| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
1 Marine Biology Research Division 0202, Scripps Institution of Oceanography, University of California San Diego, La Jolla, California 92093-0202
2 University of Guam Marine Laboratory, Mangilao, Guam 96923
3 Smithsonian Tropical Research Institute, Apartado 2072, Balboa, Republic of Panama
* To whom correspondence should be addressed. E-mail: rrowan{at}uog9.uog.edu
 |
Abstract |
|---|
 TOP Abstract IntroductionMaterials and MethodsResultsDiscussionLiterature Cited |
|---|
Abbreviations: RFLP, restriction fragment length polymorphism • rDNA, ribosomal RNA genes • srDNA, small ribosomal subunit RNA genes
|
Introduction |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionLiterature Cited |
|---|
Coral taxonomy at the species level, although occasionally frustrating (Knowlton and Jackson, 1994; Veron, 1995), has generally been sufficient to describe overall diversity and to define experimental subjects. This taxonomy seldom, however, has considered zooxanthellae, because it was widely assumed that one species of coral associates with only one species of zooxanthella—in other words, that host taxonomy identified both partners. Zooxanthellae are diverse (e.g., Schoenberg and Trench, 1980; Rowan, 1998), and it is now recognized that some species of corals associate with multiple species of zooxanthellae (Rowan and Knowlton, 1995; Rowan, 1998). Thus corals identified as members of the same species may not in fact be equivalent at the whole organism (holobiont) level, and the taxonomic identities of zooxanthellae may be as ecologically important as those of their hosts.
As far as is known, zooxanthellae in reef-building corals are members of the genus Symbiodinium (Rowan, 1998), which includes four species described as in vitro cultures (Freudenthal, 1962; Trench and Blank, 1987). Several other cultured isolates of Symbiodinium have been named informally, but most members of the genus remain uncultured and undescribed (Rowan, 1998). Nevertheless, sequences and restriction fragment length polymorphism (RFLP) of genes that encode ribosomal RNA (rDNA) can be used to distinguish some taxa of Symbiodinium and to study ecological relationships among host, symbiont, and habitat diversity (Rowan and Powers, 1991a, b; Rowan and Knowlton, 1995; Rowan et al., 1997; Baker and Rowan, 1997; Hill and Wilcox, 1998; Darius et al., 1998; Baker, 1999). The present study uses genes that encode small ribosomal subunit RNA (srDNA).
Our earlier work concerned zooxanthellae of the sibling coral species Montastraea annularis, M. faveolata, and M. franksi, which are the dominant reef-building corals in the Western Atlantic (Goreau, 1959). On an offshore reef in the San Blas Islands of Panama, we found that both M. annularis and M. faveolata associate with three distinct taxa of Symbiodinium (A, B, and C; see Rowan and Knowlton, 1995; Rowan et al., 1997). Symbiodinium A and B, or both, are predominant in tissue exposed to high irradiance (shallower water or colony tops), Symbiodinium C is predominant in shaded tissue (deeper water or colony sides), and mixtures of Symbiodinium A and/or B with C occur between these extremes. Colonies of M. franksi, in contrast, were found to host only Symbiodinium C (Rowan and Knowlton, 1995); however, this coral species was not found at shallow depths on this reef. These observations led to two questions addressed here. First, do these symbiont taxa and patterns of association occur on other types of reef? Second, does the deeper distribution of M. franksi reflect an inability by this species to host those taxa of Symbiodinium with which M. annularis and M. faveolata associate in shallow water?
We also discuss some concerns about using srDNA to identify the Symbiodinium that we collected. Although srDNA was heterogeneous in samples of Symbiodinium B, C, and E, we found no evidence to suggest that the zooxanthellae in each of these samples were heterogeneous. We suspect that srDNA in these Symbiodinium is a heterogeneous multigene family, as is rDNA in some other dinoflagellates (Scholin et al., 1993; Scholin and Anderson, 1994, 1996). We discuss practical implications of this suspicion for the use of srDNA as a taxonomic character.
|
Materials and Methods |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionLiterature Cited |
|---|
-in (#12) steel hole punch, yielding a coral core with about 1.3 cm2 of live colony surface. Cores were wrapped in aluminum foil and frozen in a cryogenic dry shipper (chilled with liquid nitrogen). Many colonies of Montastraea annularis and M. faveolata were sampled both on their tops and on their sides to obtain samples from relatively high- and low-irradiance tissues (respectively) within a colony (Rowan et al., 1997). Most colonies of M. franksi were sampled at only one location because their relatively flat morphologies made a distinction between colony top and side superfluous. Coral colonies were sampled at three sites in the Republic of Panama (Fig. 1) between October 1997 and October 1998:
|
(2) Cayos Limones, San Blas. These reefs are located 15 km north of mainland Panama and are not strongly influenced by terrestrial runoff (D’Croz et al., 1999). We sampled from a relatively steep, leeward fringing reef that ends abruptly at depths between 35 and 40 m in soft sediments (see fig. 9 in Robertson and Glynn, 1977). On this reef, M. franksi is common below 8 m, and it is the dominant coral (with Agaricia lamarcki) below 15 m. We sampled M. franksi throughout its depth range (4 to 38 m, n = 78 colony tops).
(3) Bocas del Toro. Juan Point and Cocos Point reefs are located in the semienclosed lagoon of Bahía Almirante in the Province of Bocas del Toro. Like Río Cartí, these sites are affected by high rainfall and river outflow throughout much of the year. On many of the reefs in this area, M. franksi is the most abundant member of the M. annularis complex. We made a limited collection at depths of 1-15 m for comparative purposes, consisting of 1 top sample of M. annularis, 10 of M. franksi, and 3 of M. faveolata.
Observations of srDNA heterogeneity within samples of Symbiodinium C (see Results) prompted us to investigate the stability of these genotypes under stress. We identified 11 colonies of M. annularis (each colony consisting of a cluster of columns) that hosted heterogeneous RFLP genotypes of Symbiodinium C. After an initial sample, columns (one per colony) were transplanted from their natural habitat (ca. 10-14 m depth) to 1 m depth at either Cayos Limones (n = 4 transplants) or Aguadargana (n = 7 transplants) reefs (Fig. 1), where they bleached. Columns were sampled again after 4 days (Cayos Limones) or 40 days (Aguadargana). Transplants and determinations of zooxanthellar numbers were conducted as described in Toller et al. (2001). In the present study, however, we did not sample corals further (i.e., during zooxanthellar repopulation; see Toller et al., 2001).
Identification of zooxanthellae
Zooxanthellae were isolated and identified as described previously (Rowan and Powers, 1991b; Rowan and Knowlton, 1995). srDNAs were obtained by PCR amplification with a "host-excluding" primer pair (ss5 and ss3Z) or with universal primers (ss5 and ss3), and then characterized by restriction enzyme digestion. The host-excluding primer pair does not amplify known host srDNAs (Rowan and Powers, 1991b; unpubl. obs.), but does amplify srDNAs from a phyletic group that is much larger than Symbiodinium (McNally et al., 1994; Toller et al., 2001). All samples were assayed using host-excluding primers, and about one-third of them were also analyzed with universal primers. Data obtained from the two kinds of amplifications were always in agreement.
Every sample was analyzed by digesting amplified srDNA with Dpn II and with Taq I, which differentiate Symbiodinium A, B, and C by RFLP (Rowan and Powers, 1991a; Rowan and Knowlton, 1995; Rowan et al., 1997). RFLPs were diagnosed by comparison to genotype standards, which were obtained by PCR amplification from cloned srDNAs of Symbiodinium A, B, and C, all isolated from M. annularis (Rowan and Knowlton, 1995), and from Symbiodinium E (from M. faveolata, this study). These cloned genotype standards are denoted hereafter as A0, B0, C0, and E0-1. We use the superscript zero to indicate srDNA clones, as opposed to taxa of Symbiodinium; clones obtained from different samples of the same taxon of Symbiodinium are distinguished by numbers (e.g., E0-1 and E0-2; see below). Because universal PCR primers amplify coral host srDNA when it is present (Rowan and Powers, 1991b), a cloned srDNA from M. annularis (clone H0; see below) was used as an additional standard in RFLP analyses of these amplifications. Where RFLP analyses indicated mixtures of Symbiodinium A, B, C, or E in a sample, relative abundance (greater than or less than 50% of the total) was estimated by comparison to standard mixtures prepared from cloned srDNAs (Rowan and Knowlton, 1995; Rowan et al., 1997; see Fig. 4).
|
Clones E0-1, E0-2, and E0-3 were sequenced completely, as were cloned genotype standards A0, B0, and C0 (from which only partial sequences had been obtained previously; Rowan and Knowlton, 1995) and clone H0. Plasmids were prepared using QIAprep Spin Miniprep kits (Qiagen, Inc., Valencia, CA) according to manufacturer’s recommendations, and sequences were determined for both DNA strands using Big Dye Terminator sequencing kits (PE Corporation, Norwalk, CT) with vector sequencing primers T7 and M13-Reverse, and with srDNA sequencing primers 18F1 (5'-AGCTCGTAGTTGGATTTCTG-3'), 18F2 (5'-TTAATTTGACTCAACACGGG-3'), 18R1 (5'-AGTCAAATTAAGCCGCAGGC-3') or 18-R1X (5'-GTTGAGTCAAATTAAGCCGC-3'), and 18R2 (5'-ATATACGCTATTGGAGCTGG-3'). Reactions were analyzed with an ABI 373 sequencer (Applied Biosystems, Foster City, CA) and complete sequences were assembled using SeqEd software (Applied Biosystems). RFLP genotypes of cloned srDNAs were obtained from their sequences using Gene Construction Kit software (Textco, Inc., West Lebanon, NH). Note that we used only partial srDNA sequences in some analyses (Fig. 3); the full-length srDNA sequences were deposited in GenBank (http://www.ncbi.nlm.nih.gov/; accession numbers AF238256-AF238258, AF238261-AF238263, and AF238267).
|
To investigate srDNA variation within our samples of Symbiodinium in greater detail, we selected representative samples of each Symbiodinium taxon from each host coral species (M. annularis, M. faveolata, M. franksi) and made additional RFLP analyses. Different samples (from different colonies) of Symbiodinium A (n = 10), B (n = 12), C (n = 12), and E (n = 12) were analyzed with a panel of 12 restriction enzymes, used one at a time. These enzymes were Dpn II, Taq I, Alw I, BstU I, Hae III, Hha I, Hinf I, Mse I, Msp I, Nci I, Sau96 I, and Sty I. Samples of Symbiodinium E were investigated further with the enzymes Alu I, Bsp1286 I, Mae III, Mnl I, SfaN I, and Tsp45 I. We chose the latter enzymes based on RFLP differences among clones E0-1, E0-2, and E0-3. All enzymes were purchased from New England Biolabs, Inc. (Beverly, MA) except for Mae III (Roche Diagnostics Corp., Indianapolis, IN).
|
Results |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionLiterature Cited |
|---|
|
Distribution of different taxa of Symbiodinium
At Río Cartí, M. franksi was observed with only two taxa of zooxanthellae—Symbiodinium E and C—and the same two taxa were obtained from M. faveolata and M. annularis (Fig. 5) at this reef. Symbiodinium E was the predominant zooxanthella from all three Montastraea species: it occurred in 35 of 43 corals and was the only zooxanthella detected in 18 of these. In M. franksi and M. faveolata, Symbiodinium E was more common in higher irradiance habitats (colonies at 1-3 m depth, tops of colonies at 3-6 m depth) than in lower irradiance habitats (colony sides at 3-6 m depth and generally below 6 m); Symbiodinium C exhibited the converse pattern (Fig. 5). Samples from M. annularis (n = 4) showed the same top and side pattern of zooxanthellar distribution within colonies (Fig. 5), although our small sample size precludes an examination across depth. A zonation pattern was often observed in comparisons of tops and sides from the 16 doubly sampled colonies that had the two types of zooxanthellae. In 12 of these colonies, the ratio of Symbiodinium E:C decreased from top to side, in three there was no clear difference in the ratios, and in only one colony did the ratio increase from top to side.
|
|
In our limited sample of corals from two reefs at Bocas del Toro (1-15 m depth), M. franksi was found with Symbiodinium E only (1 colony), with Symbiodinium E and C (4 colonies), with Symbiodinium C only (2 colonies), or with Symbiodinium A only (3 colonies). M. faveolata was found with Symbiodinium C only (2 colonies) or with Symbiodinium A only (1 colony). The single encountered colony of M. annularis contained Symbiodinium A. We did not observe Symbiodinium B in any of these samples.
Other diversity in zooxanthellar srDNAs
The routine RFLP analyses (with Dpn II and Taq I) reported above indicated that all samples of zooxanthellae in this study contained srDNAs of either Symbiodinium A, B, C, or E, or mixtures thereof, as defined by our standard, cloned srDNA genotypes (A0, B0, C0, E0-1). However, when zooxanthellar srDNAs were analyzed in greater detail (with additional restriction enzymes; see Methods and Materials), samples of Symbiodinium B, C, and E (but none of 10 tested samples of Symbiodinium A) were found to contain additional srDNAs that could not be attributed to genotypes A0, B0, C0, E0-1, or to host srDNA. These other srDNAs appeared as additional DNA fragments in restriction digests, as described below.
Twelve selected samples of Symbiodinium E and clones E0-1, E0-2, and E0-3 were all indistinguishable in digests with Dpn II (examples in Fig. 7, Dpn II panel) and with Taq I (not shown). In digests with Mae III, however, all of these samples had an additional DNA fragment in relatively low abundance (arrow in Fig. 7, Mae III panel) that was not part of the RFLP genotype of clones E0-1 and E0-2, but which was in the RFLP genotype of clone E0-3. Thus, these samples apparently contained at least two srDNAs—one defined in Mae III digests by clones E0-1 and E0-2, the other by clone E0-3. Similarly, an additional band in digests of sample srDNAs with Mnl I (arrow in Fig. 7, Mnl I panel) apparently represents the RFLP genotype of clone E0-1 (versus clones E0-2 and E0-3). Digestion of samples with Alu I also yielded an additional DNA fragment (arrow in Fig. 7, Alu I panel), and digestion of cloned srDNAs with Alu I showed that the genotype of clone E0-2 is unique. In all, additional bands like those shown in Figure 7 (arrows) were observed in 7 of 18 different restriction enzyme digestions (other digests not shown) of the 12 tested samples of Symbiodinium E. Therefore, srDNA in these samples of Symbiodinium E was clearly heterogeneous. This heterogeneity did not, however, vary qualitatively nor quantitatively among the tested samples (e.g., Samples 1-3 in Fig. 7). Thus, clones E0-1, E0-2, and E0-3, which are different (Fig. 7; see also Fig. 3), were obtained from indistinguishable samples of zooxanthellae.
|
|
|
|
|
Discussion |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionLiterature Cited |
|---|
Groups A, B, C, and E constitute the known diversity of coral-associated Symbiodinium (Rowan, 1998; this study), and M. annularis, M. faveolata, and M. franksi all associate with at least one member of each of these groups. This is a remarkable amount of taxonomic diversity—at least 12 distinct symbioses—in what was previously (Knowlton et al., 1992; Rowan and Knowlton, 1995) regarded as one species of coral hosting one species of zooxanthella. Moreover, this diversity is not randomly distributed, suggesting that what was once viewed as a single quintessential generalist (Connell, 1978) is in fact a complex assemblage of ecologically more specialized entities.
Our observations from Cayos Limones now enable us to refute the speculation that M. franksi associates exclusively with Symbiodinium C—this host coral can and does form symbioses with Symbiodinium A, B, and E. However, at this offshore reef, the latter host-zooxanthella combinations are observed only at the margins of this coral’s depth distribution (Fig. 6): shallow (Symbiodinium B > A > E) and very deep (Symbiodinium E; discussed further below). Otherwise, M. franksi hosts Symbiodinium C throughout nearly all of its depth range (Fig. 6), where colony growth is robust. Although in shallow water the distributions of zooxanthellae (mostly Symbiodinium A and B) are similar in M. franksi, M. faveolata, and M. annularis, the small size of M. franksi colonies in shallow water may reflect a relatively poor physiological fit between this coral host and these zooxanthellae.
The main question posed by our new results is why all three species in the Montastraea annularis species complex at a coastal site (Río Cartí) host predominantly Symbiodinium E at higher irradiance (Fig. 5), instead of Symbiodinium A or B, as found at offshore reefs (Rowan and Knowlton, 1995; Rowan et al., 1997; Fig. 6). One possible explanation is that this coastal site is characterized by environmental stress to which Symbiodinium E is more tolerant than are Symbiodinium A or B. High irradiance is a stress that may exacerbate (Brown, 1997) the many other kinds of stress found in nearshore environments (e.g., fluctuations in temperature, salinity, nutrients, sediments, and underwater irradiance; see Bowden, 1983; Kirk, 1994). All of these factors can affect the stability of coral-algal symbioses (Falkowski et al., 1993; Brown, 1997; Wesseling et al., 1999). In the San Blas Archipelago (Fig. 1), nearshore effects associated with freshwater runoff are limited to a relatively narrow coastal band and do not reach our offshore study sites at Aguadargana and Cayos Limones (D’Croz et al., 1999). Symbiodinium E was also common in Montastraea within a large coastal lagoon at Bocas del Toro, Panama (Fig. 1), an area of exceptionally high rainfall where water quality is also likely to be dominated by coastal effects.
A second (and perhaps related) question asks why Symbiodinium E was distributed differently at Cayos Limones, where it was common not at high irradiance but rather in the very deepest colonies of M. franksi (Fig. 6). Perhaps shallow and deep populations of Symbiodinium E are different species of zooxanthella, although we did not find any evidence to support this (see following section). Instead, we suggest that Symbiodinium E was actually not distributed so differently at these two sites. In both cases it was associated with marginal habitat: at great depth where M. franksi colonies are not large and where the reef itself disappears into sediment (Cayos Limones), and along the coast near a large river, where coral reefs are poorly developed or absent (Río Cartí). Bleaching-associated stress may be common in both habitats, due to occasional smothering by sediments in the former (e.g., Wesseling et al., 1999) and to near-shore conditions in the latter (see above). We propose that the Symbiodinium E we observed represents a taxon of zooxanthella that occurs in certain habitats not because it performs best in those habitats, but because it tolerates them, whereas Symbiodinium A, B, and C do not. According to this idea, Symbiodinium E is rare or absent from other habitats not because it performs poorly in them, but because Symbiodinium A, B, and C are better adapted to those habitats and somehow exclude it.
Anecdotal observations are consistent with our interpretation of Symbiodinium E as a stress-tolerant zooxanthella. We observed Symbiodinium E (diagnosed by Dpn II and Taq I digests of srDNA) in M. faveolata in the Bahamas (not shown), in four of seven colonies that were relatively unbleached during a natural bleaching event (D. Zawada, Scripps Institution of Oceanography, pers. comm.). We also found that Symbiodinium E—but not Symbiodinium B or C—was adept at repopulating severely bleached corals in experiments (Toller et al., 2001). These experimental results suggest that, in addition to tolerating stress, Symbiodinium E may also be good at colonizing corals whose zooxanthellar communities have been severely disrupted by stress.
Observations of zooxanthellae related to Symbiodinium E in other hosts and seas imply that this taxon, like the taxa Symbiodinium A, B, and C (Rowan, 1998), may represent a group (clade) of zooxanthellae. Those observations include the corals Montipora patula in Hawaii (Rowan and Powers, 1991a; Fig. 3), Acropora palifera in Australia (R. R., unpubl. obs.), Pocillopora damicornis in the eastern Pacific (Baker, 1999), Goniastrea aspera in Thailand (A. Douglas, University of York, pers. comm.), and the giant clam Hippopus hippopus in Australia (R. R., unpubl. obs.). In the context of our hypothesis that Symbiodinium E is stress tolerant in Montastraea, it is notable that G. aspera occurs on reef flats—an environment that is stressful for corals, and where coral bleaching events occur regularly (Brown et al., 2000). Similarly, in P. damicornis, Symbiodinium E was disproportionately common in unbleached colonies during an El Niño-related bleaching event (Symbiodinium D of Baker [1999] has an RFLP pattern that is indistinguishable from that of Symbiodinium E from Montastraea in three restriction enzyme digests; A. Baker, Wildlife Conservation Society, pers. comm.). These observations suggest that other members of the clade Symbiodinium E may also be stress tolerant.
The hypothesis that Symbiodinium E is a relatively stress-tolerant zooxanthella is based on circumstantial evidence, and should be tested in experiments in which environmental factors are controlled and physiological responses are measured. Descriptive studies of unmanipulated corals are, however, indispensable for framing realistic hypotheses in the first place.
Taxonomic interpretation of variation in zooxanthellar srDNA
We recognize the RFLP genotype E as a distinct taxon— Symbiodinium E—for the following reasons: (i) RFLP genotype E was common, and many samples contained only this genotype (Fig. 5); (ii) the nonrandom distribution of RFLP genotype E (Figs. 5 and 6) strongly implies that it represents a distinct organism with distinct ecological attributes; and (iii) phylogenetic analyses of genotype E srDNAs place them within Symbiodinium, but distinct from srDNAs of genotypes A, B, and C (Fig. 3), which, by the same reasoning, represent distinct taxa of Symbiodinium (Rowan, 1998). In practice, these four taxa of Symbiodinium are readily identified by comparison to cloned srDNAs (RFLP genotypes A0, B0, C0, and E0-1) digested with the enzymes Dpn II and Taq I.
By analyzing zooxanthellar srDNA with additional restriction enzymes, we found that samples containing srDNA of RFLP genotype B0, C0, or E0-1 also contained at least one additional srDNA of a different RFLP genotype (examples in Figs. 7-9). What do these additional srDNAs represent, taxonomically? Like an srDNA in genotype C* (Rowan and Knowlton, 1995), they appear to be from Symbiodinium (and not some other type of organism) because (i) they were distinguishable in fewer than one-half of different restriction digests, (ii) many of them seemed to represent simple, single restriction site changes compared to a cloned srDNA (not shown), and (iii) different srDNAs from samples of Symbiodinium E (Fig. 3) or of C* (Rowan and Knowlton, 1995) differed relatively little in sequence.
Do these additional srDNAs represent distinct species or strains of Symbiodinium? In the case of Symbiodinium E and B, no evidence suggests that they do. Specifically, these srDNAs were not observed by themselves, nor did they vary in relative abundance from sample to sample (Figs. 7 and 8). This contrasts with observations on srDNAs of RFLP genotypes A0, B0, C0 and E0-1, which occur alone, and also mix in a range of proportions (e.g., Symbiodinium C and E, Fig. 4; Rowan and Knowlton, 1995; Rowan et al., 1997).
Because srDNA is a multigene family in eukaryotes, srDNA heterogeneity (as seen within samples of Symbiodinium B and E) can reside in one organism—including dinoflagellates (Scholin et al., 1993; Scholin and Anderson, 1994, 1996)—among gene-family members (Hillis and Dixon, 1991). We favor this as an explanation for our data because it is parsimonious compared to the alternative of multiple strains of zooxanthellae that for some reason always co-occur in the same relative proportion. Testing this hypothesis requires the analysis either of one dinoflagellate (e.g., Yeung et al., 1996) or of a clonal culture (e.g., Scholin et al., 1993; Rowan et al., 1996).
Heterogeneity of srDNA within samples of Symbiodinium C was more intriguing because sample-to-sample variation was observed among colonies (Fig. 9). That observation suggested that different srDNAs within any one sample could represent different strains of Symbiodinium. If so, that sample-to-sample variation might also appear within one coral colony, either from place to place or time to time, especially before versus after an environmental change. We found no such variation (e.g., Fig. 10) in corals hosting Symbiodinium C, which again is consistent with the hypothesis that srDNA heterogeneity is a property of individual zooxanthellae. Different patterns of srDNA heterogeneity seen among samples of Symbiodinium C from different corals (Fig. 9) are different zooxanthellar genotypes, but we do not know if these differences are biologically significant (e.g., Scholin and Anderson, 1994, 1996).
Independent of its source, within-sample srDNA heterogeneity limits the information that can be obtained from srDNA sequences. This limitation is apparent in our analysis of Symbiodinium E. The sequence of clone E0-1 implies that our RFLP analyses, using 18 enzymes (examples in Fig. 7), surveyed about 220 nucleotide positions (not shown). Heterogeneity was detected with seven enzymes, which implies a within-sample srDNA sequence diversity of about 3% (7 of 220 nucleotide positions). We do not know how this diversity is distributed; possibilities range from two srDNAs that differ at 7 of 220 positions (ca. 3% different srDNAs, similar to the difference between srDNAs of Symbiodinium A and B [Rowan and Powers, 1992]) to seven srDNAs that differ from one another at 1 of 220 positions (ca. 0.4% different srDNAs). Differences among srDNA clones E0-1, E0-2, and E0-3 fall within this range, and there is no reason to expect any cloned srDNA to represent our samples of Symbiodinium E with any greater precision. Moreover, the PCR creates chimeric DNA molecules when mixed templates are amplified, and many clones obtained from those PCR products will be artifacts (Bradley and Hillis, 1997; Wintzingerode et al., 1997; Darius et al., 1998).
Sequences of srDNAs obtained (as clones) from Symbiodinium in the M. annularis species complex are summarized in Figure 3. Because we have evidence for only four taxa—A, B, C, and E—the multiple branches within groups B, C, and E represent sequence variation within, not among, taxa. An exception to this statement is the pair of sequences labeled C20-1 and C20-2, which came from an experimentally bleached M. annularis and from an unmanipulated colony of the coral Siderastrea siderea, respectively (see Toller et al., 2001). Ecological data and RFLP analyses strongly imply that C20-1 and C20-2 represent a taxon (Symbiodinium C2) that is distinct from the taxon Symbiodinium C found commonly in unmanipulated Montastraea (Toller et al., 2001). We stress that this taxonomic difference could not be inferred from srDNA sequence data alone, given the levels of srDNA heterogeneity within samples of Symbiodinium C and C2 (Toller et al., 2001).
In conclusion, the problem of fully interpreting srDNA variation in natural samples of Symbiodinium is challenging. By themselves, srDNA sequence data contributed relatively little to understanding zooxanthellar diversity in Montastraea. RFLP data were much more informative, not the least because they revealed the informational limits of srDNA sequences. Many samples of zooxanthellae from these species of coral contained more than one taxon of Symbiodinium (Figs. 4 and 5; Rowan et al., 1997), a phenomenon that would have been difficult to understand from srDNA sequences alone. RFLP data are easily obtained, at reasonable cost, from many samples of zooxanthellae, which allows ecological data to inform taxonomic decisions.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
|
Literature Cited |
|---|
|
TOPAbstractIntroductionMaterials and MethodsResultsDiscussionLiterature Cited |
|---|
This article has been cited by other articles:
|
|
 |
|
  E. M. Sampayo, T. Ridgway, P. Bongaerts, and O. Hoegh-Guldberg Bleaching susceptibility and mortality of corals are determined by fine-scale differences in symbiont type PNAS, July 29, 2008; 105(30): 10444 - 10449. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. Takahashi, S. Whitney, S. Itoh, T. Maruyama, and M. Badger Heat stress causes inhibition of the de novo synthesis of antenna proteins and photobleaching in cultured Symbiodinium PNAS, March 18, 2008; 105(11): 4203 - 4208. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. L. Wulff Sponge Systematics by Starfish: Predators Distinguish Cryptic Sympatric Species of Caribbean Fire Sponges, Tedania ignis and Tedania klausi n. sp. (Demospongiae, Poecilosclerida). Biol. Bull., August 1, 2006; 211(1): 83 - 94. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. F. Little, M. J. H. van Oppen, and B. L. Willis Flexibility in Algal Endosymbioses Shapes Growth in Reef Corals Science, June 4, 2004; 304(5676): 1492 - 1494. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
W. W. Toller, R. Rowan, and N. Knowlton Repopulation of Zooxanthellae in the Caribbean Corals Montastraea annularis and M. faveolata following Experimental and Disease-Associated Bleaching Biol. Bull., December 1, 2001; 201(3): 360 - 373. [Abstract] [Full Text] [PDF]
|
|
| |||||||||||
