Distribution of Harmful Algal Growth-Limiting Bacteria on Artiﬁcially Introduced Ulva and Natural Macroalgal Beds

: The intensity and frequency of harmful algal blooms (HABs) have increased, posing a threat to human seafood resources due to massive kills of cultured ﬁsh and toxin contamination of bivalves. In recent years, bacteria that inhibit the growth of HAB species were found to be densely populated on the bioﬁlms of some macroalgal species, indicating the possible biological control of HABs by the artiﬁcial introduction of macroalgal beds. In this study, an artiﬁcially created Ulva pertusa bed using mobile ﬂoating cages and a natural macroalgal bed were studied to elucidate the distribution of algal growth-limiting bacteria (GLB). The density of GLB a ﬀ ecting ﬁsh-killing raphidophyte Chattonella antiqua , and two harmful dinoﬂagellates, were detected between 10 6 and 10 7 CFU g − 1 wet weight on the bioﬁlm of artiﬁcially introduced U. pertusa and 10 to 10 2 CFU mL − 1 from adjacent seawater; however, GLB found from natural macroalgal species targeted all tested HAB species (ﬁve species), ranging between 10 5 and 10 6 CFU g − 1 wet weight in density. These ﬁndings provide new ecological insights of GLB at macroalgal beds, and concurrently demonstrate the possible biological control of HABs by artiﬁcially introduced Ulva beds.


Introduction
Harmful algal blooms (HABs) are naturally-occurring phenomena in fresh and marine water systems, often causing severe damage to the fishery industry worldwide, through massive kills of commercially cultured fishes and toxin-contamination of shellfish leading to closures of commercial shellfish harvests, and, in some cases, consumption of contaminated shellfish results in human illness and even death [1][2][3]. Lately, the acceleration of geographic expansion and intensity of HABs on a global scale associated with climate change have been reported [4][5][6][7]. Hence, developments of effective countermeasures to control the HAB events are even more urgent and crucial concerns. Mechanical, chemical, genetic, biological, and environmental measures to prevent HAB have been proposed [4], and clay dispersion is the only practical measure confirmed to control HABs formed by Cochlodinium polykrikoides in Korean coastal waters [8]. However, the negative impacts of clay on non-target organisms living in the seafloor and water column are also pointed out [9][10][11][12]. The control of HABs directly intervenes with surroundings in dynamic ocean environments, demanding an establishment of measures as environmentally friendly as possible. Among all proposed strategies to mitigate HABs, biological controls using naturally living organisms may have the least impact on ecosystems in the present circumstances. Numbers of biological controls using microbes to bivalves have been proposed and tested extensively for several decades [13][14][15][16][17][18][19][20]. In particular, bacteria that kill and/or inhibit the growth of HAB species often associated with bloom termination have received worldwide attention as possible biological control of HABs [21][22][23][24][25][26][27][28][29][30][31].
Although the use of natural pathogenicity could be environmentally friendly and cost-effective to control HABs, introducing potentially invasive bacteria may require a careful manner [32,33]. Recently, a biofilm of seagrass and macroalgae were found to be densely populated by algal growth-limiting bacteria (GLB), including both algicidal and growth-inhibiting bacteria [34][35][36][37], indicating their function as an enormous source of GLB for the adjacent seawater [31,38]. The latest microcosm study demonstrated that seawater in the Zostera marina bed successfully suppressed artificial Chattonella blooms [31], indicating that protection and restoration of seagrass and macroalgal beds prevent the outbreak of HABs through providing the specific microbial communities. However, these holistic approaches may sometimes not be flexible in actual field applications in that they often require specific habitat characteristics such as light, sediment or substrate types, currents, bioturbation, etc. [39,40].
In this study, Ulva pertusa beds using mobile floating cages were artificially created to study the distribution of algal growth-limiting bacteria (GLB), seeking flexible and widely applicable technologies to control HABs. A natural macroalgal bed was also studied for comparison.

Installation of Artificial Ulva Bed
Yellowish-green reproductively matured Ulva pertusa thalli were provided from Wakayama prefectural fisheries experiment station, Wakayama Prefecture, Japan. After the thallus of U. pertusa was gently wiped and dried in the shade for 1-2 h, it was placed into a water tank filled with autoclaved-filtered seawater [41]. Pumice stones were carefully rinsed with autoclaved seawater and gently set to promote attachments of motile zoospore of U. pertusa [42]. Then, the pumice stones were placed into mobile floating cages and installed on June 20, 2017, in a nearshore area by an upright seawall at Station 1 (Stn. 1), Kushimoto-cho, Wakayama Prefecture, Japan ( Figure 1). The location of installation was carefully selected after confirming the absence of any macroalgal species in the area. However, this may not guarantee attachments of naturally-occurring macroalgal propagules to the tested pumice stone and cages after installation.

Sampling
Samplings were performed on 7 to 8 September 2017, at Stn. 1 (artificial Ulva bed) and Stn. 2 (a natural macroalgal bed), Taiji-cho, Wakayama Prefecture, Japan ( Figure 1). The weather was mostly cloudy during the sampling on 7 September 2017, with surface seawater temperature and salinity of 27.9 • C, 34.0 PSU at Stn. 1. On 8 September 2017, it was sunny with surface seawater temperature and salinity of 29.2 • C, 34.7 PSU at Stn. 2. The green alga, U. pertusa, successfully grown on the pumice stone and floating cage at Stn. 1 ( Figure 2) and four dominant macroalgae (one brown alga: Sargassum dupulicatum, one red alga: Gelidium elegans, two green algae: Cladophora ohkuboana and U. pertusa) at Stn. 2 were sampled using sterile instruments in a sterile manner. Adjacent surface seawater samples were also collected from both sites. Samplings were conducted at low tide, and all the samples were collected in an autoclaved polycarbonate bottle (500 mL) filled with autoclaved, filtered (0.7 µm GF/F) seawater. All the samples were immediately transported to the laboratory and processed on the same day.
autoclaved-filtered seawater [41]. Pumice stones were carefully rinsed with autoclaved seawater and gently set to promote attachments of motile zoospore of U. pertusa [42]. Then, the pumice stones were placed into mobile floating cages and installed on June 20, 2017, in a nearshore area by an upright seawall at Station 1 (Stn. 1), Kushimoto-cho, Wakayama Prefecture, Japan ( Figure 1). The location of installation was carefully selected after confirming the absence of any macroalgal species in the area. However, this may not guarantee attachments of naturally-occurring macroalgal propagules to the tested pumice stone and cages after installation.  Appl. Sci. 2020, 10, x FOR PEER REVIEW 3 of 16

Sampling
Samplings were performed on 7 to 8 September 2017, at Stn. 1 (artificial Ulva bed) and Stn. 2 (a natural macroalgal bed), Taiji-cho, Wakayama Prefecture, Japan ( Figure 1). The weather was mostly cloudy during the sampling on 7 September 2017, with surface seawater temperature and salinity of 27.9 ℃, 34.0 PSU at Stn. 1. On 8 September 2017, it was sunny with surface seawater temperature and salinity of 29.2 ℃, 34.7 PSU at Stn. 2. The green alga, U. pertusa, successfully grown on the pumice stone and floating cage at Stn. 1 ( Figure 2) and four dominant macroalgae (one brown alga: Sargassum dupulicatum, one red alga: Gelidium elegans, two green algae: Cladophora ohkuboana and U. pertusa) at Stn. 2 were sampled using sterile instruments in a sterile manner. Adjacent surface seawater samples were also collected from both sites. Samplings were conducted at low tide, and all the samples were collected in an autoclaved polycarbonate bottle (500 mL) filled with autoclaved, filtered (0.7 µm GF/F) seawater. All the samples were immediately transported to the laboratory and processed on the same day.

Sample Processing and Bacterial Culturing
Macroalgal samples using bottles filled with 200 mL autoclaved filtered seawater were shaken 600 times by hand using a hand tally counter to detach a biofilm formed on the surface of the macroalgae. Biofilm-suspended samples were serially diluted 10-fold to 10 −4 with autoclaved filtered seawater, and an aliquot (0.1 mL) of each dilution was spread onto a ST10 −1 marine agar plate (Trypticase peptone 0.5 g, Yeast extract 0.05 g and agar 15 g in 1 L seawater [43]). The adjacent seawater, also collected from both sites, were serially diluted 10-fold to 10 −4 , and an aliquot (1 mL) was filtered through autoclaved 3 µm pore size Nucleopore Track-Etch membrane filters (Whatman,

Sample Processing and Bacterial Culturing
Macroalgal samples using bottles filled with 200 mL autoclaved filtered seawater were shaken 600 times by hand using a hand tally counter to detach a biofilm formed on the surface of the macroalgae. Biofilm-suspended samples were serially diluted 10-fold to 10 −4 with autoclaved filtered seawater, and an aliquot (0.1 mL) of each dilution was spread onto a ST10 −1 marine agar plate (Trypticase peptone 0.5 g, Yeast extract 0.05 g and agar 15 g in 1 L seawater [43]). The adjacent seawater, also collected from both sites, were serially diluted 10-fold to 10 −4 , and an aliquot (1 mL) was filtered through autoclaved 3 µm pore size Nucleopore Track-Etch membrane filters (Whatman, GE Healthcare, Parramatta, Australia). The filter was gently placed on a ST10 −1 marine agar plate to culture the particle-associated bacteria (PAB). An aliquot of 0.1 mL from the filtrate was also used to culture the free-living bacteria (FLB) on the same medium above [44]. All the plates were kept for two weeks in the dark at 20 • C and used for the culturable bacterial enumeration and isolation. For the isolation, colonies were randomly picked by sterilized toothpicks from the agar plate, inoculated onto a 48-well filled with the same bacterial medium and kept at 20 • C in the dark until the following co-culture experiment. Aliquots (10 mL) from all samples were fixed with glutaraldehyde (1% final concentration) and stored at 4 • C for later determination of total bacterial abundance by DAPI staining [45].

Co-Culture Experiment Using Bacterial Isolates
Cultured bacteria isolated from the biofilm of U. pertusa grown on the pumice stone and floating cage (Stn. 1) and four macroalgal species (Stn. 2), as well as PAB and FLB from the surrounding seawater at both sites, had their growth inhibitory properties determined through co-culturing experiments (a single axenic HAB species cultured in the algal medium with a single bacterium) toward five targeted HAB species [37]. Cell densities were diluted with the same algal medium to approximately 10 3 cells mL −1 for C. antiqua, A. catenella, and K. mikimotoi and 10 4 cells mL −1 for H. akashiwo and H. circularisquama, then 0.8 mL aliquots were incubated for two days in sterilized disposable 48-well microplates to check for any abnormal appearance of cells. Individual bacterial isolates were inoculated into the algal cultures with an approximate density of 10 5 cells mL −1 , and the plates were incubated for another two weeks with the conditions described above for the algal cultures. Bacteria-free quadruplicate wells (Controls) were also prepared in each microplate. A Nikon ECLIPSE TE200 inverted microscope was used for daily checks of algal cell mortality and/or any abnormalities of algal cells. The wells with 90% or more of algal cells killed were labeled as algicidal bacteria (AB), and the wells with 90% or more of motility reduced algal cells simultaneously were labeled as growth-inhibiting bacteria (GIB). In this study, the term algal growth-limiting bacteria (GLB) was mainly applied to express a set of bacteria limiting algal growth, including both AB and GIB, to eliminate repeated usage of those terms. The density of AB and GIB was estimated using the following formula: NAG = NV × SAG/NT (NAG is the number of AB or GIB (CFU mL −1 or CFU g −1 wet weight of macroalgae), NV is the number of viable bacteria (CFU mL −1 or CFU g −1 wet weight of macroalgae), SAG is the number of bacterial strains that expressed growth inhibitory properties, NT is the number of bacterial strains examined). All of the GLB were isolated aseptically into a 48-well plate after confirming a pure culture by the streak plate method [48].

16S rRNA Gene Sequencing for Identification
A total of 47 bacterial strains (12 strains from Stn. 1 and 35 strains from Stn. 2) were identified by partial 16S rRNA gene sequencing (500 bp). All strains were cultured in ST10 −1 liquid medium for two weeks then centrifuged (8000 rpm) for 5 min to obtain bacterial pellets. Pellets were suspended in phosphate-buffered saline and washed three times to remove the medium. Bacterial DNA was extracted using NucleoSpin Tissue XS (TaKaRa BIO Inc., Japan) and stored at −20 • C. The universal primers 27F (5 -AGAGTTTGATCMTGGCTCAG-3 [49]) and 519R (5 -GWATTACCGCGGCKGCTG-3 [50]) were applied to amplify the 16S rRNA gene fragments using Blend Taq ® -Plus-(TOYOBO, Japan). The PCR temperature cycling conditions were: initial denaturation at 94 • C for 1 min, followed by 30 cycles of denaturation at 94 • C for 30 s, annealing at 52 • C for 30 s, and extension at 72 • C for 1 min; final elongation was for 7 min at 72 • C. PCR products were purified using ISOSPIN PCR Product (NIPPON GENE CO., LTD., Japan) before sequencing. Both strands of the PCR products were sequenced using the same primers in a cycle sequencing reaction using a sequencing kit (Big Dye Terminator Cycle version 3.1, Applied Biosystems, USA) using a DNA sequencer (ABI 3130, Applied Biosystems, USA). Nucleotide sequences obtained were aligned using Chromas Pro (ver. 1.7.1). Approximately 500 bp of each consensus sequence were analyzed by A BLAST (http://blast.ncbi.nlm.nih.gov/Blast.cgi) sequence similarity search to identify similarity with previously reported bacteria.

Composition of Growth-Limiting Bacteria
The genus and class of GLB identified from different seawater fractions (FLB and PAB), and each of the host macroalgae at both sites are shown in Figure 6. All of the GLB isolated from seawater and  ) and growth-inhibiting bacteria (GIB: ). The tested HAB species and their abbreviations are displayed at the bottom.

Composition of Growth-Limiting Bacteria
The genus and class of GLB identified from different seawater fractions (FLB and PAB), and each of the host macroalgae at both sites are shown in Figure 6. All of the GLB isolated from seawater and macroalgal biofilms were comprised of the classes Alphaproteobacteria, Gammaproteobacteria, and Flavobacteria at Stns. 1 and 2. Seven GLB isolated from seawater at Stn. 1 were classified as the genera of Nautella (Alphaproteobacteria) in FLB, Alteromonas, Pseudoalteromonas (Gammaproteobacteria) and Tenacibaculum (Flavobacteria) in PAB (Figure 6a). In total, five GLB isolated from U. pertusa grown on the pumice stone and the floating cage at Stn. 1 were found to belong to the genera of Pseudoalteromonas, Vibrio (Gammaproteobacteria) and Aquimarina (Flavobacteria), respectively ( Figure  6b,c). The eight GLB detected in seawater at Stn. 2 were one Winogradskyella (Flavobacteria) as FLB, Roseobacter (Alphaproteobacteria), Alteromonas, Pseudoalteromonas and Vibrio (Gammaproteobacteria) and Aquimarina, Winogradskyella (Flavobacteria) as PAB (Figure 6d). The twenty-seven GLB strains isolated from four macroalgal biofilms consisted of two genera from Alphaproteobacteria, Phaeobacter, and  ). The tested HAB species and their abbreviations are displayed at the bottom.

Composition of Growth-Limiting Bacteria
The genus and class of GLB identified from different seawater fractions ( of the host macroalgae at both sites are shown in Figure 6. All of the GLB isol macroalgal biofilms were comprised of the classes Alphaproteobacteria, G Flavobacteria at Stns. 1 and 2. Seven GLB isolated from seawater at Stn. 1 were of Nautella (Alphaproteobacteria) in FLB, Alteromonas, Pseudoalteromonas (Ga Tenacibaculum (Flavobacteria) in PAB (Figure 6a). In total, five GLB isolated fr the pumice stone and the floating cage at Stn. 1 were found to bel Pseudoalteromonas, Vibrio (Gammaproteobacteria) and Aquimarina (Flavobacteri 6b,c). The eight GLB detected in seawater at Stn. 2 were one Winogradskyell ). The tested HAB species and their abbreviations are displayed at the bottom.

Composition of Growth-Limiting Bacteria
The genus and class of GLB identified from different seawater fractions (FLB and PAB), and each of the host macroalgae at both sites are shown in Figure 6. All of the GLB isolated from seawater and macroalgal biofilms were comprised of the classes Alphaproteobacteria, Gammaproteobacteria, and Flavobacteria at Stns. 1 and 2. Seven GLB isolated from seawater at Stn. 1 were classified as the genera of Nautella (Alphaproteobacteria) in FLB, Alteromonas, Pseudoalteromonas (Gammaproteobacteria) and Tenacibaculum (Flavobacteria) in PAB (Figure 6a). In total, five GLB isolated from U. pertusa grown on the pumice stone and the floating cage at Stn. 1 were found to belong to the genera of Pseudoalteromonas, Vibrio (Gammaproteobacteria) and Aquimarina (Flavobacteria), respectively (Figure 6b,c). The eight GLB detected in seawater at Stn. 2 were one Winogradskyella (Flavobacteria) as FLB, Roseobacter (Alphaproteobacteria), Alteromonas, Pseudoalteromonas and Vibrio (Gammaproteobacteria) and Aquimarina, Winogradskyella (Flavobacteria) as PAB (Figure 6d). The twenty-seven GLB strains isolated from four macroalgal biofilms consisted of two genera from Alphaproteobacteria, Phaeobacter, and Roseovarius, and three genera from Gammaproteobacteria, Alteromonas, Pseudoalteromonas, and Vibrio plus three genera of Flavobacteria, Aquimarina, Dokdonia and Winogradskyella (Figure 6e-h). About 67% of GLB found from four macroalgal biofilms at the natural bed (Stn. 2) were the bacteria classified into the class Flavobacteria. Aquimarina was the only GLB found from all four macroalgae at the natural bed (Stn. 2).
Appl. Sci. 2020, 10, x FOR PEER REVIEW 9 of 16 Roseovarius, and three genera from Gammaproteobacteria, Alteromonas, Pseudoalteromonas, and Vibrio plus three genera of Flavobacteria, Aquimarina, Dokdonia and Winogradskyella (Figure 6e,f,g and h). About 67% of GLB found from four macroalgal biofilms at the natural bed (Stn. 2) were the bacteria classified into the class Flavobacteria. Aquimarina was the only GLB found from all four macroalgae at the natural bed (Stn. 2).

Discussion
Recent findings of growth-limiting bacteria densely living in a biofilm of seagrass and macroalgae have offered an insight into possible future HAB control by harnessing ecosystem services [51]. Meanwhile, the loss of seagrass and macroalgal beds due to human impacts have been accelerated over the last half-century [52][53][54]. Such deterioration of coastal vegetation coincided with the increasing occurrences of HAB [51,55]. The increasing toxic blooms of the dinoflagellate A. minutum accompanied by the large-scale decline of seagrass beds around the Mediterranean coast is one of the specific cases [56]. Inaba et al. [37] documented that H. akashiwo cysts were found only from the sediments where eelgrass had disappeared and studied sites where the abundant growth-limiting bacteria detected rarely experienced blooms of paralytic shellfish toxin-producing dinoflagellate Alexandrium in Puget Sound, WA, USA. On the positive side, the efforts to protect and restore seagrass and macroalgal beds are underway worldwide [57,58]; for instance, vigorous restoration activities of seagrass bed in Hinase, Okayama, Japan increased the natural bed more than twenty times since 1985 [57]. The efforts will enhance the resistance of coastal areas against HAB events in the long run. However, this holistic approach to prevent HABs may not be as applicable as direct controls, such as the use of flocculants [59]. Thus, it is crucial to develop an environmentally friendly measure that can be widely applicable and cost-effective.
Ulva can be attractive candidate organisms revealed to harbor GLB [34,37], growing fast compared to other macroalgae or seagrass [60,61], relatively easy to culture [62], restoring water quality through the uptake of excess nutrients [63,64], releasing allelochemicals negatively affecting the growth of HAB species [65,66], and having commercial importance in that they are frequently consumed under the name "Aonori" in Asia [67,68]. Several Ulva species have already been tested to be organisms composing large scale integrated multi-trophic aquaculture (IMTA) systems [67,69]. In the present study, harmful algal growth-limiting bacteria (GLB) were found in the order of 10 6 CFU per 1 g of U. pertusa grown on the pumice stone and the mobile floating cage at Stn. 1, showing the equivalent value to the natural growing macroalgal species at Stn. 2 as well as previous studies estimating the densities of GLB on Ulva sp. and Gelidium sp. in Osaka Bay, Japan [34] and on U. lactuca in Puget Sound, USA [37]. To our best knowledge, this is the first report to reveal that algal growth-limiting bacteria can be harbored within a few months on the newly introduced Ulva thallus, suggesting the possible future field application to control HABs by introducing Ulva beds as a source of GLB.
Previous studies have shown that a higher proportion of GLB was often associated with particles [28,44]. In the study in Puget Sound, most isolated bacteria from seawater having more extensive activity ranges were in the particle-associated form [37]. In the present study, the density of PAB (Particle-associated bacteria) as GLB was higher than FLB (Free-living bacteria) at the natural macroalgal bed (Stn. 2), killing most of the tested HAB species except A. catenella (Figure 5b). Moreover, 88% of GLB detected from seawater at the natural macroalgal bed was the same bacterial genera as GLB detected from macroalgal biofilms ( Figure 6). It is known that macroalgal beds are one of the essential sources releasing particulate and dissolved organic matter (POM and DOM) to the surrounding environments [70][71][72]. The POM derived from the macroalgae at Stn. 2 may have increased substrate availability for specific bacterial communities with high exoenzymatic activities [73][74][75] that resulted in frequent detection of GLB as PAB forms in the water. This could be reasonably explained by highly dense GLB harbored on a macroalgal surface being provided synchronously with macroalgal-derived POM to the surrounding water.
Many bacteria previously reported as algicidal and/or growth-inhibiting bacteria were among the Cytophaga/Flavobacterium/Bacteroidetes (CFB) group [27]. They are also critical players for the degradation of organic matter in aquatic environments [76]. Ecologically important roles toward macroalgae of the CFB group have been demonstrated, such as controlling morphogenesis [77,78], promotion and/or inhibition of spore germination, and colonization [79,80], close association with diseases [81], etc. In the present study, more than half of GLB were identified as bacteria belonging to the class Flavobacteria. Notably, the genus Aquimarina within the class Flavobacteria were found from all of the macroalgae and 25% of GLB in seawater samples at Stn. 2 ( Figure 6). With regard to the growth-limiting activity range, all GLB isolated from seawater and Ulva samples at Stn. 1 were found to exhibit specific activities in that 92% of GLB negatively affected the growth of only one HAB species (C. antiqua or H. circularisquama) ( Figure 7a); however, GLB isolated from natural macroalgae bed showed a much broader range of activities on tested HAB species (Figure 7b). About 90% of GLB affecting the growth of more than one HAB species were found to be Flavobacteria, which consisted of Aquimarina and Dokdonia (Figure 7b). Aquimarina isolated from red alga on the coast of China is shown to exhibit diverse agarase activities, enhancing the degradation of host algal tissue [82]. These results suggest the multiple roles of the members of Flavobacteria diffusing around macroalgal beds. The physiological and biochemical characteristics of macroalgae predetermine the composition of their biofilm microbial communities [83], and the host species-specific microbial association in aquatic angiosperms is known [84]. It is worth noting that natural macroalgal beds with higher diversity may offer various inhabiting substrates for diverse GLB, functioning effectively as bloom control for multiple harmful algal species (Figure 7b).
Appl. Sci. 2020, 10, x FOR PEER REVIEW 11 of 16 growth-limiting activity range, all GLB isolated from seawater and Ulva samples at Stn. 1 were found to exhibit specific activities in that 92% of GLB negatively affected the growth of only one HAB species (C. antiqua or H. circularisquama) ( Figure 7a); however, GLB isolated from natural macroalgae bed showed a much broader range of activities on tested HAB species (Figure 7b). About 90% of GLB affecting the growth of more than one HAB species were found to be Flavobacteria, which consisted of Aquimarina and Dokdonia (Figure 7b). Aquimarina isolated from red alga on the coast of China is shown to exhibit diverse agarase activities, enhancing the degradation of host algal tissue [82]. These results suggest the multiple roles of the members of Flavobacteria diffusing around macroalgal beds. The physiological and biochemical characteristics of macroalgae predetermine the composition of their biofilm microbial communities [83], and the host species-specific microbial association in aquatic angiosperms is known [84]. It is worth noting that natural macroalgal beds with higher diversity may offer various inhabiting substrates for diverse GLB, functioning effectively as bloom control for multiple harmful algal species (Figure 7b).

Conclusions
The present study illustrated that GLB with equivalent value as naturally growing macroalgae could be harbored on the surface of artificially introduced U. pertusa for less than three months, providing novel insight into the potential use of the artificial introduction of U. pertusa beds as biological control of HABs. Results from the natural macroalgal bed suggested that higher macroalgal diversity provide diverse inhabiting substrates for GLB, functioning effectively as bloom control for multiple harmful algal species. It should be also emphasized that the protection and restoration of natural macroalgal beds may serve as not only critical habitats for numerous marine and estuary organisms but also one of the original hotspots of GLB, contributing to mitigating HABs in coastal environments. Further investigations to understand the diffusive behavior of GLB are needed to achieve the practical application.

Conclusions
The present study illustrated that GLB with equivalent value as naturally growing macroalgae could be harbored on the surface of artificially introduced U. pertusa for less than three months, providing novel insight into the potential use of the artificial introduction of U. pertusa beds as biological control of HABs. Results from the natural macroalgal bed suggested that higher macroalgal diversity provide diverse inhabiting substrates for GLB, functioning effectively as bloom control for multiple harmful algal species. It should be also emphasized that the protection and restoration of natural macroalgal beds may serve as not only critical habitats for numerous marine and estuary organisms but also one of the original hotspots of GLB, contributing to mitigating HABs in coastal environments. Further investigations to understand the diffusive behavior of GLB are needed to achieve the practical application.