1 INTRODUCTION

With deterioration of the global environment caused by anthropogenic sources of pollution, harmful algal blooms, particularly those comprising red tide algae, have frequently occurred in ocean bodies worldwide. In China, these algal blooms cover a total area of several thousand square kilometers every year, which have been destructive to the local mariculture. Current techniques for controlling red tide algae are based on physical methods, such as ultrasonic waves, chemical methods, such as hydroxyl radical treatment, or biological methods, such as bacterial treatment(Yu et al., 2004; Zhou et al., 2004; Pei et al., 2007). Biological methods are being increasingly favored as being most cost-effective and environmentally friendly(Cross, 2003; Tang et al., 2011).

Many macroalgae have been found to exert algicidal effects on red tide algae(Jeong et al., 2000; Nan et al., 2008; Laabir et al., 2013). Of particular significance are the green algae Ulva fasciata, U . linza, and Enteromorpha clathrata(Alamsjah et al., 2005; Xu et al., 2005). U . intestinalis(Enteromorpha intestinalis)belong to the phylum Chlorophyta, class Ulvophyceae, family Ulvaceae, and are dominant in the Yellow Sea and East China Sea. This genus exhibits excellent fecundity and rapid growth under optimal environmental conditions(Baeck et al., 2000; Kim et al., 2010; Wang et al., 2012). For example, an outbreak of green tide comprising Enteromorpha species was observed offthe coast of Qingdao in 2008 (Ye et al., 2008). Previous studies have shown that seaweeds might selectively inhibit harmful microalgae through excretion of algicidal compounds, including polyunsaturated fatty acids, hexadeca-4, 7, 10, 13- tetraenoic acid(HDTA), and octadeca-6, 9, 12, 15- tetraenoic acid(ODTA), as well as some types of peptides(Alamsjah et al., 2005; Park et al., 2011). Considering that some macroalgae inhibit growth of harmful algae through algicidal effects instead of nutritional competition, if these macroalgae are present in large amounts, they might represent a potentially important factor for controlling red tide algae. The purpose of this study was to screen algicidal activity within various macroalgae species and, specifically, to investigate the inhibitory effects of U . intestinalis on two species of native red tide algae, Heterosigma akashiwo and Prorocen t rum micans . In addition, three algicidal compounds were isolated and identified from green algal EtOAc extracts and their inhibition effects were determinated.

2 MATERIAL AND METHOD 2.1 Algal material

Specimens of red tide algae H . akashiwo and P . micans were collected at the Marine Biotechnology Key Laboratory of Ningbo University, China. They were maintained in flasks with f/2 medium(Guillard and Ryther, 1962)at 22°C, 50 μmol/(m² ∙s)(12 L:12 D light period)for 12 d, with the flasks agitated thrice daily to prevent algae from adhering to glass surfaces.

Specimens of twenty selected macroalgal species were collected from Fenghua coast and three islands in Zhejiang, China in March 2006(Table 1). All samples were extracted using 95% ethanol, followed by rotary evaporation, and then the resulting solutions were extracted with water and EtOAc to generate separate aqueous and EtOAc phase extracts. Sterile containers were prepared for algal cultures by autoclaving at 121°C for 20 min.

2.2 Inhibitory effects of macroalgae extracts on red tide algae

Samples of 20 macroalgae species were washed with sterile seawater, dried at room temperature, and sliced into small fragments. Individual 10 g samples were extracted thrice with 95% ethanol for 7 d and the three extracts were combined for further analysis.

During the logarithmic growth phase of H . akashiwo, 0.3 mg/mL of ethanol extracts were added to the growth medium and H . akashiwo cells were counted on the 3rd day. Using experimental and control groups, the inhibitory effects of algal extracts were further assessed on H . akashiwo growth.

2.3 Inhibitory effects of U . intestinalis fresh tissue

Fresh tissue of U . intestinalis was added at 1.0, 2.0, 4.0, 8.0, and 16.0 mg/mL and 2.0, 4.0, 8.0, 12.0, and 16.0 mg/mL to logarithmic growth phase cultures of H . akashiwo and P . micans . Culture samples(1.0 mL) were collected daily and the cells of H . akashiwo and P . micans were measured using an Olympus optical microscope and CASY particle counter, respectively. All experiments were repeated thrice.

2.4 Inhibitory effects of U . intestinalis dry powder

Dry powder of U . intestinalis was added at 0.15, 0.30, 0.60, 1.2 and 2.4 mg/mL and 0.2, 0.4, 0.8, 1.6 and 3.2 mg/mL to H . akashiwo and P . micans cultures as above, respectively. Inhibition effects were assessed as described above.

2.5 Inhibitory effects of U . intestinalis water and EtOAc extracts

Water and EtOAc extracts of U . intestinalis were included at 0, 0.01, 0.02, 0.03, 0.06, and 0.12 mg/mL and 0, 0.01, 0.02, 0.04, 0.08, and 0.16 mg/mL, respectively, in H . akashiwo medium. These extracts were also added to P . micans cultures at 0, 0.06, 0.12, 0.24, 0.48, and 0.96 mg/mL and 0, 0.02, 0.08, 0.16, 0.32, and 0.64 mg/mL, respectively. Inhibition effects were assessed as described above.

2.6 Isolation and purification of bioactive compounds

Bioactive compounds were isolated and purified from 127.0 g of EtOAc extracts(dry weight). Fourteen fractions were produced using silica gel column chromatography, and well eluted in sequence by petroleum ether(100%), then dichloromethane/ methanol(10/1, 5/1, 2/1, and 1/1, v/v). The inhibitory activities of all fractions on red tide algae were then assayed separately. The three fractions displaying the highest algicidal activity were subjected to repetitive Sephadex LH ₂₀ column chromatography, which resulted in the purification of compounds Ⅰ and Ⅱ and compound Ⅲ(TLC developed by dichloromethane/ methanol, 3/1, and dichloromethane/methanol, 8/1, respectively).

2.7 Structural identification of bioactive compounds

Compound structures were determined using a 400- MH Z nuclear magnetic resonance(NMR)spectrometer (Bruker BioSpin Corp., TX, USA), an ultraperformance liquid chromatography system with a time-of-flight high-resolution mass spectrometer (Acquity UPLC, Q-TOF Premier MS; Waters Corp., Milford, MA, USA), and infrared and UV spectroscopy. Chemical structure identifications were validated by comparison with the Chemical Abstracts database.

2.8 Algicidal activity of isolated compounds

Algicidal activity of 10 μg/mL aliquots of the three fractions from above was assessed. In addition, the algicidal activity of the three isolated and purified compounds was tested against H . akashiwo and P . micans to determine the respective IC₅₀ values. Inhibition effect was assessed as described above.

2.9 Statistical analyses

All experiments were repeated thrice and data expressed as mean±standard deviation(SD). Significance of means was evaluated by analysis of variance using the software SPSS(version 16.0). Differences were considered to be significant at P ＜0.05.

3 RESULT 3.1 Inhibitory activity of macroalgae extracts

The growth inhibitions to H . akashiwo growth by 0.3 mg/mL of macroalgal extracts were shown in Table 1. All these extracts inhibited the growth of H . akashiwo to some degree. Five of these, U . intestinalis, U . fasciata, Grateloupia ramosissima, Chondria crassicaulis, and Gracilariopsis lemaneiformis, displayed the strongest inhibitory activities, while Sargassum muticum displayed the weakest.

3.2 The inhibitory effects of U . intestinalis fresh

tissue All groups with added fresh U . intestinalis tissue showed significant inhibitory effects in comparison with the control group(P ＜0.05). Above a certain threshold concentration, fresh tissue strongly decreased the cell densities of H . akashiwo and P . micans, with the degree of inhibition positively correlated with the fresh tissue concentration(Fig. 1Ia and Ⅱa). The threshold concentrations for significant inhibition of H . akashiwo and P . micans growth corresponded to 8.0 and 12.0 mg/mL, respectively.

3.3 The inhibitory effects of U . intestinalis dry powder

Dry powder of U . intestinalis tissue showed inhibitory activities on these red tide algae(Fig. 1). At concentrations of 1.2 and 2.4 mg/mL, H . akashiwo was killed by the 2 nd day of growth. Concentrations greater than 0.6 and 1.6 mg/mL significantly inhibited the growth of H . akashiwo and P . micans, respectively. Thus, the dry powder had stronger inhibitory effects on H . akashiwo than on P . micans .

3.4 Inhibitory effects of U . intestinalis aqueous extracts

Aqueous extracts of U . intestinalis displayed significant inhibitory effects versus the control group (P ＜0.05). It was found that the higher the water extract concentration in the medium, the lower the final densities of H . akashiwo and P . micans by the 7th day of incubation(Fig. 2Ia and Ⅱa). Extract concentrations greater than 0.12 mg/mL consistently inhibited H . akashiwo growth while concentrations at or exceeding 0.96 mg/mL killed P . micans by the fifth day of incubation.

3.5 Inhibitory effects of U . intestinalis EtOAc extracts

These two red tide algae species showed significantly different responses to the presence or absence of EtOAc extracts(P ＜0.05), with U . intestinalis EtOAc extracts displaying the highest inhibitory activities among all extracts against H . akashiwo and P . micans(Fig. 2Ib and Ⅱb). At >0.08 or >0.04 mg/mL of U . intestinalis EtOAc extract, H . akashiwo was killed by the 2nd or 6th day, respectively. However, P . micans mortality occurred at 0.32 mg/ mL EtOAc extract on the 5th day. Therefore, EtOAc extract was more effective against H . akashiwo than against P . micans .

3.6 Identification of algicidal compounds

Preparative column chromatography and analysis of inhibition active fractions by NMR and mass spectroscopy allowed the identification of three compounds(Fig. 3). Compound I was a yellowish and lipophilic, with possible unsaturated bonds due to the presence of an absorption peak at 254 nm. IR (KBr)spectroscopy revealed a hydroxyl-stretching vibration(3 443, 71, broad), a carboxyl signal (1 713.64), and methyl signals(2 919.97, 2 851.53). ¹H-NMR revealed two methyl signals, δ : 0.892(s, 17- CH ₃)and 0.896(s, 18-CH ₃); six alkene hydrogen signals, δ : 5.328(br, 6-CH), 5.347(br, 13-CH), 5.363(br, 10-CH), 5.367(br, 9-CH), and 5.369(br, 7-CH, 12-CH). ¹³ C-NMR analysis revealed one carboxyl carbon signal, δ : 177.67(1-COOH); and six alkene carbon signals, δ : 128.26(7-CH), 128.86(12- CH), 129.04(9-CH), 129.21(10-CH), 130.85(13- CH), and 131.09(6-CH); two oxygenated carbon signals, δ : 64.46(16-CH ₂ -)and 73.87(15-CH-); and nine CH ₂ signals. The molecular weight was 294.207 5 and the formula was deduced to be C ₁₈ H ₃₀ O ₃, using negative ion high-resolution MS m/z. Combining this data, the compound was identified as 15-ethoxy-(6z, 9z, 12z)-hexadecatrienoic acid(Ⅰ).

Compound Ⅱ was also yellowish and lipophilic, but also water-soluble. The presence of unsaturated bonds was deduced from an absorption peak at 254 nm. IR(KBr)analyses revealed a hydroxyl stretching vibration(3 424.54 broad), a carboxylic ester signal(1 734.61)and methyl signal(2 925.25, 2 854.36). ¹ HNMR disclosed that it possessed two methyl signals δ : 0.916(s, 23-CH ₃), 2.088(d, J=5.6Hz, 26-CH ₃), several oxygenated carbon signals, δ : 3.097–4.184 and six alkene hydrogen signals δ : 5.373(br, 14-CH, 15-CH), 5.425(br, 11-CH, 18-CH), 5.574(br, 12-CH, 17-CH). ¹³ CNMR revealed two carboxyl carbon signals δ :171.98(22-COOH); 175.43(20-COOH); six alkene carbon signals, δ : 122.90(12-CH), 129.14(17-CH), 129.41(15-CH), 130.87(14-CH), 130.97(18-CH), and 135.86(11-CH); six oxygenated carbon signals, δ : 64.45(6-CH ₂ -), 69.58(3-CH-), 72.50(2-CH-), 75.15(1-CH-), 77.62(4- CH-), 100.20(5-CH-), and ¹² CH ₂ signals. The molecular weight was revealed to be 482.251 7 and the formula deduced to be C ₂₆ H ₄₂ O ₈, using negative ion high-resolution MS m/z. Based on these observations, the compound was identified as (6E, 9E, 12E)-(2-acetoxy- β - D -glucose)-octadecatrienoic acid ester(Ⅱ).

Compound Ⅲ presented as white flaky crystals soluble in methanol, with no absorption peak at 254 nm. IR(KBr)analysis revealed a hydroxyl stretching vibration(3 448.53)and carboxyl signal (1 702.45). ¹ H-NMR revealed the presence of one methyl signal, δ : 0.880(t, J=6.0Hz, 16-CH ₃); four methylene signals, δ : 1.257(s, 4-CH ₂, 5-CH ₂, 6-CH ₂, 7-CH ₂, 8-CH ₂, 9-CH ₂, 10-CH ₂, 11-CH ₂, 12-CH ₂, 13- CH ₂, and 14-CH ₂), 1.301(s, 15-CH ₂), 1.632(_t, _J=6.8Hz, 3-CH ₂), and 2.347(t, J=7.2Hz, 2-CH ₂); and one carboxyl hydrogen signal, δ : 11.319(s, 1-COOH). ¹³C-NMR revealed one carboxyl carbon signal, δ : 179.97(1-COOH); and a series of alkane carbon signals, δ : 14.13(16-CH ₃), 22.70(15-CH ₂), 24.68(3- CH ₂), 29.07(4-CH ₂), 29.25(5-CH ₂), 29.37(6-CH ₂), 29.44(7-CH ₂), 29.60(8-CH ₂, 9-CH ₂, 10-CH ₂), 29.65(11-CH ₂), 29.68(12-CH ₂), 29.70(13-CH ₂), 31.93(14-CH ₂), and 34.02(2-CH ₂). The molecular weight was calculated to be 256.229 4 and the formula deduced to be C ₁₆ H ₃₂ O ₂, using negative ion highresolution MS m/z. The compound was identified to be hexadecanoic acid(Ⅲ; Meng et al., 1999).

3.7 The algicidal activities of fractions and pure compounds from U . intestinalis

U . intestinalis EtOAc extracts at 20 μg/mL inhibited H . akashiwo and P . micans growth by 32.6% and 12.1%, respectively. Meanwhile, the inhibitory activities of the three column chromatographic fractions at 20 μ g/mL were 52.7%, 69.3%, and 47.6% for H . akashiwo, and 25.4%, 36.7%, and 21.6% for P . micans, respectively. The IC ₅₀ values of compounds I and Ⅱ were 13.4 and 24.7 μ g/mL and 4.9 and 14.1 μ g/ mL for H . akashiwo and P . micans, respectively. Compound Ⅲ displayed relatively weak activity, with IC 50 values of 50.6 and 69.5 μ g/mL, respectively, toward these red tide algae.

4 DISCUSSION

Allelopathy among aquatic plants has been a key marine research focus since it was first studied by Molish in 1937. Currently, several genera of aquatic plants, including Eichhormia, Crassipes, Ulva, and Enteromorpha, ahave been reported to display allelopathic activity, with some allelopathic compounds successfully isolated and identified(Hong and Hu, 2009). In comparison with freshwater plants, research on the allelopathic potential of marine plants towards red tide algae has not been explored in detail, and few algicidal compounds have been purified. For example, the fresh and dried tissues and aqueous extracts of Ulva inhibit growth of Heterosigma , Prorocentrum, and Alexandrium tamarense owing to the production of certain allelopathic compounds(Jin and Dong, 2003; Jin et al., 2005). Similarly, fresh and dried tissues and aqueous extracts of U . pertusa and G . lemaneiformis displayed inhibitory activity against H . akashiwo(Wang et al., 2009), as did Corallina pilulifera against growth of Cochlodinium polykrikoides . Meanwhile, dried tissue from C . pilulifera also inhibits growth of Gymnodinium mikimotoi, G . sanguineum, H . akashiwo, Prorocentrum triestinum, and Pyramimonas sp. (Jeong et al., 2000). The fresh and dry tissues of E . linza also inhibited growth of H . akashiwo(Xu et al., 2005). Among the studied algae, species in the genus Ulva exhibited the most potent inhibitory activity (Wang et al., 2007).

Aquatic allelopathic compounds are generally difficult to identify. Polyunsaturated fatty acids in U . fasciata and U . pertusa play a crucial role in allelopathy; C16:4 n-3, C18:3 n-3, and C18:4 n-3 fatty acids have been detected in high proportions in these species(Alamsjah et al., 2008). In addition, HDTA, α -linolenic acid, and ODTA have been isolated from U . fasciata and proven to strongly inhibit growth of red tide algae, with IC 50 values of 1.35, 1.13, and 0.83 μ g/mL, respectively(Alamsjah et al., 2005). Furthermore, two algicidal flavonoid glycosides have been isolated from marine cordgrass, Spartina anglica, inhibiting H . akashiwo growth by 61.20% and 73.20%, respectively(Xu et al., 2009). In this study, fresh and dry powder and aqueous and EtOAc extracts of U . intestinalis were found to inhibit H . akashiwo and P . micans, with the EtOAc extract showing the most potent activity. Compound Ⅱ, isolated from the EtOAc extract, displayed the most potent algicidal activity, with an IC ₅₀ value of 4.9 and 14.1 μ g/mL for H . akashiwo and P . micans, respectively, which were lower than values reported for HDTA, α -linolenic acid, and ODTA from Ulva species.

Methods for controlling red tide algae must be economically viable, highly effective, and ecologically safe. Compared with traditional physical and chemical methods, those based on plant allelopathy have excellent potential(Yang et al., 2008). Advantages of using Ulva species(aka Enteromorpha in old literature)to inhibit red tide algae are their rapid growth and resource-intensive metabolism, which can thus relieve seawater eutrophication through nutrient consumption. However, to avoid Ulva green tide outbreaks, the use of Ulva species as biocontrol agents should be undertaken carefully. As U . intestinalis involves growth inhibitory compounds and not nutrient competition, the following aspects should be addressed in future studies of possibly useful agents from U . intestinalis . First, other algicidal compounds not isolated and identified in the present study need to be examined and described. Second, the mechanism by which these compounds inhibit red tide algal growth should be explored.

5 CONCLUSION

In this study, twenty species of seaweed were screened for their algicidal activity against the red tide microalgae Heterosigma akashiwo and Prorocen t rummicans . Five species, U . intestinalis, U . fasciata, Grateloupia r a mosissima, Chondria crassicaulis, and Gracilariopsis lemaneiformis, were shown to possess potent algicidal activity. The red tide algal inhibitory effects of the green alga U . intestinalis were assessed and three main algicidal compounds were isolated, purified, and identified. These findings suggested that certain macroalgae could be used as effective biological control agents against red tide algae.