Antimicrobial activity of extracts of Caribbean gorgonian corals

Marine Bioiogy (1996) 125:411 419

9 Springer-Verlag 1996

P. R. J e n s e n 9 C. D . H a r v e l l 9 K. W i r t z 9 W . F e n i c a l

Antimicrobial activity of extracts of Caribbean gorgonian corals

Received: 13 September 1995/Accepted: 13 October 1995

Extracts of 39 species of Caribbean gorgonians were tested for antimicrobial activity against 15 strains of marine bacteria. The bacteria consisted of three opportunistic pathogens, Vibrio parahaemolyticus, Leucothrix mucor, and Aerococcus viridans, and 12 strains isolated from either healthy or decayed gorgonians. Overall, only 15% (79 out of 544) of the tests resulted in antibacterial activity with 33 % (13 out of 39) of the gorgonians inhibiting only one bacterial strain and 23% (9 out of 39) showing no activity. The extracts of four Pseudopterogorgia species showed relatively high levels of activity, inhibiting 43 to 86% of the bacterial strains. The potency of the active Pseudopterogorgia species was variable, however, and three additional Pseudopterogorgia species were inactive against all bacterial strains. With the exception of one sensitive strain, Vibrio species were resistant to gorgonian metabolites. Our results indicate that organic extracts of most Caribbean gorgonians do not possess potent, broad-spectrum antibacterial activity inhibitory to the growth of opportunistic marine pathogens and bacteria associated with healthy and decayed gorgonian surfaces. These findings suggest that the inhibition of bacterial growth is not the primary ecological function of gorgonian secondary metabolites and that bacteria may not be important selective agents in the evolution of gorgonian secondary chemistry. Abstract

Communicated by M.F. Strathmann, Friday Harbor P.R. Jensen ( ~ ) . W. Fenical Center for Marine Biotechnology and Biomedicine, Scripps Institution of Oceanography, University of California at San Diego, La Jolla, California 92093-0236, USA C.D. HarvellK. K. Wirtz Section of Ecology and Systematics, CorneI1 University, Ithaca, New York 14853, USA

Introduction Microorganisms such as bacteria, fungi, and viruses can be important agents in the evolution of secondary metabolites. This has certainly been the case with terrestrial plants, for which adaptive antimicrobial chemical defenses are well documented (Levin 1976). Among the best examples are the phytoalexins, antifungal substances of relatively low molecular weight, often phenols or terpenoids. Phytoalexins are induced locally in plants next to a wound site, and appear in appreciable quantities after initiation of an infection (Levin 1976; Creasy 1985). Despite the susceptibility of marine invertebrates to bacterial pathogenesis (e.g. Sparks 1985), the rote of invertebrate metabolites in antimicrobial chemical defense, and the extent to which microorganisms have influenced the evolution of these metabolites, is unknown. The lack of evidence for antimicrobial chemical defense in marine invertebrates is somewhat surprising, considering that in the last three decades thousands of novel secondary metabolites have been discovered from these organisms (Faulkner 1994 and references cited therein). Many of these substances display potent antimicrobial activities against human pathogens. If these activities are indicative of an antimicrobial chemical defense, it would help explain the high incidence and evolutionary significance of antimicrobial compounds in certain groups of soft-bodied, marine invertebrates including sponges, gorgonian corals, and ascidians. At present, however, a correlation between biomedically relevant antimicrobial activity and antimicrobial chemical defense in marine invertebrates has not been established. The potential mechanisms by which an invertebrate chemical defense can affect microorganisms include (alone or in combination): (1) toxicity (i.e., antibiotic activity: cell death or the inhibition of cell growth); (2) the induction of a negative chemotactic response; (3) the attraction of favorable epibionts to

412

the competitive exclusion of deleterious microbes. In the case of toxicity, standard agar disc-diffusion antimicrobial assays should be a good predictor of this defense mechanism if the following criteria are met: (1) ecologically relevant microbes are employed; (2) the active materials can be extracted from the animal; and (3) the concentrations at which the extracts are tested reflect the metabolite concentrations experienced by microorganisms in nature. These objectives are difficult to achieve, however, as the potential targets of invertebrate antimicrobial chemical defense are unknown, and may vary from species to species. It is also difficult to assess the distribution of antimicrobial compounds within the surface tissues of the producing organism and the extent to which environmental microbes come in contact with these substances. Given these uncertainties, antimicrobial assays utilizing taxonomically diverse marine microorganisms, including invertebrate pathogens, should provide a relatively accurate method by which to detect the existence of toxic chemical defenses. Testing extracts at concentrations equal to those obtained from the invertebrate tissues also provides a rational approach for antibiotic testing. The results of tests performed using natural extract concentrations, however, must be interpreted with caution as extract diffusion-rates in agar undoubtedly exceed the rates at which metabolites are released from animal tissues in nature. Although there is little solid evidence documenting the existence of antibacterial chemical defenses in marine invertebrates, there is ample evidence that invertebrate crude organic extracts can inhibit the growth of marine bacteria. Sponge extracts have been the most extensively studied in this regard, and have shown varying levels of antibacterial activity (Burkholder and Ruetzler 1969; Berquist and Bedford 1978; McCaffrey and Endean 1985; Thompson et al. 1985; Amade et al. 1987). In a number of cases, a correlation between sponge antimicrobial activity and reduced surface macrofouling was used to suggest that antibacterial activity may be indicative of antifouling chemical defense (e.g. McCaffrey and Endean 1985; Thompson et al. 1985). It should be noted, however, that this correlation is not always observed, as the extracts of some heavily fouled sponges are antibiotic (McCaffrey and Endean 1985). Like sponges, gorgonian octocorals are a rich source of unusual secondary metabolites, many of which display biomedically relevant biological activity. The ability of gorgonian metabolites to inhibit the growth of marine bacteria, however, has been documented in only a limited number of cases. The first such evidence was obtained by Burkholder and Burkholder (1958), who in a subsequent study suggested that antimicrobial activity may be responsible for the lack of surface fouling on gorgonians (Burkholder 1973). Recently, Kim (1994) reported that the extracts of eight gorgonian species exhibited low levels of bacterial inhibition and concluded that the ability to control microbial populations

is an important ecological characteristic of these corals. Gorgonians as a group, however, have not been comprehensively surveyed, and the extent to which organic substances produced by these chemically rich octocorals inhibit the growth of co-occurring and potentially pathogenic marine bacteria remains unclear. Although little is known about the ecological effects of gorgonian secondary compounds on marine bacteria, their effects on other microorganisms are better documented. For example, there is evidence that metabolites produced by the Pacific gorgonian Muricea fruticosa may reduce surface fouling by the marine diatom Phaeodactylum tricornutum (Bandurraga and Fenical 1985). Targett (1988) reported similar results, showing that extracts of the gorgonians Pseudopterogorgia americana and P. acerosa inhibit the growth of the pennate diatom Nitzchia sp., and substances from Leptogorgia virgulata and L. setacea inhibit growth of the marine diatom Navicula salinicola (Targett et al. 1983). Some gorgonian secondary metabolites have also been shown to inhibit motility in marine flagellates (Ciereszko and Guillard 1989). In terms of effects on invertebrates, it has been demonstrated in at least three studies (Standing et al. 1984; Rittschof et al. 1985; Gerhart et al. 1988) that low molecular weight organic substances isolated from gorgonian octocorals inhibit larval barnacle settlement (see Pawlik 1992 for a review of the chemical ecology of invertebrate settlement). The objectives of this study were to survey Caribbean gorgonian extracts for their ability to inhibit the growth of ecologically relevant marine bacteria and to determine if gorgonian antibiotic activity varied as a function of the type and source of the bacteria tested. The gorgonians studied represented a taxonomically comprehensive group of the most abundant species found in the Caribbean Sea. Extract antibiotic activity was assessed using standard agar disc-diffusion assays against opportunistic marine pathogens and strains isolated from the surfaces of either healthy or decayed gorgonians. We examined gorgonian antibiosis, because extracts of these animals are chemically rich and significant ichthyodeterrent activities have been reported from both crude extracts and purified compounds (Pawlik et al. 1987; Harvell et al. 1988; Harvell and Fenical 1989; Fenical and Pawlik 1991; Van Alstyne and Paul 1992; Harvell et al. 1996). If gorgonians constitutively produce broad-spectrum antibiotics as an antibacterial chemical defense, we would expect to detect these substances in bioassays against taxonomically diverse marine bacteria.

Materials and methods Collection Thirty-nine gorgonian species representing 14 genera were collected during a June 1991 cruise of the R.V. "Columbus Iselin" offthe coast

413 Table 1 Description of bacteria used in antimicrobial assays. Pathogens were purchased from American Type Culture Collection (ATCC); remaining strains were isolated from either healthy Briareum asbestinum or decayed gorgonians ( - negative; + positive; FAA fatty acid analysis reported as similarity index to closest species in database; Morph/biochem strain identity based on colony morphology, cell size, shape, motility, oxidase and catalase production, fermentation of glucose, and oxygen requirements)

Strain and bacterial type 1] 2 3

Gram reaction m

pathogen m

4! o] 5 Briareum

+ +

10 ] 11 12 13 14 15

m

decayed gorgonian

of Belize, Central America. The gorgonians were collected from various depths, by SCUBA divers at six different locations along the Belizian barrier reef. Attempts were made to collect the most abundant genera observed. Gorgonian species were collected only once, and in most cases three replicate colonies of each species were collected in the same habitat. Replicate colonies were separated by at least 3 m to enhance sampling from different clones. All gorgonians collected were identified to species, with the exception of eight that could, with confidence, be identified to only the genus level. Extracts of unidentified species within any one genus were verified as taxonomically distinct, and the replicates for each species verified (qualitatively) as chemically homogeneous, by silica-gel thin-layer chromatography (TLC). In two cases, species collected in the same habitat varied in either color (Pterogorgia citrina, yellow and purple form) or morphology (Pseudopterogorgia hummelincki, curly and non-descript form). TLC analyses of these growth forms indicated significant chemical variation, so three replicates of each form were collected and treated as separate species. In a third case, extracts of Plexauraflexuosa collected from two sites differed in chemical composition, hence three replicates from each location were analyzed as separate species.

Extraction Samples were air-dried in the sun for at 1east 2 h, following which branch tips (which contain a high ratio of organic tissue to endoskeleton) were chopped into small pieces. In our experience, gorgonian secondary compounds are stable to mild heat and are not photochemically reactive; thus it is unlikely that antibiotic activity was lost due to the extraction procedure. To guard against the possibility that the drying and extraction procedures generated low-level antibiotic activity due to metabolite degradation, the data were interpreted conservatively (see following subsection and "Discussion"). All replicates were processed individually; to prepare extracts of similar concentration a standard volumetric method was used, as weights could not be accurately measured aboard ship. By this method, dried gorgonian was added to 45 ml of 20% methanol in dichloromethane until the volume displaced equaled 50 ml. This resulted in a 10% (v/v) solution. Gorgonians were extracted at room temperature for 1 h, filtered, and the extracts were concentrated by evaporation. Concentrated extracts were subsequently diluted to 5 ml (the original dry volume of the gorgonian tissue) with methanol, and kept at - 2 0 ~ prior to antimicrobial testing.

m

m

m

Species

Method

Vibrio parahaemoIyticus Leucothrix mucor Aerococcus viridans

ATCC27969 ATCC25906 ATCC10400

Vibrio campbellii Unidentified Vibrio haloplanktis Vibrio parahaemolyticus Unidentified Unidentified

FAA0.44 FAA/no match FAA/0.242 FAA/0.926 FAA/no match FAA/no match

V ibrio fluvialis/filrnissii Vibrio haloplanktis Similar to Strain 6 Unidentified Similar to Strains 10-15 Vibrio furnissi

FAA/0.571/0.548 FAA/0.393 Morph/biochem FAA/no match Morph/biochem FAA/0.701

Antimicrobial assay Gorgonian extracts were tested for antimicrobial activity against a panel of 15 marine bacteria (Table 1). This panel included three strains known to cause disease in marine invertebrates (Sparks 1985): Vibrio parahaemolyticus (ATCC 27969), Leucothrix mucor (ATCC 25906), and Aerococcus viridans (ATCC 10400), as well as six strains isolated from the surfaces of healthy Briareum asbestinum, and six strains isolated from the surfaces of gorgonians that had been enclosed in jars of sterile seawater and allowed to decompose. Ten of the 12 strains isolated from healthy and decayed gorgonians were subjected to fatty acid analysis (Microbial Identification, Inc., MIDI, Newark, Delaware). Six of these strains (4, 6, 7, 10, 11, and 15) had fair to excellent matches with reference species within the genus Vibrio. Two strains (12 and 14) not subjected to fatty acid analysis had similar morphological and biochemical characteristics to strains identified as Vibrio species. The fatty acid profiles of the four remaining strains (5, 8, 9, and 13), including two Gram-positive strains (8 and 9) isolated from healthy B. asbestinum, did not match any of the reference strains in the MIDI library. Bacterial types were defined as either pathogens, isolates from healthy B. asbestinum, or isolates from decayed gorgonians. One strain (13) grew poorly in the antimicrobial assay medium; the data from this strain are shown (see Table 2), but were not included in any analyses because they are incomplete. Standard agar disc-diffusion antimicrobial assays (Barry 1980) were performed by inoculating each bacterial strain onto the surface of an agar medium (0.25% peptone, 0.15% yeast extract, 0.15% glycerol, 75% seawater, 25% deionized water). The inocula were spread with a sterile glass rod so that a film of the test strain covered the agar surface. Gorgonian extracts (three replicates for most species) were tested by pipetting 30 ~tl onto 6.5 mm x 1.0 mm circular paper discs (disc vol = 33 ~tl),allowing the solvent to evaporate, and placing the discs onto the surface of the inoculated agar. Assays were run until the bacteria developed a confluent film of growth on the agar surface. For most strains, this occurred in 1 d, however some strains required up to 5 d to reach confluency (duration of test was used as a co-variate in the statistical analyses). Solvent controls were not run, however the inactivity of most extracts indicates that sufficient time was allowed for solvent evaporation. Areas of inhibited microbial growth were observed as clear halos (zones) surrounding the discs. Zone diameters were measured, and 6.5 mm subtracted (representing the diameter of the paper disc). Gorgonian species for which at least two replicate extracts produced zones >5 mm were considered to possess antibacterial activity. Antimicrobial activities were calculated for each gorgonian species

414 as the mean zone diameter of the replicate extracts (minus the paper disc). The mean zones of inhibition for the active species are underlined in Table 2. The statistical analysis of the data were conducted in two parts. To examine the incidence of antibacterial activity, we first conducted a categorical analysis (G-test for analysis of frequencies: Sokal and Rohlf 1981) comparing the proportion of species which either did or did not display antibacterial activity as defined above. Second, ANOVAs were performed to determine if gorgonian species and genera varied in antibiotic activity. This was followed by Duncan's multiple-range tests to determine which active species and genera differed significantly. Because the incidence data were discontinuous, had a large number of zeros, and were not normally distributed, we included in the second part of the analysis only those gorgonian species and genera which met our minimum requirement for antibiotic activity (at least two replicate extracts producing zones >5 ram) against at least one of the test bacteria. This partial data set meets the requirements for parametric analyses. Pseudopterogorgia nanna was excluded from these analyses because only two replicates were collected.

Because of the large number of combinations to be tested, the shipboard assays were done in a series of four runs, each with roughly 30 extracts tested across a panel of 14 bacterial strains. The duration of the runs varied slightly, and was included as a time covariate in the ANOVA model because bacterial growth rates could affect antibiotic susceptibility. An additional shipboard run re-tested roughly 30 extracts on three strains which had grown insufficiently to produce results in earlier runs. Extracts were tested against Leucothrix mucor in a sixth run at Scripps Institution of Oceanography.

Results

Organic extracts of 39 gorgonian species, representing 14 genera, were tested for antimicrobial activity against 15 strains of marine bacteria (Table 2). One of the 15

Table 2 Gorgonian antimicrobial activity (mean diam, mm) against Bacterial Test Strains 1 to 15 (grouped by type). Active interactions (underlined) are those for which at least two replicate extracts produced zones >5 ram. Bacterial Test Strain 13 was not included in calculation of % active (or any data analyses) because this strain grew poorly in assay medium, resulting in no data (nd) for many of tests Gorgonian

Pseudopterogorgia acerosa P. americana P. rigida P. nanna P. bipinnata P. kallos P. elisabethae P. hummelincki (Morph 1) P. hummelincki (Morph 2) PseudopIexaura crucis P. porosa P. flagellosa Muriceopsis flavida Gorgonia ventalina G. mariae PIexaura homomalla P. flexuosa vat. A P.flexuosa vat. B Carijoa riisea Eunicea mammosa E. calyculata E. tourneforti Eunicea sp. A Eunicea sp. B Eunicea sp. C Eunicea sp. D Briareum asbestinum B. asbestinum (encrusting) Muricea muricata M. elongata M. pinnata Plexaurella dichotoma P. nutans Pterogorgia anceps P. citrina (yellow) P. citrina (purple) Erythropodium caribaeorum Iciligorgia schrammi Ellisellidae sp.

Pathogens

Briareum asbestinum isolates

1

2

3

2.7 1__1 0 0 0 0 0 0 0 0 0 0 0 0 0 3.5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 0

5.3 9.7 7.7 0 L0 6.7 0 0 0 0 0 0 0 3 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

1 1 . 3 10.8 5.7 3.3 5.2 5.2 0 6S 3 10.3 4.3 0 6.2 3 0 11.8 0 = 0 0= 0 0 0 2.7 0 nd 0 0 7 0 0 5 0 0 3.2 0 0 0 0 0 9.j 0 2.3 0 0 0 0 0 0 0 0 0 0 0 0 9.8 0 0 2.8 0 0 0 0 0 2.7 0 0 0 0 0 0 0 0 0 0 0 0 0 0 2.8 0 0 0 0 0 0 0 0 0 0 0 0 0 0 3 0 0 3.2 0 0 0 0 0 0 0 0 5.J 0 3.2 4.2 0 0 0 0 0 0 0 0

a Data available for one replicate only

4

5

Decayed gorgonian isolates

(% active)

6

7

8

9

10

ll

12

13

14

15

5.7 0 7.2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

6.3 2.5 3.3 0 0 0 0 0 0 0 0 0 0 2.5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

11.2 14 8.3 .5 6.2 _9 0 2.5 0 5.3 5.7 3.3 2.5 _8 0 6.2 0 0 0 6.5 5.5 0 0 0 0 2.7 0 0 0 0 0 0 2.8 0 0 0

9.5 8.2 14.5 _9a 12.5 L5 0 0 0 16.7 16.5 13.3 0 5.2 0 2.8 0 0 11.2 5.2 3.3 0 3 6.2 10.7 7.J 2.3 5 0 0 0 0 0 0 0 0

0 2.7 0 0 0 0 2.7 0 0 0 0 0 0 0 0 0 0 2.5 5.2 0 0 0 0 0 0 0 0 0 6 3.2 0 0 0 0 2.5 0

_6 0 2.5 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

5.2 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0

nd 9,27 nd nd _9 16.5 10 = 5.5 nd nd nd nd 15 3.7 2.7 0 4.3 0 0 0 0a nd 13.5 0 10 = nd L0 nd 3.5 0 0= 0 0 13.3 6.2 nd

0

0

0

0

0

0

0

11 a

0 0

0 0

0 0

0 0

9 0

0 0

0 0

nd nd

5 0 6.5 8.8 0 0 0 0 0 3.7 0 2.5 0 0 0 0 0 0 0 0 0 5.5 2.7 0 3 0 0 0 0 0 3.7 0 0 0 0 -6 5.7 8.2 2.8

5.7 3.2 2.5 8.J 0 0 0 0 0 2.5 0 2.7 6.7 8.j nd 10.7 8.8 0 6.7 6 5.5 5.2 2.7 0 0 0 5.j 0 3.3 0 5.J 7.7 5.3 0 0 2.8 2.5 _8 3

(86) (43) (43) (43) (29) (29) (0) (0) (0) (21) (21) (7) (7) (29) (0) (14) (7) (0) (29) (21) (14) (14) (0) (7) (7) (7) (7) (7) (7) (0) (7) (7) (7) (0) (0) (14) (7) (21) (0)

415 Table 3 A N O V A for species-level analysis of antibiotic activity (at least two replicates with zone sizes > 5 mm). Type III sum of squares (SS) removes significant effect of time (duration of assay) from hierarchical model (MS mean square) Source

df

S

MS

F

Model Error Corrected Total

43 1217 1260

5757.3 12457.4 18214.7

133.9 10.2

13.08

df

Type III SS

Time Gorgonian species Bacterial strain

F

>F 0.0001

p > F

1 29

70.4 2584.7

6.88 8.71

0.0088 0.0001

13

2333.4

17.53

0.0001

Table 4 Gorgonian species with highest antibiotic potency (mean zone of inhibition), presented in decreasing order of potency. Means with same letter do not differ significantly (Duncan's multiple-range test, p < 0.005) Gorgonian

Mean (mm)

Duncan group

Pseudopterogorgia acerosa P. rigida P. americana P. kallos Gorgonia ventaIina Pseudopterogorgia bipinnata Pseudoplexaura erucis Carijoa riisei

6.3 4.7 4.6

A B B

3

C

2.8 2.7 2.5 2.3

C C C C

strains tested (Strain 13), grew poorly and was not included in any of the data analyses although the results for this strain are presented in Table 2. Of the 544 tests analyzed, 79 (15%) resulted in antimicrobial activity, i.e., at least two replicate extracts produced zones of inhibition > 5 mm in diameter (Table 2). Thirtyeight percent of the antibiotic activities detected could be accounted for by the four most active species, all of

Table 5 A N O V A for genuslevel analysis of antibiotic activity (at least two replicates with zone sizes > 5 mm). Type III sum of squares removes significant effect of time (duration of assay) from heirarchical model (Abbreviations as in Table 3)

df

Source Model Error Corrected

39 1221 1260

df Time Gorgonian genus Bacterialtype a Genus x type

1 12 2 24

which belong to the genus Pseudopterogorgia. These four species (P. acerosa, P. americana, P. rigida, and P. hanna) inhibited 43 to 86% of the test bacteria, and were the only gorgonians to inhibit more than four of the bacterial strains. Extracts of gorgonian species differed widely in their antibiotic activity. The most active species, Pseudopterogorgia acerosa, inhibited 86% of the test strains, while 13 species inhibited only one strain and 9 species showed no activity. Antibiotic potency (estimated by the diameter of the zone of inhibition) differed significantly among species and bacterial test strain (Table 3). The hierarchical model, which accounted for the duration of the assay (i.e., the time required for the test bacteria to form a confluent film on the agar surface), indicates that fast- and slow-growing bacteria showed different susceptibilities to gorgonian extracts. The mean zone diameters of the most active gorgonian species are listed in Table 4. The three gorgonian species with the highest incidence of antibacterial activity were also the most potent. P. nanna was excluded from this analysis because only two replicates were collected. Gorgonian genus and bacterial type (pathogen, Briareum asbestinum isolate, or decayed gorgonian isolate) both significantly affected potency in a hierarchical model (Table 5). Furthermore, the significant interaction term between genus and bacterial type shows that different gorgonian genera tended to inhibit different bacterial types (Table 5). Extracts of Pseudopterogorgia were considerably more potent than all other genera (Table 6). Although extract antibiotic potency varied between gorgonian genera, even the most potent genera did not produce mean activities (zone diameters) of > 5 mm against any of the three bacterial types (Fig. 1). The four bacterial strains inhibited most frequently were Strains 3, 8, 9, and 15. These strains were inhibited by 27 to 41% of the extracts (Table 7). Three of these strains (3, 8, and 9) were Gram-positive cocci; these were the only Gram-positive strains on the test panel. Overall, Gram-positive strains were inhibited in 32% (37 out of 116) of the tests, while Gram-negative strains

SS 4350.9 13863.8 18214.7 Type III SS 697.8 1828.6 165.7 1416.8

MS

F

111.6 11.4

9.83

F

p > F

61.46 13.42 7.30 5.20

0.0001 0.0001 0.0007 0.0001

a Pathogen, Briareum isolate, or decayed gorgonian isolate

>F 0.0001

416 Table 6 Gorgonian genera with highest antibiotic potency (mean zone of inhibition) in decreasing order of potency. Means which share same letter do not differ significantly (Duncan's multiple-range test, p < 0.05) Genus

No. of species

Mean (mm)

Duncan group

Pseudopterogorgia Gorgonia Carijoa Pseudoplexaura lcihgorgia Plexaura Eunicea Erythropodium PlexaurelIa Muricea Muriceopsis Briareum Pterogorgia

5 1 1 3 1 2 6 1 2 2 1 2 3

4.3 2.8 2.3 2.1 t.8 1.1 0.9 0.9 0.8 0.7 0.7 0.6 0.5

A B B, C B, C, D B, C, D, E C, D, E D, E D, E E E E E E

Table 7 Number of gorgonian extracts that inhibited individual bacterial strains. Inhibitory extracts are those in which at least two replicates generated antibacterial zones >5 ram. Pathogens were purchased from American Type Culture Collection, remaining strains were isolated from healthy Briareum asbestinum or decayed gorgonians Strain and bacterial type

1

2 3 4 5 6 7 8 9 10 11 12 14 15

Pathogen

Briar eum

decayed gorgonian

Species

Inhibitory extracts

Vibrio parahaemolyticus Leucothrix mucor Aerococcus viridans Vibrio campbellii

1 (3%) 5 (13%) 9 (27%) 4 (10%) Unknown 2 (5%) V ibrio haIoplanktis 2 (5%) Vibrio parahaemolyticus 1 (3%) Unknown 12 (31%) Unknown 16 (41%) Vibrio fluvialis/furnissi 3 (8 %) Vibrio haIopIanktis 1 (3%) Similar to Strain 6 1 (3%) Similar to Strains 10 15 7 (18%) Vibriofurnissii 15 (39%)

5-

healthy Briareum asbestinum and could not be identified by fatty acid profile (MIDI). The remaining strain was Vibriofurnissii (Strain 15), isolated from a decayed gorgonian. Excluding the four sensitive strains of bacteria, the remaining ten strains (all Gram-negative) were inhibited relatively infrequently, showing sensitivity in 7% (27 out of 390) of the interactions tested. Of these 10 insensitive strains, 8 belong to the genus Vibrio. With the exception of Strain 15, Vibrio species were highly insensitive to gorgonian secondary compounds. The incidence of antimicrobial activity appeared unrelated to the source of the test strain, as marine pathogens were inhibited in 13% (15 out of 116) of the tests, B. asbestinum isolates were inhibited in 16% (37 out of 234) of the tests, and isolates from decayed gorgonians were inhibited in 14% (27 out of 194) of the tests.

[] Pathogens

4-

[] Briareumisolates

g N

[] Decayedgorgonianisolates

3-

,.Q

N

1-

o

.

.

.

.

.

.

.

.

.

.

.

.

.

t

~

n

F

~

-

-

~

Discussionand conclusions

Gorgonian genus Fig. 1 Antibiotic potency of 14 gorgonian genera against three groups of bacteria (bacterial pathogens, bacteria isolated from surfaces of Briareum asbestinum, and bacteria isolated from surfaces of decayed gorgonians). Genus potency reported as pooled mean diameter of zones of inhibition (from Table 2) for all species within a genus against all strains within a bacterial group (Table 1)

were inhibited in 10% (42 out of 427) of the tests. This difference was significant (chi 2 = 28.4, p < 0.001). Of the antibiotic-sensitive strains, Strain 3 is the opportunistic marine pathogen Aerococcus viridans (ATCC 10400); Strains 8 and 9 were isolated from

The overall objective of this study was to determine the extent to which Caribbean gorgonians produce organic soluble antibiotics inhibitory to the growth of cooccurring and potentially pathogenic marine bacteria. Based upon the examination of 39 gorgonian species, representing 14 genera, we conclude that the constitutive production of potent, broad-spectrum antibiotics by this group of invertebrates is not common. In support of this conclusion, only 15% (79 out of 544) of the tests generated antibiotic activity (defined as two or more replicate extracts generating zones of inhibition >5 mm in diameter, Table 2). Furthermore, the antimicrobial potencies (estimated by the diameter of the zone of inhibition) were generally low compared to those reported for other marine invertebrates (e.g.

417

Rinehart et al. 1981). There were, however, four species within the genus Pseudopterogorgia in which extract antibacterial activity was relatively common (43 to 86% of the tests generated antibiotic activity). Although these crude extracts were not potent by pharmacological standards, it remains possible that certain Pseudopterogorgia species produce broad-spectrum antibacterial metabolites that function in antimicrobial chemical defense. Although broad-spectrum antibiotic activity was not detected in most of the Caribbean gorgonians studied, the incidence of activity against Gram-positive strains was significantly greater than against Gram-negative strains. Similar results were obtained in at least three studies of marine sponges (McCaffrey and Endean 1985; Thompson et al. 1985; Amade et al. 1987). One interpretation of these data is to suggest that chemical defenses in some invertebrate taxa selectively target Gram-positive bacteria. This suggestion may be premature, however, as there is no ecological rationale for the maintenance of such a defense. An alternative explanation is provided by the lack of an outer membrane in the Gram-positive cell, the functions of which can include protection from antibiotics (see Costerton et al. 1974). Twelve of the bacterial strains used for antibiotic testing were isolated from either the surface of healthy Briareum asbestinum or from gorgonians allowed to decay in sterile seawater. Eight of these strains were identified as Vibrio species. This is not entirely surprising, considering that the genus Vibrio can represent a large percentage of the colony-forming bacteria associated with coral mucus (Ducklow and Mitchell 1979). The general observation that bacteria commonly associated with gorgonian surfaces are not inhibited by secondary compounds constitutively produced by these invertebrates strengthens our conclusion that the maintenance of a broad-spectrum chemical defense, toxic to co-occurring bacteria, does not appear to be the primary ecological function of gorgonian secondary compounds. It remains possible, however, that gorgonians maintain (alone or in combination) inducible chemical defenses, secondary compounds inhibitory to bacteria not tested in this study, chemical defenses that were not extracted from the animal tissues, and chemical defenses that function by mechanisms other than toxicity as detected by the agar disc-diffusion assay. Although the extracts of most of the gorgonians studied were not antibiotic, it also remains possible that the few relatively active gorgonians indicate that some species produce toxic chemical defenses targeting specific bacterial taxa. We believe our inability to detect broad-spectrum antibacterial activity in most gorgonian extracts accurately reflects the absence of these activities in nature. In support of this conclusion, the test panel included at least three diverse genera representing six species of marine bacteria. At least two additional genera are

most likely represented by the Gram-negative strain (Strain 5) and the two Gram-positive strains (8 and 9) which could not be identified by fatty acid analysis (Table 1). In addition, we believe that the extracts exposed the test bacteria to extremely high concentrations of gorgonian metabolites relative to what would be experienced in nature. The conservative nature of the results can be assumed because: (1) the extracts were tested at concentrations found in dried gorgonian tissue, thereby overestimating natural (wet tissue) concentrations, and (2) gorgonian metabolites undoubtedly diffuse in the agar assay medium at greater rates than they diffuse away from animal tissues. Although the distribution of organic metabolites in the gorgonian tissues was not determined, it is highly unlikely that surface-associated bacteria contact gorgonian metabolites at concentrations equivalent to those encountered by bacteria during the antibiotic assays. If organic materials were released in situ at concentrations equivalent to those coming off the paper discs, it would be difficult to explain how the integrity of gorgonian tissues is maintained in nature. Our conclusions differ from a recent study of eight Caribbean gorgonians in which most extracts were reported to possess antibiotic activity (Kim 1994). The difference between these two studies can be attributed to different interpretations of what constitutes "significant" antimicrobial activity as measured by the agar disc-diffusion assay. We chose a conservative interpretation of the data, and scored as active only those gorgonian species for which at least two replicate extracts generated antibiotic zones >5 mm in diameter. Kim scored zones with an average diameter as low as 0.2 mm as active. If the data reported by Kim were to be analyzed using our criteria, at most one of the extracts would be considered active, and it would be concluded that antibacterial activity was not common in the gorgonians studied. Regardless of the activity criteria used, it must be pointed out that in our study, at most three replicates of any gorgonian were collected (and in most cases all of these from one location), so little can be concluded about the antibiotic activity of individual species throughout the Caribbean. Our conservative interpretation of the antibiotic disc-diffusion assay results is based upon the concern that low levels of activity may be an artifact of the extraction and testing procedures. For example, extracted photosynthetic pigments (derived from endosymbiotic microalgae in the case of gorgonians) and fatty acids can degrade to chlorophyllides and fatty acid hydroperoxides, respectively, both of which can display weak antibacterial activity. When extracts are tested at high concentrations, fatty acids can exhibit weak antibacterial activities by virtue of their mildly acidic properties. Evidence that nutrients can inhibit the growth of marine bacteria (e.g. Buck 1974) suggests that high concentrations of primary metabolites can lead to small antibiotic zones. It is also possible that small

418

zones of inhibition are caused by physical interactions between the extract and the bacteria. The variability inherent to the disc-diffusion assay, which we estimate to be at least _+1.0 mm, must also be considered in the interpretation of antibiotic activity. Our estimates of the precision of these assays are supported by Barry (1980), who reported that zone diameter standard deviations ranged from 0.6 to 1.9 mm when antibiotics were tested against a variety of bacteria. This variability emphasizes the importance of replicate sampling, and makes it difficult to determine the accuracy of small zones if they are outside the assay's detection limits. On the basis of these and other potential complexities (e.g. residual solvent in the extract), we did not consider small, non-reproducible zones of bacterial inhibition indicative of an antimicrobial chemical defense. To gain insight into antimicrobial chemical defense, behavior-oriented bioassays have advantages over tests of extract toxicity. Although not extensively used, behavior-oriented assays have shown that marine bacteria respond chemotactically to algal products (Bell and Mitchell 1972; Kogure et al. 1982), and that echinoderm extracts can enhance or deter bacterial attachment to a surface (Bryan et al. 1996). It has also been shown that organic substances can induce negative chemotaxis at sub-lethal concentrations (Chet et al. 1975), and that non-toxic ascidian extracts can deter bacterial attachment (Wahl et al. 1994). Non-toxic metabolites capable of detering attachment or inducing negative chemotactic responses would not be detected by standard antimicrobial assays, yet could provide effective chemical defense. Expanded studies targeting the microbiological effects of non-toxic, biologically active metabolites may provide new insight into our understanding of chemical ecology at the microbial level. The gorgonian species which displayed antibacterial activity in our study were non-randomly distributed among genera. Species within the related genera Pseudopterogorgia and Gorgonia were among the most inhibitory. One hypothesis to explain this correlation among related species is that a similar class of compound was present in a common ancestor and diversified with colony diversification. The secondary metabolite chemistry of most species of Pseudopterogorgia is well known and does not provide strong support for this hypothesis. A subset of species, P. acerosa, P. bipinnata and P: kallos, do produce similar classes of diterpenoids, but the other species of Pseudopterogorgia produce a wide variety of compounds of different biosynthetic classes. An alternative hypothesis to explain the correlation among related species is that they have independently evolved antimicrobial activity due to a particular, shared susceptibility to microbial attack. This is a difficult hypothesis to evaluate because we know little about microbial-gorgonian interactions, but it certainly is a priority for future research.

The overall low level of antibacterial activity detected in gorgonian extracts is interesting given their rich arsenal of ichthyodeterrent compounds (e.g. Pawlik et al. 1987; Harvell and Harvell et al. 1988, 1996; Fenical 1989; Fenical and Pawlik 1991; Coll 1992; Pawlik 1993). It is noteworthy that the genus Pseudopterogorgia, species of which had the highest incidence of antibiotic activity, stands out both as having a relatively high proportion of biologically active compounds and as being one of the most deterrent to predators. In addition, three of the most antibiotically active gorgonians, P. acerosa, P. rigida, and the genus Gorgonia, are also ichthyodeterrent (Pawlik et al. 1987; Harvell et al. 1988; Van Alstyne and Paul 1992). It should be noted, however, that this pattern has not been consistent, as some highly ichthyodeterrent gorgonians, e.g. Briareum asbestinum (Harvell et al. 1996), Erythropodium caribaeorum (Fenical and Pawlik 1991), Eunicea, and Pterogorgia (Pawlik et al. 1987) showed no corresponding antibacterial activity. Because ichthyodeterrence far surpasses antibacterial activity in prevalence and potency among Caribbean gorgonians, the deterrence of fish predators appears to be a more important ecological function of gorgonian defensive chemistry than the inhibition of bacterial growth. These findings suggest that bacteria may not be an important selective agent in the evolution of gorgonian secondary chemistry; however, additional studies addressing non-toxic antibacterial chemical defenses are warranted. Acknowledgements We thank J. Wheaton for gorgonian identifications, and Microbial identification, Inc., Newark, Delaware, for bacterial fatty acid analyses. We also thank T. Cursar, P.D. Coley, M. Wahl, and three anonymous reviewers for helpful comments on the manuscript. This research, including ship support funding for the use of the R.V. "Columbus Iselin," was provided by the National Science Foundation, Chemistry Division, Grant CHE-9322776 to WF, Oceanography Division; Grant OCE-9012034 to CDH; an NSF-REU supplement to CDH and KW; and by a New York State Hatch grant to CDH.

References Amade P, Charroin C, Baby C, Vacelet J (1987) Antimicrobial activities of marine sponges from the Mediterranean Sea. Mar Biol 94:271-275 Bandurraga MM, Fenical W (1985) Isolation of the muricins. Evidence of a chemical adaptation against fouling in the marine octocoral Muricea fruticosa (Gorgonacea). Tetrahedron 41:1057-1065 Barry AI (1980) Procedures and theoretical considerations for testing antimicrobial agents in agar media. In: Lorian V (ed) Antibiotics in laboratory medicine. William & Wilkins, Baltimore, Md, pp 1 16 Bell W, Mitchell R (1972) Chemotactic and growth responses of marine bacteria to algal extracellular products. Biol Bull mar biol Lab, Woods Hole 143:265 277 Berquist PR, Bedford JJ (1978) The incidence of antibacterial activity in marine Demospongiae; systematic and geographic considerations. Mar Biol 46:215 221

419 Bryan PJ, Rittschof D, McClintock JB (1996) Bioactivity of echinoderm ethanolic body-wall extracts: an assessment of marine bacterial attachment and macroinvertebrate larval settlement. J exp mar Biol Ecol (in press) Buck JD (1974) Effects of medium composition on the recovery of bacteria from sea water. J exp mar Biol Ecol 15:25 34 Burkholder PR (1973) The ecology of marine antibiotics and coral reefs. In: Jones OA, Endean R (eds) Biology and geology of coral reefs. Vol. II. Biology. Academic Press, New York, pp 117-176 Burkholder PR, Burkholder LM (1958) Antimicrobial activity of horny corals. Science, NY 127:1174-1175 Burkholder PR, Ruetzler K (1969) Antimicrobial activity of some marine sponges. Nature, Lond 222:983 984 Chet I, Asketh P, Mitchell R (1975) Repulsion of bacteria from marine surfaces. Appl Microbiol 30:1043-1045 Ciereszko LS, Guillard RRL (1989) The incidence of some cembranolides from gorgonian corals on motility of marine flagellates. J exp mar Biol Ecol 127:205-210 Coll JC (1992) The chemistry and chemical ecology of octocorals (Coelenterata, Anthozoa, Octocorallia). Chem Rev 92:613 631 Costerton JW, Ingram JM, Cheng KJ (1974) Structure and function of the cell envelope of Gram-negative bacteria. Bact Rev 38: 87-110 Creasy LL (1985) Biochemical responses of plants to fungal attack. In: Cooper-Driver G, Swain AT, Conn EE (eds). Chemically mediated interactions between plants and other organisms. Plenum Press, New York, pp 47-80 Ducklow HW, Mitchell R (1979) Bacterial populations and adaptations in the mucus layers on living corals. Limnol Oceanogr 24:715 725 Faulkner DJ (1994) Marine natural products. Nat Product Rep 11: 355-395 Fenical W, Pawlik JR (1991) Defensive properties of secondary metabolites from the Caribbean gorgonian coral, Erythropodium caribaeorum. Mar Ecol Prog Ser 75:1-8 Gerhart DJ, Rittschof D, Mayo SW (1988) Chemical ecology and the search for marine antifoulants. Studies of a predato~prey symbiosis. J chem Ecol 14:1905-1917 Harvell CD, Fenical W (1989) Chemical and structural defenses of Caribbean gorgonians (Pseudopterogorgia spp.). II. Intracolony localization of defense. Limnol Oceanogr 34:382-389 HarvelI CD, Fenical W, Greene CH (1988) Chemical and structural defenses of Caribbean gorgonians (Pseudopterogorgia spp.). I. Development of an in situ feeding assay. Mar Ecol Prog Ser 49: 287-294 Harvell CD, Fenical W, Porcile P (1996) Chemical defenses of a West Indian gorgonian (Briareum asbestinum) against vertebrate, invertebrate, and microbial consumers. Mar Ecol Prog Ser (in press)

Kim K (1994) Antimicrobial activity in gorgonian corals (Coelenterata; Octocorallia). Coral Reefs 13:75-80 Kogure K, Simidu U, Taga N (1982) Chemotactic response of marine bacteria to Skeletonema costatum. Bull Jap Soc scient Fish 48:185 188 Levin DA (1976) The chemical defenses of plants to herbivores and pathogens. A Rev Ecol Syst 7:121-159 McCaffrey EJ, Endean R (1985) Antimicrobial activity of tropical and subtropical sponges. Mar Biol 89:1 8 Pawlik JR (1992) Chemical ecology of the settlement of benthic marine invertebrates. Oceanogr mar Biol A Rev 30:273-335 Pawlik JR (1993) Marine invertebrate chemical defenses. Chem Rev 93:1911-1992 Pawlik JR, Burch MT, Fenical W (1987) Patterns of chemical defense among Caribbean gorgonian corals: a preliminary survey. J exp mar Biol Ecol 108:55-60 Rinehart KL et al. (25 co-authors) (1981) Marine natural products as sources of antiviral, antimicrobial, and antineoplastic agents. Pure appl Chem 53:795 817 Rittschof D, Hooper IR, Branscomb ES, Costlow JD (1985) Inhibition of barnacle settlement and behavior by natural products from whip corals, Leptogorgia virgulata (Lamarck, 1815). J chem Ecol 11:551-563 Sokal RR, Rohlf FL (1981) Biometry. The principles and practice of statistics in biological research. 2nd edn. W.H. Freeman & Co, New York Sparks AK (1985) Synopsis of invertebrate pathology, exclusive of insects. Elsevier, New York Standing JD, Hooper IR, Costlow JD (1984) Inhibition and induction of barnacle settlement by natural products present in octocorals. J chem Ecol 10:823-834 Targett NM (1988) Mlelochemistry in marine organisms: chemical fouling and antifouling strategies. In: Thompson ME, Sarojini R, Nagabhushanam R (eds) Marine biodeterioration, advanced techniques applicable to the Indian Ocean. Oxford & IBH Publishing Co, New Delhi, pp 609 617 Targett NM, Bishop SS, McConnell OJ, Yoder JA (1983) Antifouling agents against the benthic diatom, Navicola salinicola. J chem Ecol 9:817-829 Thompson JE, Walker RP, Faulkner DJ (1985). Screening and bioassays for biologically-active substances from forty marine sponge species from San Diego, California, USA. Mar Biol 88: 11-21 Van Alstyne KL, Paul VJ (1992) Chemical and structural defenses in the sea fan Gorgonia ventalina: effects against generalist and specialist predators. Coral Reefs 11:155-159 Wahl M, Jensen PR, Fenical W (1994) Chemical control of bacterial epibiosis on ascidians. Mar Ecol Prog Ser 110:45 57