Flexibility in Algal Endosymbioses Shapes Growth in Reef Corals

See allHide authors and affiliations

Science  04 Jun 2004:
Vol. 304, Issue 5676, pp. 1492-1494
DOI: 10.1126/science.1095733


The relation between corals and their algal endosymbionts has been a key to the success of scleractinian (stony) corals as modern reef-builders, but little is known about early stages in the establishment of the symbiosis. Here, we show that initial uptake of zooxanthellae by juvenile corals during natural infection is nonspecific (a potentially adaptive trait); the association is flexible and characterized by a change in (dominant) zooxanthella strains over time; and growth rates of experimentally infected coral holobionts are partly contingent on the zooxanthella strain harbored, with clade C–infected juveniles growing two to three times as fast as those infected with clade D.

The recent discovery of the genetically diverse nature of the dinoflagellate genus Symbiodinium (zooxanthellae) that forms symbiotic associations with stony corals raises the possibility that physiological properties and tolerances of reef corals may vary according to the association established. The genus Symbiodinium consists of at least seven clades (A to G) based on sequence analysis of the internal transcribed spacer (ITS) region (15), as well as many genetic types within each clade, referred to as subclades or strains (e.g., C1, C2) (46). In most broadcast spawning corals, zooxanthellae are acquired from the environment in early ontogeny by horizontal transmission and become established in the endodermal cells of coral hosts as an endosymbiosis. This creates an opportunity for the host to establish an association with a variety of symbionts. Indeed, adults of some coral species form associations with more than one Symbiodinium strain according to the local environment (7, 8) or microhabitats within a coral (6, 9, 10). Such polymorphic symbioses suggest that corals within a species may not be physiologically uniform (11) and that the taxonomic identity of the Symbiodinium partner(s) may be as significant as that of the host in determining the physiology of the holobiont (host-symbiont partnership). A recent review (12) highlights our limited understanding of the influence of symbiont type on physiological performance of the holobiont and the importance of understanding potential flexibility in Symbiodinium symbioses in an era of global coral reef deterioration.

Acropora tenuis and A. millepora are broadcast spawning corals with horizontal transmission of symbionts (13) that, as adults, express different specificities for Symbiodinium strains at Magnetic Island (an inshore reef in the central section of the Great Barrier Reef, Australia), where adult colonies of A. millepora contain a Symbiodinium D strain, whereas A. tenuis adults contain Symbiodinium strain C1 and occasionally strain C2 (6, 10). The production of larvae free of zooxanthellae by both species provides the opportunity to observe natural patterns of zooxanthella infection and also to manipulate the strains offered for uptake in controlled experimental conditions to determine the impact of known strains on juvenile growth.

Larvae of A. tenuis were raised from spawned gametes (14) and settled onto tiles (15). Positions of juveniles on the tiles were mapped, and the tiles were then attached to the reef (Nelly Bay, Magnetic Island) in a zone where adult A. tenuis colonies were abundant (15). Thirty juvenile corals were sampled at about 1, 2, 4, and 9 months after settlement. Total DNA (both coral and algal) was extracted from the polyps, and the polymerase chain reaction (PCR) was used to amplify the rDNA ITS1 region of the Symbiodinium genome (6). We identified zooxanthella genotypes by single-stranded conformation polymorphism (SSCP), using reference samples of known genotypes. Contrary to expectations, we found that the apparent specificity for strain C1 observed in adult populations of A. tenuis is not present in the early stages of infection. Two distinct Symbiodinium strains, D and C1, were acquired by juveniles in the first month. In the subsequent 4 months, the relative abundance of these two strains within the symbiosis changed, with a clear increase (from ∼33 to >90%) in the number of juveniles harboring strain D, and a decrease in the number of juveniles harboring strain C1 (from ∼50 to 0%) or a combination of the two strains (from ∼17 to <6%) by 4.5 months (Fig. 1). The dominance of Symbiodinium D in early juveniles of A. tenuis, in contrast to the dominance of Symbiodinium C1 in adults of this species in Magnetic Island populations (6, 10), suggests that there may be “active” selection by the host to maximize symbiont effectiveness that varies with differences in physiological requirements between juvenile and adult corals. For example, corals may have a higher demand for nutrients when they reach reproductive maturity, leading to a preference for one type to meet increased energy requirements. It is possible that Symbiodinium C1 persists in very low densities and is maintained as an undetectable “background” strain, because the SSCP method cannot detect a strain with a relative abundance below ∼5% (16). Evidence supporting this interpretation is provided by zooxanthellae cultures, where cultured strains are often not the same as ones initially identified from the host used to establish the culture (5, 17). Thus, the dynamics of coral-zooxanthellae associations may vary with the changing physiological needs of the host in response to life history stage requirements or ambient environmental conditions.

Fig. 1.

Relative abundance of two Symbiodinium strains in juveniles (in the first 9 months after settlement) and adults (16) of A. tenuis at Magnetic Island. n, number of juveniles genotyped. The rDNA-ITS1 region was successfully amplified for the majority of juveniles in the last three samples (96, 63, and 100% amplified, respectively), but only for 23% of samples collected at week 4, probably because of low densities of zooxanthellae in juveniles at this early stage.

We also examined the impact of locally (i.e., relative to Magnetic Island populations) homologous and heterologous strains of zooxanthellae on growth of coral hosts as a surrogate measure of fitness. Larvae of A. tenuis and A. millepora were raised in sterile (0.5 μm filtered) seawater (15). Strain C1 and D zooxanthellae were isolated from adult corals of the two species, and each strain was added to half of the settled juveniles of each coral species (N > 1000) (15). As in the above study, the positions of juveniles on settlement tiles were mapped, and tiles were attached to the reef. Growth of juveniles was monitored for 6 months as the number of polyps per colony, a more sensitive measure than colony diameter in the earliest stages of coral growth. Juveniles of A. tenuis and A. millepora were each able to establish associations with both homologous and heterologous strains of Symbiodinium. At each sampling time, SSCP analysis verified that juveniles tested for each treatment contained only the strain initially offered (15). Some of the polyps were pale when placed in the field, which suggested that levels of experimental infection were low. Although juveniles gained normal pigmentation, none took up additional types from wild populations of zooxanthellae, which suggests that uptake of zooxanthella types is fixed at an early stage.

We found that juveniles of both species of Acropora grew fastest when associated with Symbiodinium C1 (15) (Fig. 2, A and B). For A. tenuis, the rate of polyp budding in the first 6 months was more than 2 times that in C1 juveniles (mean size = 25 ± 3.4 polyps per colony at 6 months) compared with D juveniles (9 ± 0.9 polyps per colony). Similarly, in A. millepora, growth of C1 juveniles (mean size = 27 ± 1.8 polyps per colony) during the same period was significantly greater than D juveniles (10 ± 1.5 polyps per colony). Faster growth of holobionts infected with Symbiodinium C may reflect a greater contribution of the symbiont to host nutrition through faster rates of population growth inside the host (18). For A. tenuis, the faster growth rates of C1 juveniles may explain why C1 adults are the most common at Magnetic Island (10). In contrast, the dominance of Symbiodinium D, known to be associated with greater thermal tolerance (12) in naturally infected, 6-month-old A. tenuis, may reflect distinct physiological needs of the juvenile holobiont, which recruits into populations at the beginning of summer.

Fig. 2.

Comparison of mean (±SEM) growth among coral juveniles infected with zooxanthellae strain C1 versus D over a 6-month period for (A) A. millepora (n = 2214, 1099, 810, and 203 for C1 juveniles; n = 736, 379, 150, and 13 for D juveniles in November, December, January, and May, respectively), and (B) A. tenuis (n = 2396, 1842, 1261, and 55 for C1 juveniles; n = 1740, 1473, 1026, and 45 for D juveniles in November, December, January, and May, respectively). Fused colonies were excluded. Where error bars are not visible, they are small and hidden by the symbols.

The lack of detectable (by SSCP) acquisition of additional symbiont strains from field zooxanthella populations, when experimentally infected juveniles were reared on the reef, suggests that the temporal window for symbiont acquisition is relatively narrow. The increase in the proportion of A. tenuis juveniles with strain D through time may represent greater mortality of juveniles with clade C, or competitive exclusion of clade C within juvenile hosts, at least to undetectable levels, or alternatively, selective up-regulation of strain D by the host. In combination, our results are consistent with adjustments in the ratio of already coexisting symbiont populations or holobiont types, rather than uptake of additional symbiont types. Symbiont shuffling [sensu (12)] represents a mechanism for rapid acclimatization of the holobiont to environmental change, whereas the lack of specificity in initial uptake of zooxanthellae in early ontogeny demonstrated in our study provides a mechanism for establishing associations with multiple symbionts and, hence, may be adaptive. Use of more sensitive methods for detecting potential background strains would help determine the mechanism underlying changes in detectable symbiont genotypes in adult corals when moved from one light environment to another (1921) and further our understanding of the “adaptive bleaching hypothesis” (ABH), which postulates that hosts may be repopulated by better-adapted algal endosymbionts after bleaching (22).

Our study demonstrates that coralzooxanthella associations are both dynamic and flexible and that algal endosymbionts contribute significantly to physiological attributes of the coral holobiont. As yet, little is known about host factors that contribute to this symbiosis. Further studies are required for a better understanding of the implications of these new findings for the capacity of corals to cope with global climate change.

Supporting Online Material

Materials and Methods

References and Notes

View Abstract

Stay Connected to Science

Navigate This Article