Introduction

An alarming increase in diseases of corals has been reported in the last twenty years (Gardner et al. 2003). Since corals are the main structural components of coral reefs, these diseases may threaten this ecosystem in general (Bellwood et al. 2004). Of the coral diseases reported thus far, aspergillosis of sea fans is one of the best understood. This disease was first reported as causing tissue mortality in the Caribbean sea fans Gorgonia ventalina and Gorgonia flabellum (Nagelkerken et al. 1997). Aspergillus sydowii was identified as the causative agent of aspergillosis based on sequences of the 18S nuclear ribosomal gene, morphology, and Koch’s postulates (Smith et al. 1996; Geiser et al. 1998). Three sources of inoculum have been proposed: (1) spores or hyphae of A. sydowii associated with terrestrial runoff or airborne dust; (2) physical contact between healthy and diseased sea fans; and (3) waterborne infection of healthy colonies by hyphae or spores from an infected colony (Jolles et al. 2002). Incidence of sea fan aspergillosis has been related, in part, to a reduction in host defense and to an enhancement of fungal growth as a result of increases in water temperature (Alker et al. 2001; Dube et al. 2002; Kim and Harvell 2004). Sea fans aspergillosis is also characterized by high temporal and spatial variability (Kim and Harvell 2004), which complicates its study.

The sea fan aspergillosis literature states or assumes that A. sydowii is the responsible pathogen, as reported by Smith et al. (1996) and Geiser et al. (1998). However, lack of knowledge of the fungal community associated with healthy sea fans undermines an understanding of the roles of fungi when colonies become diseased. Recently, A. sydowii and other potentially pathogenic fungi were isolated from G. ventalina colonies without signs of aspergillosis (Toledo-Hernández et al. 2007). These findings highlight our ignorance of the basic microbial ecology of sea fans.

This study characterizes the fungal community associated with healthy and diseased G. ventalina colonies to address the following questions. Is the fungal community of healthy G. ventalina colonies different in diversity and composition to that of diseased ones? Within colonies with aspergillosis, does the fungal community of healthy tissue differ from that of diseased tissue? Is A. sydowii part of the resident mycoflora of healthy sea fan colonies? Can aspergillosis also be caused by other species of Aspergillus, or other fungi?

Materials and methods

Isolation and identification of fungi

Two hundred and three G. ventalina colonies from 13 reefs distributed around Puerto Rico were surveyed between 2004–06 (Fig. 1). The number of colonies surveyed per reef ranged between 12 and 20. Of the 203 colonies surveyed, 122 were healthy colonies (i.e., no lesions, no purpling, nor any tissue overgrowth by fouling organisms), while 81 colonies showed signs of aspergillosis (necrotic area surrounded by a purple halo, as defined by Petes et al. 2003). Colonies were selected within an area of at least 1,500 m2 in each reef taking care not to select close neighbors. Diseased colonies were actively searched for, as they were relatively rare. The size of surveyed sea fans was 900–2,000 cm2. Since diseased colonies <900 cm2 were not observed, healthy colonies smaller than this were also excluded. One tissue sample of 2 × 2 cm was collected from each healthy colony while 2 tissue samples of 2 × 2 cm were collected from the diseased colonies, one from a healthy area and the other from an aspergillosis lesion. Tissue samples were immediately placed in individual sterile 50 ml tubes filled with filtered, autoclaved seawater and transported in a cooler with ice for processing within 24 h of collection. To eliminate fungi present on the surface of fan fragments but not colonizing internal tissues, each fragment was surface-sterilized individually in ethanol 70% for 30 s and then washed in filtered, autoclaved seawater for another 30 s. Sodium hypochlorite was not used for sterilization because it dissolves the mesoglea, and may kill some fungi (Koh et al. 2000). Fragments were plated on Glucose Peptone Yeast Agar (GPYA), a standard medium for marine fungi (3 g glucose, 0.3 g yeast extract, 0.3 g peptone, and 20 g agar l−1 filtered seawater), and incubated at 28°C in the dark for one month or until fungi were observed (Toledo-Hernández et al. 2007). Fungi were isolated in pure culture and (where possible) identified by morphology. However, some colonies did not sporulate and morphology-based identification was thus not possible; otherwise, identifications based on DNA sequence data agreed with morphology. Representative strains of each morphospecies were chosen for DNA extraction. For DNA extraction, emerging fungi were transferred to liquid Potato Dextrose (PD) medium made with filtered, autoclaved seawater. (PDA is more nutrient-rich than GPYA; it is less useful as an isolation medium because it encourages the growth of bacteria, but extracts grown on PDA yield more DNA than GPYA.) After five days, DNA was extracted using a Plant Mini Extraction Kit (Qiagen Sciences). The nuclear ribosomal ITS region was amplified using primers ITS1F and ITS4 and annealing temperature of 52–56oC (White et al. 1990) and sequenced. Sequences were corrected using Sequencher and the most similar sequences in GenBank were found using BLAST searches. When the top three matching BLAST hits were from the same species and ≥95% similar to the query sequence, this name was assigned to the culture.

Fig. 1
figure 1

Map of Puerto Rico showing collection sites. LP Luis Peña, CR Carlos Rosario, CB Culebrita, PT Punta Soldado, VI Vieques Island, ESC Escambrón, PN Piñones, IK Icacos, IP Isla Piñero, HU Humacao, JO Jobos, MO Mona Island

Full size image

Data analysis

Estimates of the number of fungal species in healthy tissue from healthy colonies (HH), and healthy and diseased tissues from diseased colonies (HD and DD, respectively) were made using the Chao2 and Jackknife1 estimators in EstimateS 8.0.0 (Colwell 2006). A richness/species abundance coefficient Bray-Curtis (S ab ) was estimated based on presence/absence matrix of fungi isolated from the three tissue types. An analysis of similarity (ANOSIM) and the contribution of each species to the average Bray-Curtis dissimilarity among the three tissue types (SIMPER) were performed using the statistical software PRIMER for Windows version 5.2.9. For this analysis 129 samples were used (no fungi were isolated from the other 74 colonies).

Results

Fungal diversity

Fifteen fungal genera, 30 species, and one unknown isolate with affinities to the Helotiaceae were isolated from the 203 G. ventalina colonies surveyed (Table 1). Aspergillus was the most common and the most diverse genus with 12 species and 64 isolates (Table 1). The second most common genus was Penicillium with 5 species and 33 isolates. Most of the remaining genera occurred as singletons and doubletons. Twenty-two taxa are new reports for G. ventalina and gorgonian corals in general: Aspergillus aculeatus, Aspergillus melleus, Aspergillus ochraceus, Aspergillus tamarii, Aspergillus terreus, Aspergillus versicolor, Candida sp., Chalaropsis sp., Cladosporium cladosporioides, Helotiaceae, Hypocrea lixii, Nectria haematococca, Nectria gliocladioides, Penicillium chrysogenum, Penicillium minioluteum, Pichia guilliermondii, Stachybotrys chlorohalonata, Trichoderma harzianum, and Tritirachium spp. For some fungi, identification inferred from BLAST searches could not distinguish between closely related species. This could reflect limitations of the sequences currently available in GenBank or taxonomic problems in certain species groups (i.e., Aspergillus flavus vs. Aspergillus oryzae (Geiser et al. 1998, 2000).

Table 1 Fungi isolated from Gorgonia ventalina as identified by ITS sequences
Full size table

Variation in fungal community structure among tissue types

Species richness varied among the three tissue types. HH (healthy tissue from healthy colonies) was highest (25 taxa and 88 isolates, 6 of which could not be identified) followed by DD (diseased tissue from diseased colonies, 15 taxa and 32 isolates, 13 of which could not be identified) and HD (healthy tissue from diseased colonies, 15 taxa and 22 isolates, 6 of which could not be identified) (Table 1). Although no asymptote was reached in any tissue type, the Chao2 and Jackknife 1 indexes predicted the highest species richness for HH (Fig. 2). This suggests that the highest fungal species richness in HH was not due to greater sampling effort, but to a natural pattern.

Fig. 2
figure 2

Species accumulation curves of Chao2 and Jackknife1 estimators for fungi isolated from healthy tissue of (a) healthy Gorgonia ventalina colonies, (b) healthy tissue, and (c) diseased tissue of diseased colonies

Full size image

The most common fungi differed among tissue types. A. flavus, A. sydowii, and Penicillium citrinum were the most frequently isolated fungi in HH (Table 1; Fig. 3). A. sydowii, was only isolated from healthy colonies, contrary to expectations. In contrast, A. flavus was the most common fungus in HD followed by C. cladosporoides (Table 1; Fig. 3). In DD, A. flavus, P. citrinum, and Tritirachium spp. were the most frequently isolated fungi (Table 1; Fig. 3). Of all the isolated fungi, 5 were common to all tissue types, 3 were shared exclusively between HH and HD, 6 were shared exclusively between HH and DD, while no fungi were shared exclusively between HD and DD (Table 1). Moreover, 9 fungi were exclusively isolated from HH (including A. sydowii), 7 were exclusively isolated from HD, and 4 fungi from DD (Table 1, Fig. 3). The ANOSIM test showed significant differences in fungal communities among tissue types (r = 0.034; P = 0.002). The SIMPER analysis showed 93.6% dissimilarity between HH and HD, 90% dissimilarity between HH and DD, and 91.3% dissimilarity between HD and DD. HH was distinguished from HD and DD by the greater relative abundance of A. flavus, A. sydowii, and P. citrinum whereas HD was distinguished from DD by the greater relative abundance of A. flavus, P. citrinum, and C. cladosporiodes. Only one fungus was more common in diseased tissue than in other tissue types: Tritirachium sp. (Table 1).

Fig. 3
figure 3

Pie diagram illustrating the observed frequencies in percentages of fungi isolated from healthy tissue from (a) healthy Gorgonia ventalina colonies, (b) healthy tissue, and (c) diseased tissue from diseased colonies

Full size image

Discussion

Fungal diversity associated with gorgonians

This study, in conjunction with those of Koh et al. (2000) and Toledo-Hernández et al. (2007), shows that sea fans have a diverse and variable fungal community. Overall, 65 fungal taxa have been identified from 11 gorgonian species using culture-dependent techniques. Sixteen fungal genera and 51 species (including two yeasts) were isolated from 10 species of gorgonian corals in Singapore (Koh et al. 2000). All of these fungi were new reports for gorgonians. Eight genera of fungi and 15 species were isolated from the Caribbean sea fan G. ventalina (Toledo-Hernández et al. 2007). Four of these genera, Rhodotorula, Stachybotrys, Gloeotinia, and Xylaria, and an unknown fungus with affinities to the Xylariales were new reports for any gorgonian coral. In this study, at least 35 fungal species from 15 genera were identified from 203 G. ventalina colonies. Twenty-three of these species are new reports for G. ventalina or any other gorgonian coral. Here as well as in previous studies (Koh et al. 2000; Toledo-Hernández et al. 2007), Aspergillus and Penicillium were the most frequently isolated and diverse genera. Aspergillus and Penicillium also appear to be common in other marine invertebrates such as scleractinian corals (Kendrick et al. 1982; Priess et al. 2000) and sponges (Höller et al. 2000). Some of these species (e.g., A. sydowii) are believed to be essentially terrestrial organisms, capable of growing in the sea but incapable of sporulating there (Smith et al. 1996). However, their presence in many marine invertebrates suggests they are ubiquitous in the marine environment.

Differences in fungal diversity between healthy and diseased sea fans

This study shows that (1) the fungal community of healthy sea fans is distinct from, and more diverse than, that of diseased fans and (2) within afflicted colonies the fungal community of lesions is different and more diverse than that of healthy tissue. Previous studies have reported differences in bacterial communities between healthy and diseased tissue from diseased scleractinian corals (Frías-López et al. 2002; Casas et al. 2004; Breitbart et al. 2005). However, these studies did not include healthy colonies, which were found here to be distinct from healthy tissue in diseased colonies. Only one previous study has contrasted the microbial community between healthy and diseased colonies as well as between healthy and afflicted tissue within diseased colonies (Pantos et al. 2003). Bacterial communities in Montastraea annularis followed the same pattern observed here for fungi in G. ventalina: healthy tissue in healthy colonies was most diverse, healthy tissue from diseased colonies was least diverse, and diseased tissue was intermediate (Pantos et al. 2003). There are three plausible explanations why healthy tissue from diseased colonies showed decreased microbial richness compared to diseased tissue. First, the reduction of microbial diversity in diseased colonies may be directly caused by an increase in dominance of the pathogen (Ward et al. 2007). Second, environmental changes may reduce microbial diversity, making the host more susceptible to pathogens (Pantos et al. 2003). Third, the colony may have a generalized physiological response to a local infection, which could affect commensal microorganisms as well as pathogens. Similarly, localized infections have a general effect on reproduction: even small lesions affect reproductive success of the whole colony (Petes et al. 2003).

Is A. sydowii a pathogen?

The most striking result of this study is that A. sydowii, the reported causal agent of sea fan aspergillosis, was not isolated from diseased G. ventalina colonies, but only from healthy colonies. These observations cast doubt on the role of A. sydowii as the pathogen (either primary or opportunistic) of sea fan aspergillosis, at least in the colonies studied. The most common isolates (e.g., A. flavus and P. citrinum) were found in both healthy and diseased colonies, suggesting they may be opportunistic pathogens capable of causing aspergillosis.

These two species are prime suspects. A. flavus is an opportunistic pathogen of insects, bird, humans, and plants (Yu et al. 2005) and produces secondary metabolites called aflatoxins that are toxic, carcinogenic, teratogenic, and immunosuppressive in animals when ingested (Pitt 2000). P. citrinum produces the toxin citrinin, associated with renal toxicity in animals (Malmstrom et al. 2000). It is not known if any of these mycotoxins are produced in vivo in sea fans, but if they are, it is possible they play a role in the disease.

Another possible pathogen is Tritirachium sp., the only species that was more common in diseased tissue than in other tissue types. Tritirachium has not been reported from corals previously, but based on phylogeny, this fungus might be expected to be a pathogen: it belongs to the Clavicipitaceae, all members of which are obligate pathogens of invertebrates or plants.

Alternatively, sea fan aspergillosis may be caused by pathogens other than fungi. However, little or no effort has been devoted to study the bacteria, protists, and viruses associated with sea fans, perhaps because A. sydowii is widely assumed to be the primary pathogen. Controlled infection experiments are needed to test the pathogenicity of fungi and other microorganisms found in sea fans.

Methodology and sampling issues

In this study, fungi were cultured before being identified, so any fungi that did not grow in culture were overlooked. Estimates of diversity presented here therefore under-represent total fungal diversity, though it is impossible to say by how much. Studies comparing culturable and nonculturable fungi in other communities suggest that nonculturable species are a much smaller percentage of total diversity for fungi than for bacteria. In fungal endophytes of pine, for example, culturing revealed more major groups of fungi than direct amplification, and was more effective at revealing species richness of certain groups of fungi (Arnold et al. 2007). Since each technique revealed some organisms not revealed by the other, a combination of direct amplification and culturing is the optimal approach for fungal diversity studies (Rohwer et al. 2001; Bayman 2007). Direct amplification of fungi and other microorganisms from sea fan tissue would clearly reveal more species. Nevertheless, this limitation does not affect the most important conclusion of this study, that regarding A. sydowii. Like most species of Aspergillus, A. sydowii grows readily in culture under a wide range of conditions. If anything, culture-based studies may overstate the importance of Aspergillus and Penicillium relative to fungi that do not grow so rapidly and prolifically on agar media.

The present study demonstrates that G. ventalina colonies contain a large, diverse, and mostly unknown fungal community. It also shows that the fungal community of an entire sea fan colony is affected even when only a relatively small portion of the colony suffers from aspergillosis; the fact that this pattern was also observed in bacterial communities in scleractinian corals (Pantos et al. 2003) suggests that it may be a general phenomenon, though the cause is unclear. This study did not find A. sydowii associated with diseased colonies, suggesting that A. sydowii is not the pathogen, at least in Puerto Rico. The presence of several common fungi in both diseased and healthy colonies suggests that aspergillosis may not always be a result of the arrival of new pathogens, but a change in population size or pathogenicity of fungi already present. The importance of environmental stress in the development of aspergillosis has been discussed (Lesser et al. 2007). However, population dynamics of microorganisms in corals and the interaction between microbial communities and stress are not understood.

Initial reports that A. sydowii was the pathogen responsible for aspergillosis in Gorgonia were no doubt accurate (Smith et al. 1996; Geiser et al. 1998); they reflected pathogens of gorgonians in certain areas at certain times. However, subsequent articles generalized these results to mean that A. sydowii was the cause of aspergillosis of gorgonians throughout the Caribbean, which is clearly not the case. Given that it is not clear that Aspergillus, or any fungus, is the sole pathogen, calling this disease aspergillosis is an oversimplification. Aspergillosis of sea fans has been viewed as one of the best-understood coral diseases; the fact that it turns out to be more complicated than previously thought implies that other coral diseases may be so as well.