Introduction

Carrageenans, or hydrocolloids derived from red algal cell walls, are used in numerous products (both edible and inedible) as thickening and stabilizing agents (Doty 1973; Abbott 1996). The value of the industry is estimated to be US$240 million annually (McHugh 2003), and the market for carrageenan has grown 5% per year since the 1970s (Bixler 1996; McHugh 2003). The Philippines is the leading producer of cultured Eucheuma denticulatum (N.L. Burman) F.S. Collins & Hervey and Kappaphycus spp. (Villanueva et al. 2008) and supplies approximately 70% of the world’s semi-refined carrageenan (Llana 1991). Farming outside the Philippines has been profitable in only a few countries (Hurtado and Agbayani 2002; McHugh 2003). During the 1970s, increased demand for carrageenans resulted in several species of carrageenan-producing algae being brought to Hawaii (Smith et al. 2002). Eucheuma denticulatum and two species of Kappaphycus [K. alvarezii (Doty) Doty ex P.C. Silva and K. striatum (F. Schmitz) Doty ex P.C. Silva] were legally introduced to a northwestern reef bordering Moku O Loe in Kaneohe Bay, Oahu, for growth studies (Glenn and Doty 1981). In 1983, researchers dismissed concerns about the spread and potential impacts of Kappaphycus spp. in the bay (Russell 1983). By 1996, however, Kappaphycus spp. was estimated to be spreading at a rate of 250 m per year and had spread 6 km from the initial sites of introduction (Rodgers and Cox 1999). By 2002, reports indicated that Kappaphycus was heavily dominating some patch reefs in Kaneohe Bay, competing with corals by overgrowing colonies (Smith et al. 2002). Currently, Kappaphycus spp. continue to spread and have moved outside the bay and along the east coast of Oahu up to Hauula, which is approximately 14.0 km from Waikane town in north Kaneohe Bay (B. Hauk, personal communication).

Understanding the distributional spread of the three eucheumoid species in Hawaii is complicated by their morphological plasticity and paucity of diagnostic morphological characters for identification (Lluisma and Ragan 1995; Conklin and Smith 2005; Zuccarello et al. 2006). Misidentification has led to confusion over the distribution and spread of E. denticulatum, K. alvarezii, and K. striatum in Kaneohe Bay, Oahu. Increasingly, the use of molecular approaches has clarified identification of algal species in which morphological characters are poor or contradictory (e.g., Byrne et al. 2002; Kooistra and Verbruggen 2005; Garguilo et al. 2006; Guillemin et al. 2008). While previous studies have employed a variety of molecular markers (the mitochondrial cox2-3 spacer, the plastidal RuBisCo spacer, and rbcL gene, and the nuclear small-subunit ribosomal gene) to elucidate the systematics and taxonomy of Kappaphycus and Eucheuma (Lluisma and Ragan 1995; Fredericq et al. 1999; Zuccarello et al. 2006), here we assess the usefulness of the mitochondrial cox2-3 spacer plus a system of three additional short molecular markers to aid identification and clarify the distribution of these species on Oahu. These three markers are successfully being used to characterize red algal biodiversity in the Hawaiian Islands through the establishment of DNA sequence frameworks for as many red algal species as possible (see the Hawaiian Algal Database for details; http://algae.manoa.hawaii.edu/). Here, we present the results for the genera Kappaphycus and Eucheuma—poorly understood yet important taxa that are having profound negative effects on Oahu’s coral reef ecosystems.

Materials and methods

A total of 15 collections of Eucheuma denticulatum and Kappaphycus spp. were included in the molecular analyses from various sites around the island of Oahu, Hawaii (Table 1). Five herbarium specimens representing Eucheuma and Kappaphycus from the Bernice Pauahi Bishop Museum (BISH) were sampled for DNA analysis, and ten fresh samples were collected by snorkeling. Fresh collections were cleaned of epiphytes under a dissecting scope and either frozen at −20°C or dried using silica gel as a desiccant. Formalin and herbarium vouchers were made with all remaining material. All morphological vouchers are presently located in the Sherwood Laboratory at the University of Hawaii and will ultimately be deposited at BISH.

Table 1 List of samples analyzed in this study
Full size table

DNA extraction, PCR amplification, and sequencing

Genomic DNA was extracted from herbarium material using a modified Dellaporta et al. (1983) protocol described by Hughey et al. (2001) and from silica-dried and frozen material using a DNeasy Plant Mini Kit (Qiagen, USA). Polymerase chain reaction (PCR) was performed using an Eppendorf Mastercycler ep gradient S thermal cycler (Eppendorf, Germany). Four molecular regions were amplified: a portion of the nuclear 28S rRNA gene (large subunit, or LSU), a portion of the plastid 23S rRNA gene (Universal Plastid Amplicon, or UPA), the 5′ end of the mitochondrial cytochrome oxidase subunit I gene region (COI), and the mitochondrial cox2-3 spacer region. PCR was attempted for all samples for the first three regions, and several samples were selected to represent the different genetic clusters for amplification and sequencing of the cox2-3 spacer region, to allow comparison of our collections to those from other locations. PCR reactions (26.5 µL) consisted of 2.5 µL of 10X MangoTaq reaction buffer (Bioline, MA, USA), 1.5 µL of 50 mM MgCl2, 1.5 µL of 1.0% bovine serum albumin solution, 1.0 µL of each primer (0.4 mM), 1.0 µL (20 mM) of each dNTP, 1.0 µL of MangoTaq DNA polymerase, 13.0 µL of nanopure water, and 1.0 µL of total genomic DNA. Herbarium extractions were diluted to 1:10, 1:50, or 1:100, while Qiagen extract DNA templates were used at full strength. PCR cycling conditions and primers followed those described in Sherwood and Presting (2007) for UPA, Saunders (2005) for COI, and Zuccarello et al. (2006) for the cox2-3 spacer. For the LSU region, two primers were used for amplification: nu28SF, 5′-GGAATCCGCYAAGGAGTGTG-3′ and nu28SR, 5′-GCAGGTAAGGGAAGTCGGCA-3′ (Sauvage et al., in review). These primers flank a region of approximately 668 nt (nucleotides) and are used extensively in the Sherwood Lab Rhodophyta Biodiversity Project on both florideophycean and bangiophycean red algae. Amplification conditions for the LSU region were as follows: initial denaturation at 94°C for 2 min, 40 cycles of 94°C (20 s)/55°C (30 s)/72°C (50 s), and a final cycle of 72°C for 5 min. Successful PCR products were purified using a Qiagen PCR purification kit (Qiagen). Purified PCR products were sequenced in both directions on an ABI 377XL DNA Sequencer (Applied Biosystems, USA). Sequences were assembled using Sequencher™ (Gene Codes, Ann Arbor, USA), and 21 additional cox2-3 spacer sequences of Eucheuma and Kappaphycus spp. for collections from Hawaii and other locations were retrieved from National Center of Biotechnology Information GenBank database for comparison to our newly generated cox2-3 spacer sequences (Table 1).

Analyses

Alignments for each of the four regions were generated using Clustal W (Thompson et al. 1994). All gaps and missing data were deleted from the datasets prior to analyses. We employed DNA barcode-like analyses to examine clusters of sequences as potential taxonomic units, a method that has been used in several red algal studies to estimate species boundaries (Saunders 2005, Robba et al. 2006, Sherwood et al. 2008). Neighbor-joining (NJ) analyses (Saitou and Nei 1987) were performed using MEGA version 4 (Tamura et al. 2007), based on Kimura 2-parameter distances (Kimura 1980).

Results

The final nuclear marker data set included 607–628 nt of the LSU (the number of nucleotides varied due to alignment gaps) and consisted of 12 samples. The final plastid marker data set included 370 nt of the UPA marker and consisted of 13 samples. The final COI gene data set included 617 nt (6 samples) while the cox2-3 spacer data set included 342–347 nt and consisted of 26 samples (a subset of our Eucheuma and Kappaphycus samples plus GenBank sequences).

Neighbor-joining (NJ) analyses of portions of the LSU, UPA, and COI revealed similar topologies and demonstrated that the Eucheuma samples are distinct from the Kappaphycus samples (Fig. 1). Within Kappaphycus, the nuclear and plastid markers showed a single nt difference while the mitochondrial marker showed 28 nt differences. The LSU and UPA markers showed no sequence divergence between samples in cluster B, while the COI marker yielded one nt difference between sample 03955 and samples 03954, 03956, and 03957 of cluster B.

Fig. 1
figure 1

Neighbor-joining analyses of the nuclear 28S rRNA region, or LSU (a), the plastid 23S rRNA region, or UPA (b), and the mitochondrial COI region (c) based on Kimura-2-parameter distances for Eucheuma and Kappaphycus samples from Oahu, Hawaii. Kappaphycus clusters (A and B) correlate to Zuccarello et al.’s (2006) clades based on cox2-3 spacer sequences. Scale bars indicate substitutions per site

Full size image

The NJ analysis of the cox2-3 spacer region resulted in a tight clustering of our samples with previously sequenced collections from around the world (including sites from Oahu) (Fig. 2). Samples 00888 and 03953 exhibited no sequence divergence from a wild E. denticulatum sample from Indonesia, although two distinct E. denticulatum clusters are present in the analysis. Sample 02614 showed no sequence divergence from a commercial K. alvarezii sample from Venezuela, but again, several distinct clusters of this taxon are present in the analysis. Samples 00919 and 03955 showed no sequence divergence from K. alvarezii samples previously collected from Kaneohe Bay, Oahu. The taxonomic identities of members of these clusters are not clear and will require further investigation.

Fig. 2
figure 2

Neighbor-joining analysis of cox2-3 spacer sequence data based on Kimura-2-parameter distances for Eucheuma and Kappaphycus samples from Hawaii and GenBank accessions. GenBank accession code or Sherwood Lab accession numbers are indicated in parentheses following the taxon labels. Species (K. alvarezii, K. striatum (sacol), and E. denticulatum) and cluster (A, B, C, D, and E) labels to the right of the tree refer to those from Zuccarello et al. (2006). Scale bar indicates substitutions per site

Full size image

The three newer markers employed in this study (LSU, UPA, and COI) are comparable in their separation of the genera Eucheuma and Kappaphycus using NJ analyses (Fig. 1), and clusters of these analyses are consistent with those revealed for global comparisons with the cox2-3 spacer and GenBank data (Fig. 2). This lends support for the use of the LSU, UPA, and COI markers for the broader assessment of Hawaiian red algal biodiversity.

Discussion

DNA barcoding is based on short, single-marker sequence similarity comparisons to identify species (Hebert et al. 2003). Unfortunately, single-marker identifications may be confounded by varying levels of sequence divergence among taxa (i.e., divergence “cutoffs” may not be universally applicable across even closely related taxa) and, in any case, represent signals from only one of the possible sources of genetic information. We followed a DNA barcoding-like approach to identify Eucheuma and Kappaphycus samples from Hawaii, but used three molecular markers that represent the genetic sources in an algal cell (nucleus, plastid, and mitochondrion). Much confusion surrounds the morphological identification of this eucheumoid complex, and our results showed that these three short regions support congruent and clear separation of a Eucheuma sp. cluster and two Kappaphycus spp. clusters for samples collected from Oahu and revealed new insights into their distribution in Hawaii.

Nuclear, plastidial, and mitochondrial marker congruence to detect species of Eucheuma and Kappaphycus

This effort is the first of several studies currently being conducted to employ these three markers to examine patterns of genetic variation in cryptic red algae in Hawaii. Here, we successfully distinguished morphologically confusing collections of Eucheuma and Kappaphycus that co-occur in Kaneohe Bay. The sequence comparisons based on the LSU, UPA, and COI markers were largely congruent and supported previous findings that separated the genera Eucheuma and Kappaphycus (Llana 1991; Fredericq et al. 1999; Zuccarello et al. 2006). Prior studies employed longer gene regions, such as 892–1443 nt of the rbcL gene (Fredericq et al. 1999), 1767–1781 nt of the SSU rRNA gene (Lluisma and Ragan 1995), or combined short, fast-evolving regions, such as the mitochondrial cox2-3 and plastidial RuBisCO spacers (Zuccarello et al. 2006), to elucidate the systematics and taxonomy of Kappaphycus and Eucheuma. In contrast, our markers are short regions (ranging from 320 to 628 nt) that easily amplify and sequence. These data allow us to separate our Hawaiian samples into a single Eucheuma cluster and two Kappaphycus clusters (Fig. 1), although some differences in ease of application were noted for the three markers. Amplification and sequencing of the COI marker were not successful with herbarium material from BISH when compared with fresh material. A combination of DNA degradation of herbarium specimens and a lower copy number for mitochondrial genes may account for the lower rate of amplification and sequencing success for this marker (A. Kurihara, personal communication). There were fewer problems with amplification and sequencing of the other two markers. Within Kappaphycus, the nuclear and plastidial marker alignments revealed one nt difference while the mitochondrial COI marker alignment showed 28 nt differences between two genetic clusters containing Kappaphycus samples. These data suggest that the markers have different levels of genetic resolution; a finding previously noted for members of the Batrachospermales for the UPA and COI regions (Sherwood et al. 2008). The LSU and UPA markers can be used to separate the genera Eucheuma and Kappaphycus, but may not be adequate to distinguish closely related species.

Confirmation of Eucheuma and Kappaphycus identifications using the cox2-3 spacer

The NJ analysis of the cox2-3 spacer sequences for a subset of our samples and worldwide sequences from GenBank complicate our species-level identification of Eucheuma and Kappaphycus samples (Fig. 2). Our Eucheuma samples (00888 and 03953) exhibited no sequence divergence from a E. denticulatum sample from Indonesia. This relationship was previously noted by the clustering of other Hawaiian E. denticulatum samples with Indonesian E. denticulatum samples in Zuccarello et al.’s (2006) Eucheuma and Kappaphycus study employing the cox2-3 and RuBisCO spacers. This clade is only one of three clades of E. denticulatum presented by Zuccarello et al. (2006) and recovered in our analyses (Fig. 2: C, D, and E). Most of the earlier study’s samples of Eucheuma could only be identified to genus. For these reasons, it is difficult to identify the Hawaiian Eucheuma species as E. denticulatum with confidence. Further analyses examining global molecular diversity of Eucheuma, coupled with DNA sequencing of type specimens or material from type localities, will be necessary to solve this taxonomic conundrum.

Sample 02614 showed no sequence divergence from a K. alvarezii sample from Venezuela. No Hawaiian samples included in Zuccarello et al.’s (2006) study clustered previously with this “cultivated” K. alvarezii clade (Fig. 2: A), although collections had been made from this same locality (Moku o Loe, Kaneohe Bay). This result is not surprising, however, given that Moku o Loe was the site of an experimental seaweed farm in the 1970s for at least three eucheumoid species in Kaneohe Bay (Russell 1983; Glenn and Doty 1990). In contrast, sample 00919 (identified morphologically as K. striatum) showed no sequence divergence from K. alvarezii samples, which were previously collected from Kaneohe Bay, Oahu. This second “K. alvarezii” lineage (Fig. 2: B) was described by Zuccarello et al. (2006) and consisted of haplotypes unique to Hawaii. Inconsistencies in species identifications based on morphology are probably very common for these eucheumoid taxa. Zuccarello et al. (2006) included samples in their analysis identified initially as K. striatum from Tanzania and Madagascar which clustered in their K. alvarezii clade A. In contrast, their “sacol” variety samples were initially identified as K. cottonii based on a study by Aguilan et al. (2003), but did not cluster with another sample morphologically identified as K. cottonii. Zuccarello et al. (2006) concluded that the “sacol” variety was probably a distinct Kappaphycus species, but more sampling and morphological examination was needed. The clade was re-named K. striatum, which still appears to be the commercially accepted species for the “sacol” variety (Hurtado et al. 2008) until a proper name is determined. Thus, it is not surprising to find that our 00919 sample, which was originally identified as K. striatum, has the same sequence as samples identified as K. alvarezii given the lack of distinct and consistent morphological characters for this genus. As mentioned above, given the morphological plasticity of Eucheuma and Kappaphycus species, confirmed application of taxonomic names to genetic clusters will require DNA sequencing of type specimens. To minimize confusion surrounding the Eucheuma and Kappaphycus species in Hawaii, we will use Zuccarello et al.’s (2006) clades as terminology for the Hawaiian entities (Eucheuma clade E, Kappaphycus clade A, and Kappaphycus clade B) and await the results of future studies for species identifications.

Revised distributions of Eucheuma and Kappaphycus on Oahu

Based on these results, new insights have been gained into the distribution of these species around Oahu. Degradation of the Kaneohe Bay patch reefs appears to be proceeding by overgrowth of a species that has been generally identified as Kappaphycus sp. (Rodgers and Cox 1999; Woo 2000; Conklin and Smith 2005). Here, we tentatively revise its identification to Eucheuma clade E (Fig. 2) based on analyzed sequences of samples from Kaneohe Bay. Further use of these molecular markers should be considered when sampling the dominant algal species overgrowing other patch reefs in the bay.

Movement of eucheumoid species outside Kaneohe Bay has sparked well-founded concern. The identification of the entity that has spread as far as Kaaawa (and potentially Hauula) on the east shore of Oahu and been found in drift in Maunalua Bay on the southeast shore of Oahu as Kappaphycus clade B (and not E. denticulatum) was surprising (Fig. 3). This is an important finding for understanding the interplay of invasive character display and spread of a morphologically plastic species complex. Contrary to previous assumptions, two species need to be controlled: Eucheuma clade E and its overgrowth and movement throughout Kaneohe Bay, as well as Kappaphycus clade B and its expansion outside Kaneohe Bay. Their individual reproductive strategies may explain how both species are successful in their respective ways. According to Abbott (1999), tetrasporangial and spermatangial plants of E. denticulatum have not been seen in Hawaii and cystocarpic plants are rare. However, tetrasporangial and cystocarpic materials have been found among herbarium specimens at BISH for both Kappaphycus species (Table 2). Our current understanding suggests that Eucheuma clade E is principally spreading via vegetative propagation, and this strategy has not only allowed the alga to successfully invade the patch reefs it does reach, but, to date, has also confined the alga within the bay. Kappaphycus clade B, on the other hand, has been found both north and south of Kaneohe Bay. Numerous collections of small Kappaphycus clade B plants attached to rocks in Kaaawa suggest that this species is spreading north outside the bay along the east shore of Oahu via dispersal of tetraspores or carpospores from populations inside the bay or populations outside the bay. Ocean circulation patterns along the east side of Oahu support this northern movement from Kaneohe Bay (Qiu et al. 1997). Following the single collection of a large Kappaphycus clade B plant found in the drift in Maunalua Bay on the south shore, algal surveys were done in the area of the Hawaii Kai boat ramp. No other plants were found (B. Hauk, personal communication). It is speculated that this specimen was discarded from a vessel using the boat ramp after a fishing trip to Kaneohe Bay.

Fig. 3
figure 3

Map of Oahu and Eucheuma and Kappaphycus collection sites

Full size image
Table 2 List of tetrasporangial and cystocarpic material for Kappaphycus specimens deposited in the Bernice P. Bishop Museum
Full size table

These results were unexpected. Previous studies suggested that K. alvarezii (formerly E. striatum “tambalang”) is the dominant eucheumoid alga in Kaneohe Bay. Glenn and Doty (1990) revealed that K. alvarezii had a larger average relative growth rate (5.06% per day) than both K. striatum and E. denticulatum (3.50% per day). This has recently been supported in Brazil (Bulboa and de Paula 2005), where K. alvarezii has been shown to grow faster than K. striatum under all experimental conditions. Russell (1983) provided some insight into why Eucheuma clade E (as E. denticulatum) and Kappaphycus clade B (as K. striatum) could be more successful than Kappaphycus clade A (as K. alvarezii) in the Kaneohe Bay. Russell (1983) showed that K. alvarezii (as E. striatum “tambalang”) was biologically limited because it was not found to produce viable spores, while it could reproduce by vegetative fragmentation. Kappaphycus alvarezii was speculated to have no life-history mechanism for dispersal over deep water or out of depressions and channels. Further use of these molecular markers should be considered in creating up-to-date detailed distributional maps of all three groups of Eucheuma and Kappaphycus around Oahu.

Molecular identifications play a crucial role in clarifying these algae, revealing the complex nature of introduced eucheumoid species to Oahu. There is a dire need for further Eucheuma and Kappaphycus taxonomic and ecological research. Much of the confusion surrounding the Eucheuma and Kappaphycus complex arises from the morphological plasticity of these species and the small number of diagnostic morphological characters. Molecular approaches will allow field biologists to corroborate eucheumoid species identifications and thus better associate specific invasive characteristics to a species on an ecological level.

DNA sequencing of type specimens is strongly recommended before taxonomic names are applied to genetic clusters. This approach is critical for this growing and fairly unregulated industry. Recently, Chandrasekaran et al. (2008) reported that K. alvarezii is invading and establishing on both dead and live coral in the Gulf of Mannar Marine Biosphere Reserve (GoM) in India since its introduction in 2000–2002 from the Philippines for commercial production of carrageenan. This runs counter to the assumption that the alga would restrict itself to sand-covered habitats and not compete with native corals in the GoM (Chandrasekaran et al. 2008). Molecular tools can assist in the identification of potentially invasive strains of Eucheuma and Kappaphycus before they are introduced to new sites, and focus management and control efforts on these problematic algae.