Bloom dynamics and chemical defenses of benthic cyanobacteria in the Indian River Lagoon, Florida

Harmful Algae 69 (2017) 75–82

Contents lists available at ScienceDirect

Harmful Algae journal homepage: www.elsevier.com/locate/hal

Bloom dynamics and chemical defenses of benthic cyanobacteria in the Indian River Lagoon, Florida Jennifer M. Sneeda,* , Theresa Meicklea , Niclas Engenea,b , Sherry Reeda , Sarath Gunasekeraa , Valerie J. Paula a b

Smithsonian Marine Station at Fort Pierce, 701 Seaway Dr., Ft. Pierce, FL 34949, USA Department of Biological Sciences, Florida International University, Miami, FL 33199, USA

A R T I C L E I N F O

Article history: Received 20 June 2017 Received in revised form 3 October 2017 Accepted 6 October 2017 Available online xxx Keywords: Lyngbya sp. Okeania erythroflocculosa Filamentous cyanobacteria Lyngbyoic acid Malyngolide Microcolin

A B S T R A C T

Cyanobacterial blooms are predicted to become more prominent in the future as a result of increasing seawater temperatures and the continued addition of nutrients to coastal waters. Many benthic marine cyanobacteria have potent chemical defenses that protect them from top down pressures and contribute to the persistence of blooms. Blooms of benthic cyanobacteria have been observed along the coast of Florida and within the Indian River Lagoon (IRL), a biodiverse estuary system that spans 250 km along Florida’s east coast. In this study, the cyanobacterial bloom progression at three sites within the central IRL was monitored over the course of two summers. The blooms consisted of four unique cyanobacterial species, including the recently described Okeania erythroflocculosa. The cyanobacteria produced a range of known bioactive compounds including malyngolide, lyngbyoic acid, microcolins A–B, and desacetylmicrocolin B. Ecologically-relevant assays showed that malyngolide inhibited the growth of marine fungi (Dendryphiella salina and Lindra thalassiae); microcolins A–B and desacetylmicrocolin B inhibited feeding by a generalist herbivore, the sea urchin Lytechinus variegatus; and lyngbyoic acid inhibited fungal growth and herbivore feeding. These chemical defenses likely contribute to the persistence of cyanobacterial blooms in the IRL during the summer growing period. © 2017 Published by Elsevier B.V.

1. Introduction Localized, episodic surges in the growth (blooms) of benthic marine cyanobacteria have been reported in tropical and subtropical zones worldwide (Paul et al., 2005; Ahern et al., 2007; Martin-Garcia et al., 2014; Yamashiro et al., 2014). These blooms can have negative impacts on benthic organisms and cause significant detrimental effects on ecosystem health (Watkinson et al., 2005). Cyanobacterial blooms are often associated with high nutrients and are predicted to become more pervasive as seawater temperatures increase due to climate change (Paul et al., 2005; Paerl et al., 2008; O’Neil et al., 2012; Paerl and Paul, 2012). Given optimal growth conditions (high nutrients, warm temperatures), cyanobacterial blooms will persist unless controlled by biotic factors such as grazing or infection (bacterial, fungal and/or viral) (Paerl and Otten, 2013). Saprotrophic and pathogenic fungi are prevalent among marine macrophytes and can act as driving factors in the collapse of cyanobacterial blooms (Hyde et al., 1998;

* Corresponding author. E-mail address: [email protected] (J.M. Sneed). https://doi.org/10.1016/j.hal.2017.10.002 1568-9883/© 2017 Published by Elsevier B.V.

Gerphagnon et al., 2013). Many cyanobacterial strains, however, persist in the environment by producing highly bioactive secondary metabolites that protect them from normal top-down controls (Capper et al., 2006; Soares et al., 2015). Marine benthic cyanobacteria include a variety of filamentous species that are often chemically rich and grow attached to seagrasses, corals, macroalgae, and sediment. They are found in subtropical and tropical coastal waters worldwide and can form dense blooms especially during warm summer months (Paul et al., 2005; Watkinson et al., 2005). The morphological similarities among these cyanobacteria make taxonomic identification particularly difficult, and as a result many cryptic species have been overlooked and grouped together with other taxa based on phenotypic similarities (Engene et al., 2013a). Much of the previous research on benthic cyanobacterial blooms has focused on species that were reported to belong to the genus Lyngbya. Certain species in this genus inhibit settlement and increase larval mortality in scleractinian corals and octocorals (Kuffner et al., 2006), reduce growth and photosynthetic efficiency in adult corals (Titlyanov et al., 2007), and are correlated with increased tumor growth when ingested by green sea turtles (Arthur et al., 2008). Blooms of Lyngbya spp. have also been associated with changes in benthic

76

J.M. Sneed et al. / Harmful Algae 69 (2017) 75–82

invertebrate assemblages and depth distributions within the underlying sediment (Garcia and Johnstone, 2006; Estrella et al., 2011). Recent molecular investigation into Lyngbya spp. resulted in the reclassification of some members into the newly described genera Moorea and Okeania (Engene et al., 2012, 2013b), and the taxonomy of this group needs further revision based on molecular data (Engene et al., 2013a,b). Seasonally reoccurring benthic cyanobacterial blooms have become common in southeastern Florida, especially along the south Florida reef tract (Paul et al., 2005). Reports of benthic cyanobacteria within the Indian River Lagoon (IRL) suggest that similarly persistent blooms may have taken hold here as well (Capper and Paul, 2008; Capper et al., 2016; Tiling and Proffitt, 2017). The IRL is a highly diverse estuary that extends along 40% of the east coast of Florida (Dawes et al., 1995). For this study, surveys were conducted over two growing seasons at three sites within the IRL to assess the density and composition of benthic cyanobacterial blooms within the lagoon. The production of feeding deterrent and antifungal compounds by all cyanobacteria discovered during the survey were investigated to get a better understanding of the

factors that contribute to the ability of these cyanobacteria to persist and bloom in the Indian River Lagoon. 2. Methods 2.1. Bloom dynamics The percent cover of benthic cyanobacteria was monitored throughout the summer months (April–September) in 2011 and 2012 at three sites in the Indian River Lagoon near Fort Pierce, FL. Site 1 (N 27 27.202, W 80 18.597) was a stand-alone seagrass flat with an area of 157 m  43 m. Site 2 (N 27 28.366, W 80 18.669) was a seagrass flat attached to Little Jim Island (Fort Pierce) with an area of 267 m  66 m. Site 3 (N 27 31.987, W 80 20.907) was a seagrass flat along the coastline near Harbor Branch Oceanographic Institute with an area of 65 m  21 m. In the summer of 2011, each site was monitored once a week during low tide starting on April 14 and continuing until the bloom died off in September. The sites were monitored once every two weeks during the summer of 2012 starting April 18 and ending in September when the bloom

Fig. 1. Mean (SE) percent cover of cyanobacteria at three sites within the Indian River Lagoon over the course of two summers.

J.M. Sneed et al. / Harmful Algae 69 (2017) 75–82

died off. One m2 quadrats were divided into 100 10 cm2 sections. Quadrats were tossed haphazardly 15 times within each site and percent cover was recorded as the number of sections within each quadrat that contained cyanobacteria. 2.2. Cyanobacterial collection and identification Vouchers of cyanobacteria that looked visually different were collected for molecular identification and stored in RNAlaterTM. Larger, bulk samples were collected for chemical extraction. These samples were cleaned of debris, blotted dry using paper towels and weighed before being frozen and lyophilized. Live or formalin (5% in filtered seawater) preserved samples were used to take microscopic measurements of filament width, cell width, and cell length. For each specimen, DNA was extracted using the Wizard1 Genomic DNA Purification Kit (Promega, Madison, WI, USA), and measured on a ND-1000 spectrophotometer (Thermo Scientific, Waltham, MA, USA). The 16S rRNA genes were PCR-amplified using the universal primers specified in Nübel et al. (1997). The PCR reaction volumes were 25 mL containing: 1 mL (100 ng) of DNA, 5 mL of 5X Green GoTaq1 Flexi buffer, 2.5 mL MgCl2 (10 mM), 0.5 mL (10 mM) of dNTP mix, 1.0 mL of each primer (10 mM), 0.25 mL of GoTaq1 DNA polymerase (5 u mL 1), and 14.75 mL dH2O. The PCR reactions were performed in a DNA Engine Dyad1 Peltier Thermal Cycler (Bio-Rad, Berkeley, CA, USA) as follows: initial denaturation for 2 min at 95  C, 25 cycles of 45 s at 95  C, 45 s at 50  C and 2 min at 72  C, and final elongation for 3 min at 72  C. The PCR products were purified using a MinElute1 PCR Purification Kit (Qiagen, Valencia, CA, USA) before subcloning using the pGEM1-T Easy Vector system (Promega). Plasmid DNA was isolated using the QIAprep1 Spin Miniprep Kit (Qiagen) and sequenced bidirectionally with M13 vector primers as well as the internal primers 359F, 785R, and 1509R (Nübel et al., 1997). All gene sequences are available in the DDBJ/EMBL/GenBank databases under the accession numbers: KY953144- KY953149. The 16S rRNA gene sequences were aligned with a total of 46 cyanobacterial specimens, including Gloeobacter violaceus PCC 7421R as outgroup, using the MUSCLE algorithm (Edgar, 2004). Nucleotide substitution models were selected using uncorrected/corrected Akaike Information Criterion, Bayesian Information Criterion, and the Decision-theoretic in jModeltest 0.1.1 (Posada, 2008). The Maximum likelihood (ML) inference was performed using PhyML (Guindon and Gascuel, 2003). The analysis was run using the K80 model ( Kimura, 1980) assuming heterogeneous substitution rates and gamma substitution of variable sites (proportion of invariable sites (pINV) = 0.449, shape parameter (a) = 0.477, number of rate categories = 4). Bootstrap resampling was performed on 1000 replicates. Bayesian analysis was conducted using MrBayes 3.1 ( Ronquist and Huelsenbeck, 2003). Four Metropolis-coupled MCMC chains (one cold and three heated) were run for 10,000,000 generations and the first 100,000 generations were discarded as

77

burn-in and the following data sets were sampled with a frequency of every 100 generations. 2.3. Chemical analysis Lyophilized samples were extracted exhaustively in 1:1 ethyl acetate/methanol. Extracts were separated into ethyl acetate and butanol partitions using liquid-liquid partitioning. Solvents were removed via rotary evaporation and speed-vac. Partitions were examined using nuclear magnetic resonance (NMR) spectroscopy for the presence of known compounds. Major compounds from each cyanobacterium were isolated and purified using column and high-performance liquid chromatography. 2.4. Feeding assays Feeding assays were performed in flow-through raceways at the Smithsonian Marine Station using the sea urchin Lytechinus variegatus collected from the Lake Worth Lagoon, FL following the methods outlined by Hay et al. (1998) and Capper et al. (2016). This sea urchin was selected because it is the most common urchin in seagrass habitats where the cyanobacterial blooms occurred. Extract partitions and pure compounds were dissolved in either ethyl acetate or ethanol, depending on polarity, and 500 mL added to 2 g of lyophilized, ground Ulva sp. Solvent controls were prepared with 500 mL of ethyl acetate or ethanol. All extract partitions were added at natural dry weight concentrations (Supplementary Table 1). Concentrations of pure compounds used were calculated based on the amount of compound isolated from cyanobacteria collected during this study, except for malyngolide, which was based on a previous collection and is within the natural range reported in the literature (Meickle, 2010). If a pure compound was not active at natural concentration, it was tested at 2x natural concentration. If a pure compound was active at natural concentration, it was tested at 0.5x natural concentration. Solvent was allowed to evaporate, and treatment and control coated food was mixed with 10 mL distilled water to form a slurry. 1 g agar was mixed with 20 mL distilled water and heated to boiling. When the agar solution cooled to approximately 68  C, the Ulva sp. slurries were mixed into the agar and poured into molds containing window screening. Once cooled, the window screening was cut so that each feeding strip contained a treatment and control square, each approximately 4 cm2. During feeding assays, urchins were confined in plastic containers with holes to allow water flow so that each container contained one urchin and one feeding strip that contained both treatment and control food. Urchins were allowed to feed on the strips for 2–5 h, until at least half of the control squares had been eaten. Replicates were only included in the analysis if the amount of total food eaten was >10% and <90%. The number of window screens squares consumed was recorded and treatments were compared to controls using a paired t-test with Statistix 9.

Table 1 Collection information, morphological measurements and major secondary metabolites for the six cyanobacteria collected. Strain

GenBank Acc. Nr.

Collection Date (2011)

Collection Site

Filament width (mm)

Cell width (mm)

Cell length (mm)

Major compound

IRL11-1

KY953144

14 April

Site 1

44.0  1.7

35.5  0.9

3.6  0.4

IRL11-2 IRL11-3 IRL11-4 IRL11-5 IRL11-6

KY953145 KY953146 KY953147 KY953148 KY953149

14 April 14 April 20 April 25 July 14 April

Site Site Site Site Site

31.0  1.0 32.0  0.9 31.3  0.9 33.0  0.5 n.d

27.5  1.4 26.5  1.0 30.4  0.8 31.5  0.6 10.5  0.3

3.0  0.2 3.9  0.5 3.8  0.5 3.1  0.4 3.0  0.3

microcolins A, B, & desacetylmicrocolin B lyngbyoic acid lyngbyoic acid malyngolide malyngolide n.a.

1 1 2 3 1

78

J.M. Sneed et al. / Harmful Algae 69 (2017) 75–82

Fig. 2. Phylogenetic relationship of cyanobacterial strains collected in this study to each other and other known strains. The phylogram is based on nucleotide sequences of 16S (rRNA) ribosomal genes. The support values are indicated as posterior probability for Bayesian inference (MrBayes) and bootstrap for maximum-likelihood (PhyML). The scale bar is indicated at 0.04 expected nucleotide substitutions per site using the GTR + I + G substitution model. The specimens are indicated as species, strains, and GenBank accession numbers in brackets.

2.5. Antifungal assay Antifungal assays were performed following the methods outlined by Kubanek et al. (2003). Marine fungal strains (Lindra thalassiae MP005 and Dendryphiella salina EBGJ syn. Scolecobasidium salinum) (Clipson et al., 2001) were grown on yeast (1 g L 1), peptone (1 g L 1), mannitol (2 g L 1) (YPM) agar supplemented with penicillin and streptomycin (250 mg L 1 each). Ethyl acetate and butanol partitions of organic extracts and isolated compounds were dissolved in 400 mL methanol and added to 4 mL molten YPM P/S agar at natural volumetric concentration (Supplementary Table 1). Agar solutions were pipetted into sterile 24 well plates so that each treatment filled 3 wells per plate with 600 mL per well. Duplicate plates were prepared so that n = 6 for each treatment. Each plate also contained a methanol only solvent control, a control with no solvent, and a positive control with the known antifungal agent, amphotericin B (Sigma Aldrich, 6.25 mg mL 1). Once the agar had solidified, 0.5 mm2 squares of fungal cultures were cut from agar plates and transferred to the center of each well. Fungi were incubated at 26  C until fungi had completely overgrown control wells (5–12 days). Area of growth inhibition

was measured by counting the number of window screen grids within the well that contained fungal hyphae to obtain proportion inhibition ([mean solvent control treatment]/mean solvent control; n = 6). Data were arcsin square root transformed and treatments were compared using a one-way analysis of variance (ANOVA) or a Kruskal-Wallis one-way ANOVA on ranks if data did not meet the required assumptions for parametric tests. Significant differences (p < 0.05) from solvent controls were determined using a Dunnett’s multiple comparisons test. All statistical analyses were performed using SigmaPlot. 3. Results 3.1. Bloom dynamics In both 2011 and 2012, the bloom had already started at Site 2 in early April when monitoring began (Fig. 1). This site consistently had the most cyanobacterial coverage of the three sites monitored and the bloom consisted entirely of one cyanobacterium (cf. Lyngbya majuscula strain IRL11-4). In both years, the bloom reached peak coverage by mid to late May (92  2.6% on May 17, 2011 and

J.M. Sneed et al. / Harmful Algae 69 (2017) 75–82

79

June 21, 2011 and 20  8.0% on August 9, 2012. In 2011 the bloom was completely gone from this site in early August. In 2012 the bloom persisted until early September. 3.2. Identification of cyanobacteria Phylogenetic analysis of the strains collected along with morphological data (Table 1, Fig. 2) indicate that IRL11-4 and IRL11-5 strains are closely related and the IRL11-2 and IRL11-3 strains are also closely related. IRL11-1 has been recently classified as Okeania erythroflocculosa (Engene et al., 2013b). The IRL11-6 strain was the most distantly related from the other strains. 3.3. Bioassays

Fig. 3. Mean (SE) number of food squares consumed by a generalist herbivore (Lytechinus variegatus) in paired choice assays between food containing ethyl acetate (A) or butanol (B) partitions of cyanobacterial extracts (treatment: gray bars) and solvent only control (control: white bars). Asterisks indicate significant differences between treatment and control food based on a paired t-test.

65  8.3% on May 29, 2012). After the peak, the blooms crashed in June of each year. On June 7, 2011 the average coverage at Site 2 was 2.3  2.0%. In 2012, cyanobacterial coverage was completely gone by June 29. This crash was followed by a resurgence that peaked on August 9 of both years but at a lower level than the May peak (49  11% in 2011, 45  9.6% in 2012). The bloom persisted through the beginning of September in both years at which time it died off for the winter months. The bloom dynamics were much more variable at the other two sites. At Site 1, the bloom was well underway when monitoring began in April of 2011. The bloom at this site peaked on April 27 with an average coverage of 78  5.4% and was nearly completely gone by May 13 (0.13  0.13% average coverage). This bloom consisted of a mixture of cyanobacteria identified as Okeania erythroflocculosa (strain IRL11-1), unidentified strain IRL11-6, and cf. Lyngbya majuscula (strains IRL11-2 and IRL11-3). The bloom was effectively absent in 2012. There was only one monitoring day in 2012 (June 11) in which cyanobacteria were present at this site and only at an average of 0.67  0.36%. Site 3 was dominated by one cyanobacterium (cf. Lyngbya majuscula strain IRL11-5), which began blooming on June 14, 2011 and May 29, 2012. The bloom at this site was never as dense as that at Site 2. The highest percent cover for this site was 27  12% on

The cyanobacterial extracts displayed a range of activities in bioassays. In general, the ethyl acetate (more non-polar) partitions elicited a greater response in all bioassays. The ethyl acetate partitions of IRL11-1 (p < 0.001), IRL11-2 (p < 0.001), IRL11-3 (p = 0.001), and IRL11-4 (p = 0.007) (paired t-tests) all significantly reduced the amount of food consumed by Lytechinus variegatus compared with control food (Fig. 3). For IRL11-2, the butanol (more polar) partition also inhibited feeding by L. variegatus (Fig. 3). No extracts from IRL11-5 or IRL11-6 affected L. variegatus feeding. Ethyl acetate partitions of IRL11-2, IRL11-3, IRL11-4 and IRL11-5 caused a reduction in growth of Lindra thalassiae (p < 0.05, Fig. 4). Ethyl acetate partitions of IRL11-2 and IRL 11-4 also reduced growth in Dendryphiella salina (p < 0.05, Fig. 4). The butanol partition of IRL11-2 was active against D. salina, but not L. thalassiae (p < 0.05). Strain IRL11-6 had no activity in any of the assays performed, and therefore, the chemistry of this strain was not explored. For the other five strains, the chemical diversity correlated with the phylogenetic diversity (Table 1, Fig. 2). Strains IRL11-4 and IRL11-5 both produced malyngolide (Cardellina et al., 1979) as a major compound while IRL11-2 and IRL11-3 both produced lyngbyoic acid (Kwan et al., 2011). Malyngolide is non-polar and was found exclusively in the ethyl acetate partitions. Lyngbyoic acid was split between the ethyl acetate and butanol partitions with a higher proportion of the compound in the ethyl acetate. IRL11-1 (Okeania erythroflocculosa) produced microcolins A-B and desacetylmicrocolin B (Koehn et al., 1992; Meickle et al., 2009), which were found exclusively in the ethyl acetate partition. Natural concentrations of these compounds are reported in Supplementary Table 1. Malyngolide was tested at a range of concentrations from 0.43–2.0 mg g drwt 1 and did not inhibit urchin feeding even at the highest concentration (Fig. 5, p = 0.233), but did inhibit growth of both fungal strains at natural concentration (0.11 mg mL 1, Dunnett Test, p < 0.05, Fig. 4). Lyngbyoic acid significantly reduced growth of both fungal strains at natural concentration (0.73 mg mL 1, p < 0.05, Fig. 4) and inhibited urchin feeding at 5.69 mg g drwt 1 (2x natural concentration, p = 0.002, Fig. 5). Extracts of O. erythroflocculosa (IRL11-1) did not inhibit fungal growth, therefore, microcolins were not tested for antifungal activity. Microcolin A was strongly active against urchin feeding at half its natural concentration (0.2 mg g drwt 1, p < 0.001) and the mixture of microcolin B and desacetylmicrocolin B was active at natural concentration (0.38 mg g drwt 1, p = 0.002, Fig. 5). 4. Discussion Benthic cyanobacterial blooms occur in tropical and subtropical coastal areas worldwide (Paul et al., 2005; Ahern et al., 2007; Martin-Garcia et al., 2014; Yamashiro et al., 2014). They can harm coastal ecosystems and be economically damaging due to

80

J.M. Sneed et al. / Harmful Algae 69 (2017) 75–82

Fig. 4. Mean percent inhibition (SE) of fungal growth relative to solvent controls in response to ethyl acetate (white bars) and butanol (gray bars) partitions of crude cyanobacterial extracts and pure compounds (hashed bars) isolated from active partitions. Asterisks indicate significant differences in growth area between treatments and solvent controls according to Dunnett’s test. N = 6.

reductions in tourism and fishing (Watkinson et al., 2005). Among the three sites surveyed within the central Indian River Lagoon, one site (Site 2) exhibited a consistent, single species bloom that peaked in May, died back in June, and then peaked again in late July/early August (Fig. 1). The presence, composition and density of cyanobacterial cover at the other two sites were more variable. Reports of localized benthic cyanobacterial blooms in the central Indian River Lagoon have been documented for over a decade (Capper and Paul, 2008; Tiling and Proffitt, 2017). Cyanobacterial blooms at Little Jim Island (Site 2) were documented as early as July 2005 and April–July of 2006 (Capper and Paul, 2008) and have been witnessed regularly since then (personal observation). Although they did not quantify bloom coverage, Tiling and Proffitt (2017) describe a large “Lyngbya majuscula” bloom at Harbor Branch Oceanographic Institute occurring from June to September 2006. While cyanobacteria were observed in the present study at Harbor Branch Oceanographic Institute (Site 3), the mean percent cover of the bloom always remained relatively low (<30%). Based on phylogenetic, morphological and chemical analyses, blooms at both Site 2 (Little Jim Island) and Site 3 (Harbor Branch) were composed of a single species. The difference in bloom dynamics at these two sites is likely a result of site specific environmental influences that were not measured here. Slight differences in the bioactivity of the chemical extracts from the cyanobacteria collected at these two sites could also impact bloom dynamics. Ethyl acetate partitions of IRL11-4 (Site 2) extracts significantly deterred urchin feeding while those from IRL11-5 (Site 3) did not

(Fig. 3). Similarly, IRL11-4 (Site 2) extracts were slightly more active against fungi compared to extracts from IRL11-5 (Site 3), which may reflect differences in the concentrations of active compounds (Fig. 4). Both of these strains produced the known compound malyngolide. Malyngolide has been isolated from geographically disparate (Hawaii, Guam, Florida) cf. Lyngbya majuscula samples and has a wide range of ecological activities including feeding deterrence, inhibition of fungal growth and quorum sensing interference (Cardellina et al., 1979; Thacker et al., 1997; Nagle et al., 1998; Dobretsov et al., 2010; Meickle, 2010). In this study, malyngolide inhibited fungal growth at natural concentration (0.11 mg mL 1), but did not affect urchin feeding even at nearly 5x the natural concentration (2.0 mg g drwt 1). This indicates that the feeding deterrent activity elicited by crude extracts of IRL11-4 was either caused by another compound not described here or that this strain contained a higher concentration of malyngolide than tested in this assay. Malyngolide has been shown to deter feeding in fish and sea hares at a 0.4% dry mass but not 0.2% dry mass (Thacker et al., 1997; Nagle et al., 1998). The highest concentration tested here was 2.0 mg g drwt 1 (0.2% dry mass) and was therefore below the minimum active concentration in previous feeding assays (Thacker et al., 1997; Nagle et al., 1998). The presence of malyngolide was confirmed within samples collected from Site 2 and Site 3, however, the concentration of the compound within these samples was not measured. The natural concentration of malyngolide reported here was calculated based on a previous collection of L. majuscula within the IRL and is within the range of concentrations reported for this species (Meickle, 2010). The production of secondary metabolites by cyanobacteria can vary based on environmental conditions (Arthur et al., 2009), and it is likely that the difference in bioactivity of the Little Jim (IRL11-4) and Harbor Branch (IRL11-5) strains are a result of variations in malyngolide concentration, although that was not assessed here. There was a large bloom of mixed cyanobacteria at the Site 1 that peaked in early summer (April) 2011 with a mean percent cover of nearly 80% (Fig. 1). Despite this initially high coverage, the bloom crashed by mid May 2011 and was virtually absent throughout the rest of the study. This bloom consisted of four visually different strains, however, molecular, chemical, and morphological analyses suggested that the IRL11-2 and IRL11-3 collections were the same species (Fig. 2, Table 1). Ethyl acetate partitions of extracts from these two strains strongly deterred urchin feeding and inhibited fungal growth (Figs. 3 and 4). The butanol partition of the IRL11-2 was also active in both bioassays. Microscopic observation of the IRL11-3 strain showed that this collection was a mixture of cyanobacteria and other microscopic algae, which probably led to it looking visually distinct from the IRL11-2 collection. Since bioassays were conducted at natural concentrations based on the weight of the biomass from which the compounds were extracted, the presence of other algae within the IRL11-3 collection may account for the slight decrease in biological activity by causing a dilution of the active compound’s concentration by weight. Lyngbyoic acid was the major compound found in extracts of these two strains. Lyngbyoic acid was first isolated from L. majuscula collected in Florida and has been shown to disrupt quorum sensing in the human pathogen Pseudomonas aeruginosa (Kwan et al., 2011). This compound strongly inhibited the growth of two marine fungal strains at natural concentration found within the ethyl acetate partition of the IRL11-2 collection (Fig. 3), indicating that this compound may function as a chemical defense against environmental fungi. Lyngbyoic acid also inhibited urchin feeding, but at twice the natural concentration (Fig. 5). The ethyl acetate partitions of IRL11-2 and IRL 11-3 inhibited urchin feeding at natural concentration indicating that, in addition to lyngbyoic acid, another compound or compounds are present that

J.M. Sneed et al. / Harmful Algae 69 (2017) 75–82

81

activity in either of the bioassays, and therefore the chemistry of this strain was not further explored. Cyanobacterial blooms are predicted to increase in frequency and intensity as seawater temperatures rise and nutrient loads increase in coastal waters (O’Neil et al., 2012; Paerl and Paul, 2012). Many of these predictions are based on studies of planktonic cyanobacteria, and the roles (especially of seawater temperature) on benthic cyanobacteria are less understood (Watkinson et al., 2005). Benthic cyanobacterial blooms are often associated with areas of anthropogenic nutrient loading and are postulated to be stimulated by high levels of iron, phosphorus, and nitrogen (Albert et al., 2005; Watkinson et al., 2005; Ahern et al., 2007; Arthur et al., 2009). The potent chemical defenses produced by benthic cyanobacteria protect them from feeding and infection, and allelopathic compounds allow them to outcompete other organisms for space and resources (Thacker et al., 1997; Capper et al., 2006, 2016; Dobretsov et al., 2010; Kwan et al., 2011; Yamashiro et al., 2014; Ritson-Williams et al., 2016). Given the right environmental conditions for growth, benthic cyanobacterial blooms will thrive with little top down control by generalist consumers. The antifungal and inhibition of feeding activity demonstrated by the majority of the cyanobacteria found within the seasonal blooms likely help contribute to the persistence of these cyanobacteria in the Indian River Lagoon. Acknowledgements

Fig. 5. Mean (SE) number of food squares consumed by a generalist herbivore (Lytechinus variegatus) in paired choice assays between food containing pure compounds isolated from cyanobacteria (treatment: gray bars) and solvent only control (control: white bars). Asterisks indicate significant differences between treatment and control food based on a paired t-test. Natural concentrations, shown in bold, were determined based on the amount of compound isolated from strains collected in this study with the exception of malyngolide, which is based on a previous collection.

contributed to the feeding deterrence of this strain or alternatively, the natural concentration of lyngbyoic acid within IRL11-2 reported here may be an underestimate of the true natural concentration. The nonpolar extract from the related cyanobacterium IRL 1 (Capper et al., 2016) (Fig. 2) that also contained lyngbyoic acid deterred feeding by reef fishes but had no deterrent effects on feeding by several invertebrate grazers from Florida and Belize, including two other sea urchins, Diadema antillarum and Echinometra lucunter. The IRL11-1 strain from Site 1 was identified as Okeania erythroflocculosa (Engene et al., 2013b) (Fig. 2). Extracts of this strain inhibited urchin feeding, but had no effect on fungal growth. The feeding deterrence was attributed to the major compounds microcolins A and B and desacetylmicrocolin B which also inhibited feeding at or below natural concentrations (Figs. 3 and 5). Unlike the results shown here with the generalist herbivore Lytechinus variegatus, microcolin A does not affect feeding in the specialist sea hares Stylocheilus longicauda or S. striatus (Nagle et al., 1998; Arthur et al., 2009). Microcolin B does inhibit feeding by S. longicauda similar to the feeding deterrent activity demonstrated here with a mixture of microcolin B and desacetylmicrocolin B. Nonpolar extracts of O. erythroflocculosa that contained microcolins also deterred feeding by reef fishes but not the sea urchin Diadema antillarum (Capper et al., 2016). The final strain (IRL11-6) found at Site 1 was morphologically and phylogenetically distinct from the other strains collected during this study (Fig. 2, Table 1). Chemical extracts of this strain had no

This work was supported by funds from the Smithsonian Marine Station and NSF Centers for Ocean Sciences Education Excellence (COSEE FL, OCE-1038990). We would like to thank M. Puglisi for providing fungal strains and J. Craft for collecting sea urchins and assisting with herbivore assays. We thank the Harbor Branch Oceanographic Institute at Florida Atlantic University spectroscopy facility for 600 MHz NMR spectrometer time. We would like to acknowledge K. Demet and L. Kelly for their work on this project as participants in the Research Experience for Preservice Teachers program sponsored by COSEE FL. This is contribution no. 1073 from the Smithsonian Marine Station.[CG] Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at https://doi.org/10.1016/j.hal.2017.10.002. References Ahern, K.S., Ahern, C.R., Savige, G.M., Udy, J.W., 2007. Mapping the distribution, biomass and tissue nutrient levels of a marine benthic cyanobacteria bloom (Lyngbya majuscula). Mar. Freshw. Res. 58 (10), 883–904. Albert, S., O’Neil, J.M., Udy, J.W., Ahern, K.S., O’Sullivan, C.M., Dennison, W.C., 2005. Blooms of the cyanobacterium Lyngbya majuscula in coastal Queensland, Australia: disparate sites, common factors. Mar. Pollut. Bull. 51 (1–4), 428–437. Arthur, K., Limpus, C., Balazs, G., Capper, A., Udy, J., Shaw, G., Keuper-Bennett, U., Bennett, P., 2008. The exposure of green turtles (Chelonia mydas) to tumour promoting compounds produced by the cyanobacterium Lyngbya majuscula and their potential role in the aetiology of fibropapillomatosis. Harmful Algae 7 (1), 114–125. Arthur, K.E., Paul, V.J., Paerl, H.W., O’Neil, J.M., Joyner, J., Meickle, T., 2009. Effects of nutrient enrichment of the cyanobacterium Lyngbya sp on growth, secondary metabolite concentration and feeding by the specialist grazer Stylocheilus striatus. Mar. Ecol. Prog. Ser. 394, 101–110. Capper, A., Paul, V.J., 2008. Grazer interactions with four species of Lyngbya in southeast Florida. Harmful Algae 7 (6), 717–728. Capper, A., Cruz-Rivera, E., Paul, V.J., Tibbetts, I.R., 2006. Chemical deterrence of a marine cyanobacterium against sympatric and non-sympatric consumers. Hydrobiologia 553 (1), 319. Capper, A., Erickson, A.A., Ritson-Williams, R., Becerro, M.A., Arthur, K.A., Paul, V.J., 2016. Palatability and chemical defences of benthic cyanobacteria to a suite of herbivores. J. Exp. Mar. Biol. Ecol. 474, 100–108. Cardellina, J.H., Moore, R.E., Arnold, E.V., Clardy, J., 1979. Structure and absoluteconfiguration of malyngolide, an antibiotic from the marine blue-green-alga Lyngbya majuscula Gomont. J. Org. Chem. 44 (23), 4039–4042.

82

J.M. Sneed et al. / Harmful Algae 69 (2017) 75–82

Clipson, N., Landy, E., Otte, M., 2001. Fungi. In: Costello, M.J., Emblow, C., White, R.J. (Eds.), European Register of Marine Species: A Check-List of the Marine Species in Europe and a Bibliography of Guides to Their Identification. Collection Patrimoines Naturels, Paris, pp. 15–19. Dawes, C.J., Hanisak, D., Kenworthy, W.J., 1995. Seagrass biodiversity in the Indian River Lagoon. Bull. Mar. Sci. 57 (1), 59–66. Dobretsov, S., Teplitski, M., Alagely, A., Gunasekera, S.P., Paul, V.J., 2010. Malyngolide from the cyanobacterium Lyngbya majuscula interferes with quorum sensing circuitry. Environ. Microbiol. Rep. 2 (6), 739–744. Edgar, R.C., 2004. MUSCLE: multiple sequence alignment with high accuracy and high throughput. Nucleic Acids Res 32 (5), 1792–1797. Engene, N., Rottacker, E.C., Kaštovský, J., Byrum, T., Choi, H., Ellisman, M.H., Komárek, J., Gerwick, W.H., 2012. Moorea producens gen. nov., sp. nov. and Moorea bouillonii comb. nov., tropical marine cyanobacteria rich in bioactive secondary metabolites. Int. J. Syst. Evol. Microbiol. 62 (Pt 5), 1171–1178. Engene, N., Gunasekera, S.P., Gerwick, W.H., Paul, V.J., 2013a. Phylogenetic inferences reveal a large extent of novel biodiversity in chemically rich tropical marine cyanobacteria. Appl. Environ. Microbiol. 79 (6), 1882–1888. Engene, N., Paul, V.J., Byrum, T., Gerwick, W.H., Thor, A., Ellisman, M.H., 2013b. Five chemically rich species of tropical marine cyanobacteria of the genus Okeania gen. nov. (Oscillatoriales, Cyanoprokaryota). J. Phycol. 49 (6), 1095–1106. Estrella, S.M., Storey, A.W., Pearson, G., Piersma, T., 2011. Potential effects of Lyngbya majuscula blooms on benthic invertebrate diversity and shorebird foraging ecology at Roebuck Bay, Western Australia. J. R. Soc. West. Aust. 94, 171–179. Garcia, R., Johnstone, R.W., 2006. Effects of Lyngbya majuscula (Cyanophycea) blooms on sediment nutrients and meiofaunal assemblages in seagrass beds in Moreton Bay, Australia. Mar. Freshw. Res. 57 (2), 155–165. Gerphagnon, M., Latour, D., Colombet, J., Sime-Ngando, T., 2013. Fungal parasitism: life cycle, dynamics and impact on cyanobacterial blooms. PLoS One 8 (4), e60894. Guindon, S., Gascuel, O., 2003. A simple, fast, and accurate algorithm to estimate large phylogenies by maximum likelihood. Syst Biol 52 (5), 696–704. Hay, M.E., Stachowicz, J.J., Cruz-Rivera, E., Bullard, S., Deal, M.S., Lindquist, N., 1998. Bioassays with marine and freshwater macroorganisms. In: Haynes, K.F., Millar, J.G. (Eds.), Methods in Chemical Ecology Volume 2: Bioassay Methods. Springer, US, Boston, MA, pp. 39–141. Hyde, K.D., Jones, E.B.G., Leaño, E., Pointing, S.B., Poonyth, A.D., Vrijmoed, L.L.P., 1998. Role of fungi in marine ecosystems. Biodivers. Conserv. 7 (9), 1147–1161. Kimura, M., 1980. A simple method for estimating evolutionary rates of base substitutions through comparative studies of nucleotide-sequences. J Mol Evol 16 (2), 111–120. Koehn, F.E., Longley, R.E., Reed, J.K., 1992. Microcolin A and microcolin B, new immunosuppressive peptides from the blue-green alga Lyngbya majuscula. J. Nat. Prod. 55 (5), 613–619. Kubanek, J., Jensen, P.R., Keifer, P.A., Sullards, M.C., Collins, D.O., Fenical, W., 2003. Seaweed resistance to microbial attack: a targeted chemical defense against marine fungi. Proc. Natl. Acad. Sci. U. S. A. 100 (12), 6916–6921. Kuffner, I.B., Walters, L.J., Becerro, M.A., Paul, V.J., Ritson-Williams, R., Beach, K.S., 2006. Inhibition of coral recruitment by macroalgae and cyanobacteria. Mar. Ecol. Prog. Ser. 323, 107–117. Kwan, J.C., Meickle, T., Ladwa, D., Teplitski, M., Paul, V., Luesch, H., 2011. Lyngbyoic acid, a tagged fatty acid from a marine cyanobacterium, disrupts quorum sensing in Pseudomonas aeruginosa. Mol. Biosyst. 7 (4), 1205–1216.

Martin-Garcia, L., Herrera, R., Moro-Abad, L., Sangil, C., Barquin-Diez, J., 2014. Predicting the potential habitat of the harmful cyanobacteria Lyngbya majuscula in the Canary Islands (Spain). Harmful Algae 34, 76–86. Meickle, T., Matthew, S., Ross, C., Luesch, H., Paul, V., 2009. Bioassay-guided isolation and identification of desacetylmicrocolin B from Lyngbya cf. polychroa. Planta Med. 75 (13), 1427–1430. Meickle, T., 2010. Isolation and structure elucidation of novel compounds from marine cyanobacteria. Chemistry and Biochemistry. Florida Atlantic University p. 130. Nagle, D.G., Camacho, F.T., Paul, V.J., 1998. Dietary preferences of the opisthobranch mollusc Stylocheilus longicauda for secondary metabolites produced by the tropical cyanobacterium Lyngbya majuscula. Mar. Biol. 132 (2), 267–273. Nübel, U., GarciaPichel, F., Muyzer, G., 1997. PCR primers to amplify 16S rRNA genes from cyanobacteria. Appl Environ Microb 63 (8), 3327–3332. O’Neil, J.M., Davis, T.W., Burford, M.A., Gobler, C.J., 2012. The rise of harmful cyanobacteria blooms: the potential roles of eutrophication and climate change. Harmful Algae 14, 313–334. Paerl, H.W., Otten, T.G., 2013. Harmful cyanobacterial blooms: causes, consequences, and controls. Microb. Ecol. 65 (4), 995–1010. Paerl, H.W., Paul, V.J., 2012. Climate change: links to global expansion of harmful cyanobacteria. Water Res. 46 (5), 1349–1363. Paerl, H.W., Joyner, J.J., Joyner, A.R., Arthur, K., Paul, V., O’Neil, J.M., Heil, C.A., 2008. Co-occurrence of dinoflagellate and cyanobacterial harmful algal blooms in southwest Florida coastal waters: dual nutrient (N and P) input controls. Mar. Ecol. Prog. Ser. 371, 143–153. Paul, V.J., Thacker, R.W., Banks, K., Golubic, S., 2005. Benthic cyanobacterial bloom impacts the reefs of South Florida (Broward County, USA). Coral Reefs 24 (4), 693–697. Posada, D., 2008. jModelTest: Phylogenetic model averaging. Mol Biol Evol 25 (7), 1253–1256. Ritson -Williams, R., Ross, C., Paul, V.J., 2016. Elevated temperature and allelopathy impact coral recruitment. PLoS One 11 (12), 16. Ronquist, F., Huelsenbeck, J.P., 2003. MrBayes 3: Bayesian phylogenetic inference under mixed models. Bioinformatics 19 (12), 1572–1574. Soares, A.R., Engene, N., Gunasekera, S.P., Sneed, J.M., Paul, V.J., 2015. Carriebowlinol, an antimicrobial tetrahydroquinolinol from an assemblage of marine cyanobacteria containing a novel taxon. J. Nat. Prod. 78 (3), 534–538. Thacker, R.W., Nagle, D.G., Paul, V.J., 1997. Effects of repeated exposures to marine cyanobacterial secondary metabolites on feeding by juvenile rabbitfish and parrotfish. Mar. Ecol. Prog. Ser. 147, 21–29. Tiling, K., Proffitt, C.E., 2017. Effects of Lyngbya majuscula blooms on the seagrass Halodule wrightii and resident invertebrates. Harmful Algae 62, 104–112. Titlyanov, E.A., Yakovleva, I.M., Titlyanova, T.V., 2007. Interaction between benthic algae (Lyngbya bouillonii, Dictyota dichotoma) and scleractinian coral Porites lutea in direct contact. J. Exp. Mar. Biol. Ecol. 342 (2), 282–291. Watkinson, A.J., O’Neil, J.M., Dennison, W.C., 2005. Ecophysiology of the marine cyanobacterium, Lyngbya majuscula (Oscillatoriaceae) in Moreton Bay, Australia. Harmful Algae 4 (4), 697–715. Yamashiro, H., Isomura, N., Sakai, K., 2014. Bloom of the cyanobacterium Moorea bouillonii on the gorgonian coral Annella reticulata in Japan. Sci. Rep. 4, 6032.