreated leaves. Our results suggested that SMF-pretreatment of seeds diminishes the ambient UV-induced adverse effects on soybean.

**Keywords:** phase-sensitive imaging; magnetopriming; UV exclusion; leaf venation; leaf hydraulics

#### **1. Introduction**

One of the non-ionizing parts of the electromagnetic spectrum of solar radiation is ultraviolet radiation. Ultraviolet (UV) radiations are further divided into three ranges: UV-A (315–400 nm), UV-B (280–315 nm), and UV-C (100–280 nm). The UV-C and major part of UV-B radiations are absorbed by the earth's ozone layer [1]. Even if around 20% of UV-B is able to pass through the ozone layer and reach the earth's surface, it may be harmful

**Citation:** Fatima, A.; Kataria, S.; Agrawal, A.K.; Singh, B.; Kashyap, Y.; Jain, M.; Brestic, M.; Allakhverdiev, S.I.; Rastogi, A. Use of Synchrotron Phase-Sensitive Imaging for the Investigation of Magnetopriming and Solar UV-Exclusion Impact on Soybean (*Glycine max*) Leaves. *Cells* **2021**, *10*, 1725. https://doi.org/ 10.3390/cells10071725

Academic Editor: Sara Rinalducci

Received: 30 April 2021 Accepted: 3 July 2021 Published: 8 July 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

to biological systems due to its high energy content. Anthropogenic activities resulted in the reduction of the ozone layer, due to which the percentage of UV-B reaching the earth increased [2,3]. This further resulted in an increasing interest of scientists to understand how plants with a sessile nature react to this increased level of UV-B radiation [2–6]. The different responses of high UV-B radiation on plant structure, morphology, physiology, and genetics have been intensively studied previously [2,4,5,7] where UV-B radiations have been observed to adversely impact the cell membrane and caused changes in plant photosynthesis and enzyme activities [2,8].

Seed priming methods are the pre-treatment of seeds prior to sowing for the purpose of improving the physiological state of the seeds so that the seed germinates more efficiently [9,10]. There are several seed priming methods practiced in agronomy for increasing the seed germination, crop growth, and yield [10–13]. Static magnetic field (SMF) is a seed pre-treatment method based on the interactions of electromagnetic fields with seeds which act as bio-stimulators for the growth of seeds and plants [8,14–16]. The effect of SMF on plants has been extensively studied over the past few years as magnetic field pre-treatment may provide a non-chemical solution to the plants [16–18]. Some of the previous studies reported stimulatory effects of SMF treatment on crops including rice, maize, soybean, and sunflower [15,18–21], whereas the others reported slow development [22]. It is thus predicted that various plant species respond in different ways to varied frequencies and intensities of the magnetic field [23–25]. Plants showed reactions to magnetic fields based on the intensity, flux density, and exposure time [16,25,26]. The enhanced germination percentage improved plant growth, photosynthesis and yield were observed due to SMF pre-treatment of seeds as compared to the untreated seeds under non-stress as well as under abiotic stresses such as salt, water, UV-B, and arsenic toxicity [8,15–18,27–29]. The effect of magnetopriming on plants can be best understood in the framework of two mechanisms, namely the ion cyclotron-resonance (ICR) and the radical-pair models (RPM) [16,30]. The RPM is currently the only possible mechanism demonstrating the function of cryptochromes as a candidate for magneto-reception [16]. The experimental and theoretical studies provide evidence that the application of magnetic fields increases the average radical concentration, increases radical lifetime, and escalates the probability of radical reactions with cellular components [30]. The radical pair intermediates, triplet yields, and emission intensity that occur in Photosystem I and II of green plants can be modulated by an external magnetic field. The increased water uptake compared to untreated seeds is explained by the assumption that the magnetic field interacts with ionic currents in the cell membrane of the plant embryo [31]. In addition to these mechanisms, the interaction between environmental impacts such as ionizing radiation (ultraviolet-UV) and the magnetic field influence as a repair mechanism has also been reported previously in chick embryos [32].

Magnetic field treatment with low flux densities and the exclusion of solar UV radiation are the two parts of radiation biology that have positive stimulating effects on leaf growth, venation, and photosynthesis [8,18,29]. The network of leaf venation is composed of minor veins and a midrib (major conducting vein), which provides mechanical stability to the leaf structure. The venation network has the important function of transportation of water, nutrients, and carbon to different plant tissues [33–35]. The hydraulic system associated with plant leaf veins plays a key part in photosynthetic gas exchange and growth determination [36]. The width of midrib and minor veins, leaf vein density (LVD) (known as the vein length per leaf area), and the vein number density (which is the number of veins per leaf area) are all directly related to leaf hydraulic conductivity and photosynthesis [37–39]. Both magnetic treatment and exclusion of solar UV radiation change plant photosynthetic function which is related to the midrib of the leaf venation. The positive effects of solar UV exclusion and SMF on the leaf venation (midrib width) have been individually studied using synchrotron-based X-ray phase-contrast imaging [25,40]. However, there have been no reports on X-ray imaging of leaf venation to the combination of SMF pre-treatment of seeds and the exclusion of solar UV radiation.

The relationship of leaf venation and hence leaf hydraulics with photosynthesis is not yet explored completely. Advancements in non-destructive X-ray imaging techniques have overcome the limitations of manual sectioning and staining of leaves for imaging. So far, X-ray imaging studies for various parts of the plant have been reported [34]. Xray radiography and micro-computed tomography (μCT) studies of intact plant parts with synchrotron radiation have contributed to the understanding of plant anatomical structures [37,41–46].

The phase-contrast imaging (PCI) technique relies on phase variations which occur when the X-ray wave front transmits through a sample [47–50]. The technique overcomes the limitations of conventional absorption-based techniques. It is well suited for imaging weakly absorbing samples like leaves in non-destructive ways [37]. In the present study, we have used the soybean (*Glycine max*) variety JS-335 an economically important crop to investigate the effects of exclusion of solar UV radiation in plants grown from the seeds pre-treated with SMF for 1 h. The aim of the present study was to determine the changes in the width of the midrib and minor veins, length of minor veins (2◦ and 3◦) of leaves, and leaf vein density through high-resolution X-ray imaging and relate it to leaf growth, photosynthetic rate, and stomatal conductance.

#### **2. Materials and Method**

The soybean (*Glycine max* (L.) var. JS-335) seeds were procured from the Indian Institute of Soybean Research in Indore, India. The experiment was conducted under natural sunlight at the open terrace of the School of Biochemistry, in Devi Ahilya Vishwavidyalaya, Indore (22◦44 N, 75◦50 E), India. The experimental period was between October 2018 to December 2018. After moistening the SMF-pretreated (MT) and untreated (UT) soybean seeds were further mixed with recommended fungicides *viz*Bevistin and Diathane M at 2 gm kg−<sup>1</sup> seeds and *Rhizobium* culture (provided by National Fertilizer limited, New-Delhi, India) at 3 g kg−<sup>1</sup> seeds before sowing. The uniform shape and size of seeds were sown in plastic nursery bags of 34 × 34 cm. The nursery bags were filled with a mixture of soil, sand, and organic manure in a 2:2:1 ratio, and ten seeds of soybean were sown; three bags were prepared for each treatment. In each bag, six plants of uniform size were maintained after germination.

#### *2.1. Magnetic Field Generation*

An electromagnetic field generator ("AETec" Academy of Embedded Technology, Delhi, India) was used for the generation of magnetic field for seed pre-treatment, as previously described by Kataria et al. [51].

#### *2.2. Magnetic Treatment*

For the experiments, the seeds were exposed to SMF treatment of 200 mT for 1 h (MT) on the basis of our previous study on soybeans [25]. Through the Gauss meter, we can measure the magnetic field generated between the poles. The current in coils was regulated to obtain the exact magnetic field for the SMF pretreatment. At 50 mT, the variation in the applied field was observed to be 0.6% in the horizontal and 1.6% in the vertical direction, whereas, at 300 mT the variation decreased to 0.4% and 1.2% in both directions, respectively. A temperature of 25 ± 5 ◦C was maintained during seed exposure to SMF. The seeds from the same lot were kept under conditions without any influence of the magnetic field served as untreated (UT) seeds.

#### *2.3. UV-A+B and UV-B Exclusion*

The UV-A+B and UV-B radiations were cut-off from solar radiation by using bandpass polyester filters (Garware polyester Ltd., Mumbai) with cut-offs of <315 nm and <400 nm radiation. Two controls were designed for this study; one with a polythene filter transparent to all ambient light (filter control, FC) and the other grown on the terrace without any filter (open control, OC). Figure 1 shows the transparency of the filters used

in the experiments. The transmission spectra of the filters were measured according to the method of Kataria et al. [8]. The filters were continuously used from seed germination to maturity, with a regular exchange of filters every two weeks due to the solar radiation effect on the filters. For proper ventilation, the lower sides of the cage (0.35 m above the surface) holding the filter were not covered. The experiments were placed in the corner where sunlight was available throughout the day without any shading. The temperature inside and outside the cage was monitored through thermometers. During the growing period, average temperature was raised from 25 ◦C to 32 ◦C. No significant difference in the inside and outside temperatures was observed due to proper ventilation.

**Figure 1.** Transmission spectra of UV cut-off filters and polythene filter used for raising soybean plants under iron mesh cages [8] (Kataria et al. 2017a).

#### *2.4. Radiation Measurement*

At midday (around noon), a radiometer (Solar light Co. Inc. (PMA 2100), Glenside, PA, USA) was used to measure the intensity of solar spectra. The average photosynthetic active radiation (PAR) value at midday was observed to be 1450 μmol m−<sup>2</sup> s−<sup>1</sup> for the non-filter control, which decreased by 12.5% (1270 μmol m−<sup>2</sup> s−1) under the UV-B filter and 11.8% (1280 μmol m-2 s−1) under the UV-A+B filter, whereas a decrease of 4.2% (1390 μmol m−<sup>2</sup> s<sup>−</sup>1) was observed for the filter control.

#### *2.5. Growth Data Collection and Analysis*

A random selection of plants was done after 45 days of seed germination (DAE). At least three plants in triplicates from each treatment were harvested and transferred to the laboratory for growth data analysis. The soil particles from roots were washed and different parts of the plant were measured through a portable laser leaf area meter CID-202 scanning leaf area meter (CID Inc., Camas, WA, USA).

#### *2.6. Photosynthesis and Stomatal Conductance*

The LI-COR photosynthetic system (Li-6200, LI-COR Inc., Lincoln, NE, Serial No. PPS 1332 USA) was used to measure net photosynthesis (*Pn*, μmol CO2 m−<sup>2</sup> s<sup>−</sup>1) and stomatal conductance (*gs*, mol H2O m−<sup>2</sup> s−1) for intact soybean plants from each experimental condition after 45 DAE. Photosynthetic measurements were performed on fully expanded third trifoliate leaves of soybean plants under ambient temperature and CO2 concentration, on clear days. The photosynthetic photon flux density (PPFD) was observed to be in between 1300–1600 μmol m−<sup>2</sup> s−<sup>1</sup> with airflow of 500 μmol s−<sup>1</sup> and CO2 concentration of 350–380 ppm.

#### *2.7. Phase Contrast Imaging Technique*

The Imaging Beamline (BL-4), Indus-2 synchrotron radiation source [40,52] was used to generate the phase-contrast images. The experimental setup was previously described in [25].

The third trifoliate leaves of soybeans from all the groups were pressed flat and dried for two days at room temperature. The whole leaflets of the third trifoliate leaves were mounted in a rectangular metallic frame and phase-contrast images were acquired for middle regions in each leaf. The high-resolution X-ray microscope with 1.8 μm resolution (20 μm thick YAG-Ce scintillator, 4× objective, and PCO-2000 CCD camera) was used for image acquisition at 12keV energy, with a sample to detector distance (SDD) 50mm and an exposure time of 5 min.

#### *2.8. Leaf Midrib Width Quantification*

From the synchrotron images of the middle leaflet of third trifoliate leaves of soybeans, the midrib width was quantified at six places in the direction perpendicular to the length at fixed intervals with ImageJ [53]. The average width of the midrib vein and the adjoining minor vein (2◦) was obtained for all the leaflets in the third trifoliate and an average value for the leaf was then calculated [25,40].

#### *2.9. Leaf Minor Vein Length and Leaf Vein Density Quantification*

The length of the minor vein (2◦) was obtained using a freehand line in Image J. To obtain the total length and number of the (3◦) minor vein in the entire phase contrast image of 2048 × 2048 pixel size, the objectJ plugin was used (plant-imageanalysis.org/software/object (accessed on 12 April 2019). In the phase-contrast images, the vascular region above the midrib was selected with the freehand selection tool in Image J, and the area was measured. Similarly, the area of the vascular region below the midrib was acquired. To find the vascular area in the whole image, the area of the two regions measured were combined. Leaf vein density (LVD) was found by dividing the total length of all 3◦ veins (marked with red) in the image with the total area of the image. The total number of 3◦ veins in the images was divided with the total area to calculate the vein number density using ObjectJ.

#### *2.10. Statistical Analysis*

All data are presented in triplicate (*n* = 3); from each replica five plants were randomly taken for each treatment. The statistical analysis was performed on Microsoft Excel and Prism 4 (GrafPad Software, La Jolla, CA, USA) software where mean and standard errors were calculated, and the analysis of variance (ANOVA) followed by post hoc Newman– Keuls Multiple Comparison Test was performed. ### *p* < 0.001; ## *p* < 0.01; # *p* < 0.05 denotes statistically significant differences between seedlings that emerged from untreated (UT) seeds of OC with seedlings that emerged from untreated (UT) seeds of different treatment conditions-FC, UV-B and UV-A+B cutoff filters. \*\*\* *p* < 0.001; \*\* *p* < 0.01; \* *p* < 0.05 denotes statistically significant differences between seedlings that emerged from SMF-pretreated (MT) and untreated (UT) seeds under each treatment.

#### **3. Results and Discussion**

In the present study, the individual effects of the exclusion of solar UV-A+B, UV-B radiation, and SMF pre-treatment as well as their combination were investigated on the growth, photosynthesis, and development of soybean leaves. Individual and joint exclusion of solar UV-A+B, UV-B radiation, and SMF pre-treatment significantly enhanced all leaf growth parameters studied in the present study, but the extent of enhancement was greater when the plants pre-treated with SMF were grown under ambient UV stress (OC and FC conditions).

A prominent increase was observed in the area and length of the middle leaflet of the third trifoliate leaves of soybean plants raised after SMF (200 mT for 1 h) priming with or without ambient UV radiations (Figure 2a,b). Similarly, solar UV exclusion also enhanced the area and length of middle leaflets of third trifoliate leaves of plants that emerged from untreated (UT) seeds (Figure 2a,b). The area of the middle leaflet increased by 44% and 50% through SMF-treatment respectively under OC and FC conditions as compared to their UT ones (Figure 2a).

**Figure 2.** *Cont.*

**Figure 2.** Leaf area (**a**), leaf length (**b**), stomatal conductance (**c**) and rate of photosynthesis (**d**) in middle leaflets of third trifoliate leaves of soybean after SMF pretreatment and solar UV exclusion in soybean. ## *p* < 0.01; # *p* < 0.05 denotes statistically significant differences between seedlings emerged from untreated (UT) seeds of OC with the seedlings emerged from untreated (UT) seeds of different treatments conditions-FC, UV-B and UV-A+B cutoff filters, \*\*\* *p* < 0.001; \*\* *p* < 0.01; \* *p* < 0.05 denotes statistically significant differences between seedlings emerged from SMF-pretreated (MT) and untreated (UT) seeds under each treatment.

The enhancement in the length of middle leaflets of third trifoliate leaves of soybean after SMF treatment was 34% in OC and 30% in FC conditions as compared to their UT ones (Figure 2b). A significant increase in leaf length by 41% under solar UV-B exclusion and 37% under UV-A+B exclusion in UT was observed as compared to the plants from UT seeds under OC conditions (Figure 2b).

A significant enhancement in stomatal conductance and photosynthetic rate was observed for the plants pretreated with SMF of 200 mT for 1 h (Figure 2c,d). SMF caused a 28% and 26% increase in stomatal conductance and a 70% and 69% increase in the net photosynthetic rate as compared with untreated controls respectively in OC and FC (presence of ambient UV stress) conditions (Figure 2c,d). Enhancement of leaf area along with an increase in the rate of photosynthesis and stomatal conductance after the SMF pretreatment (200 mT for 1 h) has been previously reported in soybean and maize [8,15,18,21].

A qualitative and quantitative comparison of phase-contrast images of untreated and SMF pre-treated leaves in OC, FC, UV-A+B, and UV-B showed enhancement in the midrib width, minor vein width, and leaf vascular region near the midrib (Figures 3–9). In the OC group which received all the ambient solar radiation (280–400 nm), the quantification of leaf veins in the phase-contrast images showed an enhancement of 44% in the width of the midrib in the plants grown from the SMF pre-treated seeds as compared to untreated seeds (Figures 3 and 4a). The visibility of vascular structures comprising of higher-order veins (3◦) has also been improved in SMF pre-treated leaves (Figure 3b), which is due to a thinning effect [54].

**Figure 3.** Phase contrast images of soybean leaves under open control (OC) receiving all ambient solar radiation: (**a**) emerging from untreated seeds, (**b**) emerging from seeds pre-treated with static magnetic field (SMF) of 200 mT strength for 1 h. The vascular region below the midrib region is highlighted in red and zoomed images are shown below the respective images.

**Figure 4.** *Cont.*

**Figure 4.** Width of midrib (**a**), width of minor veins (**b**) and length of minor veins (**c**) from X-ray images after SMF pretreatment and solar UV exclusion in middle leaflets of the third trifoliate leaves of soybean. ## *p* < 0.01; # *p* < 0.05 denotes statistically significant differences between seedlings emerged from untreated (UT) seeds of OC with the seedlings emerged from untreated (UT) seeds of different treatments conditions-FC, UV-B and UV-A+B cutoff filters. \*\*\* *p* < 0.001; \*\* *p* < 0.01 denotes statistically significant differences between seedlings that have emerged from SMF-pretreated (MT) and untreated (UT) seeds under each treatment.

**Figure 5.** Phase contrast images of filter control (FC) soybean leaves grown with polythene filters which transmitted solar radiation: (**a**) emerging from untreated seeds, (**b**) emerging from seeds pretreated with static magnetic field (SMF) of 200 mT strength for 1 h. The midrib regions enclosed with the red square in the images are zoomed to show midrib enhancement. The midrib quantification was done as shown with the vertical line in the zoomed filter control of the magnetically treated leaf (FCMT) image.

**Figure 6.** Leaf vein density of tertiary veins (**a**) and number density of veins (**b**) from X-ray images after SMF pretreatment and solar UV exclusion in soybeans. ## *p* < 0.01; # *p* < 0.05 denotes statistically significant differences between seedlings that emerged from untreated (UT) seeds of OC with the seedlings that emerged from untreated (UT) seeds of different treatment conditions; FC, UV-B, and UV-A+B cutoff filters. \*\* *p*< 0.01 denotes statistically significant differences between seedlings that emerged from SMF-pretreated (MT) and untreated (UT) seeds under each treatment.

**Figure 7.** Phase-contrast images of soybean leaves from open control (OC) showing the 3◦ veins marked with red to obtain the total length of 3◦ veins and thus the leaf vein density (LVD) with the ObjectJ plugin. The number of minor veins in the images has been used to find the number density of the (3◦) minor vein: (**a**) emerging from untreated seeds, (**b**) emerging from seeds pre-treated with a static magnetic field (SMF) of 200 mT strength for 1 h showing greater numbers of minor veins. Similar images for the quantification of 3◦ veins in other leaf groups have been obtained with ObjectJ.

**Figure 8.** Phase contrast images of ultraviolet radiation excluded (UV-A+B) soybean leaves: (**a**) emerging from untreated seeds, (**b**) emerging from seeds pre-treated with static magnetic field (SMF) of 200 mT strength for 1 h.

**Figure 9.** Phase contrast images of ultraviolet-B radiation excluded (UV-B) soybean leaves: (**a**) emerging from untreated seeds, (**b**) emerging from seeds pre-treated with static magnetic field (SMF) of 200 mT strength for 1 h.

The second-order (2◦) minor veins also showed an increase of 27% in width and 8% in length by SMF treatment in the OC group (Figure 4b,c). Similar midrib enhancement in the SMF pre-treated group has been observed in the filter control leaves grown with polythene filters which received all the ambient solar radiation and also with UV cut-off filters (Figure 4a, Figure 5a,b, Figures 8 and 9a,b). A 28% increase by UV-A+B and 31% by UV-B filters in the average width of major veins was observed after SMF treatment as compared to their UT ones (Figure 4a).

The zoomed images of the midrib region enclosed with rectangles in red (Figure 5a,b) show enhancement of the midrib structure in the SMF pre-treated leaves. Apart from the first- and second-order leaf veins, quantification of the tertiary veins (3◦) has also been done with the ObjectJ plugin to obtain leaf vein density (LVD) (μm mm<sup>−</sup>2) and the number density of veins (mm<sup>−</sup>2) in leaves of all groups (Figure 6a,b). The tertiary veins, which are visible in the untreated and SMF pre-treated open control leaf images, are shown in red color (Figure 7a,b).

Comparison showed a higher LVD and a higher number of 3◦ veins in the SMF pretreated group compared to the untreated group (Figure 6a,b and Figure 7a,b). In the OC and FC groups receiving all the solar radiation, SMF pre-treatment led to better growth of the plants, as observed from the synchrotron imaging results and also supported by the area and length of leaves and along with rate of photosynthesis in the plants. Thus, it indicated that SMF pre-treatment alleviated the UV stress in plants grown under OC and FC conditions receiving ambient solar radiation.

In the UV-A+B and UV-B excluded group, the plants from untreated seeds (Figures 8a and 9a) showed enhancement as compared to plants receiving ambient solar radiation (OC and FC) in terms of the width of the midrib and 2◦ vein, length of the 2◦ vein, LVD and number density of 3◦veins (Figures 4 and 6). The phase-contrast images for the combination of SMF pre-treatment and exclusion of solar UV-A+B and UV-B radiation (Figures 8b and 9b) have also shown significant enhancement in the width of the midrib by 28% and 31% respectively, as compared to leaves that emerged from untreated seeds under UV exclusion filters (Figures 4a, 8a and 9a). The enhancement in the width of the midrib observed in UV-excluded along with SMF pre-treated leaves is lesser than the

enhancements of 44% and 38% which were obtained in leaves of SMF pre-treated plants receiving all solar radiation respectively in OC and FC conditions (Figures 3, 4a and 5).

An increase in the leaf vein density and number density of minor (3◦) veins was seen in the SMF pre-treated control leaves receiving all UV and also in the UV-A+B, UV-B excluded leaves (Figure 6a,b). Leaf vein density, which is the total length of minor veins per unit area, accounts for >80% of the total vein length [34]. The increase in the LVD of a minor (3◦) veins indicates increased hydraulic activity in the SMF pre-treated leaves as reported [34].

High LVD can enable higher stomatal conductance and also indicates higher rates of gas exchange per unit leaf area and photosynthesis [25,39]. The vein density, leaf mid rib and minor vein thickness, were strongly correlated with the hydraulic conductivity and higher photosynthetic rate of the leaves. Thus, the observation showed that SMF pretreatment and solar UV exclusion individually and together enhanced leaf hydraulic efficiency, which can be observed through the changes in leaf venation architecture. The leaves were observed to be expanded with thicker veins from SMF-treated and UV excluded plants which give good mechanical support, whereas transpiration cooling and improved photosynthesis were observed because of higher water transportation due to higher vein length per unit area of the leaves [39,55]. The mechanism by which plants perceive MFs and regulate the signal transduction pathway is not fully understood. It has been suggested that MF perception/signaling in plants is regulated by blue light photoreceptors-cryptochromes. It has also been found that reactive oxygen species (ROS) and nitric oxide (NO) are the signaling molecules for magnetopriming-induced seed germination, plant growth, and photosynthesis [29,56]. The participation of NO through nitric oxide synthase enzyme was confirmed in SMF-induced tolerance towards UV-B stress in soybean [56]. However, this aspect of magneto biology still deserves in-depth investigation during leaf growth and photosynthesis.

#### **4. Conclusions**

The exclusion of UV-A+B and UV-B radiation is advantageous, as it was suggested that plant growth, leaf area, and photosynthesis were inhibited by ambient UV-B stress. The exposure of seeds to SMF treatment prior to sowing is an eco-friendly method with the potential to alleviate the adverse effects of UV-B stress in the plants. Looking into the correlation between leaf venation and leaf hydraulic conductivity, we used X-ray imaging to study leaf venation (major and minor vein up to 3◦) under UV-exclusion, SMF pre-treatment, and the combined effect of both. UV exclusion and SMF pre-treatment individually and jointly showed positive effects on plant growth, development, photosynthesis, and leaf venation parameters obtained from the X-ray images. To our knowledge, this is the first study on X-ray imaging of leaf venation under the combined effects of solar UV exclusion and SMF pre-treatment.

**Author Contributions:** Conceptualization, A.F., S.K., M.B. and S.I.A.; methodology, A.F. and S.K.; software, A.F.; validation, A.F., S.K., A.K.A., Y.K. and A.R.; formal analysis, S.K. and A.F.; investigation, S.K. and A.F.; resources, A.F., Y.K., M.J., A.K.A. and B.S.; data curation, A.F., A.R. and S.K.; writing—original draft preparation, A.F. and S.K.; writing—review and editing, A.F., S.K. and A.R.; visualization, A.F., S.K. and A.R.; supervision, Y.K., M.J., M.B. and S.I.A.; project administration, A.F., S.K., Y.K., M.J.; funding acquisition; A.F., S.K., A.R., M.B. and S.I.A. All authors have read and agreed to the published version of the manuscript.

**Funding:** Financial support by DST SERB National Post-Doctoral Fellowship-NPDF (PDF/2017/000643) to AF, SIA supported by the grant RFBR-NSFC (21-54-53015) and Department of Science Technology Women Scientists-A Scheme (SR/WOS-A/LS-17/2017) to SK are thankfully acknowledged. AR has performed this work while being on NAWA Bekker Programme under project No. PPN/BEK/2019/1/ 00090/U/00001.

**Institutional Review Board Statement:** Not applicable.

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Users can obtain the datasets by being in touch with Anis Fatima (anees349@gmail.com), or Sunita Kataria (sunita\_kataria@yahoo.com).

**Conflicts of Interest:** On behalf of all authors, the corresponding authors state that there is no conflict of interest.

#### **References**


### *Article* **Features of the Duckweed** *Lemna* **That Support Rapid Growth under Extremes of Light Intensity**

**Jared J. Stewart 1,\* , William W. Adams III <sup>1</sup> , Marina López-Pozo 1, Naiara Doherty Garcia 1, Maureen McNamara <sup>1</sup> , Christine M. Escobar 2,3 and Barbara Demmig-Adams 1,\***


**Abstract:** This study addresses the unique functional features of duckweed via comparison of *Lemna gibba* grown under controlled conditions of 50 versus 1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> and of a *L. minor* population in a local pond with a nearby population of the biennial weed *Malva neglecta*. Principal component analysis of foliar pigment composition revealed that *Malva* was similar to fast-growing annuals, while *Lemna* was similar to slow-growing evergreens. Overall, *Lemna* exhibited traits reminiscent of those of its close relatives in the family Araceae, with a remarkable ability to acclimate to both deep shade and full sunlight. Specific features contributing to duckweed's shade tolerance included a foliar pigment composition indicative of large peripheral light-harvesting complexes. Conversely, features contributing to duckweed's tolerance of high light included the ability to convert a large fraction of the xanthophyll cycle pool to zeaxanthin and dissipate a large fraction of absorbed light non-photochemically. Overall, duckweed exhibited a combination of traits of fast-growing annuals and slow-growing evergreens with foliar pigment features that represented an exaggerated version of that of terrestrial perennials combined with an unusually high growth rate. Duckweed's ability to thrive under a wide range of light intensities can support success in a dynamic light environment with periodic cycles of rapid expansion.

**Keywords:** antioxidants; carotenoids; chlorophyll fluorescence; photochemical efficiency; protein; tocopherol; xanthophyll cycle; zeaxanthin

#### **1. Introduction**

Small, floating plant species in the duckweed family (Lemnaceae) possess attractive nutritional features as they accumulate large quantities of high-quality protein (with all essential amino acids for humans) throughout the plant [1]. Furthermore, our group has highlighted the exceptional ability of *Lemna gibba* to accumulate high levels of the carotenoid zeaxanthin under conditions when the plant is growing rapidly [2,3]. Zeaxanthin (and its close isomer lutein) is an essential human micronutrient required to support brain function and fight systemic inflammation [4]. Duckweed also has potential uses in sustainable agricultural systems as food for humans, feed for animals (via conversion of wastewater to feed [5,6]), for other valuable products [7], or to improve nitrogen-use efficiency and yield of crops like rice [8]. Here, we present further insight into how *L. gibba* is able to combine fast growth across a range of environments with high nutritional content including pronounced zeaxanthin accumulation as well as other essential nutrients for humans or livestock.

We previously reported a notable ability of *L. gibba* to maintain uniformly high growth rates, paired with profound modulation of photoprotection, over a range of growth photon flux densities (PFDs) from 100 to 700 μmol m−<sup>2</sup> s−<sup>1</sup> of continuous light [2]. Plants

**Citation:** Stewart, J.J.; Adams, W.W., III; López-Pozo, M.; Doherty Garcia, N.; McNamara, M.; Escobar, C.M.; Demmig-Adams, B. Features of the Duckweed *Lemna* That Support Rapid Growth under Extremes of Light Intensity. *Cells* **2021**, *10*, 1481. https://doi.org/10.3390/ cells10061481

Academic Editors: Suleyman Allakhverdiev and Marian Brestic

Received: 1 May 2021 Accepted: 9 June 2021 Published: 12 June 2021

**Publisher's Note:** MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.

**Copyright:** © 2021 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (https:// creativecommons.org/licenses/by/ 4.0/).

grown under higher PFDs exhibited higher levels of the interconvertible xanthophyll cycle carotenoids (violaxanthin, antheraxanthin, and zeaxanthin) and pronounced conversion to zeaxanthin that dissipates potentially harmful excess absorbed light [3,9,10]. In the present study, we further broadened the range of growth PFDs to test whether duckweed's phenotypic plasticity with respect to photoprotective capacity and maintenance of a high growth rate may extend to even more extreme growth PFDs. We compared features of *L. gibba* grown under very low (50 μmol photons m−<sup>2</sup> s<sup>−</sup>1) or very high (1000 μmol photons m−<sup>2</sup> s−1) intensity of continuous light under otherwise common, controlled conditions. Continuous exposure (24 h per day) to the high growth PFD represented a greater total daily photon flux than that on the longest, brightest day on Earth. Beyond extending the range of PFDs versus the previous study [2], additional parameters were characterized in the present study including the light-use efficiency of biomass production as well as the production of protein as a key macronutrient and α-tocopherol (vitamin E) as an additional micronutrient. Moreover, CO2-saturated photosynthetic capacity was characterized under both saturating light and the respective contrasting growth PFDs, and photosynthesis as well as protein and all micronutrients were expressed on multiple reference bases (per frond area, biomass, and chlorophyll [Chl] content) for a fuller evaluation of both plant function and nutritional quality for the consumer.

Furthermore, the present study tested the hypothesis that the combination of exceptionally rapid growth with a remarkable ability to grow under a wide range of light intensities in duckweed may be associated with pigment patterns not seen in other fastgrowing species. In particular, prior studies of leaf pigment composition in slow-growing evergreens or perennials versus fast-growing annual species often reported an inverse relationship between growth rate and accumulation of photoprotective pigments (for a review, see [3]). The present study compared a population of *Lemna minor* in an open outdoor pond with a nearby population of the fast-growing terrestrial biennial weed *Malva neglecta* that was previously shown to exhibit a pigment composition and photoprotective capacity similar to that of fast-growing annual crop species [11]. Foliar pigment composition of *M. neglecta* and *L. minor* growing in full sun outdoors as well as that of *L. gibba* grown in low versus high PFD under controlled conditions were compared via principal component analysis to foliar pigment data for other species groups (including annuals as well as evergreens and other perennials).

#### **2. Materials and Methods**

#### *2.1. Plant Material and Growth Conditions*

#### 2.1.1. Controlled Conditions

Cultures of *Lemna gibba* L. 7741 (G3) obtained from Rutgers Duckweed Stock Cooperative (http://ruduckweed.org; accessed on 10 June 2021) were grown under controlled conditions in Conviron PGR15 and E15 growth chambers (Controlled Environments Ltd., Winnipeg, MB, Canada). Plants were grown in 150 × 75 mm PYREX Crystallizing Dishes (Corning Inc., Corning, NY, USA) that contained 1000 mL of freshly prepared Schenk and Hildebrandt medium (bioWORLD, Dublin, OH, USA [12]) at a concentration of 1.6 g L−<sup>1</sup> (pH adjusted to 5.5 via 1% [*w*/*v*] KOH). Plants were cultivated continuously under 50 μmol photons m−<sup>2</sup> s−<sup>1</sup> (provided by F72T12/CW/HO fluorescent bulbs [Philips, Somerset, NJ, USA] and 100 W, 130 V incandescent bulbs [EiKO, Shawnee, KS, USA]) at 25 ◦C, and a subset of plants from each dish were transferred to clean dishes containing freshly prepared media at least once per week.

Plants from the cultures under 50 μmol photons m−<sup>2</sup> s−<sup>1</sup> were acclimated to 1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> (provided by M47/E 1000 W metal halide bulbs; Philips, Somerset, NJ, USA) following the procedure described by Stewart et al. [2]. A subset of plants from the stock cultures (approximately 20 fronds per dish) were transferred to 200 μmol photons m−<sup>2</sup> s−<sup>1</sup> (supplied by C503C-WAN white LEDs; CREE Inc., Durham, NC, USA) for three days and then transferred to 1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> for three days. After each of these three-day acclimation phases, a subset of plants (approximately 20 fronds per dish)

that had developed under the prevailing light intensities were transferred to clean dishes with freshly prepared and filtered media. This process ensured that the characterized plant material developed under 1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> and had not developed under a lower PFD and then been transferred to 1000 μmol photons m−<sup>2</sup> s–1. Following this six-day acclimation process, plants were characterized over the course of four days while growing under 1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> and a media temperature of 25 ◦C.

#### 2.1.2. Field Conditions

Populations of *Lemna minor* L. and *Malva neglecta* Wallr. plants (identification based on morphology and geographic distribution) growing naturally in Superior, CO, USA (39◦56 28 N, 105◦09 02 W) were characterized. *Malva neglecta* is a relatively fast-growing herbaceous biennial species that remains photosynthetically active throughout the year in this area (see [13]). The *L. minor* plants were growing in a slow-moving section of a small stream, and the *M. neglecta* plants were growing on a south-facing slope immediately north of the stream. Both locations received direct, midday sunlight. Samples for pigment analysis were collected during exposure to full sun (maximal PFD of 1600 μmol m−<sup>2</sup> s−1) prior to solar noon on 17 May 2019. Samples were imaged (for quantification of frond/leaf area via ImageJ [14]) and then submerged and stored in liquid nitrogen at the field site. The four characterized samples of *L. minor* each consisted of multiple fronds from multiple plants (i.e., multiple biological replicates per sample), whereas the four characterized samples from *M. neglecta* each consisted of one leaf segment from four separate plants (i.e., four biological replicates).

#### *2.2. Growth Metrics*

Under controlled conditions, the dishes containing *L. gibba* plants were imaged from directly overhead once per day during the four-day period of characterization, and the frond area was quantified from these images using MATLAB Image Processing Toolbox (MathWorks, Natick, MA, USA) as previously described [2]. Dry mass of whole plants (i.e., fronds with intact roots) and only fronds (i.e., fronds with excised roots) was measured from samples that had been dried at 70 ◦C for seven days. Prior to drying, each sample was imaged from directly overhead, and the frond area was determined from these images using ImageJ [14].

Relative growth rate was calculated by dividing the difference in natural logarithm– transformed frond areas at the end and the beginning of the four-day experimental period by the time elapsed between the two measurements. Doubling time was calculated as the natural logarithm of 2 divided by relative growth rate. Light-use efficiency of frond area production was calculated as the accumulated frond area (i.e., the difference between the final and initial frond areas) divided by the number of incident photons on the frond surface (calculated as described in detail by Stewart et al. [2]). Light-use efficiency of biomass production was estimated as the accumulated biomass (i.e., product of accumulated frond area [m2] and whole-plant dry mass per unit frond area [g m−2]) divided by the number of available photons during this time period.

#### *2.3. Photosynthesis and Respiration*

Rates of photosynthetic oxygen evolution were determined as described by Stewart et al. [2] with saturating CO2 (5% CO2, 21% O2, balance N2) using leaf disc oxygen electrodes (Hansatech Instruments Ltd., Norfolk, United Kingdom; see [15]) and a circulating water bath set to 25 ◦C. Fronds from *L. gibba* plants grown under 50 μmol photons m−<sup>2</sup> s−<sup>1</sup> and 1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> were assayed under their respective growth PFDs as well as a saturating PFD of 1500 μmol photons m−<sup>2</sup> s−1. Respiration rates were determined following measurements of photosynthesis as the rate of oxygen consumption in darkness.

Photochemical and photoprotective processes were assessed via measurements of chlorophyll fluorescence with a PAM-101 chlorophyll fluorometer (Walz, Effeltrich, Germany) following the procedures described in detail by Stewart et al. [2] and using calculations described by Demmig-Adams et al. [16]. To ascertain the maximal level of fluorescence (Fm or Fm ) in high-light flashes, two or three flashes were given in rapid succession [17], which revealed that the maximal attainable level was reached during the first flash under all conditions used here.

#### *2.4. Protein and Starch*

Total protein content was quantified spectrophotometrically with the Total Protein Kit, Micro Lowry, Peterson's Modification (Sigma-Aldrich, Saint Louis, MO, USA), which follows a modified version [18] of the procedure described by Lowry et al. [19]. Whole plants with approximately three fronds per dish, which had been imaged and immediately frozen in liquid nitrogen, were homogenized via mortar and pestle, combined with 1 mL of water, vortexed, and centrifuged for 10 min at 10,000 rpm. The resulting supernatant was decanted, combined with 0.1 mL of deoxycholate, vortexed, and maintained at room temperature for 10 min. Subsequently, 0.1 mL of trichloroacetic acid was added, and this solution was vortexed and centrifuged for 10 min at 10,000 rpm. The resulting pellet was re-suspended in 1 mL of Lowry reagent, transferred to cuvettes, and then mixed with an additional 1 mL of water that was used to rinse the microcentrifuge tube. After 20 min of incubation at room temperature, 0.5 mL of the Folin–Ciocalteu phenol reagent was added, and this solution was mixed via pipette and incubated at room temperature for 30 min. Absorbance at 660 nm was determined with a Beckman DU 640 Spectrophotometer (Beckman Instruments, Inc., Fullterton, CA, USA) and these values were converted to protein levels (in μg mL−1) using a standard calibration curve based on a gradient of bovine serum albumin.

The abundance of starch in *L. gibba* plants was detected qualitatively with a diluted iodine-potassium iodide solution (Lugol's solution; Sigma-Aldrich, St. Louis, MO, USA). Plants were cleared in 70% (*v*/*v*) ethanol, stained for 5 min, and then immediately mounted and imaged with a high-resolution scanner (Perfection 3200 Photo; Epson America, Inc., Long Beach, CA, USA).

#### *2.5. Chlorophyll, Carotenoid, and α-Tocopherol Levels*

Chlorophylls *a* & *b*, lutein, zeaxanthin (Z), antheraxanthin (A), violaxanthin (V), neoxanthin, β-carotene, and α-tocopherol levels were quantified via high-performance liquid chromatography as previously described in detail [2,20]. Under controlled conditions, samples of approximately 10 fronds per dish were collected under the respective growth PFDs, imaged, and then frozen and stored in liquid nitrogen.

Pigment data from a previously conducted survey of multiple species in the same area by Demmig-Adams and Adams [21], which included *M. neglecta*, were used in the present study and divided into three groups: (i) shade-grown perennials; (ii) sun-grown perennials; and (iii) sun-grown annuals (for detail, see Table S1). To ensure the datasets were comparable, pigments were characterized from samples of *Vinca minor* plants growing in complete shade (shade-grown perennials) as well as plants growing exposed to full sun (sun-grown perennials) on the University of Colorado campus (Boulder, CO, USA) collected during the afternoon of 16 April 2019.

#### *2.6. Statistical Analyses*

Comparisons of means were preceded by Levene's tests to assess the equality of variances. Comparison of two means were made with Student's (equal variances) or Welch's (unequal variances) *t*-tests, and comparisons of more than two means were made with one-way ANOVAs and post-hoc Tukey–Kramer HSD. Linear relationships between two variables were evaluated with Pearson correlations. Comparisons with multiple parameters were conducted with principal component analysis on correlations. All statistical analyses were made using JMP Pro 15.0.0 (SAS Institute Inc., Cary, NC, USA).

#### **3. Results**

#### *3.1. Growth and Photosynthesis of Lemna gibba under Extremes in Growth PFD*

Despite vastly different light availability (a 20-fold difference between 50 versus 1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> of continuous light) during growth, a similar amount of duckweed frond area accumulated over time in the two controlled conditions (Figure 1A). This represented a similar frond area doubling time of just under 1.5 days for either of the two growth PFDs (Figure 1B), which also corresponded to similar average relative growth rates of 0.48 ± 0.04 and 0.50 ± 0.03 day−<sup>1</sup> (average daily increase in natural logarithm–adjusted frond area over the four-day experimental phase) for plants grown under 50 and 1000 μmol photons m−<sup>2</sup> s−1, respectively. As a consequence of the similar frond area accumulation under the two vastly different growth PFDs, the ratio of frond area produced per incident PFD, which can be considered the light-use efficiency of frond area production, was dramatically greater (1733%) in fronds grown under 50 versus 1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> (Figure 1C), in other words, almost proportional to the 20-fold (or 2000%) difference in incident PFD.

**Figure 1.** (**A**) Accumulation of frond area (% relative to frond area at the beginning of the experiment) over four days of growth, (**B**) average doubling time in frond area over this four-day period, and (**C**) light-use efficiency of frond area production (total frond area produced relative to incident PFD during the four-day growth period) in *Lemna gibba* plants under growth PFDs of 50 (green) or 1000 (yellow) <sup>μ</sup>mol m−<sup>2</sup> <sup>s</sup>−1. Mean values <sup>±</sup> standard deviations; *<sup>n</sup>* = 7 for 50 <sup>μ</sup>mol photons m−<sup>2</sup> s<sup>−</sup>1; *n* = 3 for 1000 μmol photons m−<sup>2</sup> s<sup>−</sup>1. A significant difference between the growth PFDs is denoted by asterisks in (**C**); \*\*\* = *p* < 0.001; *n.s.* = not significantly different.

A greater amount of dry biomass (Figure 2A) and protein (Figure 2B) was accumulated on a frond area basis in plants growing under the higher PFD, but the fraction of dry biomass (% biomass in g g−1) that consisted of protein (Figure 2C) was greater under the lower PFD, which resulted in a remarkable 46% of dry biomass consisting of protein under the low growth PFD. There was also evidence for greater starch accumulation (Figure 2D,E) under the higher PFD. Another contributing factor to the difference in total dry biomass in plants grown under 1000 versus 50 μmol photons m−<sup>2</sup> s−<sup>1</sup> was a difference in root dry biomass, which accounted for 16% versus 6%, respectively. As was the case for area production, light-use efficiency of total biomass production on an incident light basis was much greater in fronds grown under 50 versus 1000 μmol photons m−<sup>2</sup> s−1, albeit at a less pronounced percentage (672%) than seen for frond area (Figure 1C) due to the fact that, unlike frond area accumulation, dry mass accumulation was almost 3× greater at the higher growth PFD (Figure 2A).

**Figure 2.** (**A**) Dry biomass per frond area, (**B**) protein per frond area, (**C**) protein per dry biomass, and (**D**,**E**) presence of starch (detected via diluted iodine-potassium iodide solution) in *Lemna gibba* plants grown under 50 (green) or 1000 (yellow) <sup>μ</sup>mol photons m−<sup>2</sup> <sup>s</sup>−1. Mean values <sup>±</sup> standard deviations; *<sup>n</sup>* = 4. Significant differences between the growth PFDs are denoted by asterisks; \*\*\* = *p* < 0.001. Dimensions of each image (**D**,**E**) are 1 cm × 1 cm.

CO2-saturated photosynthesis rate was determined for plants grown under their respective growth PFDs of 50 and 1000 μmol m−<sup>2</sup> s−<sup>1</sup> as well as under a common saturating PFD of 1500 μmol m−<sup>2</sup> s−<sup>1</sup> (Figure 3). Furthermore, the resulting photosynthesis rates were expressed on three different reference bases (i.e., per frond area (Figure 3A), per frond dry biomass (Figure 3B), and per frond Chl content (Figure 3C)), revealing different trends. Photosynthesis on a frond area basis was higher in plants grown under the high PFD (Figure 3A) when measured under the respective PFDs (1000 versus 50 μmol m−<sup>2</sup> s−1). Since fronds grown under high PFD accumulated a greater amount of biomass than fronds grown under the low PFD, the photosynthesis rate on a dry mass basis was similar in the fronds when measured at their respective growth PFDs (Figure 3B). Likewise, respiration rates were higher on an area basis (3.36 ± 0.39 versus 1.42 ± 0.82 <sup>μ</sup>mol O2 <sup>m</sup>−<sup>2</sup> <sup>s</sup><sup>−</sup>1, *<sup>p</sup>* < 0.05) and similar on a dry mass basis (0.08 ± 0.01 versus 0.07 ± 0.04 <sup>μ</sup>mol O2 <sup>g</sup>−<sup>1</sup> <sup>s</sup>−1, *<sup>p</sup>* > 0.05) in fronds grown under 1000 versus 50 μmol photons m−<sup>2</sup> s<sup>−</sup>1, respectively.

**Figure 3.** CO2-saturated rates of photosynthetic oxygen evolution of fronds on the bases of (**A**) area, (**B**) dry mass (DM), and (**C**) chlorophyll (Chl) *a* + *b* levels from *Lemna gibba* plants grown under PFDs of 50 (green) or 1000 (yellow) μmol m−<sup>2</sup> s−<sup>1</sup> and measured under the respective growth PFDs (left columns) as well as a common saturating PFD of 1500 μmol m−<sup>2</sup> s−<sup>1</sup> (right columns). Mean values ± standard deviations; *n* = 3. Significant differences between growth PFDs are denoted by asterisks; \*\* = *p* < 0.01, \*\*\* = *p* < 0.001, *n.s.* = not significantly different.

In contrast, light- and CO2-saturated photosynthetic capacity on an area basis (and even more so on a dry mass basis) measured under a saturating PFD of 1500 μmol m−<sup>2</sup> s−<sup>1</sup> was higher in the fronds grown under the low versus the high PFD (Figure 3A,B). Finally, photosynthesis rate on a Chl basis was much higher in the fronds grown under the high versus the low PFD when measured either under the respective growth PFDs or under saturating PFD (Figure 3C).

#### *3.2. Pigment Composition, Light-Use Efficiency, and Photoprotection of Lemna gibba under Extremes in Growth PFD*

The fronds grown under the low PFD were green (Figure 4A) whereas those grown under the high PFD were bright yellow (Figure 4B). This difference in visual appearance was associated with a much lower Chl content, but similar total carotenoid content, on a frond area basis in the fronds grown under 1000 versus 50 μmol photons m−<sup>2</sup> s−<sup>1</sup> (Figure 4C). This difference in Chl content is, furthermore, consistent with the much higher ratios of O2 evolution relative to Chl (Figure 3C) in the fronds grown under the high versus the low PFD (which contrasted with the similar or lower rates of O2 evolution on the frond area or dry mass bases, respectively). Despite the lower Chl content, the ratio of O2 evolution per absorbed photons would likely be considerably lower at the high versus the low PFD, but this ratio cannot be computed since the fraction of absorbed light available to photosynthesis (by way of chlorophyll as opposed to carotenoids, at least some of which may not harvest light for photosynthesis) in a yellow frond is unknown.

**Figure 4.** (**A**,**B**) Images of crystallizing dishes (1L volume) with representative cultures (starting from 20 fronds per dish) at the end of the four-day growth period, (**C**) chlorophyll *a* + *b* and carotenoid levels per frond area, and (**D**) the molar ratio of chlorophyll *a* to chlorophyll *b* in fronds of *Lemna gibba* grown under PFDs of 50 (green) or 1000 (yellow) <sup>μ</sup>mol m−<sup>2</sup> <sup>s</sup><sup>−</sup>1. Mean values <sup>±</sup> standard deviations; *n* = 3 or 4. Significant differences between growth PFDs are denoted by asterisks; \*\*\* = *p* < 0.001; *n.s.* = not significantly different.

Figure 5A shows an estimation (from chlorophyll fluorescence) of the allocation of absorbed light to PSII photochemistry (Photochemistry) and non-photochemical routes (Dissipation) as well as the fraction of excitation energy dissipated neither via photochemical or regulated non-photochemical routes (Remainder). The combination of photochemical and non-photochemical routes of excitation energy utilization or dissipation apparently was sufficient to prevent any major build-up of excitation energy.

**Figure 5.** (**A**) Estimated percentages of absorbed light (under the respective growth PFDs) allocated to PSII photochemistry (Fv /Fm × qP), dissipation (0.8 − Fv /Fm ), and the fraction of excitation energy dissipated neither via photochemical or regulated non-photochemical routes (remainder; Fv /Fm × [1 − qP]) and (**B**) levels of xanthophyll cycle pool constituents violaxanthin, antheraxanthin, and zeaxanthin relative to chlorophyll (Chl) *a* + *b* in fronds of *Lemna gibba* plants grown under PFDs of 50 (green) or 1000 (yellow) μmol m−<sup>2</sup> s−1. The conversion state of the xanthophyll cycle to zeaxanthin and antheraxanthin in (**B**) should be compared to the level of dissipation in (**A**). Mean values ± standard deviations; *n* = 3 or 4. Significant differences between growth PFDs are denoted by asterisks; \*\* = *p* < 0.01, \*\*\* = *p* < 0.001.

The much greater estimated dissipation of absorbed light via regulated non-photochemical routes (Figure 5A, Dissipation) in fronds grown under 1000 versus 50 μmol photons m−<sup>2</sup> s−<sup>1</sup> was mirrored by strong accumulation of the photoprotective xanthophylls zeaxanthin and antheraxanthin relative to Chl *a* + *b* (Figure 5B). Furthermore, Figure 6 shows that either the estimated fraction of absorbed light allocated to PSII photochemistry (Figure 6A) or the fraction of the interconvertible xanthophyll cycle pool (violaxanthin + antheraxanthin + zeaxanthin, V + A + Z) converted to the de-epoxidized forms zeaxanthin and antheraxanthin (Figure 6B) exhibited a significant positive relationship with the energy-use efficiency of total frond dry biomass production over a range of six growth PFDs from 50 to 1000 μmol m−<sup>2</sup> s−<sup>1</sup> [2]. For the energy-use efficiency of frond area production, significant correlations were likewise seen (not shown) with the estimated fraction of absorbed light allocated to PSII photochemistry (*p* = 0.008; *r*<sup>2</sup> = 0.862) and xanthophyll cycle pool conversion state (*p* = 0.039; *r*<sup>2</sup> = 0.694), respectively.

**Figure 6.** Correlations between the light-use efficiency of biomass production (total dry biomass produced per incident PFD during the four-day growth period) and (**A**) (Fm − F)/Fm = Fv /Fm × qP (as an estimate of the light-use efficiency of PSII photochemistry) or (**B**) the percent of the xanthophyll cycle pool converted to zeaxanthin (Z) and antheraxanthin (A) in *Lemna gibba* fronds grown under six PFDs ranging from 50 to 1000 <sup>μ</sup>mol m−<sup>2</sup> <sup>s</sup><sup>−</sup>1. Mean values <sup>±</sup> standard deviations, *<sup>n</sup>* = 3 or 4. Data for PFDs of 100 to 700 μmol m−<sup>2</sup> s−<sup>1</sup> were calculated from [2]. V = violaxanthin.

Figure 7 shows that the estimated fraction of absorbed light allocated to photosystem II photochemistry in fronds grown under 1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> was lowered to less than 20% (Fv /Fm < 0.2) under this growth PFD and rose rather quickly upon transfer of fronds to 10 μmol photons m−<sup>2</sup> s−<sup>1</sup> to over 60% (Fv/Fm > 0.6) within 30 min. In contrast, there was only a small (albeit significant; *p* < 0.001) difference between the estimated fraction of absorbed light allocated to photosystem II photochemistry (not shown) under the growth PFD of 50 <sup>μ</sup>mol photons m−<sup>2</sup> <sup>s</sup>−<sup>1</sup> (at 72.2 ± 1.2%) versus 30 min of recovery (78.1 ± 0.8%) in 10 <sup>μ</sup>mol photons m−<sup>2</sup> <sup>s</sup><sup>−</sup>1.

**Figure 7.** Fv /Fm (as an estimate of intrinsic PSII efficiency during exposure to growth PFD) and Fv/Fm (as an estimate of intrinsic PSII efficiency after 5 min darkness) for *L. gibba* fronds grown under 1000 μmol photons m−<sup>2</sup> s−1. Fv/Fm was determined at two time points (i.e., after 5 min of darkness upon removal from growth PFD and following a recovery period of 30 min under 10 μmol photons m−<sup>2</sup> <sup>s</sup>−<sup>1</sup> followed by 5 min darkness). Mean values <sup>±</sup> standard deviations; *<sup>n</sup>* = 3. Significant differences between time points are denoted by different lower-case letters.

A full characterization of pigment and α-tocopherol composition of *L. gibba* fronds grown under the extremes of 50 and 1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> is presented in Tables 1 and 2. Table 1 shows the levels of carotenoids and α-tocopherol on both frond area and dry biomass (DM) bases. Table 2 shows the ratios of carotenoids and α-tocopherol to Chl *a* + *b* or Chl *b* only as well as other ratios. While total carotenoid concentration under the high versus the low growth PFD was the same on a frond area basis (Figure 4), it was only about half on a DM basis (Table 1), but about 4× higher on a Chl *a* + *b* basis (Table 2). All individual carotenoids except zeaxanthin and antheraxanthin were present at lower concentrations on an area basis, and even lower on a DM basis (Table 1), while being enhanced relative to total Chl *a* + *b*, and even more so relative to Chl *b* alone, at the high versus the low growth PFD (Table 2). Zeaxanthin, antheraxanthin, and the total xanthophyll cycle pool (V + A + Z) were all greater on an area basis (Table 1) and relative to Chl *a* + *b* or Chl *b* alone (Figure 5B, Table 2). Zeaxanthin and antheraxanthin (Z + A), but not the total xanthophyll cycle pool, were also greater on a DM basis (Table 1). Moreover, higher ratios were seen at the high growth PFD for zeaxanthin or the total xanthophyll cycle pool to lutein, all xanthophylls to β-carotene, and the proportion of the total xanthophyll cycle pool that was converted to either zeaxanthin alone or the sum of zeaxanthin and antheraxanthin (Table 2). The only carotenoid present in the same ratio to Chl *a* + *b*, thus exhibiting a proportional decline as Chl *a* + *b* at the high growth PFD, was neoxanthin (Table 2). However, the neoxanthin level was not lowered as much as the Chl *b* level at the high growth PFD (Table 2). The only compound among those considered in Table 1 that was not significantly different in concentration on an area basis between the two growth PFDs was α-tocopherol, which resulted in a strong increase in the ratio of α-tocopherol to Chl *a* + *b* (Table 2). α-Tocopherol concentration was about half per DM at the high versus the low growth PFD (Table 1).


**Table 1.** Levels of carotenoids and α-tocopherol per frond area and dry mass (DM) in fronds of *Lemna gibba* grown under PFDs of 50 or 1000 μmol m−<sup>2</sup> s<sup>−</sup>1.

Mean values ± standard deviations; *n* = 3 or 4. Significant differences between growth PFDs are denoted by asterisks; \*\* = *p* < 0.01; \*\*\* = *p* < 0.001; *n.s.* = not significantly different. A = antheraxanthin, V = violaxanthin, Z = zeaxanthin.


**Table 2.** Levels of carotenoids and α-tocopherol relative to chlorophylls (Chl) or other carotenoids in fronds of *Lemna gibba* grown under PFDs of 50 or 1000 μmol m−<sup>2</sup> s<sup>−</sup>1.

Mean values ± standard deviations; *n* = 3 or 4. Significant differences between growth PFDs are denoted by asterisks; \*\* = *p* < 0.01; \*\*\* = *p* < 0.001; *n.s.* = not significantly different. A = antheraxanthin, V = violaxanthin, Z = zeaxanthin.

#### *3.3. Unique Pigment Patterns in Lemna Compared to Other Species*

Under exposure to full sunlight, fronds of *L. minor* floating on a local pond exhibited conversion of a greater percentage of the total pool of interconvertible xanthophylls (V + A + Z) to the photoprotective forms zeaxanthin or zeaxanthin + antheraxanthin compared to leaves of a nearby population of the weed *M. neglecta* (Figure 8). Whereas the conversion state to the photoprotective, de-epoxidized components was higher in *L. minor* (Figure 8), the total pool of the xanthophyll cycle relative to Chl *a* + *b* was smaller in *L. minor* compared to *M. neglecta* (Figure 9A).

Lutein was present at a greater level on a Chl basis in *L. minor* compared to *M. neglecta* (Figure 9B). In contrast, β-carotene per Chl (Figure 9C), α-tocopherol per Chl (Figure 9D), Chl *a* + *b* per area (Figure 9E), and the Chl *a*/*b* molar ratio (Figure 9F) were all lower in *L. minor* than in *M. neglecta*.

The herbaceous weed *M. neglecta* (a biennial) fell into a cluster comprised of fastgrowing annual species that accumulated rather modest amounts of zeaxanthin in full sun, while *L. minor* fell into a cluster of slow-growing perennial species (Figure 10A) using principal component analysis based on foliar pigment composition (Figure 10B). *Lemna minor*'s pigment composition was thus similar to that of slow-growing perennials, which accumulated large amounts of zeaxanthin and lutein in full sun. Remarkably, even *L. gibba* grown under 50 μmol photons m−<sup>2</sup> s−<sup>1</sup> fell into this cluster (Figure 10A), despite the fact that it did not accumulate zeaxanthin under this low growth PFD (Figure 5B). The yellow fronds of *L. gibba* grown under 1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> that maintained carotenoids while strongly lowering Chl content (Figure 4C) fell far to the right of all three clusters (Figure 10A).

**Figure 8.** Conversion state of the xanthophyll cycle pool to zeaxanthin (Z) or zeaxanthin + antheraxanthin (A) under full sun (1600 μmol photons m−<sup>2</sup> s<sup>−</sup>1) in fronds of *Lemna minor* (orange) and leaves of *Malva neglecta* (green) growing naturally in Superior, CO, USA. Mean values ± standard deviations; *n* = 4. Significant differences are denoted by asterisks; \*\* = *p* < 0.01, \*\*\* = *p* < 0.001. V = violaxanthin.

**Figure 9.** Levels of (**A**) the xanthophyll cycle pool constituents, violaxanthin, antheraxanthin, and zeaxanthin (V + A + Z), (**B**) lutein, (**C**) β-carotene, and (**D**) α-tocopherol per chlorophyll (Chl) *a* + *b*, (**E**) chlorophyll *a* + *b* levels per frond/leaf area, and (**F**) the molar ratio of chlorophylls *a* to *b* in fronds of *Lemna minor* (orange) and leaves of *Malva neglecta* (green) growing naturally in Superior, Colorado, USA. Mean values ± standard deviations; *n* = 4. Significant differences are denoted by asterisks; \*\* = *p* < 0.01, \*\*\* = *p* < 0.001.

**Figure 10.** (**A**) Score and (**B**) loading plots of the first two principal components from principal component analysis on correlations of chlorophyll (Chl) *a* + *b* per leaf/frond area, Chl *a*/*b* molar ratio, and levels of violaxanthin (V), β-carotene (β-C), antheraxanthin (A), zeaxanthin (Z), lutein (L), and neoxanthin (N) per Chl *a* + *b* in fronds of *Lemna gibba* grown under PFDs of 50 (dark green circles) and 1000 (bright yellow circles) μmol m−<sup>2</sup> s−<sup>1</sup> in growth chambers, and fronds of sun-grown *Lemna minor* (bright orange circles), leaves of sun-grown *Malva neglecta* (bright green circles), and a variety of sun-grown annuals (light green diamonds), shade-grown perennials (gray squares), and sun-grown perennials (light orange triangles). For details, see Tables S1–S3 and [21].

#### **4. Discussion**

The results of this study extend the conclusions reported in [2] of a notable ability of the duckweed *L. gibba* to grow rapidly across a range of growth PFDs from 100 to 700 μmol m−<sup>2</sup> s<sup>−</sup>1. The present study extended this growth PFD range to include an even lower intensity of 50 μmol photons m−<sup>2</sup> s−<sup>1</sup> and an even higher intensity of 1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> and documented the same high growth rate under the latter two extremes. This ability of duckweed to thrive under a wide range of light intensities makes sense in the context of duckweed ecology. Duckweed thrives in ponds where light environments can range from deep shade (at the pond's edge where emergent macrophytes and/or overhanging willows, etc., may provide considerable shade) to full sun in the middle of an open pond, with rapid cycles of vegetative expansion (e.g., during spring upon activation of overwintering turions, after a pond is disrupted by inclement weather involving wind and/or recharge with a major precipitation event, etc.) experienced periodically.

#### *4.1. Interspecies Comparison*

Foliar pigment composition of a closely related species, *L*. *minor*, growing on an open pond in full sunlight was similar to those of slow-growing evergreens, in particular with respect to the high maximal conversion of the xanthophyll cycle pool to zeaxanthin and antheraxanthin at midday in sunny habitats (see, e.g., [21,22]). Duckweed thus exhibited a combination of the traits of fast-growing annuals and slow-growing evergreens with foliar pigment features reminiscent of evergreens but coupled with a growth rate that exceeds that of the fastest-growing terrestrial plants [23].

Duckweeds are fast-growing aquatic plants that are members of the monocotyledonous order Arales; duckweeds were previously considered a subfamily (Lemnoideae) of the Araceae, but are now grouped as their own family (Lemnaceae), which is closely related to the Araceae (for recent reviews of its taxonomy, see [24,25]). Terrestrial Araceae are common in shaded rainforest environments, possessing adaptations for high shade tolerance that make them suitable as house plants (e.g., genera such as *Alocasia*, *Dieffenbachia*, *Monstera*, *Philodendron* [26–29]). The foliar pigment composition, and its response to growth PFD, seen here in duckweed, was reminiscent of a member of the Araceae, *Monstera deliciosa* (subfamily Monsteroideae), which possesses a remarkable ability to acclimate to both deep shade and full sunlight [30]. This makes sense in the context of the ecology of some aroids

as hemi-epiphytic vines that germinate in soil in deep shade on the rainforest floor, grow toward the darkest area (a tree trunk), climb into the forest canopy, lose connection to soil/ground, and thrive in dappled to full sunlight within the canopy [31–33].

On the other hand, their exceptionally high growth rates set duckweeds apart from terrestrial Araceae. The high growth rates of duckweed may be associated with genome reductions in these diminutive, floating plants that are associated with loss of controls placed on growth and stomatal conductance in terrestrial species [34,35]. Terrestrial species typically curb growth and stomatal opening under limiting water as a defense strategy [36]. This would appear less necessary for species that float on water, and duckweeds indeed impose much less control on either growth or stomata [35]. In addition, terrestrial plants curb growth under limiting nitrogen supply in the soil [37,38], whereas duckweeds have a particular propensity for effective nitrogen uptake from the growth medium [39] and an expanded set of genes for nitrogen uptake and amino acid synthesis [34].

#### *4.2. Specific Features That May Contribute to Duckweed's High Shade Tolerance*

As noted by Stewart et al. [2], *L. gibba* cultivated under favorable controlled conditions exhibited thin leaves with apparent minimal self-shading.

Concerning foliar pigment composition, principal component analysis revealed that neither of the two *Lemna* species examined here clustered with other fast-growing species, and instead clustered with slow-growing, highly shade-tolerant evergreens, and perennials growing in full sun (pond *L. minor* and *L. gibba* grown under the low PFD) or fell outside either cluster (*L. gibba* grown under the extremely high PFD). Foliar pigment patterns of evergreen and perennial species can be differentiated from those of fast-growing, herbaceous annuals, and biennials by comparatively high total Chl contents with lower levels of VAZ pool carotenoids and β-carotene relative to total Chl, and lower Chl *a*/*b* ratios but comparatively greater levels of lutein relative to Chl in the evergreens and perennials [21]. These features are all consistent with a high light-harvesting capacity associated with comparatively large Chl *b*- and lutein-containing light-harvesting antennae and smaller β-carotene-binding inner antennae [40]. One can thus describe *Lemna* as being unusual in combining fast growth with high shade tolerance. The shade tolerance of duckweeds and their ability to maintain a high growth rate under very low growth PFD may also be due to the fact that floating plants with much-reduced root structures do not need to support a significant proportion of non-photosynthetic tissue, which is challenging in very low light.

Another aspect of duckweed physiology that would support high growth rates in low light, where photosynthetic light-use efficiency must be high, is a rapid return to high photochemical efficiency upon transfer from high to low PFD. This was seen in the present study in the form of a rapid lowering of the rate of thermal energy dissipation upon transfer of high-light-grown *L. gibba* to low light with little to no sustained depression of photosystem II photochemical efficiency or photoinhibition. Notably, such rapid return to a high PSII photochemical efficiency is also seen in shade-tolerant species subsequent to exposure to rapid sunflecks in natural understory settings [3,41] as well as in sun-grown plants of terrestrial Araceae upon return to low light after extended exposure to high light ([42]; see also [43]).

Yet another feature *L. gibba* shares with terrestrial Araceae is maintenance of a similar photosynthetic capacity on an area basis across a wide range of growth PFDs, as reported by Stewart et al. [2] for *L. gibba* grown under PFDs from 100 to 700 μmol photons m−<sup>2</sup> s<sup>−</sup>1. This trend is also reminiscent of what was reported for *Monstera deliciosa*. When grown under high versus low PFDs, *M. deliciosa* maintained a similar photosynthetic capacity and adjusted its capacity for regulated, photoprotective dissipation of excess absorbed light (not utilized in photochemistry), whereas fast-growing annuals strongly adjusted their photosynthetic capacities with little to no difference in the capacity for photoprotective energy dissipation [30]. The finding of the present study that photosynthetic capacity as well as relative growth rate in fronds grown under 50 μmol photons m−<sup>2</sup> s−<sup>1</sup> was similarly high as those of fronds grown over a range of 100 to 700 μmol photons m−<sup>2</sup> s−<sup>1</sup> [2] further

supports *L. gibba*'s tendency to maintain a similar photosynthetic capacity across a wide range of growth PFDs (see more below on what that means for its high-light tolerance). Concerning *L. gibba*'s shade tolerance, one could speculate that its ability to support a high photosynthetic capacity under low growth PFD may contribute to its high growth rate under very low PFD. It is possible that a high level of the CO2-fixing enzyme Rubisco may be associated with duckweed's propensity to accumulate vegetative storage protein throughout the plant rather than storing protein only in the seed like soybean (duckweed can produce 20× more protein per hectare than soybean [6]). Martindale and Bowes [44] described an unusual propensity of duckweed to accumulate high levels of Rubisco across a range of growth PFDs. While *L. gibba* plants contained more protein on an area basis under 1000 versus 50 μmol photons m−<sup>2</sup> s−1, protein level relative to dry biomass was actually higher under the low growth PFD (biomass was 46% protein) compared to the high growth PFD (biomass was 25% protein on a gram per gram basis). High-quality plant-based protein from duckweeds could thus be produced highly efficiently under low growth PFD.

#### *4.3. Features That Likely Contribute to Duckweed's Tolerance of High Light*

Evergreens and perennials often exhibit relatively lower maximal electron transport rates associated with very high fractions (around 90%) of absorbed light allocated to nonphotochemical routes as well as very high fractions of the xanthophyll cycle pool converted to zeaxanthin at midday in sunny, but otherwise favorable, habitats [3,16,22,45]. In contrast, annuals and biennials often exhibit relatively higher maximal electron transport rates and relatively lower fractions (around 50%) of absorbed light allocated to non-photochemical routes and fractions of the xanthophyll cycle pool converted to zeaxanthin at midday in the same habitats [3,16,22,45]. In a comparison of the response of terrestrial annual species with the evergreen *M. deliciosa* to a range of growth PFDs, the annuals exhibited pronounced differences in photosynthetic capacity on a leaf area basis with no or only modest differences in photoprotective dissipation of excess excitation energy over a wide range of growth PFDs (with midday peaks of 300 versus 1500 μmol m−<sup>2</sup> s<sup>−</sup>1), whereas *M. deliciosa* showed no difference in photosynthetic capacity on a leaf area basis but a higher level of both thermal energy dissipation and zeaxanthin content at the higher growth PFD [30]. Duckweed exhibited similar features as *M. deliciosa*, with a pronounced increase in the fraction of absorbed light allocated to energy dissipation via regulated non-photochemical routes and of zeaxanthin accumulation, but no change in photosynthetic capacity, with increased growth PFD. This profoundly greater non-photochemical dissipation of absorbed light in the fronds grown under 1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> was apparently highly effective in limiting the build-up of excitation energy (absorbed light not utilized via either photochemistry or non-photochemical routes), as demonstrated in the present study.

At the same time, foliar pigment composition of *Lemna* was distinctive; even shadegrown fronds not containing zeaxanthin exhibited an overall pigment pattern similar to that of sun-grown terrestrial perennials. Furthermore, yellow fronds grown under an extremely high light supply exhibited a much-exaggerated version of this pattern. These features further illustrate that *Lemna* is unusual in combining fast growth with a distinct pigment composition. The low Chl *a* + *b* content, and concomitant high Chl *a*/*b* ratio, in *L gibba* grown under continuous light of 1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> is consistent with a strong downregulation of antenna size, which is different from the response of evergreens that, as their name indicates, exhibit limited variation of chlorophyll content. Further support for a small antenna size in *L. gibba* grown under the high PFD comes from the much lower neoxanthin concentration (lowered in proportion of Chl *a* + *b*) and the lower levels of β-carotene and lutein on a frond area basis. The fact that xanthophyll cycle pool, the concentrations of antheraxanthin and zeaxanthin, and the ratio of total xanthophylls to β-carotene were all greater on a frond area basis at the high PFD is presumably due to strong upregulation of zeaxanthin-based photoprotection that may take place not only in pigment-binding protein complexes, but also in the membrane phospholipid bilayer [46].

Since light supply was 20× greater at 1000 versus 50 <sup>μ</sup>mol photons m−<sup>2</sup> <sup>s</sup>−<sup>1</sup> and total dry biomass produced was only just under 3× greater, the biomass produced per mol photons (i.e., the light-use efficiency of biomass production) was dramatically lower at the high PFD. The significant linear relationships between the light-use efficiency of total biomass production and either the fluorescence-derived parameter Fv /Fm × qP or the fraction of the xanthophyll cycle pool converted to its de-epoxidized components showed that features associated with primary photosynthetic processes can serve as indicators of duckweed productivity across a range of growth PFDs, irrespective of possible variations of biomass composition with respect to the proportion of, for example, protein, starch, or pigments. Duckweed biomass is particularly valuable with high levels of protein and starch, as previously noted [2]. These relationships may also be as straight-forward in duckweed because this species consists of a one-layer canopy of fronds with minimal non-photosynthetic tissue.

Both Fv/Fm and Fv /Fm × qP were also correlated with measures of productivity in rice (see [47]). Prediction of productivity of other systems including whole ecosystems, from parameters associated with primary photosynthetic events is possible but requires consideration of additional features [48,49]. Whereas dark Fv/Fm was shown to be closely correlated with light-use efficiency of photosynthetic electron transport (from O2 evolution [50,51]), and Fv /Fm × qP is frequently used to estimate photochemical efficiency under illumination [52], these relationships can be tenuous [17] as was also recently discussed by Sipka et al. [53]. Nevertheless, our result that either Fv /Fm × qP or xanthophyll cycle pool conversion correlated closely with the light-use efficiency of plant productivity in duckweed is consistent with the assumption that the activity of any additional dissipative processes varies in proportion with the regulated non-photochemical dissipation of excitation energy associated with de-epoxidized xanthophyll cycle components in this species. Future research should examine a possible involvement of alternative photochemical sinks for excitation energy (other than carbon fixation [54]; see also [55,56]) such as oxygen reduction by electron transport, photorespiration, and nitrogen reduction (especially by plastid nitrite reductase). In *Lemna*, nitrogen reduction could be of interest because of the demonstrated enrichment in the duckweed genome of core enzymes in amino acid synthesis [34] and the propensity of duckweed to produce vegetative storage protein.

In the present study, light- and CO2-saturated maximal photosynthetic capacity on a frond area or dry mass basis was lower in fronds grown under 1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> compared to fronds grown under either 50 μmol photons m−<sup>2</sup> s−<sup>1</sup> (this report) or under 100 to 700 μmol photons m−<sup>2</sup> s−<sup>1</sup> [2]. This lower maximal photosynthetic capacity did not, however, lead to a lower growth rate in the fronds grown under 1000 versus 50 μmol photons m−<sup>2</sup> s<sup>−</sup>1, which indicates that at the very high growth PFD of 1000 μmol m−<sup>2</sup> s−<sup>1</sup> of continuous light, the somewhat lower photosynthetic capacity was sufficient to support the same high growth rate as at 50 μmol photons m−<sup>2</sup> s−1. The lower photosynthetic capacity on an area or dry mass basis in fronds grown under 1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> thus represents an adjustment that allows the plants to maintain a similar growth rate with a lesser photosynthetic capacity and a much lower chlorophyll content in a growth environment with a very high light supply. This feature could also be associated with duckweed accumulating Rubisco levels in excess of what is needed for CO2 fixation at a given time. Total Rubisco level may be associated with light- and CO2-saturated photosynthetic capacity, but varying proportions of this capacity may be sufficient to support growth under different growth PFDs [44]. The observed extremely high capacity of photosynthetic O2 evolution on a Chl basis in fronds grown under 1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> of continuous light is consistent with a strong preferential downregulation of antenna size relative to the components of photosynthetic electron transport under this high growth PFD.

Furthermore, *L. gibba* accumulated considerable starch at higher growth PFDs including under 700 μmol photons m−<sup>2</sup> s−<sup>1</sup> where no decline in photosynthetic capacity was seen [2], which suggests that the lower photosynthetic capacity in plants grown under

1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> may not be due to downregulation associated with starch accumulation. *Lemna*'s propensity for unabated growth and photosynthetic activity over a wide range of growth environments is consistent with the reported reduction of control by water and nutrient level in duckweeds coupled with their genome reduction and associated permanently open stomates and highly effective nutrient acquisition [35].

In a nutshell, *L. gibba* plants grown under continuous high light of 1000 μmol photons m−<sup>2</sup> s−<sup>1</sup> effectively counteracted build-up of potentially dangerous excess excitation through a combination of strong downregulation of antenna size (i.e., how much light is absorbed) with strong non-photochemical dissipation of excess absorbed light under the high growth PFD. These photoprotective processes did not interfere with the ability of photosynthetic electron transport to support similar area production and higher dry biomass production in the plants grown under 1000 versus 50 μmol photons m−<sup>2</sup> s<sup>−</sup>1. This high light tolerance of duckweed is reminiscent to that of the alga *Chlorella ohadii*, which can thrive under a light intensity equivalent to twice that of sunlight [57]. Under such high light conditions, this alga also exhibited unabated fast growth, a high photosynthesis rate, a small antenna size (a constitutive small antenna and an extremely high Chl *a*/*b* ratio of 13:1), and photoprotective energy dissipation that, however, relied on mechanisms other than zeaxanthin-associated non-photochemical energy dissipation [57]. The latter results and those reported here illustrate the existence of photosynthetic systems that grow at high rates under extremely high light and use unique combinations of photoprotective mechanisms. Duckweed is presumably well adapted to a range of natural environments that include predominantly either continuously shaded or high-light-exposed sites (where antenna size modulation should be particularly beneficial) as well as sites with rapidly fluctuating light (where the rapid reversibility of non-photochemical dissipation should be particularly beneficial).

In conclusion, the duckweed *L. gibba* was evidently able to acclimate to very high growth light intensity through a combination of a high growth rate with pronounced starch and protein accumulation, decreased light absorption (presumably by downregulation of antenna size), pronounced non-photochemical dissipation of excess light associated with zeaxanthin as well as other forms of photoprotection provided by the xanthophyll lutein (that can remove excitation energy from Chl in the triplet state [58]), β-carotene (that can contribute to the photoprotection of photosystem I [59]), and α-tocopherol (that can scavenge singlet oxygen and lipid peroxy radicals [60,61]). The greater levels of especially zeaxanthin, α-tocopherol, lutein, and to some extent β-carotene relative to Chl under high growth PFD are consistent with an enhanced need for photoprotection on part of the plant. From the standpoint of human/animal nutrition, however, production of micronutrients per area or as a percent of biomass matters most. Whereas zeaxanthin production required high light irrespective of reference basis, α-tocopherol (vitamin E), lutein, β-carotene (provitamin A), and protein levels as a percent of biomass were all lower under the high growth PFD. Therefore, a mixed lighting protocol with mainly low background PFD, supplemented with brief exposures to high light (see, e.g., [62]) might be attractive to produce high-quality nutrition for the consumer.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/cells10061481/s1, Table S1: Source, species, and grouping of data used for principal component analysis of pigment composition (see Figure 10), as well as scores for all eight of the resulting principal components; Table S2: Loading values of the eight pigment parameters used for principal component analysis (see Figure 10) for the eight resulting principal components; Table S3: Correlation matrix for the eight variables used in the principal component analysis of pigment composition (see Figure 10). **Author Contributions:** B.D.-A., C.M.E. and W.W.A.III wrote the grant on which this study is based. B.D.-A., W.W.A.III and J.J.S. planned the experiments with contributions from C.M.E., J.J.S., W.W.A.III, M.L.-P., N.D.G. and M.M. carried out the experiments and biochemical assays with contributions from B.D.-A., C.M.E., B.D.-A., W.W.A.III, J.J.S. and M.L.-P. analyzed and interpreted the data. B.D.-A., W.W.A.III and J.J.S. wrote the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was funded by the Translational Research Institute for Space Health through Cooperative Agreement NNX16AO69A, the National Science Foundation award number IOS-1907338, and the University of Colorado.

**Institutional Review Board Statement:** Not applicable for this work using only plants.

**Informed Consent Statement:** Not applicable for this work using only plants.

**Data Availability Statement:** The data presented in this study are available from the corresponding author upon reasonable request.

**Acknowledgments:** We thank Adam Escobar and Paul Fourounjian for the valuable discussion and feedback, and Gabrielle Glime for assistance with the data collection.

**Conflicts of Interest:** C.M.E. has a financial interest in Space Lab Technologies, LLC, a company that may be affected by this research. A Memorandum of Understanding approved by the University of Colorado manages potential conflicts arising from this relationship. The remaining authors declare no conflicts of interest. The sponsors had no role in the design, execution, interpretation, or writing of the study.

#### **References**

