, 2004 ( Ulvophyceae , Chlorophyta ) , a new report and likely nonnative species in the Gulf of Mexico and Atlantic Florida , USA

Species level identifications of morphologically simple marine algae have undoubtedly caused biodiversity assessments to be an arduous task. The green algal genus Ulva L., 1753, is notorious for morphological plasticity and cryptic speciation. We used two chloroplast-encoded (rbcL and tufA) molecular markers and the nuclear internal transcribed spacer 1 (ITS1) of the ribosomal cistron to detect Ulva ohnoi M. Hiraoka and S. Shimada, 2004, a species known for forming green tides in Japan, as a new record for the Western Atlantic, including the Gulf of Mexico (GoMX) and Atlantic coast of Florida. All rbcL sequences from this investigation were identical to reports for U. ohnoi. The Western Atlantic isolates showed relatively low genetic diversity in tufA and ITS1 sequences, which suggests that this species is not native to the GoMX and Atlantic Florida. Furthermore, we have identified U. ohnoi as the species that formed an ephemeral, localized overgrowth during July of 2013 in Biscayne Bay, Florida, an area with a persistent bloom of two other green algal species, Anadyomene stellata J. V. Lamouroux, 1812, and Anadyomene sp., due to eutrophication from anthropogenic nutrient loading near canals. A tissue nutrient analysis of samples from this overgrowth of Ulva showed that this species has a high affinity for nitrogen, especially ∂15N, which suggests anthropogenic sources of N. Further investigations are needed to assess the geographical ranges of this species in this region as well as the potential invasiveness of this alga in the Western Atlantic. It is highly recommended to monitor the abundance of this species in response to nutrient discharges in Biscayne Bay.


Introduction
Accurate biodiversity assessments of marine macrophytes are essential for conservation, monitoring, and management of biological introductions and invasions.Environments with high shipping traffic are constantly under pressure of introductions of non-native species through anthropogenic transportation of, in this case, marine algae through vectors such as fouling on ships (Mineur et al. 2008), ballast water (Flagella et al. 2007;Flagella et al. 2010), and transportation in the aquaculture and aquarium trades (Boudouresque and Verlaque 2002;Booth et al. 2007;Geller et al. 2010).The simple morphology, morphological plasticity, and cryptic diversity of many algal species means introductions of these organisms may go unnoticed.To further complicate the issue, an understanding of native ranges is lacking for many algae.Fortunately, molecular genetics techniques allow for a more complete understanding of the diversity of marine macrophytes and can help unravel native ranges (Geller et al. 2010).
Species of the green algal genus Ulva L., 1753, are some of the most commonly transported organisms through various shipping vectors (Flagella et al. 2007;Mineur et al. 2008;Flagella et al. 2010).Ulva species can survive severe environmental conditions, including over ten months of light deprivation (Santelices et al. 2002).Furthermore, some of these species exhibit invasive traits such as rapid growth, high reproductive potential, vegetative reproduction, and a high affinity to nutrient surplus (Andreakis and Schaffelke 2012).To help determine whether specimens of Ulva are native or nonindigenous species (NIS), previous studies have used the chloroplast-encoded rbcL (Heesch et al. 2009) and tufA (Kirkendale et al. 2013) molecular markers and the nuclear internal transcribed spacer region 1 (ITS1) of the ribosomal cistron (Couceiro et al. 2011).A similar approach was used in this study.
Some of the largest green tides caused by Ulva occur in Qingdao, China, beginning infamously during the 2008 Summer Olympics (Leliaert et al. 2009).In 2008, the bloom in the Yellow Sea covered miles of shoreline and delayed sailing events (Liu et al. 2009;Ye et al. 2011).The species was first identified it as a member of the Ulva linza-prolifera-procera clade or LPP-clade based on rbcL and ITS data (Leliaert et al. 2009).Further analysis of the 5S rDNA and cross breeding tests revealed that the massive growth was caused by Ulva prolifera (Hiraoka et al. 2011).The cost of dealing with this algal bloom was approximately US$100 million (Wang et al. 2009).A green tide the size of that in Qingdao has not been reported in the Western Atlantic; however, blooms   (Hofmann et al. 2010), and three of these blade-forming species (U. compressa, U. lactuca, and U. rigida) were also found blooming in Narrangansett Bay, Rhode Island, USA (Guidone et al. 2013).
The extent and effects of the bloom in Qingdao, China, provided a reason to investigate Ulva incursions in other areas.During the summer of 2013, a localized, ephemeral overgrowth consisting of an unidentified Ulva species with large, freefloating blades was found in Deering Estate, Biscayne Bay, Florida, USA.Blades of approximately 1.0 m length by 0.5 m width were covering an area approximately 500 m long by 4-5 m wide in mangrove habitat (Figure 1).Despite having the lowest water column nutrient concentrations in Florida, parts of Biscayne Bay are subjected to high nutrient loading from canals that discharge into the bay (Caccia and Boyer 2007;Stalker et al. 2009, Swart et al. 2013).Since 2005, there has been a persistent algal bloom of Anadyomene stellata (Wulfen) C. Agardh, 1823, and Anadyomene sp. that covers approximately 60 km 2 of sea grass near the canals, with severe negative effects on the sea grass beds (Collado-Vides et al. 2013).
The goals of this study were to identify the species that formed the overgrowth in Biscayne Bay and explore the geographical distribution of the species elsewhere in the Gulf of Mexico (GoMX) and Atlantic coast of Florida by employing a molecular approach using chloroplast (rbcL and tufA) and nuclear (ITS1) molecular markers.By using the molecular data, we also attempted to determine the native or non-indigenous nature of this species in the Western Atlantic.Tissue nutrient content in algae and seagrass has been used as indicators of nutrient availability in coastal waters (Collado-Vides et al. 2011).Rapidly growing species usually have a high affinity for nitrogen (N) making them opportunistic species in high N availability systems (Lin and Fong 2010).High levels of δ 15 N have been used as signature of anthropogenic source of N (Swart et al. 2013).Tissue nutrient concentrations of three dominant macrophyte species in Biscayne Bay samples were determined and used to evaluate whether excess anthropogenic nutrient inputs enhanced the invasive potential of the Ulva species.

Sampling
Most of the Ulva samples were collected from the intertidal zone during spring and summer months (2012)(2013)(2014)(2015) in the GoMX (Texas, Alabama, and Florida), Atlantic coast of Florida (Biscayne Bay, Indian River Lagoon, and Florida Bay), and Yucatan, Mexico (MX) (Figure 2; Table S1).Herbarium vouchers were made for each sample, and a part of each sample was preserved in silica gel for molecular analyses.Herbarium vouchers were submitted to The University of Alabama Herbarium (UNA) and Fairchild Tropical Botanic Garden (FTBG), and also entered in the online Macroalgal Herbarium Portal (macroalgae.org)where images of the specimens were made available.

DNA extraction and sequencing
This investigation involved molecular analyses of 33 distromatic blades.DNA was extracted from samples desiccated in silica gel using the OMEGA E.Z.N.A. Plant DNA Extraction Kit (Omega Bio-tek Inc., GA, USA).RbcL and ITS1 were amplified by PCR (polymerase chain reaction) with MangoTaq (Bioline US Inc., MA, USA) with the primers in Table 1.For rbcL and ITS1, 3.625 l PCR water, 2.5 l 5x reaction buffer, 0.75 l of 50mM MgCl 2 , 2 l dNTP's, 0.5 l of 10mM forward primer, 0.5 l of 10mM reverse primer, 0.125 l Mango Taq polymerase, 2.5 l of 0.5M Betaine, and 1 l of genomic DNA were used for amplification.The PCR conditions of all primer combinations consisted of an initial denaturation at 94C for 5 min, followed by 35 cycles of 94C for 1 min, annealing at 45C for 1 min, extension at 72C for 2 min, followed by a final extension of 5 min at 72C.The tufA marker was amplified following Saunders and Kucera (2010).The resulting PCR products were visualized by gel electrophoresis.Bands at the appropriate length were cut from the gel and extracted with the OMEGA gel extraction kit (Omega Bio-tek Inc., Georgia, USA).After measuring the concentration of DNA with a NanoDrop ND-1000 Spectrophotometer (Thermo Fisher Scientific, Waltham, MA, USA), DNA products with concentrations of approximately 7 ng/l or greater were sent to AGTC at the University of Kentucky (UK) to be sequenced on a Sanger platform.

Molecular analyses
Sequences were edited and assembled in Geneious R7 (Kearse et al. 2012).A total of 23, 26, and 20 sequences of rbcL, tufA, and ITS1, respectively, from the GoMX and Atlantic Florida were used along with sequences from previously published sequences.GenBank Accession numbers of samples from this study can be found in Table S1.GenBank Accession numbers of previously published Ulva ohnoi sequences can be found in Table 2, and other GenBank Accession numbers, including out-groups, can be found in Table S2.Alignments were made with MUSCLE (Edgar 2004) implemented in Geneious.A maximum likelihood (ML) analysis was performed on each dataset separately with MEGA6 using a GTR + G model of evolution and 1000 bootstrap replicates (Tamura et al. 2013).TCS haplotype networks (Clement et al. 2002) for tufA and ITS1 were made using POPART (Leigh and Bryant 2015).

Nonindigenous species (NIS) investigation
Sequences of Ulva ohnoi from the GoMX and Atlantic coast of Florida were compared to the global genetic diversity of this species to assess the native or nonindigenous nature of this species.We considered three previously used different options for results of this part of the study (as in Heesch et al. 2009;Couceiro et al. 2011;Kirkendale et al. 2013): 1) If the genetic diversity was greater in the GoMX and Atlantic coast of Florida than the global pool, then U. ohnoi from our study area would be considered as a source pool for the global diversity; 2) If the genetic diversity was greater in the global pool, then U. ohnoi would be considered as nonnative to the GoMX and Atlantic Florida; 3) If there was no difference in the genetic diversity in the global pool compared to the GoMX and Atlantic Florida samples, then the native or nonindigenous nature of U. ohnoi in the GoMX and Atlantic Florida would remain questionable until more samples and/or more molecular markers suitable for population genetics can be tested.

Nutrient analysis
Three macrophyte species were collected at Deering Estate, Biscayne Bay, FL, USA, to compare their concentration of nitrogen (N) and phosphorous (P); further, δ 15 N was estimated to detect the potential anthropogenic signature of the incorporated nitrogen by the macrophytes.The seagrass Thalassia testudinum K. D. Koenig, 1805, the dominant species in the area (Lirman et al. 2014), Anadyomene stellata, a persistent blooming species in the area (Collado-Vides et al. 2013), and Ulva ohnoi, an ephemeral but blooming species in the area (this study), were analyzed.Samples in triplicate for each species were collected by hand and stored in a cooler with ice until processed.In the laboratory, each sample was examined for epiphytes and invertebrates, cleaned, dried at 60 °C for 48 h, ground using a Fritsch Analysette 3 Spartan grinder, and stored in 10 ml vials.Elemental P tissue content was measured by the dry-oxidation-acid hydrolysis extraction followed by a colorimetric analysis of phosphate concentration method in the Seagrass Laboratory at Florida International University, and expressed as elemental P% dry weight (P%DW) (Fourqurean et al. 1992).Estimation of 14 N and 15 N was performed using the Thermo Delta C EA-IRMS (elemental analyzer isotope ratio mass spectrometry) system in the Nutrient Analysis Laboratory at Florida International University (Fourqurean et al. 2005).
The EA combusted the organic material and reduced the formed gases into N 2 , which were then measured in a continuous flow mode in the IRMS.The isotopic ratios, or relative abundance of the heavier isotope 15 N to 14 N per sample (R), are reported in the delta standard notation (‰) where R is the ratio of 15 N/ 14 N, and the reference is atmospheric nitrogen (Air N2) (Peterson and Fry 1987).

Statistical analysis
All data were tested for homogeneity of variance using the Levene's test.If homogeneity of variance was met, an analysis of variance (ANOVA) was applied if more than two species were compared, and significant differences between species were detected  Other samples with rbcL sequences 100% identical to the Ulva ohnoi sequences from the Gulf of Mexico and Atlantic Florida not included in the phylogenetic tree include the following collection numbers: TM210, TM145, TM150, TM162, TM182, TM184, TM308, TM316, TM325, TM348, TM366, TM387B, TM391, TM392, TM393, TM232, TM253, TM297.
using a posthoc Tukey test.A t-test was applied to compare means between the two species when homogeneity of variance was met.Three values for each species were recorded for P%DW.Only two values were obtained for the N%DW and δ 15 N‰ analysis of Thalassia testudinum; therefore, no statistical analysis was performed for this species.Three values for Anadyomene stellata and Ulva ohnoi were recorded for N% DW and δ 15 N‰.

Molecular identification
All DNA sequences of rbcL, tufA, and ITS1 corresponded to sequences of Ulva ohnoi M. Hiraoka & S. Shimada, 2004.The specimens of U. ohnoi showed a range of morphologies of free-floating and attached blades, i.e., rosettes, perforated and reticulated blades (Figure 3).Overgrowths of this species were present in Deering Estate, Biscayne Bay, FL, in Florida Bay, FL, and near Brazos Santiago Pass, South Padre Island, TX.
A total of 26 tufA sequences of Ulva ohnoi were recovered from samples from the GoMX (21 sequences) and Atlantic Florida (five sequences).In addition to 23 sequences from previous research in Australia, a 773 bp alignment was formed.The tufA molecular marker was slightly more variable in sequence divergence than rbcL.Three Australian haplotypes were represented by KF195532 (haplotype 1), JN029329 (haplotype 2), and JN029328 (haplotype 3) (Figure 5).Two haplotypes were present in the GoMX and Atlantic Florida sequences.The two haplotypes from this study were identical to the haplotypes represented by KF195532 (haplotype 1) and JN029329 (haplotype 2).Haplotype 2 only varied by a single base pair from haplotype 1 and haplotype 3, while haplotype 1 and 3 differed by two base pairs from each other.A phylogenetic tree of the tufA sequences and their genetic distances are provided as supplemental material (Figure S1 and Table S4, respectively).
An alignment of 47 sequences of ITS1 (178 bp) of Ulva ohnoi from the GoMX and Atlantic Florida (20 sequences) and previous research (Japan: six, Hawaiian Islands: eight, Australia: 12, Italy: one) was used for phylogenetic inference (Figure S2) and to form a haplotype network (Figure 6).This molecular marker was the most variable in this investigation.A total of five haplotypes of ITS1 were found from previous research.All five of these haplotypes were found in 12 sequences from Australia, which were represented by KF195504 (haplotype 1), KF195506 (haplotype 2), KF195495 (haplotype 3), KF195489 (haplotype 4), and KF195496 (haplotype 5).All sequences from Japan (six sequences) and the Hawaiian Islands (eight sequences) were exact matches to haplotype 3 from Australia.Two haplotypes were present in the samples from the GoMX and Atlantic Florida.These two haplotypes were exact matches to the haplotypes represented by KF195495 (haplotype 3) and KF195496 (haplotype 5).Haplotype 3 was present in the GoMX and Atlantic Florida, and haplotype 5 was only found in samples from the northern GoMX.Genetic distances of these samples are provided as supplemental material (Table S5).

Nutrient analysis
Significant differences were found for P%DW between species (ANOVA, F 2,6 = 43.75,p< 0.001) with significantly lower values for Anadyomene stellata compared to the other two species.The means values for Thalassia testudinum and Ulva ohnoi did not differ significantly (Table 3).
Three values of N and δ 15 N‰ were obtained for Ulva ohnoi and Anadyomene stellata, and two values were recorded for Thalassia testudinum.A t-test (df = 4) was applied for U. ohnoi and A. stellata.No significant difference was found for mean N%DW (p > 0.743) between the two species; however, δ 15 N‰ was significantly higher in U. ohnoi (p < 0.007).
The average values for T. testudinum were higher for N%DW and lower for δ 15 N‰ compared to the values of the other two species; however, these differences were not statistically tested due to the small number of samples of T. testudinum (Table 3).

Discussion
Our study reports Ulva ohnoi M. Hiraoka & S. Shimada, 2004, in the Gulf of Mexico and Atlantic coast of Florida for the first time.This species was not previously reported in recent species checklists for this area (Dawes and Mathieson 2008;Littler et al. 2008;Fredericq et al. 2009 Ulva ohnoi was first reported blooming in Tosa Bay, Japan, as "Ulva sp." (Ohno 1988).In 2004 Ulva ohnoi was described from this site based on a molecular study of rbcL and ITS1-5.8S-ITS2, a morphological analysis, and cross breeding tests, which showed that this species is reproductively isolated from two of its closest relatives, Ulva fasciata Delile, 1813, and Ulva reticulata Forsskål, 1775, and also is distinct from the closely related asexual species Ulva spinulosa Okamura & Segawa, 1936(Hiraoka et al. 2004a,b).Subsequent work has also detected U. ohnoi in other localities around Japan (Kawai et al. 2007;Yabe et al. 2009), South Korea (Wang et al. 2010), Australia (Kirkendale et al. 2013;Lawton et al. 2013), and the Hawaiian Islands (O'Kelly et al. 2010).This species has also been identified in ballast water from ships in the Mediterranean Sea; however, colonization by this species has not been detected in this area (Flagella et al. 2010).
Since the genetic diversity of tufA and ITS1 was greater in the global pool than in the GoMX and Atlantic Florida isolates, the hypothesis that Ulva ohnoi is nonnative in the GoMX and Atlantic Florida was supported.Additionally, the phylogenetic trees based on tufA and ITS1 showed that earlier diverging haplotypes of U. ohnoi were not present in the GoMX and Atlantic Florida, which provides further evidence that this species is non-native.Performing molecular work on historical herbarium vouchers that have previously been collected from this area would provide insights into how long this species has been present.The sample collected from Yucatan, Mexico, in 1999 and the broad distribution throughout the GoMX and most of Atlantic Florida indicates the introduction of U. ohnoi to the GoMX and Atlantic Florida is not recent.It would not be surprising if U. ohnoi was widely introduced to the Western Atlantic by a shipping vector given that this species has been documented to be transported in ballast water (Flagella et al. 2007;Flagella et al. 2010).Further studies are needed to investigate the distribution of this species in the Western Atlantic.
Ulva species are notorious for their ability to form green tides (Largo et al. 2004;Leliaert et al. 2009).The green tides of Ulva ohnoi in Japan, reported as Ulva sp. 1 in Hiraoka et al. (2004a), that occurred during the summer months in bays in Kochi and Hakata reached their greatest biomass in mid-August.The localized overgrowth in Biscayne Bay, FL, occurred during July 2013.Rapid and massive growth, the ability to asexually reproduce (including by fragmentation), and opportunistic usage of resources are some characteristics of an invasive species (Andreakis and Schaffelke 2012;Ruesink and Collado-Vides 2006;Geller et al. 2010).It is well documented that species of Ulva are able to metabolize different sources of N from the environment (Teichberg et al. 2007) resulting in the formation of large green tides (Buapet et al. 2008), and can increase their growth rate rapidly as nutrients become available (Pérez-Mayorga et al. 2011).Therefore, Ulva species have a high potential for introduction and invasiveness if they arrive to altered environments with high availability of nutrients.In this study from Biscayne Bay, U. ohnoi showed a high affinity for nutrients, particularly for 15 N, compared with the seagrass Thalassia testudinum, and more importantly, higher than the bloom-forming green alga Anadyomene stellata.
Due to their high affinity for N, Ulva species have been used as indicators of nutrient pollution in coastal waters (Teichberg et al. 2010;Orlandi et al. 2014).Areas affected by anthropogenic nutrient input can have δ 15 N%o values as high as 14.4%o in Boston Harbor, MA, USA, near the sewage effluent from treatment facilities (Tucker et al. 1999), 15%o at Narragansett Bay, RI, USA (Pruell et al. 2006), 13%o in areas affected by sewage and groundwater discharges in east central Florida (Barile 2004), and 10%o from areas that were highly affected by nutrient discharges on the Ivory Coast (Granger et al. 2012).Thus, the high δ 15 N (up to 15%o) of Ulva ohnoi in this study indicates heavy nutrient enrichment of Biscayne Bay, which likely is facilitating the rapid growth of this species.
The ecological and economic effects that green tides can have on coastal ecosystems are well documented (Valiela et al. 1997;Teichberg et al. 2010;Wang et al. 2009).The occurrences of green coenocytic algal blooms, such as Codium isthmocladum Vickers, 1905, and Caulerpa brachypus f. parvifolia (Harvey) A.B. Cribb, 1958, have been reported in the Caribbean and south Florida coral reef systems linked with increase in nutrient availability with several ecological costs (Lapointe 1997;Lapointe et al. 2005;Lapointe and Bedford 2010).On a local basis, the single cell layer blade Anadyomene spp.bloom in Biscayne Bay has already negatively affected seagrass beds due to coverage of up to 75% for more than 10 years making this bloom a major threat to this bay (Collado-Vides et al. 2013).The presence of Ulva ohnoi, a simple twocell layer blade, could add stress to the fragile seagrass ecosystems of the bay.Furthermore, the overgrowths of U. ohnoi near Brazos Santiago Pass, TX, in 2013 and more recently the overgrowth of in Florida Bay, FL, highlight this species' capacity to bloom and the need to be monitored, especially if it becomes invasive.
While Ulva blooms may be considered as a nuisance, many Ulva species have been studied as biofilters (Msuya et al. 2005;Neori et al. 2003;Mata et al. 2010).Ulva ohnoi was specifically identified as a strong suitor to be used in the bioremediation of aquaculture wastewater in Australia due to its fast growth rates and broad geographical distribution (Lawton et al. 2013), and thus, could be an ecologically important species in the Western Atlantic.

Figure 1 .
Figure 1.Ulva ohnoi in Deering Estate, Biscayne Bay, FL, during July 2013.(A) free-floating blades in the intertidal and tangled in mangrove roots and limbs; (B) attached blades near mangroves; and (C) free-floating blade measure about 1.0 × 0.5 m.Photographs by Ligia Collado-Vides.

Figure 2 .
Figure 2. Collection sites of Ulva ohnoi in the GoMX and Atlantic Florida (represented by black dots) made in ArcGIS (ESRI 2009).The collection site of the sample collected from Yucatan, MX, was excluded from the map since the exact collection site is unknown.

Figure 3 .
Figure 3. Variation in forms of Ulva ohnoi collected from the Gulf of Mexico and Atlantic coast of Florida.The scale bars represent 2 cm.Specimens (TM collection numbers) were archived at The University of Alabama Herbarium (UNA) and Fairchild Tropical Botanic Garden (FTBG).The main forms were specimens with: (A) an attached blade with a few marginal reticulations or perforations (TM391); (B) a perforated, free-floating blade (TM291); (C) blades forming a small rosette (TM306); and (D) an attached blade without reticulations or perforations (TM293).Photographs by James T. Melton III.

Table 1 .
Primers used in this study.

Table 2 .
GenBank Accession numbers of sequences of Ulva ohnoi used in this study.

Table 3 .
Nutrient tissue content (P%DW: percentage of elemental phosphorus dry weight; N%DW: percentage of elemental nitrogen dry weight; and δ 15 N: 15 N/ 14 N in a delta standard notation) of three species of subaquatic vegetation (Anadyomene stellata, Ulva ohnoi, and Thalassia testudinum) collected at Deering Estate, Biscayne Bay, FL.For P%DW, means sharing the same superscript did not differ significantly (Tukeys test, P > 0.05).