Journal of Invertebrate Pathology 78, 109–118 (2001)doi:10.1006/jipa.2001.5047, available online at http://www.idealibrary.com on

Chemical Defense in the Egg Masses of Benthic Invertebrates: AnAssessment of Antibacterial Activity in 39 Mollusks and 4 Polychaetes

Kirsten Benkendorff,*,†,1 Andrew R. Davis,* and John B. Bremner†*Department of Biological Sciences and †Department of Chemistry, University of Wollongong, New South Wales 2522, Australia

Received April 30, 2001; accepted July 30, 2001

trntaW

1oqtsips

Many marine invertebrates deposit benthic eggmasses that are potentially vulnerable to microbial in-fection. To help counter this threat these species mayhave evolved some form of chemical protection for theirencapsulated embryos. In this study the egg masses from7 marine mollusks were tested for antibacterial activityagainst 4 marine pathogens: Enterococcus sericolicida,Vibrio anguillarum, Vibrio alginolyticus, and Vibrioharveyi. Extracts from all of these egg masses werefound to inhibit the growth of at least 1 marine bacte-rium at concentrations that approximate the naturalconcentration of extract in the egg masses. The eggmasses of 39 mollusks and 4 polychaetes were thentested for antibacterial activity against 3 human patho-genic bacteria; Escherichia coli, Staphylococcus aureus,and Pseudomonas aeruginosa. Activity was detected inthe egg masses from 34 species, including 2 polychaetesand mollusks from two classes and 18 families. Antibac-terial activity in molluskan egg masses was found toextend across the marine, estuarine, freshwater, andterrestrial environments. Both gelatinous egg massesand tough egg capsules were found to inhibit microbialgrowth, suggesting that physical protection alone maynot be sufficient to protect the eggs. Antimicrobial activ-ity was observed in the fresh egg masses but not in thewell-developed egg masses of a subset of species. Theresults of this study indicate that a wide range of inver-tebrates use chemical defense to protect their earlystage embryos against bacterial infection. © 2001 Academic

Press

Key Words: antibacterial; egg mass; mollusk;polychaete; Enterococcus sericolicida; Vibrio anguilla-rum; Vibrio alginolyticus; Vibrio harveyi; Escherichiacoli, Staphylococcus aureus; Pseudomonas aeruginosa.

INTRODUCTION

The benthic marine environment can effectively bedescribed as a soup of microorganisms. Typically, bac-

1 To whom correspondence should be addressed. Fax: 61 242

214135. E-mail: kirsten@uow.edu.au.

109

terial populations range from 103 to 106 per milliliter inthe ocean, with as many as 109 per milliliter in marinesediments (Austin, 1988). Furthermore, bacteria havebeen found to rapidly colonize a wide range of bioticand abiotic substrata in the ocean (Zobell and Allan,1935; Davis et al., 1989; Wahl et al., 1994). In responseo this selective pressure, a wide range of sessile ma-ine organisms have evolved chemical defense mecha-isms, which involve the production of secondary me-abolites with antimicrobial activity (e.g., Burkholdernd Burkholder, 1958; McCaffrey and Endean, 1985;alls et al., 1993; Wahl et al., 1994; Koh, 1997).Many marine mollusks and polychaetes have

evolved a reproductive strategy that involves the dep-osition of fertilized embryos in benthic egg masses.These egg masses appear to be subject to selectivepressures similar to those of many of the sessile inver-tebrates that are known to use chemical defense. Eggmasses provide a rich source of nutrition and mostspecies do not provide parental protection for theirdeveloping embryos, although their development cantake many months (e.g., Havenhand, 1993). The eggmasses are composed primarily of proteins and poly-saccharides (Hunt, 1966; Flower et al., 1969; Karuso,987), making them highly suitable for the settlementf microorganisms. These egg masses must be ade-uately protected to ensure the long-term survival ofhe species. However, there have been no comprehen-ive studies on the antimicrobial properties of benthicnvertebrate egg masses. Nevertheless, toxic com-ounds have been reported in the eggs of some free-pawning marine invertebrates (e.g., Gil-Turnes et al.,

1989; McClintock and Vernon, 1990; Coll, 1992; Gil-Turnes and Fenical, 1992). Elevated chemical defenceof reproductive tissue has also been documented for abroad range of taxa, including plants (Thompson,1982), bryozoans (Harvell, 1984), and terrestrial inver-tebrates (Eisner et al., 1996).

Benthic egg masses are generally thought to play aprotective role during the embryonic development ofmarine invertebrates (e.g., Thorson, 1950; Mileik-

ovsky, 1971; Smith et al., 1989). The energy expendi-

0022-2011/01 $35.00Copyright © 2001 by Academic Press

All rights of reproduction in any form reserved.

110 BENKENDORFF, DAVIS, AND BREMNER

ture associated with encapsulation appears to be high,suggesting that encapsulation must have substantialsurvival value (Pechenik, 1979). Furthermore, the in-stantaneous mortality rates of eggs retained in benthicegg capsules appear to be much lower than the esti-mated mortality rates for planktonic development(Strathman, 1985). Thorson (1950) suggested that thebenthic egg masses of marine invertebrates could pro-vide protection against infection, although his asser-tion does not appear to have ever been tested. How-ever, several compounds with antibacterial activityhave been isolated from the egg masses of some marinemollusks (Kamiya et al., 1984, 1988; Matsunaga et al.,1986; Yamazaki, 1993; Benkendorff et al., 2000a).

It seems unlikely that newly fertilized eggs will des-ignate some of their limited nutritive reserves to theproduction of defense compounds. It is more likely thatfreshly deposited egg masses are enclosed with antimi-crobial compounds derived from the adults. Conse-quently, we predict that antimicrobial metabolites aremore likely to be found in the freshly laid egg massesthan in the well-developed or hatching egg masses. Thegelatinous egg masses of the sea hare Aplysia julianahave been shown to lose antimicrobial activity duringembryonic development (Kamiya et al., 1988). A formof chemical ripening was observed in the egg masses ofDicathais orbita, where the toxic compound, tyrindo-leninone, is converted into tyriverdin, which is bacte-riostatic but not cytolytic, and then, at the time ofhatching, tyriverdin is converted into Tyrian purple,which shows no antibiotic properties (Benkendorff etal., 2000a). Furthermore, there is evidence that therequirement for antimicrobial agents in some mollus-kan egg masses diminishes as the embryos develop.The early stage, shell-less embryos of the muricidsNucella emarginata and Nucella lapillus were bothfound to suffer high mortality outside their leatheryegg capsules (Pechenik et al., 1984; Rawlings, 1995),which was apparently due to bacterial infection (Lord,1986). However, N. lapillus embryos can be rearedoutside the capsule in nonsterile seawater after theyhave developed shells (Pechenik et al., 1984), indicat-ing that they have acquired resistance to disease.

In this study, the benthic egg masses of 7 species ofmarine mollusks from a broad range of taxonomicgroups were screened for antimicrobial activity againsta suite of marine pathogenic bacteria. To determinewhether antimicrobial activity is a general phenomenain benthic invertebrate egg masses, a further 32 mol-lusks and 4 polychaetes were screened for activityagainst 3 human pathogenic bacteria (Escherichia coli,Staphylococcus aureus, and Pseudomonas aeruginosa),which are routinely used for the testing of disinfectants(Olsen, 1979). These include the gelatinous egg massesand leathery egg capsules from 23 families of mollusks.The majority of species were marine, although 3 estu-

arine, 2 freshwater, and 1 terrestrial species were also

tested. A subset of marine mollusks was tested forantimicrobial activity at two stages of embryonic de-velopment to test the prediction that freshly laid eggshave a greater requirement for chemical protectionthan well-developed embryos.

METHODS

Collection of Egg Material

The egg material from a total of 39 mollusks (33marine, 3 estuarine, 2 freshwater, and 1 terrestrial)and 4 polychaetes were collected for antimicrobial test-ing (Table 1). Voucher specimens and/or photographsare held in a reference collection at the University ofWollongong. Most of the specimens were collected fromintertidal reefs along the New South Wales coast, nearWollongong (34°309, 151°509) Australia. However, theegg masses from 3 gastropod and 1 polychaete specieswere collected from the Mediterranean coast, Spain(Table 1). These specimens were freeze-dried prior tobeing shipped to Australia. The egg masses were ex-amined and classified either as gelatinous when theeggs were embedded in a jelly matrix or as toughcapsules when the eggs were enclosed in a tough leath-ery outer casing or embedded in a firm matrix. Severalindependent collections of the egg masses of some com-mon species were made to test for any variation inantimicrobial activity. A summary of all the speciestested, the type of egg mass, the number of collections,and the total amount collected is provided in Table 1.

In general, the egg masses were collected shortlyafter they were laid and were used immediately orplaced in the freezer (220°C) until used. Antimicrobialactivity was also assessed at a later stage of develop-ment for the egg masses of seven species. The develop-mental stage was determined by changes in the colora-tion or firmness of the egg masses and confirmed byinspection of the embryos or veligers under a dissectingmicroscope. An open aperture at the top of leathery eggcapsules also indicated the commencement of hatch-ing.

Preparation of Egg Material

The egg masses were prepared differently for twoantimicrobial assays (see below). The assay used de-pended on the availability of egg material, and a sub-sample of 15 mollusks and 2 polychaetes were tested inboth assays (Table 1). When there was sufficient ma-terial, organic solvent extracts were prepared andtested for activity in the fluorescein diacetate (FD)assay. Egg masses that were rare, difficult to find, ordifficult to collect were tested in the zone of inhibition(ZI) assay.

For use in the ZI assay, gelatinous egg masses were

cut into approximately 1-cm pieces and ground with a

111ANTIBACTERIAL ACTIVITY IN BENTHIC INVERTEBRATE EGG MASSES

TABLE 1Species of Invertebrates, Arranged Phylogenetically, That Were Screened for Antibacterial Activity in Their Egg Masses

Family SpeciesEgg

massAmountcollected Assay

Natural Concentration

CHCl3 MeOH

Phylum: Mollusca—Class Gastropoda

Neritopsidae Nerita atramentosa TC 25 capsules ZI (c) NA NALittorinidae Bembicium nanum G 161.5 g ZI (c), FD 1.3–5.8 NCNaticidae Conuber sordidus G 150 g ZI (c), FD 0.7–0.8 4.4Ranellidae Cabestana spengleri TC 15 g FD 20.5 38Muricidae Dicathais orbita TC 292 g ZI, FD 14.2–18.6 24

Agnewia tritoniformis TC 10.5 g ZI, FD 8.6–11.3 NCMorula marginalba TC 0.26 g FD NC NCBedeva hanleyi TC 1.56 g FD NC NCLepsiella reticularis TC 1.2 g FD NC NCTrunculariopsis trunculus* TC 8.2 g FD 11.3 NACeratostoma erinaceum* TC 1.2 g FD 23.9 NA

Mitridae Mitra carbonaria TC 10.4 g ZI, FD NC NCConidae Conus papilliferus TC 4.3 g ZI, FD NC NCAmphibolidae Salinator fragilis G 25.4 g ZI (c), FD 5.6 12.8

Salinator solida G 52.9 g ZI (c), FD 3.2 NCSiphonariidae Siphonaria denticulata G 24.3 g ZI, FD NC NC

Siphonaria zelandica G 5.6 g FD NC NCPlanorbidae Isidorella hainesi** G 28.8 g ZI, FD 6.3 19.9

Glyptophysa gibbosa** G 8 masses ZI NA NACamaenidae Meridolum gulosum** TC 6 capsules ZI NA NAAplysiidae Aplysia juliana G 170 g ZI, FD 5–8.3 10

Aplysia sydneyensis G 11.7 g ZI, FD 6.8 11.3Aplysia parvula G 2 g ZI NA NAAplysia dactylomela G 1 ribbon ZI NA NAStylocheilus longicauda G 30.6 g FD 12.4 14.5Dolabrifera dolabrifera G 1/6 ribbon ZI NA NA

Philinidae Philine angasi G 178.4 g ZI, FD NC NCPleurobranchidae Pleurobranchus peroni** G 1 ribbon ZI NA NA

Pleurobranchea sp.** G 2 ribbons ZI NA NABullinidae Bullina lineata** G 2 ribbons ZI (c) NA NAHydatinidae Hydatina physis** G 1 ribbon ZI (c) NA NAOxynoidae Oxynoe viridis G 1 ribbon ZI NA NADendrodorididae Dendrodoris fumata** G 2 ribbons ZI NA NADorididae Platyodoris galbannus** G 2 ribbons ZI NA NA

Jorunna pantherina** G 1 ribbon ZI NA NAGoniodorididae Goniodoris sp. G 4 ribbons ZI NA NAPolyceridae Plocampherus imperialis** G 1 ribbon ZI NA NAGlaucidae Spurilla neopolitana* G 1.4 g FD 38.9 NA

Phylum: Mollusca—Class Cephalopoda

Loliginidae Sepioteuthis australis TC 67 g FD 8.2–12 NA

Phylum: Annelidia — Class Polychaeta

Eupolymnia sp.* G 1.8 g ZI, FD NC NCUnidentified sp. 1 G 6 masses ZI NA NAUnidentified sp. 2 G 2 masses ZI NA NAUnidentified sp. 3 G 7.2 g ZI (c), FD NC NC

Note. All species are from the intertidal area on the Wollongong Coast, N.S.W, Australia, except for those marked *, which wereobtained from the Mediterranean Sea, Spain. The egg masses were collected from the field except those marked **, which were collectedfrom aquaria. The types of egg mass are TC, tough capsules and G, gelatinous egg mass. The total amount of material collected forantimicrobial testing is recorded per egg mass for species that were tested only in the zone of inhibition assay but is provided in gramsfor material that was solvent-extracted (wet weight except for the Mediterranean specimens *, which were freeze-dried). Theantimicrobial assays used are the zone of inhibition (ZI) and fluorescein diacetate (FD) assays. In the ZI assay some of the egg masseswere contaminated by foreign microbial growth when placed on agar plates, which has been indicated by (c). The natural concentrationrefer to the concentration of cholorform (CHCl3) and methanol/water (MeOH) extracts obtained from the egg masses (in mg extract perml of egg mass). NA refers to not applicable for egg masses that were not extracted and tested in the FD assay. NC indicates that the

natural concentration was not calculated.

ISaCmc(timam(

msBtac

bmppcrbci

fmwwescaotagrarFtr(p

a

112 BENKENDORFF, DAVIS, AND BREMNER

mortar and pestle. Leathery egg capsules were cutopen with a pair of sterile scissors to release the intra-capsular fluid and then ground with a mortar andpestle to homogenize the sample. Prior to use with eachnew sample, the mortar and pestle were washed,rinsed in 100% ethanol, and completely dried under afume hood.

For the FD assay, the egg masses were blended insolvent with a Waring blender or cut open with a pairof scissors. Organic extracts were prepared by thesoaking of macerated egg masses in chloroform/meth-anol (1:1, v:v). The solvent was decanted and replacedafter 3 h and after a further 6 h. The sample was thenleft to soak overnight and the extracts were combinedand filtered. A small volume of water was added to theextract when necessary to induce the formation of twolayers. The chloroform layers were separated from thewater/methanol layers in a separating funnel. The sol-vent was then removed by rotary evaporation undervacuum. The dried extract was transferred to a smallweighed vial by sequential resuspension in small vol-umes (;0.5 ml) of methanol (MeOH) and dichlorometh-ane. The bulk of the remaining solvent was removedunder a stream of nitrogen gas and the last traces wereremoved by evaporation in vacuo (0.1 mm Hg, 30 min,room temperature). The natural concentration of theextracts was recorded as milligrams of extract per mil-liliter of fresh egg mass. The natural concentration ofextract was not calculated for every species due toincomplete extraction and/or the insolubility of thedried extract.

Antimicrobial Assays

Four marine pathogens, Enterococcus sericolicida,Vibrio anguillarum, Vibrio alginolyticus, and Vibrioharveyi were provided by the Department of Primaryndustries and Fisheries, Launceston, Tasmania.tock cultures of E. coli ACM845, S. aureus ACM844,nd P. aeruginosa ACM846 were obtained from theulture Collection at the University of Queensland andaintained at 278°C in 15% glycerol. The microbial

ultures were prepared according to Benkendorff et al.2000b). Activity against the marine pathogens wasested with the FD assay (Chand et al., 1994), accord-ng to Benkendorff et al. (2000b). Activity against hu-

an pathogens was assessed with both the FD assaynd the ZI assay (Benkendorff et al., 2000b), which is aodification of the traditional agar diffusion assay

e.g., Spooner and Skyes, 1972).Molluskan egg masses were tested directly for anti-icrobial activity in the ZI assay by the placing of

amples of the egg material onto a lawn of bacteria (seeenkendorff et al., 2000b). Exponentially growing cul-

ures (1 ml) of the bacteria were first pipetted onto thegar and evenly spread across the surface. The excess

ulture liquid was removed with a pipette and the

acterial lawn was allowed to dry for approximately 5in. Duplicate samples of the egg material were then

laced onto the microbial lawn with sterile forceps. Thelates were incubated overnight at 37°C before beinghecked for apparent zones of inhibition in the bacte-ial lawn surrounding the egg material. Zones of inhi-ition were recorded only if there was a clear andonsistent ring that exceeded 2 mm in width surround-ng the egg material.

Chloroform and polar (methanol/water) extractsrom the molluskan egg masses were tested for anti-

icrobial activity in the FD assay. The crude extractsere tested for activity at 1 and 10 mg/ml in acetone,hich were in the range of natural concentrations ofxtract found in the fresh egg masses (Table 1). Ithould be noted, however, that the extracts did notompletely dissolve at the notional 10 mg/ml, and thectual test concentration was found to be in the rangef 3–7 mg/ml for a subsample of species. To calculatehe actual test concentrations, the extracts remainingfter the FD assay were dried under a stream of nitro-en gas, followed by 1 h in vacuo (see above). Theemaining dried extract was reweighed to calculate themount used, which was then divided by the number ofeplicates for which the sample was used during theD assay. The water/methanol extracts were alsoested at higher concentrations, which were within theange of natural concentrations for these polar extracts20–50 mg/ml), depending on the availability of sam-le.Seawater was used as a control in both antimicrobial

ssays. In the ZI assay, 100 ml of seawater was pipetteddirectly onto the microbial lawn. Organic extracts fromseawater and residues from large volumes (300 ml) ofevaporated solvent were prepared in the same way asthe egg extracts for the FD assay.

Statistical Analysis

The relative frequency of antimicrobial activity intough egg capsules compared to gelatinous egg masseswas analyzed by the Fisher exact test (Zar, 1984) withthe JMP statistical package. Two-tailed tests wereused, with a set at 5%.

RESULTS

Antimicrobial Activity against Marine Pathogens

Organic extracts from the egg masses of seven ma-rine mollusks from five orders were tested for bacteri-ostatic activity against four marine pathogenic bacte-ria. The chloroform extracts from all of these specieswere found to have some inhibitory activity toward allof the marine pathogens (Table 2). In general, activitywas observed at concentrations similar to the natural

concentration of extract in the fresh egg mass (Table 2).

p

113ANTIBACTERIAL ACTIVITY IN BENTHIC INVERTEBRATE EGG MASSES

However, V. alginolyticus appears to be relatively re-sistant to extracts from the gelatinous egg mass ofBembicium nanum and the tough egg capsules of Se-pioteuthis australis. E. sericolicida was also relativelyresistant to extracts from S. australis egg masses. The

olar extracts from the egg masses of D. orbita and A.juliana showed little activity against any of the marinepathogens (Table 2).

Antimicrobial Activity against Terrestrial Pathogens

The freshly laid egg masses from 39 molluscs and 4polychaetes were also tested for antimicrobial activityagainst 3 human pathogenic bacteria, and 34 of thesespecies effectively inhibited the growth of at least 1 testmicroorganism in at least one assay (Table 3). Twopolychaetes and species from 18 of the 23 families ofmollusk tested were found to have antimicrobial prop-erties in their egg masses. These included representa-tives of estuarine (Naticidae and Amphibolidae), fresh-water (Planorbidae), and terrestrial families (Ca-maenidae), in addition to marine species from twoclasses (Gastropoda and Cephalopoda).

The full details of the antimicrobial activity in theegg mass of each species in the two assays are availablein Benkendorff (1999). In the ZI assay, a total of 16 of38 species were found to unambiguously inhibit bacte-rial growth (Table 3). Consistent results were observedwith separate collections of egg masses from the samespecies. No zones of inhibition were produced by sea-water controls. The egg ribbons of the sea hares, Aply-sia spp., consistently produced the largest zones ofinhibition against all the test organisms. Two uniden-tified polychaetes, 1 Dorid nudibranch, and Philine

TABBacteriostatic Activity in the Egg Masses of Seven Marin

anguillarum (V. ang.), Vibrio alginolyticus (V. alg.), Vibrio hathe Fluorescein Diacetate (FD) Hydrolysis Assay

Species ExtractNatural Co

(mg/ml)

Bembicium nanum CHCl3 5.8Cabestana spengleri CHCl3 20.5Dicathais orbita CHCl3 17.6

MeOH/water 24Salinator fragilis CHCl3 7.8Aplysia juliana CHCl3 8.3

MeOH/water 10Stylocheilus longicauda CHCl3 12.4Sepioteuthis australis CHCl3 12

Note. The natural concentration of extract in the egg mass andminimum inhibitory concentration refers to the lowest concentratipresented as falling within a range of test concentrations when painhibition was observed at the higher concentration, which was ininhibited the hydrolysis of FD at the maximum test concentration; —test concentration (10 mg/ml).

angasi produced inconsistent zones of inhibition

around the egg masses. These have all been recordedas not active in Table 3 due to the ambiguity of theresults. Foreign microbial growth was found to sur-round the egg masses of 1 polycheate and 7 molluskswhen these were placed on agar plates on the ZI assay(Table 1).

In the FD assay, antimicrobial activity was detectedin the egg masses of 22 mollusks and 2 polychaetes.The activity was apparent in all replicated collectionsof the egg masses. However, there was some variationin the minimum inhibitory concentration of samplesfrom some species that were collected from the sameheadland on different days and between samples fromdifferent locations. Most extracts were active only attheir maximum solubility, although S. aureus and E.coli were susceptible to some of the extracts at 1 mg/ml.The active concentrations were generally within therange of natural concentrations. The main exceptionwas the egg mass of Conuber sordidus, which is com-posed primarily of water and produced a very smallamount of organic extract. The lipophilic (chloroform)extracts from the molluskan egg masses showed moreantimicrobial activity than the polar extracts, withactivity in 24 of 25 species. Strong antimicrobial activ-ity was detected in the polar extracts from 3 species inthe family Aplysiidae and 1 fresh water pulmonate,Isidorella hainesi. No antimicrobial activity was ob-served with the seawater or solvent controls.

Consistent results were found in the two assays for 8of the 15 species that were tested in both (Table 3).However, antimicrobial activity was detected in theegg masses from 6 species in the FD assay but not withany certainty in the ZI assay. The egg masses from 5 of

2Mollusks against Four Marine Pathogenic Bacteria, Vibrioeyi (V. har.), and Enterococcus sericolicida (E. ser.) Tested in

.Minimum inhibitory concentration

V. ang. V. alg. V. har. E. ser.

1–10 .10 1–10 1–105 5–10 5 51–10 1–10 1–10 1–10— — — —1–10 1–10 1–10 1–101–10 1–10 1–10 10

.10 — — —5–10 5–10 5 5

10 .10 10 .10

e minimum inhibitory concentrations are recorded as mg/ml. Theat which the hydrolysis of FD was completely suppressed and isl inhibition was observed at the lower concentration and completepletely soluble in acetone; . indicates that the extracts partiallyicates that there was no apparent inhibition of FD at the maximum

LEerv

ncn

thonrtiacomind

these species were contaminated by epibiotic microor-

pttabs

114 BENKENDORFF, DAVIS, AND BREMNER

ganisms on the agar plates. The remaining speciesproduced inconsistent areas of inhibition around thetissue. The egg mass from 1 species of Mitridae clearlyinhibited microbial growth in the ZI assay but extractsfrom this egg mass did not inhibit the hydrolysis offluorescein diacetate (Table 3).

Of the human pathogens, the Gram-positive bacte-rium S. aureus was the most susceptible to componentsof the molluskan egg masses. The growth of S. aureuswas inhibited by 79% of the egg masses, whereas E. coliwas inhibited by only 69% and P. aeruginosa was in-hibited by 72% of species (Table 4). Furthermore, S.aureus was inhibited at lower concentrations of extractfor several species in the FD assay and the largestzones of inhibition were produced against this bacte-rium in the ZI assay.

Gelatinous Egg Masses vs Tough Egg Capsules

Antimicrobial activity against at least one ecologi-cally relevant (marine) pathogen was observed in bothgelatinous egg masses and tough capsules of mollusks(Table 4). A Fisher exact test found no significant dif-ference in the overall frequency of antimicrobial activ-ity against the marine pathogens in these two types ofegg mass (P 5 0.5604 ; df 5 27). Antimicrobial activ-ity was also observed in both tough egg capsules and

TABProportion of Invertebrate Egg Masses Found to Have A

(E. coli, S. aureus, P. aeuriginosa) in the Zone of I

Superorder Order/infraorder

Phylum: MolluscaNeritopsina NeritodeaCaenogastropoda Littorinimorpha

Neogastropoda

Heterobranchia Basommatophora

EupulmonataAnaspideaCephalaspidea

NotaspideaSaccoglossaNudibranchia

Debrachia TeuthoideaPhylum: Annelidia—PolychaetaTotal egg masses

gelatinous egg masses against the human pathogens r

(Table 4). A Fisher exact test showed that there was nosignificant difference in the overall frequency of anti-microbial activity for the two types of egg mass (P 5

3imicrobial Activity against at Least One Microorganismbition (ZI) and Fluorescein Diacetate (FD) Assays

Family

Proportion of species active

ZI FD Total

Neritidae 0/1 0/1Littorinidae 0/1 1/1 1/1Naticidae 0/1 1/1 1/1Ranellidae 1/1 1/1Muricidae 3/4 7/7 7/7Mitridae 1/1 0/1 1/1Conidae 1/1 1/1 1/1Amphibolidae 0/2 2/2 2/2Siphonariidae 1/1 2/2 2/2Planorbidae 2/2 1/1 2/2Camaenidae 1/1 1/1Aplysiidae 5/6 3/3 6/6Bullinidae 0/1 0/1Hydatinidae 0/1 0/1Philinidae 0/1 1/1 1/1Pleurobranchidae 1/2 1/2Oxynoidae 1/1 1/1Dendrodorididae 1/1 1/1Dorididae 1/2 1/2Gonidorididae 0/1 0/1Polyceridae 0/1 0/1Glaucidae 0/1 1/1 1/1Loliginidae 0/1 1/1 1/1

1/4 2/2 2/416/38 24/25 34/43

TABLE 4Antibacterial Activity in Tough Egg Capsules Compared to

Gelatinous Egg Masses of Mollusks

Origin ofpathogen Species

Toughcapsules

Gelatinousmasses

All eggmasses

Marine V. anguillarum 3/3 4/4 7/7V. alginolyticus 2/3 3/4 5/7V. harveyi 3/3 4/4 7/7E. sericolicida 2/3 4/4 7/7Combined 10/12 15/16 25/28At least one 3/3 4/4 7/7

Terrestrial S. aureus 12/13 18/25 30/38E. coli 10/13 17/26 27/39P. aeuriginosa 6/9 15/20 21/29Combined 28/35 50/71 75/103At least one 12/13 20/26 32/39

Note. The frequency of antibacterial activity is recorded as theroportion of egg masses that effectively inhibited the marine anderrestrial pathogens in at least one antibacterial assay. The propor-ion of egg masses with activity has been combined for the marinend terrestrial pathogens to give the overall incidence of antimicro-ial activity against these two groups of pathogens. The proportion ofpecies showing activity against at least one pathogen has also been

LEnt

nhi

ecorded for the bacteria of marine and terrestrial origin.

A

tolgoht(

TZ

ADMCSIA

g—e

115ANTIBACTERIAL ACTIVITY IN BENTHIC INVERTEBRATE EGG MASSES

0.3539, df 5 105) or the proportion of species thatwere active against at least one human pathogen (P 50.3884, df 5 38).

ntimicrobial Properties at Different Stages ofDevelopment

The egg masses from seven species of mollusk wereested for antimicrobial activity at two different stagesf development (freshly laid and hatching). The freshlyaid egg masses from all of these species inhibited therowth of at least one test microorganism, in at leastne antimicrobial assay (Table 5). However, none of theatching egg masses were found to completely inhibithe growth of any test microorganisms in either assayTable 5).

DISCUSSION

The benthic egg masses of invertebrates are poten-tially vulnerable to microbial infection, which mayhave led to the evolution of chemical defence mecha-nisms. In support of this hypothesis, antimicrobial ac-tivity was recorded in 79% of the invertebrate eggmasses tested in this study (Table 3). The egg massesof mollusks from 18 families, in addition to 2 species ofpolychaetes, were found to inhibit the growth of humanpathogenic bacteria. The antibacterial activity in mol-luskan egg masses appears to extend across marineand terrestrial environments, with 26 marine, 3 estu-arine, 2 freshwater, and 1 terrestrial species testingpositive. Extracts from the egg masses of 7 marinemollusks were also found to inhibit the growth of 4marine pathogenic bacteria (Table 2), 3 of which areknown to infect mollusks (Austin, 1988; Sutton and

TABLE 5Antibacterial Activity of Seven Molluskan Egg Masses

ested at Different Stages of Embryonic Development in theone of Inhibition (ZI) and Fluorescein Diacetate (FD) Assays

Species

Developmental stage

Freshly laidWell

developed

ZI FD ZI FD

gnewia tritoniformis 1 1 — —icathais orbita 1 1 — —itra carbonaria 1 — — —onus papilliferus 1 1 — —iphonaria denticulata 1 1 — —sidorella hainesi 1 1 — —plysia juliana 1 1 — —

Note. 1 Indicates that the specimen unambiguously inhibited therowth of at least one terrestrial microorganism in the given assay;

indicates that the specimen showed no clear antibacterial prop-rties.

Garrick, 1993; Koh, 1997). The active concentrations

approximate the natural concentrations of extract inthe egg masses, particularly when the incomplete sol-ubility of the extracts at 10 mg/ml is taken into con-sideration. Consequently, the antimicrobial compo-nents in these egg masses could realistically play anecological role in the prevention of microbial infection.The idea that egg masses of marine invertebrates pro-vide protection against microbial infection was firstsuggested by Thorson (1950) but until now has notbeen rigorously tested.

Antimicrobial properties in benthic egg masses ap-pear to have evolved in two different invertebratephyla and at least twice within the Mollusca. The dep-osition of benthic egg masses has evolved in two sepa-rate classes of mollusk, Gastropoda and Cephalopoda(Smith et al., 1989), and species from both of theseclasses were found to have antimicrobial activity intheir egg masses (Table 1). Within the class Gas-tropoda, egg masses have evolved in one subclass, theOrthogastropoda, but could have evolved indepen-dently in two clades, the apogastropods (including cae-nogastropods and heterobranchs) and the neritopsines(including Neritoidea) (Ponder and Linberg, 1997).Most of the species tested in this study are apogastro-pods. The egg capsules from one species of Neritoidea(Nerita atramentosa) were tested for antimicrobial ac-tivity in the ZI assay but did not show clear antimicro-bial activity due to contamination by foreign microor-ganisms surrounding the eggs (Table 2). Future re-search should include the testing of a greater range ofspecies from the gastropod superorder Neritopsina andthe class Cephalopoda against marine pathogens.

It remains possible that antimicrobial activity inbenthic invertebrate egg masses is even more wide-spread than indicated by this study. All of the speciesthat did not exhibit clear activity were tested only inthe zone of inhibition assay, which has a number ofassociated limitations (Chand et al., 1994; Benkendorffet al., 2000b). Several species tested negative in the ZIassay but showed clear activity in the FD assay (Table3). Interestingly, most of these egg masses were con-taminated by the growth of unidentified microorgan-isms when placed on agar in the ZI assay. Furtherexperiments are required to determine whether thesemicroorganisms are epibiotic. However, it is possiblethat there is a commensal relationship between thebenthic egg masses of some invertebrates and theepibiotic microbes. Many organisms are protected frompathogenic microorganisms as long as they retain theirnormal associated microflora (Austin, 1988). Further-more, symbiotic microorganisms on the surface of eggsfrom some crustaceans have been shown to produceantimicrobial compounds (Gil-Turnes and Fenical,1992; Gil-Turnes et al., 1989).

It is also significant that the egg masses from severalspecies that tested negative in this study were collected

from aquaria (Table 1). The deposition of egg masses by

tdamganoie

cipgglattltfcrqwCfiTeepr

pebimidttdkikdP

biimro

maTpatwf4ivs

sotldGlw

ecc

116 BENKENDORFF, DAVIS, AND BREMNER

Opisthobranchs in aquaria is thought to be a stressresponse (W. Rudman, pers. com.) and therefore thesespecies may not have accumulated the compounds thatare normally produced to protect the egg mass. Signif-icantly, soft corals have been shown to accumulate thedefensive compounds found in the eggs only a fewweeks prior to spawning (Coll et al., 1985). It is likelyhat antimicrobial compounds in some species are dieterived, as appears to be the case for the biologicallyctive compounds that have been found in the eggasses of the Opisthobranchs Hexabranchus san-

uineus (Faulkner, 1992) and Elysia halimedae (Paulnd van Alstyne, 1988). Nevertheless, this is clearlyot the case with the antibacterial components from D.rbita egg masses, which are derived from a precursorn the hypobranchial gland of the adults (Benkendorfft al., 2000a).

It appears that a physical barrier may not be suffi-ient to protect reproductive material from microbialnfection in the marine environment. Antimicrobialroperties against both marine and terrestrial patho-ens were found to occur at similar frequencies in bothelatinous egg masses and tough egg capsules of mol-usks (Table 4). All of the egg masses showed somectivity against at least one marine pathogen andhere was no significant difference in the proportion ofhe two types of egg mass that were active against ateast one human pathogen. Lord (1986) demonstratedhat the contents of N. lapillus egg capsules were freerom bacterial contamination and suggested that theapsule walls were impermeable to bacteria. However,eproductive tracts free of bacteria would also be re-uired to produce egg capsules with axenic contents,here physical defense would be more problematic.onsequently, a bacteria-free environment may be dif-cult to achieve by physical defence mechanisms alone.his study indicates that antibacterial compounds arequally common in tough egg capsules and gelatinousgg masses, suggesting that chemical defence is therimary mechanism used to maintain an axenic envi-onment for embryonic development.

The requirement for antimicrobial protection ap-ears to be greatest at the early stages of molluskanmbryonic development. An assessment of antimicro-ial activity in the egg masses of seven mollusks dur-ng two stages of development indicates that egg

asses have significantly greater antimicrobial activ-ty when first laid and then lose activity as the larvaeevelop and hatch (Table 5). These findings are consis-ent with previous studies on molluskan egg masseshat have reported the loss of antimicrobial activityuring larval development (Kamiya et al., 1988; Ben-endorff et al., 2000a). The lack of antimicrobial activ-ty at late stages of development suggests that mollus-an embryos lose their susceptibility to bacteria duringevelopment. This is supported by the observations of

echenik et al. (1984) on the survivorship of embryos

from the dog whelk N. lapillus. The loss of antimicro-ial activity during larval development could also be anmportant component of the hatching process in gelat-nous egg masses. Microbial degradation of the jelly

atrix around the time of hatching may facilitate theelease of the larvae or juveniles into the water columnr onto the substratum (Harris, 1975; Eyster, 1986).Antimicrobial activity in the egg masses of mostarine mollusks tested in this study appears to be

ttributable to components of low to moderate polarity.he activity in the egg masses of the mollusks wasrimarily restricted to the lipophilic extracts in the FDssay. This would effectively reduce the loss of bioac-ive components into the water column. However, theater/methanol-soluble components from extracts of

our species showed clear antimicrobial activity (Table). Three of these species are from the family Aplysi-dae and several cytolytic glycoproteins have been pre-iously isolated from eggs and adult sea hares in salineolutions (Yamazaki et al., 1984, 1985, 1990; Kisugi et

al., 1987; Kamiya et al., 1984, 1988). Nevertheless, thelipophilic extracts of Aplysia eggs were also found tohave antimicrobial properties in this study and, signif-icantly, bioactive compounds have been isolated fromthe chloroform extracts of adult sea hares (e.g., Hollen-beak et al., 1979; Miyamoto et al., 1995).

The egg masses of marine invertebrates appear tohave broad-spectrum antimicrobial activity. The eggmasses of the seven mollusks found to inhibit thegrowth of marine pathogens were also found to havesome activity against human pathogens of consider-able medical importance. Both Gram-negative andGram-positive marine and human pathogens were sus-ceptible to extracts from the egg masses. Overall, how-ever, the Gram-positive bacterium S. aureus was moreusceptible to the toxic compounds in the eggs than thether test organisms (Table 4). S. aureus is an impor-ant human pathogen and has evolved resistance to aarge number of conventional antibiotics in the lastecade (Stinson, 1996). The greater resistance of theram-negative bacteria to the molluskan egg masses is

ikely due to the greater reinforcement in their cellalls (e.g., Martin and Beveridge, 1986).Overall, our data indicate that benthic invertebrate

gg masses provide a novel source of potential antimi-robial agents. We have recently isolated a potent non-ytolytic antibiotic from the egg mass of D. orbita (Ben-

kendorff et al., 2000a), which is bacteriostatic at thetherapeutically significant concentration of 0.5 mg/mland has a chemical structure different from that of allantibiotics in current use. Previous studies on mollus-kan egg masses have reported agglutinating activity(Kamiya and Shimizu, 1981), antineoplastic and cyto-lytic activity (Yamazaki et al., 1984, 1985; Roesner andScheuer, 1986; Kisugi et al., 1987; Kamiya et al., 1988),antifeedant activity (Pennings, 1994), and tranquiliz-

ing effects (Marthy, 1976). These studies all demon-

L

M

M

M

M

M

M

M

O

P

P

P

P

117ANTIBACTERIAL ACTIVITY IN BENTHIC INVERTEBRATE EGG MASSES

strate that invertebrate egg masses are suitable tar-gets for further chemical, ecological, and pharmacolog-ical studies.

ACKNOWLEDGMENTS

The assistance of Bill Rudman and Ian Loch from the AustralianMuseum in the species identification is greatly appreciated. We aregrateful to Dr. Manual Ballesteros (University of Barcelona) forassistance in the collection and identification of the egg masses fromthree Mediterranean mollusks and one polychaete. Dr. Jeremy Car-son from the Fish Health Unit, Department of Primary Industry andFisheries, Tasmania kindly provided the strains of marine bacteria.We also thank the Australian Flora and Fauna Research Centre, theBiomolecular Research Centre, and the Environmental ResearchInstitute, University of Wollongong for a supplementary postgradu-ate scholarship to K.B. This is contribution No. from the Ecology andGenetics Group at the University of Wollongong.

REFERENCES

Austin, B. 1988. “Marine Microbiology.” Cambridge Univ. Press,Melbourne.

Benkendorff, K. 1999. “Bioactive Molluscan Resources and TheirConservation: Chemical and Biological Studies on the Egg Massesof Marine Molluscs.” Ph.D. thesis, University of Wollongong.

Benkendorff, K., Bremner, J. B., and Davis, A. R. 2000a. A putativerole for the precursors of Tyrian Purple in the egg masses of theAustralian Muricid, Dicathais orbita. J. Chem. Ecol. 26, 1037–1050.

Benkendorff, K., Davis, A. R., and Bremner, J. B. 2000b. Rapidscreening for antimicrobial agents in the egg masses of marinemolluscs. J. Med. Appl. Malac. 10, 211–223.

Burkholder, P. R., and Burkholder, L. M. 1958. Antimicrobial activ-ity of horny corals. Science 127, 1174.

Chand, S., Lusunzi, I., Veal, D. A., Williams, L. R., and Karuso, P.1994. Rapid screening of the antimicrobial activity of extracts andnatural products. J. Antibiot. 47, 1295–1304.

Coll, J. C. 1992. The chemistry and chemical ecology of octocorals(Coelenterata, Anthozoa, Octocorallia). Chem. Rev. 92, 613–631.

Coll, J. C., Bowden, B. F., Tapiolas, D. M., Willas, R. H., Djura, P.,Streamer, M., and Trott, L. 1985. Studies of Australian soft cor-als—XXXV. The terpenoid chemistry of soft corals and its impli-cations. Tetrahedron 41, 1085–1092.

Davis, A. R., Targett, N. M., McConnell, O. J., and Young, C. M.1989. Epibiosis of marine algae and benthic invertebrates: Naturalproducts chemistry and other mechanisms inhibiting settlementand overgrowth. In “Bioorganic Marine Chemistry” (P. J. Scheuer,Ed.), Vol. 3, pp. 85–114, Springer-Verlag. Heidelberg.

Eisner, T., Attygalle, A. B., Conner, W. E., Eisner, M., MacLeod, E.,and Meinwald, J. 1996. Chemical egg defense in a green lacewing(Ceraeochrysa smithi). Proc. Natl. Acad. Sci. USA 93, 3280–3283.

Eyster, L. S. 1986. The embryonic capsules of nudibranch molluscs:Literature review and new studies on albumen and capsule wallultrastructure. Am. Malac. Bull. 4, 205–216.

Faulkner, D. J. 1992. Chemical defenses of marine molluscs. In“Ecological Roles of Marine Natural Products” (V. J. Paul, Ed.), pp.119–163. Cornell Univ. Press, Ithaca, NY.

Flower, N. E., Geddes, A. J., and Rudall, K. M. 1969. Ultrastructureof the fibrous protein from the egg capsules of the whelk Buccinumundatum. J. Ultrastructure Research 26, 262–273.

Gil-Turnes, M. S., and Fenical, W. 1992. Embryos of Homarus ameri-canus are protected by epibiotic bacteria. Biol. Bull. 182, 105–108.

Gil-Turnes, M. S., Hay, M. E., and Fenical, W. 1989. Symbioticmarine bacteria chemically defend crustacean embryos from a

pathogenic fungus. Science 246, 116–118.

Harris, L. G. 1975. Studies on the life history of two coral-eatingnudibranchs of the genus Phestilla. Biol. Bull. 149, 539–550.

Harvell, C. D. 1984. Why nudibranchs are partial predators: Intra-colonial variability in bryozoan palatability. Ecology 65, 716–724.

Havenhand, J. N. 1993. Egg to juvenile period, generation time, andthe evolution of larval type in marine invertebrates. Mar. Ecol.Prog. Ser. 97, 247–260.

Hollenbeak, K. H., Schmitz, F. J., Hossain, M. B., and Van Der Helm,D. 1979. Deodactol, antineoplastic sequiterpenoid from the seahare Aplysia dactylomela. Tetrahedron 35, 541–545.

Hunt, S. 1966. Carbohydrate and amino acid composition of the eggcapsule of the whelk, Buccinum undatum. Nature 210, 436–437.

Kamiya, H., Muramoto, K., Goto, R., and Yamazaki, M. 1988. Char-acterisation of the antibacterial and antineoplastic glycoproteinsin a sea hare Aplysia juliana. Nippon Suisan Gakkaishi 54, 773–777.

Kamiya, H., Muramoto, K., and Ogata, K. 1984. Antibacterial activ-ity in the egg mass of a sea hare. Experientia 40, 947.

Kamiya, H., and Shimizu, Y. 1981. A natural agglutinin inhibitableby D-galacturonic acid in the sea hare Aplysia eggs: Characterisa-tion and purification. Bull. Jap. Soc. Sci. Fish. 47, 255–259.

Karuso, P. 1987. Chemical ecology of the nudibranchs. In “BioorganicMarine Chemistry,” Vol. 1, pp. 31–60, Springer-Verlag Berlin.

Kisugi, J., Kamiya, H., and Yamazaki, M. 1987. Purification andcharacterisation of Aplysianin E, an antitumour factor from seahare eggs. Cancer Res. 47, 5649–5653.

Koh, E. G. L. 1997. Do Scleractinian corals engage in chemicalwarfare against microbes? J. Chem. Ecol. 23, 379–398.

ord, A. 1986. Are the contents of egg capsules of the marine gas-tropod Nucella lapillus (L.) axenic? Am. Malac. Bull. 4, 201–203.arthy, H. J. 1976. Natural tranquilizer in cephalopod eggs. Nature261, 496–497.artin, N., and Beveridge, T. 1986. Gentamicin interaction withPseudomonas aeruginosa cell envelope. Antimicrob. Agents Che-mother. 29, 1079–1087.atsunaga, S., Fusetani, N., and Hashimoto, K. 1986. KabiramideC, a novel antifungal macrolide from nudibranch eggmasses.J. Am. Chem. Soc. 108, 847–849.cCaffrey, E. J., and Endean, R. 1985. Antimicrobial activity oftropical and subtropical sponges. Mar. Biol. 89, 1–8.cClintock, J. B., and Vernon, J. D. 1990. Chemical defence in theeggs and embryos of antarctic sea stars (Echinodermata). Mar.Biol. 105, 491–495.ileikovsky, S. A. 1971. Types of larval development in marinebottom invertebrates, their distribution and ecological signifi-cance: A reevaluation. Mar. Biol. 10, 193–213.iyamoto, T., Ebisawa, Y., and Higuchi, R. 1995. Aplyparvunin: Abioactive acetogenin from the sea hare Aplysia parvula. Tetrahe-dron Lett 36, 6073–6074.lsen, W. A. 1979. The improved Kelsey–Skyes test for disinfectants:An outline of its development and proposed modifications. Aust. J.Hosp. Pharm. 9, 127–133.

aul, V. J., and van Alstyne, K. L. 1988. The use of injested algalditerpenoids by Elysia halimeda Macnae (Opisthobranchia: Asco-glossa) as antipredator defences. J. Exp. Mar. Biol. Ecol. 119,15–29.

echenik, J. A. 1979. The role of encapsulation in invertebrate lifehistories. Am. Nat. 114, 859–870.

echenik, J. A., Chang, S. C., and Lord, A. 1984. Encapsulateddevelopment of the marine prosobranch gastropod Nucella lapil-lus. Mar. Biol. 78, 223–229.

ennings, S. C. 1994. Interspecific variation in chemical defenses inthe sea hares (Opisthobranchia: Anaspidea). J. Exp. Mar. Biol.

Ecol. 180, 203–219.

S

S

S

S

S

T

T

W

W

Y

Y

Y

Y

Z

Z

118 BENKENDORFF, DAVIS, AND BREMNER

Ponder, W. F., and Lindberg, D. R. 1997. Towards a phylogeny ofgastropod molluscs: An analysis using morphological characters.Zool. J. Linn. Soc. 119, 83–265.

Rawlings, T. A. 1995. Direct observation of encapsulated develop-ment in Muricid gastropods. Veliger 38, 54–60.

Roesener, J. A., and Scheuer, P. J. 1986. Ulapualide A and B, ex-traordinary antitumor macrolides from nudibranch eggmasses.J. Am. Chem. Soc. 108, 846–847.

mith, B. J., Black, J. H., and Shepherd, S. A. 1989. Molluscan eggmasses and capsules. In “Marine Invertebrates of Southern Aus-tralia Part 2” (S. Shepherd and I. Thomas, Eds.), pp. 841–891,Southern Australian Government Printing Division, Adelaide.

pooner, D. F., and Skyes, G. 1972. Laboratory assessment of anti-microbial activity. In “Methods in Microbiology 7B” (J. R. Norrisand D. W. Ribbons, Eds.), pp. 211–227, Academic Press, London.

tinson, S. C. 1996. Drug firms restock antibacterial arsenal. Chem.Engineer. News Sept. 23, 1996, pp. 75–100.

trathmann, R. R. 1985. Feeding and nonfeeding larval developmentand life-history evolution in marine invertebrates. Annu. Rev.Ecol. Syst. 16, 339–361.

utton, D. C., and Garrick, R. 1993. Bacterial disease of culturedgiant clam Tridacna gigas larvae. Dis. Aquat. Org. 16, 47–53.

hompson, J. 1982. “Interaction and Coevolution.” Wiley–Inter-

science, New York.

horson, G. 1950. Reproductive and larval ecology of marine bottominvertebrates. Biol. Rev. 25, 1–45.ahl, M., Jensen, P. R., and Fenical, W. 1994. Chemical control ofbacterial epibiosis on ascidians. Mar. Ecol. Prog. Ser. 110, 45–57.alls, J. T., Ritz, D. A., and Blackman, A. J. 1993. Fouling surfacebacteria and antibacterial agents of four bryozoan species found inTasmania, Australia. J. Exp. Mar. Biol. Ecol. 169, 1–13.

amazaki, M. 1993. Antitumor and antimicrobial glycoproteins fromsea hares. Comp. Biochem. Physiol. 105C, 141–146.

amazaki, M., Kisugi, J., Ikenami, M., Kamiya, H., and Mizuno, D.1984. Cytolytic factor in eggs of the sea hare Aplysia kurodai.Gann 75, 269–274.

amazaki, M., Kisugi, J., Kimura, H., and Mizuno, D. 1985. Purifi-cation of antineoplastic factor from eggs of a sea hare. FEBS Lett.185, 295–298.

amazaki, M., Ohye, H., Kisugi, J., and Kamiya, H. 1990. Bacterio-static and cytolytic activity of the purple fluid from the sea hare.Dev. Comp. Immun. 14, 379–383.

ar, J. H. 1984. “Biostatistical Analysis,” 2nd ed. Prentice-Hall ofAustralia, Sydney.

obell, C. E., and Allen, E. C. 1935. The significance of marinesurfaces in the fouling of submerged surfaces. J. Bacteriol. 29,

239–251.