Undaria pinnatifida gametophyte clones were established from mature sporophytes sampled at Geoje and Jindo Islands, Korea on Mar 1, 2013 and May 5, 2016, respectively. Sporulation, isolation, and culture were performed as previously described (Avila-Peltroche et al. 2020). Female (from Jindo Island) and male (from Geoje Island) vegetative gametophyte cultures in the early (4-day-old) and mid-exponential phase (6-day-old), respectively, were used for subsequent experiments.

Protoplast isolation was performed following the protocols described by Benet et al. (1997) and Coelho et al. (2012) with some modifications (Avila-Peltroche et al. 2020). Approximately 100 to 200 mg of U. pinnatifida gametophytes from 1-L round flasks were incubated in a 0.22 μm filter-sterilized enzymatic solution (640 mM NaCl, 208 mM MgCl₂·6H₂O, 35.2 mM MgSO₄, 256 mM KCl, 3.2 mM CaCl₂, 16 mM MES [pH 6], 2,512 mOsm L⁻¹ H₂O) at 20°C with shaking at 70 rpm for 4 h (female) or 6 h (male) in the dark. The solution was supplemented with 1% cellulase RS (Yakult Co. Ltd., Tokyo, Japan), 4 U mL⁻¹ alginate lyase, and 1% driselase from Basidiomycetes sp. (Sigma-Aldrich, St. Louis, MO, USA). Chelation pre-treatment was conducted with a calcium-chelating solution (665 mM NaCl, 30 mM MgCl₂·6H₂O, 30 mM MgSO₄, 20 mM KCl, and 20 mM ethylene glycol-bis (β-amino-ethyl ether)-N,N,N′,N′-tetraacetic acid tetrasodium salt (EGTA-Na₄) [pH 5.5] as calcium chelator) for 20 min prior to enzymatic digestion (Coelho et al. 2012). Protoplasts were filtered using a 25-μm nylon mesh to remove undigested filaments and further washed three times by centrifugation at 100 ×g for 10 min. The pellets were resuspended in an enzymatic solution. The average number of gametophyte-isolated protoplasts was calculated with a hemocytometer. For male gametophytes, the fresh weight was approximately (3.12 ± 0.51) × 10⁶ protoplasts g⁻¹ and for male gametophytes, (2.11 ± 0.08) × 10⁶ protoplasts g⁻¹. The viability of protoplasts and cell wall regeneration were assessed by red chlorophyll autofluorescence and staining with calcofluor white M2R (Sigma-Aldrich), respectively, as previously described (Avila-Peltroche et al. 2019). Calcofluor white was chosen for monitoring cell wall formation because it binds to β-linked fibrillar polymers such as cellulose but does not block its synthesis (Murgui et al. 1985).

The effects of LED wavelengths on regeneration of the gametophyte-isolated protoplasts were investigated using a DYLED 44V system (DyneBio Co., Seongnam, Korea). Five different LEDs were installed in the growth chambers: (1) white (all wavelengths), (2) red (peak wavelength at 660 nm), (3) blue (peak wavelength at 460 nm), (4) mixed red and blue (1 : 2; RB), and (5) green (peak wavelength at 530 nm). Fig. 1 shows the emission spectra of the LEDs. Spectral wavelengths were measured using the USB2000 UV-VIS spectrometer (Ocean Optics Inc., Dunedin, FL, USA), while the LED irradiation intensities were measured using a quantum meter (MQ-500; Apogee Instruments, Logan, UT, USA). The protoplasts were grown under LEDs on a 14 : 10 h light : dark cycle.

Protoplasts were dispensed into at least 4 wells of a 12-well tissue culture plate containing 2 mL of regeneration medium (PES with 285 mM NaCl and 5 mM CaCl₂). The medium was supplemented with an antibiotic mix (50 mg L⁻¹ penicillin G, 25 mg L⁻¹ streptomycin, and 5 mg L⁻¹ chloramphenicol). The plates were incubated at 20°C in the dark and with an initial protoplast density of 9 × 10³ protoplasts. Previous experiments showed that 20°C was suitable for successful protoplast regeneration. Also, this temperature promotes vegetative growth of U. pinnatifida gametophytes (Morita et al. 2003). After 4 days in the dark, osmotic pressure was reduced slowly using a PES medium, and cultures were gradually exposed to a final intensity of 40 μmol photons m⁻² s⁻¹. The well plates were separately irradiated using white (control), red, blue, RB, or green LEDs. The culture medium was replaced weekly. The responses of cultured protoplasts were assessed during the early and late stages of regeneration. In the early stages, the percentage of dividing cells was calculated at 8 (female) and 11 (male) days in culture. In addition, the percentage of asymmetric cells formed by cell divisions was determined for all conditions because LED lighting promotes the formation of asymmetric cells in protoplasts of the moss Physcomitrella (Jenkins and Cove 1983). The percentage of branched filaments (BFs) or multiple rhizoid-like protrusions (MRLPs) was determined at 15 or 18 days in culture, respectively. Final growth was represented as the gametophyte area at the end of the experiment (female = 22 days in culture; male = 26 days in culture) and analyzed using image analysis software (ImageJ; US National Institutes of Health, Bethesda, MD, USA). For percentage calculations, at least 100 cells per well were counted. For area measurements, at least 30 gametophytes per well were analyzed.

Chlorophyll a (Chl a), chlorophyll c (Chl c), and fucoxanthin (Fx) were extracted from regenerated female gametophytes (10–20 mg fresh weight) at the end of the experiments using the protocol and equations of Seely et al. (1972) with dimethyl sulfoxide as the solvent. Regenerated male gametophytes were further cultured under the same conditions for 3 months to obtain sufficient biomass for reliable measurements. A microplate reader (SpectraMax Paradigm; Molecular Devices, San Jose, CA, USA) was used to quantify the pigments from at least three wells per LED treatment. Pathlength correction was performed according to the method described by Warren (2008).

The effect of LED light intensity on protoplast regeneration was evaluated under RB LED treatment. The gametophyte-isolated protoplasts were cultured as described above in at least 4 wells of a 12-well plate. They were placed at different distances from the LED unit to achieve different intensities: 20, 40, 60, and 80 μmol photons m⁻² s⁻¹. The light intensities were measured using a quantum meter. The regenerative responses of the cultured protoplasts were assessed as previously described.

The percentage of dividing cells, asymmetric cells, BFs, or MRLPs, were analyzed as proportions with beta regressions since beta distribution provides a flexible model for continuous variables restricted to the interval (0, 1) (Ferrari and Cribari-Neto 2004). The analyses were performed using the R package betareg (Cribari-Neto and Zeileis 2010). The proportional reduction of error (PRE) statistic was used as the overall model effect size (Smithson and Verkuilen 2006).

Two separated one-way ANOVA tests were used to evaluate the differences in gametophyte area and pigment composition. Normality and homoscedasticity were checked using the Shapiro-Wilk and Levene test, respectively, before conducting ANOVA. Kruskal-Wallis or Welch’s ANOVA tests were used when data did not meet normality or homoscedasticity assumptions, respectively. Effect sizes (Sullivan and Feinn 2012) were presented as ω² (Field 2009, Lakens 2013) in the case of significant results. The formula for ω² calculation was as follows:

, where SS_M is the between-group effect, df_M is the degrees of freedom for the effect, MS_R is the regression mean square, and SS_T is the total amount of variance in the data. All the analyses were performed using the car (Fox and Weisberg 2019) and userfriendlyscience packages (Peters 2018) in R.

Tukey’s post hoc test was used when the results were significant. When data were heteroscedastic, the Games-Howell test was performed instead. The post hoc comparisons were conducted using the multcomp (Hothorn et al. 2008) or userfriendlyscience package (Peters 2018) in R.

The significance threshold was set at p = 0.01 to reduce the true type I error rate (at least 7% but typically close to 15%) (Sellke et al. 2001). It is worth mentioning that this precautionary approach also reduces the chances to detect false negatives (type II error). All graphs were created using GraphPad Prism 6.0 (GraphPad Software, La Jolla, CA, USA).

Freshly isolated protoplasts showed 90 to 95% viability according to red chlorophyll autofluorescence. They also had no cell wall, as shown by the absence of a blue fluorescent signal after calcofluor white staining. After 3 hours of culture, cell wall formation started as blue fluorescent spots (positive staining) on the protoplast surface (Fig. 2A). These spots grew and connected each other forming an extensive area of cell wall deposition after 16 h of culture (Fig. 2B) The entire surface of the protoplast was covered by a uniform cell wall after 2 to 6 d of culture (Fig. 2C).

Protoplast regeneration started as a single bud on one pole before the first asymmetric cell division (Fig. 2D & G). Symmetric cell division was also observed among the regenerated protoplasts but at lower percentages (14–20%). After 15 d in culture, most regenerated protoplasts isolated from female gametophytes followed unipolar germination (81%). The 15-day-old female gametophytes were visible as uniseriate BFs (Fig. 2E). New thalli were distinguished after 22 d of culture. Protoplasts isolated from male gametophytes produced MRLPs after 18 d in culture (Fig. 2H). New thalli were visible after 26 d of culture. Both regenerated gametophytes developed morphologically normal reproductive tissues (Fig. 2F & I).

Overall, all analyzed LED conditions induced protoplast cell divisions. In the early stages, LEDs affected the percentage of dividing cells (p < 0.001, PRE = 0.133). Moreover, blue and RB LEDs significantly increased these values by 1.5 and 1.7 times, respectively, compared with the control (Fig. 3A). LED treatments influenced the number of asymmetric cells (p < 0.001, PRE = 0.088). More asymmetric cells formed under the blue and RB conditions; however, they represented a small increase (1.1–1.2 times) compared with the control (Fig. 3B). In the late stages, LED treatments did not affect the percentage of BFs (p < 0.01) (Fig. 3C), but RB LEDs promoted a significant increase in the final gametophyte area by three times compared with white LEDs (p < 0.001, ω² = 0.78) (Fig. 3D). LEDs also affected the final gametophyte morphology. Under red LEDs, gametophytes showed longer filaments and more branches than those under the blue, which only presented one or none. The RB treatment produced gametophytes with longer filaments and more branches than those under the blue treatment. They were also slightly shorter than gametophytes under the red condition. Gametophytes under green LEDs showed a morphology resembling white LED conditions (Fig. 5A). Chl a, Chl c, and Fx contents were not affected by the treatments (p > 0.01) (Fig. 6A–C).

All LED conditions allowed protoplast cell divisions. In the early stages, LEDs affected the percentage of dividing cells (p < 0.001, PRE = 0.255). Red, blue, and RB LEDs significantly increased these values by 2–2.5 times, respectively, compared with the control (Fig. 4A). LED treatments also influenced the number of asymmetric cells (p < 0.001, PRE = 0.041). More asymmetric cells were formed under the green condition; however, this represented a small increase (1.1 times) compared with the control (Fig. 4B). In the late stages, only RB LEDs increased the percentage of MRLP (p < 0.001; PRE = 0.138) and final gametophyte area (p < 0.001, ω² = 0.87) by 1.2 and 3.6 times, respectively, compared with white LEDs (Fig. 4C & D). LEDs also affected the final gametophyte morphology. Under red LEDs, gametophytes showed longer filaments with less profuse branching than those under the blue. Gametophytes grown under the RB treatment had an “intermediate” morphology. They had longer filaments than the gametophytes under the blue and more profuse branching than those under the red treatment. Gametophytes under green LEDs showed a morphology resembling white and red conditions (Fig. 5B). The Chl a content was not affected by LED treatments (p = 0.293). The Chl c content dropped under all conditions compared with the white (p < 0.001, ω² = 0.46), whereas the Fx content decreased under blue, RB, and green conditions compared with the control (p = 0.001, ω² = 0.48) (Fig. 6D–F).

RB LEDs were chosen for testing different light intensities on protoplast regeneration because they improved it in both protoplast types. In the early stages, the percentage of dividing cells was higher under 40 and 60 μmol photons m⁻² s⁻¹ than at 20 or 80 μmol photons m⁻² s⁻¹ for both gametophytes (p < 0.001; PRE [female] = 0.134; PRE [male] = 0.236) (Fig. 7A & D). In the late stages, the percentage of BFs increased at 80 μmol photons m⁻² s⁻¹ compared with the other light intensities (p = 0.002, PRE = 0.118) (Fig. 7B), while the percentage of MLRPs was higher at 40 and 60 μmol photons m⁻² s⁻¹ than at 20 or 80 μmol photons m⁻² s⁻¹ (p < 0.001, PRE = 0.814) (Fig. 7E). The final female gametophyte area peaked at 80 μmol photons m⁻² s⁻¹ (p = 0.002, ω² = 0.46), representing 1.8 times increase compared with the lowest value at 20 μmol photons m⁻² s⁻¹ (Fig. 7C). The final male gametophyte area showed the highest value at 60 μmol photons m⁻² s⁻¹ (p < 0.001, ω² = 0.87), representing 2.9 times increase compared with the lowest value at 20 μmol photons m⁻² s⁻¹ (Fig. 7F).