Host specificity of the symbiotic cyanobacterium ... - Springer Link

17 downloads 2570 Views 366KB Size Report
sequences to construct phylogenies for host sponges and their symbiotic cyanobacteria ... collected from the Great Barrier Reef, Australia (Hinde et al. 1994) and ...
Marine Biology (2003) 142: 643–648 DOI 10.1007/s00227-002-0971-x

R.W. Thacker Æ S. Starnes

Host specificity of the symbiotic cyanobacterium Oscillatoria spongeliae in marine sponges, Dysidea spp.

Received: 24 January 2002 / Accepted: 8 October 2002 / Published online: 7 February 2003  Springer-Verlag 2003

Abstract Marine sponges can host a variety of cyanobacterial and bacterial symbionts, but it is often unclear whether these symbionts are generalists that occur in many host species or specialists that occur only in certain species or populations of sponges. The filamentous cyanobacterium Oscillatoria spongeliae is found in the sponges Dysidea n. sp. aff. herbacea 1A and 1B, and similar cyanobacteria are found in D. n. sp. aff. granulosa. We amplified and sequenced sponge nuclear ribosomal DNA (rDNA) and cyanobacterial 16S rDNA from specimens of these three sponges. We then used these sequences to construct phylogenies for host sponges and their symbiotic cyanobacteria. Each of these three sponge species hosts a distinct cyanobacterial clade, suggesting a high degree of host specificity and potential coevolution between symbiotic cyanobacteria and their host sponges.

Introduction Many marine sponges contain symbiotic cyanobacteria and heterotrophic bacteria (Sara` et al. 1998; Friedrich et al. 1999), but it is often unclear whether these symbionts are generalists that occur in several host species or specialists that occur only in certain species or populations of sponges. Although these micro-organisms may benefit their host sponge by providing fixed carbon or nitrogen (Rai 1990; Sara` et al. 1998; Diaz and Ru¨tzler 2001), cyanobacterial symbionts have been observed to overgrow and kill their host sponge (Ru¨tzler

Communicated by P.W. Sammarco, Chauvin R.W. Thacker (&) Æ S. Starnes Department of Biology, University of Alabama at Birmingham, Birmingham, AL 35294, USA E-mail: [email protected] Fax: +1-205-9756097

1988) and a pathogenic bacterium has been isolated from the sponge Rhophaloeides odorabile (Webster et al. 2002). Microbial symbionts may produce many of the pharmaceutically active compounds isolated from marine sponges (Unson et al. 1994; Flowers et al. 1998; Schmidt et al. 2000). These compounds can serve a variety of ecological functions, including predator and competitor deterrence (Pawlik et al. 1995; Thacker et al. 1998; Engel and Pawlik 2000) and resistance to malignant microbial infections (Garson 2001). The filamentous cyanobacterium Oscillatoria spongeliae has been isolated from the sponge Dysidea herbacea collected from the Great Barrier Reef, Australia (Hinde et al. 1994) and implicated in the production of the polybrominated biphenyl ethers (Unson et al. 1994) and chlorinated diketopiperazines (Flowers et al. 1998) found in this sponge. Within some sponge species, including D. herbacea, there are many color forms and morphological variations (Colin and Arneson 1995). These variations in color and morphology may be correlated with the presence of different cyanobacteria and bacteria and their production of different secondary metabolites. Cyanobacteria that appear similar to O. spongeliae are associated with three species of Dysidea found around Guam: D. n. sp. aff. herbacea (1A, ridged form), D. n. sp. aff. herbacea (1B, smooth form), and D. n. sp. aff. granulosa. We hypothesized that each of these three sponge species hosts a genetically distinct strain of Oscillatoria. Alternatively, a single strain of Oscillatoria may form associations with all three of these sponge species, or different cyanobacterial strains may colonize sponges in different locations or environmental conditions. In addition, we collected two color forms of D. n. sp. aff. herbacea (1B, smooth form), green and gray, and hypothesized that each color form may host a separate strain of Oscillatoria. To test these hypotheses of host specificity, we constructed phylogenies based on sponge and cyanobacterial ribosomal DNA (rDNA) sequences and determined whether more closely related sponges host more closely related cyanobacteria.

644

Materials and methods Sample collection and DNA extraction Specimens of three Dysidea species were collected from a variety of locations in Guam, including Anae Island, Piti Bombholes, Tumon Bay, Double Reef, and Pago Bay (Fig. 1). All three species lack official taxonomic names, but have been described as D. n. sp. aff. herbacea 1A (ridged form), D. n. sp. aff. herbacea 1B (smooth form), and D. n. sp. aff. granulosa (Kelly et al. 2003). Complete taxonomic descriptions of these species are currently in progress (M. Kelly, personal communication). D. n. sp. aff. herbacea 1A (ridged form) has a purple-gray coloration, grows in spreading mats consisting of soft, interlocking lamellae, and has rows of low, cream-colored conules across its surface (Fig. 2C). D. n. sp. aff. herbacea 1B (smooth form) ranges in color from purple-gray to yellow-green, growing in spreading mats with vertical digitate extensions, but with a smooth surface texture and very few visible conules (Fig. 2A, B). Both species are similar to D. herbacea (Keller 1889) as described by Bergquist (1965). D. n. sp. aff. granulosa ranges in external color from light gray to purple, growing as cylindrical, ramose branches or irregular mounds (Fig. 2D). The sponge is heavily packed with detritus and has a granular surface texture, similar to that of D. granulosa Bergquist (1965). Specimens were placed in 70% ethanol and stored at room temperature for 7 months prior to DNA extraction. Total genomic DNA was extracted using the Wizard Genomic DNA Purification Kit (Promega), following the manufacturer’s protocol for animal tissue. Extracts were immediately cleaned using the Wizard DNA Clean-up System (Promega), again following the manufacturer’s protocol.

Amplification of sponge nuclear rDNA We designed PCR primers based on known sponge nuclear rDNA sequences (Borchiellini et al. 2000) to amplify a portion of the 5.8S rDNA subunit, the second internal transcribed spacer (ITS-2), and a portion of the 28S rDNA subunit from the total genomic extracts, yielding a 700 bp product. We used 25 pmol of forward primer SP58bF (5¢-AATCATCGAGTCTTTGAACG-3¢) and reverse primer SP28cR (5¢-CTTTTCACCTTTCCCTCA-3¢) in each 50 ll PCR reaction. Other components included the supplier’s reaction buffer and additives (TaqMaster, Eppendorf) and 8 nmol of each dNTP. Templates were denatured for 4 min at 94C, after which 0.5 units of MasterTaq DNA polymerase (Eppendorf) were added to the reaction mixture. The mixtures were subjected to 35 cycles of denaturation at 94C for 30 s, annealing at 45C for 30 s and extension at 72C for 60 s.

Amplification of cyanobacterial 16S rDNA We used cyanobacterial-specific primers (Nu¨bel et al. 1997) to amplify approximately 1,300 bp of cyanobacterial 16S rDNA from these same genomic extracts. These reactions differed from the previous reactions only in the annealing temperature (60C) and the primers used. Forward primer CYA106F (5¢-CGGACGGGTGAGTAACGCGT GA-3¢) and reverse primer CYA781R (an equimolar mixture of 5¢-GACTACTGGGGTATCTAATCCCATT-3¢ and 5¢-GACTACAGGGGAATCTAATCCCTTT-3¢) amplified the upstream 650 bp of the sequence, while forward primer CYA359F (5¢-GGGGAATTTTCCGCAATGGG-3¢) and reverse primer 1509R (5¢-GGTTACCTTGTTACGACTT-3¢; Martı´ nezMurcia et al. 1995) amplified approximately 1,150 bp downstream.

DNA sequencing, sequence alignment and analyses PCR products were gel-purified, cleaned with Wizard PCR Preps (Promega), and directly sequenced on an ABI PRISM 377 automated sequencer by the UAB DNA Sequencing Core Facility. The amplification primers were used as sequencing primers, with the addition of CYA781F (an equimolar mixture of 5¢-AATGGGATTAGATACCCCAGTAGTC-3¢ and 5¢-AAAGGGATTAGATTCCCCTGTAGTC-3¢) as an additional primer for the downstream portion of the cyanobacterial sequence. Sequences were aligned using the Pileup subroutine of the Wisconsin Package (Genetics Computer Group) and Se–Al (Rambaut, University of Oxford). After excluding ambiguous base calls near the terminal primers, 683 bp of sponge nrDNA and 1265 bp of cyanobacterial 16S rDNA were analyzed using the PAUP* program (Swofford 1999). For each data set, 500 bootstrap replicates of a branch-and-bound maximum-parsimony search were evaluated. Maximum likelihood searches were also performed for both the sponge and cyanobacterial data, using a series of nested models (Swofford et al. 1996). A log-determinant distance matrix was calculated for each data set, and used to construct a neighbor-joining tree. A Mantel test of cross-matrix correlation was used to determine whether sponge and cyanobacterial log-determinant distances were significantly correlated (Parker and Spoerke 1998).

Results

Fig. 1 Dysidea n. sp. aff. herbacea 1A (ridged form) was collected on Guam in Tumon Bay. Collection locations for D. n. sp. aff. herbacea 1B (smooth form) included Double Reef and Anae Island, while D. n. sp. aff. granulosa was collected at Double Reef, Piti Bombholes, and Pago Bay

Sequences obtained from host sponges and symbiotic cyanobacteria are stored in GenBank under accession numbers AF420441–6 and AF534685–702. A total of 683 bases were aligned from the sponge ITS-2 and 28S ribosomal DNA sequences, including 100 (14.6%) parsimony-informative bases. Of the 318 bp in the ITS-2

645 Fig. 2A–D Photographs of collected sponges. A The yellow-green color morph of Dysidea n. sp. aff. herbacea 1B (smooth form). B The purplegray color morph of D. n. sp. aff. herbacea 1B (smooth form). C D. n. sp. aff. herbacea 1A (ridged form; photograph courtesy of G. Paulay, Florida Museum of Natural History). D D. n. sp. aff. granulosa

region, 82 (25.8%) were parsimony-informative. Of the 365 bp in the 28S subunit, 18 (4.9%) were parsimonyinformative. A single maximum parsimony tree of length = 111 and consistency index = 0.99 (Fig. 3) shared an identical topology with distance and maximum-likelihood (general time-reversible model) trees. The sponge phylogram indicates that these sequences effectively differentiate among the three sponge species, with little variation among locations, color forms, or individual specimens (pairwise sequence divergence among species: 7.74–14.26%; within species: 0–0.53%). A total of 1,265 bases were aligned from the cyanobacterial 16S rDNA sequences, including 34 (2.7%) parsimony-informative bases. The consensus of three equivalent maximum-parsimony trees with length = 79 and consistency index = 0.98 (Fig. 3) shared an identical topology with distance and maximum likelihood (Felsenstein 1981 model) trees. A distinct clade of cyanobacteria was associated with each sponge species (Fig. 3). Cyanobacterial symbionts displayed less overall genetic variability than their host sponges (pairwise sequence divergence among clades: 1.67–3.02%), but displayed more variability within clades than did sponges (pairwise sequence divergence within clades: 0–1.44%).

A GenBank search revealed that the O. spongeliae sequences were most similar to the 16S rDNA sequence reported for the marine filamentous cyanobacterium O. corallinae (Nelissen et al. 1996). We found no consistent genetic differentiation of cyanobacteria among locations or among sponge color forms (Fig. 3). However, log-determinant distance matrices for sponges and cyanobacteria were significantly positively correlated (Mantel cross-matrix correlation = 0.805, P=0.002), indicating that more closely related sponges host more closely related cyanobacteria. The ratios of transitions to transversions were different between these taxa, with a ratio of 1.26 for sponges and a ratio of 0.59 for cyanobacteria.

Discussion Our evidence supports the hypothesis that each species of Dysidea hosts a distinct strain of Oscillatoria, with little genetic variation within sponge species or cyanobacterial strains. Most previous studies of poriferan molecular systematics have focused on family-level classifications or higher, with few studies examining relationships within

646

Fig. 3 Phylogenies constructed for sponges, Dysidea spp., and their cyanobacterial symbionts, Oscillatoria spp. Color forms and collection locations are noted after sponge names. For clarity, these trees have been rooted with D. n. sp. aff. herbacea 1A (ridged form), without implying polarity. The topologies shown were obtained using maximum parsimony search criteria; log-determinant distances and maximum likelihood models yielded congruent topologies. Numbers above branches indicate percentage bootstrap values from 500 replicates of a branch-and-bound maximumparsimony search; *