evidence for the microbial basis of a chemoautotrophic invertebrate community at a whale fall on the...

Evidence for the Microbial Basis of a ChemoautotrophicInvertebrate Community at a Whale Fall on the DeepSeafloor: Bone-Colonizing Bacteriaand Invertebrate EndosymbiontsJODYW. DEMING,1* ANNA-LOUISE REYSENBACH,2 STEPHENA. MACKO,3 AND CRAIG R. SMITH4

1School of Oceanography, University of Washington, Seattle, Washington 981952Institute of Marine and Coastal Sciences, Rutgers University, New Brunswick, New Jersey 089033Department of Environmental Sciences, University of Virginia, Charlottesville, Virginia 229034Department of Oceanography, University of Hawaii, Honolulu, Hawaii 96822

KEY WORDS: sulfide oxidation; mixotrophy; stable isotopes; Beggiatoa; nitrifiers; metha-notrophs; hydrothermal vents

ABSTRACT To explore the microbial basis for a remarkable macrofaunal community at the siteof a whale skeleton on the seafloor of the Santa Catalina Basin, we obtained samples of whale bone,bone-colonizing invertebrates, microbial mats, and the dominant fauna in the adjacent sulfide-richsediments during Alvin expeditions in 1988 and 1991. Invertebrate tissues were examined bytransmission electronmicroscopy (TEM) andmats and bone-penetrating bacteria by epifluorescencemicroscopy (EM). Tissues from the dominant bivalveVesicomya c.f. gigas, themytilid mussel Idasolawashingtonia, and selected gastropods and limpets were also assayed chemically for enzymesdiagnostic of sulfur- and methane-based chemoautotrophy and for stable carbon isotopic composi-tion. Results of all analyses were consistent with dominant sulfur-based endosymbioses in the clamand mussel (the first record of endosymbiosis in the genus Idasola) and the general absence ofmethane symbioses at the site, strengthening the analogy of the whale-skeleton faunal communityto those known from distant Pacific hydrothermal vent sites. Examples of minor endosymbionts,either nitrifying or methanotrophic cells according to internal membrane structures by TEM, raisedthe possibility of a supplemental mode of nutrition to the clam, or means to remove ammonia in thegill tissue, in the event of significant changes in the chemical environment. Microsc. Res. Tech.37:162–170, 1997. r 1997 Wiley-Liss, Inc.

INTRODUCTIONIn late 1987 during a dive of the deep-sea submers-

ible Alvin at a depth of 1,240 m in the Santa CatalinaBasin (SCB) 33° 148 N, 118° 308 W), the intact skeletonof a 21 m long blue or fin whale was discovered on theseafloor (Smith et al., 1989). Although partial skeletalremains of other whales have since been trawled frombathyal depths off central California (Bennett et al.,1994), this was the first such sighting in the mul-tidecadal history of manned submersible research. Pur-suant to the discovery, the SCBwhale site was revisitedby Alvin on two occasions, in 1988 and in 1991. Nowhale soft tissue was ever detected at the site, butdetailed photographic mapping of the whale bonesduring these expeditions (Bennett et al., 1994) and thedegree of skeletal articulation indicated that the car-cass had arrived at the seafloor intact (Allison et al.,1991). Other information, including analysis of bomb14C in collagen from bone and the inferred ages ofbone-colonizing fauna, constrain the age of the skeletonsite to between 10 and 40 years (Bennett et al., 1994; C.Smith and M. Stuiver, unpublished).The whale skeleton on the SCB seafloor has produced

a remarkable ‘‘habitat island’’ for faunal communities,as documented in detail by Bennett et al. (1994). The

partially buried bones, up to 60% lipid by weight (S.A.Macko, unpublished data from a vertebra collected in1988), have produced an unexpectedly strong reducingenvironment—a sulfide-rich habitat for fauna and mi-croflora quite distinct from those inhabiting the sur-rounding basin floor. In the spectrum of known reduc-ing environments in the deep sea, from hydrothermalvents to wood accumulations, the whale-island commu-nity composition most closely resembles those knownfrom geographically distant (1,800 km) hydrothermalvent communities in the Pacific Ocean (Bennett et al.,1994). Approximately 50% of the bone surfaces visiblein flyover photographs in 1988 were covered by yellowand white mats of filamentous microorganisms (Ben-nett et al., 1994), much like those seen at GuaymasBasin vents (Nelson et al., 1989). Other portions of thebones were colonized by a variety of macrofauna,including mytilid mussels (Idasola washingtonia), ar-cheogastropod limpets (Cocculina craigsmithi), snails(Mitrella permadesta), lucinid clams, serpulids, poly-

Received 3 February 1995; Accepted in revised form 14 June 1995*Correspondence to: Jody W. Deming, School of Oceanography, Box 357940,

University of Washington, Seattle, WA98195. Email: [email protected]

MICROSCOPY RESEARCH AND TECHNIQUE 37:162–170 (1997)

r 1997 WILEY-LISS, INC.

noids, and amphipods (Bennett et al., 1994). However,the dominant megafaunal species, the vesicomyid clamVesicomya c.f. gigas, occurred in great abundance in thesediment, always within 20 cm of the skeleton, andsometimes on the bones (Bennett et al., 1994). Vesico-myid and lucinid clams similar to those of the whale-skeleton habitat are known from hydrothermal andother reducing environments to supplement their nutri-tion or derive all of it from either sulfur- or methane-oxiding endosymbionts (Cavanaugh, 1985; Childressand Fisher, 1992; Fisher, 1990).Here we elaborate on our original findings, summa-

rized only briefly earlier (Deming et al., 1990; Smith etal., 1989), by documenting morphological and nutri-tional features of the invertebrate endosymbionts andbone-colonizing microbial mats. Morphological aspectsare revealed by transmission electronmicroscopy (TEM)and epifluorescence microscopy (EM). Physiology isinferred from enzyme assays, stable isotope measure-ments, and morphological features detected by TEM.We also provide the results of EM observations of thebone-penetrating microbes assumed to provide the pri-mary reductant (sulfide) for the endosymbiont-carryinginvertebrates that colonize this remarkable skeletonisland on the deep seafloor.

MATERIALS AND METHODSSample Collection

All samples were collected using the submersibleAlvin, primarily during a series of dives in November1988 but also during follow-up dives in February 1991(Bennett et al., 1994). Whale vertebrae were placed inseparate compartments of an insulated cooler mountedon the basket of the submersible to protect againstwashing and temperature changes during Alvin ascentand recovery in warm surface waters. From the upper-wardly oriented surfaces of the bones that had beenexposed to flowing seawater in situ (the remainder ofthe bone surfaces had been buried in sediment), samplesof microbial mat and specimens of the snail Mitrellapermodesta, the mussel Idasola washingtonia, and thelimpet Cocculina craigsmithi (shown in Fig. 3 of Ben-nett et al., 1994) were removed for subsequent analysis.A rib bone too large for the cooler was also pulled fromthe sediments and recovered in the Alvin basket. Thetip of the rib, the only portion that had been exposed toflowing seawater in situ, yielded additional specimensof the mussel I. washingtonia and the limpet C. craig-smithi (as described in Bennett et al., 1994). Specimensof the vesicomyid clam Vesicomya c.f. gigas were recov-ered by the Alvin manipulator from sulfide-rich sedi-ments (Bennett et al., 1994) adjacent to the whalebones using a scoop net to a depth of 10–20 cm. Within30 min of shipboard recovery of the submersible, allspecimens used in our analyses were removed from thebones or sediments and kept on ice until furtherprocessing or storage, at either 4°C or 220°C or inliquid nitrogen, as indicated below.One 16 cmwide vertebra was cut in half on shipboard

to reveal the bone interior in cross-section. Bone mate-rial was subsampled at an angle perpendicular to thefreshly cut interior surface using a sterile (for bacterialanalysis) or solvent-cleaned (for lipid analysis) 3 mmcork-borer. Eight samples were taken along a line at 2cm intervals, starting 1 cm from the bone surface that

had been oriented upwards in situ (as determined byBennett et al., 1994) and ending 1 cm from the oppositesurface that had been buried in sediment (see inset,Fig. 1). A ninth sample was taken from the bottom ofthe bone to include the sediment-exposed exterior. Allsamples were fixed immediately, as described below,and stored at 4°C in the dark until further processing.

Fixation and Processing Protocolsfor EM and TEM

Portions of microbial mat were fixed for epifluores-cence microscopy (EM) in a sterile prefiltered artificialseawater solution of 2% formaldehyde and stored at4°C until further processing. The samples were viewedusing a Zeiss epifluorescence microscope after stainingwith a 0.001% solution of the DNA-specific stain 486-diamidino-2-phenylindole (DAPI) and gently filteringfilaments onto a 0.2 µm black Nuclepore filter (Porterand Feig, 1980). Samples of the microbial mat andvarious macrofaunal tissues (as described below forenzyme assay procedures) were also fixed for transmis-sion electron microscopy (TEM), using standard proce-dures as described by Fisher et al. (1987) for similarsample types. Tissues were fixed in phosphate-buffered3% glutaraldehyde, stored at 4°C for 2 weeks, postfixedin 1% osmium for 1 h, dehydrated through a gradedethanol series, and embedded in Spurr’s embeddingmedium. Thin sections were stained with uranyl ac-etate and lead citrate and examined with a JEOL 100Btransmission electron microscope.

Analyses of Bone MaterialSamples of bone material from the interior of the

vertebra examined in cross-section were fixed in astandard 0.1 M phosphate-buffered solution of 2%formaldehyde and 2% glutaraldehyde in anticipation ofqualitative analysis by scanning electron microscopy.However, we later opted for more quantitative esti-mates of the bacterial content of the samples using EM.In preliminary work, we determined that the bonematerial readily disaggregated by vortex into a rela-tively homogeneous mixture of microscopic particles(mean diameter of 22 µm) and that greater than 99% ofthe bacteria remained attached to the particles aftervortex treatment. Removing bacteria for counting pur-poses could not be achieved efficiently or reproduciblyby common methods (sonication and/or treatment withthe detergent Triton-X). We settled on a routine count-ing procedure whereby all bacteria attached to twentyrandomly observed particles (after vortex treatmentonly) were counted. All samples were counted within 2months of sample fixation. No correction was made forbacteria obscured on the underside of the particles, sofinal counts represent underestimates of the total num-ber. Bacterial counts were normalized to volume of bonematerial, calculated from estimates of particle volumebased on dimensional measurements using a microme-ter and fine-scaled focus. The best staining method forvisualizing bacteria attached to the bone particlesproved to be a dual staining approach, using acridineorange stain (AO) according to Hobbie et al. (1977)followed by DAPI stain (0.001%). Individual bacteriawere counted rountinely using standard optical filtersfor AO; switching to optical filters for DAPI allowed

163MICROBIAL SUPPORT OF WHALE FALL COMMUNITY

confirmation that AO-fluorescing particles were micro-organisms.Lipids were extracted from replicate, initially frozen

(220°C) samples of the interior bone material using amodified Bligh-Dyer technique (Dobbs and Findlay,1993). Mass was determined gravimetrically. Lipiddata were expressed as a percentage of total mass of thebone sample.

EnzymeAssay ProceduresSamples of microbial mat removed from a whale

vertebra were quick-frozen shipboard in liquid nitrogenuntil further analysis. Whole tissue from ten specimenseach ofMitrella permodesta and Cocculina craigsmithiwere separated from the shell, using aseptic technique,and also quick-frozen in liquid nitrogen. Two specimensof the dominant large bivalve at the site, Vesicomya c.f.gigas, and ten specimens of the smaller mussel, Idasolawashingtonia, were first dissected aseptically to re-move the gill (andmantle, in the case of the clam) tissuefor separate quick-freezing and later analysis.

Frozen tissues were analyzed 3–6 months later at22°C for the activity of several enzymes that arediagnostic for chemoautotrophic potential. Thawedsamples were gently homogenized in a chilled, sterileglass homogenizer. Control tissues (e.g., mantle tissuenot expected to carry endosymbionts) were also pro-cessed. For sulfur-based chemoautotrophy, the lysateswere assayed for two enzymes: adenosine 58-phosphosul-fate reductase (APS), important in the oxidation ofsulfite, and ribulose-1,5-biphosphate carboxylase/oxy-genase (Rubisco), the enzyme diagnostic for net fixationof carbon dioxide via the Calvin-Benson cycle. Theformer was assayed according to techniques describedby Peck et al. (1965), the latter using 14C-bicarbonateaccording to themethods of Glover andMorris (1979) asmodified by Tuttle (1985). For methane-based nutri-tion, the activity of methanol dehydrogenase, a keyenzyme used in the oxidation of methanol by methylo-trophic bacteria, was assayed according to the methodof Weaver and Lidstrom (1985) and expressed as nano-

Fig. 1. Bacterial content (number cm23, determined by EM) andlipid content (percentage by weight) of interior bone material as afunction of distance from the center of the whale vertebra (0 cm) to theuppermost surface of the bone exposed to seawater (18 cm) and thelowermost surface buried in sediment (28 cm). Samples of the bone

material were taken with a sterile (for bacteria) or solvent-cleaned (forlipids) 3 mm cork-borer at an angle perpendicular to a freshly cutcross-section of the 16 cmwide vertebra, originally buried to a depth of7 cm in SCB sediments. Inset: Bone orientation in situ and subsam-pling points; shading indicates portion originally buried in sediment.

164 J.W. DEMING ET AL.

moles substrate hydrolyzed minute21 milligram21 pro-tein.

Stable Isotope MeasurementsPortions of the same sample types collected for

enzyme assays (above) were also frozen for subsequentdetermination of stable carbon isotope composition, asdescribed by Fisher et al. (1994). Carbon isotopic analy-ses of freeze-dried and acidified samples were con-ducted by standardmethods using closed-vessel combus-tion techniques (Macko et al., 1987). Carbon dioxidefrom the combusted tissue samples was analyzed on aV.G. Micromass Prism mass spectrometer. All carbonisotopic results were expressed relative to the ChicagoPDB (PeeDee Belemnite) carbonate (13C) with a prede-termined precision of 60.2% (Macko et al., 1987).

RESULTSBone-Penetrating Bacteria

The vertebra we examined in cross-section was foundto contain highest concentrations of bacteria in theperipheral region of the bone (Fig. 1). The differencesbetween densities in the seawater-exposed portion ofthe bone and the sediment-buried portion (Fig. 1) werenot statistically significant (P . .01, using T8 un-plannedmultiple comparison test among pairs of meanswith unequal sample sizes [Sokal and Rohlf, 1995]).Morphologically, the bacteria were typical of sedimen-tary communities, with a predominance of submicron-sized cocci and less frequent rod-shaped forms. Thecentral portion of the bone was relatively devoid ofbacteria by comparison (a reduction of four orders ofmagnitude and near the sensitivity limit of the count-ing method employed). An inverse pattern was ob-served for lipid content of the bone material. Theinterior of the bone was rich in lipids, up to 70% byweight, whereas the heavily colonized peripheral zoneswere virtually depleted in lipids (Fig. 1). The stablecarbon isotope composition of the interior bonematerial(Table 1) was 223.8%; for lipid extracted from the boneand for the postextraction residue, it was 224.8 and220.3%, respectively. Preliminary efforts to assay lipiddegradative activity within the bone, using 14C-butyr-ate and -oleate, were not successful (W. Ritzrau and J.Deming, unpublished); however, we did not havemeansto protect the bone from oxygenation during subsam-pling shipboard or to prepare and incubate spikedsubsamples anaerobically and thus to mimic the reduc-ing environment known by pyrite deposition to exist inthe peripheral zones of the whale vertebrae (Allison etal., 1991; Smith, 1992).

Bone-Colonizing Microbial MatsMicroscopic observations of microbial mats from the

whale bones revealed a variety of morphological forms,including the major component of filamentous bacteriaof mixed sizes, varying in width from 3–100 µm (Fig. 2),and a minor component of rods and cocci. The dominantfilament in the samples examined by epifluorescencemicroscopy was approximately 100 µm wide (Fig. 2)and 2 cm long, although greater lengths were observedin situ and on bones in the collection cooler (Bennett etal., 1994). The results of enzyme assays (Table 1)indicated that enzymes diagnostic for sulfur-based but

not methane-based chemoautotrophy were present inthe microbial mats. Stable carbon isotope analysis ofmat material indicated a d13C composition of 221.0%(Table 1).

Invertebrate EndosymbiontsTEM examination of gill tissue from the abundant

bone-colonizing mussel, Idasola washingtonia (Fig. 3),and from the dominant invertebrate at the whaleskeleton site, the clam Vesicomya c.f. gigas (Fig. 4),revealed significant densities of prokaryotic endosymbi-onts with double-unit membranes indicative of Gram-negative bacteria. In surveying thin sections of theclam gill tissue, we observed several distinctive prokary-otic cells, single and in a state of cell division, withintracytoplasmic membranes stacked concentrically atthe periphery of the cell (Fig. 5). The results of enzymeassays indicated the presence of gill endosymbionts inboth the clams and mussels that were capable ofchemoautotrophy based on sulfide oxidization (Table 1).No Rubisco or APS-reductase activity was detected inthe clam mantle tissue or in any tissues from the otherinvertebrates (the snail Mitrella permodesta and thelimpet Cocculina craigsmithi) that were examined(Table 1). No methanol dehydrogenase activity wasdetected in any of the invertebrate tissues examined(Table 1). Stable isotope analyses (Table 1) revealedisotopically light d13C compositions for the clams (237.6to 235.2%; n 5 6) and mussels (233.7 to 226.8%;n 5 4) and heavier compositions for the snails (228.7 to219.6%; n 5 6) and limpets (215.3 and 214.4%).

DISCUSSIONChemoautotrophic invertebrate communities thrive

at hydrothermal vent and cold seep environments as aresult of geothermal or subsurface supplies of thereductants sulfide and methane (Cavanaugh, 1985;Childress and Fisher, 1992; Fisher, 1990). Although thecomposition of the invertebrate community at the is-land habitat of the whale skeleton in the Santa Catalina

TABLE 1. Results of enzyme assays and stable carbon isotopecompositions of samples collected at the SCB whale site

Sample

Enzyme activities1(nmol min21 mg21 protein)

d13C2Rubisco APS Meth

Whale vertebra n.d.3 n.d. n.d. 224.8 to 220.3 (n 5 3)Microbial mat 3.96 0.06 0 221.0Cocculinacraigsmithi4 0 0 0 215.3 and 214.4

Mitrellapermodestra4 0 0 0 228.7 to 219.6 (n 5 6)

Idasolawashingtonia4 0.34 0.04 0 233.7 to 226.8 (n 5 4)

Vesicomya c.f.gigasGill tissue5 4.04 0.04 0 236.5 to 235.4 (n 5 3)Gill tissue5 3.72 0.03 0 236.0 and 235.2Mantle tissue 0 0 0 237.6

1Rubisco and APS-reductase are diagnostic for sulfur-based chemoautotrophy;methanol dehydrogenase (Meth) has been used to infer methane-based chemoau-totrophy (see text).2Determined on individual specimens (number provided parenthetically, ifn . 2); see text for further distinctions between samples.3n.d., not determined.4Enzyme assays made on pooled whole tissue from ten specimens.5From separate specimens.


Basin most closely resembles those found at Pacifichydrothermal vent sites (Bennett et al., 1994), nogeothermal source of reductants is available on site.Rather, we (Bennett et al., 1994; Smith et al., 1989)have suggested that anaerobic heterotrophic bacterialproduction of sulfide at the expense of a persistentsource of lipids in the whale bones drives the chemoau-totrophic symbioses prevalent at the site. The results ofour microscopic and chemical analyses of the interior ofa whale vertebra, showing tight spatial correspondencebetween bacterial penetration and lipid depletion (aswell as earlier documentation of pyrite deposition in theinterior of another vertebra [Allison et al., 1991]),support this idea. Additional research is required toprovide quantitative measures of sulfide flux from thebones and to assess variability in the process amongbones and over time.Also in support of bacterially produced sulfide as the

dominant energy source are our findings regarding themicrobial mats that colonize the surfaces of the bones.Both microscopic and enzymatic analyses of these mats

suggest that the dominant filamentous component is amember of the genus Beggiatoa, most likely Beggiatoagigantea, for only this genus includes filaments ap-proaching the extraordinary sizes observed (see Nelsonet al., 1989, and citations therein). Beggiatoa spp.typically dwell at an anaerobic-aerobic interface, sincethey require sulfide as an energy source for CO2 fixationand oxygen for the oxidation of the sulfide. The surfacesof SCB whale bones, bathed by oxygenated seawaterwith sulfide presumably diffusing from the bone inte-rior, provide a favorable interface for this chemoautotro-phic mode of nutrition. The similarity between d13Ccomposition of the microbial mat (221.0%) and that ofthe interior of the vertebra (220.3 to 224.8%) isreasonably consistent with bone carbon as the source ofmat carbon (i.e., carbon being remineralized by hetero-trophic bone-penetrating bacteria to CO2 and then fixedby the mat filaments). However, the bone-colonizingmats (and bone-penetrating bacteria) were comprisedof multiple microbial morphologies and possibly physi-ologies, and we have no direct measures of the isotopic

Fig. 2. A: EM (DAPI-stained) of putative Beggiatoa gigantia filaments, approximately 100 µm inwidth, from a sample of the microbial mat removed from a whale vertebra. B: TEM of a smaller, partiallycollapsed filament from the same sample. Bar 5 5 µm.


composition of the CO2; thus, comparative stable iso-tope values provide only inferential information aboutthe potential source of carbon to the mats (Van Doverand Fry, 1994). The slightly heavier isotopic composi-tion obtained for the mat sample may reflect a trophicshift from heterotrophic activity within the mat (De-Niro and Epstein, 1978), a mixotrophic mode of nutri-tion for some of the mat components, or consumption ofa heavier source of carbon in the surrounding seawater.Sulfide is clearly the major energy source for the

invertebrate endosymbioses prevalent at the whaleskeleton, as indicated by the diagnostic enzyme assayson clam and mussel tissue. These two sulfur-basedendosymbioses and their light carbon isotope signa-tures, especially those for the SCB clam endosymbiosis(236.5 to 235.2%) which are reminiscent of vent clams(Fisher et al., 1994), strengthen the whale-vent analogydrawn by Bennett et al. (1994) based solely on inverte-brate community structure. The heavier isotope signa-tures obtained for the mussels inhabiting the whaleskeleton (233.7 to 226.8%) may reflect a mixotrophiclife style (supplementation of chemoautotrophy withconsumption of photosynthetically fixed particulateorganic matter), a different source of carbon than thatused by the clams, or endosymbionts that fractionatecarbon differently than the clam symbionts. SinceIdasola washingtonia has only a moderately enlargedgill compared to the grossly modified gill morphology ofvesicomyids (R. Turner, personal communication), itappears to have retained the capacity for heterotrophicfeeding on suspended particles. In fact, this is the firststudy to document chemoautotrophic endosymbiosis inthe genus Idasola. Also, since the clams at the whale

skeleton were found largely at depth in the underlyingsulfide-rich sediments and the mussels largely in bonecrevices (Bennett et al., 1994), isotopically distinctcarbon sources may pertain. The relatively heavy iso-tope signatures for the snails (228.7 to 219.6%) andlimpets (215.3 and 214.4%) are consistent with aheterotrophic grazing lifestyle. A more detailed consid-eration of both carbon and nitrogen stable isotopeinformation at the whale site, coupled with observa-tions of macrofaunal grazing patterns, will furtherclarify trophic transfers within and beyond the uniquewhale-island community (Bennett et al., 1994; Fisher etal., 1994; C. Smith,A. Baco, S. Macko and J. Deming, inpreparation).No evidence for methane-driven endosymbiosis at

the SCB whale site was obtained in this study. Itsabsence is perhaps surprising, since anaerobic marinesediments rich in organic material (and, by analogy,whale bones) typically support not only sulfate-reduc-ing bacteria but also a community of methanogens(Deming and Baross, 1993). We cannot rule out thepossibility that the methods used were not sensitiveenough to detect the relevant enzyme activity (reportedsensitivity is on the order of nanomoles milligram21

protein [Weaver and Lidstrom, 1985]) or that samplesfrozen for severalmonthsmay have lost activity (Cavan-augh et al., 1992). The absence of positive controls formethanol dehydrogenase activity also leaves open thequestion of methanotrophy as assessed by this method(in contrast to the cases for Rubisco and APS activitymeasurements). Alternatively (assuming no method-ological problems), if a microbial source of methaneexists in the most anaerobic regions of the bones and

Fig. 3. A: TEM of sulfide-oxidizing endosymbionts (B 5 bacteria) in gill tissue (N 5 nucleus;V 5 vacuole) of the mussel Idasola washingtonia. Bar 5 3 µm. B: TEM of selected endosymbionts(denoted by arrow in panelA) at higher magnification showing double-unit membranes (arrow) indicativeof Gram-negative bacteria. Bar 5 0.2 µm.


sediments, it may be insufficient to support an inverte-brate endosymbiosis or else be consumed competitivelyby free-living methane oxidizers. Shipboard efforts todetect free-living methane-oxidizing bacteria on vari-ous surfaces at the whale site using 14C-methane, inanalogy to the findings of de Angelis et al. (1991) athydrothermal vents (and using identical protocols),yielded positive activity in association with limpetshells but not with tissues from Idasola washingtonia

and Vesicomya c.f. gigas (A.-L. Reysenbach, unpub-lished). The latter findings support the validity of thenegative results obtained by the enzyme assay formethanotrophy.Anaerobic heterotrophic bacterial activity in organic-

rich marine sediments (and possibly whale bones) canalso be expected to produce ammonia from nitratereduction and the putrefaction of the organic material.In fact, these anaerobic processes yield the smaller

Fig. 4. A,B:TEMof sulfide-oxidizing endosymbionts (B 5 bacteria)in gill tissue (N 5 nucleus; V 5 vacuole) of the clam Vesicomya c.f.gigas. Bars 5 5 µm. C: TEM of selected endosymbionts at highermagnification adjacent to microvilli of the gill tissue. B, bacteria.

Bar 5 1 µm. D: TEM of selected endosymbionts at still highermagnification showing double-unit membranes (arrow) indicative ofGram-negative bacteria. B, bacteria. Bar 5 0.4 µm.


molecular-weight organic molecules like acetate pre-ferred by sulfate-reducing (sulfide-producing) bacteria(Nedwell, 1984). Although we did not attempt to mea-sure ammonia flux from whale bones or adjacent sedi-ments, the possibility of ammonia consumption by aminor endosymbiont arises from TEM observations ofclam gill tissue (Fig. 5). The concentric stacking ofintracellular membranes at the periphery of these cellsresembles the distinctive intracellular morphology ofnitrifying bacteria in the genera Nitrosomonas andNitrosococcus (Watson et al., 1989), bacteria that con-vert ammonia to nitrite and fix carbon chemoautotrophi-cally. Such TEM observations alone are insufficient todiagnose nitrifiers, since intracellularmembranes (vari-ously organized in vivo) also characterize other meta-bolic types of bacteria, particularly the methanotrophs.We were unable to detect endosymbiotic activity ofmethanotrophs in these gill tissues, although variouscaveats pertain to these negative results, as discussedearlier. Definitive identification of this minor endosym-biont will require either host specimens with a largerpopulation of the cells or immunogold detection of keyenzymes.If the minor endosymbionts we observed are nitrifi-

ers, they would represent a novel type of endosymbiosiswith the potential to detoxify ammonia from gill tissueas well as supplement host nutrition. To our knowledge,there are no prior reports or speculation on a chemoau-totrophic endosymbiosis based on ammonia oxidation;rather, attention has been focused on determiningnitrogen sources to the invertebrate host, includingnitrogen gas (nitrogen fixation has not been detected)

and ammonia or nitrate assimilated heterotrophicallyinto organic matter (Hentschel and Felbeck, 1993; Leeand Childress, 1994; and citations therein). Regardlessof whether the minor endosymbionts are ammonia- ormethane-oxidizing bacteria (or some other metabolicform), their overall activity in Vesicomya c.f. gigastissues must not have been very significant at the timeof our study, given their low frequency of occurrence.The possibility that changes in the chemical environ-ment over time might induce changes in the propor-tional significance of such endosymbionts warrantsfurther research.In summary, the microbial studies reported here,

relying onmicroscopic techniques and supportive chemi-cal measurements, verify the close resemblance of theSCB whale-skeleton community and modes of nutritionto those of hydrothermal vents and other reducinghabitats in the deep Pacific Ocean, as predicted bySmith et al. (1989) and documented at the faunalcommunity level by Bennett et al. (1994). They alsoprovide the first record of endosymbiosis in the genusIdasola. From TEM analyses of gill tissue removedfrom the dominant clam Vesicomya c.f. gigas, we ob-servedminor endosymbionts with possible ammonia- ormethane-oxidizing potential. A test of the significanceof their contributions to host well-being may best bepursued among invertebrates that thrive at hydrother-mal vents characterized by unusually high concentra-tions of ammonia (Lilley et al., 1993) or at otherreducing sites with variable supplies of ammonia,methane, and sulfide.

Fig. 5. A: TEM of minor endosymbiont in gill tissue of the clam Vesicomya c.f. gigas showing internalmembranes, concentrically stacked at the periphery of the cell, an arrangement that resembles that inchemoautotrophic nitrifying bacteria or possibly methanotrophs (see text). Bar 5 0.2 µm. B: TEM ofsimilar cells fixed in the process of dividing. Bar 5 0.2 µm.


NOTE ADDED IN PROOFSince this manuscript was accepted (spring 1995),

the following significant papers have been published:

Lee, R.W., and Childress, J.J. (1996) Inorganic N assimilation andammonium pools in a deep-sea mussel containing methanotrophicendosymbionts. Biol. Bull., 190:373–384.

Naganuma, T., Wada, H., and Fujioka, K. (1996) Biological communityand sediment fatty acids associated with the deep-sea whale skel-eton at the Torishima Seamount. J. Oceanogr., 52:1–15.

Smith, C.R., Maybaum, L.L., Baco, A.R., Carpenter, S.D., Yager, P.L.,Macko, S.A., and Deming, J.W. (1997) Sediment community struc-ture around whale bones in the deep northeast Pacific Ocean:Macrofaunal, microbial and bioturbation effects. Deep-Sea Res.(submitted).

ACKNOWLEDGMENTSThis work was supported by an ONR grant and NSF

PYI award to J.W.D. and NSF grants to C.R.S. Wethank F. Dobbs for measurements of lipid content of thewhale vertebra collected in 1991, S. Carpenter for EManalyses and generation of figures, D. Niemer forpreparation of TEM samples, D. Penry for assistancewith invertebrate dissections at sea, M. de Angelis forhelp with methane oxidation studies, and M. Orellanafor useful discussions regarding enzyme assays. We arealso grateful to anonymous reviewers for constructivecomments on an earlier version of the manuscript andto the many people who assisted shipboard in samplecollection and processing, including B. Bennett, H.Kukert, W. Ritzrau, Y.-A. Vetter, and P. Yager. Thisworkwould not have been possiblewithout the outstand-ing support of the Alvin-Atlantis II crews.

REFERENCESAllison, P.A., Smith, C.R., Kukert, H., Deming, J.W., and Bennett, B.A.(1991) Deep-water taphonomy of vertebrate carcasses: A whaleskeleton in the bathyal Santa Catalina Basin. Paleobiology, 17:78–89.

Bennett, B.A., Smith, C.R., Glaser, B., and Maybaum, H.L. (1994)Faunal community structure of a chemoautotrophic assemblage onwhale bones in the deep northeast Pacific Ocean. Mar. Ecol. Prog.Ser., 108:205–223.

Cavanaugh, C.M. (1985) Symbiosis of chemoautotrophic bacteria andmarine invertebrates from hydrothermal vents and reducing sedi-ments. Bull. Biol. Soc. Wash., 6:373–388.

Cavanaugh, C.M., Wirsen, C.O., and Jannasch, H.W. (1992) Evidencefor methylotrophic symbionts in a hydrothermal vent mussel (Bival-via: Mytilidae) from the Mid-Atlantic Ridge. Appl. Environ. Micro-biol., 58:3799–3803.

Childress, J.J., and Fisher, C.R. (1992) The biology of hydrothermalvent animals: Physiology, biochemistry and autotrophic symbiosis.Oceanogr. Mar. Biol. A. Rev., 30:337–441.

de Angelis, M.A., Reysenbach, A.-L., and Baross, J.A. (1991) Surfacesof hydrothermal vent invertebrates: Sites of elevated microbial CH4oxidation activity. Limnol. Oceanogr., 36:570–577.

Deming, J.W., and Baross, J.A. (1993) The early diagenesis of organicmatter: Bacterial activity. In: Organic Geochemistry. M.H. Engeland S.A. Macko, eds. Plenum Press, NewYork, pp. 119–144.

Deming, J.W., Reysenbach, A.-L., Neimer, D.J., Macko, S.A., and

Smith, C.R. (1990) Microbial associations with megafauna coloniz-ing a whale carcass on the sea floor: Specific analogs to hydrother-mal vent environments. EOS Transactions, 71:92.

DeNiro, M.J., and Epstein, S. (1978) Influence of diet on the distribu-tion of carbon isotopes in animals. Geochim. Cosmochim. Acta,42:495–506.

Dobbs, F.C., and Findlay, R.H. (1993) Analysis of microbial lipids todetermine biomass and detect the response of sedimentary microor-ganisms to disturbance. In: Handbook of Methods in AquaticMicrobial Ecology. P.F. Kemp, B.F. Sherr, E.B. Sherr, and J.J. Cole,eds. Lewis Publishers, Boca Raton, FL, pp. 347–358.

Fisher, C.R. (1990) Chemoautotrophic and methanotrophic symbiosesin marine invertebrates. Rev. Aquat. Sci., 2:399–436.

Fisher, C.R., Childress, J.J., Oremland, R.S., and Bidigare, R.R. (1987)The importance of methane and thiosulfate in the metabolism of thebacterial symbionts of two deep-sea mussels. Mar. Biol., 96:59–71.

Fisher, C.R., Childress, J.J., Macko, S.A., and Brooks, J.M. (1994)Nutritional interactions in Galapagos Rift hydrothermal vent com-munities: Inferences from stable carbon and nitrogen isotope analy-ses. Mar. Ecol. Prog. Ser., 103:45–55.

Glover, H.E., and Morris, I. (1979) Photosynthetic carboxylatingenzymes in marine phytoplankton. Limnol. Oceanogr., 24:510–519.

Hentschel, U., and Felbeck, H. (1993) Nitrate respiration in thehydrothermal vent tubeworm Riftia pachyptila. Nature, 366:338–340.

Hobbie, J.E., Daley, R.J., and Jasper, S. (1977) Use of Nuclepore filtersfor counting bacteria by fluorescence microscopy. Appl. Environ.Microbiol., 33:1225–1228.

Lee, R.W., and Childress, J.J. (1994) Assimilation of inorganic nitro-gen bymarine invertebrates and their chemoautotrophic andmetha-notrophic symbionts. Appl. Environ. Microbiol., 60:1852–1858.

Lilley, M.D., Butterfield, D.A., Lupton, J.E., Macko, S.A., and McDuff,R.E. (1993) Anomalous CH4 and NH41 concentrations at a mid-ocean ridge hydrothermal system: Combined effects of sedimentsand phase separation. Nature, 364:45–47.

Macko, S.A., Fogel, M.L., Hare, P.E., and Hoering, T.C. (1987) Isotopicfractionation of nitrogen and carbon in the synthesis of amino acidsby microorganisms. Chem. Geol. (Isot. Geosci. Sect.), 65:79–82.

Nedwell, D.B. (1984) The input and mineralization of organic carbonin aerobic aquatic sediments. Adv. Microb. Ecol., 7:93–131.

Nelson, D.C., Wirsen, C.O., and Jannasch, H.W. (1989) Characteriza-tion of large, autotrophicBeggiatoa spp. abundance at hydrothermalvents of the Guaymas Basin. Appl. Environ. Microbiol., 55:2909–2917.

Peck, H.D., Deacon, T.E., and Davidson, J.T. (1965) Studies on APSreductase fromDesulfovibrio desulfuricans andThiobacillus thiopa-rus. I. Assay and purification. Biochim. Biophys. Acta, 96:429–446.

Porter, K.G., and Feig, Y.S. (1980) The use of DAPI for identifying andcounting aquatic microflora. Limnol. Oceanogr., 25:943–948.

Smith, C.R. (1992) Whale falls: Chemosynthesis on the deep seafloor.Oceanus, 35:74–78.

Smith, C.R., Kukert, H., Wheatcroft, R.A., Jumars, P.A., and Deming,J.W. (1989) Vent fauna on whale remains. Nature, 341:27–28.

Sokal, R.R., and Rohlf, F.J. (1995) Biometry: The Principles andPractice of Statistics in Biological Research, 3rd ed. W.H. Freemanand Co., NewYork, 887 pp.

Tuttle, J.H. (1985) The role of sulfur-oxidizing bacteria at deep-seahydrothermal vents. Biol. Soc. Wash. Bull., 6:335–345.

Van Dover, C.L., and Fry, B. (1994) Microorganisms as food resourcesat deep-sea hydrothermal vents. Limnol. Oceanogr., 39:51–57.

Watson, S.W., Bock, E., Harms, H., Koops, H.-P., and Hooper, A.B.(1989) Nitrifying bacteria. In: Bergey’s Manual of Systematic Bacte-riology, Vol. 3. J.T. Staley, M.P. Bryant, N. Pfenning, and J.G. Holt,eds. Williams andWilkins, Baltimore, pp. 1808–1834.

Weaver, C.A., and Lidstrom, M.E. (1985) Methanol dissimilation inXanthobacter denitrificans. J. Gen. Microbiol., 131:2183–2197.


evidence for the microbial basis of a chemoautotrophic invertebrate community at a whale fall on the...

Documents