CO₂ Uptake and Fixation by Endosymbiotic Chemoautotrophs from the Bivalve Solemya velum

Abstract

Chemoautotrophic symbioses, in which endosymbiotic bacteria are the major source of organic carbon for the host, are found in marine habitats where sulfide and oxygen coexist. The purpose of this study was to determine the influence of pH, alternate sulfur sources, and electron acceptors on carbon fixation and to investigate which form(s) of inorganic carbon is taken up and fixed by the gamma-proteobacterial endosymbionts of the protobranch bivalve Solemya velum. Symbiont-enriched suspensions were generated by homogenization of S. velum gills, followed by velocity centrifugation to pellet the symbiont cells. Carbon fixation was measured by incubating the cells with ¹⁴C-labeled dissolved inorganic carbon. When oxygen was present, both sulfide and thiosulfate stimulated carbon fixation; however, elevated levels of either sulfide (>0.5 mM) or oxygen (1 mM) were inhibitory. In the absence of oxygen, nitrate did not enhance carbon fixation rates when sulfide was present. Symbionts fixed carbon most rapidly between pH 7.5 and 8.5. Under optimal pH, sulfide, and oxygen conditions, symbiont carbon fixation rates correlated with the concentrations of extracellular CO₂ and not with HCO₃⁻ concentrations. The half-saturation constant for carbon fixation with respect to extracellular dissolved CO₂ was 28 ± 3 μM, and the average maximal velocity was 50.8 ± 7.1 μmol min⁻¹ g of protein⁻¹. The reliance of S. velum symbionts on extracellular CO₂ is consistent with their intracellular lifestyle, since HCO₃⁻ utilization would require protein-mediated transport across the bacteriocyte membrane, perisymbiont vacuole membrane, and symbiont outer and inner membranes. The use of CO₂ may be a general trait shared with many symbioses with an intracellular chemoautotrophic partner.

Nutritive symbioses between chemoautotrophic bacteria and invertebrates are found in many marine habitats where oxygen and sulfide coexist, including coastal reducing sediments, deep-sea hydrothermal vents, and hydrocarbon seeps (2, 37). Many of the invertebrate partners have a reduced or absent digestive tract and rely to various degrees on their bacterial symbionts for organic carbon and nitrogen (17, 41). In most cases, the symbiont-containing organs of these associations have moderate to high activities of ribulose 1,5-bisphosphate carboxylase/oxygenase (RubisCO) (3, 17). RubisCO utilizes CO₂, and not HCO₃⁻ (11), for carbon fixation. At acidic pHs, CO₂ is the most abundant form of dissolved inorganic carbon (DIC, comprising CO₂, HCO₃⁻, and CO₃²⁻), but at physiological and marine pH values, HCO₃⁻ dominates (29). Many marine photoautotrophs cope with the environmental abundance of HCO₃⁻ and paucity of CO₂ by transporting HCO₃⁻ across their cell membranes and converting it to CO₂ in the cytosol with carbonic anhydrase (24). Recently, a marine sulfur chemoautotroph has also been demonstrated to use extracellular HCO₃⁻ (12). Whether symbiotic marine chemoautotrophs have similar adaptations has been determined only for the endosymbionts of hydrothermal vent vestimentiferans, which cannot use extracellular HCO₃⁻ (36). Symbionts from other associations, particularly those present in habitats with CO₂ concentrations lower than those at the vents, might be able to transport HCO₃⁻.

Solemya velum is a protobranch bivalve that lives in symbioses with sulfur-oxidizing chemoautotrophic gamma-proteobacteria (5, 14). This bivalve inhabits Y-shaped burrows in sulfide-containing silty sediments near eelgrass beds off the coast of the northeastern United States (40). S. velum clams position themselves above the downward-pointing stem of their burrows and pump water both from the oxygenated upper portion of the burrow and from the anoxic, sulfide-containing lower stem (27, 44). Thiosulfate resulting from chemical sulfide oxidation may be entrained from the surrounding suboxic sediments (16, 19, 42) by burrow ventilation. Thiosulfate may also be generated by the clams themselves via mitochondrial sulfide oxidation, as has been documented for Solemya reidi (31, 33). As a result of the behavioral and biochemical activities of the clam, the symbionts are provided with a supply of sulfide, thiosulfate, and oxygen.

The symbionts are found within epithelial cells, termed bacteriocytes, which line the gill filaments (5). RubisCO activity is high in S. velum gills (5); further, this enzyme has been localized within the cytoplasm of the symbiotic bacteria by immunocytohistochemistry (6). Enzymes consistent with the oxidation of reduced inorganic sulfur compounds are present in the gills (7), and carbon fixation by intact gills is stimulated by the presence of sulfide and thiosulfate (5), indicating that the symbionts are sulfur-oxidizing chemoautotrophs. Based on the bivalve's reduced gut (34), its diminished ability to filter feed (26), and the similarity in the lipid and isotopic compositions of the hosts and symbionts (9, 10), it is well supported that S. velum obtains most of its organic carbon from symbiont carbon fixation (26).

The mechanism by which the symbionts obtain CO₂ from their environment was not known, and the influence of pH, as well as the presence and concentrations of reduced sulfur compounds, oxygen, and nitrate, had not been characterized. CO₂ concentrations in the interstitial water surrounding S. velum burrows range from 20 to 200 μM (39), elevated relative to that in surface seawater (∼20 μM) but low compared to that in hydrothermal vents (20 to 1,000 μM) (22), which may necessitate adaptations to supply the symbionts with adequate dissolved inorganic carbon for carbon fixation. With respect to electron acceptors, symbionts may experience periods of diminished oxygen availability, when the clam is not actively ventilating its burrow. Under these circumstances, it would be helpful for them to be able to use nitrate as an electron acceptor in order to sustain themselves. Given that prior studies similar to this one had focused primarily on symbionts from hydrothermal vent organisms (8, 36, 38), this study was performed to determine whether “nonvent” symbionts might have unique requirements for redox substrates and dissolved inorganic carbon.

The purposes of this study were (i) to find concentrations of inorganic sulfur compounds (sulfide and thiosulfate) and oxidants (oxygen and nitrate) and pH values that would sustain high rates of symbiont chemoautotrophy and (ii) to determine whether S. velum symbionts utilize both CO₂ and HCO₃⁻. S. velum clams were collected, and enriched suspensions of symbionts were generated from gill homogenates. The effects of pH and the concentrations of sulfide, thiosulfate, oxygen, and nitrate on symbiont carbon fixation were measured radiometrically with [¹⁴C]DIC. Optimum conditions determined from these experiments were then utilized to elucidate which form of dissolved inorganic carbon the symbionts were capable of using for carbon fixation.

MATERIALS AND METHODS

Solemya velum collection.

Solemya velum clams were collected in June and July 2003 from eelgrass bed sediments near Woods Hole, MA, and kept in chilled seawater until use within 24 h of collection, since symbiont abundance and activity diminish upon collection (4).

Reagents and buffers.

Incubations were conducted in mixtures of unbuffered and buffered artificial seawater (ASW and BASW, respectively; ASW consists of 0.36 M NaCl, 0.01 M KCl, 0.055 M MgCl₂, 0.01 M CaCl₂, and 0.03 M Na₂SO₄; 0.05 M HEPES is added for BASW). To remove DIC, ASW and BASW were brought to pH 2 with HCl and sparged for 10 to 20 min with mixtures of CO₂-free air and N₂. CO₂-free air was prepared by passing air through a 30-cm column of soda lime. After sparging, the pHs of ASW and BASW were adjusted appropriately for a given experiment using HCl or freshly prepared DIC-free NaOH solutions.

Stock solutions of DIC (0.5 M), KNO₃ (0.01 M), Na₂S (0.01 M), and Na₂S₂O₃ (0.01 M) were prepared to generate appropriate dissolved gas and nutrient concentrations. For the DIC and KNO₃ solutions, ASW was sparged with N₂, NaHCO₃ or KNO₃ was added, and the solutions were sealed in glass serum vials. For the sulfide stock solution, Na₂S·9H₂O crystals were added to N₂-sparged distilled deionized water and sealed in glass serum vials beneath a N₂ headspace to prevent spontaneous oxidation. Na₂S₂O₃ solutions were prepared in the same manner for comparisons of symbiont utilization of Na₂S and Na₂S₂O₃ as electron donors.

A stock NaH¹⁴CO₃ solution (2 mCi/ml; pH 9.5; MP Biomedicals catalog no. 17441) was divided into 0.5-ml aliquots immediately upon arrival, sealed in glass ampoules, and refrigerated until use.

Symbiont enrichments.

Symbionts were enriched from Solemya velum gills via differential centrifugation. Gills from five to eight bivalves were excised for a given experiment, placed in BASW (pH 8), and homogenized with a loose-fitting Dounce homogenizer. The homogenate was filtered through a 100-μm-mesh nylon screen and centrifuged (5,000 × g, 5 min, 4°C). After the supernatant was decanted, the pellets were resuspended in ∼0.5 ml BASW or ASW, depending on the experiment (see below). Symbiont-enriched suspensions (“symbionts”) were kept on ice until use, which occurred within 10 min. Enrichments were examined by phase-contrast microscopy to verify that symbiont cells were abundant and intact.

To verify that intact symbiont cells were responsible for carbon fixation, a pilot experiment was conducted in which a symbiont suspension was split in half. One aliquot was centrifuged to pellet the symbionts, and the supernatant was retained. Carbon fixation by both aliquots was then measured. The supernatant sample did not fix carbon, indicating that carbon fixation measured in the symbiont-enriched suspensions was not due to host enzymes or RubisCO released by lysed symbiont cells.

Symbiont carbon fixation.

ASW or BASW (volumes as appropriate for each experiment; see below) was injected into 2-ml glass autosampler vials fitted with gas-impermeable septa and stir bars, which had been flushed with mixtures of CO₂-free air and N₂ to match the target oxygen concentration for the incubations. DIC and [¹⁴C]DIC were added by injecting portions of unlabeled and ¹⁴C-labeled NaHCO₃⁻ solutions with a gas-tight syringe to generate a specific activity of ∼2 μCi μmol⁻¹. Immediately before the start of an experiment, Na₂S or Na₂S₂O₃ solutions were injected into each vial (volumes as appropriate for each experiment). The experiments began by injecting 10 μl of symbiont-enriched suspensions into each vial, which brought the final volume of the incubation to 1 ml. Subsamples (200 μl) were removed at 5-min intervals using gas-tight syringes, injected into scintillation vials primed with 200 μl hot glacial acetic acid, sparged gently with air for 20 s, and left beneath a fume hood for 30 min before addition of 3 ml scintillation cocktail. Initial activities of [¹⁴C]DIC were measured by removing 10 μl from each incubation and adding it to a vial primed with 3 ml scintillation cocktail supplemented with 50 μl β-phenethylamine. ¹⁴C disintegrations per minute (DPM) in each time point and each initial activity vial were measured with a scintillation counter.

Headspace gas compositions were maintained during sampling by piercing the septum with a needle fitted to a balloon filled with gas mixes of CO₂-free air and N₂ (as appropriate to the experiment); this measure also prevented the outgassing of the incubation due to the buildup of a vacuum from removal of subsamples. Carbon fixation rates were normalized to the amount of protein present in each incubation, which was determined using the Bradford assay (Bio-Rad). This overall procedure was used for all symbiont incubation experiments, with variations in the factors being tested (e.g., pH, oxygen concentration, and DIC concentration) as described below.

Two subsamples were taken at 5 and 10 min (determined by the intensive manipulations necessary to sample each of the 8 to 10 vials per time point). All incubations in a particular experiment were initiated within 5 min to maintain the viability of symbionts once they were separated from the host. Longer incubations in the presence of sulfide and oxygen demonstrated a progressive decrease in carbon fixation over time, presumably due to spontaneous sulfide oxidation and dwindling symbiont viability (data not shown). Symbiont carbon fixation (in millimoles) was calculated as [DIC] × Vol_inc × DPM_sample × Vol_ia × DPM_ia⁻¹ × Vol_sample⁻¹, where [DIC] is the concentration of dissolved inorganic carbon (in millimoles per liter), Vol_inc, Vol_ia, and Vol_sample are the volumes of the incubation, the initial activity subsample, and time course samples (in liters), respectively, and DPM_sample and DPM_ia are the DPM measured for a time point sample and the initial activity sample, respectively. Carbon fixation rates were calculated by regressing carbon fixed against time (0, 5, and 10 min). These rates were consistently quite high, exceeding those measured for vestimentiferan symbionts (36).

Reduced sulfur compounds.

To determine which concentrations of reduced inorganic sulfur compounds would stimulate symbiont carbon fixation, stock solutions of Na₂S or Na₂S₂O₃ were added to the incubation vials immediately prior to symbiont addition, to give final concentrations from 0 to 1 mM. The DIC concentration in each vial was brought to 5 mM by injecting 10 μl of a stock NaHCO₃ solution.

Oxygen and nitrate.

To determine whether the symbionts were capable of using either oxygen or nitrate as an electron acceptor, carbon fixation was measured as described above, except that portions of BASW were bubbled with a variety of gas mixtures before being injected into autosampler vials that had been flushed with the same gas mix used to sparge the BASW. Ten microliters of stock NaHCO₃ solution was injected into each vial to bring the final [HCO₃⁻] to 5 mM. Immediately before the symbionts were injected into the vials, a sulfide stock solution was added to generate a final concentration of 0.5 mM. For determining the optimal concentration of oxygen, BASW was sparged with mixtures of CO₂-free air and N₂, or pure oxygen, that resulted in oxygen concentrations ranging from 0 to 1 mM. To maintain headspace and dissolved concentrations as subsamples were removed, the vial septa were pierced with needles fitted to balloons as described above, but filled with the appropriate gas mix. For incubations in the presence of nitrate, BASW was sparged with pure N₂. Various amounts of KNO₃ stock solution were injected into five of the vials with N₂ headspace to generate nitrate concentrations ranging from 0 to 5 mM.

pH optimum.

To determine the optimum pH for symbiont autotrophy, carbon fixation rates were measured as described above at a range of pH values. To generate these pH values, ASW in a flask was brought to pH 2 by addition of HCl and then sparged with CO₂-free air:N₂ (1:1) for 15 min. Sodium HEPES was added to a final concentration of 50 mM (which alkalinized the solution to a pH of ∼10.5), and the pH was adjusted to 10 by addition of HCl. An 890-μl subsample was removed and injected into a sealed vial that had been flushed with N₂:CO₂-free air (1:1). HCl was added in increments to the buffer remaining in the flask, and 890-μl subsamples were removed to generate a series of vials containing BASW with pH values ranging from 10 to 6. A 50-μl portion of ¹⁴C-labeled NaHCO₃⁻ in a 0.5 M stock DIC solution was injected into the vials and allowed to equilibrate for ∼10 min. To begin the assay, 50 μl of stock sulfide solution and 10 μl of symbionts (in ASW instead of BASW) were added to each vial. The final concentration of DIC in these incubations was very high (25 mM) to ensure that the concentration of CO₂ was saturating and would not influence the rate of symbiont carbon fixation. Adding this high concentration of bicarbonate, however, also altered the pH for incubations with pH values substantially more acidic or basic than the pK_a of HEPES (∼7.6). To account for this, the pH was measured for 8.9-ml portions of BASW (pH 10 to 6, as above) before and after addition of 0.5 ml of 0.5 M NaHCO₃⁻. For BASW with pH values from 7 to 7.75, the change was less than 0.05 pH unit. For BASW with pH values outside of this range, the pH values changed more substantially, and the pH reported for each incubation reflects the value after bicarbonate addition.

CO₂ and HCO₃⁻ use.

To determine whether the symbionts were utilizing extracellular CO₂ or HCO₃⁻, experiments were conducted in which the HCO₃⁻ was held constant at a range of CO₂ concentrations. Two independent experiments were conducted (i.e., with different preparations of symbionts) in which the HCO₃⁻ concentration was held at three constant levels while the CO₂ concentration was altered by modifying the pH (Table 1). In the first experiment, BASW (pH 7.5 and 7.75) was supplemented with DIC to final concentrations of 0.25, 0.7, and 2 mM. Due to the increase in [CO₂] at pH 7.5 and the increase in [CO₃²⁻] at pH 7.75, the [HCO₃⁻] at pH 7.5 and 7.75 are approximately the same. In the second experiment, three pH values (7.5, 7.75, and 8) were used with three [DIC] (0.52, 1.56, and 4.68 mM) to expand the range of [CO₂] tested. [CO₂] and [HCO₃⁻] were calculated from [DIC] and National Buffer Standards pH values with dissociation constants from reference 28.

TABLE 1.

CO₂ and HCO₃⁻ concentrations of the incubations used to determine the form of DIC utilized by Solemya velum symbionts^a

Exptb and DIC concn (mM)	pH	Concn (mM) of:
		HCO3−	CO2
Expt 1
0.25	7.75	0.24	0.004
	7.50		0.008
0.7	7.75	0.66	0.013
	7.50		0.022
2	7.75	1.89	0.036
	7.50		0.064
Expt 2
0.52	8.00	0.48	0.005
	7.75		0.009
	7.50		0.017
1.56	8.00	1.46	0.015
	7.75		0.028
	7.50		0.050
4.58	8.00	4.39	0.046
	7.75		0.084

^aFor these experiments, the [CO₂] at a particular [DIC] was adjusted by changing the pH. At the range of pH values used, the changes in [HCO₃⁻] were minor.

^bResults from two different symbiont preparations are presented here.

In order to estimate the half-saturation constant () and maximal velocity of symbiont carbon fixation with respect to extracellular [CO₂], incubations were conducted at a range of [DIC] at pH 7.75. BASW was prepared by sparging ASW (pH 2) with CO₂-free air:N₂ (1:1), adding sodium HEPES to a final concentration of 50 mM, and adjusting the pH to 7.75 with HCl. After portions of BASW were injected into sealed autosampler vials, DIC was brought to concentrations ranging from 0.018 to 9.8 mM by injecting stock NaHCO₃ solution and [¹⁴C]DIC into the vials. The average half-saturation constant and maximal velocity (± standard deviation [SD]) for carbon fixation from three independent experiments were calculated by the direct linear plot method (15).

RESULTS

Reduced sulfur compounds.

Carbon fixation by S. velum symbionts was stimulated substantially more by sulfide than by thiosulfate (Fig. 1), with the highest rates (61 to 69 μmol min⁻¹ g of protein⁻¹) measured at 0.1 and 0.5 mM sulfide, and was inhibited at 1 mM sulfide. For thiosulfate, rates increased over the full range of concentrations tested (0 to 1 mM), with a maximal value of 18 μmol min⁻¹ g of protein⁻¹. Measured rates demonstrated more variability with sulfide than with thiosulfate, likely due to changes in sulfide and oxygen concentrations during these incubations resulting from the rapid chemical oxidation of sulfide (38).

FIG. 1.

Influence of sulfide and thiosulfate concentrations on Solemya velum symbiont carbon fixation rates in the presence of 0.11 mM oxygen and 5 mM DIC at pH 7.3. Error bars, SDs of carbon fixation rates (3 time points per rate).

Oxygen and nitrate.

Carbon fixation rates were highest in the presence of oxygen concentrations ranging from 44 to 220 μM and were inhibited at 1 mM oxygen when 0.5 mM sulfide was provided as the electron donor and nitrate was absent (Fig. 2). The presence of nitrate did not stimulate symbiont carbon fixation above background at any concentration tested and was inhibitory at 5 mM.

FIG. 2.

Effects of oxygen and nitrate concentrations on Solemya velum symbiont carbon fixation rates in the presence of 0.5 mM sulfide and 5 mM DIC at pH 7.3. Error bars, SDs of carbon fixation rates (3 time points per rate).

pH optimum.

The most rapid rates of symbiont carbon fixation were measured between pH 7.5 and 8.4 when sulfide (0.5 mM) and oxygen (0.11 mM) were present (Fig. 3).

FIG. 3.

Effect of pH on Solemya velum symbiont carbon fixation rates in the presence of 0.5 mM sulfide, 0.11 mM oxygen, and 25 mM DIC. Error bars, SDs of carbon fixation rates (3 time points per rate).

CO₂ and HCO₃⁻ use.

Symbiont carbon fixation rates correlated with the concentration of CO₂ and became saturated at high [CO₂] when [HCO₃⁻] was held constant and [CO₂] was increased by lowering the pH of BASW (Table 1; Fig. 4). For example, in experiment 2, when [HCO₃⁻] was 0.49 mM and the [CO₂] was raised from 5 μM to 17 μM by dropping the pH from 8 to 7.5 (Table 1), symbiont carbon fixation increased from 16 to 25 μmol min⁻¹ mg of protein⁻¹ (Fig. 4). This correlation of symbiont carbon fixation rates with the concentration of CO₂, and not with that of HCO₃⁻, is consistent with symbiont reliance on extracellular CO₂.

FIG. 4.

Carbon fixation by Solemya velum symbionts as a function of [CO₂] in the presence of constant HCO₃⁻, 0.5 mM sulfide, and 0.11 mM oxygen. Panels A and B show results of two independent experiments, for each of which three different [HCO₃⁻] were used. The [CO₂] at a given [HCO₃⁻] was varied as a function of pH. (A) pHs of 7.5 and 7.75 were used. (B) pHs of 7.5, 7.75, and 8 were used. For the [HCO₃⁻] of 4.41 mM, the third incubation at pH 7.5 has been omitted because the CO₂ visibly outgassed during the experiment. Error bars, SDs of carbon fixation rates (3 time points per rate).

From a separate series of incubations at a constant pH of 7.75 and a range of DIC concentrations, symbiont carbon fixation rates could be modeled with rectangular hyperbolae (Fig. 5). The average half-saturation constant was 28 ± 3 μM, and the average maximal velocity was 50.8 ± 7.1 μmol min⁻¹ g of protein⁻¹ (Fig. 5).

FIG. 5.

Solemya velum symbiont carbon fixation versus [CO₂] at 0.5 mM sulfide and 0.11 mM oxygen at pH 7.75. Results from three independent experiments are shown with the rectangular hyperbolae derived from the data (, 28 ± 3 μM; V_max, 50.8 ± 7.1 μmol min⁻¹ g of protein⁻¹). [CO₂] was calculated from [DIC] and National Buffer Standards pH values with dissociation constants from reference 28. Error bars, SDs of carbon fixation rates (3 time points per rate).

DISCUSSION

Despite the difference in their habitats, S. velum symbionts have several parallels with symbionts from deep-sea hydrothermal vent symbioses. They are similar to symbionts from the hydrothermal vent mussel Bathymodiolus thermophilus and the hydrothermal vent clam Calyptogena magnifica in their ability to use both thiosulfate and sulfide as electron donors (8, 30), which explains why carbon fixation by intact S. velum gills is stimulated by both substrates (5). In contrast, symbionts from the vent tubeworm Riftia pachyptila can use sulfide, but not thiosulfate, as an electron donor (18). Thiosulfate utilization may be an adaptation to fluctuations in sulfide availability. If S. velum, like its congener Solemya reidi, has mitochondria capable of oxidizing sulfide to thiosulfate (1, 31, 43), it could “stockpile” a supply of reduced inorganic sulfur in a nontoxic form in its hemolymph, as C. magnifica does (8). Because symbiont carbon fixation is inhibited by 1 mM sulfide, under circumstances where environmental sulfide concentrations are extraordinarily high, both the host and symbionts would benefit from the host converting this toxic substrate to thiosulfate. Alternatively, thiosulfate may be present in the interstitial water of S. velum's sedimented habitat, since thiosulfate is a major by-product of sulfide oxidation in sediments (16). The relative importance of environmental versus host-generated thiosulfate remains to be elucidated for this symbiosis.

The higher rates of carbon fixation observed when sulfide was provided as the electron donor rather than thiosulfate highlights the importance of this form of reduced sulfur. The clam can acquire this toxic dissolved gas directly from the environment and deliver it to the symbionts using gill hemoglobins as the carriers (13). Sulfide has the added feature that in protonated form it can diffuse across host and symbiont membranes without the assistance of a transporter. Alternatively, the host may carry sulfide to the symbionts by using the amino acid taurine as a carrier. Upon exposure to sulfide, S. velum accumulates thiotaurine in its gills (23). When sulfide concentrations decrease, this sulfide may be released from the thiotaurine for use by the symbionts (R. Lee, personal communication).

The reliance of S. velum symbionts on oxygen, and not nitrate, to support carbon fixation is interesting, given that it necessitates adaptations to meet both host and symbiont oxygen demands. Intracellular hemoglobins that bind oxygen and sulfide are present in S. velum gills and facilitate the delivery of both substrates to the symbionts by maintaining a large total (bound plus free) pool of these substrates near the symbionts (13). The extraordinarily high mantle-flushing rates observed for this bivalve, the fastest measured for a bivalve of its size (C. M. Cavanaugh, unpublished data), may also be necessary to deliver oxygen at a rate to match the metabolic demands of both symbiont and host.

The observed optimum for symbiont carbon fixation at pH 8 and the sensitivity to pH values more acidic and basic than this may make it necessary for the host clam to regulate gill pH values. The vestimentiferan tubeworms R. pachyptila and Lamellibrachia luymesi demonstrate high rates of proton extrusion, presumably to prevent the sulfuric acid produced by symbiont sulfide oxidation from acidifying their tissues (20), and have elevated levels of proton-ATPases in their major gas exchange organs (21). Perhaps S. velum gill epithelial cells also have high proton-ATPase activities, which would prevent gill acidification.

S. velum symbionts rely on CO₂ and not HCO₃⁻. This preference for the uncharged form of dissolved inorganic carbon is consistent with the symbionts' intracellular lifestyle. Reliance on HCO₃⁻ would require that every membrane between the environmental source of DIC and the symbionts (which includes intercalary cells, the bacteriocyte membrane, the perisymbiont vacuole, and the symbiont cell envelope itself) would be studded with protein facilitators for HCO₃⁻ entry. Equipping every membrane in this “Russian matryoshka doll” of a system with bicarbonate transporters would be energetically expensive. Reliance on CO₂ and host adaptations may be a simpler solution. Indeed, carbonic anhydrase activity is present in S. velum and is elevated in the symbiont-containing gills compared to the symbiont-free feet (K. M. Scott, unpublished data), where, as in other chemoautotrophic symbioses, it may function to facilitate the diffusion of CO₂ through the bacteriocytes to the symbiont cells (25).

The  for the S. velum symbiont RubisCO (40 μM [39]) is larger than the half-saturation constant for the intact symbionts (28 μM). This may reflect differences in the ribulose 1,5-bisphosphate (RuBP) supply. The RubisCO assays (39) were conducted in the presence of a large excess of RuBP (0.4 mM). For the carbon fixation assays reported here, the symbionts were generating RuBP via the Calvin-Benson cycle. At high [CO₂], intracellular RuBP concentrations may limit the rate of carbon fixation, which would result in an apparent half-saturation constant for cellular carbon fixation that is smaller than the  for RubisCO.

The apparent  for S. velum symbionts (28 μM) is lower than CO₂ concentrations measured from sediments where S. velum was collected (up to 200 μM [39]). Provided boundary layers and other diffusive distances between the symbionts and their environmental source of CO₂ are small enough to prevent the intracellular CO₂ concentration from being substantially lower than the sediment CO₂ concentration, the rate of symbiont carbon fixation should not be limited by CO₂ availability. A carbon flux model, derived to estimate the sizes of the CO₂ concentration gradients between the environment and the symbiont cytosol for R. pachyptila tubeworms (35), allows similar estimates to be made for S. velum. This model consists of a mass balance equation to describe the influences of symbiont carbon fixation and host respiration on the isotope ratio of carbon dioxide within the symbiont cells: R_b = {R_e[CO_2(e)]/[CO_2(s)]}/{α_R + α_D([CO_2(e)]/[CO_2(s)] − 1)}, where R_b is the stable carbon isotope ratio (¹³C/¹²C) of S. velum biomass, R_e is the isotope ratio of environmental CO₂, [CO_2(e)] is the environmental CO₂ concentration, [CO_2(s)] is the concentration of CO₂ within the symbiont cells, α_R is the kinetic isotope effect for carbon fixation by S. velum symbiont RubisCO, and α_D is the kinetic isotope effect for CO₂ diffusion. Using published values for the isotope ratios of S. velum biomass (0.01088 to 0.01085 [10]) and environmental CO₂ (0.01113 to 0.01108 [39]), as well as for the kinetic isotope effects for S. velum symbiont RubisCO (1.0245 [39]) and diffusion (1.0007 [32]), the calculated symbiont intracellular CO₂ concentration is at least 70% of that present in the bivalve's environment. CO₂ gradients between the environment and the symbiont cells are likely to be mitigated by rapid flushing of the mantle cavity and gill carbonic anhydrase. These behavioral and biochemical adaptations may also explain why the symbionts appear not to have bicarbonate transporters or extraordinarily high affinities for CO₂.

These experiments have illuminated the necessity for adaptations by the host bivalves to support the metabolism of their sulfur-oxidizing chemoautotrophic symbionts. S. velum is faced with the daunting challenges of maintaining an intracellular pH that sustains rapid symbiont carbon fixation rates, supplying sufficient oxygen to meet the demands of host heterotrophy and symbiont autotrophy, and preventing sizeable CO₂ gradients from forming and hindering symbiont carbon fixation. Indeed, given these challenges, it is truly remarkable how ubiquitous these chemoautotroph-invertebrate symbioses are in oceanic reducing habitats.