*Article* **Coagulant Plus** *Bacillus nitratireducens* **Fermentation Broth Technique Provides a Rapid Algicidal Effect of Toxic Red Tide Dinoflagellate**

**Barathan Balaji Prasath 1,2,†, Ying Wang 1,†, Yuping Su 1,2,\*, Wanning Zheng 1, Hong Lin <sup>1</sup> and Hong Yang <sup>3</sup>**


**Abstract:** When the toxic red tide alga *Gymnodinium catenatum* H.W. Graham accumulates in sediment through sexual reproduction, it provides the provenance of a periodic outbreak of red tide, a potential threat to the marine environment. In our study, the flocculation effects of four coagulants were compared. Bacteria fermentation (Ba3) broth and coagulant were combined with Ba3 to reduce the vegetative cells of *G. catenatum*, inhibit the cystic germination in the sediment, and control the red tide outbreak. To promote a more efficient and environmentally friendly algae suppression method, we studied these four coagulants combined with algae suppression bacteria for their effect on *G. catenatum*. The results show that polyaluminum chloride (PAC) is more efficient than other coagulants when used alone because it had a more substantial inhibitory effect. Ba3 broth also had a beneficial removal effect on the vegetative cells of *G. catenatum*. The inhibition efficiency of 2-day fermentation liquid was higher than that of 1-day and 3-day fermentation liquids. When combined, the PAC and Ba3 broth produced a pronounced algae inhibition effect that effectively hindered the germination of algae cysts. We conclude that this combination provides a scientific reference for the prevention and control of marine red tide. Our results suggest that designing environmentally friendly methods for the management of harmful algae is quite feasible.

**Keywords:** *Bacillus nitratireducens*; fermentation broth; polyaluminum chloride coagulation (PAC); *Gymnodinium catenatum*; cysts

#### **1. Introduction**

At present, red tides have become one of the ocean's most catastrophic global disasters. The occurrence frequency, outbreak intensity, and impact range of red tides worldwide have been increasing and causing various degrees of harm to many countries' coastal regions. It is conservatively estimated that the annual loss of fishery and tourism caused by harmful algae blooms (HABs) in Europe is up to EUR 862 million, while the annual loss caused by HAB in the United States is up to USD 82 million. Between 1995 and 2004, an average of USD 1.31 million was lost annually in South Korea's fisheries due to harmful algal blooms. Red tide refers to the rapid proliferation or accumulation of various dinoflagellates and diatoms under external environmental conditions, mainly in marine environments such as coastlines or estuaries [1].The algal cell density reaches a certain level, causing water discoloration and affecting coastal areas and aquatic ecosystems, which, in turn, cause severe effects and hinder tourism [2]. The dinoflagellate *Gymnodinium catenatum* is a dominant harmful algae bloom (HAB)-forming species along coasts worldwide [3,4]. For example, *G. catenatum* is a bloom-forming species that forms HABs in China's Fujian

**Citation:** Balaji Prasath, B.; Wang, Y.; Su, Y.; Zheng, W.; Lin, H.; Yang, H. Coagulant Plus *Bacillus nitratireducens* Fermentation Broth Technique Provides a Rapid Algicidal Effect of Toxic Red Tide Dinoflagellate. *J. Mar. Sci. Eng.* **2021**, *9*, 395. https:// doi.org/10.3390/jmse9040395

Academic Editor: Bum Soo Park

Received: 12 March 2021 Accepted: 27 March 2021 Published: 8 April 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/).

coastal waters almost every year [5]. In 2017, a red tide of *G. catenatum* broke out in the coastal waters near Quanzhou and the Zhangzhou Sea. The continuous eruption of *Gymnodinium* species and cystic settlements formed by sexual reproduction in sediments poses a significant threat to the aquaculture industry and people's health. Therefore, dinoflagellate cyst deposits in coastal areas, the distribution patterns, and the abundance of species are an urgent need to control the harmful effects generated by HABs [6].

Different methods have been developed to prevent and eliminate red tides, including physical, chemical, and biological methods in recent decades. These three methods have their advantages and disadvantages in red tide management [7]. Among these methods, physical methods will not harm the original ecosystem, but they are also inefficient, expensive, not suitable for large-scale red tides, and can only suppress red tides quickly. Chemical methods of algae removal are very efficient and can destroy algae cells. Regrettably, these methods are quite costly and may lead to secondary pollution [8–10]. Biological methods, however, are economical, effective, and environmentally friendly. In particular, microbial algae suppression methods offer many advantages such as simple operation, complete algae killing, and no secondary pollution to the environment [11]. The management of current red tide hotspot areas is of great significance, and the prospect of controlling HAB algae-killing microorganisms has also been in rapid development. In particular, these algaecide bacteria can lyse algae by directly or indirectly attacking cells [12–15]. Certain substances secreted by microorganisms will cause cell lysis and death by invading and contacting algae cells, although the action time is longer.

Nowadays, related research has discovered that the combined algae removal effect of multiple methods is more effective than that of a single method. However, the inhibitory effect of combined methods on coastal HAB species is much less studied. In the present study, chemical algae removal can reduce algae cells' density for a temporary period during a red tide outbreak. Among these particular methods, the coagulation method has relatively high safety [16]. Polyaluminum chloride (PAC), aluminum sulfate [Al2(SO4)3], ferric sulfate [Fe2(SO4)3], and ferric chloride (FeCl3) can also increase HAB removal efficiency. Besides biological algae removal methods, algae suppression bacteria and fermentation broth algae suppression methods have become new research directions in recent years. Few studies have demonstrated that *Bacillus* sp. can suppress the growth of harmful algal bloom species [17]. However, the effect of the combined approach on micro-algae is little known. To the best of our knowledge, there have been no reports so far. In this study, we first compared four coagulants, namely PAC, Al2(SO4)3, Fe2(SO4)3, and FeCl3, against *G. catenatum*. The coagulant we selected has the best flocculation effect and the best algae inhibition effect; on this basis, we explored the algaecide effect of *Bacillus* fermentation broth (Ba3) and further studied the inhibitory effect of the combination of coagulant and Ba3 on *G. catenatum*. Furthermore, we used this method to inhibit the germination of algae cysts in the sediments. We planned a systematic study of the comprehensive effect of the coagulant Ba3 on the red tide of algae and its germination as controlled from the source red tide outbreak, which may have crucial effective management strategies. The framework of the study design is shown schematically in Figure S1.

#### **2. Material and Methods**

#### *2.1. Cultivation of the Dinoflagellate and Bacteria*

*Gymnodinium catenatum* was obtained from the State Key Laboratory of Marine Environmental Science at Xiamen University, China. We maintained the axenic algal culture at 20 ± 2 ◦C in a sterile L1 medium prepared with natural seawater filtered to 0.45 μm and maintained under a 12:12 h light/dark cycle. We also counted cell numbers under a microscope. The algal culture was transferred once a week to a fresh, sterilized medium, which ensured that experiments were always conducted with cultures during the exponential growth phase.

We previously identified the algicidal bacteria preserved in the College of Life Sciences, Fujian Normal University, as *Bacillus nitratireducens*. The strain was initially cultivated

with a modified Bacillus Medium (Peptone 10 g/L, sodium chloride 5 g/L, beef paste 5 g/L, pH value 7.2~7.4, dissolved in deionized water) in a rotary shaker (30 ◦C, 180 rpm), and the resultant mixture kept in our laboratory was thoroughly stirred for subsequent experiments.

#### *2.2. Cultivation of Gymnodinium catenatum Cysts*

Dinoflagellate has a high cyst formation rate under low nitrogen and phosphorus environments [18]. Usually, the limitation of nutrients in the water phase is an effective way to induce vegetative cells to form cysts. Therefore, an L1 medium with low phosphate and nitrate was used to prepare *G. catenatum* cysts. The dilution ratio of phosphate and nitrate was 1:15 (named L15). In brief, we added 20 mL of 10<sup>6</sup> cells/L of *G. catenatum* in a 50-mL centrifuge at 3000 rpm for 10 min, discarded the supernatant, and then slowly added 20 mL of L15 medium. After adding the medium, the algae cells in the centrifuge tube were mixed with the medium and then transferred to a sterile Erlenmeyer flask for cultivation.

#### *2.3. Selecting Coagulants and Preparing Concentration*

Four agents, including PAC, Al2(SO4)3, Fe2(SO4)3, and FeCl3, were used in the experiments. The coagulant concentration was set to 0, 10, 20, 30, 50, 70, and 90 mg/L; the first group (0 mg/L) was used as the control, and the following six were used as the treatment groups. Each treatment was established in triplicate.

#### *2.4. Cell Inhibition Efficiency*

#### 2.4.1. First Experiment

The efficiency of the cell removal experiment was tested in 25-mL sterile test tubes. A 20-mL aliquot of the algal culture was placed in each tube with a density of 2 × 106~ <sup>4</sup> × <sup>10</sup><sup>6</sup> cells/L; then, 0, 10, 20, 30, 50, 70, or 90 mg/L of the four agents' stock solution was evenly added to the tube. After adding the agent, each tube was mixed thoroughly, and each concentration was tested in triplicate on the alga cells. After three periods (3, 24, and 48 h), three replicate samples from each group were pipetted from the upper-middle region of the collected liquid surface, and each sample's concentration of algal cell was determined under the microscope; photographs were taken for the bottom flocs of each treatment. In addition, the supernatant's pH value was measured to determine the total removal rate of zeta potential, turbidity, and UV254; then the positive effects and the best dose of the coagulant were noted for further experiments.

#### 2.4.2. Second Experiment

*B. nitratireducens* bacteria were inoculated into the culture medium and grown to the stationary phase (30 ◦C at 180 rpm for 24 h). The extraction of bacterial fermentation broth (Ba3) was collected using centrifugation (10,000, 15 min) over three days (1, 2, and 3 d). The supernatants were filtrated through 0.22-μm Millipore membrane filters and then used. Ba3 volume ratios of 0.3, 0.7, 1.0, and 2.0% were added to the 30-mL exponential phase *G. catenatum* algae. Each group was placed in a light incubator for culturing, shaken twice a day, and sampled once every two days until the end of the experiment. The group with no Ba3 served as a control for the experiments, and all experiments were repeated in triplicate.

#### 2.4.3. Third Experiment

In the first and second experiments, a more effective dosage of mixed coagulant and Ba3 fermentation broth was selected. This experiment determined whether it was possible to improve the previous experiments results using combinations of more effective agents and Ba3. The above experiments' dosage of the various coagulants and the fermentation broth dosage for two days were 0.3, 0.7, 1.0, and 2.0%. The combined coagulants and fermentation broth were added to 20 mL *G. catenatum* alga in triplicate samples for each group for better cell inhibition efficiency. Each group was placed in a light incubator for cultivation, and samples were assessed at 3, 24, 48, and 96 h for density determination to ensure that they had significant inhibition efficiency.

#### *2.5. Calculation Method of Algae Removal Rate*

The algae removal rate was monitored by estimating cell numbers utilizing a microscope and calculated according to the following formula.

$$\text{RE (\%)} = (1 - \text{N}\_{\text{t}}/\text{N}\_{\text{0}}) \times 100\%$$

where RE is the removal rate of algal cells; N0: vegetative cell density before adding algae inhibitor; and Nt is the vegetative cell density after adding the algae inhibitor.

#### *2.6. Analytical Methods*

An optical microscope was used to observe the effect of fermentation broth on the morphological characteristics of algal cells. After the addition of algal inhibitor, 1 mL of the supernatant was removed, placed on a slide, and observed under a microscope. The size of the coagulant and the flocs as well as the compactness of the flocs were photographed for further characterization of the coagulation effect of different coagulants on the cells of dinoflagellates. UV254 was determined by using the spectrometer as an indicator to reflect the content of organic pollutants in water. The spectrometer reading does not represent a specific organic substance, but rather the total amount of multiple organic substances; therefore, the spectrometer reading refers to the many soluble fine particles and various inorganic compounds present in the experiment. In this experiment, after stirring, the supernatant was taken from 2 cm below the liquid level after standing for 24 h. The supernatant was filtered using a 0.45-μm acetate fiber membrane, and then, pure water was used as the reference solution to determine the results of each experiment group. A spectrophotometer analyzed turbidity with a wavelength of 660 nm. Zeta potential is one of the indexes that reflect the stability of suspended matter or colloid in water. The zeta potential was monitored using Zetasizer software. In this experiment, 2 mL of algal sedimentation floc was taken and injected into the sample pool with an injection needle, and the sample pool was put into the card slot to ensure that the algal sedimentation floc in the sample pool did not exceed the electrodes at both ends of the sample pool. Zetasizer software was used to measure and record the results. Dissolved inorganic phosphorus (DIP) and dissolved inorganic nitrogen (DIN) were determined using a spectrophotometer at a wavelength of 882 nm.

#### *2.7. Provenance Control Experiment*

This experiment was divided into two groups; one group was marine sediments— PAC + Ba3 + sediment (marked as I-1, I-2, and I-3)—and the other group was marine sediments + pure-breed *Gymnodinium* cysts 25 ± 2 cysts/g DW—PAC + Ba3 + sediments + pure-breed (marked as II-1, II-2, and II-3). The framework of the study design is shown schematically in Table S1. The dosages of the treatment used in the study were as follows: PAC 50 mg/L, and PAC + Ba3 50 mg/L + 0.3%.

#### *2.8. A Calculation Method of Algae Cell Abundance*

We calculated the algae cell abundance rate after germination in the sediment were calculated by visual observation method using a microscope and calculated according to the following formula.

$$\mathbf{N} = (40/30) \times (\mathbf{V}\_0/\mathbf{V}) \times \mathbf{n}$$

where N is the abundance of algal cells after germination, with the unit of cells/mL; V is the count volume taken in each observation, with the unit of ml; V0 is the volume of the sample after concentration, with the unit of mL; n is the number of algae in each sample.

#### *2.9. Data Processing*

The experimental data were analyzed using SPSS 22.0 software, and the difference between the data (*p* < 0.05) was significant (*p* < 0.01)**.**

#### **3. Results**

#### *3.1. The Effect of Different Coagulants on the Algae Inhibition*

The four kinds of agents have noticeable removal effects on *G. catenatum*, but different degrees of the "back-dissolving" phenomenon appeared over time. Figure 1 shows that when the reagent dosage was low (10 mg/L) at 3 h, the PAC, Fe2(SO4)3, Al2(SO4)3, and FeCl3 algae removal rates achieved were 80, 68.5, 77.5, and 80%, respectively. When the dosage was increased, the difference between PAC, Fe2(SO4)3, Al2(SO4)3, and FeCl3 in algae removal efficiency rate gradually narrowed, while that of the FeCl3 group was relatively low. The removal rate of *G. catenatum* in the PAC group increased when the dosage was increased. When the dosage was 50 mg/L, the removal rate reached more than 95.1%; when the dosage was 70 and 90 mg/L, the removal rate reached 100%.

**Figure 1.** Removal effect of four different coagulants on *G. catenatum.*

After 24 h of dosing, the removal rates of different dosing amounts were between 65.0 and 92.9%. When the dosage was 50 mg/L, the removal rate of *G. catenatum* decreased to 80.3%. After 48 h of dosing, the resolution phenomenon was more marked when the dosage was 10 mg/L, which reduced the removal rate to 52.9%. After 96 h of dosing, the removal effect of *G. catenatum* increased compared with that at 48 h, and the removal rate was 69.4% at 10 mg/L. However, when the dosage was greater than 50 mg/L, the removal rate of *G. catenatum* was still higher than 80% over time. After 3 h of dosing, the different dosages' removal rates differed significantly from those between 24 and 96 h (*p* < 0.05). When the dosage of Fe2(SO4)3 was higher than 20 mg/L, the removal rate was higher than 85%. After 48 h of administration, the removal rate decreased; after 96 h of administration, the removal rates of the 10, 20, and 30 mg/L groups increased, but when the dosage of ferric sulfate increased to 90 mg/L, the removal rate decreased to 69.4%. The removal rate was not altered between 24 and 96 h. When the Al2(SO4)3 dosage was 30 mg/L, the removal rate reached 85% or more; above this dosage, the removal rate did not change significantly (*p* > 0.05). When the dosage was 90 mg/L, the removal rate of *G. catenatum* between 24 and 96 h was slightly higher than that after 3 h. The FeCl3 group activity was relatively low with the increase in the dosage, and the removal rate did not increase significantly (*p* > 0.05); when the dosage was 90 mg/L, the removal rate was 90.2%. With

time, the removal rate decreased significantly at a low dosage (10~50 mg/L) (*p* < 0.05). When the dosage was 70 mg/L, the removal rate was stable (Figure 1). After adding different dosages, the removal rate did not change significantly between 24 and 96 h.

Figure 2A shows the four coagulants' influence the pH range, and the control group's pH value was 8.1. With the addition of coagulants, the pH values of the four groups showed a downward trend. FeCl3 had the smallest effect on pH at 90 mg/L, and the pH value was the lowest, which was 7.6. The pH range of the Al2(SO4)3 group was 8.1~6.8. The influence trend of the PAC and Fe2(SO4)3 groups affected *G. catenatum* in almost the same way, and the variation range was 8.1~6.8. The pH range between 6.8 and 8.1 exhibited a good algal removal effect at both a higher dosage of coagulant and a lower dosage of coagulant, a pH decline, and the efficiency was slightly less reactive (Figure 2A).

**Figure 2.** (**A**) Effect of coagulants on pH; (**B**) changes in zeta potential; (**C**) removal effects of coagulants on turbidity.

Zeta potential is one of the evaluation indices of the coagulation method's water treatment effect, showing the stability of colloids or suspended solids in a solution system. Coagulation and algae removal usually use the positively charged aluminum or iron hydrolyzed cations formed by the coagulant in water and the negative charge on the algae cells' surface to attract each other. Figure 2B shows the potential change in the surface of *G. catenatum* with different coagulants added. The potential of the four coagulants decreased first and then increased with the increase in the dosage. The lowest point of the potential appeared when the dosage was 20 mg/L. Among them, the rising trend of FeCl3 was the most obvious. When the dosage was 50 mg/L, the potential was the highest at −6.1 mV, but the fluctuation range was also higher. The Al2(SO4)3 group fluctuated slowly with the dosage increase, and the dosage potential at 20 mg/L was higher than other coagulant groups. The Fe2(SO4)3 group's potential decreased the most, and at a high dose (90 mg/L), the potential exceeded the other coagulant groups. In the PAC group, after the dosage was higher than 20 mg/L, and with the increase in the dosage, the surface potential of the algae cells increased rapidly. The highest point appeared at 90 mg/L, with a potential of −5.3 mV (Figure 2B). The turbidity of the system reflects the sedimentation effect of algae cells, and different dosages of different coagulants have different effects on the removal of turbidity.

Figure 2C illustrates the dosage at 10 mg/L. The turbidity of the four retardants showed significant differences, among which the turbidity of Fe2(SO4)3 was the lowest, followed by the turbidity of Al2(SO4)3, FeCl3, and PAC. That said, the removal rate was higher except for PAC; the other three coagulant groups' turbidity removal rates were all negative, thus increasing the system's turbidity. When the dosage was greater than 30 mg/L, the turbidity's changing trend was stable (Figure 2C). Comprehensive analysis of

the removal effect of the density, turbidity, and organic matter of *G. catenatum* when using the four coagulants PAC, FeCl3, Fe2(SO4)3, and Al2(SO4)3 showed that PAC had the most robust ability to remove algae on *G. catenatum* and had a low re-solubilization rate. Large flocs made of a large number of flocculated algal cells were formed, and PAC proved to have a good removal effect on their turbidity and organic matter.

An optical microscope was used to observe the settled flocs to observe coagulant and dinoflagellate coagulation effects. Figure 3a–d show the coagulation effects of the four groups of coagulants on *G. catenatum*. The micrograph suggests that the algae cells in the flocs, formed by the coagulation of aluminum sulfate and algae cells, were not tight enough, and there were few algae cells fixed in the flocs; however, the cell morphology of *G. catenatum* was also observed, and the cell shape was still active. The floc's micrograph that settled on the bottom after the FeCl3 group was added for 96 h (Figure 3b). It is also evident that ferric chloride has a good coagulation effect on *G. catenatum* because the flocs were large, and few algal cells were swimming outside the flocs. The algae cells in the flocs were many and dense, but the flocs were not tight. The swimming algae cells could easily break away from the flocs and return to the water body. The morphology of algae cells in the flocs was almost unchanged at 400×. A micrograph of the floc that settled to the bottom after the Fe2(SO4)3 group, which was added for 96 h, is shown in Figure 3c.

**Figure 3.** (**a**–**d**) Micrographs of flocs after sedimentation of (**a**) Al2 (SO4)3, (**b**) FeCl3, (**c**) Fe2 (SO4)3, and (**d**) polyaluminum chloride (PAC) group for 96 h.

Clearly, the flocs were small and scattered, and algae cells were swimming outside the flocs. As the floc is small, it distributes the algae cells on the edge of the floc. As the algae cells swim, the algal cells may break free from the flocs, causing a "back-dissolution" phenomenon. A micrograph of a floc that settled to the bottom after PAC administration for 96 h is shown in Figure 3d. At 40×, we could see that the flocs formed by the flocculation and *G. catenatum* cells were more extensive, and there were fewer algal cells outside the flocs. Under the 100× microscope, we observed that the algae cells in the flocs were relatively dense, there were few algae cells at the edges of the flocs, and the coagulation effect was better. PAC has a minor effect on the cell morphology of *G. catenatum,* and the four coagulants have different effects on the removal of algae. Although the surface potential of the algal cells increased after PAC addition, the fluctuations were larger.

#### *3.2. The Effect of Bacterial Fermentation Broth (Ba3) on the Algae Inhibition*

Ba3 fermented broth from 1, 2, and 3 d was added to an algal concentration of 1200~1500 cells/mL at four dosage levels of 0.3, 0.7, 1.0, and 2.0% (*v/v*). The experimental results are displayed in Figure 4A. With the increase in the action time, the concentration of *G. catenatum* continued to decrease. After the second day of dosing, the concentration of algal cells decreased to below 500 cells/mL. The removal rates of all groups were at least 68.1%, and the highest one was 92.9%. On days 2–6, the concentration was stable. After adding Ba3 bacteria fermentation broth, compared with the control group, all concentrations of fermentation broth had a more obvious removal effect. From the results of adding

Ba3 bacteria 1-d fermentation broth, we observed that with the increase in the addition, the concentration of *G. catenatum* showed a downward trend. When the addition was 2.0%, the cell number reached the lowest value. The removal rate was as high as 82.1%. Ba3 2-d bacteria fermentation broth and 3-d fermentation broth also increased the algaecide effect with the dosage. The fermentation broth usage at different fermentation times also has a certain impact on the algaecide effect. The figure clearly shows that the effect of Ba3 bacteria 2-d fermentation broth was higher than that of the 1-d fermentation broth and 3-d fermentation broth but was not affected by the dosage or influence of time. Figure 4B shows the effect of removing algae from fermentation broth on turbidity. We designed this experiment with two groups—with and without algae—to observe the turbidity changes by adding bacteria fermentation liquid. The turbidity of the algal and algal-free groups increased with the addition of the fermentation broth. Among them, the turbidity of each dosage of Ba3 bacteria 2-d fermentation broth was slightly lower than that of the 1-d and 3-d fermentation broths. Except for the control group, the algae-containing group's turbidity and the non-algae-containing group in each experimental group differed, showing that after the bacterial fermentation broth destroyed the algal cells, a large amount of content was released, leading to a significant increase in turbidity.

**Figure 4.** (**A**) Effect of fermentation broth on the growth of *G. catenatum;* (**B**) effect of algal removal from fermentation broth on turbidity; (**C**) effect of algal removal from fermentation broth on UV254.

The bacterial fermentation broth consisted of a variety of active products secreted by the bacterial body. Therefore, the various organic substances present in the bacterial fermentation broth not only increased the turbidity of the water body but also increased the UV254 value of the water body. The algae cells ruptured, and the organics inside the cells flowed out. In short, the fermentation broth can destroy the integrity of algal cells, but the action time is longer. After the algae cells dissolve, the body's organic matter will be released, and the bacterial fermentation broth will maintain certain turbidity. As a result, the water phase's turbidity and organic content will increase after processing the bacterial fermentation broth (Figure 4C).

Figure 5 shows the influence of algal cell morphology. The surface of the algae cells of *G. catenatum* was smooth and complete before the bacterial fermentation broth treatment, with prominent horizontal grooves and nuclei (Figure 5A). After 12 h of Ba3 treatment (Figure 5B), the upper shell of algal cells had become transparent, and the nucleus was visible. The cell wall showed damage to a certain extent, but the cell membrane was still intact and no content flowed out. Over 24 h (Figure 5C), the algae cells' surfaces were loose and ruptured, and the cell membrane was damaged. At 36 h (Figure 5D), the cell

membrane was revealed to be severely damaged. Many granular materials of different sizes appeared around the cells. It is possible that after the cell membrane of the cell wall is ruptured, the contents of the cell overflow, and the algal cell's morphological structure becomes relatively blurred. After 96 h of observation, *G. catenatum* had lost its morphology entirely, and the algal cells gradually decomposed and ruptured into small particles that were almost unrecognizable (Figure 5E,F). From the analysis, the algicidal effect of Ba3 on *G. catenatum* is triggered via indirect algaecide activity. The bacteria secrete some active substances for algae cells to lyse the algae cells, achieving the algae-killing effect.

**Figure 5.** Effect of fermentation broth on *G. catenatum* ((**A**)—0 h; (**B**)—12 h; (**C**)—24 h; (**D**)—36 h; (**E**) and (**F**)—96 h).

#### *3.3. The Effect of Combined Coagulant and Ba3 on the Algae Inhibition*

In this study, the coagulant and Ba3 fermentation broth were combined to eliminate algal cells in order to increase the removal effects on *G. catenatum*. As shown in Figure 6A, the removal effect of the four coagulants combined with four concentrations of Ba3 broth on *G. catenatum* occurred at different times. The PAC and Ba3 fermented broth group had the best algae removal effect. The removal rate reached 100%, and the number of algal cells in the overlying water did not increase over time. The effect of other coagulants and algae cells combined with the algae inhibition method was lower than that of the PAC and Ba3 fermented broth group. Figure 6B shows that the group of Al2(SO4)3, combined with the Ba3 fermented broth, caused the algae density to reach the highest value at 24 h; the removal rate of *G. catenatum* was 82.3%. Figure 6C shows that with the Fe2(SO4)3 and Ba3 broth group, the removal of algal cells reached the highest rate at 80.7%. Figure 6D shows the FeCl3 and Ba3 fermented broth group and the changing trend; all combinations reach the highest value at 24 h. Overall, the combination of PAC and Ba3 broth has the best algae inhibition effect, and when the volume ratio of bacterial fermentation broth is 0.3%, the removal rate can reach 100%. The combination method of algae suppression can improve algae suppression efficiency quickly and reduce the effect of algae cell re-dissolution. This method can be combined with the addition of the coagulant to act on the algae cells to form flocs and precipitate to the bottom; the bacterial fermentation broth also acts on the algae cells, which are gradually lysed under the stimulation of the active substance, and the exercise ability gradually decreases until the cells rupture.

**Figure 6.** Effects of coagulant and fermentation broth on growth of *G. catenatum*. (**A**) Combination of four coagulants and fermentation broth to inhibit algae; (**B**) Al2(SO4)3 + Ba3 broth; (**C**) Fe2(SO4)3 + Ba3 broth; and (**D**) FeCl3 + Ba3 broth.

Figure 7 shows the effect of the combination method of coagulant and Ba3 fermented broth on UV254. With the increase in Ba3 broth dosage, each group's organic removal effect showed a downward trend. In the PAC-combined Ba3 broth group, when the dosage of the bacterial fermentation broth was 0.3% (*v/v*), the maximum removal rate was 41.0%, and the lowest removal rate was 1.0%. This removal rate confirms that when the dosage of Ba3 broth is higher than 1.0%, the content of organic matter in water is higher than the flocculation effect of PAC. The removal rate of Al2(SO4)3-combined Ba3 broth group was negative when the dosage was higher than 0.3%, demonstrating that the removal effect of Al2(SO4)3 on the organic matter is lower than that of PAC. In the fermented Fe2(SO4)3 and FeCl3-combined Ba3 broth groups, the two groups' removal rates were negative for different bacterial fermentation broth dosages. Therefore, the effects of the four coagulants combined with different concentrations of Ba3 broth on the removal of organic matter followed the order of PAC + Ba3 fermented broth > Al2(SO4)3 + Ba3 broth > Fe2(SO4)3 + Ba3 broth > FeCl3 + Ba3 fermented broth.

**Figure 7.** Removal effects on UV254 by the combination of coagulants and fermentation broth.

Figure 8 shows the effect of the combination method of coagulant and Ba3 broth on turbidity. Each group's turbidity removal effect showed a downward trend with the increase in bacterial fermentation broth dosage. Each group had the highest removal rate when the bacterial fermentation broth was added at 0.3%. The removal rates of turbidity in the fermentation broth group of PAC combination Ba3 broth were between 46.3 and 89.1%; the removal rates of turbidity in the fermentation broth group of Al2(SO4)3 combination bacteria were between 55.8 and 83.7%; with Fe2(SO4)3, the turbidity removal rate of the combined Ba3 broth group was between 58.2 and 89.1%; the turbidity removal rates of the FeCl3-combined Ba3 broth group were between 57.8 and 88.4%.

**Figure 8.** Removal effects on turbidity by the combination of coagulants and fermentation broth.

#### *3.4. The Effect of Combined Coagulant and Ba3 on Cyst Germination Inhibition*

The sediment used for algae germination in the simulated sediment was from the Quanzhou section where the red tide of *G. catenatum* had occurred, and the total abundance of phytoplankton in the sediment was 8.04 × <sup>10</sup><sup>2</sup> cells/g, mainly diatoms and dinoflagellate cysts. The in situ experiment sediment samples were taken at 5, 10, and 15 d to observe the phytoplankton species and abundance in the water phase after germination (Figure 9). The sediment's overlying water was L1 medium with no algae, and the proportion of primary algae diatoms germinated in the three tests was higher. After five days of culture, the abundance of algae in the overlying water of the in situ sediment group (I-1) was approximately 20.3 cells/mL, in which the diatoms reached 17.2 cells/mL. Only an insignificant amount of dinoflagellates germinated. In the group (I-2) with the PAC group, the algae's germination rate was low; only diatoms emerged, and the inhibition rate reached over 70%. However, the phytoplankton abundance of the PAC group (I-2) was 32.5 cells/mL, indicating that mainly diatoms and dinoflagellates still existed. The inhibition rate of this group decreased to 57.2%. After 15 days of culture, the algal abundance of the in situ sediments group was slightly lower than that of the 10-d culture, and the inhibition rate was 61.3%. In the whole experiment period, no algal cell germination was observed in the group of PAC and Ba3 fermented broth (I-3).

In the in situ sediment with added *G. catenatum* cyst (II-1) group culture, diatoms were dominant in the overlying water and we detected only a few dinoflagellates. After 10 d of cultivation, the total abundance of germinated phytoplankton reached 90.5 cells/mL, and the density of *G. catenatum* was significantly higher than that in group I-1, which reached 18.0 cells/mL; the addition of PAC (II-2), compared with the II-1 group, has a particular inhibitory effect on algal cell germination, especially diatoms. After 15 days of cultivation, the algae density in the overlying water was slightly lower than that after 10 days of cultivation, which may be related to the content of nutrients in the water. With the increase in cultivation time, the water's nutrients were gradually consumed, which ultimately decreased the water's nutrients. The PAC-combined Ba3 fermented broth group (II-3) was only detected on the 10th day with an insignificant amount of diatoms appearing, with a removal rate of over 98% (Figure 9). The PAC had a better inhibitory effect on germination in the sediment; there was almost no algae germination. This lack of germination may be

because the active substances in the bacterial fermentation broth have a good destruction effect on the algae cells, the experimental water body has a weak exchange flow ability, and the concentration of the bacterial fermentation broth in the water body remains almost unchanged; therefore, the provenance in the sediment will likely be affected by the bacteria to a certain extent; the destruction of active substances will not occur.

**Figure 9.** Effect of different algal inhibition method on microcapsule germination in sediment (I) and In situ sediment added with *G. catenatum* cyst (II).

After the L1 medium was added to the sediment, the nitrogen and phosphorus elements in water were essential for the phytoplankton's growth. The initial DIP concentration was 0.13 mg/L, the DIN concentration was 5.333 mg/L, of which the nitrite concentration was 0.042 mg/L, and the ammonium salt concentration was 0.007 mg/L of nitrate. The concentration was 5.283 mg/L. Figure 10 displays the effects of different algae suppression methods on algae cyst germination in the in situ sediments. After five days of PAC administration and PAC-combined Ba3 fermented broth, they dissolved inorganic phosphorus in the water body. The removal effect of the PAC group (I-2) on DIP reached 86.4%. On the 10th day, the removal rate was 87.3%; with the extension of the cultivation time, the removal effect showed a tendency to decrease. On the 15th day, the DIP concentration was 0.011 mg/L and the removal rate dropped to 77.4%. The PAC-combined Ba3 group (I-3) also had a beneficial removal effect on DIP, and the removal effect was slightly higher than that of the PAC group; there was no reduction in removal rate on the 15th day (Figure 10A). Each control group's removal affected the in situ sediment + *G. catenatum* cyst (group II) germination group. Within 5-10 days of the PAC group's control, the removal rate was higher than 85% on the 5th day; on the 15th day, the DIP concentration was 0.009 mg/L, and the removal rate dropped to 81.3%. Compared with group I-2, the DIP content was lower in group II-2, which may consume DIP due to the higher abundance of algae. PAC's removal rate in the combined Ba3 group broth group (II-3) reached more than 87% (Figure 10B).

Figure 10C, D show the effects of different algae inhibitors on dissolved inorganic nitrogen in water and in the in situ sediment germination group. The dissolved inorganic nitrogen in both groups showed a downward trend with time. The cultivation did not transform in each group until the 5th day, but the I-3 group (PAC + Ba3 broth) inorganic nitrogen concentration is higher than that of group I-2 (PAC); when cultured to day 10, group I-3 shows a higher value than those of groups I-2 and I-1. On the 15th day of cultivation, the difference between the groups was more prominent. The dissolved inorganic nitrogen concentration in group I-3 was 2.30 mg/L, while the dissolved inorganic nitrogen concentration in group I-1 was 1.34 mg/L. The lower of the two groups was 0.79 mg/L. Compared with the initial inorganic nitrogen concentration, the decrease rate of inorganic nitrogen was 59.0~85.2% in the entire cultivation cycle under the control of PAC. Figure 10D shows, in the germination group of in situ sediments + *G. catenatum* cysts, that the three groups all

have the same tendency of change, and the concentration of dissolved inorganic nitrogen shows a tendency to decrease over time. The concentration of dissolved inorganic nitrogen in the three groups at different incubation times was in the order II-1 > II-3 > II-2 during the entire cultivation cycle. Regarding the decrease rate of group II-2 compared with the initial inorganic nitrogen concentration, the range was between 61.6 and 80.8%, and the removal rate of group II-3 was between 56.6 and 77.8%. Using PAC and PAC combined with Ba3 broth had no noticeable effect on removing dissolved inorganic nitrogen in the water. Compared with group II-2, the addition of bacterial fermentation broth in II-3 increased the content of dissolved inorganic nitrogen in the water body, which was mainly related to the nitrogen-containing compounds in the bacteria's fermentation product. The PAC-combined Ba3 broth group had a beneficial removal effect on the soluble inorganic phosphorus, and the removal rate reached over 87%. The active bacteria substance has nitrogen-containing compounds; therefore, this group's soluble inorganic nitrogen content is higher than that of PAC.

**Figure 10.** Effect of different algal inhibition methods on (**A**) dissolved inorganic phosphorus (DIP) and (**C**) dissolved inorganic nitrogen (DIN) in sediment; and effect of different algal inhibition methods on (**B**) DIP and (**D**) DIN in sediment (added *G. catenatum* cyst).

#### **4. Discussion**

Our experimental study results revealed that all four coagulant chemicals, PAC, Al2 (SO4)3, Fe2(SO4)3, and FeCl3, exhibited a high removal efficiency against *G. catenatum*, with some differences. However, PAC has the best removal effect, with a low re-solubilization rate, large floc formation, and the most considerable amount of flocculated algal cells. It also has a beneficial removal effect on turbidity and organic matter. Recent studies showed that the high removal efficiencies with five kinds of coagulants are comparable to the results of Liu Lijuan et al. for the control of a lake containing algae bloom [16]. They concluded that PAC has a more impressive algae removal effect by changing the dosage and pH conditions. The algae removal mechanism of polyaluminum salts showed that PAC and algae cells in the water first underwent adsorption and an electrical neutralization reaction, and a bridging network was trapped, so that the flocs were more extensive and, therefore, more prone to settle [19]. Moreover, a pilot test of enhanced coagulation of raw water in the Yangtze River with PAC showed that PAC had a beneficial coagulation effect on the Yangtze River water, forming large flocs and a rapid settling speed [20].

Previous studies have shown that specific bacterial populations can inhibit HAB species' growth through microbial algae suppression into direct algae suppression and indirect algae suppression [21,22]. In one of these studies, Ba3 (*Bacillus* fermentation broth) had a high algae inhibition effect, reaching over 90%. To date, several publications have proven that *Bacillus* sp. can inhibit the growth of harmful algal blooms [17,23]. Compared with other algaecide bacteria isolated from aquatic water, *Bacillus* showed similar or more

potent algicidal activity against algae [24]. Zhao et al. [25] isolated four algicidal metabolites from a fermentation broth of *Bacillus* B1 strain against *Phaeocystis globosa*, and all metabolites had a strong alga dissolving effect. In this study, a Ba3 *Bacillus* strain of fermentation broth was used to influence algal growth, and the results showed that the 2-d fermentation broth had the best removal effect on *G catenatum*. Under the microscope, we observed that the cell wall gradually loses its ability to move under the action of the Ba3 broth, the algae cells gradually become transparent, and the contents of the cell overflow due to rupture, though there is no complete algal cell morphology. In terms of their biological safety, anti-compounds are biodegradable. Moreover, when used to control blooms, they appear to be harmless to the environment [26]. These results strongly bolster the claim that Ba3 has potential applications in controlling algae outbreaks.

The coagulant removes algae quickly, and it can coagulate with algae cells to form flocs and settle to the bottom immediately after addition, but the coagulant algae suppression will have a certain back-dissolving phenomenon. The algae killing effect of the fermentation broth is thorough, but the action time is relatively long. However, previous research revealed that PAC combined with algae-lysing bacteria has a high inhibitory effect on *Microcytis aeruginosa* and nutrients in the water [27]. Furthermore, Wang et al. [28] studied the inhibition of the growth of *Scrippsiella trochoidea* by combining algae-suppressing bacteria with two modified clays. The results show that the combined algae suppression method can improve the algae suppression effect, increase the algae suppression time, and reduce algae cells' rebound effect. In this study, the combination of coagulants and Ba3 fermented broth was used to suppress algae and remove algae cells in a short time while killing algae cells. The results of the four coagulants combined with *Bacillus* Ba3 fermentation broth, when combined with algae inhibition, showed that the combination of PAC and Ba3 fermented broth had the best effect on inhibiting *G. catenatum* with an inhibition rate of 100% and no rebound. The bacterial fermentation broth is indeed yellow and contains many organic substances. An increased dosage of bacterial fermentation broth leads to increased water turbidity and organic matter content; when the Ba3 broth dosage in the PAC-combined bacterial fermentation broth group was 0.3%, it had the highest removal rate of turbidity and organic matter in water.

Although the red tide community's increase may be significant and alarming, this issue should not be linked to on-site production. Researchers have also proposed other HABs from resting cysts germinating from the bottom sediments, which seriously affect many aspects of the red tide phenomenon [29]. Cysts are also particularly efficacious for community spread. They enable species to survive under adverse conditions, and because their development generally involves sexuality, they promote genetic recombination [30]. In addition, the high abundance of cysts in the sediment may reflect the recent blooming of these species in the area. According to previous studies, many resting cysts will be produced by the end of the blooming cycle. The formation of cysts is one of the predominant factors in bloom termination [31,32]. Several researchers believe that the vegetative cell inoculum size from cyst germination is critical to the beginning of blooming. However, the final bloom size cannot measure the effect of altering cell growth or the aggregation of complex natural conditions because unpredictable variables control these effects. Therefore, it is necessary to control the sediment cyst before germination. Nevertheless, there have only been a few studies on the coagulant effect with combine bacterial fermented broth on cyst formation and germination to control HAB species.

We divided the germination experiment into two groups: in situ sediments and in situ sediments after adding *G. catenatum* cysts. In this study, treatment with the PAC-combined Ba3 fermented broth group significantly removed over 90% of the *G. catenatum* vegetative cells. Recent research has also found that nutrient changes caused by modifying clay can be re-released from the algae matrix into the water, which may contribute to the formation of resting cysts [33]. These results offer a contrast to this study where the PAC-combined Ba3 broth group had a good removal effect on the soluble inorganic phosphorus, and the removal rate reached over 87%. The active bacteria substance has nitrogen-containing compounds; therefore, the content of soluble inorganic nitrogen in this group was higher than that of the PAC group. These treatments not only were effective for removing the vegetative cells but also lowered the nutrients level.

#### **5. Conclusions**

In conclusion, our results show that the feasibility of using appropriate concentrations of PAC combined with Ba3 fermented broth is potentially useful for controlling *G. catenatum* blooms. The effect of bacterial fermentation broth on killing algae is complete, but the time needed for effective action is longer. The active substances in the fermentation broth act on the algae cells to lyse them. The four coagulants combined with the bacterial fermentation broth group had an obvious inhibitory effect on *G*. *catenatum*. Through the comprehensive analysis of the removal effects of algal cells, turbidity, and organic matter, it was found that after the action of the polyaluminum chloride (50 mg/L) combined with the bacterial fermentation broth (0.3%, *v/v*) for 3 h, the algal inhibition rate was 100%, and the algal cells did not rebound. The inhibition rate of the bacterial fermentation broth and polyaluminum chloride group was more than 98%. Our findings will integrate well into the future studies of controlling toxic dinoflagellate cysts in the benthic environment. Our research also shows that dinoflagellate organisms use nitrogen and phosphorus cues to determine when vegetative growth occurs. Understanding the environmental cues related to algae dormancy is essential for the understanding and management of HABs in aquatic ecosystems.

**Supplementary Materials:** The following are available online at https://www.mdpi.com/article/10 .3390/jmse9040395/s1, Figure S1—A schematic design of the study on the proliferation regulation of *Bacillus* sp. Fermentation broth combined with coagulant on *G. catenatum*, Table S1—A schematic design of the study on the control of cyst germination of *Gymnodinium catenum* by the combination of coagulant and fermentation broth

**Author Contributions:** Conceptualization, Y.S. and B.B.P.; methodology, B.B.P. and W.Z.; validation, Y.S.; formal analysis, B.B.P. and W.Z.; investigation, Y.S.; data curation, B.B.P., H.L. and W.Z; writing—original draft preparation, B.B.P. and Y.W.; writing—review and editing, Y.S., B.B.P. and H.Y.; supervision, Y.S.; project administration, Y.S.; funding acquisition, Y.S. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by the National Key Research & Development Plan "Strategic International Scientific and Technological Innovation Cooperation" project (2016YFE0202100), the National Natural Science Foundation of China (41573075), and Minjiang Scholar Program.

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

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** The data presented in this study are available on request from the corresponding author. The data are not publicly available.

**Acknowledgments:** The authors are also indebted to the anonymous reviewers for their constructive comments and suggestions for the improvement of this manuscript.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **Identification of New Sub-Fossil Diatoms Flora in the Sediments of Suncheonman Bay, Korea**

**Mirye Park 1, Sang Deuk Lee 2,\*, Hoil Lee 3, Jin-Young Lee 4, Daeryul Kwon <sup>1</sup> and Jeong-Min Choi <sup>5</sup>**


Nakdonggang National Institute of Biological Resources (NNIBR), 137, Donam 2-gil, Sangju-si 37182, Korea <sup>3</sup> Center for Active Tectonics, Geology Division, Korea Institute of Geoscience and Mineral Resources (KIGAM), 124, Gwahak-ro, Yuseong-gu, Daejeon 34132, Korea; hoillee@kigam.re.kr


**Abstract:** Suncheonman Bay, Korea's most representative estuary, is an invasive coastal wetland composed of 22.6 km2 of tidal flats surrounded by the Yeosu and Goheung Peninsulas. In January 2006, this region was registered in the Ramsar Convention list in Korea, representing the first registered wetland. Estuaries are generally known to have high species diversity. In particular, several studies have been conducted on planktonic and epipelic diatoms as primary producers. Suncheonman Bay has already been involved in many biological and geochemical studies, but fossil diatoms have not been evaluated. Therefore, we investigated fossil diatoms in Suncheonman Bay and introduced sub-fossil diatoms recorded in Korea. One sedimentary core has been extracted in 2018. We identified 87 diatom taxa from 52 genera in the SCW03 core sample. Of these, six species represent new records in Korea: *Cymatonitzschia marina*, *Fallacia hodgeana*, *Navicula mannii*, *Metascolioneis tumida*, *Surirella recedens*, and *Thalassionema synedriforme*. These six newly recorded diatom species were examined by light microscopy and scanning electron microscopy. The ecological habitats for all the investigated taxa are presented.

**Keywords:** sub-fossil diatom; sediment; Suncheonman Bay; new record

#### **1. Introduction**

An estuary can be defined as a semi-enclosed coastal body of water that has a free connection with the open ocean, within which seawater is diluted with freshwater derived from land drainage [1,2]. River mouths, coastal bays, tidal marsh systems, and sounds all fit this definition. Estuaries are transitional zones between freshwater and marine habitats. Due to tides and storms, the water level and salinity vary in estuaries [3]. They are most commonly located in low-relief coastal regions. Estuaries and wetlands are among the most productive aquatic ecosystems, providing a home for both freshwater and marine plants, and a source of nutrients for a variety of animal communities adapted to brackish waters [4,5]. Moreover, they filter out pollutants supplied to the ocean [4,6,7]. Thus, many animals rely on estuaries that have abundant species diversity for food, places to breed, and migration stopovers [8,9] (https://oceanservice.noaa.gov/facts/estuary.html (accessed on 8 February 2021)).

In South Korea, where three sides are surrounded by the sea, there are numerous estuaries such as the Nakdonggang, Keumgang, and Seomjingang. Coastal wetlands cover approximately 2800 km2, which represents approximately 3% of the total land area [10–12]. Suncheonman Bay, Korea's most representative estuary, is an invasive coastal wetland

**Citation:** Park, M.; Lee, S.D.; Lee, H.; Lee, J.-Y.; Kwon, D.; Choi, J.-M. Identification of New Sub-Fossil Diatoms Flora in the Sediments of Suncheonman Bay, Korea. *J. Mar. Sci. Eng.* **2021**, *9*, 591. https://doi.org/10.3390/jmse9060591

Academic Editor: Azizur Rahman

Received: 1 May 2021 Accepted: 22 May 2021 Published: 29 May 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/).

composed of 21.6 km2 of tidal flats and 5.4 km<sup>2</sup> of reed fields surrounded by the Yeosu and Goheung Peninsulas [10]. In January 2006, it became the first registered wetland in the Ramsar Convention list in Korea, as it was designated as a "wetland protected area" by the Ministry of Land, Transport and Maritime Affairs in December 2003 [10,13]. In June 2008, Suncheonman Bay was designated as national cultural property, "Myeongseung" number 41, and in 2010, south-western tidal flats in Korea, including Suncheonman Bay (south-western coast tidal flats), were included in the UNESCO World Heritage Tentative List. Such registrations and designations demonstrate the recognition of its ecological and environmental value [14] (https://whc.unesco.org/en/tentativelists/5482 (accessed on 20 April 2021)). In particular, the natural environment and ecosystems within Suncheonman Bay are well preserved, making them a habitat favorable for many species of marine organisms [15]. In these regions, primary producers such as diatoms play a critical role as a food source for large invertebrates and fishes [16].

Diatoms are unicellular algae characterized by a biomineralized (opaline) cell wall that may fossilize and be preserved in the sedimentary record [17,18]. The sub-fossil diatoms herein described consist of Holocene diatom remains not fully involved in the fossilization process. Diatoms thrive in very different environments (e.g., hot springs, polar regions, and fresh, brackish, and marine waters) and are extremely sensitive to physical and chemical changes (e.g., temperature, salinity, and nutrients) in the water [18–23]. Therefore, fossil diatoms represent an excellent source of information about past climate change and its effect on aquatic ecosystems. There have been relatively few studies on sub-fossil diatoms along the southern Korean coast. Marine to brackish sub-fossil diatom assemblages were initially studied in the Pohang and Gampo sediments in 1975 [24], then they were extended to the regions of Bukpyeong [25] and Pohang [26,27] in the East Sea and the regions of the Mankyung-Dongin river estuary [28], Dodaecheon River [29], Ilsan estuary [30], Chollipo [31], Isanpo [32], and Sabsi-do and Kunsan in the Yellow Sea [33].

The Suncheonman Bay has been used to study various environmental characteristics, including grain size and organic matter in tidal flat sediments [10], seasonal water quality, pollution, environmental safety [13,34], and other geochemical characteristics, as well as local inhabitants such as benthic invertebrates, plants, fishes, birds, bacteria, and fungi [6,14,35–39]. Among the studies conducted to date, the investigation of phytoplankton communities in the Dong Cheon River and Isa Cheon River stream into Suncheonman bay was the most interesting [40]. In this study, we describe a newly recorded sub-fossil diatom assemblage recovered in the sediments of the Suncheonman Bay.

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

#### *2.1. Coring and Sampling of Sediment*

Drilling was carried out using a peat core sampler (52 mm diameter; Peat Sampler, Eijkelkamp Soil & Water, Giesbeek, Netherlands). One sediment core (= SCW03 with a length of 6.0 m) was retrieved from Suncheonman Bay in Korea on 11 June 2018 (Figure 1, Table 1). The core was transported to the laboratory after vacuum packing in a plastic bag to prevent drying and oxidation. Sedimentological description and subsampling were performed after the sediment core profile was cut vertically in half [41]. Shell fragments in the SCW03 core were selected for the analysis of chronology and diatoms because the sediment layer was well preserved.

**Figure 1.** Sampling sites in Suncheonman Bay.

**Table 1.** Information of sampling sites.


#### *2.2. Analysis of Chronology*

Age dating was performed using an accelerator mass spectrometer (AMS) at the Korean Institute of Geoscience and Mineral Resources (KIGAM), Korea. The estimated ages were calibrated by the OxCal statistical analysis program (http://c14.arch.ox.ac.uk (accessed on 20 May 2021)).

#### *2.3. Sample Preparation for Diatom Identification*

Thirteen samples of diatoms were collected every 0.5 m along the SCW03 core. Their analysis was conducted according to the following steps: 1 g of sediment was dried at 60 ◦C for 24 h; the siliceous material was boiled with 20 mL of 30% hydrogen peroxide (H2O2) and washed with distilled water to remove organic matter; the treated samples were mounted with Pleurax (Mountmedia, Wako, Japan) and briefly heated using an alcohol lamp for subsequent analysis using a light microscope (LM; Eclipse Ni, Nikon, Tokyo, Japan). Photomicrographs were taken using a digital camera (DS-Ri2, Nikon, Tokyo, Japan). Some remaining peroxide-cleaned samples were filtered using 2.0-μm polycarbonate membrane filters (Nuclepore, Whatman, Maidstone, UK). The membranes were placed on stubs and coated with gold-palladium for analysis using a field emission scanning electron microscope (FE-SEM; MIRA 3, TESCAN, Brno-Kohoutovice, Czech Republic). SEM photomicrographs of all the samples were used to identify the diatoms. Morphological analyses of diatoms were performed using ImageJ v1.32 software (NIS-Elements BR4.50.00, Nikon, Tokyo, Japan) [42]. Taxonomical nomenclature was based on recent taxonomic information guidelines [43].

#### **3. Results**

#### *3.1. Sedimentary Facies Analysis*

The core SCW03 mostly consists of greenish-grey silty clay and can be divided distinguished into three sedimentary facies according to color, fossils content, and sedimentary structure (Figure 2A,C). Facies A is characterized by yellowish-brown mottling structures. Shell fragments are not observed in this facies. In Facies B, the yellowish-brown mottling

structures are less abundant and shell fragments are observed to occur sporadically. The size of shell fragments are about 2 mm in diameter. Facies C is represented by highly concentrated shells and shell fragments in several-centimeters-thick intervals.

**Figure 2.** (**A**) Stratigraphic section, (**B**) result of age dating, and (**C**) photographs of core SCW03.

The sediments of core SCW03 are interpreted as deposited in a tidal flat [44,45]. It is very similar to the current tidal flat environment. Abundant mottling structures in facies A originate from an oxygen-rich environment, indicating that the sea level has gradually decreased slightly in facies B. In the meantime, highly concentrated shells and shell fragments of facies C are interpreted as sedimentation caused by flooding or storm events [45].

#### *3.2. Age Dating*

The results of age dating for five samples in the core SCW03 shows a range from 1170 to 1560 cal. yr BP (Table 2 and Figure 2B).


**Table 2.** Results of AMS 14C dating and calibrated dates for core SCW03.

#### *3.3. Diatom Assemblages*

A total of 87 diatom species belonging to 52 different genera were identified in the sediments from Suncheonman Bay in Korea (Table 3); of these, six species, namely, *Cymatonitzschia marina*, *Fallacia hodgeana*, *Navicula mannii*, *Metascolioneis tumida*, *Surirella recedens*, and *Thalassionema synedriforme,* were never described before in this area. The diatom flora

encountered in this survey was composed of 87 taxa, which were classified into 52 genera. We present information on the valve shape, occurrence depth in sediment, and habitat of 87 species, including the six newly recorded species. Information collected about the newly recorded diatoms identified here included taxonomic information, illustrations, basionyms, synonyms, original description references, depth in the core, distribution, and diagnosis (Table 3).

**Table 3.** Occurrence (black squares) and habitat of diatom species by depth. A total of 72 diatom species belonging to 52 different genera were identified. Star marks on the specific name are newly recorded species in Korea (6 species: *Cymatonitzschia marina*, *Fallacia hodgeana*, *Navicula mannii*, *Metascolioneis tumida*, *Surirella recedens*, and *Thalassionema synedriforme*).



**Table 3.** *Cont.*


**Table 3.** *Cont.*

*Cymatonitzschia marina* (F.W.Lewis) Simonsen 1974 (Figure 3A,B)

Basionym: *Cymatopleura marina* F.W.Lewis 1861 [107]

Synonym: *Cymatopleura marina* F.W.Lewis 1861 [107]

Original description: Simonsen 1974: 56, pl. 41: Figures 5–9 [108]

Description: Valves are observed to be solitary, usually lying in the valve view. Valves are linear lanceolate, with very acute ends. Valves are strictly isopolar, and not constricted in the middle. Overall dimensions include a valve length ranges from 58.42 to 67.94 μm and valve width from 9.14 to 11.28 μm. The valve faces have numerous undulations (9–11), with a distance between two undulations in the ranges from 4.81 to 7.97 μm. Undulations are found to have a nearly trapezoidal shape (Figure 3A). The valve surface has irregular punctate on the undulations. A raphe system is observed running around one side of the valve margin. Striae uniseriate are found to be densely spaced, with approximately 28–29 per 10 μm observed, on one side of the valve margin (Figure 3B).

Depth occurrence in the core: 2.0 m.

Distribution: This species is reported from brackish water to marine environments mainly [61–66]. This taxon is reported from some estuary, e.g., East River, New York and Long Island Sound [109]. *Cymatopleura marina* was first recorded from the Indian Ocean [108].

Differential diagnosis: This genus differs from *Cymatopleura*. The genus *Cymatopleura*, as a member of the Surirellaceae, has a completely different raphe morphology, which runs along the edge of the valve around the entire margin, whereas in *Cymatonitzschia* it is, as in *Nitzschia*, limited to one of the sides [108]

Remarks of raphe: *Cymatonitzschia marina* has an eccentric keeled raphe placed through the edge of the valve, and it appears on one of the sides [108,110].

*Fallacia hodgeana* (R.M.Patrick and Freese) Y.H.Li and H.Suzuki 2014 (Figure 3C–F)

Basionym: *Navicula hodgeana* R.M.Patrick and Freese 1961 [111]

Synonym: *Navicula hodgeana* R.M.Patrick and Freese 1961 [111]

Original description: Li et al., 2014 in p. 33 [74]

Description: Valves are observed to be solitary, usually lying in the valve view. Valves are naviculoid and linear-elliptic, with bluntly rounded ends. Overall dimensions include length ranges from 14.73 to 15.42 μm and width ranges from 4.21 to 4.61 μm. The valve face is nearly flat with a slightly curved raphe (Figure 3C,D, arrow). Central raphe endings are proximately hooked (Figure 3E,F, arrowheads). Terminal raphe fissures exhibit a sickleshaped curve in the same direction. Some parts of the striae are covered with a thin siliceous covering, or conopeum, on the external valve surface. Tow slits opening of the canal, present near the terminal raphe fissures, are also observed (Figure 3F, arrowhead). Areolae are found to be curved upward and were directly connected to the mantle. The

finely porous conopeum extends outward from the outer edge of the raphe sterna, running through the surface, and connect to the proximal edge of the mantle. Numerous peg-shaped structures are observed in the nonporous margin of the conopeum along the proximal edge of the mantle. The elongated areolae, with an approximate length and width of 0.28–0.42 μm and 0.15–0.18 μm, respectively, are found on the hyaline area of the valve surface with an undulated junction. The elongated areolae were found in groupings of 12 per 5 μm. Peg-shaped structures, in groups of 12 per 5 μm, are also observed, finely porous on the conopeum 12–13 per 1 μm transversely.

**Figure 3.** Light microscope (**A**,**G**,**H**) and field emission scanning electron microscopic (**B**–**F**,**I**,**J**) photomicrographs of diatoms: (**A**,**B**) *Cymatonitzschia marina* (the arrow points to raphe), (**C**–**F**) *Fallacia hodgeana* (in **C** the arrow points to raphe; in **F** the arrow points to the terminal openings), (**G**–**J**) *Navicula mannii* with raphe and ribbon-shaped areola (the arrow points to raphe and ribbon-shaped areola). Scale bar = 10 μm.

Depth occurrence in the core: 6.0 m.

Distribution: This species lives in fresh- to brackish-water environments. It was first reported from scrapings of small rock submerged at the edge of a lagoon as *Navicula hodgeana* [111]. Li et al. 2014 collected this species from Edogawa River, Japan. This taxon is known to benthic diatom [74,112,113].

Differential diagnosis: *Fallacia hodgeana* possesses morphological features such as a single H-shaped plastid, depressed lateral sterna interrupting striae that contain round areolae enclosed by hymen; well-developed, finely porous conopeum; and a canal system between the primary silica layer and the conopeum. These features indicate that this species does not belong to the genus *Navicula* or *Pseudofallacia* [74]. This species is related to *Navicula dissipata*. The length-to-breadth ratio is similar, although *N. dissipata* is a larger taxon. The clear central area is narrower in our taxon, and the striae are composed of many fine puncta instead of a few large ones. The median ends of the raphe are close together as in *N. dissipata* [111].

Remarks of raphe: *Fallacia hodgeana* has a slightly curved raphe that terminates at fissures curved in the same way. The distal end fissures were sickle-like and curved in the same direction (Figure 3C arrow). The central endings lie close to each other, and slightly curved slits seem to be promoted from the general valve [74].

*Navicula mannii* Hagelstein 1939 (Figure 3G–J)

Synonym: *Navicula elegantissima* Meister 1935 [114]

Original description: Hagelstein 1939, p. 388, pl. 7, Figures 7 and 8 [115]

Description: Valves are observed to be solitary. Valves are broadly lanceolate, and abruptly constricted toward the ends (Figure 3G–I). Overall dimensions include average length and width ranges from 28.28 to 30.09 μm and from 8.57 to 9.44 μm, respectively. The axial area is narrow, and becomes gradually wider, larger, and rounded toward the central area (Figure 3I). Raphe is observed to be very slightly curved filiform style with a very thickened and hyaline sternum (Figure 3I,J, arrow). Striae are very coarse and of low density (9–11 in 10 μm); they are observed to strongly radiate in the middle and then become parallel towards the ends. The central area striae alternate between longer and shorter forms (longer striae 4 areolae, and shorter striae 2 areolae). Areolae are observed to be ribbon-shaped, and approximately 5–6 are found in 2 μm sections (Figure 3J, arrow).

Depth occurrence in the core: 2.7 cm.

Distribution: *Navicula mannii* was reported in brackish water or marine environments [70,116,117]. Navarro (1983) reported the taxon in tropical temperate waters from the southwestern coast of Puerto Rico [116]. This species was known to neritic, pantropical, and cosmopolitan [116]. Ohtsuka (2005) collected the species from a muddy tidal flat in the Ariake Sea in south-western Japan [117].

Differential diagnosis: Hagelstein (1939) described that the *Navicula mannii* have minutely punctate areolae, but we found the ribbon-shaped areolae on the striae based on SEM observation in this study [115]. Ultrastructural studies of this species are rarely performed using a SEM; this study represents the first example of this approach.

Remarks of raphe: *Navicula mannii* has a straight raphe. Proximal raphe ends have an expanded pore-like shape and bent distal raphe ends (Figure 3G,H,J, arrows). Normally, *Navicula* spp. have a straight raphe system, unlike the raphe shapes of *Navicula cryptocephala* and *Navicula gregaria*, which commonly occur in Korea. *N. mannii* and *N. cryptocephala* both have drop-like internal ends, but *N. mannii* has a more pore-like end than *N. cryptocephala* [83,118], with a T-shaped structure. In contrast, *Navicula gregaria* has a different raphe shape than *N. mannii*, which is bent in the same direction as the raphe and exhibits asymmetrical thickening, beside the proximal raphe ends and beside the raphe rib [118].

*Metascolioneis tumida* (Brébisson ex Kützing) Blanco and Wetzel 2016 (Figure 4A,B)

**Figure 4.** Field emission scanning electron microscopic photomicrographs of diatoms. (**A**,**B**) *Metascolioneis tumida* (the arrow points to raphe), (**C**,**D**) *Surirella recedens* (the arrow points to raphe), (**E**–**I**) *Thalassionema synedriforme*. Scale bar = 10 μm.

> Basionym: *Navicula tumida* Brébisson ex Kützing 1849 [119] Synonym: *Navicula tumida* Brébisson ex Kützing 1849 [119] *Scoliopleura tumida* (Brébisson ex Kützing) Rabenhorst 1864 [120] *Microstigma tumida* (Brébisson) Meister 1919 [121] *Scoliotropis tumida* (Brébisson ex Kützing) R.M.Patrick and Freese 1961 [111] *Scolioneis tumida* (Brébisson ex Kützing) D.G.Mann 1990 [122]

*Navicula jenneri* W.Smith 1853 [84]

Scoliopleura jenneri Grunow 1860 [123]

Original description: Blanco, S. and Wetzel, C.E., 2016, pp. 195–205 [124].

Description: Valves are found to be solitary with one-layered valves. Valves are linear lanceolate with bluntly rounded apices (Figure 4A). Overall dimensions include an average length and width ranges from 41.72 to 52.5 μm and from 6.05 to 6.50 μm, respectively. Cells are usually found in a girdle view and twisted about the apical axis (Figure 4B). The valve mantle is relatively deep, and its face curved moderately into mantles. Striae uniseriate (16–17 in 10 μm) with small poroids (12–15 μm in 5 μm) are observed. The raphe system is twisted and sigmoidal in shape (Figure 4A, arrow). The raphe sternum is generally narrow and slightly wider in thcenterre. The raphe is found to be straight with simple raphe endings and straight terminal fissures extending to the valve margin. The girdle consists of several open bands. The band closest to the valve bears one transverse row of poroids 31–32 at 10 μm.

Depth occurrence in the core: 1.0, 1.5, 2.0 m.

Distribution: This species has been found in marine habitats. Stoermer et al. (1999) presented this taxon in a checklist of diatoms as *Navicula tumida* from the marine environment in the Laurentian Great Lakes [125]. Vilicic et al. (2002) reported the species as *Scoliopleura tumida* from the eastern Adriatic Sea [93]. This species was listed to the British marine diatoms as *Scoliopleura tumida* [126,127]. Méléder et al. (2007) reported the taxon as *Scolioneis tumida* in a sediment of mudflat from Bourgneuf Bay, France [128].

Differential diagnosis: Formerly included in *Scolipleura* taxon, but lacking the offset central raphe endings and longitudinal canals of that genus. Distinguishable from *Scoliotropis* by having fewer plastids, which lie against the valves rather than the girdle, by the simple uniseriate striae and raphe structure [122].

Remarks of raphe: *Metascolioneis tumida* (syn. *Scolioneis tumida*, *Navicula tumida*) has a slightly twisted raphe and raphe sternum that is normally narrow and becomes expanded in the center (Figure 4A, arrow). External central raphe endings have straight fissures along the valve margin. Furthermore, internal central endings are T-shaped and elongated [122]. *Surirella recedens* A.W.F.Schmidt 1875 (Figure 4C,D)

Homotypic synonym: *Surirella fastuosa* var. *recedens* (A.W. F. Schmidt) Cleve 1878 [129] *Surirella fastuosa* var. *typica* f. *recedens* (A. W. F. Schmidt) Deby 1897 [130]

Original description: Schmidt and Fricke 1875, pls 17–20. [131]

Description: Valves are found to be solitary, strongly silicified, and lying in the valve or girdle view. Valves are heteropolar with a broadly rounded headpole and cuneate footpole. Overall dimensions include a length and width of 35.08 μm and 22.83 μm, respectively. The valve surface has four costae of 10 μm in length. The valve margin includes fibulae, siliceous braces, 7–8 fibulae, 10 μm. The raphe system runs around the entire valve margin and is located within a canal (Figure 4C,D; arrows). The canal is raised above the valve's surface. One or more potulae are located between the two fibulae. Four or more fibulae are located between the two costae.

Occurring depth in Core: 2.7, 5.5 m.

Distribution: This taxon was known to marine species [95,96]. López-Fuerte and Siqueiros-Beltrones (2016) reported the species as a benthic diatom from coastal waters in Mexico [132] and the Nanaura mudflat in Ariake Sea, Japan [117]. However, *Surirella recedens* was found in brackish waters from Cochin Backwater south in the Indian Ocean [133].

Differential diagnosis: *S. recedens* is composed of a heteropolar valve, whereas *Surirella fastuosa* has an isopolar valve [134]. *S. recedens* is smaller overall, in comparison to *S. fastuosa*, and more lanceolate in shape. In addition, *S. fastuosa* has more noticeable apices than *S. recedens* [135]. Goldman (1990) identified the *Surirella* cf. *fastuosa* based on the outline, length, infundibula, and circular pattern of the valve [136].

Remarks of raphe: *Surirella recedens* (Syn. *Surirella fastuosa* var. *recedens*, *Surella fastuosa* var. *typical* f. *recedens*) *Surirella* sp. has a raphe that is located along the margin of the valve (Figure 4D, arrow). A raphe positioned within a canal might be elevated above the valve surface in several species [137]. *S. recedens* shows a representative *Surirellaceae* raphe system, positioned along the margin of the valve (Figure 4D, arrow).

*Thalassionema synedriforme* (Greville) G.R.Hasle 1999 (Figure 4E–I)

Basionym: *Asterionella synedriformis* Greville 1865 [138]

Synonym: *Asterionella synedriformis* Greville 1865 [138]

*Thalassionema javanicum* Grunow Hasle in Hasle and Syvertsen 1996 [139]

Depth occurrence in the core: Hasle, G.R. 1999, pp. 54–59, 23 figures [140].

Description: The valves are heteropolar spatulate, linear, long, slightly wider in the middle, and constricted towards the head pole, rather than towards the foot pole (Figure 4E). Valve width ranges from 2.66 to 2.87 μm in the middle part of the valve and from 4.03 to 4.40 μm in the middle part of the footpole. An apical spine is located in the head pole, not the foot-pole part of the valve. The valve face is flat, with a wide sternum and slight undulation (5–6 in 5 μm) at the foot pole (Figure 4F). Areolae are placed within the valve face and valve mantle. Areolae are heart-shaped (5–6 in 5 μm) near the foot pole part, but similar to Y- or flower-shaped occlusions (5–6 in 5 μm) toward the middle part of the valve (Figure 4F–I). Labiate processes are placed at each pole of the valve. An opening of the labiate process places the external apex in the foot pole [141].

Depth occurrence in the core: 3.0 m.

Distribution: *Thalassionema synedriforme* was known to marine species. Hasle (2001) mentioned the species is restricted in tropical and subtropical waters [98]. This species was recorded for the first time from Argentinean coastal waters [142].

Differential diagnosis: *Asterionella synedriformis* Greville is the basionym of *Thalassionema synedriforme* [98]. Frenguelli (1941) mentioned that he found valves with 9–10 areolae within 10 μm and illustrated a linear, fragmentary specimen that, according to the valve outline and areola density, might also be attributed to *Thalassionema frauenfeldii*, but not to *T. synedriforme* [143]. Additionally, in the case of *Thalassiothrix javanica* (Grunow), Hustedt and Frenguelli illustrated a specimen that was slightly heteropolar with 6–7 areolae within 10 μm, which differs in areolae density and valve outline from *Thalassionema synedriforme* (12–16 areolae in 10 μm, according to Hasle 2001) [98].

#### **4. Discussion**

#### *4.1. Diatoms in SCW03*

In this study, the analysis of sub-fossil diatoms among core samples obtained from Suncheon Bay, Korea, was carried out. Within the SCW03 core sample, a total of 52 genera and 87 species of sub-fossil diatoms were identified, with locations ranging from the surface to within 6 m of the basement, and among them, six species of newly recorded sub-fossil diatom never recorded before were found. At a depth of 4.0 m, the maximum variety of diatom samples was observed, with 23 genera 34 species, while depths of 0.5 m and 3.5 m revealed the lowest variety of diatoms, with 12 genera 14 species and 10 genera 14 species, respectively (Table 3). The highest and lowest taxonomic richness occur at depths of 4.0 m (23 genera, 34 species) and 3.5 m (10 genera, 14 species), respectively (Table 3).

Among the diatoms observed, *Amphora* sp., *Auliscus sculptus,* and *Fragilaria capucina* were found only at 0.1 m depths; therefore, they are hypothesized to have only recently entered the Suncheonman Bay (Table 3). *Cymatosira lorenziana*, *Fallacia hodgeana*, *Pleurosigma* sp., *Semiorbis* sp., and *Trachyneis aspera* are no longer observed since they appeared to only occur at 6.0 m depth. It is hypothesized that this is due to climate, environmental, or topographical changes in the Suncheonman Bay habitat. Intriguingly, *C. lorenziana* is mainly found in warm waters, whereas *T. aspera* is mainly distributed in the Antarctic area, with almost opposite habitat characteristics; however, the causative environmental changes in Suncheonman Bay could not be identified in this study [144–146]. Nevertheless, identifying past environmental changes in Suncheon Bay is an important source of information for the prediction of future environmental changes, and should be investigated further.

Most of the marine and brackish species occur at a depth of 5.0 to 6.0 m of the SCW03 core sample, and the emergence of freshwater species gradually increased within depth

ranges of 3.0 to 4.5 m. Interestingly, only marine and brackish species appeared at the 2.5 m section (Table 3). Marine and some freshwater species appeared within occur the range of 1.0 m to 2.0 m depths, while only marine and brackish species appeared at 0.5 m depth; finally, freshwater species increased again at 0.1 m depth. These changes in the flora of diatoms assemblage composition observed suggest that the sediments recovered from SCW03 were deposited in a marine area that was scarcely influenced by freshwater inflow in the past (3.0–4.5 m depth; about 1260–1830 yr BP) and experienced a renewed influx of fresh water in recent times (0.1 m depth; about 1340 ± 20 yr BP) (Tables 2 and 3, Figure 2) [147,148]. More accurate results and solid interpretations of the triggers responsible for such environmental changes require further studies of comprehensive environmental change, including a quantitative analysis of diatoms and a chronological analysis of core samples.

#### *4.2. New Recorded Taxa from Korea*

In this study, we identified six species newly recorded in Suncheonman Bay area: Cymatonitzschia marina, Fallacia Hodgeana, Navicula mannii, Metascolioneis tumida, Surirella recedens, and Thalassionema synedriforme. N. mannii was first identified by light microscopy in 2005; however, in this study, we observed the ultrastructure of N. mannii and discovered a ribbon-shaped areolae by Fe-SEM [117]. The ribbon-shaped areolae of N. mannii are the first to ever be recorded.

Unrecorded sub-fossil diatoms in Suncheonman Bay were discovered at 1.0 m depths. This study is therefore meaningful because if we study the modern composition of diatoms in Suncheonman Bay or Korea, we would estimate six previously unrecorded diatoms were extinct in this area.

**Author Contributions:** Data curation, formal analysis, writing—original draft, and writing—review and editing, M.P., D.K., and S.D.L.; funding acquisition, S.D.L.; field investigation, J.-Y.L. and J.-M.C.; writing—review and editing, M.P., S.D.L., H.L., J.-Y.L., D.K., and J.-M.C. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by grants from the Nakdonggang National Institute of Biological Resources (NNIBR) (NNIBR202101108) projects.

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

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Acknowledgments:** Thanks to Seung Won Nam, Suk Min Yun, and Pyo Yun Cho for helping with the coring of sediment.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **First Report of the Marine Benthic Dinoflagellate** *Bysmatrum subsalsum* **from Korean Tidal Pools**

**Joon Sang Park 1,†, Zhun Li 2,†, Hyun Jung Kim 1,3, Ki Hyun Kim 2, Kyun Woo Lee 4,\*, Joo Yeon Youn 1, Kyeong Yoon Kwak <sup>1</sup> and Hyeon Ho Shin 1,\***


**Abstract:** Dense patches were observed in the tidal pools of the southern area of Korea. To clarify the causative organisms, the cells were collected and their morphological features were examined using light and scanning electron microscopy (SEM). In addition, after establishing strains for the cells the molecular phylogeny was inferred with concatenated small subunit (SSU) and large subunit (LSU) rRNA sequences. The cells were characterized by a nucleus in the hypotheca, strong reticulations in thecal plates, the separation of plates 2a and 3a, the tear-shaped apical pore complex, an elongated rectangular 1a plate and the absence of the right sulcal list. The thecal plate formula was Po, X, 4- , 3a, 7", 6c, 4S, 5---, 2----. Based on these morphological features, the cells were identified as *Bysmatrum subsalsum.* In the culture, the spherical cysts of *B. subsalsum* without thecal plates were observed. Molecular phylogeny revealed two ribotypes of *B. subsalsum* are identified; The Korean isolates were nested within the ribotype B consisting of the isolates from China, Malaysia and the French Atlantic, whereas the ribotype A includes only the isolates from the Mediterranean Sea. In the phylogeny, *B. subsalsum* and *B. austrafrum* were grouped. This can be supported by the morphological similarity between the two species, indicating that the two species may be conspecific, however *B. subsalsum* may distinguish from *B. austrafrum,* because of differences in the types of eyespots reported in previous studies. These findings support the idea that there is cryptic diversity within *B. subsalsum.*

**Keywords:** *Bysmatrum*; cyst; eyespot; morphology; ribotype

#### **1. Introduction**

The genus *Bysmatrum* M.A. Faust and K.A. Steidinger was erected to separate three benthic *Scrippsiella* species, *Scrippsiella arenicola* T. Horiguchi and R.N. Pienaar, *S. subsalsa* (Ostenfeld) K.A. Steidinger and Balech and *S. caponii* T. Horiguchi and R.N. Pienaar [1]. These species share a number of morphological characterizations of thecal plates: a lack of contact between 2a and 3a, the presence of six cingular plates, and the posterior sulcal plate that does not touch the cingulum [1]. Based on such morphological characteristics, six *Bysmatrum* species (including *Bysmatrum subsalsum* (Ostenfeld) M.A. Faust and K.A. Steidinger as the type species have been reported so far [1–6] and, within the genus, the *Bysmatrum* species) can be distinguished from each other by the differences in cell size and shape, plate ornamentation, cingulum displacement, the morphology of the apical pore complex (APC), nucleus position, habitat and habitus [1,4,5,7,8]. However, molecular data for *Bysmatrum* species are as yet available for only five species; *B. arenicola* T. Horiguchi and R. N. Pienaar, *B. austrafrum* Dawut, Sym, Suda and Horiguchi, *B. granulosum* L. Ten–Hage,

**Citation:** Park, J.S.; Li, Z.; Kim, H.J.; Kim, K.H.; Lee, K.W.; Youn, J.Y.; Kwak, K.Y.; Shin, H.H. First Report of the Marine Benthic Dinoflagellate *Bysmatrum subsalsum* from Korean Tidal Pools. *J. Mar. Sci. Eng.* **2021**, *9*, 649. https://doi.org/10.3390/ jmse9060649

Academic Editor: Carmela Caroppo

Received: 17 May 2021 Accepted: 8 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/).

J.P. Quod, J. Turquet and Couté, *B. gregarium* (E. H. Lombard and B. Capon) T. Horiguchi and *Hoppenrath* and *B. subsalsum* [6–9].

Since the morphological descriptions of *Bysmatrum* species by Faust and Steidinger [1], Anglès et al. [7] documented the morphological details and molecular phylogeny of *B. subsalsum* collected from different locations of the Mediterranean Basin, and concluded that the strains of *B. subsalsum* are morphologically indistinguishable but genetically distinct, probably due to the differences in habitat, physiology or life-history traits. Since then, Luo et al. [8] documented that *B. subsalsum* strains from Malaysia and the French Atlantic formed a subclade (ribotype B) that can be distinguished from a clade (ribotype A) consisting of stains from the Mediterranean Sea and that there is cryptic diversity within *B. subsalsum,* based on the genetic distance shown in ITS sequences among *Bysmatrum* species [7]. This indicates that additional molecular data (including morphological descriptions) are needed to clarify the cryptic diversity of *B. subsalsum,* with strains established from water bodies of various geographical regions.

*Bysmatrum subsalsum* is a cosmopolitan species [10], and its occurrences have been recorded in the Aral Sea, the Mediterranean Sea, the Caribbean Sea, the Gulf of Mexico, the French and Portuguese Atlantic coast and the East and South China Sea [7,8]. However, in the Korean coastal area, this species has not been found, and only *B. gregarium* (as *B. caponii*) isolated from plankton samples has been recorded, with its morphology and molecular information [9]. In the tidal pools of the southern area of Korea, we observed dense patches caused by unidentified organisms. These were isolated, and strains were established and the morphological features were examined using light and scanning electron microscopy (SEM). The observations revealed that the species was identical to *B. subsalsum*. In this study, we describe the morphological details of Korean strains of *B. subsalsum* and report on their molecular characterization, based on concatenated small subunit (SSU) and large subunit (LSU) rRNA gene sequences.

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

#### *2.1. Sampling and Culture*

In June 2017, dense patches of phytoplankton were observed in the tidal pools of Geoje Island (34◦59- 34.3" N, 128◦41- 44.1" E) and Jeju Island (33◦29- 26.5" N, 126◦25- 12.8" E), Korea. Water temperature and salinity were 23.0 ◦C and 34 psu in the tidal pool of Geoje Island, and 10.5 ◦C and 24.5 psu in the tidal pool of Jeju Island, respectively. The water samples from the pools were collected using the 50 mL conical tubes and were then transported to the laboratory for observation of the causative organism. In the laboratory, the single mass colonies from the samples were isolated using a glass micropipette on an inverted microscope (Eclipse Ts2R, Nikon, Tokyo, Japan), and transferred into a 24-well tissue plate containing f/2 medium adjusted to a salinity of 32. The isolated colonies were maintained at 24 ◦C in a 12:12 LD cycle under a photon irradiance of 100 <sup>μ</sup>mol photons·m−2·s−1. After sufficient growth, the cells were transferred to a culture flask. A monoclonal culture was successfully established from the cells collected in the tidal pool of Geoje Island and deposited in the Library of Marine Samples, Korea Institute of Ocean Science and Technology, as strain number LIMS-PS-2685 (=MABIK PD00002006). The isolate from the tidal pool of Jeju Island was also established as a strain; however, it is currently unavailable because of unexpected cell death.

#### *2.2. Morphological Observation*

Live cells were isolated and photographed at 1000X magnification using an AxioCam 512c digital camera (Carl Zeiss, Göttingen, Germany) on an upright microscope (Zeiss Axio Imager2, Carl Zeiss, Göttingen, Germany). For fluorescence microscopy, approximately 1 mL of cell culture was transferred to a 1.5-mL microcentrifuge tube, and 4- ,6-diamidino-2-phenylindole (DAPI) stain (Sigma-Aldrich, St. Louis, MO, USA) was added at a final concentration of 10 μg mL−1. Cells were then incubated in the dark at room temperature

for 1 h, and viewed and photographed through a Zeiss Filterset 49 (emission: BP 365–445; beamsplitter: FT 395).

For scanning electron microscopy, a 20 mL aliquot of a dense culture was fixed in glutaraldehyde and paraformaldehyde with a final concentration of 2% (*w/v*). The aliquot containing fixed cells was filtered through a polycarbonate membrane filter (5 μm pore size), without applying additional pressure and rinsed three times with distilled water to remove the salt. After rinsing, the sample was dehydrated in an ethanol series (10, 30, 50, 70, 90 and 100% ethanol, followed by two 100% ethanol steps) and dried using a critical point dryer (BAL-TEC, CPD 300, Balzers, Germany). Finally, the sample was coated with gold-palladium and examined using a field emission-scanning electron microscope (JSM-7610F, Jeol, Japan).

#### *2.3. DNA Extraction and Sequencing*

The mass colonies were harvested via centrifugation for 10 min at room temperature, and the supernatant was discarded. Genomic DNA was extracted from cell pellets using a DNeasy Plant Mini Kit (Qiagen, Hilden, Germany), according to the manufacturer's protocol. The SSU rRNA gene was amplified using primers (ATF01: 5- -YAC CTG GTT GAT CCT GCC AGT AG-3 and ATR01: 5- -RMW TGA TCC TTC YGC AGG TTC ACC-3- ) [11], and the D1–D3 regions of the LSU rRNA gene were amplified using previously described primers (D1R: 5- -ACC CGC TGA ATT TAA GCA TA-3- [12] and 28r691: 5- -CTT GGT CCG TGT TTC AAG AC-3- ) [11]. PCR reactions were conducted in a volume of 50 μL; 2.0 μL gDNA; 5 μL 10x buffer; 0.2 mM dNTP; 0.1 μM each primer; 0.25 U Taq polymerase (Takara Ex Taq, Takara, Seoul, Korea); and ddH2O to a final volume of 50 μL. The PCR condition was as follows: 94 ◦C for 4 min, 30 cycles of 94 ◦C for 1 min, 54 ◦C for 1 min, and 72 ◦C for 1 min and then 1 extension cycle at 72 ◦C for 10 min. PCR products were purified with QIAqucik PCR purification kit (Qiagen, Hilden, Germany). All rRNA gene sequencing was performed using an ABI PRISM 3700 DNA Analyzer (Applied Biosystems, Foster City, CA, USA). The sequences were trimmed and assembled into contigs using Geneious Prime (https://www.geneious.com (accessed on 2 January 2020)).

#### *2.4. Phylogenetic Analysis*

The sequences were aligned using MAFFT in Geneious plugin with default settings, and some regions manually adjusted using the Geneious Prime. The phylogenetic analysis includes all the published sequences of *Bysmatrum* species. A total of 50 sequences of the order Peridinales were selected, while the family Peridiniaceae and Protoperidiniaceae were used as outgroups. The sequence alignments of SSU and LSU rRNA were concatenated with introduced gaps. The phylogenetic tree for the concatenated sequence alignment was inferred using maximum likelihood (ML) analyses via RAxML version 8 [13], and using Bayesian inference (BI) through MrBayes version 3.2 [14]. The general time reversible (GTR) model with parameters accounting for γ-distributed rate variation across sites (G) was used in all analyses, taking into account 6-class gamma. The GTR+G substitution model was selected using the Akaike information criterion (AIC) as implemented in jModelTest version 2.1.4 [15]. Bootstrap analyses for ML were carried out with 1000 replicates to evaluate statistical reliability. The Markov chain Monte Carlo (MCMC) method for BI was used with four runs for 10 million generations, sampling every 100 generations. The first 10% of trees were deleted as burn-in, and a majority rule consensus tree was constructed to examine the posterior probabilities of each clade. The final trees were visualized with FigTree v1.4.4 (http://tree.bio.ed.ac.uk/software/figtree/ (accessed on 5 January 2020)).

#### **3. Results**

#### *3.1. Morphology of Vegetative Cell and Resting Cyst of Bysmatrum subsalsum*

Cells had a conical epitheca and a round hypotheca, were slightly obliquely dorsoventrally flattened and the longitudinal flagella were visible (Figure 1a–f). They were 37–52 μm in length and 31–44 μm in width (*n* = 50). An apical stalk consisting of a transparent gelatinous matrix was observed on the apex of the cells (Figure 1b,c). A yellowish eyespot was located on the right side of the sulcus (Figure 1d,e). Rod-shaped chloroplasts were observed in the periphery of the cell (Figure 1g,h). The DAPI-stained nucleus was large, commonly elongated (Figure 1h) and sometimes horseshoe-shaped by the different observation angle (Figure 1i). The position of a nucleus was posteriorly in the hypotheca (Figure 1g–i).

**Figure 1.** Light micrographs of *Bysmatrum subsalsum* (Strain: LIMS-2685). (**a**) Surface focus of ventral view showing the longitudinal flagella (arrow), sulcus and cingulum. (**b**) Deeper focus of ventral view showing the outline of the cell and apical stalk (arrow). (**c**) Surface focus of ventral-apical view showing the apical stalk (arrow). (**d**) Deeper focus of ventral-lateral view showing a red eyespot (arrow). (**e**) Surface focus of ventral-right lateral view showing the longitudinal flagella (arrow). (**f**) Deeper focus of dorsal view. (**g**) Epifluorescence image of ventral view of DAPI-stained cell showing the position of the nucleus. (**h**) Epifluorescence image of dorsal view of DAPI-stained cell showing the position of the nucleus. (**i**) Epifluorescence image of antapical view of DAPI-stained cell showing the shape of the nucleus. Scale bars = 5 μm.

SEM observation revealed that the cells have a plate formula of Po, Cp, X, 4- , 3a, 7", 6C, 4S, 5---, 2---- (Figure 2). The thecal plates were covered with strong reticulations (Figure 2a–e). Apical pore complex (APC) was tear-shaped and included a Po plate, a round cover plate (Cp) and a canal plate (X) with thick margins formed by the raised borders of the apical plates (Figure 2f). The first apical plate (1- ) was pentagonal and asymmetrical with shorter anterior sutures than the posterior ones, and surrounded five plates: 2- , 4- , 1--, 7- and Sa (Figure 2a,d,e). Three intercalary plates (1a, 2a and 3a) were similar in size, and 1a and 2a contacted each other; however, 3a was separated from 1a and 2a. (Figure 2b,c,e). Plate 1a was elongate and rectangular, whereas plates 2a and 3a were hexagonal and pentagonal, respectively (Figure 2b,c,e). The precingular plates were symmetrically distributed (Figure 2e). The first precingular plate (1--) was pentagonal and smaller than the others (Figure 2a,e). The cingulum was deeply excavated and descended by about one cingulum width (Figure 2a–d). Six cingulum plates were observed; plates C1, C2 and C3 were much smaller than the C4, C5 and C6 plates (Figure 2a–d). In post cingular series, plates 1---, 2-- and 4-- were tetragonal, whereas plates 3-- and 5-- are pentagonal in shape (Figure 2g). The plate 1-- was much smaller than the other plate (Figure 2a–c,g). Two antapical plates (1--- and 2----) have pentagonal shapes, and plate 2--- is larger than the 1---- plate (Figure 2a–c,g). The sulcus was wide and did not contact the antapex (Figure 2a,g), and consisted of four major plates with inconspicuous lists: the anterior sulcal plate (Sa) is narrow and elongated; the left sulcal plate (Sl) in narrow and right sulcal (Sr) plate are triangular (Figure 2a,h); the posterior sulcal plate (Sp) is the largest sulcal plate, wider than long, and does not contact the plate (Figure 2g,h). There were internal sulcal lists (Isl) emerging from left side of the Sr plate (Figure 2h). The left sulcal list (Lsl) that had emerged from the lower side of plate 1--was visible (Figure 2h).

**Figure 2.** Scanning electron micrographs of vegetative cells of *Bysmatrum subsalsum* strain LIMS-2685 from Korea. (**a**) Ventral view. (**b**) Dorsal view. (**c**) Dorsal-left lateral view. (**d**) Ventral-right lateral view. (**e**) Apical view, showing a centrally located raised dome (apical stalk: AS) and epithecal plate pattern. (**f**) Detail of apical pore complex showing the apical pore (AP), cover plate (Cp) and canal plate (X). (**g**) Antapical view, showing hypothecal plate pattern. (**h**) Detail of the sulcal plates showing anterior sulcal plate (Sa), right sulcal (Sr) and left sulcal (Sl) plates, posterior sulcal plate (Sp), left (Lsl; red arrows) and internal (Isl; blue arrows) sulcal lists. Scale bars = 10 μm (**a**–**e**,**g**); 5 μm (**f**,**h**).

Spherical cysts were observed under culture conditions (Figure 3). The cysts were 29.5–34.6 μm in diameter (*n* = 25). The cyst was grayish in color, and an orange accumulation body was visible (Figure 3a–c). The cyst wall was smooth, without any distinguishing features on the surface (Figure 3d).

**Figure 3.** Light and scanning electron micrographs of cysts of *Bysmatrum subsalsum* strain LIMS-2685. (**a**) Light micrographs of spherical cyst. (**b**,**c**) Light micrographs of cysts showing large, orange, accumulation body (arrows). (**d**) Scanning electron micrograph showing a smooth organic wall without any ornamentation. Scale bars = 5 μm.

#### *3.2. Molecular Phylogeny*

Bayesian inference (BI) and maximum likelihood (ML) based on the concatenated SSU and LSU rRNA gene sequences yielded similar phylogenetic trees. The genus *Bysmatrum* was a monophyletic group, with maximum support (Figure 4). *Bysmatrum arenicola* was the base of *Bysmatrum* species, and *B. subsalsum* was branched from *B. gregarium,* with moderate supports (BT/PP = 68/0.92). Two ribotypes (ribotype A and B) of *B. subsalsum* were identified; ribotype A was comprised only of strains from the Mediterranean Sea with maximum support (BT/PP = 100/1.00), whereas ribotype B included the strains from China, Malaysia, the French Atlantic and Korea. *Bysmatrum austrafrum* (HG321) was nested between the two ribotypes. In ribotype B, the Korean strains of *B. subsalsum* were closely related to the Malaysian strain (TBBYS03) (BT/PP = 80/1.00).

**Figure 4.** Phylogeny of *Bysmatrum subsalsum* inferred from concatenated SSU and partial LSU rRNA gene sequences using maximum-likelihood (ML). Ribotypes are labeled according to designations by Luo et al. [8]. Numbers on branches are statistical support values to clusters on the right of them (left: ML bootstrap support (BT) values; right: Bayesian posterior probabilities (PP)). Bootstrap support values > 50% and Bayesian posterior probabilities > 0.7 are shown. Branch lengths are drawn to scale, with the scale bar indicating the number of nucleotide substitutions per site.

#### **4. Discussion**

#### *4.1. Morphological Comparisons of Korean Isolates of Bysmatrum subsalsum with Other Isolates of B. subsalsum, and B. austrafrum*

According to Anglès et al. [7] and Luo et al. [8], the genetic sequences of *B. subsalsum* show large intraspecific differences, clustering two well-differentiated clades (ribotype A and B). The two clades of *B. subsalsum* based on SSU and LSU sequences were also shown in this study, and the Korean isolates of *B. subsalsum* were nested in the ribotype B and clustered with the isolates from China, Malaysia and the French Atlantic (Figure 4). The Korean isolates of *B. subsalsum* were morphologically characterized by the separation of plates 2a and 3a, the tear-shaped APC, an elongated rectangular 1a plate and the general morphology, such as cell size, plate ornamentation and nucleus position, which coincides with that of *B. subsalsum* in previous studies (Luo et al. [8] and reference therein). Luo et al. [8] reported the differences in the number of sulcal lists among strains of *B. subsalsum* in two clades; the French strain has the right sulcal list (Three sucal lists), whereas the Malaysian strains are characterized by the absence of the right sulcal list (Two sulcal lists). In Korean strains, the right sulcal list was not present. In addition, the number of sulcal lists varied among specimens collected from other geographical regions [16–19]. Anglès et al. [7] concluded that despite a certain degree of morphological variation (such

as cell size, APC morphology and size, and cingulum displacement), cells from the two clades of *B. subsalsum* exhibit similar morphological characteristics. Luo et al. [8] also confirmed the morphological similarities in the Malaysian and French strains of *B. subsalsum*. Consequently, the prominent morphological characteristics for clarifying the two clades of *B. subsalsum* still remain unclear.

Recently, Luo et al. [20] recorded *B. austrafrum* between two clades of *B. subsalsum*, with strong support in the phylogeny, which is in agreement with our result. *Bysmatrum austrafrum* was first described by Dawut et al. [6]. This species has a typical plate of Po, X, 4- , 3a, 7", 6C, 4S, 5---, 2--- for the genus *Bysmatrum* and is morphologically characterized by dimensions of 25–45 μm long and 20–42.5 μm wide, arranged reticulation in the thecal plates, the possession of an equatorially positioned cingulum and a cingulum displaced by a distance exceeding its own width. Based on these morphological features, Dawut et al. [6] reported that *B. austrafrum* is similar to *B. subsalsum*, and concluded that *B. austrafrum* can be distinguished from *B. subsalsum* by differences in cell size, the shape of the APC and apical plate 1- . However, cell sizes recorded in *B. austrafrum* were recorded in the other specimens of *B. subsalsum* (see Table 2 in Luo et al. [8]), and the shapes of the APC and apical plate 1 of *B. austrafrum* quite resembled those of Korean isolates and the specimens of *B. subsalsum* recorded by Luo et al. [8]. In addition, although Dawut et al. [6] did not consider the absence or presence of the right sulcal list, *B. austrafrum* does not have the right sulcal list in their description. Consequently, it is quite difficult to distinguish *B. subsalsum* from *B. austrafrum*, based on their morphological features. Nevertheless, different types of eyespots have been reported in *B. subsalsum* and *B. austrafrum* [6,8]. Based on the types of eyespots suggested by Moestrup and Daugbjerg [21], *B. subsalsum* has the Type B eyespot, whereas *B. austrafrum* presents the Type A eyespot. This may be a useful characteristic for distinguishing *B. subsalsum* from *B. austrafrum*. However, as the Type A eyespot has not been reported in other *Bysmatrum* species, such as *B. granulosum* and *B. gregarium*, more isolates of *B. austrafrum* need to be examined for clarifying the type of eyespot within *Bysmatrum* species.

#### *4.2. Morphology of Bysmatrum subsalsum Cyst*

Cyst morphology can be helpful to understand the diversity within the genus (e.g., Li et al. [22]). In *Bysmatrum* species, a cyst–theca stage relationship has only been established through germination experiments for Mediterranean *B. subsalsum* [7]. Two types of cysts of *B. subsalsum* have been described in culture and sediments; in the culture spherical to ovoidal cysts without any ornamentations were observed [7,19], whereas cysts with the typical plate pattern, which are morphologically similar to vegetative cells of *B. subsalsum*, were identified in sediments [8,23]. In our study, the spherical cysts were also observed from the culture. *Bysmatrum subsalsum* from the Mediterranean Sea (ribotype A) could produce both cyst types in culture [7,19], and from the French strain of *B. subsalsum* (ribotype B), only cysts with the thecal plate were described [8]. In previous studies, differences in cyst types have been recorded between cultures and natural sediments. For example, unarmored dinoflagellate *Margalefidinium polykrikoides* (formerly *Cochlodinium polykrikoides*) produced two different types of cyst (a spherical cyst without ornamentation in culture and an ornamented cyst in sediments) [24,25]. Tang and Gobler [24] suggested that the ornaments and spines of *M. polykrikoides* might be caused by biotic or chemical processes in sediments. However, this does not seem to be in accordance with *B. subsalsum*, because two different types of cysts were reported in sediments.

Similar morphological features between cyst and vegetative cell of unarmored dinoflagellate *M. polykrikoides* have been reported [26,27], and the cyst was identified as a temporary cyst that can be the short-term stage. The temporary cyst of *M. polykrikoides* is surrounded by a transparent and thin hyaline membrane, indicating in sediments the unarmored temporary cysts may be destroyed because of geochemical processes related to organic matter degradation or the attack of viruses or bacteria into the cell. However, temporary cysts with thecal plates (armored cyst) (such as cyst of *B. subsalsum*) may be

protected from the environmental conditions in sediments. If so, the cyst with typical thecal plates of *B. subsalsum* from sediments or culture may be temporary, because the resting cysts of dinoflagellate usually have distinct morphological features that can be distinguished from their vegetative cell (e.g., Matsuoka and Fukuyo [28]).

#### *4.3. Phylogenetic Position of Bysmatrum subsalsum*

In the phylogenetic tree, the Korean isolates within *B. subsalsum* were nested in the ribotype B consisting of the isolates from China, Malaysia and the French Atlantic, whereas the ribotype A includes only the isolates from the Mediterranean Sea. Iwataki et al. [29] documented that in the phylogenetic tree for *M. polykrikoides* isolates, the ribotypes can be useful for characterizing the geographical distribution pattern of *M. polykrikoides*. In our study, the ribotype A also represents the isolates of *B. subsalsum* from the Mediterranean Sea, which is in agreement with the conclusion by Iwataki et al. [29]. However, the ribotype B is a mixture of isolates originated from different geographic regions, although it mostly includes the Asian isolates. This indicates that the cells originating from France might be transferred from the Asian areas. Benthic dinoflagellates such as *Bysmatrum* species usually have a restricted distribution, possibly because of benthic, epiphytic behavior. Nevertheless, *B. susalsum* has been reported from various samples, including plankton samples, and their resting cysts are also present in the sediments (Luo et al. [8] and reference therein). This suggests that the vegetative cells and resting cysts of *B. subsalsum* can artificially be introduced into other coastal areas, possibly caused by ballast ship waters (e.g., Hallegraeff [30]).

Since the discovery of two ribotypes of *B. subsalsum*, Luo et al. [20] identified *B. austrafrum* between two clades of *B. subsalsum* in a phylogenetic tree based on concatenated SSU and LSU rDNA sequences, which is in agreement with our result. In the phylogenetic tree, *B. subsalsum* and *B. austrafrum* were grouped, with strong support (BT/PP = 95/1.00) (Figure 4), indicating that the two species may be conspecific. This result can be supported by morphological similarity between the two species. However, as there are differences in the types of eyespots between the two species, *B. subsalsum* can also be a species that distinguished from *B. austrafrum.* If so, it is possible that *B. subsalsum* is not monophyletic. This finding supports the idea that there is cryptic diversity within *B. subsalsum* [7].

#### *4.4. Environmental Conditions in Relation to the Growth of Bysmtrum subsalsum*

Although *Bysmatrum* species are considered benthic dinoflagellates, the occurrence of *B. subsalsum* has been reported in the plankton, with floating detritus and on sand and macroalgae [18]. Luo et al. [8] recorded *B. subsalsum* in plankton and sediment samples, and Anglès et al. [7] found *B. subsalsum* in mangrove detritus and salt marshes. In Korean isolates of *B. subsalsum*, the species is considered to inhabit tidal pools, and *B. austrafrum* was also discovered in tidal pools in South Africa [6]. In comparison with *B. subsalsum*, other *Bysmatrum* species, such as *B. teres, B. granulosum, B. arenicola* and *B. gregarium* have been reported in the restricted habitats (Luo et al. [8] and reference therein). This indicates that *B. subsalsum* in contrast to other *Bysmatrum*, may have broad environmental tolerance and a worldwide distribution.

Environmental conditions in relation to the growth of *B. subsalsum* have been rarely reported. Anglès et al. [7] documented that, based on the occurrence of *B. subsalsum* reported in previous studies, the species has a wide salinity tolerance and is able to grow under a wide range of sea water temperatures, with a preference for salinities > 30 and temperatures > 20 ◦C. López-Flores et al. [31] recorded that the cell abundances of *B. subsalsum* (as *Scrippsiella subsalsa*) decreased sharply in October, when salinity reached values > 30 and water temperature was low (15.6 ◦C). In this study, however, the dense patch of *B. subsalsum* was observed at 10.5 ◦C, indicating that it has a wider temperature tolerance than previously known. Further studies are also needed to clarifying the morphology (vegetative cell and cyst) of the genus *Bysmatrum* and, in particular, of *B. subsalsum*.

**Author Contributions:** J.S.P. and Z.L. conceived the research; J.S.P. and K.W.L. performed field work; J.S.P., Z.L., H.J.K., K.H.K., J.Y.Y., K.Y.K., H.H.S. analyzed the data; J.S.P., Z.L., H.H.S. and K.W.L. wrote and edited the manuscript. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by grants from Korea Institute of Ocean Science & Technology (PE99921), the National Marine Biodiversity Institute of Korea (2021M01100), the National Research Foundation of Korea (NRF) (No. 2018R1A6A3A01012375), and by the KRIBB Research Initiative Program (KGM5232113) and the National Research Foundation of Korea (NRF) grant funded by the Korea government (MSIT) (No. 2021R1C1C1008377).

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

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**


## *Article* **New Records of the Diatoms (Bacillariophyceae) from the Coastal Lagoons in Korea**

**Daeryul Kwon 1, Mirye Park 1, Chang Soo Lee 1, Chaehong Park <sup>2</sup> and Sang Deuk Lee 3,\***


**Abstract:** Lagoons are natural bodies of water that are isolated from the sea due to the development of a sand bar or spit. Each lagoon has distinct ecological characteristics, and these sites also serve as popular tourist attractions because they are common habitats for migratory birds and are characterized by beautiful natural scenery. Lagoons also have distinct ecological characteristics from those of their associated estuaries, and there are active research efforts to classify, qualify, and quantify the high biodiversity of lagoons. The lagoons in Korea are primarily distributed in the East Sea, and are represented by Hwajinpo, Yeongrangho, and Gyeongpoho. Here, we report the discovery of 11 unrecorded diatom species (*Diploneis didyma*, *Mastogloia elliptica*, *Cosmioneis citriformis*, *Haslea crucigera*, *Pinnularia bertrandii*, *Pinnularia nodosa* var. *percapitata*, *Gyrosigma sinense*, *Gomphonema guaraniarum*, *Gomphonema italicum*, *Navicula freesei*, *Trybionella littoralis* var. *tergestina*) among samples collected from the Hwajinpo, Hyangho, Maeho, Gapyeongri wetland, Cheonjinho, and Gyeongpoho lagoons in Korea during a survey from 2018–2020. We present the taxonomic characteristics, ecological information, habitat environmental conditions, and references for these 11 species.

**Keywords:** lagoon; new record diatoms; taxonomic; ecological; habitat

#### **1. Introduction**

Lagoons are a form of riparian terrain where the river and sea water meet, but they have unique ecological characteristics that distinguish them from estuaries because, in lagoons, the entrance to the coast is blocked by sand dunes [1,2]. The lagoons distributed along the Korean coast are known to have been formed by a combination of rising sea levels during the postglacial age and the development of sandbars or sand dunes [3,4].

The East Sea coast boasts the cleanest marine environments in Korea and is blessed with natural resources to sustain the area. There are 18 lagoons distributed along a 112 km stretch of the East Sea coast [5]. These lagoons provide beautiful natural scenery and have unique ecological value with their brackish water lakes, important migratory bird habitats, and are also economically valuable as tourist attractions [4,6].

Hwajinpo Lake is the largest lake in Korea with a circumference of 16 km and is a typical high-salinity brackish lake with a high ecological value [4,7]. Hyangho Lagoon hosts a community of aquatic plants covering approximately 20% of its area. It belongs to a heavily landlocked lake, and, unlike other lagoons, freshwater lakes are distributed behind it, serving as a sheltered habitat for organisms living in the lake [8]. Cheonjinho Lake's characteristic brackish water has disappeared due to the blocked inflow of seawater; it has a wide variety of free-floating and floating-leaved plants, and a high proportion of aquatic and wetting plants [9,10]. Gyeongpoho is the representative lagoon of the East Sea coast for the endangered prickly waterlily, *Euryale ferox*, which appeared 40 years ago and is

**Citation:** Kwon, D.; Park, M.; Lee, C.S.; Park, C.; Lee, S.D. New Records of the Diatoms (Bacillariophyceae) from the Coastal Lagoons in Korea. *J. Mar. Sci. Eng.* **2021**, *9*, 694. https://doi.org/10.3390/ jmse9070694

Academic Editors: Wonho Yih, Milva Pepi and Magnus Wahlberg

Received: 1 May 2021 Accepted: 16 June 2021 Published: 24 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/).

known as a favored habitat for various migratory birds [11,12]. The Maeho lake has a high ecological conservation value due to its diverse distribution of biological resources and ecosystem services, such as being designated as a cultural heritage reserve. However, the lake area is decreasing due to the nearby expansion of farmland and subsequent soil erosion, causing sedimentation [4]. Gapyeongri wetland is one of the smallest lagoons, well-hidden behind a widely developed coastal sand dune [13]. Yeomgaeho Lake is also small, but is a typical natural lagoon, suitable as a migratory bird habitat; however, land reclamation is occurring rapidly in the area [13]. Regardless, the lagoons are worthy of research because they have ecologically diverse characteristics and high biodiversity [14,15].

Diatoms are recognized worldwide as one of the most suitable biological components for water quality assessments because they continuously occur in aquatic ecosystems and respond quickly to environmental changes [16,17]. Diatoms can occur wherever there is water, can attach to all substrates, such as gravel and plants, and can live in various water environments [18,19]. However, research on diatoms in lagoons is poor. Even though data from low-salinity environments are readily available, those from high-salinity environments are limited [20].

There are many reports of new and unrecorded diatoms in Korea [21–23]. Species collected in environments such as lagoons, estuaries, and sedimentary soils with unique habitats other than freshwater environments are being reported [23–25].

In this study, we describe taxonomic information on 11 previously unrecorded diatom species found in six lagoons in Korea, and provide details on the ecological characteristics, reference specimen, basionym, synonym, and distribution of each.

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

#### *2.1. Collecting and Fixation of Samples*

Samples were collected from six lagoons: Hwajinpo, Hyangho, Maeho, Gapyeongri wetland, Cheonjinho, and Gyeongpoho, adjacent to the East Sea in Korea from 2018–2020. The epilithic diatoms were scraped off the upper surface of the gravel using a small chisel or toothbrush. The collected wild cells were immediately fixed with Lugol's solution (iodine-iodide solution), 4% neutralized formalin, or 2% glutaraldehyde [26]. Environmental parameters such as water temperature (◦C), pH, dissolved oxygen (mg/L), electric conductivity (μs/cm3), and salinity (PSU) were measured using a portable multiparameter water quality meter (Pro DSS, YSI, Yellow Springs, OH, USA) in the field.

#### *2.2. Pretreatment and Observation of Diatoms*

To remove organic matter, the fixed samples were boiled in equal amounts of mixed HCl and KMnO4 at 70 ◦C until the samples were slightly colored, and were then rinsed in distilled water to remove residual acid, following a modified method of Hasle & Fryxell [27]. The samples were then rinsed in distilled water to remove any residual acid. The samples, some fixed and some cleaned, were mounted in Pleurax (Mountmedia, Wako, Japan) for observation using a light microscope (LM) (Eclipse Ni; Nikon, Tokyo, Japan), equipped with Nomarski differential interference contrast optics (DIC) and a digital camera (DS-Ri2; Nikon, Tokyo, Japan). In addition, the morphological characteristics of various foci on the valve face on *Mastogloia elliptica* were observed with a microscope.

The terminology was according to that of Anonymous [28], Ross et al. [29], and Round et al. [30].

#### **3. Results**

#### *3.1. Identification of Diatom Species in Korean Lagoons*

A total of 11 previously unrecorded diatom species were identified from the six lagoon sites surveyed in this study (Figure 1). Two species each belonged to the genera *Gomphonema* and *Pinnularia*, with one species represented by the genera *Diploneis*, *Mastogloia*, *Cosmioneis*, *Haslea*, *Gyrosigma*, *Navicula*, and *Tryblionella*, respectively. According to site, four of the unrecorded species (*Mastogloia elliptica*, *Cosmioneis citriformis*, *Haslea crucigera*, and *Pinnularia*

*bertrandii*) were detected at Hyangho, the two *Gomphonema* species (*G. guaraniarum* and *G. italicum*) were detected in the Gapyeongri wetland, and two species (*Navicula freesei* and *Tryblionella littoralis* var. *tergestina*) appeared in Yeomgaeho. One unrecorded species was, respectively, detected at Gyeongpoho (*Diploneis didyma*), Cheonjinho (*Pinnularia nodosa* var. *percapitata*), and Maeho (*Gyrosigma sinense*) (Table 1).

**Figure 1.** Sampling sites along the lagoons (CJH: Chunjinho, GPW: Gapyeongri wetland, YGH: Yeomgaeho, MH: Maeho, HH: Hyangho, GPH: Gyeongpoho).


**Table 1.** Information of sampling sites.

#### *3.2. Environmental Characteristics at the Sampling Sites*

The pH of the six sites ranged from 7.6 to 8.5, showing a weakly basic range. The salinity of Gyeongpoho was in the range of 22.1–25.4 PSU and the electrical conductivity was higher than 35,000 μS/cm3. By contrast, Yeomgaeho and Cheonjinho had a salinity of less than 0.2 PSU and an electrical conductivity of less than 360 μS/cm3, which are more similar to the conditions of a freshwater environment (Table 2).


**Table 2.** Field water quality of sampling sites.

#### *3.3. Species Descriptions*

3.3.1. *Cosmioneis citriformis* R.L. Lowe & A.R. Sherwood 2010

Original description: Lowe and Sherwood 2010, p. 24, Figures 10–18 [31].

Dimensions: Length Range 29–33 μm, width range 12–15 μm, 17–20 striae in 10 μm. Description: The valves are citriform with broadly rounded margins. The ends are short, narrow, and rostrate. The axial area is narrow and expands to a rounded central area. The proximal raphe ends are slightly expanded. The distal raphe ends are curved in the same direction. Internally, the raphe ends are anchor-shaped. Striae are punctate and radiate, curving from the margin towards the center of the valve. There are 17 striae in 10 μm in the centre that become finer at the ends (20 in 10 μm). There are 16 areolae in 10 μm, which are round towards the valve ends, becoming transversely elongated near the centre (Figure 2A).

**Figure 2.** Light micrographs of diatoms. (**A**) *Cosmioneis citriformis* (**B**) *Diploneis didyma* (**C**,**D**) *Gomphonema guaraniarum* (**E**) *Gomphonema italicum* (**F**,**G**) *Gyrosigma sinense* (**H**) *Haslea crucigera*. Scale bars. 2–6, 8, 9:10 μm; 7:50 μm.

Ecology and distribution: Distributed in freshwater habitats, benthic taxon. Hyangho lake is the focus in this study (Table 1).

3.3.2. *Diploneis didyma* (Ehrenberg) Ehrenberg 1839

Original description: Ehrenberg 1839, pls. 1–4 [32].

Basionym: *Pinnularia didyma* Ehrenberg 1844 [33].

Dimensions: Length Range 60–74 μm, width range 25–28 μm, 8–10 striae in 10 μm.

Description: The cells are solitary and free, the valves are elliptical or linear-elliptical, with or without a median constriction, and the apices are rounded or broadly cuneate. The central nodule is prominent, often large, quadrate, and strongly formed; the central area is small, reduced, and produces longitudinal extensions that are usually described as "horns" which lie on either side of the raphe and enclose it as solid ribs. Beyond the horns, there are thinner, usually narrow, and depressed areas, typically referred to as "furrows." These may be hyaline and structureless, may contain a row of large puncta, or may be crossed by faintly transverse costae. Beyond the furrows, on each side of the raphe, some specimens have a lunate area in each segment, which is usually referred to as the "lunula." This may be crossed by costae or alveoli, which may or may not bear a single or double row of puncta; these are frequently more developed and closer to the valve margin. In some specimens, transverse costae may be absent. Chromatophores are usually found as two deeply crenulated bodies that lie along the girdle. The valves are panduriform and slightly constricted in the middle, dividing the valve surface into two tongue-shaped segments. The central nodule is subquadrate or almost circular in some cases and protrudes to form two horns on either side of the raphe or median line, respectively. The valve surface is costate, transverse in the middle, but slightly curving. Radiating lines are found towards the apices that are crossed by numerous undulating longitudinal lines (Figure 2B).

Ecology and distribution: Distributed in marine-brackish habitats, benthic taxon. Gyeongpoho lake is the focus in this study (Table 1).

#### 3.3.3. *Gomphonema guaraniarum* Metzeltin & Lange-Bertalot 2007

Original description: Metzeltin & Lange-Bertalot 2007,p. 147, pl. 212, Figures 9–14 [34]. Dimensions: Length Range 58–77 μm, width range 10–12 μm, 10–12 striae in 10 μm.

Description: The valves are rhombic-lanceolate, with less rounding of the apical and basal ends. The raphe-sternum is narrow and linear. The central area is unilaterally expanded, which is limited by a shortened median striae. The raphe is slightly sinuous with proximal ends that are punctuated and are slightly curved towards the stigma. The transapical striae slightly radiate parallel to the central region. Areolae are inconspicuous. There is a stigma at the end of the central stria. Under scanning electron microscopy (SEM), the stigma appears delicate and rounded, and striae are uniseriate with the areolae rounded to elongate longitudinally. The ends of the raphe dilate into pores, and the distal ends are curved, extending to the valve mantle. The pore field is formed by rounded poroids disposed on both sides of the terminal raphe fissure (Figure 2C,D).

Ecology and distribution: Distributed in freshwater habitats, epiphytic taxon. Gapyeongri Wetland is the focus in this study (Table 1).

#### 3.3.4. *Gomphonema italicum* Kützing 1844

Original description: Kützing 1844, p. 85, pl. 30, Figure 75 [35].

Dimensions: Length Range 24–50 μm, width range 10–12 μm, 11–14 striae in 10 μm.

Description: The frustules in girdle view are wedge-shaped. The valves are strongly heteropolar and clavate, with the largest valve width in the upper valve half. The valves are tumid at the center, gradually narrowing towards the footpole and are slightly constricted towards the broadly rounded headpole. The axial area is moderately broad and linear. The central area is small and irregular in shape, bordered on each margin by a few irregularly shortened striae. One isolated pore is present at the end of the long central stria. The external isolated pore opening is small and rounded. The raphe is distinctly lateral and strongly undulated, with simple and slightly expanded proximal endings. The external proximal raphe endings are teardrop-shaped and are deflected towards the isolated pore. At the headpole, the distal raphe ends first deflect towards the pore-bearing side and then towards the opposite side, extending onto the valve mantle. At the footpole, the distal raphe dissects into a well-developed pore field. The transapical striae are not interrupted near the valve face/mantle junction but rather continue onto the valve mantle (Figure 2E).

Ecology and distribution: Distributed in freshwater and terrestrial habitats, epiphytic taxon. Gapyeongri Wetland is the focus in this study (Table 1).

#### 3.3.5. *Gyrosigma sinense* (Ehrenberg) Desikachary 1988

Original description: Desikachary 1988, pp. 1–13, pls. 401–621 [36].

Basionym: *Navicula sinensis* Ehrenberg 1847 [37].

Description: The valves are linear-sigmoid, with inflated central and distal portions. The color in resin and standardized dark-field microscopy is bright blue. The raphe sternum has a double curvature and is rotated towards the internal central raphe node, strongly eccentric at the ends, where it is markedly displaced owing to its concavity. The central area is rhombic and rotated. The terminal areas are triangular and strongly displaced away from the apices so that they are in a completely lateral position. The central external raphe fissures have isomorphic deflection patterns crossing the striae. The apical structure shows one very long and one short apical microforamina segment on opposite sides of the raphe sternum (Figure 2F,G).

Ecology and distribution: Distributed in marine habitats, benthic taxon. Maeho lake is the focus in this study (Table 1).

#### 3.3.6. *Haslea crucigera* (W. Smith) Simonsen 1974

Original description: Simonsen 1974, pp. 47 [38].

Basionym: *Schizonema crucigerum* W. Smith 1856 [39].

Dimensions: Length Range 95–97 μm, width range 11–12 μm, 17–20 striae in 10 μm.

Description: Living cells have slightly curved, narrowly rectangular frustules in girdle view. The valves are lanceolate to linear–lanceolate. Two band-like plastids lie against the girdle on each side of the cell. The internal margin of the plastids usually appears to be slightly undulating because of the presence of small, obliquely inserted, and rod-shaped pyrenoids. In the cleaned cells, the raphe appears straight and central. Transapical striae are visible under light microscopy and are crossed by more delicate longitudinal striae. The central two or three transapical virgae are thickened, forming a pseudostauros. Under SEM, the external valve surface is covered with closely spaced longitudinal strips of silica separated by narrow slits, which merge with a continuous peripheral slit near the apices. The external raphe fissures are slightly expanded and turn to one side centrally before sharply deflecting to the same side at the poles. Internally, the raphe slits open laterally in the raphe sternum, except at the center where the endings are straight and approximate, and at the apices, where they are slightly expanded in a slightly raised helictoglossa. An accessory rib on the primary side of the valve flanges over the raphe sternum, obscuring it for much of its length. The internal areola arrangement is similar to that of the other taxa, but with fewer longitudinal striae. On both sides of the raphe, the three central virgae thicken, forming a pseudostauros. The thickened virgae are fused with an accessory rib on the primary side of the valve and with a shorter thinner rib on the secondary side of the valve. The thickening of the virgae extends further across the valve and is more even compared with that in *H. salstonica* (Figure 2H).

Ecology and distribution: Distributed in brackish inland water habitats, benthic taxon. Hyangho lake is the focus in this study (Table 1).

#### 3.3.7. *Mastogloia elliptica* (C. Agardh) Cleve 1983

Original description: (C. Agardh) Cleve 1983, pl. 185, Figures 24–27 [40]. Basionym: *Frustulia elliptica* C. Agardh 1824 [41].

Dimensions: Length Range 17–57 μm, width range 8–12 μm, 15–18 striae in 10 μm. Description: The valves are elliptical to linear-elliptical with convex to nearly parallel sides, and blunt, obtusely rounded apices. The axial area is narrow and barely wider than the raphe, and the central area is circular. The raphe branches are lateral and sinuous, with weakly expanded proximal ends. There are numerous tecta of approximately the same size. The striae consist of single rows of areolae that radiate throughout (Figure 3A–I). The raphe is a straight line or an undulate shape. The striae of *Mastogloia elliptica* are parallel in shape, and the raphe is straight (Figure 3A). In addition, due to the complex valvocopula structure, pertaca and striae are visible at the same time (Figure 3B–D), with seven symmetrical pertaca on each side (Figure 3E–I).

**Figure 3.** Light micrographs of *Mastogloia elliptica* in various depth of field ranges in microscope. (**A**–**C**,**I**) surface focus of valve face with uniseriate striae (**D**–**H**) Deeper focus with similar sized numerous partecta (arrows). Scale bar:10 μm.

Ecology and distribution: Distributed in marine and freshwater habitats, benthic taxon. Hyangho lake is the focus in this study (Table 1).

3.3.8. *Navicula freesei* R.M. Patrick & Freese 1961

Original description: Patrick, R.M. & Freese 1961, p. 206, pl. 2, Figure 14 [42]. Dimensions: Length Range 61–77 μm, width range 13–15 μm, 8–9 striae in 10 μm.

Description: The valves are lanceolate with rounded apices. The striae are radiated at the center, and are parallel and slightly convergent at the apices. The central area is asymmetrically rounded, whereas the axial area is narrow and straight. Striae on one or both sides of the central area are irregularly spaced. The raphe is lateral and straight. The distal raphe fissures form curved hooks onto the mantle. The proximal raphe ends form an elongated central poroid. A thickened central nodule is evident. Girdle bands have not yet been defined (Figure 4A).

**Figure 4.** Light micrographs of diatoms. (**A**) *Navicula freesei* (**B**) *Pinnularia bertrandii* (**C**) *Pinnularia nodosa* var. *percapitata* (**D**,**E**) *Tryblionella littoralis* var. *tergestina*. Scale bars:10 μm.

Ecology and distribution: Distributed in freshwater habitats, benthic taxon. Yeomgaeho is the focus in this study (Table 1).

#### 3.3.9. *Pinnularia bertrandii* Krammer 2000

Original description: Krammer 2000, pp. 122, 226, pl. 91, Figures 22–30 [43].

Dimensions: Length Range 14–27 μm, width range 4–6 μm, 17–18 striae in 10 μm.

Description: The valves are linear-elliptical to linear-lanceolate with weakly convex sides. The ends are subcapitate nearly covering the breadth of the valve body, which are broadly rounded. The raphe is filiform to slightly lateral, the central pores are small and slightly laterally bent, and the terminal fissures are distinct. The axial area is narrowed and the lanceolate is widened from the ends to the fascia. The striae in the middle weakly radiate and moderately converge at the ends and longitudinal bands are absent (Figure 4B).

Ecology and distribution: Distributed in freshwater habitats, benthic taxon. Hyangho lake is the focus in this study (Table 1).

#### 3.3.10. *Pinnularia nodosa* var. *percapitata* Krammer 2000

Original description: Krammer 2000, pp. 57, Figures 26:9-12; 27:9, 10 [43].

Dimensions: Length Range 47–61 μm, width range 9–10 μm, 8–10 striae in 10 μm.

Description: The valves are linear with triundulate margins. In the largest specimens, the central undulation is wider than the distal undulations. Apices are distinctly capitated. Axial areas are about one-third of the valve's width and widen from the apices towards the valve center. The central area is a bilateral fascia. The surface of the valve is mottled along either side of the raphe and continues into the central area. The raphe is straight. The proximal raphe ends are bent to one side and terminate in small, tear-shaped pores. Distal raphe fissures are shaped like question marks. The striae are weakly radiated at the valve center and become strongly convergent at the apices (Figure 4C).

Ecology and distribution: Distributed in freshwater habitats, benthic taxon. Chunjinho lake is the focus in this study (Table 1).

#### 3.3.11. *Tryblionella littoralis* var. *tergestina* (Grunow) Snoeijs 1998

Original description: Snoeijs & Balashova 1998, pp. 1–144, Figure 1, pls. 101 [44]. Basionym: *Nitzschia littoralis* var. *tergestina* Grunow 1880 [45].

Dimensions: Length Range 30–100 μm, width range 12–30 μm, 30–38 striae in 10 μm. Description: The valves are broadly elliptical-lanceolate to linear-elliptical, with a very slight central constriction of the keel. Apices are cuneate, narrowed, and rounded. Striae are difficult to resolve under light microscopy (Figure 4D,E).

Ecology and distribution: Distributed in freshwater habitats, benthic taxon. Yeomgaeho is the focus in this study (Table 1).

#### **4. Discussion**

Reservoir salt concentrations fluctuate over time; thus, lagoons create a unique ecosystem by mixing inland freshwater with intrusions of seawater. Accordingly, freshwater life with resistance to salt mixes with marine life in the lagoon ecosystem [3]. In this study, 11 previously unrecorded diatom species were discovered in six lagoons in Korea. *Diploneis didyma* was detected in Gyeongpoho. This species is mainly found in harbors but has also been reported in freshwater. At the time of the survey, Gyeongpoho had a high electrical conductivity above 30,000 μS/cm3 and the salinity was 22 PSU, demonstrating environmental conditions similar to those of seawater (Tables 2 and 3). *Cosmioneis* predominantly inhabits brackish water zones [46–48] but has also been found in alkaline freshwater environments [30,49]. We detected *Cosmioneis citriformis* in Hyangho, which has very low salinity (0.03 PSU) and an electrical conductivity of 60 μS/cm3, demonstrating similarity to the conditions of a freshwater environment. However, the measurements taken in the survey of 2018 showed a salinity of 5.0 PSU and an electrical conductivity of 8925 μS/cm<sup>3</sup> at this site, and *C. citriformis* was not found at that time (Tables 2 and 3). *Haslea crucigera* is a benthic species and is mainly found in high-salinity water [50]; however, during the 2018 survey, the salinity at Hyangho was 5.0 PSU and the electrical conductivity was 8926 μS/cm3. Therefore, *H. crucigera* appears to prefer a water environment with some salt, although the salinity at this site remained lower than that of seawater throughout the survey period (Tables 2 and 3). *Pinnularia bertrandii* is a benthic species that is mainly detected in freshwater environments [51]. During the survey in 2020, the environmental conditions of the water at Hyangho were similar to those of a freshwater environment with a salinity of 0.03 PSU and an electrical conductivity of 60 μS/cm3 (Tables 2 and 3). *Pinnularia nodosa* var. *percapitata* is also a benthic species that is mainly found in freshwater environments, mostly in streams, reservoirs, and small lakes, with a pH ranging from 6.1 to 7.8 (slightly acidic to neutral), and an electric conductivity from 22 to 169 μS/cm3 [43]. At the time of sampling, the water of Chunjinho had very low salinity of 0.1–0.2 PSU, with a pH of 8.1 to 8.7 and a relatively low electrical conductivity of 152–360 μS/cm3 (Tables 2 and 3). *Gyrosimga sinense* is a benthic species that has a seawater-based and widespread distribution under high water temperatures [52]. At the time of sampling, the water temperature of Maeho was relatively high, ranging from 15.8 ◦C to 20.2 ◦C, with salinity ranging from 3.9 to 9.4 PSU, and electrical conductivity ranging from 5767 to 15,969 μS/cm3 (Tables 2 and 3). The genus *Gomphonema* is mostly attachable and is mainly found in freshwater environments, representing the largest genus in freshwater environments among diatoms of the world's most broadly distributed species [53,54]. In particular, *Gomphonema* can grow by attaching to aquatic plants because of the mucous stem secreted from the pore fields at the end of the valve [30,55,56]. *G. guaraniarum* is a freshwater species and its ecological properties are rarely described in the literature. *G. italicum* is also a freshwater species with a relatively high electrical conductivity of 13,250 μS/cm<sup>3</sup> and is known to emerge in environments with a weakly basic pH [57]. In this study, *G. guaraniarum* and *G. italicum* were detected at the Gapyeong-ri wetland, with a water temperature of 13.3 ◦C, a DO of 16.2 mg/L, a pH of 8.0, a salinity of 0.9 PSU, and an electrical conductivity of 1528 μS/cm3, indicating that they inhabit an environment with freshwater-like conditions (Tables 2 and 3). *Navicula freesei* was found in freshwater and brackish environments, mainly with neutral

to weakly basic conditions [42]. *Tryblionella littoralis* var. *tergestina* is an epipelic diatom and marine species [44]. In this study, these two species were identified at Yeomgaeho, with the water environment being similar to a freshwater environment at the time of sampling (a salinity 0.1 PSU, a pH of 7.6, and an electrical conductivity of 288 μS/cm3) (Tables 2 and 3). This suggests that *Tryblionella littoralis* var. *tergestina* adapted to the desalinated lagoon. *Mastogloia* is a benthic and epiphytic diatom, which was identified in both freshwater and marine environments; however, most species of this genus appear in marine environments and prefer weakly basic water bodies [58,59]. Patrick and Reimer [60] report *M. elliptica* as a halophilic to mesohalobic taxon characteristic of coastal areas, but is also found in inland lakes with some salinity. In Europe, Krammer and Lange-Bertalot [61] report *M. elliptica* from brackish waters in coastal areas and from saline inland waters. At the time of sampling, the pH of Hyangho was 8.2 and the salinity was 0.03 PSU, exhibiting a water environment close to that of the freshwater environment (Tables 2 and 3). *Mastogloia* has oval to linear oval convex valves, with complex silica chambers on both sides, called pertaca, that secrete mucus, and living cells that have two plastids [39,62]. This genus appears to be closely related to the genus *Aneumastus* but with a more complex valvocopula [63]. Patrick and Reimer [60] reported that *M. elliptica* is a halophilic to mesohalobic taxon characteristic of coastal areas but is also found in inland lakes with some salinity. In Europe, Krammer and Lange-Bertalot [61] reported *M. elliptica* from brackish waters in coastal areas and from saline inland waters.



**Author Contributions:** Data curation, formal analysis, writing-original draft, and writing-review and editing, D.K. and S.D.L.; funding acquisition, C.S.L.; field investigation, D.K., M.P., C.S.L. and S.D.L.; writing—review and editing, D.K., M.P., C.S.L., C.P. and S.D.L. All authors have read and agreed to the published version of the manuscript.

**Funding:** This work was supported by grants from the Nakdonggang National Institute of Biological Resources (NNIBR) (NNIBR202101103) projects.

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

**Informed Consent Statement:** Not applicable.

**Data Availability Statement:** Not applicable.

**Conflicts of Interest:** The authors declare no conflict of interest.

#### **References**

