Physiological Proteomics of the Uncultured Endosymbiont of Riftia pachyptila

See allHide authors and affiliations

Science  12 Jan 2007:
Vol. 315, Issue 5809, pp. 247-250
DOI: 10.1126/science.1132913


The bacterial endosymbiont of the deep-sea tube worm Riftia pachyptila has never been successfully cultivated outside its host. In the absence of cultivation data, we have taken a proteomic approach based on the metagenome sequence to study the metabolism of this peculiar microorganism in detail. As one result, we found that three major sulfide oxidation proteins constitute ∼12% of the total cytosolic proteome, which highlights the essential role of these enzymes for the symbiont's energy metabolism. Unexpectedly, the symbiont uses the reductive tricarboxylic acid cycle in addition to the previously identified Calvin cycle for CO2 fixation.

Riftia pachyptila inhabits deep-sea hydrothermal vent areas along mid-ocean ridges in the East Pacific (1). Instead of containing a digestive system, the worm'scoelomic cavity is densely populated by a single species of sulfide-oxidizing gamma-proteobacteria that provide for their host's carbon and energy supply by fixing CO2 from the surrounding water (24). Microbial chemosynthesis is sustained by the presence of H2S originating from reduced hydrothermal fluids and oxygen in the seawater (5). Compared with free-living sulfur oxidizers, the symbionts benefit from high nutrient concentrations within the worm's body (6), which allow for a high metabolic activity. The microbially produced carbon compounds are transferred to the host, making R. pachyptila one of the fastest-growing marine invertebrates known (7).

We took a functional genomics approach to analyze the proteome and hence to derive information on the physiology of the uncultured Riftia symbionts. The sequencing of the symbiont genome was conducted by a metagenome approach. Bacterial symbionts were isolated and separated from the host tissue, the so-called trophosome (8). Using one- and two-dimensional (2D) gel electrophoresis, we established reference maps (master gels) of the soluble intracellular and the membrane-associated bacterial proteome, on which more than 220 identified proteins have been registered so far (Fig. 1 and fig. S1). Our procedures were reproducible and showed the presence of a single endosymbiont species (9) with no indication of other bacterial or host tissue contaminants. The proteomic approach provided evidence regarding which of the predicted genes are expressed under natural growth conditions. Levels of protein synthesis reflect the relative proportion of translational capacity that is invested in the individual metabolic paths, and we can deduce the relevance of particular proteins for the symbionts' physiology. On the basis of our proteome data, we were thus able to deduce major pathways of the symbionts' metabolism, e.g., the sulfide oxidation pathway (fig. S2) and the reverse tricarboxylic acid (TCA) cycle (fig. S3), as well as to make inferences about the symbionts' response to oxidative stress.

Fig. 1.

Reference map of the Riftia pachyptila endosymbionts' intracellular proteome. Identified symbiont proteins are indicated. Protein functions are listed in table S1.

Although great advances have been made in our understanding of the sulfur oxidation pathways in a variety of sulfur oxidizers in recent years, this pathway remains incompletely characterized in Riftia symbionts. In this proteome study, the enzymes DsrA (dissimilatory sulfite reductase), AprA/AprB (adenosine phosphosulfate reductase), and SopT [adenosine triphosphate (ATP) sulfurylase] were identified, which suggested the reactions displayed in fig. S2: H2S is oxidized to SO 2–3 in a six-electron step; this is subsequently converted to adenosine phosphosulfate (APS), with consumption of adenosine monophosphate (AMP). The final oxidation step yields ATP by substrate-level phosphorylation and simultaneously creates the end product, SO 2–4. Although these enzymes were originally described as part of the reverse pathway in sulfate-reducing bacteria (as reflected in their names), it has long been proposed that they can function just as well in the opposite direction (10). Quantitative analyses of the respective protein spots in our study revealed that the three major sulfide oxidation proteins DsrA, AprA/AprB, and SopT constitute more than 12% of the total cytosolic symbiont proteome in a pH range of 4 to 7. This highlights the essential role of these enzymes for the symbiont's energy metabolism.

Using our intracellular protein master gel, we also examined the Riftia symbiont's carbon metabolism. We identified most of the enzymes involved in the energy-generating TCA cycle. Moreover, we found the enzyme 2-oxoglutarate: ferredoxin oxidoreductase β subunit (KorB), several copies of a pyruvate:ferredoxin oxidoreductase (subunits Por1, Por2, and PorA/PorG), a putative fumarate reductase (subunits SdhA/SdhC), and a protein highly similar to a citryl–coenzyme A (CoA) synthetase subunit (CcsB). These enzymes could run the TCA cycle in the reductive direction (fig. S3). This cycle represents an alternative CO2 fixation mechanism that requires less energy than the Calvin cycle for each three-carbon unit formed. Considering the great abundance of these four key enzymes on the protein gels (see relative spot volumes in table S1), we suggest that the reductive TCA cycle is not only a possibility, but a very important feature of the Riftia symbiont's carbon metabolism.

To verify the symbiont'scapabilityforusingthe reductive TCA cycle for CO2 fixation, we tested cell extracts of isolated bacteria for activity of the above-mentioned key enzymes: The assays clearly revealed specific enzyme activities in all four cases that are higher than those previously reported for ribulose-1,5-bisphosphate carboxylase-oxygenase (RuBisCO) (11) (table S2).

The results of our proteome analysis, in combination with the measured enzyme activities, provide strong evidence that the Riftia symbionts, which have been considered a prime example for chemolithoautotrophic carbon fixation via the Calvin cycle, use, at least partly, the reductive TCA cycle for autotrophic carbon fixation as well. This might explain the long-standing dilemma that the stable carbon isotopic composition of the Riftia symbiont is substantially heavier (Δ13Cof –9 to –16‰) than would be expected by the use of the Calvin cycle alone (12). The observed isotopic variation most likely results from varying contributions by the reductive TCA cycle and the Calvin cycle. It is also interesting to note that, concordant with previous findings (12), RuBisCO constitutes only ∼1% of the Riftia symbiont'stotal protein on our gels. This is a small percentage compared with other bacteria that use only the Calvin cycle for CO2 fixation, where RuBisCO is usually the major protein and can account for 4 to 50% of the total soluble protein (13). The apparent discrepancy between the low concentration and the relatively high activity of RuBisCO in the trophosome (4) and the possible occurrence of two carbon fixation pathways in one organism certainly warrant further investigations.

Riftia symbionts can store elemental sulfur in their periplasm if high concentrations of H2S are available (14); these stores result in a light green, almost yellow trophosome. The tissue appears dark green or black (15) when sulfur is limiting. To analyze the symbionts' potential responses to changing environmental conditions, we compared bacterial protein patterns from naturally occurring sulfur-rich and from sulfur-depleted trophosome tissues. Under high-sulfide conditions, the resulting 2D gels revealed a distinctly higher spot intensity for enzymes involved in sulfide oxidation and for RuBisCO, compared with low-sulfur conditions (Table 1, Fig. 2). Spot intensities of the sulfide oxidation enzymes AprA and DsrA were about eight and four times, respectively, those under low-sulfur conditions. This indicates that Riftia symbionts are capable of adjusting the production of enzymes needed for energy metabolism to the prevailing environmental conditions. In bacteria from tissue with little or no stored sulfur (dark trophosomes), the protein spot volume of the putative fumarate reductase subunit SdhA was about 24 times as high as, and the spot volume of the putative pyruvate:ferredoxin oxidoreductase Por1 was about 9 times as high as that in sulfur-rich, light trophosome (Fig. 2, Table 1). Several other enzymes involved in the reductive TCA cycle, including two other pyruvate:ferredoxin oxidoreductase subunits, PorG and PorA, and the 2-oxoglutarate synthase subunit KorB, could also be detected with elevated spot volumes in samples from sulfur-depleted trophosomes. This suggests that the Riftia symbiont might be capable of adapting to a temporary low-energy situation by adjusting its way of carbon fixation: The use of the Calvin cycle might be reduced in favor of the up-regulation of the energetically more favorable reductive TCA cycle. Because Riftia symbionts have also been shown to produce glycogen as a carbon storage compound (16), it might even be speculated that—for example, under long-lasting or severe low-energy conditions—their metabolism switches from an autotrophic mode to a heterotrophic mode. In this case, the symbiont might revert to burning carbon reserves through glycolysis and the oxidative TCA cycle to generate energy and precursors for biosynthesis. The symbiont might thus be able to use the TCA cycle in the oxidative and reductive direction, depending on the environmental conditions, which would allow for a high metabolic flexibility. Apparently, the true nature of the Riftia symbiont's metabolic strategies is much more complex than expected. Further detailed investigations are needed to explicitly evaluate and validate these hypotheses.

Fig. 2.

Comparison of protein patterns under high- and low-sulfur conditions. Images of the 2D gels were colored (sulfur-rich, green, and sulfur-depleted tissue, red spots) and overlaid. Selected proteins with distinct variations in their relative spot volumes from one gel to the other are indicated (see SOM for details).

Table 1.

Comparison of Riftia symbiont protein spot intensities under high- and low-sulfur conditions. Spot intensities of selected proteins were compared in 2D gel images from sulfur-rich and sulfur-depleted trophosome samples (see SOM for details). Positive ratios represent the factor by which the respective spot intensity was higher on the gels from sulfur-rich trophosomes. Negative ratios give the factor by which spot intensity was higher under low-sulfur conditions. The listed protein spots are highlighted in Fig. 2.

ProteinFunctionSpot ratio
Proteins with higher spot intensity in sulfur-rich trophosome samples
AprA Adenylylsulfate reductase, α subunit 8.20
CbbM RuBisCO form II, large subunit 4.29
DsrA Dissimilatory sulfite reductase, α subunit 3.55
HdrA2 Heterodisulfide reductase, subunit A 2.26
DsrL Predicted NADPH:acceptor oxidoreductase 1.92
DsrC Dissimilatory sulfite reductase, γ subunit 1.47
AprB Adenylylsulfate reductase, β subunit 1.30
Proteins with higher spot intensity in sulfur-depleted trophosome samples
SdhA Putative fumarate reductase, flavoprotein subunit -23.83
OadA Oxaloacetate decarboxylase, α subunit -15.00
Por1 Pyruvate:flavodoxin/ferredoxin oxidoreductase -8.58
SucC Succinyl CoA ligase, β subunit -6.75
PorG Pyruvate:ferredoxin/flavodoxin oxidoreductase, γ subunit -6.13
KorB 2-Oxoglutarate synthase, β subunit -4.57
AcnA Aconitase A -4.50
Mdh1 Malate dehydrogenase -2.36
SucD Succinyl-CoA synthetase, α subunit -2.12
PorA Pyruvate:flavodoxin/ferredoxin oxidoreductase, α subunit -2.03

Physiological tests revealed that Riftia symbionts do not have a catalase to protect the cells from H2O2 (17). This result is supported by the lack of the respective gene in the metagenome sequence. Because Riftia symbionts do not tolerate high oxygen concentrations (18), we tested their strategy of coping with stress situations caused by hydrogen peroxide. Our experiments revealed a strong induction of the alkyl hydroperoxide reductase AhpC (Fig. 3). This enzyme is present in large amounts in the cytoplasmic protein fraction (Fig. 1) and reduces organic hydroperoxides caused by H2O2. Our results indicate that AhpC plays a crucial role in the resistance of these microaerophilic bacteria to oxidative stress.

Fig. 3.

Comparison of protein patterns in response to oxidative stress. The 2D gels show proteins from samples taken before and 60 min after the addition of 2 mM H2O2 to a suspension of isolated Riftia symbionts. Both images were colored (control, green spots, and stress sample, red spots) and overlaid. The oxidative stress protein AhpC is indicated. Because of the delayed deformylation of the newly synthesized protein during the massive induction of AhpC production, the protein spot shifts to a slightly more acidic pH. The “new” AhpC, produced after the stress with H2O2 is thus visible as red spots right next to the “old” AhpC, visible as green spots.

This study shows that a comparative proteomic view of the Riftia symbionts' cell physiology allows for a complex physiological description without their cultivation. It reaches beyond the mere prediction of putative metabolic functions as coded in the genome sequence.

Supporting Online Material

Materials and Methods

SOM Text

Figs. S1 to S3

Tables S1 and S2

References and Notes

References and Notes

View Abstract

Navigate This Article