Sulfide-Driven Autotrophic Balance in the Bacterial Symbiont-Containing Hydrothermal Vent Tubeworm, Riftia pachyptila Jones

19 Pages·1991·8 MB·English

by  J J Childress

Reference: Biol Bull 180: L35-153.(February,
Sulfide-Driven Autotrophic Balance in the Bacterial
Symbiont-Containing Hydrothermal Vent Tubeworm,
Jones
Riftia pachyptila
J. J. CHILDRESSN.1. CK.. RS.ANFIDSEHRERS-3. AJ.NAD. AF.AVMU.ZAZLIA1,YSR.E4E. KOCHEVAR1.
,
*Department ofBiologicalSciences andMarine Science Institute. University ofCalifornia, Santa
Barbara, California 93106, ^Department ofBiological Sciences. Pennsylvania State University.
University Park. Pennsylvania 16802. 'Bam/ielilMarine Station, Bamfield. British Coiimhia.
Canada I OR IBO. and4Department Environment Profond. 1FREMER.
Centre de Brest, B. P. 70-29263 Plouiane. France
Abstract. Hydrothermal vent tubeworms, Riftia pa- all vestimentiferan tubeworms, adults ofthis species lack
chyptila Jones, were maintained alive and studied on amouthandagut(Jones, 1981, 1988;JonesandGardiner,
board ship using flow-through pressure aquaria. Simul- 1988; Southward, 1988). The adult worms appear to de-
taneous measurements ofO2, SCO:, -H2S fluxes showed rive their nutritional needs from the large population of
that theintact symbiosesreach maximum ratesofuptake sulfur-oxidizing chemolithoautotrophic bacterial sym-
of2CO2 (>2 Mmoleg~' h~') at about 90 \iM2H2S. Mea- bionts that live in cells within a specialized organ the
surements were made ofhemolymph and coelomic fluid trophosome in theirtrunk(Cavanaugh elal.. 1981; Fel-
2CO2. 2H2S, thiosulfate, pH, and hemoglobin concen- beck, 1981). The trophosome is a highly vascularized or-
trations in worms kept under various conditions of O2 gan lying between two coelomic cavities that contain a
and 2H2S. Normal hemolymph pH appears to be about hemoglobin-rich fluid (Jones, 1988). This anatomy re-
7.5and isnot affectedby 2H2Sand 2CO2concentrations quires that the animal supply the needs ofthe symbionts
within the ranges observed. We conclude that Riftia is through its circulatory system (Arp et al.. 1985; Felbeck
specialized to provide sulfide to its symbionts with min- and Childress. 1988). Because these symbiontsare sulfide-
imal interaction ofsulfide with the animal metabolism. oxidizing autotrophs (Felbeck, 1981; Belkin et al.. 1986;
The uptake ofsulfide is apparently by diffusion into the Fisheretal.. 1989), theworm musttakeupsulfide,oxygen,
hemolymph, facilitated by the sulfide-binding properties andcarbondioxide from themedium andtransportthem
ofthe hemoglobins. Both 2CO2 and PCo, are elevated in tothe symbionts. These substancescan be taken up from
the hemolymph above their levels in the medium, al- the water by the large obturacular plume, a highly vas-
though they are reduced under autotrophic conditions. cularized organ that has a large surface area and brings
Thus inorganic carbon is apparently concentrated from the hemolymph very close tothesurroundingwater(Arp
the medium into the hemolymph by an unknown mech- et al.. 1985; Jones, 1988). The hemolymph and the coe-
anism. lomic fluid both haveabundant extracellularhemoglobins
which are believed to play a key role in the transport of
Introduction all three ofthese metabolites (Childress et al.. 1984; Arp
et al.. 1985).
The giant hydrothermal vent tubeworm, Riftia pa- Two hemoglobins are found in the extracellular fluids
chyptila Jones, is perhaps the most distinctive ofthe an- ofthese worms. One has a molecular weight ofabout 1.7
imalslivingaround thedeep-sea hydrothermal vents. Like X 106Mrandisfoundprimarily iMnthehemolymph, while
the second is smaller (0.4 X 106 r) and is found in both
Received 10September 1990;accepted 24 November 1990. the coelomic and vascular compartments (Terwilliger et
135
136 CHILDRESS ET AL
J J
a/.. 1980; Arp and Childress, 1981; Terwilliger and Ter- site). Both submersibles pulled the worms offthe rocks
williger, 1985; Arp el ai. 1987). Both hemoglobins bind using their manipulators, placed them in thermally in-
oxygenandsulfidereversiblywith ahigh affinity(Arpand sulatedcontainers, andbroughtthem tothesurfaceabout
Childress, 1981, 1983; Childress et ai, 1984; Arp el ai. 2 to8 h aftercapture. Once atthesurfacethewormswere
1987; Fisher et ai, 1988a). The sulfide binding does not quickly transferred tocold seawater(7C) where undam-
affect the simultaneous binding of oxygen, and appears aged worms were set apart for the whole animal experi-
to occur at a site removed from the heme (Childress et mentsdescribed here. The wormschosen were then care-
ai. 1984; Arp et ai. 1987). When sulfide and oxygen are fully removed from their natural tubes and placed in
below saturation in the hemolymph, their normally rapid, straight plastic tubes ofappropriate size so they could be
spontaneous reaction issuppressed (FisherandChildress, fitted into the pressure vessels necessary fortheir mainte-
1984). Further, the hemoglobin can protect the animal nance. Theywerethen quickly placed in pressure aquaria.
tissues from sulfide toxicity by binding the sulfide with a The worms were routinely maintained in flowing-water
higher affinity than does the site of toxic effects, cyto- pressureaquaria (Quetin and Childress, 1980)at 200atm
chrome-c-oxidase (Powell and Somero, 1983, 1986). The pressure, 8C, and more than 100 nAI O2. Water was
hemoglobin does, however, release sulfide to the sym- pumped through these stainless steel pressure vessels at
biontswhilesimultaneously protectingthem from sulfide about 12 1/h. Previousstudies (Childress et ai. 1984) and
toxicity by holding free sulfide concentrations down preliminary observations during this study indicate that
(Fisher and Childress, 1984; Fisher el ai. 1988a, 1989). although thewormsliveatahydrostaticpressureofabout
The hemoglobins also buffer the hemolymph for carbon 260atm at thesesites, theyare abletosurviveanddisplay
dioxide transport (Childress et ai, 1984). Thus, the he- apparently "normal" behavioratpressuresaslowasabout
molymph apparently has the properties required to take 100 atm. The symbionts themselves do not show signif-
oxygen, carbon dioxide, and sulfide from the medium icant effects of pressure on carbon fixation rates within
and to transport them to the endosymbionts. the pressure range used here (Fisher et ai, 1989). In the
Most ofthe experiments described in this paper were present study pressures as low as 120 atm were used in
performed totest the roleofthe hemolymph in gas uptake someexperiments, but theexperience cited above suggests
andtransportin intact,livingRiftiapachyptilaindividuals. that these lower pressures should have little effect on the
In particular, we were concerned with demonstrating the results.
continuous uptake and oxidation ofsulfide by the intact Studies involving the maintenance of the worms at
organisms, evaluatingthe roleofthe hemoglobinsincon- known sulfideconcentrationswerecarriedoutin flowing-
centrating sulfide from the medium, looking for the pos- water aquaria (120 atm, about 4 1/h) using transparent
sible roles ofother forms ofsulfur in the symbiosis, ex- acrylic pressure vessels (Quetin and Childress, 1980), al-
amining the impact ofsulfide and symbiont autotrophy lowing the activity ofthe worms to be observed during
on internal CO2 poolsand pH, and observingthe pattern themexMperiments. Anaerobic sulfide stock solution (5 or
ofexchange ofgases between the coelomic fluid and the 10 sodium sulfide in seawater at ph 7.0 or 7.5) was
hemolymph. added continuously at the intakes ofthe pressure pumps
A consistent chemical terminology will be used withlowpressure meteringpumpstoachievestablesulfide
throughout this paper. Sulfide and inorganic carbon refer concentrations in the pressure vessels. The effluent water
tothesesubstanceswithoutspecifyingthechemical species from the vessels was periodically sampled with a 0.5 ml
involved. 2H:S and 2CO2 refer to the amounts ofthese glass syringe, and the gases were analyzed by gas chro-
gasesanalyzed from acidifiedsamples usingtheanalytical matography(Childressetai. 1984). ThepH oftheeffluent
methods described below. They are measures ofthe sum waterwas measured withadoublejunction electrodeand
ofthe various chemical forms in which these substances was between 7. and 8.1, depending on the experiment.
are found. H:S, HS~, S2~, S, CO2, HCO, and any other
chemical formulae refer only to the chemical species Metabolism measurements
symbolized. "Free" refers to that fraction ofa substance
in the body fluids that is not bound to the hemoglobins. Measurements ofwhole animal metabolism were made
in a flowingwatersystem similartothat used by Anderson
Materials and Methods et ai. (1987), but adapted for use at the high pressures
required for the survival of the worms. The system
The tubeworms used in these studies were collected pumped seawaterthrough the respirometerchambers us-
from depths ofabout 2600 m at sites on the Galapagos ing HPLC pressure pumps with small acrylic pressure
Rift(0048.247'N, 86 13.478^)andtheEast Pacific Rise vesselsasrespirometerchambers. Thewaterinthissystem
(1248'N, 10857'W) by deep submersibles (Alvin at the wasfirst passed through aseriesoffilters(5.0and 0.2 yum)
Galapagos Rift site and Nautile at the East Pacific Rise and a UV sterilizer. It was then continuously mixed by
AUTOTROPHIC FUNCTION IN RlH'l.l 137
means ofmetering pumps with an antibiotic solution to Analytical methods
achieve a final concentration of 150 mg penicillin-G per
liter and with a sulfide solution (pH 7.5 in seawater) to Gas chromatographic methods similar to those de-
achieve the desired sulfide concentration. It then went to scribed by Childress el al. (1984) were used to analyze
a vertically oriented column measuring 1 X 0.1 m, with gases in body fluids and seawater. Briefly, water samples
the seawaterentering at the top and exiting nearthe bot- were acidified with phosphoric acid, and gases were
tom. The pH ofthe water in the column was maintained stripped from them using a glassand teflon extractor, in-
vaotirab7.at5sheeb(ybNoaatOtpHoHm)cioonnfttortohtlehleeccrooltlhuuammtnnp.umOmixxpyeegddent1haeAnfdwaaNtcie2drbu(ibHnbCltLeh)de CsliyHnset4we,imatnhadlalCtohwOeerdcmoatnlhceceonantndraualctytisiiovsnistoyfingtahfsleuciOhdr:so,ammCaptOloeg:sr, aHop^fhS.0,.T2Nh-it,so,
column while maintaining the desired O concentration. 1.0 ml. The limit ofsensitivity forthesegaseswasbetween
2
The water was then pumped through the respirometer 1 and 20 nM. depending on the gas and the sample size.
chambers to a gas chromatograph for analysis. Two res- Throughout this paper, the terms 2H;S and SCO, refer
pirometer streams were continuously used in these mea- to the amounts measured using this analytical method
surements, onewithanimalsintherespirometerchamber without regard forthechemicalspeciespresentatthevery
and the otheran identical system withoutanimals, which different pH values and conditions in the worms.
served as a control for spontaneous oxidation ofsulfide. To measure pH, a sample ofhemolymph or coelomic
Fluxes of the measured gases due to the animals were fluid wasdrawn from an animal with a syringe. Thedead
calculated from the differences in gas concentrations in space ofthe syringe was filled with blood by drawing a
thewaterexitingtheexperimental andcontrol chambers. small amount ofsample into the syringe and then expel-
Theseexperimentswerecarried outat 130atm hydrostatic ling the air and excess blood before drawing the sample
pressure. for analysis. Without air exposure, the sample was im-
Ammonium fluxwasmeasured forseveral wormswhile mediately injected into a Radiometerglasscapillary elec-
they were in the respirometer system described above. trode (Radiometer America G298A) used in conjunction
The ammonium concentrations in the effluents from the with a reference electrode (Radiometer K171) in a water
two chambers were measured by flow injection analysis jacketed chamber. Precision buffers (Radiometer SI500
&
(Willason and Johnson, 1986). S1510) were used to calibrate the electrode.
The abundances of the two hemoglobins in the he-
Dissectionprocedure molymph were quantified by separating them by HPLC
Worms were dissected so that samples ofhemolymph, gel filtration and measuringtheabsorbanceastheyeluted
coelomic fluid, and trophosome could be obtained for from the column (Arp et al., 1987). A TSK-50 column,
mm mm
further analysis. Worms to be sacrificed were quickly re- 7.5 in diameter and 300 long, was used with a
moved from the pressure aquaria and the plastic tubes TSKguard column (7.5 mm by 75 mm). The eluent was
abenldotwhetnhestvreesttchiemdenouttuminwaasdiscsaercetfiunlglytrsalyit. Tfohreabfoedwycweanll- gacKitHri2cPaOci4d/1/)pahtosppHhat7e.5b,ufpfuemrp(e1.d63atgc0i.t3ricmlac/imdianndat265.C1.7
timeters parallel to the main axis of the worm on the The run time was about 40 min, and an undiluted 1-yul
ventral side. A sample of coelomic fluid (1-5 ml) was sample was used. The absorbance was measured at 415
nm
quickly drawn, with a blunt needle, from the pool ofthis as the eluent left the column.
fluidin thecoelomicspaceand placedon ice. Subsamples Determinations ofthiosulfateand otherunbound thiols
forthevariousanalyseswerequicklytaken. Theremaining in thebody fluidsweremadeby HPLCanalysisofsamples
coelomic fluid was then drained from the worm, and a derivatized by monobromobimane using the methods of
1-ml syringe with a 30-ga needle was used to remove he- Newton et al. (1981) and Fahey et al. (1983) as modified
molymph from the major dorsal vessel leading from the by Vetter et al. (1989). Derivatives were separated on a
trophosome to the plume ofthe worm. Aliquots ofthis 15 cm C-18 reversed phase column and detected using a
post-trophosome(pre-branchial) hemolymph samplewere 235 nm filter forexcitation and a 442 nm filter fordetec-
quickly taken for the various analyses. Samples of tro- tion offluorescence. The eluent flow rate was 1.5 ml per
phosome tissue were also frozen for later analysis ofele- min, using an increasing hydrophobic gradient ofHPLC
mental sulfur. Ifthe trophosome appeared "unhealthy" grade methanoland 2% aceticacid, startingat 10% meth-
[the pinkish appearance correlated with lack ofCO: fix- anol and increasing to 100% during the run.
ation in trophosome preparations (Fisher et a/., 1989), Elemental sulfur in the extracts was quantified by gas
occurred in 7 ofthe 50 animals used] for an individual chromatography according to the method ofRichard et
worm, the data from that worm were excluded from fur- al. (1977) as modified by Fisher et al. (1988b). Pieces of
ther consideration. These unhealthy worms were always tissue(0.5-2.0gwetweight)weredriedfor 18hina 100C
characterized by low (<7.0) hemolymph pH values. dryingoven, andthen extracted for24h with cyclohexane
138 J. J. CHILDRESS ET AL.
in a micro-Soxhlet apparatus. Theextractswere"cleaned 6.784 (Millero. 1986; Millero et ai. 1988) and the pH
up"bypassingthem throughafluorosil column toremove measured in that particular sample.
lipids, and concentrated by evaporation. The injector
temperature was 240C, and the initial column temper- Estimation ofPCo:
ature was 150C, programmed to 220C during the sep- Because 2CO pH, and hemoglobin contents were
mm :.
aration. A six foot (1.8 m) glass column with a 2 measured, itwaspossible, usingpreviously publisheddata,
bore, packed with 5% SP2401 on 100/120 mesh Supel- to estimate the PCo, in the fluids. The data relating pH.
coport,wasusedtoseparatesulfur. Thesulfurwasdetected 2CO:,and PCO;inRifiiacoelomicfluidat 10C(Childress
and quantified usinga thermal conductivity detector. The et ai. 1984) were used as the basis ofa family ofcurves
detection limit forelemental sulfurwasca. 0.001% ofthe that predict Pco, from pH and 2CO2. However, although
dryweight ofthe sample (dependingsomewhat on sample the coelomic fluid and hemolymph are quite similar in
size). The identity ofthe separated sulfur was confirmed ioniccomposition, thehemolymph often hasmuch higher
by the distinctive smell ofsulfur vaporcoming out ofthe hemoglobin content. Because hemoglobin is the only
gas chromatograph detector at the time of the putative protein in any concentration in the hemolymph (Arp et
sulfur peak. ai, 1987), we used the concentration ofheme as an in-
dicatorofproteincontentinthese fluids. Anapproximate
correction factor for hemoglobin concentration was de-
Estimation offree ~ZH:Sand H}S veloped by equilibratingsubsamples, brought todifferent
concentrations in Riftia saline, ofthe same hemolymph
Because 2H2S, pH, and hemoglobin contents were sample with gases ofknown Pco, andthen measuring the
measured, it was possible, using previously published SCO2 inthesesubsamples usingthegaschromamtoMgraphic
data, to estimate the concentration of free (unbound) method. These subsamples (0.909 and 3.554 heme)
2H2S aswell asthe variousspeciesofsulfide. Free 2H2S were equilibrated with 2.09 torr Pco, and a final pH of
wasestimated by using the Hill equation describingthe 7.70 (Arp el ai, 1987). These measurements indicated
relationship between fractional saturation and free sul- that the effect of heme concentration on 2CO2 in this
fide measured at 6C, pH 7.5 in a mixture ofcoelomic range was 0.50 mmole 2CO2/mmole heme. This was
fulruaitdiaonn/d(1h0e0mo-ly%mpsahtu(rFaitsihoenr)]et=ill.0,.713978(8lan):fIrnee[%2Hsa2tS- afadcdteodr tthoatthcehafningaeldetqhueatsiloopneuosfetdhetoreclaaltciuolnastheipPbCeot;waesena
fiM)- 1.778. Pco and SCO-, at different pH values. Theequation was:
To use thisequation, the capacity ofeach fluid sample PCO*, = (4.199 - 0.537 pH) + SCO2[(eA(-1.845 pH
to bind sulfide was estimated by multiplying the small + 13.396)] + 0.0667(heme - 0.79). PCO: in the medium
hemoglobin aggregateconcentration byonesulnde/heme was estimated from the medium pH and 2CO2 with the
ahenmdet.heThlaersgee aegsgtriemgaatteescwoenrceentdreartiivoend bfyrotmhreae msuulltnidpesl/e p(1K98ap4p),easntidmatCeOd2f(r0o.m063t4h5e) eatqu8atCio(nSkgiirvreonw.b1y975H)e.isler
regression ofthe sulfide concentrations in nine coelomic
fluid samples dialyzed at saturating sulfide concentrations Statistical methods
against the concentrations ofthe two aggregates in those Statistical analyseswerecarried out usingStatview SE+
samples (data from Arp, 1987). This regression had an r2 and SuperANOVA (AbacusConcepts) and Fastat(Systat
of0.97 and gave coefficients of0.90 0.27 (95% C. I.) Inc.). The Kendall rank correlation was used to test fora
and 2.97 0.86, respectively, for the two hemoglobins. relationship between two parameters without any as-
From the estimated capacity for binding sulfide and the sumptionsabout the form orlinearity ofthe relationship.
measured ZH2S in the fluid, the percent saturation was Testing for differences in the medians in paired data sets
approximated, andtheaboveequationwassolvedforfree employed the Wilcoxon signed rank test. The Mann-
sulfide. This approximation offree sulfide was then used Whitney U testwasusedtotestfordifferencesin medians
with the estimated sulfide binding capacity [% saturation between unpaired datasets. Simple and multiple linear
= 100 (bound sulfide/bindingcapacity)] in the Hill equa- regressions ofraw and in transformed data were used to
tion toestimate thesulfideboundtothe hemoglobin. This describe the relationships between parameters.
procedurewasthen carried through several iterations until
the estimates converged on a single value for free sulfide. Results
This value was then used to calculate the percentage sat- H'hole animal metabolism
uration ofthe fluid.
The free H2S in each sample was calculated from the Duetoavarietyofequipment problems, onlyonesuch
free 2H2S using a pK, value (8C, 35%o and 120 atm) of experiment was successfully conducted. In this expert-
AUTOTROPHIC FUNCTION IN Rll-TIA 139
ment, twoworms(8.7 and 5.0g)were run in theirnatural
tubes in one chamber for 68 h. This experiment was
started at 13.5C without sulfide. After 6 h, sulfide was
added continuously and 14 h later the animals showed CD
net 2CO2 uptake (autotrophy). For the next 20 h the ef- _O<D
fects ofdifferent sulfide concentrations on the fluxes of E
O2, 2H2S and 2CO2 were measured while maintaining x
O2 between 105 and 209 ^M (Fig. IB). After that time 13
the temperature ofthe system was lowered to 8.4C over
2 h, and a similar set of measurements repeated at O2
concentrations between 72 and 21 1 nM during the next
24 h (Fig. 1A). The set ofobservations at 8.4C started
at 92 nAt 2H2S, decreased in steps to 0.0 nM 2H2S, and CD
then was then raised to 41-49 pM 2H:S for 6 h. As can _OO
be seen in Figure 1A, the lower 2H2S concentrations re- E
sulted in less uptake of2CO2, and without added sulfide
the 2CO2 balance was fully heterotrophic (+3.05 /umoles
2CO2 g~'h~'). For the first three hours after the reintro-
duction of sulfide, this balance remained heterotrophic
(+1.93 Mmoles 2CO2 g~'h~'. high point at 43 nM 2H2S
in Fig. 1A), but autotrophy was reached in the next 3 h
(-0.64Mmoles2CO2g ' h~', at 49 ^/2H2S in Fig. 1A).
The worms were then removed and their tubes replaced
in the vessels. The tubes alone did not show significant
2CO2 flux (<0.1 /umole 2CO2 g ' worm h"') in the pres-
ence of 130 ju/U 2H2S. When the worms were dissected
aftertheexperiment, S wasvisible in theirtrophosomes.
These data demonstrate that these worms were depen-
dent on 2H2S levels greater than about 50 nM to break
even on carbon flux and more than 90 nM was required
for maximum uptake of2CO2. They also show that the
lag-time for changes in 2CO2 flux when 2H2S was re-
moved wasshort, suggesting that useofstored S was not
quantitatively very important. In contrast, when sulfide
wasintroducedafteran absence, the lagtimewasrelatively
long (3-14 h).
Both O2 and 2CO2 flux were significantly dependent
on the 2H2S flux (Fig. 1C). The slope ofthe line relating
2CO2 flux to 2H2S flux was 0.92 0.18 (95% C. I.) in-
dicatingthat 0.92 mole CO2 was fixed foreach mole H2S
consumed. The slope ofthe line relating O2 flux to 2H2S
flux (Fig. 1C)was 1.14 0.17 (95% C. 1.), indicatingthat
1.14 mole O2 was consumed for each mole 2H2S con-
sumed. The lines relating 2CO: and O2 fluxes to 2H2S
flux both intercept the y-axisat virtually the fluxes found
for the worms in the absence of sulfide. This indicates
that sulfide does not interact with the metabolism ofcar-
bon orO2 by theanimal tissues. The R. Q. in the absence
ofsulfide is0.83 suggestinga metabolism based on amix-
ture ofthe major substrates.
The autotrophic Riftia experiment described above
failed to show net uptake ofN2, supportingother negative
data that this species' symbiontsdo not fix N2. A prelim-
inary study has also been carried out on ammonia flux
140 J. J. CH1LDRESS /;/ AL
Fivewormsweresacrificed immediately aftercapture, and always very low (less than 36 pM, average = 24 ^Af) and
their hemolymph and coelomic fluid pH, 2CO2, 2H2S, did not decline duringthe five days in captivity (Table I).
and S2O32 aswell astrophosome S concentrationswere To examine the hypothesis that the pattern of high
measured. Nine other R. pachyptila were placed in high- 2CO: and low pH found in the hemolymph of freshly
pressure, flowing-water aquaria immediately after recovery recovered worms resulted from oxygen deprivation, we
and maintained, under pressure (120 atm), in seawater maintained two individuals for 24 h in the flowing water
without added sulfide for varying periods oftime before aquarium system at 14 juA/ O2 and 15 nM 2H2S. Prior
sacrifice and analysis. The initial values found (Table I) to this experiment these worms had been kept in the
werecomparabletothose found previously forthisspecies aquarium system for 2 days with no sulfide and more
(Childress el at.. 1984) with 2CO2 being quite elevated than 100nAIO2. Thehemolymph pH wasdepressed(6.48
and pH values being quite low. This indicated that the and 6.82), supporting the suggestion that depressed pH
worms were probably withdrawn into theirtubesand an- values after recovery are the result ofanaerobic metabo-
aerobic while theywerebeingbrought to the surface. Data lism (Childress cl a/.. 1984). The 2CO2 values were low
following recovery in the aquarium system supports the (3.357 and 3.280 mAl), but at the low pH values these
sameconclusion. Hemolymph andcoelomicpH roseand represent high Pco, values(13.9 and 8.0torr). The failure
2CO2concentration declined (from very high levelsfound ofthese worms to accumulate the higher 2CO2 concen-
immediately after capture) after the animals were main- trations found in freshly recovered worms(Table I) prob-
tained for one or more days under pressure (Table I). ably resulted from their plumes remaining extended and
Hemolymph 2H2S concentrations in the freshly collected thuscontinuingtoexchange CO2 with themedium during
animals were substantial, ranging up to 1.75 mA/(mean the experiment. In contrast, during recovery from the
of0.71 mA/Table I), decreased rapidly in animals main- bottom, worms were constrained in a box and could not
tained underpressureintheabsenceofadded sulfide,and extend their plumes to exchange gases. This isconsistent
was undetectable afterthree and five days (Table I). Tro- with observationsthat Riftiapachyptilaindividualsrelease
phosome S declined significantly with time as well, ap- substantial amounts of2CO2 to the medium under hyp-
proaching zero after 3 to 5 days (Table I). Thiosulfate oxic conditions (Childress et a/., 1984). The hemolymph
concentrations in the hemolymph ofR. pachyptila were 2H2S contents were substantial (5.497 and 5.013 mAl)
Table I
Riftiapachyptila hemolymph, coelomicfluid, andtrophosomeparameters immediatelyaftercaptureandaltermaintenance
intheah\ence<>/sulfideinflowing water, pressure(120atm) aquaria
Daysafter
AUTOTROPHIC FUNCTION IN Rll-'TlA 141
Table II
Coelomic/hudhemoglobinconcentrationsas/'unctionsofheinolymph hemoglobinconcentrations in Riftiapachyptilaaltermaintenance(24 h) in
high-pressure(120aim),flowing-wateraquariaill variousfixed~2<H:Sconcentrations >0.0and<S(W^M.
DataonfreshlycollectedwormsfromArpetal. (1987)
Wilcoxon signed-rank
[coelomic] - a + b[hemolymph]
142 J J. CHILDRESS ET AL
- y=00133*0176x.R=090
1.2 -
O 8 H
'E
0.4 H
0o1
4 6 8 10 12
Hemoglobin Fraction I (mM) Hemolymph XH S (mM)
2
Figure 2. Frequencydistributionsoftheconcentrationsofthelarge
hemoglobin[FI, 1.7 X 10"Mr,(Terwilligerelal. 1980;Arpelat.. 1987)] Figure3. Coelomic fluid -H,S asa function ofhemolymph iH:S
in coelomic fluid and hemolydmph ofRi/tiapachvptila kept for24 h at inRifiapachypltlakeptfor24hatdifferentexternalsulfideconcentrations
differentexternal sultide and concentrations. and>42
confluent at the size ofthe smaller hemoglobin, because parameters, except thiosulfate, between the compart-
the concentrations in the two compartments were signif- ments, apparently resulting from their interactions with
icantly correlated. However, re-analysis showed that while the hemoglobins. 2H2S was in much higher concentra-
the concentrations were significantly correlated in their tions in the hemolymph (Fig. 3) than in the coelomic
data, the coelomic fluid had higher concentrations pre- fluid because of the much higher concentrations of he-
cluding confluence for this size molecule (Table II). moglobin FI, which binds 3 moles ofsulfide per mole of
Although the two compartments do not exchange he- heme, although this binding is not to the heme group
moglobin molecules, it is apparent that small dissolved itself(Arp el al., 1987), versus 1 mole ofsulfide per mole
molecules are readily exchanged, because the higher O2, of heme for F II (Fig. 2). These correlations and distri-
higher pH group had highly significant correlations be- butions indicate that these worms were circulating their
tween the values of all of the measured parameters be- blood effectively. In contrast, the low O2, low pH group
tween thetwocompartments(Table III). Inaddition, there had no significant correlations between the two com-
were significant differences in the values of all of these partments for these parameters, and the values were sig-
Table III
Riftia pachyptilacoelomicfluidparametersasa function ofthesameparametersin liemolymph aftermaintmancc(24h) ofthewormsinflowing
water,pressure(120atmlaquariaat variousfixed^.H;Sconcentrationsbetween 0.0and600^M
withexternalO2concentrations> 42^M amihemolymphpH> 7.2
AUTOTROPH1C FUNCTION IN R1FTIA 143
nificantlydifferent foronlythreeoftheparameters(Table In contrast, thelowO2. lowpH groupshowsonlythree
IV). Such lack ofequilibration indicates a lack ofoppor- barely significant correlations between external SH2S and
tunity for exchange between the two fluids, suggesting any of the hemolymph sulfide parameters (Table VI).
that circulation was impaired in this group ofworms. Further, the hemolymph sulfide saturation isclosetothat
Because the hemolymph and coelomic fluid parameters expected in vitro, and 50%. saturation is close to the //;
wtheereloawlwOa2y,slopawraplHlelwoarnmdscldoosenloyt,cofrorretlhaetemdo,satnpdarbt,ecsahuoswe Tvihturso,vatlhueea(b3i.l3itaynodf1th1e.s2epwMor2mHs2St,orceosnpcecetnitvrealtye, F2igH.25SA)i.n
sliygmnpshofpaaruatmoettreorpshyfraonmdtehfefehcitgivheerciOrc:u,lhaitgihoen,r pthHewhoermmos- tbhienidrihngemooflsyumlfpihdeabpypetahreshtoemboegleoxbpilnasin.edInenatdidrietliyonb,y tthhee
wiinltlerbnealepmaprhaamseitzeersdtionexctoenrsniadlersiulnfgidteh(eTarbelsepoVn)s.esThoefltohwe ehqeumiolilbyrmipumhdfirseteri2buHti2oSnsanwidthfrteheeHco2Srrweesrpeonedsisnegnteixatlelrynailn
O2, low pH group(Table VI)will beconsidered primarily parameters (Fig. 5B, C) indicating that no gradient for
winerIcenontcthorerarshetilgathtoeertdhOew2iogttrhhoeuerpx,tgerhroenuamplo.l2yHmp2Sh. aTnhdeycoweleormeicat2lHea2sSt ocpaoxsnisddiiavtteiioounnpsthsaaokdetahepexpihasertsemnoitgnllyotbheeisssneesntwciooarulmllysd.cneoSayslemodbnigueonrndtfeursnuctlthfieiosdnee
one order ofmagnitude higher than the external concen-
to depress free sulfide concentrations.
tornasttiroantiinngatlhlecaasbielsit(yTaobfltehiVs,wVoIrI,mFtiog.co4nAc)e,ntclreaatrelysudlefimd-e It isalso apparent from thedata that sulfide isthe only
sulfur compound ofimportance in the hemolymph. Al-
from its environment. However, their hemoglobin was
maintained well below sulfide saturation at all external though thiosulfatewasfound inbothgroups, and increases
suurlaftiidoencaotncaennterxatteironnasl t2esHteSd (oFfig1.225A^)Mshaoswcinogmp5a0r%edsatto- sitiginsitfiycpainctallylyinletshsetphraensehnecmeoolfyemxtpehrn2alHs2uSlfbidyem(oTarbelethVa)n,
2
a(HFn2iSs/'h/;earlvsieotlr<oi//n.a.cfrf1ie9na8is8teayd).ofwTi5hthe0%hexestmaetorunlrayaltmip2ohnHfa2rtSe,e12b1.uH2t2fSrj.eAafmna2diHnfre:edSe dounceTehoderdOahetremohofiglmhyaemgrnpiehxttu2edrCenOa(lF2iga2.nH4d):.SPccoo,nvcaelnutersatairoensboitnhtrhee-
about an orderofmagnitude lowerthan theexternal con- higher 2 group (Table V, VII, Fig. 6A, B), suggestingthe
centrations(Table V,Fig. 5B,C).Thus,although the 2H:S removal of inorganic carbon by the autotrophic sym-
concentration in the hemolymph was much higher than bionts. There were no significant relations between these
outside the worm, there was a significant gradient from parameters for the low O2, low pH group (Table V, Fig.
theoutsidetothe inside forthe freechemical species. The 6A. B). In addition, the internal PCo, was higher than the
latter gradient could only be maintained by the con- external under virtually all conditions (Fig. 6B, Table V,
sumption ofsulfide within the worm, presumably by the VI) precluding uptake by passive diffusion into the he-
symbionts. molymph.
Table IV
CoelomicfluidparametersaxaJunction ojthesameparameters inhemolymph ojRiftia pachyptilaaftermaintenance{24 h) in high-pressure(120
aim),flowing-wateraquariaat variousfixedsulfideconcentralicinx between 13andSOI)^M
with O:concentrations < 42/iM orhemolymphpH < 7.2
Wilcoxon signed-rank
[coelomic] = a + b[hemolymph]
Coel:Vasc
Parameter b 95%CI
pH
144 J. J. CHILDRESS ET Al.
Table V
Riftia pachyptila liemolvmphparametersandS in trophosomeasJunctions ofexternalconditionsaftermaintenance(24h) inhigh-pressure112/1
aim),flowing-wateraquaria in variousfixed2//,5concentrationsgreaterthan 0.0andlessthan600/jM (externalO:concentrations > 42^M and
hcniolvmphpH> 7.2). kcndallcorrelations, but not theregressionsor Wilcoxontests, includeseven individualsat0.0S//i
X Vanable