Metabolite uptake, stoichiometry and chemoautotrophic function of the hydrothermal vent tubeworm Riftia pachyptila: responses to environmental variations in substrate concentrations and temperature

Riftia pachyptila (hereafter referred to solely as Riftia) is a monospecific genus within the family Siboglinidae(Rouse, 2001) and is indigenous to the vent fields of the Eastern and Southeastern Pacific(Shank et al., 1998). Riftia is the dominant megafaunal species at many sites, often growing in enormous aggregations and hosting numerous other species such as mussels, polychaete worms, limpets and crabs(Hessler et al., 1988; Shank et al., 1998; Tunnicliffe, 1991; Govenar et al., 2005). Riftia is devoid of a mouth or digestive tract, and possesses intracellular chemoautotrophic bacteria within a vascularized organ called the trophosome (Cavanaugh et al.,1981; Felbeck,1981). Riftia cannot ingest particulate organic matter,and flourishes where dissolved organic carbon and nitrogen concentrations are too low to support the observed biomass(Johnson et al., 1988; Gaill et al., 1997). As such, Riftia relies entirely on its symbionts for nutrition. Because the symbionts are not in contact with the external milieu, all their substrates and waste products are provided for or eliminated by the host Riftia.

Accordingly, Riftia must acquire both reduced and oxidized substrates for chemoautotrophic metabolism and therefore thrives in diffuse flow regimes, positioning its plume-like gill at the interface of vent flow and bottom-water mixing (Childress et al.,1991). However, this niche is spatially and temporally heterogenous (Johnson et al.,1986). Environmental chemistry in diffuse flows is wildly variable on short-time scales (Johnson et al.,1986; Johnson et al.,1988), with dissolved inorganic carbon concentrations ranging from 2 to >12 mmol l^-1, hydrogen sulfide from undetectable to 725μmol l^-1, and dissolved oxygen and nitrate concentrations ranging from 0 to 110 μmol l^-1 and 0 to 40 μmol l^-1, respectively (Shank et al., 1998; Luther et al.,2001; Mullineaux et al.,2003; Le Bris et al.,2006). Temperatures at diffuse flow sites have been observed to vary between 2 to 25°C, and also to change rapidly over time(Chevaldonne et al., 1991; Johnson et al., 1988).

Prior physiological studies of Riftia have largely focused on characterizing the physiological and biochemical adaptations of host to symbiont, as well as elucidating which metabolites are used by the symbioses(Arp, 1988; Arp and Childress, 1983; Childress et al., 1984; Childress and Fisher, 1992; Childress et al., 1993; Felbeck et al., 1981; Fisher and Childress, 1984). To date, little is known about how environmental conditions such as metabolite concentrations, pH and temperature influence the metabolism (and ultimately growth) of Riftia and its symbionts. The aforementioned spatial and temporal environmental variability makes it impractical to ascertain such relationships in situ. Accordingly, the experiments presented here examined the relation between Riftia metabolite uptake, symbiont chemoautotrophic function, seawater metabolite concentrations, pH and temperature using a shipboard high-pressure respirometry system. We conducted our experiments over a range of environmentally relevant chemical concentrations and temperatures to examine how thermal and chemical fluctuations in situ might influence host metabolite uptake and symbiont autotrophic function. We also examined the stoichiometric relations among the major metabolites, as well as which chemical species are preferentially acquired by Riftia.

All experiments were conducted on board the R/V Wecoma or R/V New Horizon during expeditions in April and May 1996 (HOT 96), November and December 1997 (HOT 97) and November and December 1998 (LARVE 98). Riftia pachyptila Jones tubeworms were collected from two hydrothermal vent fields along the East Pacific Rise(12°48′N,103°56′W, and 9°50′N,104°18′W), at a depth of about 2300 m. Tubeworms were collected daily by the DSV Alvin or DSV Nautile and brought to the surface in a thermally insulated container(Mickel and Childress, 1982). After arrival on board ship, the tubeworms most responsive to touch were immediately placed into flow-through, high-pressure respirometer aquaria,where they were maintained in 0.2-μm filter-sterilized flowing seawater for 1-2 h at 10°C and 27.5 MPa (Girguis et al., 2000).

In all respirometry experiments, two of the pressurized aquaria contained tubeworms while the third served as the control. To simulate the seawater chemistry found in situ, 0.2-μm filter-sterilized seawater was pumped into an acrylic gas equilibration column and bubbled with carbon dioxide, hydrogen sulfide, oxygen and nitrogen or helium to achieve the desired dissolved gas concentrations(Kochevar et al., 1992). Seawater pH was adjusted by using a proportional pH controller and isoosmotic HCl and NaOH solutions (Prominent Inc., Pittsburgh, PA, USA). A sodium nitrate solution (in 0.2-μm filtered-sterilized seawater) was pumped into the equilibration column to produce final seawater nitrate concentrations between 40 and 65 μmol l^-1. Seawater from the equilibration column was delivered to the three aquaria by high-pressure pumps (American Lewa, Inc.,Holliston, MA, USA). High-pressure aquaria temperatures were maintained at 15°C by immersing them in a circulating waterbath, while aquaria pressures were maintained at 27.5 MPa via diaphragm backpressure valves (Circle Seal, Inc., Corona, CA, USA). Vessel effluents were directed through a computer-controlled stream-selection valve that diverted one stream to the analytical instrumentation every 7 min.

During the HOT 96 and HOT 97 expeditions, the analytical system consisted of a membrane-inlet quadrupole mass spectrometer to determine all dissolved gas concentrations, an inline pH electrode and a spectrophotometer for nitrate analyses (Girguis et al.,2002; Girguis et al.,2000). During the LARVE 98 expedition, inorganic carbon concentrations were measured using a carbon dioxide specific electrode(pHoenix, Inc., Houston, TX, USA) mounted in a water-jacketed flow-through cell. Hydrogen sulfide concentrations were determined by a quantitative spectrophotometric assay (Guenther et al.,2001) using a Gilson spectrophotometer with a 250 μl flow-through cell. Oxygen concentrations were determined by a silver/silver chloride electrode (Cameron Instruments Inc., Guelph, ON, Canada) mounted in a 2 ml flow-through cell. All carbon dioxide, hydrogen sulfide and oxygen measurements were confirmed and calibrated using a Hewlett-Packard 5890 Series II gas chromatograph (Childress et al.,1984). During both experiments, pH was measured using a double-junction pH electrode mounted in a water-jacketed flow-through cell,and connected to Orion model 920A or Radiometer PHM 93 pH meter, while nitrate was analyzed from discrete samples collected every 30 min using a quantitative spectrophotometric assay (Girguis et al.,2000; Karlsson et al.,1995). Temperature was measured and recorded by a digital thermometer (Fisher, Inc., Hampton, NH, USA).

Prior to all experiments, Riftia were placed in the respirometer aquaria, and were maintained in conditions typical of those in situ.These `typical' conditions are: total dissolved inorganic carbon (i.e.ΣCO<sub>2</sub>)=5.5-6 mmol l<sup>-1</sup>, total dissolved sulfide(i.e. ΣH<sub>2</sub>S)=250-300 μmol l<sup>-1</sup>, dissolved O<sub>2</sub>=90-180 μmol l<sup>-1</sup>, dissolved NO<sub>3</sub>=40-50μmol l<sup>-1</sup>, pH=6.5, temperature=12°C, pressure=27.5 MPa. Riftia were maintained in these conditions until `autotrophic'. Autotrophic Riftia exhibit a net uptake of dissolved inorganic carbon(DIC), oxygen and sulfide, as well as net elimination of proton equivalents. This regularly occurs after 12 h following incubation.

During each experiment, while one or more factors were being varied, all other dissolved substrate concentrations, as well as pH and temperature, were held at the `typical' conditions previously described. Also during all experiments, tubeworms were maintained at each interval for at least 1 h, or until uptake rate reached a steady state. At the end of each experiment, worms were promptly removed, weighed on a motion-compensated shipboard balance(Childress and Mickel, 1980),dissected and frozen in liquid nitrogen for later analysis. In some cases,empty worm tubes were returned to the pressure vessels for several hours, and subjected to the same experimental conditions to determine what fraction, if any, of the observed flux rates are attributable to bacterial growth or other phenomena associated with the tubes. No significant contribution of bacteria to the observed metabolite flux rates was measured in this or prior studies(Girguis et al., 2000). All mass-specific rates are expressed in terms of wet mass.

To determine which chemical species of hydrogen sulfide is taken up by Riftia (sulfide, H₂S, or bisulfide, HS^-), as well as the duration of uptake, four Riftia weighing 11.9-18.1 g each were placed into two of the high-pressure aquaria immediately after being collected during both the HOT 96 and HOT 97 expeditions (two worms were placed into each vessel). During the HOT 96 expedition, tubeworms were maintained until autotrophic and then ΣH₂S was reduced to 50 μmol l^-1, while seawater pH was reduced to 5.66 over a 4 h period. Afterwards, seawater ΣH₂S was increased to 465 μmol l^-1 over a period of 18 h, at increments of 25-50 μmol l^-1. Next, seawater ΣH₂S was again lowered to 50μmol l^-1 for 4 h while the pH was increased to 7.4 and seawaterΣCO₂ was increased to 24 mmol l^-1 (to maintain an equivalent dissolved carbon dioxide concentration)(Childress et al., 1993; Goffredi et al., 1997b). Seawater ΣH₂S was then increased incrementally to 480 μmol l^-1 over a period of 11 h. During the HOT 97 expedition, tubeworms were maintained until autotrophic and then seawater pH was maintained at 5.8 while ΣH₂S was held at 359.3±8.26 μmol l^-1. pH was then increased and maintained at 7.48 over a 4 h period while ΣH₂S was maintained at 346.7±14.16 μmol l^-1.

To examine the relation between seawater ΣH₂S concentrations and ΣH₂S uptake, four Riftia weighing 13.4-15.1 g were maintained at typical in situ conditions for 10 h during the HOT 97 expedition. Seawater ΣH₂S was then increased from 0 to 870 μmol l^-1 over a period of 12 h, at increments between 50 and 100 μmol l^-1.

To determine the duration that chemoautotrophy can be sustained by blood-bound sulfide, three Riftia weighing 12.5-14.5 g each were placed into high-pressure aquaria during the HOT 97 expedition (the two smaller worms were placed in one vessel). Seawater ΣH₂S was lowered by decreasing the flow of sulfide gas into the equilibration column. Seawater ΣH₂S was monitored constantly until it decreased to below our level of detection (ca. 5 μmol l^-1)(Childress et al., 1984). pH was maintained at 6.1, and all other factors were held at the `typical'conditions previously described.

To examine the relation between oxygen and ΣH₂S uptake over a range of experimental ΣH₂S concentrations, two experiments were conducted during the HOT97 expedition. In the first experiment, two Riftia weighing 12.2-17 g each were placed into the high-pressure respirometers (one worm per chamber) and maintained in 80μmol l^-1 ΣH₂S. All other substrates were held at `typical' concentrations. Next, ΣH<sub>2</sub>S was incrementally increased from 80 μmol l<sup>-1</sup> to 208 μmol l<sup>-1</sup> over 23 h. In the second experiment, two Riftia weighing 9.1-13.6 g each were placed into the high-pressure respirometers (one worm per chamber) and maintained in 200 μmol l<sup>-1</sup> ΣH<sub>2</sub>S. All other substrates were held at `typical' concentrations. Next, ΣH₂S was incrementally increased from 200 μmol l^-1 to 843 μmol l^-1 over 26 h. During both experiments, pH was maintained at 6.1.

To examine the stoichiometric relation between the other major substrates and oxygen uptake, five Riftia weighing 7.3-12.1 g each were placed into two of the high-pressure aquaria during the HOT 96 expedition (three worms were placed into one vessel and two worms were placed into the other vessel). Seawater oxygen concentration was increased from 40 to 210 μmol l^-1 over a period of 23 h, at increments between 15 and 40 μmol l^-1. pH was maintained at 5.9, while all other factors were held at the `typical' conditions previously described.

To determine the duration that chemoautotrophy can be sustained by blood-bound oxygen, three Riftia weighing 11.4-14.2 g each were placed into high-pressure aquaria during the HOT 97 expedition. Seawater oxygen was then quickly decreased to below our level of detection by gas chromatography (about 5 μmol l^-1) by stopping the flow of oxygen and increasing the flow of N₂ into the equilibration column. pH was maintained at 5.9, while all other factors were maintained at typical in situ conditions.

To examine the relation between Riftia CO₂ uptake and experimental ΣCO₂ concentrations, three Riftiaweighing 16-17 g each were placed into high-pressure aquaria during the HOT 96 expedition. Seawater ΣCO₂ was increased from 2.1 to 16.5 mmol l^-1 over a period of 25 h (at 1 to 2.5 mmol l^-1increments) while maintaining pH at 5.9 via proportional pH control.

To examine the relation between Riftia bicarbonate uptake and experimental ΣCO₂ concentrations, the aforementioned Riftia were subject to the same experiment previously described,except that seawater pH was maintained at 6.6 for the duration of the experiment. In both these experiments, all other substrates were maintained at typical in situ concentrations.

To examine the effects of temperature on Riftia host and symbiont metabolism, two experiments were conducted during the HOT 96 and LARVE 98 expeditions. During the HOT 96 experiment, four Riftia weighing 10.5-15 g each were placed into two of the high-pressure aquaria (two worms in each aquaria). Initially, temperature was decreased to 5°C for 7 h,increased to 10°C for 4 h, and then to 20°C for 3 h. During the LARVE 98 expedition, four Riftia weighing 12-16 g each were placed in high-pressure aquaria (two in each aquaria). After the onset of autotrophy,temperature was increased to 20°C for 2 h, 27.5°C for 3 h, 30°C for 2 h and 35°C for 2 h.

In order to assess the variation in substrate uptake rates among individual Riftia collected from different sites, we collected twelve Riftia weighing 12.2-14.1 g, from three different geographical locales, during our HOT 96, HOT 97 and LARVE 98 expeditions. The HOT 96, HOT 97 and LARVE 98 worms were collected from tubeworm clumps located near 12.48N,103.56W, 9.46N, 104.16W, and 9.50N, 104.17W, respectively, at approximately 2250 m. All worms were collected via the DSV Alvin, and brought to the surface in a thermally insulated container. All Riftiawere maintained in our high-pressure respirometry system at typical conditions until autotrophic, during which time metabolite uptake and elimination were recorded for 7 h or more.

To examine the relation between environmental substrate concentration and Riftia net carbon fixation (primary productivity), three experiments were conducted during the HOT 97 expedition in which Riftia were maintained in different experimental conditions that mimic `typical', `better'and `best' habitats for chemoautotrophic function. In each experiment, four Riftia were placed in the high-pressure flow-through aquaria until the onset of autotrophy. Seawater conditions were then adjusted to simulate the conditions at different diffuse flow sites. `Typical' conditions wereΣCO₂=3.5±0.5 mmol l^-1,ΣH₂S=67±12.1 μmol l^-1,O₂=97±9.2 μmol l^-1, temperature= 10°C,NO₃^-=40 μmol l^-1, pressure=27.5 MPa.`Better' conditions were ΣCO<sub>2</sub>=4.6±0.7 mmol l<sup>-1</sup>, ΣH<sub>2</sub>S=167± 14.2 μmol l<sup>-1</sup>,O<sub>2</sub>=112±7.2 μmol l<sup>-1</sup>, temperature=15°C,NO<sub>3</sub><sup>-</sup>=40 μmol l<sup>-1</sup>, pressure=27.5 MPa.`Best' conditions were ΣCO₂=10.8±0.5 mmol l^-1, ΣH₂S=256±12.7 μmol l^-1,O₂= 197±24 μmol l^-1, temperature=15°C,NO₃^-=40 μmol l^-1, pressure=27.5 MPa. All conditions were maintained for at least 15 h. Metabolite uptake rates recorded after the first 8 h were used to calculate mean metabolite uptake rates. All rates are expressed in terms of wet mass.

Data were collected by Labview 4.0. Rates were calculated using Microsoft Excel, and all statistical analyses and regression plots were rendered on Statview 5.0 (SAS Inc., Cary, NC, USA). 3-dimensional plots were rendered on Transform (Fortner, Inc., Boulder, CO, USA)

Data from the HOT 96 expedition demonstrate that RiftiaΣH₂S uptake occurs over a wide range of ΣH₂S concentrations at both pH 5.66 and 7.48 and increases with increasing seawaterΣH₂S concentration (Fig. 1A,B). At pH 5.66, Riftia's ΣH₂S uptake rate is more responsive to increasing sulfide concentrations (as seen by the steeper slope in Fig. 1A). At both pH values the relation between ΣH₂S uptake and increasing environmental ΣH₂S concentration appeared linear between 100 and 450 μmol l^-1(Fig. 1A,B). In a separate experiment, when pH was maintained at either 5.73 or 7.73 and seawaterΣH₂S concentrations were held constant, RiftiaΣH₂S uptake was continuous for over 14 h but there were no significant differences in proton elimination rates or oxygen uptake rates(P>0.05; Spearman correlation and Mann-Whitney U-test, Table 1).

While increasing seawater sulfide concentrations up to 600 μmol l^-1 stimulated sulfide and oxygen uptake as well as proton elimination (Fig. 2), higher sulfide concentrations resulted in the diminishment of bothΣH₂S and O₂ uptake rates(Fig. 2). During all experiments, ΣH₂S uptake rate correlated to oxygen uptake rate (P=0.0017; Spearman correlation). ΣCO₂ uptake,however, did not linearly correlate to ΣH₂S or O₂uptake, and appeared to decrease at higher seawater ΣH₂S concentrations.

When ΣH₂S concentrations were reduced to below the limits of detection (BLD) (Childress et al.,1984), O₂ uptake rates were reduced to 2.88±0.89μmol g^-1 h^-1(Table 2). However,ΣCO₂ uptake was sustained for 5.3 h and O₂ uptake was sustained for 3.5 h (after which O₂ uptake continued at approximately 20% of its original rate; Table 2).

O₂ uptake strongly correlated with seawater oxygen concentration(P=0.0001; Spearman correlation; Fig. 3). TotalΣH₂S uptake also strongly correlated with oxygen uptake rate at oxygen concentrations between 50 and 200 μmol l^-1(P=0.0001, Spearman correlation; Fig. 3). Proton elimination rate also correlated with seawater oxygen concentration (P=0.04;Spearman correlation; Fig. 3). No significant linear correlation was found between ΣCO₂uptake and O₂ uptake rate (Fig. 3).

In two experiments, the ratio of Riftia's O₂ uptake perΣH₂S uptake was >2 at environmental ΣH₂S concentrations between 100 and 200 μmol l^-1 (autotrophic O₂ uptake was determined by subtracting the heterotrophic O₂ uptake rates measured prior to the onset of autotrophy; Fig. 4A). At higher concentrations of environmental ΣH₂S, however, the ratio dropped to <2 (Fig. 4B).

Riftia ΣCO₂ uptake correlated with CO₂but not HCO₃^- concentrations(Fig. 5). ΣCO₂uptake appeared to plateau at 8 mmol l^-1 CO₂concentrations, or approximately 16 mmol l^-1 total inorganic carbon(Fig. 5) with a maximumΣCO₂ uptake rate of about 34 μmol g^-1h^-1 between 7 and 8 mmol l^-1(Fig. 5). Experiments at higherΣCO₂ were attempted but not completed due to problems with gas solubility and decreased analytical resolution.

Riftia maintained at 5°C had O₂ andΣH₂S uptake rates of 3.88±0.66 and 1.18±0.73μmol g^-1 h^-1, respectively, and netΣCO₂ production (not uptake) of 1.09±0.80 μmol g^-1 h^-1 (Fig. 6). At 10°C ΣCO₂ uptake was 1.75±0.52μmol g^-1 h^-1. From 10 to 25°C,ΣCO₂, ΣH₂S and O₂ uptake rates increased, with a Q₁₀ of approximately 2.3. The sharp increase ofΣCO₂, ΣH₂S and oxygen uptake that occurs at 25°C is a marked departure from the trend at lower temperatures. Optimal temperature for maximal Riftia ΣCO₂ uptake is approximately 27°C. Temperatures above 28°C resulted in sublethal reductions in all three measured metabolite uptake rates. Lethal temperature was reached between 30 and 35°C.

Riftia maintained in three different environmental conditions (`typical',`better' and `best') exhibited significant differences in metabolite uptake rates as well as proton elimination rates(Table 3). Molar ratios ofΣCO₂ uptake to ΣH₂S uptake ranged from 0.42 at the lower conditions to 1.06 at optimal conditions. Percent energy devoted to carbon fixation was calculated from the energy required to reduce the inorganic carbon to sucrose (-495 kJ mol^-1)(Kelly, 1982) and from the energy available from the oxidation of bisulfide to sulfate viaoxygen (-995 kJ mol^-1) (Kelly,1982), and ranged from 21% to 53% at the typical and optimal conditions, respectively.

No significant differences in ΣH₂S, ΣCO₂,and oxygen uptake rates were observed between the Riftia collected from the `BIOTRANSECT 2' site and the `13 North' site(Table 4). However, Riftia collected from the `BIOTRANSECT 1' site exhibited metabolic uptake and elimination rates that were significantly different from the other two individuals (P=0.0001, Mann-Whitney U-test) and were on average 40-66% lower than the other two individual Riftia(Table 4). Significant differences in proton elimination rates were observed among all three Riftia tubeworms (Table 4). While there were no superficial differences among the Riftia, during subsequent dissections post-experimentation the Riftia collected during the LARVE 98 cruise were found to contain blackened trophosome in stark contrast to the green and red trophosomes of the other worms (black trophosomes likely indicate poor symbiont health)(Fisher et al., 1988a).

While we experimented on worms ranging from 4 g to 19 g, this range was not to examine the effects of scaling on metabolic processes.

Chemosynthetic production depends upon the oxidation of a reduced substrate. In the current experiments, ΣH₂S and O₂uptake are highly correlated to one another when seawaterΣH₂S and O₂ concentrations are between 10 and 400μmol l^-1 (Fig. 3). These concentrations are largely representative of those found in situ (Shank et al.,1998; Luther et al.,2001; Mullineaux et al.,2003; Le Bris et al.,2006). Prior studies have shown that symbiont sulfide oxidation is stimulated by oxygen (Fisher and Childress, 1984; Girguis et al., 2000). The data shown here(Fig. 3) demonstrate that symbiont sulfide oxidation is the primary factor influencing Riftiaoxygen uptake, and is likely responsible for consuming the majority of acquired oxygen. Although nitrate is present at 40 μmol l^-1in situ and may serve as a terminal electron acceptor for symbiont sulfide oxidation (Hentschel and Felbeck,1993), a prior study found no correlation between Riftianitrate and ΣH₂S uptake, and that sulfide oxidation cannot be sustained solely by nitrate reduction(Girguis et al., 2000).

Because Riftia flourishes in the vent-seawater mixing regimes,simultaneous exposure to both sulfide and oxygen is not likely to be continuous and there may be periods of time in which Riftia is not exposed to one substrate or the other(Arndt et al., 1998). However,when we reduced seawater ΣH₂S concentrations to below our level of detection, Riftia ΣCO₂ uptake was sustained for 5.3 h, after which Riftia exhibited net ΣCO₂production and decreased oxygen consumption (the remaining oxygen uptake is the host's aerobic respiration; Table 2). We believe this lag time reflects the consumption of hemoglobin-bound sulfide in the vascular and coelomic hemolymph pools(Arp and Childress, 1983; Childress et al., 1984; Zal, 1998). A prior study found that Riftia typically consists of 13.9% vascular blood and 19.5% coelomic fluid (Childress et al.,1984), and recent studies have found that vascular and coelomic hemoglobin concentrations in Riftia are 3 and 0.5 mmol l^-1, respectively (J.J.C., unpublished). Using these values and binding stoichiometries, and assuming that sulfide uptake rates prior to the removal of sulfide reflect symbiont sulfide usage, we estimate that a worm weighing 15 g should have enough bound sulfide to sustain autotrophy for approximately 6 h, a figure comparable to our experimentally determined value.

Although the presence of sulfide is a prerequisite to successful colonization and growth of Riftia, exposure to elevatedΣH₂S concentrations in situ may be detrimental to Riftia's survival. Our data suggest that inhibition of symbiont metabolism may have occurred after seawater ΣH₂S concentrations reached 700 μmol l^-1(Fig. 2). This is higher than the 300 μmol l^-1 ΣH₂S concentrations that inhibited isolated symbionts in a prior study(Fisher and Childress, 1984). After exposure to high sulfide concentrations, Riftia's oxygen uptake rate was comparable to the oxygen uptake rates measured after eliminating sulfide, further suggesting that symbiont sulfide oxidation was diminished. Accordingly, the remaining oxygen uptake likely represents the heterotrophic contribution of the host to total oxygen consumption. While we cannot precisely ascertain the effect of elevated sulfide concentrations on the Riftia's aerobic respiration, the similarity to the rates observed in the absence of sulfide suggests that Riftia is not prone to sulfide toxicity at 700 μmol l^-1 seawater ΣH₂S concentrations (Fig. 2). Nevertheless, we observed that all Riftia exposed toΣH₂S concentrations greater than 1.7 mmol l^-1 in our high-pressure aquaria quickly die. While prior measurements of sulfide concentrations around Riftia clumps in situ have shown that concentrations vary from 0 to 500 μmol l^-1 in the water surrounding the worms (Johnson et al.,1988), other studies have measured sulfide concentrations around Riftia clumps of approximately 2 mmol l^-1(Shank et al., 1998). These observations suggest that Riftia may be exposed to higher levels of sulfide in situ than previously thought, and may experience symbiont sulfide inhibition in situ.

Oxygen inhibition of symbiotic function was not observed to occur at environmentally relevant oxygen concentrations(Fig. 3) even though these are much higher than the concentrations shown to inhibit such function in symbiont preparations (Fisher and Childress,1984; Fisher et al.,1989; Scott et al.,1994). While these prior studies demonstrated the role of oxygen in sustaining sulfide oxidation by isolated symbionts, they also showed that they are microaerophilic, using oxygen as a terminal electron acceptor in sulfide oxidation but being inhibited at low concentrations of free oxygen. Although no data are available on the free oxygen concentrations (i.e. unbound oxygen) within the bacteriocytes of intact associations, the present data support previous suggestions that free oxygen concentrations within the trophsosome are very low due to the high concentrations of very high oxygen affinity hemoglobins in Riftia vascular and coelomic fluids.

Prior studies of Riftia have suggested that the species of sulfide acquired by the worm is bisulfide(Goffredi et al., 1997a). In our experiments, Riftia sustained similar ΣH₂S uptake rates over a range of environmental ΣH₂S concentrations at both acidic and basic pH values, (ca. 5.5 and ca. 7.6; Fig. 1 and Table 1). At pH 5.5,approximately 99% of the ΣH₂S is hydrogen sulfide (the pK_a is approximately 6.8 at the conditions in our respirometer system). At pH 7.6, approximately 90% of the ΣH₂S is bisulfide. Riftia's uptake rates at each pH demonstrate that both H₂S and HS^- can be acquired because the uptake of the minor sulfide species could not support the observed mass-specificΣH₂S uptake rates. Although Riftia may possess mechanisms that reduce the influx of membrane-permeable hydrogen sulfide in order to limit sulfide toxicity (Menon et al., 1995), such a mechanism(s) does not entirely uncouple H₂S uptake and CO₂ uptake by Riftia. Rather, it buffers the passive diffusion of hydrogen sulfide into the tissues. However,rapid conversion of H₂S to HS^- within the gill epithelia[where intracellular pH is about 7.4(Goffredi et al., 1999)] could significantly reduce the risk of mitochondrial sulfide poisoning as HS^- is the species bound by Riftia's hemoglobins(Childress et al., 1984; Goffredi et al., 1997a; Flores et al., 2005).

While the precise role of each of these mechanisms remains to be determined, epithelial mitigation of H₂S diffusion, together with the high-affinity HS^- binding hemoglobins, may be the most effective means by which Riftia minimizes the effects of sulfide toxicity on host aerobic pathways while maintaining a large pool of sulfide available for symbiont metabolism.

We also observed higher ratios of oxygen uptake to ΣH₂S uptake at lower seawater ΣH₂S concentrations(Fig. 4A). Because the oxidation of sulfide to sulfate stoichiometrically requires 2 O₂per sulfide, O₂ uptake:ΣH₂S uptake ratios >2 can support the complete oxidation of ΣH₂S to sulfate. This observation supports the prior hypotheses that sulfide oxidation yields sulfate as an end product (Girguis et al.,2002; Goffredi et al.,1997a; Wilmot and Vetter,1990). However at higher seawater ΣH₂S concentrations, when the ratio dropped to <2(Fig. 4B), we posit that a fraction of the ΣH₂S may be oxidized to elemental sulfur. Elemental sulfur is commonly found in high concentrations in the trophosome of healthy Riftia, and is thought to be a means of storing substrate(Fisher et al., 1988a; Childress et al., 1991). In addition, some fraction of the reductive potential from sulfide oxidation is likely used in inorganic carbon fixation, and that may lead to a shift in the ratio of oxygen uptake to ΣH₂S uptake. Thus, the data shown in Fig. 4 suggest that the end product of sulfide oxidation gradually shifts from sulfate to sulfur at higher seawater ΣH₂S concentrations, and that carbon fixation may increase as the reductive potential from sulfide is more available.

Our data demonstrate that CO₂ is the chemical species of inorganic carbon that is acquired (Fig. 5), which is consistent with previous in vitro and whole animal studies (Childress et al.,1993; Fisher et al.,1988b; Fisher et al.,1990; Goffredi et al.,1997b; Scott,2003). There is no indication that bicarbonate is acquired, even at higher pH. The ΣCO₂ uptake rate appears highly correlated to environmental CO₂ concentrations(Fig. 5). Because higher environmental CO₂ concentrations would provide a larger gradient and thus more rapid diffusion (Goffredi et al., 1997b), the asymptote of ΣCO₂ uptake rates at 8 mmol l^-1 environmental carbon dioxide concentrations may reflect a physiological or biochemical limitation in symbiont carbon fixation,although further studies would be required to verify this hypothesis.

Although linear correlations between Riftia ΣCO₂uptake rate and ΣH₂S or oxygen uptake rate were never observed (Fig. 5), our data show that Riftia carbon uptake is stimulated by exceeding `threshold'seawater ΣH₂S and oxygen concentrations(Fig. 7). Future studies should continue to interrogate the relation between energy production (viasulfide oxidation) and carbon fixation.

The issue of carbon limitation in Riftia has been debated for some time (Fisher et al., 1988b; Fisher et al., 1990; Scott, 2003). While our data show that Riftia acquires only 12.5% of the available CO₂(implying that Riftia is not carbon limited), Riftia'sΣCO₂ uptake rate consistently responded to increasing seawater ΣCO₂ throughout the duration of the experiment (up to 16 mmol l^-1 ΣCO₂). However, this observation that Riftia ΣCO₂ uptake is, strictly speaking,responsive to changes in seawater ΣCO₂ concentrations does not imply that Riftia is carbon limited. This may be attributable to limitations in another substrate besides DIC. Furthermore, we did not determine if increasing seawater ΣCO₂ concentrations led to biomass accumulation or, alternatively, glycogen accumulation, so the precise relation between increased carbon uptake (and presumably fixation) and growth remains unresolved. This too warrants further investigation.

The strongest determinant of metabolite flux, besides limiting substrate concentrations, was temperature. Fig. 6 shows that a sharp increase in ΣCO₂,ΣH₂S and oxygen uptake occurs at 25°C, a marked departure from the trend at lower temperatures. These data suggest that optimal temperature for maximal Riftia uptake, presumably a reflection of symbiont primary productivity, is between 25 and 27°C. Prolonged exposure to temperatures above 32 to 35°C appears to be lethal as all Riftia maintained at these temperatures were dead after 2 h. While a prior study suggested that Riftia tubeworms were growing rapidly in diffuse vent flows with temperatures ca. 35°C(Shank et al., 1998), diffuse vents are complex thermal regimes and it is unlikely that Riftiaencounters chronic exposure to these high temperatures. Instead, Riftia may tolerate acute exposure to high temperature in order to acquire the sulfide necessary to sustain symbiont autotrophic metabolism. It is notable that Riftia's maximal metabolite uptake (and therefore symbiont primary production) occurs at temperatures near their maximal thermal tolerance.

At steady state, when both the experimental conditions and Riftiametabolite uptake have remained constant for several hours,ΣCO₂ and ΣH₂S uptake are reliable proxies for carbon fixation and sulfide oxidation rates because they represent the continuous rate of substrate utilization by the symbionts. Accordingly, we determined the mean molar ratios of Riftia ΣCO₂ andΣH₂S uptake to examine the stoichiometric relation between carbon fixation and substrate oxidation by the chemoautotrophic symbionts(Table 3). In our experiments, Riftia's ΣCO₂:ΣH₂S uptake ratio varied from 0.42 to approximately 1.06 over a range of environmentally relevant substrate concentrations (Table 3). In a prior study of the bivalve Solemya reidii, a clam with chemoautotrophic symbionts in its gill filaments,ΣCO₂ and ΣH₂S uptake molar ratios of 0.86-0.92 were measured (Anderson et al.,1987). These ratios can also be expressed as `efficiencies', in which the energy utilized in carbon fixation (the conversion of CO<sub>2</sub>to organic carbon) is expressed as a percentage of the total energy available from the oxidation of sulfide to sulphate(Kelly, 1982). Riftiaefficiencies range from 21% to 53% at `typical' and `best' conditions,respectively. In general, it has been observed that more than 80% of the total energy budget of non-hydrogen-oxidizing chemolithotrophs is used in converting carbon dioxide to carbohydrates (Kelly,1982). The allocation of this energy has been used to explain why the growth yields of chemolithotrophs (already limited by the relatively low molar energy yield of their substrates) are in general rather meager(Kelly, 1990). However, our data demonstrate that Riftia symbionts allocate a smaller percentage of their total energy to carbon fixation and nitrate reduction when compared to free-living chemolithotrophic bacteria(Kelly, 1990). This may be attributable to their symbiotic lifestyle since these bacteria do not have to support a myriad of other energy intensive tasks (e.g. spinning flagellae)common among free-living bacteria. These data, as well as the high rates of substrate utilization by Riftia, may explain how Riftiasustains its rapid growth.

We observed substantial individual variation in metabolite flux. Under identical experimental conditions, individual Riftia exhibited differences in ΣCO₂ uptake that ranged from 4.3 to 15.7μmol g^-1 h^-1(Table 4). The differences in these carbon uptake rates may reflect the history of the habitat at different collection sites, and those worms with the highest ΣCO₂uptake rates may have been collected from tubeworm clumps growing atop ample diffuse flow and as such have more metabolically active symbionts. Our observation of blackish trophosomes in the worms with the lowest chemoautotrophic metabolic rates supports this supposition. In situconditions are highly variable and as such can strongly affect Riftiasymbiont metabolism.

At steady state, Riftia net ΣCO₂ uptake reflects the rate of chemoautotrophic carbon fixation, and can be considered net primary productivity. After placing freshly collected Riftia into the high-pressure aquaria and prior to the onset of autotrophy, we measured the response of ΣCO₂ flux to increases in eitherΣH₂S or O₂ uptake, and observed no discernable change in ΣCO₂ flux. However, we observed that concomitant increases in both ΣH₂S and O₂ uptake correlated with ΣCO₂ uptake. Specifically, ΣCO₂ uptake drastically increases after seawater ΣH₂S and oxygen concentrations exceed 86 and 95 μmol l^-1, respectively(Fig. 7). In every respirometry experiment conducted to date, the onset of autotrophy was preceded by a rapid increase in Riftia ΣH₂S and O₂ uptake(enough to consume much of the dissolved metabolite in the aquaria). Whereas in a prior study Riftia required ΣH₂S concentrations greater than 90 μmol l^-1 to support net carbon fixation(Childress et al., 1991), the current study measured net ΣCO₂ uptake occurring at substantially lower levels of sulfide and oxygen, e.g. 50 μmol l^-1 and 70 μmol l^-1 respectively(Fig. 3), but only after the threshold ΣH₂S and oxygen concentrations had been exceeded prior to being reduced. The observed phenomenon suggests that (i) carbon fixation is directly mediated by the binding and loading of oxygen and sulfide by Riftia hemoglobins or, alternatively, (ii) that Riftia(or its symbionts) actively modulates inorganic carbon uptake in response to seawater substrate concentrations, maintaining modest carbon fixation until seawater substrate concentrations are sufficient to support elevated primary productivity. Further studies are required to better address these hypotheses.

In concert, these data demonstrate that Riftia metabolite uptake is strongly governed by environmental substrate availability and temperature. Riftia symbiont carbon fixation was observed to be highest after sufficient oxygen and sulfide has been acquired by Riftia, and when temperatures are relatively high. While the relation between symbiotic function and environmental variability is both facilitated and complicated by the presence of the host, the ultimate constraint on symbiont autotrophic function is the availability of substrates from the environment, and in general Riftia is extremely well-poised to buffer the spatial and temporal variations that are characteristic of diffuse flow regimes. Future studies using longer time-averages of metabolite flux may allow us to develop predictive models of environmental conditions based upon biota observations and, conversely, models of Riftia primary productivity based on in situ chemical and temperature measurements.

The authors would like to thank the captains and crew of the RV New Horizon, RV Atlantis II and RV Atlantis, DSRV Alvin, RV Wecoma, RV Nadir and DSRV Nautile. We also thank Dr S. Goffredi for her help and humor, Dr E. F. DeLong for his support and patience, as well as Dr R. Trench for his editorial efforts. Special thanks go to Dr F. Gaill, Dr A. Chave and Dr D. Manahan, chief scientists of the 1996, 1997 and 1998 expeditions. Funding for this project was provided by NSF grants OCE-9301374,OCE-9632861 and OCE-002464.