Photosynthesis and growth of Ulva ohnoi and Ulva pertusa (Ulvophyceae) under high light and high temperature conditions, and implications for green tide in Japan

The macroalga Ulva ohnoi constitutes a considerable fraction of green tides in coastal areas of Japan, but little is known about the physiological characteristics of this species. To investigate the environmental factors that promote the formation of green tides, we tested the responses of U. ohnoi and another common Japanese species, Ulva pertusa, to various levels of irradiance at different water temperatures. Because the two species are morphologically similar, we identified them using the PCR‐restriction fragment length polymorphism method. Under laboratory conditions, we evaluated the photosynthetic, dark respiration, and relative growth rate at a range of water temperatures (5 to 35°C) and photosynthetically active radiation (0 to 1000 μmol photons m−2 s−1). The maximum gross photosynthetic rate of U. ohnoi was larger than that of U. pertusa. The dark respiration rates revealed no significant differences among the species and temperature conditions. At 500 μmol photons m−2 s−1, the relative growth rate of U. ohnoi was larger than that of U. pertusa in higher temperature and the difference was the largest at 20°C. The estimated compensation irradiance and estimated saturation irradiance of U. ohnoi and U. pertusa ranged from 0.709 to 5.510 and 40.530 to 58.674 μmol photons m−2 s−1, which were lower than those in other intertidal green macroalgae, from 6 to 11 and 50 to 82 μmol photons m−2 s−1, respectively. Thus, U. ohnoi which exists as free‐floating near the water surface and accumulating inside the green tide can survive extensively in the water column of the intertidal zone, furthermore, the species can maintain rapid growth in this situation. Therefore, as a result of this study, it is suggested that the ecological success of U. ohnoi in shallow waters such as the tidal flats, estuarine, and coasts of the inner bay in comparison with U. pertusa.


INTRODUCTION
In recent years, blooms of green macroalgae, known as 'green tide,' have occurred in eutrophic inner bays around the world (Fletcher 1996). Although several definitions of green tide have been proposed (Fletcher 1996;Ohno 1999;Taylor et al. 2001;Nelson et al. 2003), in general, it refers to large blooms and accumulation of green macroalgae of the Ulvophyceae (e.g., Ulva pertusa Kjellman; Yamochi 2013, Ulva prolifera Müller and Ulva intestinalis Linnaeus; Wang et al. 2012) in shallow waters (e.g., tidal flats, estuaries, and coasts of inner bays). In Japan, green tide has been reported in various locations, such as Tokyo Bay, Osaka Bay, Hiroshima Bay, and Hakata Bay Uchimura et al. 2004;Yabe et al. 2009;Yamochi 2013). Green tide adversely affects other primary producers in shallow waters because green macroalgae completely cover these species, inhibiting photosynthesis (Ohno 1999). The anoxic conditions resulting from decomposition of the green macroalgae cause a release of gaseous sulfur compounds such as hydrogen sulfide, carbon disulfide, and exposure to these toxic gases poses health risks to both humans and wildlife (Charlier et al. 2007;Peu et al. 2011). In addition, the accumulation and degradation of a large amount of Ulva blooms have serious impacts on local biodiversity and biogeochemistry (Frankenstein 2000).
Ulva pertusa is a common inhabitant of intertidal areas in Japan and is reported to form green tide (Okamura 1921). Arasaki (1984) reported that an invasive species of Ulva, other than U. pertusa, is involved in green tide formation. Ohno (1988) investigated which Ulva species form green tide in Uranouchi Bay (Kochi Prefecture) and concluded that green tide is mainly composed of a newly invasive Ulva species from subtropical marine environments. Hiraoka et al. (2004) showed that the species found in Uranouchi Bay is identical to Ulva ohnoi Hiraoka et Shimada, based on both morphological characteristics and DNA sequence data. They reported that U. ohnoi is also present in coastal areas in Fukuoka and Okinawa. Ulva ohnoi is considered to be the major green tide-forming species in the tidal flats of Tokyo Bay (Yabe & Ishii 2015) and Osaka Bay (Yoshimura & Yamochi 2011).
To clarify the causes of green tide, many studies have investigated the physiological characteristics of green macroalgae under experimental conditions. Some studies have shown that nitrogen application increases both photosynthetic rate and growth of Ulva (Sfriso et al. 1988;Fong et al. 1993;Ménesguen & Piriou 1995;Pedersen 1995), suggesting that nitrogen availability is a major limiting factor in the generation of green tide. In the past 20 years, the nitrogen concentrations of inner bays adjacent to urban areas have decreased in Japan (Kazama & Ando 2010;Tada et al. 2010). However, the generation of green tide is still observed in these areas. Thus, we focused on the photosynthesis and growth responses of Ulva species to understand recent green tide formation in Japan.
As in the case of U. pertusa, the previous study showed the light intensity is one of the most important factors of growth in the species (Floreto et al. 1993). However, these characteristics of U. ohnoi have not been investigated (Yabe et al. 2009), and this lack of basic knowledge hinders our understanding of how U. ohnoi has formed green tide in recent years. In green tides, accumulated Ulva spp. in the surface layer are exposed to intense light and high temperature, whereas those in the bottom layer are in dark conditions. In this study, we investigated the photosynthesis and growth responses of two common Ulva species in Japan, U. ohnoi and U. pertusa, under experimental conditions, and discussed how U. ohnoi becomes a dominant green tide species in these tidal flats.

Sampling site and stock cultures
Samples were collected from the Yatsu tidal flat (35 40 0 34.1 00 N, 140 00 0 31.8 00 E) in inner Tokyo Bay. The green tide at this tidal flat was first reported in 1995 and was caused mainly by freefloating Ulva (Yabe et al. 2009). Samples were stored in a plastic bag and transported directly to the laboratory, where they were immediately briefly rinsed with fresh water to remove any surface contaminants and detritus. Collected Ulva was initially cultivated in a 100-L culture tank (SBF-100 TO, Tanaka Sanjiro, Fukuoka, Japan) containing artificial seawater (Marine Art SF-1, Tomita Pharmaceutical, Tokushima, Japan; 36.4 g L −1 ) and nutrients (KW 21, Daiichi Kogyo, Kumamoto, Japan; 0.5 mL L −1 ) in a naturally lit glass room controlled at 25 C for 2 weeks according to Kishida and Baba (2003).
Because Ulva species forming green tides are morphologically similar, genetic analysis is needed to identify them; yet, the previous studies on the physiological properties of green tide species (Murase et al. 1993;Kishida & Baba 2003) have been performed without genetic analysis. In this study, PCRrestriction fragment length polymorphism method (Yabe & Ishii 2015) was used for species identification. After the identification, U. ohnoi and U. pertusa were cultured separately as described above.

Photosynthesis and respiration rates
Measurements of photosynthesis and respiration were carried out by measuring dissolved oxygen (DO) concentration using a set of three punched round pieces (25 mm in diameter) from each thallus except the parts of the outer edge, scratches, and wrinkles. Fresh weight (gFW) of round pieces of Ulva was measured, and three of those were placed in a 100-mL DO bottle filled with artificial seawater. Dark bottles were covered with aluminum foil. Samples were incubated for 2 h at different temperatures 10, 20 and 30 C. A multi-step light receiver ( Fig. 1) was placed in a growth chamber (FLI-2000A, Tokyo Rikakikai, Tokyo, Japan) equipped with a high-intensity HID light unit (HLU-50, Tokyo Rikakikai). As shown in Figure 1, bottles were placed on plates at different distances from HID light. Photosynthetically active radiation (PAR) was measured with a quantum sensor (MQ-200, Apogee Instruments, UT, USA), and adjusted PAR 0, 50, 100, 250, 500, 750, and 1000 μmol photons m −2 s −1 . In this experiment, three DO bottles were attached to a plate that rotated at 5 rpm and was irradiated.
Dissolved oxygen concentrations were measured before and after incubation by the light and dark bottle method (Ikusima 1965), as follows. Ulva samples were acclimated in the experimental temperatures overnight in the dark to avoid cut influence (Serisawa et al. 2001;Watanabe et al. 2014). Dissolved oxygen concentration was measured using a fluorescent oxygen sensor and a DO meter (MicroxTX3, PreSens, Bayern, Germany) at the beginning and end of incubation. The net photosynthetic rate (P n ) and dark respiration rate (Rd) were normalized in two ways, to dry weight (gDW) and to chlorophyll a (Chl a) concentration. To obtain the wet/dry weight ratio, dry weight was determined after drying at 80 C for at least 48 h. After drying, samples were kept in a desiccator until it became a constant weight.
To measure Chl a content, other samples from the same thallus used in photosynthetic measurement were cut into circles with an 11 mm diameter, homogenized with a mortar and pestle, and extracted in 10 mL of absolute methanol under refrigeration. After 24 h, the absorbance of the methanol extract was measured using a spectrometer (UV-1800, The P n was obtained by the following equation (Yamochi 2013): where P n is the net photosynthetic rate of the sample.
, and W is the amount of Chl a in the explant used for culture (mg Chl a) or dry weight (gDW).
To obtain and compare parameters for photosynthesisirradiance (P-I) curve among different conditions (e.g. two or more species; different temperature conditions), many of earlier studies take a two-step analysis as follow; P-I curve is obtained using the intact data of photosynthetic rate and subsequently perform the analysis of variance on P-I parameters derived from the P-I curve (e.g. Murakami et al. 2004;Lideman et al., 2011;Zou & Gao 2014). The two-step analysis requires an assumption that the photosynthetic data are regarded as a paired-samples across light intensity (e.g. a series of samples are taken from the same individual, or the same samples are repeatedly used across light intensity). However, some recent studies take more general solution that can be applied to random independent samples, e.g. fitting the data to a non-linear P-I model and estimating P-I parameters and their differences at a time (e.g. Fujimoto et al. 2014;Borlongan et al. 2017). The data in our study were not paired and therefore we applied the latter approach by using the following P-I curve (Webb et al. 1974;Jassby & Platt 1976;Platt et al. 1980;Henley 1993) as the basal equation to fit P n : where P max is the maximum gross photosynthetic rate, α is the initial slope of the photosynthesis, Ir is light intensity, and Rd is the dark respiration rate. To analyze photosynthesis rate with the differences between species, among temperature levels and their interactions (i.e., two-way non-linear model of the analysis of covariance), we subdivided each of the parameters in Equation 3 as: Here, we coded the species and temperature as zero and one as follows: Sp: U. pertusa = 0 and U. ohnoi = 1; T 1 : 20 C = 1, and 0 for others; T 2 : 30 C = 1, and 0 for others. There are 18 parameters to be estimated: P max ', α', and Rd' as baseline parameters (U. pertusa at 10 C), and their sub-parameters a k , b k , and c k (k = 1-5), which subdivide P max ', α', and Rd', respectively. k = 1, 2, and 3 represent difference in each of the parameters between U. pertusa and U. ohnoi, at 20 C and at 30 C, respectively. k = 4 and k = 5 indicate the interactions for U. ohnoi at 20 C and at 30 C conditions, respectively.
Since Rd is the dark respiration rate and only additive in Equation 3, Rd' and its sub-parameters (c k ) should be estimated solely based on the data at 0 μmol photons m −2 s −1 , separately from the other parameters (particularly α), to avoid correlated estimations. Therefore, we estimated the P-I parameters by using the following two-parts of models. The two-way model of the analysis of variance was used for Rd' and its sub-parameters (c k ), based on the data at 0 μmol photons m −2 s −1 with the differences of species and temperature including their interactions. Adding the separately estimated Rd, we fitted the data to Equation 3. Equation 3 was fitted for both of the P n per dry weight and Chl a content of Ulva species. The saturation irradiance (Ik) and compensation irradiance (Ic) were also calculated as P max /α and P max log (P max / (P max -Rd))/α, respectively (Schubert et al. 2006).

Relative growth rate
Measurements of growth were performed using these round pieces taken from each thallus in the same way as the experiments for measurement of photosynthesis and respiration described above. Samples from stock cultures were preincubated under each experimental condition overnight. Samples were placed into a culture bottle (CB-2, AS ONE Co., Osaka, Japan) containing 300 mL of artificial seawater. Culture bottles were randomly placed in the incubation chambers and cultured in triplicate for 7 days at low and high PAR (60 and 500 μmol photons m −2 s −1 , respectively; a photoperiod of 12 : 12 h LD) and several water temperatures (60 μmol photons m −2 s −1 : 5 to 35 C at an interval 5 C, 500 μmol photons m −2 s −1 : 10 to 30 C at an interval 10 C). For low PAR experiments, a temperature gradient growth chamber (TG-280CCFL-5LD, NK system, Osaka, Japan) with LED light source was used and the other growth chamber (FLI-2000A) was used for the high PAR experiments. The samples were manually stirred three times daily during experiments.
The fresh weight of samples was measured at the beginning and end of the culture period, and growth was calculated as the relative growth rate (RGR) (Takada et al. 2011): where W 2 is the final fresh weight (gFW), W 1 is the initial fresh weight (gFW) and 7 is the number of days of culture. The RGRs were analyzed using a quadratic model of the analysis of covariance (ANCOVA) with species as the factor and temperature as the covariance each for the two light levels (60 and 500 μmol photons m −2 s −1 ).

Statistical assumptions and computations
We performed a nonlinear least-squares estimation via a Gauss-Newton algorithm using 'nls' function of R statistics to fit Equation 3, which were successfully converged. The R code to fit Equation 3 is shown in Appendix S1. The datasets of P-I and RGR are also shown in Appendix S2 and S3, respectively. Normality assumptions were checked graphically with quantile-quantile plots of residuals against fitted values. The points distributed around a diagonal line, indicating normality. All of the statistical analyses were carried out by using R version 3.6.1 (R Development Core Team 2019).

Effect of temperature on P-I curves
The P-I curves and parameters of P-I curves in U. ohnoi and U. pertusa grown at different temperatures are presented as per gDW (Fig. 2 and Tables 1 and 2). The P max changed with temperature and was the highest at 20 C (Tables 1 and 2). The P max of U. ohnoi were larger than that of U. pertusa under all temperature conditions, and the species difference in P max was the largest at 20 C and still large at 30 C but marginal at 10 C (Table 1). The slope α was significantly larger than zero and at the largest at 20 C, but there were no other significant differences among the species and temperature conditions (Table 1). For the Rd, there were also no significant differences among the species and temperature conditions ( Table 1). The Ic of U. ohnoi and U. pertusa ranged from 1.291 to 3.040 and 3.476 to 4.345 μmol photons m −2 s −1 , respectively (Table 2). Moreover, the estimated Ik of U. ohnoi and U. pertusa were 48.205 to 60.697 and 37.077 to 49.318 μmol photons m −2 s −1 , respectively (Table 2). These results show that the P max reached its saturation around 50 μmol photons m −2 s −1 in both species regardless of growth temperature.
The results of P-I curves and estimated P-I curve parameters normalized to Chl a are also shown in Appendix S1 (Fig. S1, Tables S1 and S2).

Effect of temperature on relative growth rate
We compared the RGR of U. ohnoi and U. pertusa at different growth temperatures and a low and high PAR (Fig. 3a,b). Under the low light experiment (60 μmol photons m −2 s −1 ), the RGR showed a convex pattern with its maximum around 20 C, however the species difference in RGR pattern was negligible (Table 3). No increase in weight was observed in either species at 5 and 35 C (Fig. 3a). Under the high PAR experiment (500 μmol photons m −2 s −1 ), the RGR also showed a convex pattern with its maximum around 20 C, though the RGR of U. ohnoi was larger than that of U. pertusa in higher temperature and the difference was the largest at 20 C ( Fig. 3b and Tables 3 and 4).

DISCUSSION
Photosynthetic features of U. ohnoi and U. pertusa based on the functional form Littler and Arnold (1982) studied the primary productivity of 62 species of macroalgae and classified them into six functional form groups. All Ulva species were classified into the thin, tubular and sheet-like group, which showed the highest productivity among the six groups. In this study, P max values of U. ohnoi (19.09 O 2 mg gDW −1 h −1 ) and U. pertusa (16.92 mg O 2 mg gDW −1 h −1 ) were higher than those of intertidal macroalgae species such as Chaetomorpha linum (8.14 mg O 2 gDW −1 h −1 ), Gracilaria verrucosa Hudson (coarsely-branched group; 13.58 mg O 2 gDW −1 h −1 ), and Ulva sp. (16.54 mg O 2 gDW −1 h −1 ) (Menéndez et al. 2001). In addition, the Ic and Ik of U. ohnoi and U. pertusa ranged from 1.291 to 4.345 and 37.077 to 60.697 μmol photons m −2 s −1 , respectively. Ulva species is well known to adapt to low irradiance of 0.6 to 60 μmol photons m −2 s −1 (Fortes & Lüning 1980;Vermaat & Sand-Jensen 1987;Sand-Jensen & Borum 1988;Pérez-Lloréns et al. 1996;Taylor et al. 2001). Furthermore, Merceron et al. (2007) concluded that Ulva In the case of other intertidal green macroalgae (Chlorophyta), Ic, and Ik values in C. linum Müller (delicatelybranched group), U. intestinalis, U. lobata Kützing, and U. rigida C. Agardh have been studied in the literature, and shown that Ic and Ik in these species are, 6 to 11 and 50 to 82 μmol photons m −2 s −1 , respectively (Arnold & Murray 1980). In the case of other intertidal red macroalgae (Rhodophyta), Ic and Ik values in Pyropia tenera Kjellman (thin, tubular and sheet-like group) were 9 and 46 μmol photons m −2 s −1 , respectively (Watanabe et al. 2014). Here, we show that U. ohnoi and U. pertusa have lower Ic and Ik in comparison to other intertidal macroalgae species, suggesting that U. ohnoi and U. pertusa can maintain the sufficient primary production for their survival even if they are accumulated on top of each other in green tide.
It is noteworthy to write that U. ohnoi and U. pertusa showed the low Ic, and the P max of U. ohnoi and U. pertusa were higher than other intertidal algae. These indicate that U. ohnoi and U. pertusa survive under the lower light condition inside the green tide and maintained the higher primary productivity under high light conditions near the water surface, compared to other intertidal macroalgae.
Interaction among species and growth temperature based on P-I curve parameters and relative growth rate It is noteworthy that we succeed in showing that the RGR of U. ohnoi at temperatures higher than 20 C are significantly higher than those of U. pertusa, because the previous study only showed the intertidal macroalgae can maintain their *P < 0.05. **P < 0.01. ***P < 0.001. Ulva pertusa at 10 C was used as the basal group for intercept and slope. P max ', α', and Rd' as basal parameters (U. pertusa at 10 C), and their sub-parameters a k , b k , and c k (k = 1-5), which subdivide P max ', α', and Rd', respectively. k = 1, 2, and 3 represent difference in each of the parameters between U. pertusa and U. ohnoi, at 20 C, and at 30 C, respectively. k = 4 and k = 5 indicate the interactions for U. ohnoi on 20 C and on 30 C conditions, respectively. Estimated, estimated P-I curve parameters; SE, standard error.  (Maegawa & Sugiyama 1995). Taken together, these results indicate that U. ohnoi can adapt successfully to high water temperature than U. pertusa. This indication may support a possible distribution of Ulva species in Japan studied by Hiraoka et al. (2004). They described that U. ohnoi mainly forms green tide in warm-temperate coasts of the southwestern region, while U. pertusa forms green tide on a little more north-eastern region of Japan.
In this study, we showed that the P max value and RGR of U. ohnoi were significantly higher than that of U. pertusa. The results indicate that U. ohnoi has larger rates of photosynthesis and growth than U. pertusa. This suggests that U. ohnoi may survive, grow extensively and dominate in the natural habitats with severe environmental conditions, such as shallow tidal flats. Moreover, it is also suggested that U. pertusa may only weakly respond to temperature changes as well as U. ohnoi, because the RGR curve for U. pertusa was flatter than that of U. ohnoi in the high PAR experiment. Both channels and waterways distributed on tidal flats seem to be good habitats for U. pertusa because there are no severe environmental conditions for the species. The co-occurrence of two Ulva species was also observed in the Yellow Sea: the floating-form of U. prolifera and the attached-form of U. intestinalis (Wang et al. 2012).
In the near future, water temperatures and nutrients are expected to increase in the shallow waters of Japan according to global climate change forecasts, which are suitable physiological conditions for the expansion of green tide (Wells et al.  Ulva pertusa at 10 C was used as the basal group for intercept and slope. Estimated, estimated P-I curve parameters; SE, standard error; temperature, first-order terms of quadratic function; Temperature 2 , second-order terms of quadratic function.

2015)
, and the expansion is expected to be caused by U. ohnoi. Nonetheless, there are many other factors that can affect the expansion of green tide including phenology of Ulva spp., competition with other primary producers and herbivory in natural habitat. Therefore, further studies including field manipulative experiments and nutrient uptake experiments are also needed to project future expansions in green tide.
In conclusion, we showed that U. ohnoi which exists as free-floating near the water surface and accumulating inside the green tide can survive extensively in the water column of the intertidal zone, and the species maintains rapid growth in this situation. Therefore, this study suggests that U. ohnoi will be more ecologically successful than U. pertusa, in shallow waters such as tidal flats, estuaries, and the coastal areas of the inner parts of bays. Saco et al. (2018) observed similar phenomena in two benthic green algae (Monostroma angicava Kjellman and Protomonostroma undulatum Vinogradova), which coexist in intertidal habitats but express different photosynthetic characteristics. Characteristics of the dominant species, U. ohnoi, clarified in this study may help to suppress green tide formation. Appendix S1. The R code to fit Equation 3. Table S1. Parameter estimates of the P-I curve per Chl a. Table S2. Extended parameters of the P-I curve per Chl a.